ORIGINAL ARTICLE
Prospective, Randomized Ponseti Treatment for Clubfoot:
Orthopaedic Surgeons Versus Physical Therapists
Stephanie N. Chen, MD,* Tyler D. Ragsdale, MD,† Leslie N. Rhodes, DNP, PPCNP-BC,‡§
Lindsey L. Locke, MSN, CPNP-PC,§ Alice Moisan, BSN, RN,∥ and Derek M. Kelly, MD*
Background: Clubfoot is a common congenital foot deformity in
children. The Ponseti method of serial casting has become the
standard of care in clubfoot treatment. Clubfoot casting is per-
formed in many centers by both orthopaedic surgeons and physical
therapists (PTs); however, direct comparison of outcomes and
complications of this treatment between these providers is limited.
This study prospectively compared the outcomes of patients with
clubfoot treated by these 2 groups of specialists.
Methods: Between January 2010 and December 2014, all patients
under the age of 12 months with a diagnosis of clubfoot were
included. Patients were randomized to an orthopaedic surgeon
(MD) group or a PT group for weekly serial casting. Main
outcome measures included the number of casts required to
achieve correction, clinical recurrence of the deformity, and the
need for additional surgical intervention.
Results: One hundred twenty-six infants were included in the
study. Patient demographics and characteristics (sex, race, family
history of clubfoot, laterality, and severity of deformity) were
similar between treatment groups, with the only significant
difference being the mean age of entry into the study (5.2 weeks in
the MD group and 9.2 weeks in the PT group, P = 0.01). Mean
length of follow-up was 2.6 years. The number of casts required
trended to a lower number in the MD group. There was no sig-
nificant difference in the rates of clinical recurrence or additional
surgical intervention between groups.
Conclusions: Ponseti casting for treatment of clubfoot performed
by orthopaedic surgeons and PTs results in equivalent outcomes
without any difference in complications. Although the number of
casts required trended to a lower number in the MD group, this
From the *University of Tennessee Health Science Center-Campbell Clinic
Department of Orthopaedic Surgery and Biomedical Engineering;
‡University of Tennessee Health Science Center, College of Nursing;
§LeBonheur Children’s Hospital, Memphis; ∥ICON PLC, Brentwood,
TN; and †St. Louis University, St. Louis, MO.
D.M.K. reports a financial relationship with Elsevier Publishing and
serves on the board of the Pediatric Orthopaedic Society of North
America.
The authors declare no conflicts of interest.
Reprints: Derek M. Kelly, MD, University of Tennessee Health Science
Center-Campbell Clinic, Department of Orthopaedic Surgery and
Biomedical Engineering, 1211 Union Avenue, Suite 510, Memphis,
TN 38104. E-mail: dkelly@campbellclinic.com.
Supplemental Digital Content is available for this article. Direct URL
citations appear in the printed text and are provided in the HTML
and PDF versions of this article on the journal’s website, www.
pedorthopaedics.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002291
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
likely did not result in any clinical significance, as the difference
in cast number equaled <1 week’s difference in the overall
duration of serial casting.
Level of Evidence: Level I—therapeutic.
Key Words: clubfoot, Ponseti, serial casting
(J Pediatr Orthop 2023;43:e93–e99)
C lubfoot is a common congenital foot deformity in
children. Early diagnosis and treatment improve
outcomes and decrease complications. Over the past 3
decades, the Ponseti method1,2 has become the standard of
care in clubfoot treatment, consisting of serial manipu-
lations and casting, followed in most cases by percuta-
neous tendo-Achilles tenotomy (percTAT) and abduction
bracing. Despite initial correction rates of around 90% in
the literature,1,3,4 recurrences remain common and may
require repeat casting and/or tenotomy, tendon transfer,
or surgical reconstruction.
Treatment of clubfoot with the Ponseti method re-
quires specialized training to evaluate and manage the
disorder. Casting is performed by a variety of providers,
including orthopaedic surgeons, advanced practice pro-
viders, and physical therapists (PTs), who specializes in the
treatment of children and receive training in the Ponseti
method. Multiple studies have been conducted on the
Ponseti method and on the treatment of clubfoot by a
variety of providers and methods of manipulation and
casting,5–8 but we found no direct prospective comparison
of outcomes and complications of this treatment between
orthopaedic surgeons and PTs. This study prospectively
compared outcomes of patients with clubfoot treated by
these 2 groups of specialists.
METHODS
Study Design and Patient Selection
After ethical approval (#09-00496-XP), all patients
from January 2010 through December 2014 under the age of
12 months with a diagnosis of clubfoot, including idiopathic,
teratologic, and neurological, were included. Patients who
were 12 months old at initial presentation, had foot
deformities other than clubfoot, or had prior manipulation
or casting at another institution were excluded.
www.pedorthopaedics.com | e93
Chen et al
Included patients were randomized into 2 treatment
groups: the orthopaedic surgeon (MD) group (a fellow-
ship-trained pediatric orthopaedic surgeon assisted by a
pediatric orthopaedic fellow, orthopaedic resident, or
nurse practitioner; feet manipulated and held by the sur-
geon whereas the fellow/resident/nurse practitioner ap-
plied the cast) or the PT group (3 pediatric PTs with
special training and extensive prior clinical experience in
Ponseti casting. The same therapists served throughout the
entirety of this study). Randomization was by a coin flip
for the first patient; subsequent consented patients were
assigned to alternating providers to maintain a balanced
clinic workflow (Fig. 1). All patients were treated at the
same location and monitored throughout treatment by the
senior author.
Management and Interventions
Both treatment teams agreed on the diagnosis and
scoring of each clubfoot deformity at the initial visit.
Scoring included both Pirani9 and Dimeglio10 scoring
systems (See Supplemental Fig. 1 and 2, Supplemental
Digital Content 1, http://links.lww.com/BPO/A555). After
randomization, patients received weekly serial casting by
the same treatment team until the complete correction or
until a percTAT procedure was performed to correct any
residual equinus. The decision to proceed with percTAT
was agreed upon by both teams and performed by the
surgeon in an operating room under sedation or, in the
rare case of an insensate foot, in a clinic procedure room.
After percTAT, a long leg cast was used for 3 weeks to
allow soft-tissue healing.
FIGURE 1. Consolidated standards of reporting trials (CONSORT) flow diagram.
e94 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
Patients were then treated with a foot abduction
orthosis after correction of deformity through initial serial
casting or percTAT. Brace-wear recommendations were
23 hours/day for 3 months followed by part-time (night-
time and naptime) wear until age 3 years. The Denis-
Browne bar was the predominant brace type used in both
groups, but other foot abduction orthoses were utilized
when different shoe/bar combinations could improve
compliance or comfort. Brace-wear compliance was based
on caregiver-reported wear and provider and orthotist
brace assessment at follow-up visits.
Patients were followed weekly until deformity cor-
rection, or until percTAT was performed, and then 3
weeks after percTAT. All patients had in-brace checks at
1 month and 4 months and then at 1 year of age and every
6 months from ages 1 to 3 years. Each treated foot was
scored using Pirani and Dimeglio scoring systems at in-
brace follow-up visits. Duration of follow-up was the time
until patients reached 3 years of age or additional surgical
intervention if performed.
Patients with unsuccessful deformity correction with
initial casting or with recurrence despite successful initial
casting were treated with repeat casting, tibialis anterior
tendon transfer, limited posterior release, or comprehensive
clubfoot release based on the degree of residual deformity
or degree of recurrence. The senior author decided on
surgical intervention in both groups.
Outcome Measures
Outcome measures included the number of casts
required for correction or until percTAT was performed,
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Ponseti Treatment for Clubfoot

TABLE 1. Baseline Characteristics

Total (126 Patients, MD Group (61 Patients, PT Group (65 Patients, Difference
185 Feet) 88 Feet) 97 Feet) (P)
Mean age at presentation (wk) 7.2 5.2 9.2 0.01
M:F (% M) 82:44 (65) 39:22 (64) 43:22 (66) 0.853
Caucasion:non-Caucasian (% Caucasian) 77:49 (61) 37:24 (61) 40:25 (62) > 0.999
No. (%) with positive family history 40 (32) 20 (33) 20 (31) 0.85
Unilateral:bilateral (% bilateral) 67:59 (53) 34:27 (56) 33:32 (51) 0.743
Mean initial Pirani score 4.0 3.8 4.2 0.208
Mean initial Dimeglio score 13.4 13.1 13.6 0.526
MD indicates medical doctor; PT, physical therapist.

appointment compliance, brace-wear compliance, skin or (83%) of 185 feet]. The rate of percTAT was similar
cast complications, clinical recurrence of deformity between groups (Table 3).
(defined as the increase in Pirani or Dimeglio scores after
the completion of casting and/or percTAT), and surgical Clinical Recurrence
intervention beyond(SBT) percTAT. Thirty-nine patients had increased Pirani scores at
6 months, 51 at 12 months, and 67 at 3 years (Table 4).
Analysis Thirty patients had increased Dimeglio scores at
Groups were compared for differences in baseline 6 months, 46 at 12 months, and 70 at 3 years. There
characteristics or outcomes, including the number of casts were no demographic or patient-related differences
required for correction, rates of clinical recurrence, and SBT. between patients with clinical recurrence and those
The Fisher exact test and Welch t test were used to compare without. At the final follow-up, higher initial Pirani and
outcome measures between groups. A P value < 0.05 was Dimeglio scores and a greater number of casts during
considered significant. Sample size was determined by the initial serial casting were associated with clinical
number of patients who met inclusion criteria during the recurrence (Table 5). Brace-wear noncompliance was
time of the study. associated with increased Dimeglio scores at 6 and
For this study, we chose to include and randomize all 12 months (P = 0.005, P = 0.01).
patients regardless of foot severity or underlying diagnosis
to more closely mimic a typical clubfoot practice. We have Surgical Intervention Beyond PercTAT
included a detailed analysis of the entire patient population. There were no demographic or patient-related dif-
But, we also performed a subgroup analysis of only the ferences between patients requiring SBT at any time point
idiopathic clubfeet, excluding all patients with known ar- and those who did not in either group. At 6 months, there
throgryposis, myelomeningocele, or other neuromuscular were 3 SBTs out of 120 patients. At 12 months, there were
disorder (11 patients). In this subgroup analysis of only 11 SBT out of the remaining 100 patients, and at 3 years,
idiopathic clubfeet, we also found nearly identical com- 25 SBT out of 100 (Table 4). At both 12 months and
parisons in the group demographics, treatment parameters, 3 years, a greater number of casts were applied in patients
and outcomes between those casted by each of the 2 who eventually required SBT (12 mo: 5.8 casts in the SBT
provider groups. vs 4.1 casts in the non-SBT group, P = 0.003; 3 years: 5.3
casts in the SBT vs 3.8 casts in the non-SBT group,
RESULTS
One hundred twenty-six infants (185 feet) diagnosed
TABLE 2. Distribution of Diagnoses
with clubfoot were enrolled in the study. Baseline char-
acteristics of the 61 patients (88 feet) in the MD group and Total (126 MD Group (61 PT Group (65
65 patients (97 feet) in the PT group are summarized in Patients, 185 Patients, 88 Patients, 97
Feet) Feet) Feet)
Table 1. The 2 groups were similar in sex, race, family
history of clubfoot, laterality, and severity of the Idiopathic 115 (165) 57 (81) 58 (84)
deformity. Mean age at entry was 7.2 weeks (range 1 to [No. patients
(feet)]
50.6 wk) with a slight difference in age between the 2 Nonidiopathic [No. patients (feet)]
groups (5.2 wk and 9.2 wk in the MD and PT groups, Arthrogryposis 5 (10) 2 (4) 3 (6)
respectively; P = 0.01). The distribution of idiopathic Myelodysplasia 2 (4) 0 2 (4)
versus nonidiopathic diagnoses is shown in Table 2. The Down 1 (1) 0 1 (1)
mean length of follow-up was 2.6 ± 1.5years (< 3 y syndrome
Sacral agenesis 1 (2) 0 1 (2)
reflected continued follow-up data not collected on feet Other 2 (3) 2 (3) 0
undergoing SBT after the procedure). Twenty-six patients chromosome
(37 feet) were lost at a 3-year final follow-up. An average abnormality
of 4.1 casts were applied during initial serial casting. A MD indicates medical doctor; PT, physical therapist.
percTAT was performed in 106 (84%) of 126 patients [153

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e95

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Chen et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 3. Comparison of Outcome Measures

Total (126 Patients, MD Group (61 Patients, PT Group (65 Patients, Difference
185 Feet), n (%) 88 Feet), n (%) 97 Feet), n (%) (P)
Mean no. casts applied
Right side 4.1 3.8 4.3 0.042
Left side 4.2 3.9 4.5 0.099
No. patients with missed casting appointments 13 (10) 9 (15) 4 (6) 0.146
No. patients with missed in-brace appointments 2 (1.6) 1 (1.6) 1 (1.5) > 0.999
Mean length of follow-up (mo) 30.65 29.91 31.34 0.659
No. patients with skin/cast complications 20 (16) 8 (13) 12 (18) 0.471
No. patients with percTAT: no. feet with percTAT 106 (84):153 (83) 51 (84):70 (80) 55 (85):83 (86)
MD indicates medical doctor; percTAT, percutaneous tendo-Achilles tenotomy; PT, physical therapist.

P < 0.001). Brace-wear noncompliance was higher in
patients requiring SBT at 12 months (P = 0.047). Higher
initial Pirani and Dimeglio scores were associated with
SBT but did not reach consistent significance at all time
points. Clinical recurrence was significantly associated
with SBT at all time points (Table 6).
Orthopaedic Surgeon (MD) Versus Physical
Therapists
The number of casts required to achieve correction
trended lower in the MD group (3.8 casts in the MD
group vs 4.3 in the PT group for right-sided casting,
P = 0.04; 3.9 in the MD group vs 4.5 in the PT group for
left-sided casting, P = 0.09). Casting and in-brace follow-
up appointment compliance, length of follow-up, and skin
or cast complications were equivalent between groups
(Table 3). Clinical recurrence and SBT rates for both
groups are summarized in Table 4. No significant
differences in rates of recurrence or SBT were found
between groups at any time point.
DISCUSSION
Ponseti serial casting is widely accepted as the
preferred treatment for clubfoot worldwide.11,12 Despite
excellent initial correction rates, recurrences may require
repeat casting or additional surgical intervention. Experi-
ence with the Ponseti method and attention to detail in the
manipulation and molding of casts are necessary for desired
results. Brace-wear compliance and parental education are
key factors in the maintenance of deformity correction.13,14
Serial casting is performed around the world by a variety of
specially trained providers. Several studies confirmed that
Ponseti casting by PTs and even nonmedical personnel
yields similar successful outcomes,15–18 but prospective,
direct comparison of outcomes between providers is lim-
ited. This prospective, randomized study found that Ponseti
casting performed by orthopaedic surgeons and PTs
resulted in equivalent outcomes without differences in
complications.
Dr. Ponseti originally described applying 5 to 10
plaster casts worn for 5 to 12 weeks as necessary for ap-
propriate correction of clubfoot deformity.1,2 In the Iowa
series of 157 patients (256 clubfeet), 90% of patients re-
quired 5 or fewer casts to achieve correction.19 In our study,
an average of 4.1 casts were applied during initial serial
casting. Our initial correction rate of 97.5% compared fa-
vorably with reported success rates of 83% to 98%.1–-
4,16,18–21 Our percTAT rate of 84% was similar to Iowa 87%
rate.19 Primary percTAT is an integral part of the Ponseti
method1,2 and was not counted as an additional surgical
procedure.
Clinical recurrence was defined as any increase in
Pirani and/or Dimeglio scores after the conclusion of ini-
tial serial casting and percTAT, in patients who underwent
percTAT. Longer follow-up can lead to higher recurrence
rates,22,23 as the natural history of treated clubfoot has
shown many feet will stiffen over time. To capture
all recurrent deformities, we chose a strict definition of

TABLE 4. Comparison of Clinical Recurrence and Surgical Intervention Beyond Tenotomy (SBT) Rates
Follow-up Total, no. Patients/ MD Group, no. Patients/ PT Group, no. Patients/ Difference
Interval N (%) N (%) N (%) (P)
Recurrence based on Pirani 6 mo 39/120 (33) 18/58 (31) 21/62 (34) 0.846
score
12 mo 51/103 (50) 22/50 (44) 29/53 (55) 0.326
3y 67/106 (63) 29/51 (57) 38/55 (69) 0.229
Recurrence based on 6 mo 30/120 (25) 12/58 (21) 18/62 (29) 0.399
Dimeglio score
12 mo 46/102 (45) 19/50 (38) 27/52 (52) 0.170
3y 70/105 (67) 31/50 (62) 39/55 (71) 0.408
SBT 6 mo 3/120 (2.5) 0/58 (0) 3/62 (5) 0.245
12 mo 11/100 (11) 3/49 (6) 8/51 (16) 0.201
3y 25/100 (25) 10/48 (21) 15/52 (29) 0.489
MD indicates medical doctor; N, total number in group; PT, physical therapist; SBT, surgical intervention beyond tenotomy.

e96 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Ponseti Treatment for Clubfoot

Difference recurrence as any increase in scores beyond the initial

corrected values. Recurrence rates in the literature have

0.396
0.249
0.388

0.178

0.714
0.454
0.408
0.005
0.070

0.010
0.076
0.260
0.013
0.024
(P)
ranged from 11% to 83%.24 Ponseti and Smoley1 reported
a 56% recurrence rate in their 5 to 12-year follow-up. A
recent systematic review of long-term outcomes of the
Ponseti method found recurrence rates of 47% with ad-
ditional surgery required in 79% of recurrences.25 Our
No Recurrence
(35 Patients)

recurrence rate gradually increased over time, approach-

24:11 (69)
25:10 (71)

22:13 (37)
19:16 (54)

ing two-thirds of patients at final follow-up. The relatively

Dimeglio criteria

13 (37)

22 (63)

2 (5.7)
12.26
12.19
3.50
3.67
8.8

3.6
4.0
high recurrence rate in our series was likely due to the
unforgiving criteria of any net increase in Pirani or
Dimeglio scores.
The decision to perform SBT in recurrences is
highly dependent on the degree and specific components
of recurrence, surgeon preference, compliance expect-
ations, and discussions with caregivers. Rates vary
(70 Patients)
Recurrence

widely in the literature and, like recurrences, have been

41:29 (59)
43:27 (61)

35:35 (50)
31:39 (44)
17 (24)

52 (74)

7 (10)
14.32
14.11
4.62
4.22

found to increase with longer follow-ups.22,23 Reviews

6.3

4.4
4.6

have described ranges from 1.4% to 53.3%.23 In our

series, 2.5% of patients underwent SBT at 6 months,
which increased to 11% at 12 months and 25% at 3 years.
Severity of initial deformity can be prognostic for
recurrence and additional surgery. The Pirani9 and
Difference

Dimeglio10 scoring systems are 2 commonly used classi-

0.837

0.333
0.595
0.533

0.503

0.229
0.033
0.005

0.001
0.005
0.012
0.827

0.722
(P)

fications assessing clinical severity and have been validated

0.04

with good interobserver reliability and predictive value.26,27

Hemo et al28 found higher initial Pirani scores and a greater
number of cast changes as the strongest predictors for fu-
ture surgery. Sangiorgio et al29 reported higher Dimeglio
No Recurrence

scores led to a more likely relapse. In our series, higher

(39 Patients)

25:14 (64)

25:14 (36)
27:12 (69)

22:17 (56)

initial Pirani and Dimeglio scores significantly correlated

27 (69)
9 (23)

4 (10)
12.59
11.88
3.71
3.54
Pirani criteria

with clinical recurrence at the final follow-up. Similarly, a

7.7

3.6
3.8

greater number of casts were required to achieve correction

in those feet experiencing recurrence, also indicating the
likelihood of more severe initial deformity. As expected,
clinical recurrence was associated with SBT at all time
points; however, several patients who underwent SBT at
12 months and 3 years did not fall under the category of
(67 Patients)

clinical recurrence based on our definition. We attribute this

Recurrence

33:34 (51)
41:26 (61)
42:25 (63)

29:38 (43)
21 (31)

48 (72)

5 (7.5)

to including nonidiopathic clubfeet, many of whom had

14.16
14.40
4.52
4.32
TABLE 5. Predictive Factors for Recurrence at 3 Years

6.7

4.4
4.7

more severe deformities and higher initial scores, and some

did not have complete correction through serial casting and
percTAT. Brace-wear noncompliance was associated with
both clinical recurrence and the likelihood of requiring
MD indicates medical doctor; PT, physical therapist.

SBT, as has been thoroughly established in numerous

studies.4,13,19,29–32
Caucasion:non-Caucasian (% Caucasian)

Direct comparison of Ponseti casting between MDs

Mean initial Dimeglio score, right feet

and PTs is limited. A retrospective study by Janicki et al33

Mean initial Dimeglio score, left feet
MD group:PT group (% MD group)

No. (%) patients with missed casting

No. (%) with positive family history

Mean initial Pirani score, right feet

found no significant differences in the number of initial

Mean no. casts applied, right side
Mean initial Pirani score, left feet

No. (%) patients with brace-wear

Mean no. casts applied, left side
Unilateral:bilateral (% bilateral)

casts, Achilles tenotomy rate, and initial correction rate in

Mean age at presentation (wk)

Ponseti treatment by surgeons and physiotherapists in

Toronto. Interestingly, they reported lower recurrence
rates and the need for additional surgical intervention in
their physiotherapist group. Their series included 120 pa-
noncompliance

appointments

tients with 171 clubfeet and, although retrospective in

M:F (% M)

nature, was the only comparative study we found in the

English language literature. In our study, clinical re-
currence rates and the need for additional surgical inter-
vention were equivalents between MD and PT groups.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e97

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Chen et al
TABLE 6. Association of Clinical Recurrence With SBT
Follow-up Total, no. Patients/
Interval N (%)
Recurrence based on Pirani 6 mo 39/120 (32.5)
score
12 mo 48/100 (48)
3y 60/100 (60)
Recurrence based on Dimeglio 6 mo 30/120 (25)
score
12 mo 44/100 (44)
3y 64/100 (64)
N indicates total number of group; SBT, surgical intervention beyond tenotomy.
Although the number of casts required for deformity
correction or until percTAT trended to a lower number in
the MD group, this likely did not result in any clinical
significance, as the difference in cast number equaled <1
week’s difference in the overall duration of serial casting.
Our study had some limitations. First, as a single
surgeon, single-institution study, it is possible that a larger
series or addition of other centers could result in significant
differences in outcomes. A post hoc power analysis in-
dicated that a sample size of ~1000 patients was necessary
to find a significant difference based on our observed find-
ings. In addition, patients were enrolled during the senior
author’s early practice and likely represented a portion of
his learning curve. Therefore, surgical rates and types of
surgeries performed may not represent those of more ex-
perienced surgeons. Second, randomization was conducted
by a coin flip for the first patient and then alternated groups
to maintain an equal number of patients in the casting stage
of treatment throughout the study period. This method was
used to provide more efficient care during the casting
process; other means of randomization could result in 2 or
more patients in a row going to the same group. Third,
caregiver-reported wear as well as provider and orthotist
evaluation of the brace at follow-up visits was used to assess
brace-wear compliance. Caregiver-reported rates can be
misleading and significantly higher than actual wear rates
measured by pressure or temperature monitors,34,35 which
were not commonly available at the time of this study.
Although these devices may have provided more objective
and accurate measurements, this likely did not affect
comparative results with the same method of assessing
brace-wear for both groups.
This prospective, randomized comparison of Ponseti
casting by MDs and PTs yielded comparable outcomes and
recurrence rates. As treatment of clubfoot has shifted from
largely surgical to primarily nonoperative using this meth-
od, these reassuring results support the training of PTs or
other providers in the Ponseti method with expectations of
similar high-quality outcomes. This increase in appropri-
ately trained providers could allow for more timely access
to care for patients, improved clinic efficiency, and reduc-
tion of surgeon-led conservative management. Our study
gives confidence to centers worldwide that properly trained
providers can achieve equivalent results for babies with
clubfoot deformity.
e98 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
SBT, no. Patients/ No SBT, no. Patients/ Difference
N (%) N (%) (P)
3/3 (100) 36/117 (31) 0.033
9/11 (82) 39/89 (44) 0.024
22/25 (88) 38/75 (51) < 0.001
3/3 (100) 27/117 (23) 0.014
9/11 (82) 35/89 (39) 0.010
23/25 (92) 41/75 (55) < 0.001
ACKNOWLEDGMENTS
The authors would like to acknowledge David
Heinsch, MD, Leigh Holland Trotti, NP, Shannon Knierim
Pruitt, PT, Lauren McGee, PT, Konstance Hines, PT, Jodi
Beth Haberman, PT, Aristea Wells, DPT-PCS, and
McKinsey Juarez, DPT for their assistance in this study.
REFERENCES
1. Ponseti IV, Smoley EN. The classic: congenital club foot: the results
of treatment. 1963. Clin Orthop Relat Res. 2009;467:1133–1145.
2. Ponseti IV. Treatment of congenital club foot. J Bone Joint Surg Am.
1992;74:448–454.
3. Laaveg SJ, Ponseti IV. Long-term results of treatment of congenital
club foot. J Bone Joint Surg Am. 1980;62:23–31.
4. Jowett CR, Morcuende JA, Ramachandran M. Management of
congenital talipes equinovarus using the Ponseti method: a system-
atic review. J Bone Joint Surg Br. 2011;93:1160–1164.
5. Matos MA, de Oliveira LA. Comparison between Ponseti’s and
Kite’s clubfoot treatment methods: a meta-analysis. J Foot Ankle
Surg. 2010;49:395–397.
6. de La Taille E, Sales de Gauzy J, Gaubert Noirot M. Idiopathic
clubfoot treatment and heterogeneity of current therapeutic strat-
egies: the Ponseti method versus the French functional method (a
systematic review. Arch Pediatr. 2021;28:422–428.
7. Bina S, Pacey V, Barnes EH, et al. Interventions for congenital
talipes equinovarus (clubfoot). Cochrane Database Syst Rev. 2020;5:
CD008602.
8. Aulie VS, Halvorsen VB, Brox JI. Motor abilities in 182 children
treated for idiopathic clubfoot: a comparison between the traditional
and the Ponseti method and controls. J Child Orthop. 2018;12:383–389.
9. Pirani S, Outerbridge H, Sawatzky B, et al. A reliable method of
clinically evaluating a virgin clubfoot evaluation. 21st SICOT
congress; 1999.
10. Dimeglio A, Bensahel H, Souchet P, et al. Classification of clubfoot.
J Pediatr Orthop B. 1995;4:129–136.
11. Zionts LE, Sangiorgio SN, Ebramzadeh E, et al. The current
management of idiopathic clubfoot revisited: results of a survey of
the POSNA membership. J Pediatr Orthop. 2012;32:515–520.
12. Radler C. The Ponseti method for the treatment of congenital club
foot: review of the current literature and treatment recommenda-
tions. Int Orthop. 2013;37:1747–1753.
13. Dobbs MB, Rudzki JR, Purcell DB, et al. Factors predictive of
outcome after use of the Ponseti method for the treatment of
idiopathic clubfeet. J Bone Joint Surg Am. 2004;86:22–27.
14. Zhao D, Li H, Zhao L, et al. Prognosticating factors of relapse in
clubfoot management by Ponseti method. J Pediatr Orthop. 2018;38:
514–520.
15. Dunkley M, Gelfer Y, Jackson D, et al. Mid-term results of a
physiotherapist-led Ponseti service for the management of non-
idiopathic and idiopathic clubfoot. J Child Orthop. 2015;9:183–189.
16. Shack N, Eastwood DM. Early results of a physiotherapist-delivered
Ponseti service for the management of idiopathic congenital talipes
equinovarus foot deformity. J Bone Joint Surg Br. 2006;88:1085–1089.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
17. Kampa R, Binks K, Dunkley M, et al. Multidisciplinary manage-
ment of clubfeet using the Ponseti method in a district general
hospital setting. J Child Orthop. 2008;2:463–467.
18. Tindall AJ, Steinlechner CW, Lavy CB, et al. Results of manipulation
of idiopathic clubfoot deformity in Malawi by orthopaedic clinical
officers using the Ponseti method: a realistic alternative for the
developing world? J Pediatr Orthop. 2005;25:627–629.
19. Morcuende JA, Dolan LA, Dietz FR, et al. Radical reduction in the
rate of extensive corrective surgery for clubfoot using the Ponseti
method. Pediatrics. 2004;113:376–380.
20. Herzenberg JE, Radler C, Bor N. Ponseti versus traditional methods of
casting for idiopathic clubfoot. J Pediatr Orthop. 2002;22:517–521.
21. Ganesan B, Luximon A, Al-Jumaily A, et al. Ponseti method in the
management of clubfoot under 2 years of age: a systematic review.
PLoS One. 2017;12:e0178299.
22. Thomas HM, Sangiorgio SN, Ebramzadeh E, et al. Relapse rates in
patients with clubfoot treated using the Ponseti method increase with
time: a systematic review. JBJS Rev. 2019;7:e6.
23. Gelfer Y, Wientroub S, Hughes K, et al. Congenital talipes
equinovarus: a systematic review of relapse as a primary outcome
of the Ponseti method. Bone Joint J. 2019;101-B:639–645.
24. Wallace J, White H, Eastman J, et al. Reoccurrence rate in Ponseti
treated clubfeet: a meta-regression. Foot (Edinb). 2019;40:59–63.
25. Rastogi A, Agarwal A. Long-term outcomes of the Ponseti method
for treatment of clubfoot: a systematic review. Int Orthop. 2021;45:
2599–2608.
26. Flynn JM, Donohoe M, Mackenzie WG. An independent assessment
of two clubfoot-classification systems. J Pediatr Orthop. 1998;18:
323–327.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Ponseti Treatment for Clubfoot
27. Dyer PJ, Davis N. The role of the Pirani scoring system in the
management of club foot by the Ponseti method. J Bone Joint Surg
Br. 2006;88:1082–1084.
28. Hemo Y, Yavor A, Kalish M, et al. Ponseti treated idiopathic
clubfoot—outcome predictive factors in the test of time: analysis of
500 feet followed for 5 to 20 years. J Child Orthop. 2021;15:426–432.
29. Sangiorgio SN, Ebramzadeh E, Morgan RD, et al. The timing and
relevance of relapsed deformity in patients with idiopathic clubfoot.
J Am Acad Orthop Surg. 2017;25:536–545.
30. Kuzma AL, Talwalkar VR, Muchow RD, et al. Brace yourselves:
outcomes of Ponseti casting and foot abduction orthosis bracing in
idiopathic congenital talipes equinovarus. J Pediatr Orthop. 2020;40:
e25–e29.
31. Alves C. Bracing in clubfoot: do we know enough? J Child Orthop.
2019;13:258–264.
32. Goldstein RY, Seehausen DA, Chu A, et al. Predicting the need for
surgical intervention in patients with idiopathic clubfoot. J Pediatr
Orthop. 2015;35:395–402.
33. Janicki JA, Narayanan UG, Harvey BJ, et al. Comparison of
surgeon and physiotherapist-directed Ponseti treatment of idiopathic
clubfoot. J Bone Joint Surg Am. 2009;91:1101–1108.
34. Morgenstein A, Davis R, Talwalkar V, et al. A randomized clinical
trial comparing reported and measured wear rates in clubfoot
bracing using a novel pressure sensor. J Pediatr Orthop. 2015;35:
185–191.
35. Richards BS, Faulks S, Felton K, et al. Objective measurement of
brace wear in successfully Ponseti-treated clubfeet: pattern of
decreasing use in the first 2 years. J Am Acad Orthop Surg.
2020;28:383–387.
www.pedorthopaedics.com | e99
 ORIGINAL ARTICLE

Efficacy of 2 Nonpharmaceutical (Non-nutritive Sucking

and Human Care Contact) Pain Relief Measures for
Idiopathic Clubfoot Casted Using Ponseti Technique
Ankur Upadhyay, MS,* Sitanshu Barik, MCh,† Anil Agarwal, MS,* and Yogesh Patel, MS*

(J Pediatr Orthop 2023;43:e100–e105)

Purpose: This study aimed at exploring the pain and physio-
logical responses exhibited during Ponseti manipulation and
casting in clubfoot infants. In addition, we compared the efficacy
of 2 nonpharmaceutical techniques (non-nutritive sucking and
human care contact) for tackling these responses.
P onseti technique is the current standard modality of
treatment for idiopathic clubfoot.1 During this ma-
nipulation and casting, the infants can become fussy and
Methods: The study included children with unilateral and bi-
irritable and may exhibit a pain response.2 The pain can
lateral idiopathic clubfeet between 15 days to 6 months of age.
invoke discomfort as well as adverse physiological con-
For comparisons, children were divided into control group
sequences, such as hypoxemia, bradycardia, and hyper-
without any intervention (group A), non-nutritive sucking group
tension in young children. When such stimuli are repeated,
(group B), and human care contact group (group C). Pain score
they can also sustain potential long-term behavior and
(Neonatal Infant Pain Score), heart rate (HR), and oxygen sat-
neurological consequences.3,4
uration (SpO2) was assessed before, during and 1 minute after
Several quantitative parameters are now available
casting. These measurements were compared using statistical
which can quantify pain and associated physiological,
methods.
biochemical, behavioral, and psychological alterations in
Results: There were 16 children (11 bilateral) in group A, 17 (10
infants.5 Broadly, pain management strategies consist of
bilateral) in group B, and 18 (8 bilateral) in group C. Before
pharmaceutical and nonpharmaceutical interventions.
casting, the baseline parameters (Neonatal Infant Pain Score,
Pharmaceutical interventions typically carry the risk of
HR, and SpO2) of the 3 groups were comparable. Groups B and
adverse drug effects. Nonpharmaceutical interventions act
C had a significant reduction in pain score at casting and in
by enhancing descending inhibitory systems. They de-
postcasting period when compared with group A (P < 0.05).
crease pain perception and may also modulate response
Group B (at casting—mean: 174.1/min, postcasting—mean:
through changes in attention and decreasing appre-
168.2/min) had the lowest HR both during and after cast appli-
hension. Non-nutritive pacifiers are one such agent used in
cation. Group B had the highest SpO2 among all the 3 groups,
common practice common in neonatal care units.6,7 In
both during casting (mean: 95.7%) and after casting (mean:
healthy-term neonates, cuddling with mother has also
97.4%) (P < 0.05).
been shown to reduce both physiological and behavioral
Conclusions: Infants exhibit moderate pain response and altered
pain response.8
physiological responses during and after Ponseti casting. Non-
The child getting irritable during the plaster appli-
nutritive sucking emerged as a better method to lessen these
cation is considered an indirect sign of pain response.2
parameters when compared with the conventional technique and
A previous trial evaluated the use of sucrose and milk as
human care contact.
pain relievers during clubfoot treatment and the same
Level of Evidence: Level II.
were shown to be effective.2 Other projected advantages of
Key Words: clubfoot, Ponseti technique, anxiety, pain using these nonpharmaceutical pain relievers were a
calmer child, better patient-doctor relationship, less anx-
ious parents, and reduced number of casting sessions.2 Our
From the *Department of Pediatric Orthopedics, Chacha Nehru Bal study was specifically designed to further explore the pain
Chikitsalaya, New Delhi, Delhi; and †Department of Orthopedics, and physiological responses exhibited during Ponseti ma-
All India Institute of Medical Sciences, Deoghar, Jharkhand, India. nipulation and casting. In addition, we compared efficacy
Material preparation, data collection, and analysis were performed by
A.U. and A.A. The first draft of the manuscript was written by S.B. of 2 nonpharmaceutical techniques (non-nutritive sucking
The authors declare that no funds, grants, or other support were received and human care contact) for tackling these responses.
during the preparation of this manuscript.
The authors declare no conflicts of interest. METHODS
Reprints: Anil Agarwal, MS, Department of Pediatric Orthopedics,
Chacha Nehru Bal Chikitsalaya, New Delhi 110031, Delhi, India. Study Details
E-mail: anilrachna@gmail.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. The short-term study (March-October 2021) was con-
DOI: 10.1097/BPO.0000000000002299 ducted at a tertiary care pediatric hospital. Ethical approval

e100 | www.pedorthopaedics.com J Pediatr Orthop  Volume 43, Number 2, February 2023

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
for the same was obtained from the Institutional Ethics
Committee of Chacha Nehru Bal Chikitsalaya (No). F.1/
AEC/CNBC/15/02/2021/Protocol no. 92/4256. Written in-
formed consent was obtained from the parents. It included
children with unilateral and bilateral idiopathic clubfeet be-
tween 15 days to 6 months of age. The following children
were excluded: (1) surgically intervened/operated clubfeet, (2)
complex/atypical clubfeet, (3) inability to suck due to any
medical or surgical cause, (4) on any analgesic within 6 hours
of data collection and (5) those missing regular scheduled
weekly Ponseti casting session.
Study Groups
Subjects of idiopathic clubfoot fulfilling above criteria
were enrolled for study. In bilateral feet, right side was casted
first and the one included for observations. For comparisons,
children were divided into control group without any inter-
vention (group A), non-nutritive sucking group (group B),
and human care contact group (group C). The allocation of
child to a group was done through a computer-based random
number generator. Once the child was allotted to a group, he/
she remained in the same group till the end of the cast
treatment. All enrolled children were managed by the
standard Ponseti technique of serial casting using plaster of
Paris followed by percutaneous tendoachilles tenotomy, when
required. During casting, the manipulated foot was held in
maximally corrected position and plaster applied till mid-leg.
After the plaster has set, extension till the root of thigh was
done keeping the knee flexed at 90 degrees. Percutaneous
tendoachilles tenotomy was timed as per the following con-
ditions: a midfoot Pirani score of <1 (with a talar head score
of 0), foot hyperabduction of minimum 60 degrees, and a
hindfoot Pirani of > 1.
For group A, the child was casted by the conven-
tional method practiced in our clinic wherein the child is
laid supine on the couch. For group B, the child was ad-
ditionally provided a fresh sealed non-nutritive sucking
unit before initiation of casting. The unit was discarded
soon after the casting and subsequent observation period
was over. A fresh unit was used for each child at every
session. For group C, the casting was done with the child
secure in caretaker’s lap and arms (usually mother).
Study Methodology
Each cast session was videotaped by the study personnel
before, during, and after casting. Readings were taken at
1 minute before the procedure, just before application of cast
material after manipulation was over and at 1 minute after the
procedure. The child however was kept in clinic and under
observation for a minimum 1 hour postcasting. These were
later reviewed for an objective evaluation of pain using the
Neonatal Infant Pain Score (NIPS) (Table 1).9 Pain scoring
was done by a trained professional not part of the study.
During each casting session, clinically used objective measures
[heart rate (HR) and oxygen saturation (SpO2)] were recorded
with use of a pulse oximeter sensor (Nellcor Bedside SpO2
patient monitoring system, Covidien; Medtronic, Minneapolis,
MN) at the above-described time frames. The sensor was
placed on infant’s right palm and secured with the help of tape.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Pain Relief Measures in Idiopathic Clubfoot
Statistical Analysis
The quantitative variables were expressed as the
mean ± SD. NIPS values before, during, and after the casting
were compared using the Kruskal-Wallis test (intergroup)
and Friedman test (intragroup). The responses for HR and
SpO2 before, during, and after the casting was compared
using analysis of variance [1 way (intergroup) and repeated
measures (intragroup)]. The qualitative data was analyzed by
χ2 test. A P value <0.05 was considered significant.
RESULTS
Participants
Fifty-four children with idiopathic clubfoot were
originally enrolled during the study period who gave
consent and satisfied inclusion criteria. Three children
missed the regular scheduled weekly casting protocol and
therefore excluded from the study. One child from group
A defaulted after 1 cast and 1 child each from groups A
and B defaulted after 3 casts. There were no defaults from
group C. So, at the time of final analysis after the duration
of 8 months, the study had 51 children.
Demographic Data
There were 16 children (11 bilateral) in group A, 17
(10 bilateral) in group B, and 18 (8 bilateral) in group C.
The mean age of enrolled children was 44 days (SD: 37 d).
The age and sex distribution among the 3 groups matched
(Table 2). The mean number of casts applied before
tenotomy in group A were 4.9 (SD: 1.3), 4.4 (SD: 1.3) in
group B, and 4.5 (SD: 1.3) in group C. There was no
significant difference in number of casts applied in 3
groups (P = 0.588). Tenotomy was done in all except
1 case.
Outcome
Before casting, the baseline parameters (NIPS, HR,
and SpO2) of the 3 groups were comparable. The mean
NIPS score in the group A during casting was 4/7 (mod-
erate pain). Overall, children in group B experienced lesser
pain compared with the other 2 groups during casting
(Table 3). Both group B and C had a significant reduction
in pain score at casting (group B—2/7, group C—3/7) and
in postcasting period (group B—0/7, group C—0/7) when
compared with group A (P < 0.05). Children in group A
had significant residual pain {NIPS 2 [interquartile range
(IQR) = 4]} even during postcasting period, while the
children in group B [NIPS 0 (IQR = 0)] and C [NIPS 0
(IQR = 1)] had near normal NIPS scores (P < 0.00001). On
intragroup comparison, all the 3 groups had a significant
increase in pain score during casting [group A: NIPS 4
(IQR = 1), group B: NIPS 2 (IQR = 2), group C: NIPS 3
(IQR = 1)] followed by a fall in postcasting period
(P < 0.05) (Fig. 1).
There was a significant difference in mean HR be-
tween the 3 groups at casting and in postcasting period.
Highest HR at casting were seen in group C [mean: 191.3/
min (SD: 21.9/min)] and in postcasting period in group A
[mean: 178.3/min (SD: 28.4/min)] (Table 3). group B [At
www.pedorthopaedics.com | e101
Upadhyay et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 1. Neonatal Infant Pain Score TABLE 3. Results and Comparisons of the Parameters of
Parameters Score Interest Among the 3 Groups
Mean (SD)
Facial expressions
0—Relaxed Restful face; neutral expression Parameters Group Precasting At Casting Postcasting P
1—Grimace Tight facial muscles; negative facial
expression NIPS* A 0 (0) 4 (1) 2 (4) 0.52
Cry B 0 (0) 2 (2) 0 (0) < 0.05
0—No cry Quite; not crying C 0 (0) 3 (1) 0 (1) < 0.05
1—Whimper Intermittent moaning HR† A 161.6 (22.7) 189.4 (24.4) 178.3 (28.4) 0.77
2—Vigorous cry Continuous loud scream or shrill B 160.3 (21.9) 174.1 (20.8) 168.2 (22.5) < 0.05
Breathing patterns C 159.1 (21.2) 191.3 (21.9) 173.3 (23.2) < 0.05
0—Relaxed Usual pattern for the patient SpO2† A 98.2 (1.5) 92.3 (4.3) 96.0 (3.1) < 0.05
1—Change in breathing Indrawing; irregular, fast; gagging, breath B 96.8 (2.4) 95.7 (2.6) 97.4 (1.9) < 0.05
holding C 97.1 (2.3) 94.2 (4) 96.8 (2.4) < 0.05
Arms Significant values in bold.
0—Relaxed No rigidity; occasional random arm *Kruskal-Wallis test.
movement †One-way analysis of variance test.
1—Flexed/extended Tense, rigid; rapid extension or flexion HR indicates heart rate; NIPS, Neonatal Infant Pain Score; SpO2, oxygen
Legs saturation.
0—Relaxed No rigidity; occasional random leg
movement
1—Flexed/extended Tense, rigid; rapid extension or flexion On intragroup comparison, there was a significant fall in
State of arousal
0—Sleepy/awake Quiet, peaceful, settled; alert random limb
mean SpO2 in all the 3 groups at casting followed by
movement increase in mean SpO2 in postcasting period (P < 0.05)
1—Fussy Restless, thrashing (Fig. 3).
Total score (sum) Maximum score 7
Category
0-2 Mild to no pain DISCUSSION
3-4 Mild to moderate pain Ponseti technique involves stretching of the soft tissues
>4 Severe pain especially the ligaments and musculotendinous units of the
posterior and medial ankle. The act is repeated at each cast
session to steadily overcome the retractile tissues and
casting—mean: 174.1/min (SD: 20.8/min), postcasting—
ligaments.10 The last cast additionally hyperabducts the foot.
mean: 168.2/min (SD: 22.5/min)] had the lowest HR both
Although the technique aims at gentle and gradual correction
during and after cast application. On intragroup com-
of the clubfoot deformity, it has been observed that the
parison, children in all 3 groups had a significant increase
practise may cause discomfort and pain in some infants.2
in mean HR at casting with maximum increase seen in
Evaluation of pain in neonates via established scores and
group C [mean: 191.3/min (SD: 21.9/min)] followed by
quantitative recording of physiological parameters during
group A [mean: 189.4/min (SD: 24.4/min)] and group B
casting have substantiated the above supposition.2 Fur-
[mean: 174.1/min (SD: 20.8/min)] (P < 0.00001). The mean
thermore, it has been shown that the pain-relieving measures,
HR did not return to baseline in postcasting period in any
if used during Ponseti casting, help soothe the child and re-
of the 3 groups (P < 0.02) although a fall was noticed in all
duce physiological abnormalities.
(Fig. 2). The lowest postcasting HR were observed in
The present study aimed to further investigate the dis-
group B [mean: 168.2/min (SD: 22.5/min)] (P < 0.005).
comfort and pain responses incurred by the clubfoot infants
Group B had the highest SpO2 among all the 3
during the Ponseti casting sessions. We also compared the
groups, both during casting [mean: 95.7% (SD: 2.6%)] and
efficacy of 2 nonpharmaceutical techniques (non-nutritive
after casting [mean: 97.4% (SD: 1.9%)] (P < 0.05, Table 3).
sucking and human care contact) for tackling such pain re-
sponses. The findings from our study provided more evidence
TABLE 2. Demographic Data of the Study Participants that infants experienced pain and exhibited altered physio-
logical responses during casting and even after casting was
Group A Group B Group C P
over. The infant with clubfoot could experience up to mod-
Age (d)* erate pain during casting as graded by NIPS. A heightened
Mean ± SD 45.3 ± 43.6 47.4 ± 34.8 39.8 ± 34.8 0.858 HR and lowered SpO2 indicated altered physiological re-
Minimum-maximum 15-150 15-120 15-150
Sex [n (%)]† sponses during casting. These parameters continued to remain
Female (n = 10) 6 (37.5) 2 (11.8) 2 (11.1) 0.094 abnormal in the postcasting period as well indicating the ex-
Male (n = 41) 10 (62.5) 15 (88.2) 16 (88.9) tended bodily response to Ponseti manipulations. When non-
Side [n (%)]† pharmaceutical pain-relieving interventions were offered to
Bilateral (n = 29) 11 (68.8) 10 (58.8) 8 (44.4)
Left (n = 14) 2 (12.5) 5 (29.4) 7 (38.9) 0.5
these children, lesser pain was experienced by them. Pain relief
Right (n = 8) 3 (18.8) 2 (11.8) 3 (16.7) in non-nutritive sucking group (group B) fared significantly
better than the human care contact (group C) and control
*Kruskal-Wallis test.
†χ2 test. group (group A). In addition, with above interventions, relief
from pain continued in the postcasting period. The effect of

e102 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
Pain score (NIPS)
4.5
3.5
3
Median NIPS
2.5
1.5
0.5
0
Pre casting NIPS
Group A
FIGURE 1. Intragroup comparison of the changes in Neonatal Infant Pain Score (NIPS) in the 3 study groups.
nonpharmaceutical pain-relieving interventions was detectable
on accompanying physiological parameters also. Children in
all 3 groups experienced an increase in mean HR during
casting. However, the HR rise during casting in non-nutritive
sucking group tended to be less intense and pointed towards a
calmer child. Human care contact was unable to control rises
in HR during casting. SpO2 changes almost mirrored the HR,
with highest values observed in the non-nutritive sucking
group, both during and in postcasting period.
Monitoring of pain responses to painful stimuli in ne-
onates in intensive care units and interventions thereof is al-
ready an established practice. Caracal and colleagues5–7
noticed a significantly reduced pain score in term and preterm
neonates during the painful procedure of venepuncture with
the use of pacifier alone or when the pacifier was combined
with sucrose solution. In the studies done by other inves-
tigators, a significant decrease in pain score during the painful
procedure of heel prick in preterm and term neonates was
observed following use of nonpharmaceutical pain-relieving
measures.11–13 The use of pacifier alone or pacifier with breast
milk or pacifier with sucrose effectively controlled pain when
compared with controls.11–14
Cuddling or holding the child in mother’s lap (hu-
man care contact) is another method with potential to
control pain responses in a neonate.15 In a randomized
controlled trial conducted by Beiranvand et al,16 in which
cuddling (n = 30) and control group (n = 30) were com-
pared for control of pain response during venepuncture in
neonates, better pain relief was reported in the ex-
perimental group. Hoarau et al8 compared the efficacy of
oral sucrose with non-nutritive sucking with and without
cuddling during venepuncture in neonates (34 in each
group) and noticed that the rate of subjects experiencing
high pain scores at 60 seconds after the procedure was
significantly lower in the experimental group.
Only one previous study has investigated the pain re-
sponse incurred by neonates during Ponseti casting.2 This
Mayo Clinic group in a double-blinded randomized con-
trolled trial on 33 children assessed NIPS, HR, and SpO2
before, during, and after the casting. The pain-relieving in-
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Pain Relief Measures in Idiopathic Clubfoot
At casting NIPS Post casting NIPS
Group B Group C
terventions used were oral 20% sucrose solution, water, and
milk. Although there was no control group in this research,
the mean pain score for water was significantly more than
milk (P = 0.0005) or sucrose (P < 0.0001) during casting. The
pain score for milk and sucrose matched (P = 0.33). For the
postcasting period, the mean pain score was highest for milk
and least for sucrose. Milk and water postcasting pain scores
were comparable (P = 0.28). Thus, sucrose solution and milk
effectively reduced the pain response during casting sessions.
The sucrose solution administration continued the pain relief
into the postcasting period. Changes in HR and SpO2 during
casting and postcasting did not vary by pain-relieving method
in this study. Our study although used different pain-relieving
interventions, added to the evidence that neonates experience
pain during Ponseti casting and the same can be eased with
nonpharmaceutical measures. Some measures can have
postcasting pain-relieving effect as well.
The clinical implications of this study are far-reaching.
Foremost is the rebuttal of a long-accepted claim that Ponseti
manipulations and casting are practically painless being so
gentle. The safer nonpharmaceutical interventions can po-
tentially lessen the pain and physiological responses exhibited
by the clubfoot children undergoing treatment. The relaxed
and calmer neonate permits better casting experience for both
caregivers and receivers. Although our study failed to show
any difference in the absolute number of treatment casts in
the evaluated groups, advantages could reflect as a better-
applied cast and lesser time spent on each cast. Overall
compliance to the treatment protocol can possibly be better
with less anxious parents.
There are numerous physiological, behavioral, and bio-
chemical markers accompanying pain response in an infant.17
From a research point of view, other physiological parameters
which correlate best with pain response can be explored in
future studies. This can supplement monitoring of child during
and postcasting. The precise timelines of the bodily responses
to manipulation and casting and their return to baseline values
can be another area of future research. Last, but not the least,
the better pain-relieving measures for the purpose also need to
be determined.
www.pedorthopaedics.com | e103
Upadhyay et al J Pediatr Orthop  Volume 43, Number 2, February 2023

Heart rate
200

195

Mean Heat rate (beats/min)

190

185

180

175

170

165

160

155

150
{Mean (SD)} {Mean (SD)} {Mean (SD)}
Pre casting HR At casting HR Post casting HR

Group A Group B Group C

FIGURE 2. Intragroup comparison of the changes in heart rate (HR) in the 3 study groups.

We would like to list the limitations of our study. The
short-term project was undertaken as a pilot study in absence
of relevant statistical values from previous literature. The
study was not blinded as non-nutritive sucking unit (pacifier)
and human care contact was obvious to both caregivers and
receivers. We recorded our end readings at 1 minute post-
casting due to considerable patient load in our clubfoot
clinics. This smaller duration may not be adequate for the
physiological parameters to return to normal. Evaluation of
care receiver’s anxiety did not form part of this study. This
study used plaster of Paris for casting but there are studies
which have opined that use of fiberglass has provided a better
parental experience as well as it can be a viable alternative for
casting.18,19 The strengths were it being first of its kind study
to investigate the pain response in idiopathic clubfoot chil-
dren during Ponseti casting with and without interventions.
Our methodology of randomizing individuals rather than
casting sessions into distinct groups permitted evaluation of
overall effect of intervention in terms of number of casts. As
such, child once allotted to one intervention group remained
in the same group till treatment was over. This instilled
confidence in participating care receivers and guardians as
child was not shuffled in various intervention methods. All
manipulations and castings were done by a single pro-
fessional trained in Ponseti technique, eliminating handling
by different operators. The choice of nonpharmaceutical in-
terventions was carefully decided taking into considerations
of the established cultural practices of the region. Non-nu-
tritive sucking is commonly practiced both in rural and urban
households and the unit is available as an over the counter
product. The use of non-nutritive sucking was restricted to
casting sessions only and not promoted for any further use.
The practice of breastfeeding the child during Ponseti casting
sessions so prevalent in African continent is seldom observed
in our region because of cultural inhibitions and the limited
privacy of busy outpatient clinics.20 We, therefore, included
child placed in guardian’s (mostly mother) lap as the method
of human care contact. The videotaped recordings were de-
ciphered by an independent observer further eliminating bias.
The findings of our study indicate that the Ponseti casting
may not be entirely innocuous despite being gentle. Use of
nonpharmaceutical pain-relieving interventions during cast-

Oxygen saturation
99

97

95
Mean SpO2 (%)

93

91

89

87

85
{Mean (SD)} {Mean (SD)} {Mean (SD)}
Pre casting SpO2 At casting SpO2 Post casting SpO2
Group A Group B Group C

FIGURE 3. Intragroup comparison of the changes in oxygen saturation (SpO2) in the 3 study groups.

e104 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
ing sessions could provide better experience for the child and
turn the sessions calmer. Further large-scale randomized
studies on the issue are warranted.
CONCLUSIONS
Infants exhibit moderate pain response and altered
physiological responses during and after Ponseti casting.
Non-nutritive sucking used as nonpharmaceutical pain-
relieving measure emerged as a better method to lessen
these parameters when compared wit the conventional
technique and human care contact.
REFERENCES
1. Barik S, Nazeer M, Mani BT. Accelerated Ponseti technique:
efficacy in the management of CTEV. Eur J Orthop Surg Traumatol.
2019;29:919–924.
2. Milbrandt T, Kryscio R, Muchow R, et al. Oral sucrose for pain
relief during clubfoot casting: a double-blinded randomized con-
trolled trial. J Pediatr Orthop. 2018;38:430–435.
3. Cignacco E, Hamers JPH, Stoffel L, et al. The efficacy of non-
pharmacological interventions in the management of procedural pain
in preterm and term neonates. a systematic literature review. Eur J
Pain. 2007;11:139–152.
4. Meaney MJ, Aitken DH. The effects of early postnatal handling on
hippocampal glucocorticoid receptor concentrations: temporal pa-
rameters. Brain Res. 1985;354:301–304.
5. Carbajal R, Chauvet X, Couderc S, et al. Randomised trial of
analgesic effects of sucrose, glucose, and pacifiers in term neonates.
BMJ. 1999;319:1393–1397.
6. Elserafy FA, Alsaedi SA, Louwrens J, et al. Oral sucrose and a
pacifier for pain relief during simple procedures in preterm infants: a
randomized controlled trial. Ann Saudi Med. 2009;29:184–188.
7. Curtis SJ, Jou H, Ali S, et al. A randomized controlled trial of
sucrose and/or pacifier as analgesia for infants receiving venipuncture
in a pediatric emergency department. BMC Pediatr. 2007;7:27.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Pain Relief Measures in Idiopathic Clubfoot
8. Hoarau K, Payet ML, Zamidio L, et al. “Holding–Cuddling” and
sucrose for pain relief during venepuncture in newborn infants: a
randomized, controlled trial (CÂSA). Front Pediatr. 2021;8:607900.
9. Hudson-Barr D, Capper-Michel B, Lambert S, et al. Validation of
the Pain Assessment in Neonates (PAIN) scale with the Neonatal
Infant Pain Scale (NIPS). Neonatal Netw. 2002;21:15–21.
10. Zhao D, Liu J, Zhao L, et al. Relapse of clubfoot after treatment
with the ponseti method and the function of the foot abduction
orthosis. Clin Orthop Surg. 2014;6:245–252.
11. Liaw JJ, Yang L, Ti Y, et al. Non-nutritive sucking relieves pain for
preterm infants during heel stick procedures in Taiwan. J Clin Nurs.
2010;19(19–20):2741–2751.
12. Vu-Ngoc H, Uyen NCM, Thinh OP, et al. Analgesic effect of non-
nutritive sucking in term neonates: a randomized controlled trial.
Pediatr Neonatol. 2020;61:106–113.
13. Peng HF, Yin T, Yang L, et al. Non-nutritive sucking, oral breast
milk, and facilitated tucking relieve preterm infant pain during heel-
stick procedures: a prospective, randomized controlled trial. Int J
Nurs Stud. 2018;77:162–170.
14. Gao R, Tomlinson M, Walker C. Correlation of Pirani and Dimeglio
scores with number of Ponseti casts required for clubfoot correction.
J Pediatr Orthop. 2014;34:639–642.
15. Harrison D, Bueno M, Reszel J. Prevention and management of pain
and stress in the neonate. Res Rep Neonatol. 2015;5:9–16.
16. Beiranvand S, Faraji GM, Firouzi M. The effects of caressing and
hugging infants to manage the pain during venipuncture. Compr
Child Adolesc Nurs. 2020;43:142–150.
17. Mathew P, Mathew J. Assessment and management of pain in
infants. Postgrad Med J. 2003;79:438–443.
18. Hui C, Joughin E, Nettel-Aguirre A, et al. Comparison of cast
materials for the treatment of congenital idiopathic clubfoot using
the Ponseti method: a prospective randomized controlled trial. Can J
Surg. 2014;57:247–253.
19. Williams B, Gil JN, Oduwole S, et al. Semirigid fiberglass casting for
the early management of clubfoot: a single-center experience. Cureus.
2022;14:e22683.
20. Pirani S, Staheli L, Naddumba E. Ponseti Clubfoot Management:
Teaching Manual For Healthcare Providers in Uganda, 1st ed. GHO
Publications; 2008:48.
www.pedorthopaedics.com | e105
 ORIGINAL ARTICLE
The Ponseti Method for the Treatment of Clubfeet
Associated With Down Syndrome: A Single-institution
18-year Experience
Edwin Portalatin Perez, MD,* Sarah Parenti, BA,* Jordan Polk, BA,* Chanhee Jo, PhD,*
and Anthony I. Riccio, MD*†
Background: Although the Ponseti method has been used with
great success in a variety of nonidiopathic clubfoot deformities,
the efficacy of this treatment in clubfeet associated with Down
syndrome remains unreported. The purpose of this study is,
therefore, to compare treatment characteristics and outcomes of
clubfoot patients with Down syndrome to those with idiopathic
clubfoot treated with the Ponseti method.
Methods: An Institutional Review Board–approved, retro-
spective review of prospectively gathered data were performed at
a single pediatric hospital over an 18-year period. Patients with
either idiopathic clubfeet or clubfeet associated with Down
syndrome who were less than 1 year of age at the outset of
treatment were treated by the Ponseti method, and had a mini-
mum of 2 year’s follow-up were included. Initial Dimeglio score,
number of casts, need for heel cord tenotomy, recurrence, and
need for further surgery were recorded. Outcomes were classified
using the Richards classification system: “good” (plantigrade
foot +/− heel cord tenotomy), “fair” (need for a limited proce-
dure), or “poor” (need for a full posteromedial release).
Results: Twenty clubfeet in 13 patients with Down syndrome and
320 idiopathic clubfeet in 215 patients were identified. Average
follow-up was 73 months for the Down syndrome cohort and
62 months for the idiopathic cohort. Down syndrome patients
presented for treatment at a significantly older age (61 vs. 16 d,
P = 0.00) and with significantly lower average initial Dimeglio
scores than the idiopathic cohort (11.3 vs. 13.4, P = 0.02). Heel
cord tenotomy was performed in 80% of the Down syndrome
cohort and 79% of the idiopathic cohort (P = 1.00). Recurrence
rates were higher in the Down syndrome cohort (60%) compared
with the idiopathic group (37%), but this difference was not
statistically significant (P = 0.06). Need for later surgical proce-
dures was similar between the 2 cohorts, though recurrences in
the Down syndrome group were significantly less likely to require
intra-articular surgery (8.3% vs. 65.5%, P = 0.00). Clinical out-
comes were 95% “good,” 0% “fair,” and 5% “poor” in the Down
From the *Scottish Rite for Children, Department of Orthopaedic Sur-
gery; and †University of Texas Southwestern School of Medicine,
Dallas, TX.
The authors declare no conflicts of interest.
Reprints: Anthony I. Riccio, MD, Texas Scottish Rite Hosiptal for Children,
Dallas, TX. E-mail: anthony.riccio@tsrh.org.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002293
e106 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
syndrome cohort and 69% “good,” 27% “fair,” and 4% “poor” in
the idiopathic cohort (P = 0.01).
Conclusions: Despite the milder deformity and an older age at
presentation, clubfeet associated with Down syndrome have
similar rates of recurrence and may have better clinical outcomes
when compared with their idiopathic counterparts. When de-
formities do relapse in Down syndrome patients, significantly less
intra-articular surgery is required than for idiopathic clubfeet.
Level of Evidence: Level III.
Key Words: Down syndrome, clubfoot, Ponseti, foot, syndromic,
idiopathic
(J Pediatr Orthop 2023;43:e106–e110)
M ost musculoskeletal abnormalities in children with
trisomy 21 (Down syndrome) are related to gener-
alized ligamentous laxity and joint hypermobility.1,2 At-
lantooccipital instability, hip dislocations, patellar
instability, hallux valgus, and planovalgus foot deform-
ities are commonly encountered in this patient population.
Each of these are directly associated with excessive liga-
mentous laxity which, in turn, is believed to be a causative
factor of relatively high recurrence rates and complica-
tions when surgical management is undertaken.1–3
It is therefore interesting, and essentially paradoxical,
that a small percentage of children with Down syndrome
are born with clubfoot deformities.4,5 Unlike many of the
musculoskeletal manifestations of Down syndrome, both
idiopathic and nonidiopathic clubfoot deformities, are
typically characterized by ligamentous, tendinous, and
capsular rigidity and/or contracture. Nonidiopathic and
syndromic clubfeet are often resistant to the Ponseti method
of treatment, and therefore require a higher rate of surgical
correction than their idiopathic counterparts.6,7 Although
this is believed to be, at least in part, due to greater degrees
of soft tissue contracture and rigidity in these nonidiopathic
deformities, such rigidity is otherwise extremely un-
characteristic in Down syndrome and it is therefore unclear
whether clubfoot deformities in these children behave sim-
ilarly to other nonidiopathic clubfeet.
To the best of our knowledge, there is no literature
characterizing clubfoot deformities in infants with Down
syndrome nor has the efficacy of the Ponseti method been
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
studied in Down syndrome–related clubfeet. The purpose
of this study was therefore to evaluate the outcome of the
Ponseti method in the treatment of patients whose clubfeet
are associated with Down syndrome and to compare these
with an idiopathic clubfeet cohort.
METHODS
An Institutional Review Board–approved, retro-
spective review of a prospective clubfoot registry was
performed on patients with concurrent diagnoses of
clubfoot and Down syndrome treated at a single pediatric
orthopaedic hospital from 2000 to 2018. All clubfeet
presenting during this period were identified. Genetically
confirmed Down syndrome patients with clubfeet man-
aged by the Ponseti method who were less than 1 year of
age at treatment onset, had no prior treatments, and had a
minimum of 2 year’s follow-up met inclusion criteria. This
cohort was compared with a cohort of idiopathic clubfoot
patients identified using the same inclusion criteria.
Over the period studied, 11 pediatric orthopaedic
surgeons treated clubfeet at our institution according to
Ponseti protocol.8 A percutaneous tenotomy of the
Achilles tendon was performed to address any residual
equinus following after the of other component deform-
ities. Tenotomies were typically performed in the out-
patient clinic for patients 3 months of age or less and in the
operating room for older children. A post-tenotomy cast
was maintained for 3 weeks, at which time patients were
immediately transitioned into a foot abduction orthosis.
This brace was prescribed for 23 hours per day for a
minimum of 3 months, at which point brace use was de-
creased to night and naptime use for a minimum of
2 years. In both the idiopathic and Down syndrome pa-
tients, foot abduction orthoses were set to 60 to 70 degrees
of abduction on the clubfoot side in unilateral cases and
bilaterally for children with bilateral deformities.
Medical records were reviewed to document age at
presentation, sex, laterality (unilateral or bilateral), initial
Dimeglio score, number of casts, need for Achilles tenot-
omy, time to final follow-up, recurrence, and need for
further surgery. Outcomes were classified as previously
described by Richards et al9 as “good” (plantigrade foot
with or without need for Achilles tenotomy), “fair” (need
for a limited procedure such as tibialis anterior transfer,
lateral column shortening, or posterior release), or “poor”
(need for a posteromedial release). Recurrence was defined
TABLE 1. Demographic Parameters
Down Idiopathic
Syndrome Clubfoot
Age at initial presentation (d) 60.9 (14-308) 16 (3-96)
Follow-up (mo) 72.8 (30-162) 61.9 (24-112)
Male:female (% male) 10:3 (76.9) 143:72 (66.5)
Unilateral:bilateral 6:7 (53.9) 110:105 (48.8)
(% bilateral)
Initial Dimeglio score 11.3 (6-17) 13.4 (5-20)
Bold values has statistical significance.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Clubfoot in Downs Syndrome
as any deformity that required additional casting and/or
surgical procedures to restore satisfactory position in a
child who had completed the Ponseti method. Failure of
Ponseti treatment was defined as a clubfoot that failed to
correct fully with initial serial casting and/or a tenotomy.
Statistical Analysis
Age at presentation, age at final follow-up, Dimeglio
scores, and number of casts were compared between the
Down syndrome and idiopathic cohorts using the Mann-
Whitney test. A χ2 test and the Fisher exact test (as
appropriate) were used to compare categorical values in-
cluding sex, laterality, occurrence of surgical procedures,
recurrences, and clinical outcome. The results were de-
termined to be statistically significant for P-values < 0.05.
RESULTS
A total of 20 Down syndrome clubfeet (13 patients)
and 320 idiopathic clubfeet (215 patients) were included
(Table 1). Patients with Down syndrome were, on average,
older at the time of initial presentation (60.9 d, range: 14
to 308 d) than their idiopathic counterparts (16 d, range: 3
to 96 d) and this difference was statistically significant
(P = 0.00). Average Dimeglio score at initial presentation
was significantly lower in Down syndrome–associated
clubfeet (11.3, range: 6 to 17) than idiopathic clubfeet
(13.4, range: 5 to 20) (P = 0.02). No statistically significant
differences were identified between the cohorts regarding
length of follow-up, sex, or laterality.
Table 2 summarizes the treatment parameters and
outcomes. No statistically significant differences were
identified between the Down syndrome and idiopathic
cohorts regarding number of casts required for initial
correction (4.1 vs. 4.4) the need for an Achilles tenotomy
(80% vs. 79%), or the success with obtaining initial
correction using the Ponseti method (95% vs. 96.3%). No
feet in the Down Syndrome cohort were noted to have
overcorrection into a planovalgus position at final follow-
up. Recurrence rates were similar between the 2 groups,
but recurrences developed at a significantly older age in
the Down syndrome cohort (5.6 vs. 2.7 y, P = 0.001).
There was no significant difference in the incidence of
recurrence between patients braced using the traditional
protocol (3-mo full time followed by night/nap race use
until age 2) and those braced more conservatively (full
time until pulling to stand followed by nighttime bracing
until age 4) in either the idiopathic (37.0% vs. 37.2%,
P = 0.79) or Down syndrome (55.6% vs. 100%, P = 0.19)
group. The percentage of good, fair, and poor outcomes
also differed, with the Down syndrome group having
significantly less fair or poor outcomes (P = 0.01).
P
Of the 12 feet with recurrent deformity in the Down
0.000 syndrome cohort, 4 (33.3%) underwent further treatment:
0.962 25% were managed with repeat casting and a repeat
0.553
0.772 Achilles tenotomy, whereas 1 (8.3%) ultimately required a
posteromedial release. None of these feet developed a
0.017 second recurrence. Of the remaining 8 feet with recurrence
in this cohort, a tibialis anterior tendon transfer was rec-
ommended in 5 but these families declined surgical
www.pedorthopaedics.com | e107
Portalatin Perez et al J Pediatr Orthop  Volume 43, Number 2, February 2023

nonidiopathic, the present study suggests that such feet in

TABLE 2. Treatment Parameters, Recurrence Rates, and
Outcomes patients with Down syndrome demonstrate a response to
the Ponseti method and the rates of recurrence similar to
Down Idiopathic
idiopathic clubfeet. This speaks to the fact that non-
Syndrome Clubfoot P
idiopathic clubfeet represent a fairly heterogeneous group
No. of casts for initial 4.1 (1-6) 4.4 (2-10) 0.237 as associated connective tissue rigidity or laxity certainly
correction
Initial TAT* n/N (%) 16/20 (80) 253/320 (79.1) 1.000
varies between differing diagnoses. All Down syndrome
Outcome (feet) clubfeet achieved correction via the Ponseti method with
Good 19 222 0.011 83% of these maintaining a good outcome at nearly
Fair 0 86 — 5.5 year’s follow-up. Although clinical outcomes were
Poor 1 12 — found to be superior in Down syndrome patients, it is
Feet with recurrence 12/20 (60) 119/320 (37.2) 0.064
n/N (%) important to note that this difference was based upon
Additional procedures (feet)† n/N (%) actual repeat interventions rather than intent to treat as
Repeat TAT 3/12 (25) 28/119 (23.5) 0.999 will be discussed in more detail below.
Tibialis anterior transfer 0/12 56/119 (47.1) 0.001 Down syndrome–associated clubfeet were found to
Posterior release 0/12 66/119 (55.5) 0.000
Posteromedial release 1/12 (8.3) 12/119 (10.1) 0.999
present for initial treatment later in infancy than infants with
Intra-articular surgeries 1/12 (8.3) 78/119 (65.5) 0.000 idiopathic deformities. Patients with Down syndrome were,
(feet) on average, over 1 month older than their idiopathic coun-
Bold values has statistical significance.
terparts at the time of initial presentation. The reasons for this
*TAT (percutaneous tendoachilles tenotomy). delay in time to treatment are unclear, but may be because of
†Of the 119 feet that recurred in the idiopathic group, 45 underwent a single the need for evaluation and management of other co-
surgical intervention and 53 underwent multiple procedures. Twenty-one of these
feet only underwent additional casting.
morbidities (ie, cardiac anomalies, hypotonia, and persistent
pulmonary hypertension) associated with the syndrome that
take precedence over an extremity deformity. In addition,
intervention and 3 were felt to be minor and warrant no newborns with Down syndrome have been shown to have
further treatment. Idiopathic clubfoot recurrences were higher rates of admission to neonatal intensive care units and
managed with a wider variety, and more evenly dis- significantly longer hospitalizations immediately following
tributed number of surgical procedures after repeat serial birth than unaffected infants, which would obviously result in
casting (Table 2). When recurrences managed only with a delay of outpatient orthopaedic assessment.11 Although
repeat casting and tenotomies were excluded, the most surgeons advocate for initiation of Ponseti casting
performance of more major surgical procedures was within the first 2 weeks of life,8,12 the age at initiation of
found to be significantly lower in the Down syndrome treatment has not been found to be predictive of successful
cohort (8.3% vs. 81.5% of patients, P = 0.0006). Although deformity correction or recurrence13,14 in idiopathic clubfeet.
96% of feet in both groups had either a “good” or “fair” Although this has not been investigated in nonidiopathic
outcome, only 1 of the 12 (8.3%) Down syndrome– deformities, the present data suggests that in the Down syn-
associated clubfeet that recurred required intra-articular drome population, success with initial deformity correction,
surgery versus 65.5% of idiopathic clubfeet that recurred outcomes, and recurrences mirror idiopathic clubfeet even
(P = 0.00). Of course, these analyses do not factor in the 5 though casting was started later in infancy.
feet in which the recommendation for a tibialis anterior The importance of initial Dimeglio score is controversial,
tendon transfer declined. Were “intent to treat” rather as some studies have reported that it is not predictive of
than “actual surgeries performed” analyzed, the statistical treatment success or outcome measures,14–17 whereas others
difference identified for the rate of tibialis anterior tendon have reported correlations between initial Dimeglio score, the
transfers between the Down syndrome and idiopathic number of casts required to achieve correction, and
cohorts would be lost (41.7% vs. 55.5%, P = 0.8) and the recurrence.18–21 Although Down syndrome–associated clubfeet
distribution of good, fair, and poor outcomes would be had lower initial Dimeglio scores than idiopathic clubfeet, no
equivalent between the 2 groups (P = 0.8). The need for differences were found regarding the number of casts required
intra-articular surgery for recurrence, however, would re- for initial correction, need for an Achilles tenotomy, recurrence
main significantly higher in the idiopathic group (8.5% vs. rates, or outcomes. These findings contrast with previous lit-
65.5%, P = 0.000) (Table 3). erature suggesting that nonidiopathic clubfeet typically require
more casts to achieve initial correction and have worse out-
comes than idiopathic feet.6,22 This difference between Down
DISCUSSION syndrome clubfeet and other nonidiopathic clubfoot deform-
The Ponseti method has been proven to be effective ities may be because of the inherent soft tissue laxity associated
for the treatment of nonidiopathic clubfeet.6,7,10 None- with the condition, which in turn might explain the lower
theless, it is well reported that such feet often require a severity by Dimeglio scoring at initial presentation.
greater number of casts to achieve initial correction, have Contrary to our hypothesis, Down syndrome club-
higher rates of recurrence, and are more likely to require feet had comparable rates of recurrence after Ponseti
surgical intervention than idiopathic clubfeet.6,10 Al- treatment when compared with the idiopathic cohort.
though syndromic clubfeet are appropriately considered Surgical intervention for recurrence, however, was lower

e108 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

 J Pediatr Orthop  Volume 43, Number 2, February 2023
TABLE 3. Recurrence Outcomes (Intent to Treat)
Down Idiopathic
Syndrome Clubfoot
Additional procedures (feet)† n/N (%)
Repeat TAT 3/12 (25) 28/119 (23.5)
Tibialis anterior transfer 5/12 (41.7) 56/119 (47.1)
Posterior release 0/12 66/119 (55.5)
Posteromedial release 1/12 (8.3) 12/119 (10.1)
Intra-articular surgeries 1/12 (8.3) 78/119 (65.5)
(feet)
Bold values has statistical significance.
TAT indicate, percutaneous tendoachilles tenotomy.
†Of the 119 feet that recurred in the idiopathic group, 45 underwent a single
surgical intervention and 53 underwent multiple procedures. Twenty-one of these
feet only underwent additional casting.
in the Down syndrome feet. Surgery for recurrence was
performed in 33% of Down syndrome clubfeet, whereas
82% of idiopathic relapses underwent 1 or multiple sur-
gical procedures to achieve correction. This difference was
more remarkable when comparing only intra-articular
surgical procedures, which were performed significantly
less in Down syndrome clubfeet (8.3% vs. 65.5%). This
contrasts with literature suggesting that the nonidiopathic
clubfeet undergo more subsequent surgical treatment for
relapses after Ponseti casting.6,10 Perhaps most im-
portantly, however, was the fact that Down syndrome
clubfeet required dramatically less intra-articular surgery
(posterior release or posteromedial release) when com-
pared with idiopathic clubfeet. A fundamental tenet of the
Ponseti method is the reliance on repeat casting for club-
foot recurrence to eliminate or obviate the need for
subsequent surgery and to avoid, whenever possible, intra-
articular surgery, which is believed to carry a greater risk
of long-term morbidity from stiffness and the potential for
creation of iatrogenic deformity. The soft tissue laxity
inherent in the Down syndrome population may explain
the greater responsiveness to repeat casting for relapsed
feet and, in turn, the subsequent need for less invasive
surgical interventions.
In terms of major surgical procedures, the Down syn-
drome population underwent less tibialis anterior tendon
transfers and posterior releases when treating the recurrence of
deformity. Of note, 3 patients (5 feet) in the Down syndrome
cohort demonstrated dynamic supination for which tibialis
anterior tendon transfers were recommended. These proce-
dures were not performed because of the family’s decision to
not pursue any major surgical management. Were our data
reanalyzed by intent to treat rather than actual surgeries per-
formed, the statistical difference identified for the rate of tibialis
anterior tendon transfers between the Down syndrome and
idiopathic cohorts would be lost (41.7% vs. 55.5%, P = 0.8)
and the distribution of good, fair, and poor outcomes would be
equivalent between the 2 groups (P = 0.8).
There are several limitations to our study. In addition to
its retrospective nature, the patients included in this study were
treated over an 18-year period by 11 pediatric orthopaedic
surgeons. This almost certainly resulted in variability in treat-
ment protocols between these physicians. That being stated,
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Clubfoot in Downs Syndrome
the comparison group was treated by the same providers and
any variability in treatment protocols was consistent regardless
of whether the deformity was associated with Down syndrome
P
or not. There is also variability in the postcasting bracing
protocols used over the 18 years during which data were col-
0.999 lected. Recommendations for full-time wear of abduction
0.8
0.000 orthoses ranged from a minimum of 3 months to the age at
0.999 which the patient is able to pull to a stand. Part-time brace
0.000 wear recommendations ranged from a minimum of 2 to
4 years for some providers. It is unclear whether these differing
bracing regimens may have had an undetected effect on the
study results. That noted, there was no statistically significant
difference in the incidence of recurrence between patients
braced using the traditional protocol and those braced more
conservatively in either the idiopathic or Down syndrome
group. More importantly, is the fact that bracing compliance
with any of these protocols could not be verified in this ret-
rospective format. Brace compliance is critical to the success of
the Ponseti method; it would therefore have been ideal to
verify brace use using any of the modern compliance mon-
itoring systems available in the present day.
Although this study is the largest cohort of Down
syndrome clubfeet treated with the Ponseti method to our
knowledge, it is still limited by a very small number of
patients. This highlights the need for a larger, multicenter
analysis, should we wish to learn more about clubfeet in
this patient population. it must also be stated that the
Dimeglio classification system was developed for idio-
pathic deformities and has not been validated for non-
idiopathic clubfeet. Although routinely used to grade
nonidiopathic clubfeet, the utility of the Dimeglio score in
patients with Down syndrome certainly remains ques-
tionable. In addition, postcasting Dimeglio scores were
not recorded which might have quantified “completeness”
of correction before bracing. It should also be noted that
though our average follow-up time was over 5 years in
both cohorts, the minimum follow-up time of 2 years is
somewhat short and additional recurrences might have
been detected with a longer follow-up. Lastly, clinical
outcomes were dependent upon the treatment required to
obtain a plantigrade foot along with the provider and
parent’s assessment of what defined an acceptably planti-
grade foot in any given patient. Although this outcome
classification has routinely been used in clubfoot research,
the evaluation of outcomes using functional assessment,
gait analysis, and/or validated foot-specific patient re-
ported outcome tools would have provided far more
meaningful objective data with which to assess outcomes.
In conclusion, this is the largest series of Down
syndrome–associated clubfeet patients treated by the
Ponseti method to date, to our knowledge. The Ponseti
method of casting is an effective means of obtaining de-
formity correction in these patients. Although classified as
nonidiopathic, clubfeet in Down syndrome patients are
less severe at the time of presentation and have similar
rates of recurrence and clinical outcomes when compared
with idiopathic clubfeet. When Down syndrome clubfeet
do relapse, these recurrences are less likely to require open
or intra-articular surgical intervention.
www.pedorthopaedics.com | e109
Portalatin Perez et al
REFERENCES
1. Caird MS, Wills BP, Dormans JP. Down syndrome in children: the
role of the orthopaedic surgeon. J Am Acad Orthop Surg. 2006;14:
610–619.
2. Mik G, Gholve PA, Scher DM, et al. Down syndrome: orthopedic
issues. Curr Opin Pediatr. 2008;20:30–36.
3. Foley C, Killeen OG. Musculoskeletal anomalies in children with Down
syndrome: an observational study. Arch Dis Child. 2019;104:482–487.
4. Mulpruek P, Jirasirikul A. Down’s syndrome presented with clubfoot
deformity: a case report. J Med Assoc Thai. 1999;82:1254–1256.
5. Miller PR, Kuo KN, Lubicky JP. Clubfoot deformity in Down’s
syndrome. Orthopedics. 1995;18:449–452.
6. Janicki JA, Narayanan UG, Harvey B, et al. Treatment of
neuromuscular and syndrome-associated (nonidiopathic) clubfeet
using the Ponseti method. J Pediatr Orthop. 2009;29:393–397.
7. Zionts LE, Habell B. The use of the Ponseti method to treat clubfeet
associated with congenital annular band syndrome. J Pediatr Orthop.
2013;33:563–568.
8. Ponseti IV. Congenital Clubfoot: Fundamentals of Treatment. Oxford
University Press; 1996.
9. Richards BS, Faulks S, Rathjen KE, et al. A comparison of two
nonoperative methods of idiopathic clubfoot correction: the Ponseti
method and the French functional (physiotherapy) method. J Bone Jt
Surg. 2009;91:299–312.
10. Gurnett CA, Boehm S, Connolly A, et al. Impact of congenital
talipes equinovarus etiology on treatment outcomes. Dev Med Child
Neurol. 2008;50:498–502.
11. Morris J, Mann J, Statnikov E, et al. G373 Management and
outcomes of neonates with down syndrome admitted to neonatal
units. Arch Dis Child. 2016;101:A218.
12. Ponseti IV. Treatment of congenital club foot. J Bone Joint Surg Am.
1992;74:448–454.
e110 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
13. Alves C, Escalda C, Fernandes P, et al. Ponseti method: does age at
the beginning of treatment make a difference? Clin Orthop Relat Res.
2009;467:1271–1277.
14. Dobbs MB, Rudzki JR, Purcell DB, et al. Factors predictive of
outcome after use of the Ponseti method for the treatment of
idiopathic clubfeet. J Bone Joint Surg Am. 2004;86:22–27.
15. Chu A, Labar AS, Sala DA, et al. Clubfoot Classification: Corre-
lation with Ponseti Cast Treatment. J Pediatr Orthop. 2010;30:-
695–699.
16. Gao R, Tomlinson M, Walker C. Correlation of Pirani and Dimeglio
scores with number of Ponseti casts required for clubfoot correction.
J Pediatr Orthop. 2014;34:639–642.
17. Azarpira MR, Emami MJ, Vosoughi AR, et al. Factors associated
with recurrence of clubfoot treated by the Ponseti method. World J
Clin Cases. 2016;4:318–322.
18. Noh H, Park SS. Predictive factors for residual equinovarus
deformity following Ponseti treatment and percutaneous Achilles
tenotomy for idiopathic clubfoot: a retrospective review of 50 cases
followed for median 2 years. Acta Orthop. 2013;84:213–217.
19. Zhang W, Richards BS, Faulks ST, et al. Initial severity rating of
idiopathic clubfeet is an outcome predictor at age two years.
J Pediatr Orthop B. 2012;21:16–19.
20. Goldstein RY, Seehausen DA, Chu A, et al. Predicting the need for
surgical intervention in patients with idiopathic clubfoot. J Pediatr
Orthop. 2015;35:396–402.
21. Brazell C, Carry PM, Jones A, et al. Dimeglio score predicts
treatment difficulty during Ponseti casting for isolated clubfoot.
J Pediatr Orthop. 2019;39:e402–e405.
22. Esparza M, Tran E, Richards BS, et al. The Ponseti method for the
treatment of clubfeet associated with amniotic band syndrome: a
single institution 20-year experience. J Pediatr Orthop. 2021;41:
301–305.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Gait Analysis Characteristics in Relapsed Clubfoot
Melissa Bent, MD,*† Maia Hauschild, BA,* Susan A. Rethlefsen, PT, DPT,*
Tishya A.L. Wren, PhD,*† Adriana Liang, MS,* Rachel Y. Goldstein, MD,*†
and Robert M. Kay, MD*†
Background: Relapse rates of clubfoot deformity after initial
correction range between 19% and 68% regardless of treatment
approach. Most studies focus on relapse before age 4. Little re-
search has focused on late clubfoot relapse. The purpose of this
study was to compare the gait characteristics of children with late
clubfoot relapse (age ≥ 5 y) following treatment with the Ponseti
method only compared with intra-articular and extra-articular
surgeries.
Methods: A retrospective review was conducted of all patients
with idiopathic clubfoot ≥ 5 years old who underwent compu-
terized gait analysis for clubfoot relapse between 2001 and 2021.
Joint range of motion, muscle strength, gait kinematics, and ki-
netics were compared among 3 groups based on prior clubfoot
treatment: (1) Ponseti casting, (2) Extra-articular (EA) surgery,
and (3) Intra-articular (IA) surgery.
Results: Sixty-eight subjects (107 feet) were included (39 bi-
lateral). Thirty-one percent of feet had been treated with Ponseti
casting alone; 57% had IA surgery, and 12% had EA surgery.
The average age when presenting with late relapse was 8.2 years,
9.0 years and 10.7 years for the Ponseti, and IA and EA groups,
respectively. The IA group had greater passive dorsiflexion than
the other 2 groups (P < 0.002), greater inversion weakness than
the other 2 groups (P < 0.0001), greater dorsiflexion during the
stance phase of gait compared with the Ponseti group
(P = 0.001), and lower maximum power production at push-off
compared with the other 2 groups (P = 0.009).
Conclusion: Late relapse can occur after all types of clubfoot
correction. Consistent with existing literature, patients who have
undergone posteromedial release surgery have significantly
greater plantarflexor weakness resulting in poorer plantarflexor
moment and power production during gait.
Level of Evidence: Level III, retrospective comparative study.
Key Words: clubfoot, relapse, gait analysis
(J Pediatr Orthop 2023;43:65–69)
From the *Jackie and Gene Autry Orthopedic Center, Children’s Hos-
pital Los Angeles; and †Keck School of Medicine, University of
Southern California, Los Angeles, CA.
R.M.K., MD owns stock in Zimmer-Biomet, Medtronic, Pfizer, and
Johnson and Johnson. The remaining authors declare no conflicts of
interests.
Reprints: Susan A. Rethlefsen, PT, DPT, Jackie and Gene Autry Or-
thopedic Center, Children’s Hospital Los Angeles, 4650 Sunset Blvd,
M/S 69, Los Angeles, CA 90027. E-mail: srethlefsen@chla.usc.edu.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002314
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
I diopathic clubfoot occurs in an estimated 1 to 2 per 1000
live births.1 Before the widespread adoption of Ponseti
manipulation and casting to correct deformities, compre-
hensive surgical clubfoot correction was commonly nec-
essary due to insufficient correction obtained with other
casting methods. Though satisfactory in the short-term,
long-term results of comprehensive surgical release of
clubfoot deformities have been poor, with many studies
showing significantly greater stiffness, pain, and gait kin-
ematic and kinetic problems in surgically treated clubfoot
patients than those treated nonsurgically.2–5 The Ponseti
method of casting for clubfeet has proven more successful,
with the preservation of a range of motion and muscle
strength resulting in better gait outcomes, and is now the
treatment of choice for orthopedists treating infants
with clubfeet.6
Relapse of deformity after initial correction of idi-
opathic clubfoot, including Ponseti casting, has been re-
ported with rates ranging between 19% and 68%.3–5,7–11
Most studies focus on relapse before age 4.3,8,11–13 Fewer
have reported on relapse in later childhood.4,7,14,15 In
addition, little research has focused on late relapse, com-
paring those treated with the Ponseti method with those
treated with joint sparing and joint invasive surgeries after
initial casting treatment. The purpose of this study was to
compare gait characteristics of children with late clubfoot
relapse (defined as ≥ 5 y of age) following treatment with
the Ponseti method compared with intra-articular (post-
eromedial release) and extra-articular (joint sparing)
surgery. We hypothesized that there would be significant
differences in passive range of motion, muscle stren-
gth, and ankle kinematics and kinetics between the
three groups.
METHODS
Institutional review board approval was obtained
before study initiation, with a waiver of informed consent
to access existing medical records. A retrospective review
was conducted of all ambulatory patients with a diagnosis
of idiopathic clubfoot ≥ 5 years of age who underwent a
gait study for clubfoot relapse in preparation for surgical
intervention between May 21, 2001 to March 9, 2021 at
our institution. Relapse was defined as the presence of
any component of the original clubfoot deformity for
which the patient and/or parents sought further medical
attention.
www.pedorthopaedics.com | 65
Bent et al J Pediatr Orthop  Volume 43, Number 2, February 2023

Computerized gait analysis had been performed for
all participants. Three-D joint kinematic and kinetic data
were collected using a Vicon motion capture system (Vi-
con Motion Systems Ltd., Oxford, UK) and 4 force plates
embedded in the floor (AMTI, Inc., Watertown, MA).
Fifteen to 19 retro-reflective markers were placed on the
subject’s lower body according to the Plug-In-Gait Model.
Subjects walked barefoot along a 15-meter walkway at a
self-selected speed. For each patient, 5 to 10 trials were
recorded, and data from a representative stride were used
for analysis. Kinematic and kinetic data were collected at
120 and 2400 Hz, respectively. Physical examination was
performed by a gait laboratory physical therapist as part
of the gait analysis testing.
Preoperative and postoperative gait analysis records
were reviewed and data were extracted for analysis, in-
cluding ankle joint range of motion and strength, tempo-
ral-spatial parameters of gait, maximum, minimum and
average hip, knee and ankle joint motions during gait and
joint net internal moments and powers in stance phase
when available. Clubfoot treatments performed were de-
termined through a review of medical records and oper-
ative notes if done at the authors’ institution. If the
treatment was done elsewhere, the type of treatments re-
corded in the current physician’s notes was used.
Analysis of variance (ANOVA) with Bonferroni
post-hoc tests and Fisher exact tests were used to compare
patient-level variables (demographics, temporal-spatial
parameters) among 3 groups based on prior clubfoot
treatment: (1) Ponseti casting, (2) Extra-articular surgery,
and (3) Intra-articular surgery. Similar analysis was per-
formed for limb-level variables (joint range of motion,
muscle strength, gait kinematics, and kinetics), except that
linear mixed models were used to account for the inclusion
of 2 limbs for bilaterally involved subjects.16 These models
included a random effect for the subject and a fixed effect
for the side. Statistical significance was set at P < 0.05.
RESULTS
Sixty-eight subjects (107 feet) were included (38 bi-
lateral). There were 24 females and 44 males. Thirty-three
feet (31%) had undergone Ponseti casting with percuta-
neous tendo-Achilles lengthening (TAL) alone (Ponseti
group), 13 feet (12%) had undergone casting, and sub-
sequent extra-articular foot surgery (EA group), and 61
feet (57%) had undergone casting with subsequent intra-
articular foot surgery (IA group) for recurrent clubfoot
deformity (91% before age 5) before being referred for gait
analysis. Initial casting treatment was confirmed by med-
ical record review for all subjects in the Ponseti group.
Initial casting (Ponseti or other technique) was also con-
firmed by the medical record review for subjects in the IA
and EA groups who had initial treatment at the authors’
institution. Subjects in the IA and EA groups who had
initial treatment elsewhere or at the authors’ institution
but before the electronic medical record system was in use
were presumed to have had initial serial casting treatment
before surgical intervention as all were born after 1990, at
which point serial clubfoot casting was the standard of
care. Extra-articular surgeries included tendo-achilles
lengthening, plantar fasciotomy, anterior tibialis tendon
surgery (either split or whole tendon), abductor hallucis
lengthening, and midfoot procedures (including osteoto-
mies). Intra-articular surgeries included the posterior re-
lease and postero-medial release (PMR).

TABLE 1. Participant Characteristics and Clinical Exam Measures

Ponseti Extra-articular Intra-articular
N = 22 Subjects, n (%) N = 7 Subjects, n (%) N = 39 Subjects, n (%) P
Sex — — — 0.62
Female 9 (41) 3 (43) 12 (31) —
Male 13 (59) 4 (57) 27 (69) —
Age (y) 8.2 (2.5) 10.7 (4.4) 9.0 (3.2) 0.17
N = 33 feet N = 13 feet N = 61 feet —
Passive dorsiflexion with knee flexed, hindfoot inverted (°) −0.8 (9.9) −3.5 (9.3) 8.0 (13.1) 0.0017*
Passive dorsiflexion with knee extended, hindfoot inverted (°) −4.6 (10.4) −5.4 (7.7) 5.1 (12.4) 0.0008†
Plantarflexion strength — — — 0.10
0-2 3 (12) 1 (9) 17 (32) —
3 14 (56) 5 (46) 25 (47) —
4 2 (8) 2 (18) 11 (21) —
5 6 (24) 3 (27) 0 —
Inversion strength — — — < 0.0001‡
0-2 0 0 18 (31%) —
3 1 (3) 3 (23) 14 (24) —
4 18 (60) 7 (54) 20 (35) —
5 11 (37) 3 (23) 6 (10) —
Categorical variables are presented as N (%).
Continuous variables are presented as mean (SD).
Significance set at P < 0.05 with Bonferroni correction.
*Significant difference IA versus Ponseti (P < 0.0001), IA versus EA (P = 0.01).
†Significant difference IA versus Ponseti (P < 0.0001), IA versus EA (P = 0.03).
‡Significant difference IA versus Ponseti (P < 0.0001), IA versus EA (P = 0.02).

66 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
The average age when presenting with late relapse
tended to be younger for the Ponseti group (8.2 y) com-
pared with the IA (9.0 y) and EA (10.7 y) groups, but the
difference was not statistically significant (P = 0.17)
(Table 1).
The IA group had significantly greater passive dor-
siflexion with the knee flexed and extended than the other
2 groups (averaging 8 degrees knee flexed, 5 degrees knee
extended; P ≤ 0.002). The Ponseti and EA groups tended
to have mild plantarflexion contractures (averaging −1°/
−4° for Ponseti, −4°/−5° for EA, knee flexed/extended).
The IA group tended to have a higher proportion of limbs
with severe plantarflexor weakness (grades 0-2/5) and had
no limbs grade 5 plantarflexor strength as compared with
the other 2 groups, though this difference was not statis-
tically significant. The IA group has greater inversion
weakness than the Ponseti and EA groups (P ≤ 0.0001).
There was no difference in these variables between the EA
and Ponseti groups (Table 1). Inversion weakness was
more common in limbs that underwent PMR than
posterior release (P < 0.04) (Table 2).
Gait kinematics and kinetics differed significantly
between groups. The Ponseti group had less external hip
rotation throughout the gait cycle than the IA and EA
groups (P = 0.003). The IA group had greater maximum
ankle dorsiflexion in stance (P = 0.001) and less power
production at push-off than both the Ponseti and EA
groups (P = 0.009). The IA group had a higher cadence
and shorter gait cycle time than the Ponseti group
(P ≤ 0.04) (Table 3).
DISCUSSION
The current study illustrated that late clubfoot re-
lapse can occur regardless of the initial treatment ap-
proach (Ponseti vs. casting and PMR). Our IA group was
larger than the Ponseti and EA groups, likely reflecting the
fact that this group was primarily treated before the
widespread adoption of the Ponseti method of clubfoot
correction at our institution. Although relapse is reported
to be rare over the age of 5 years, studies have shown that
the rate of clubfoot relapse increases with time, under-
scoring the importance of follow-up until skeletal maturity
in these patients to allow identification and treatment once
TABLE 2. Ankle Plantarflexion and Inversion Strength
Comparison Between Confirmed Posterior Versus
Posteromedial Release (PMR)
Posterior Release (n = 8 Feet), PMR (n = 31 Feet),
n (%) n (%)
Plantarflexion strength
0-2 1 (13) 10 (40)
3 3 (37) 13 (52)
4 4 (50) 2 (8)
5 0 0 —
Inversion strength
0-2 0 12 (39)
3 1 (14) 8 (26)
4 4 (57) 8 (26)
5 2 (29) 3 (10)
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Gait Analysis Characteristics in Relapsed Clubfoot
relapse develops.17 The average age at which our patients
sought treatment for late relapse was 8.5 to 10.7 years.
Loof et al18 found relapse between 5 and 10 years of age in
20% of patients. A study by Jeans et al4 included some
patients with relapse through age 10.
Our results show IA surgery results in greater dor-
siflexion passively and dynamically and decreased power
at push-off, owing to greater weakness in the ankle plan-
tarflexors and invertors. Plantarflexor weakness and poor
push-off power place extra demand on the knee and hip
extensor musculature during gait and other functional
activities.19,20 The long-term effects of this are unknown.
However, research has shown more pain in patients
treated with IA surgery compared with Ponseti
casting.2,19,21 IA surgeries typically result in stiffness and
decreased plantarflexion, inversion and eversion range of
motion as compared with Ponseti casting.2,19,22 This may
have also been the case in the current study, though we
focused only on the dorsiflexion range of motion.
In contrast, the Ponseti and EA groups in the current
study had mild plantarflexion contractures but better
strength preservation than the IA group. The study by
Jeans et al is the only one in the literature comparing the
same 3 groups, and found that clubfeet treated with EA
surgery and Ponseti casting alone were more similar to
those of typically developing children than those treated
with IA surgery.4 In the current study, plantarflexor
strength was rated 0 to 2 in 32% of feet in the IA group
compared with 9% to 11% in the other 2 groups. Similarly,
kinetic gait data revealed that maximum push-off power
was more than 50% greater in both the Ponseti and EA
groups than in the IA group. The current study also
demonstrated marked weakness in inversion in the IA
group compared with the Ponseti and EA groups, with 0
to 2 strength in 32% of the feet in the IA group compared
with 0% in both the Ponseti and EA groups. Among feet
whose IA surgery type (posterior release vs. PMR) was
confirmed by operative reports, the inversion strength of <
3/5 was seen in 20/31 (65%) of PMR-treated feet, and 1
foot (14%) of those treated with posterior release. This
significant inversion weakness may be 1 of the reasons that
overcorrection is seen in patients following comprehensive
clubfoot release.
Plantarflexor tightness in the Ponseti and EA
groups, as compared with the IA group, is likely related to
the fact that these feet had not undergone the extensive
release of posterior structures that feet in the IA group
had. Plantarflexion contracture is 1 of the most common
initial signs of clubfoot relapse, and it is possible that this
P was the case for patients in our Ponseti and EA groups. It
could be argued that it is preferable to have recurrent
0.13 plantarflexion, which can be treated with nonoperative or
— operative treatment, rather than a calcaneus deformity,
—
which is difficult, if not impossible, to correct and requires
more extensive subsequent surgery, often with poor out-
0.04 comes.
— We did not attempt to describe or characterize the
— various components of our patients’ relapsed deformities.
—
As noted above, the Ponseti and EA groups had mild
www.pedorthopaedics.com | 67
Bent et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 3. Comparison of Gait Parameters Between Groups

Ponseti Extra-articular Intra-articular
Gait N = 35 Feet, n (%) N = 13 Feet, n (%) N = 61 Feet, n (%) P
Max hip extension in stance (°) −9.5 (6.3) −5.2 (6.5) −6.0 (7.2) 0.28
Max knee extension in stance (°) −0.20 (5.4) −1.5 (6.3) 0.13 (5.6) 0.63
Max knee moment in stance (°) 0.32 (0.20) 0.29 (0.21) 0.39 (0.25) 0.48
Min knee moment in stance (°) −0.40 (0.14) −0.38 (0.10) −0.31 (0.17) 0.07†
Ave pelvic rotation in stance (°) −1.1 (4.7) 0.32 (3.7) −1.8 (5.2) 0.35
Ave hip rotation in stance (°) −0.9 (4.6) −6.0 (9.3) −7.6 (8.0) 0.003‡
Ave foot progression in stance(°) 1.2 (10.6) 3.2 (10.9) 4.0 (18.3) 0.71
Max dorsiflexion (°) 6.3 (10.6) 9.9 (8.5) 13.8 (7.2) 0.004§
Min dorsiflexion (°) −16.2 (10.9) −14.8 (12.7) −9.0 (10.8) 0.11
Max plantarflexor moment (Nm/kg) 0.98 (0.20) 0.93 (0.36) 0.82 (0.35) 0.18
Max ankle push-off power (W/kg) 2.5 (1.0) 2.6 (1.4) 1.71 (0.90) 0.009‖
N = 24 subjects N = 7 subjects N = 39 subjects —
Velocity* 88.1 (11.3) 84.6 (8.4) 84.6 (15.7) 0.88
Cadence* 99 (11.6) 106 (9) 108 (12) 0.06
Stride length* 89.7 (9.8) 79.8 (7.9) 85.4 (15.7) 0.33
Gait cycle time* 104.8 (17.3) 95 (8.5) 94 (10) 0.05
Double limb stance time* 106.7 (32.6) 105.8 (10.7) 108.3 (18.6) 0.97
Data are presented as mean (SD).
Negative values indicate extension/hyperextension, external rotation/foot progression, plantarflexion, and a flexion net internal joint moment.
Significance set at P < 0.05 with Bonferroni correction.
*% normal for age.
†Significant difference IA versus Ponseti (P < 0.03).
‡Significant difference IA versus Ponseti (P = 0.001), IA versus EA (P = 0.03).
§Significant difference IA versus Ponseti (P = 0.001).
‖Significant difference IA versus Ponseti (P < 0.005), IA versus EA (P = 0.03).

plantarflexion contractures. The IA group had greater
external hip rotation than the other 2 groups, which may
be related to ipsilateral metatarsus adductus or internal
tibial torsion. Future study is recommended to explore the
nature of relapse in the 3 groups in more detail.
Limitations include the fact that the study was ret-
rospective and the groups were not randomized to treat-
ments. The treatment provided was influenced by the era
in which the patient was seen. Since we do not do routine
gait analysis tests on all clubfoot patients at our facility, all
of the patients in the current study were seen for evalua-
tion of relapsed deformity. Therefore, we were not able to
assess rates of relapse among patients seen at our hospital.
We were also unable to determine exactly when relapses
occurred. They could have occurred earlier but only be-
come intolerable at the time the patient and family sought
treatment. It is possible that patients in the IA and EA
groups had more severe or resistant deformities than the
Ponseti group. Dimeglio classification values were not
available for any of the study subjects, so we were unable
to control for clubfoot severity. Lastly, since we did not
see patients after their prior clubfoot correction proce-
dures, we relied on a review of medical records and parent
reports to classify subjects as having a relapse. There is a
possibility that some patients may have had an under-
correction rather than a true relapse.
In conclusion, our findings indicate that late relapse
can occur after all types of clubfoot correction. Consistent
with existing literature, patients who have undergone
posteromedial release surgery have significantly greater
plantarflexor and inverter weakness resulting in poorer
plantarflexor moment and power production during gait.
The next steps will include the characterization of out-
comes following treatment for late clubfoot relapse.
ACKNOWLEDGMENTS
The authors thank Wendy Mack for her assistance
with the statistical analysis.
REFERENCES
1. Smythe T, Kuper H, Macleod D, et al. Birth prevalence of congenital
talipes equinovarus in low- and middle-income countries: a system-
atic review and meta-analysis. Trop Med Int Health. 2017;22:
269–285.
2. Smith PA, Kuo KN, Graf AN, et al. Long-term results of
comprehensive clubfoot release versus the Ponseti method: which is
better? Clin Orthop Relat Res. 2014;472:1281–1290.
3. Recordon JAF, Halanski MA, Boocock MG, et al. A prospective,
median 15-Year comparison of ponseti casting and surgical treat-
ment of clubfoot. J Bone Joint Surg Am. 2021;103:1986–95.
4. Jeans KA, Karol LA, Erdman AL, et al. Functional outcomes
following treatment for clubfoot: Ten-year follow-up. J Bone Joint
Surg Am. 2018;100:2015–2023.
5. Ippolito E, Farsetti P, Caterini R, et al. Long-term comparative
results in patients with congenital clubfoot treated with two different
protocols. J Bone Joint Surg Am. 2003;85:1286–1294.
6. Hosseinzadeh P, Kiebzak GM, Dolan L, et al. Management of
clubfoot relapses with the ponseti method: Results of a survey of the
POSNA members. J Pediatr Orthop. 2019;39:38–41.
7. Hu W, Ke B, Niansu X, et al. Factors associated with the relapse in
Ponseti treated congenital clubfoot. BMC Musculoskelet Disord.
2022;23:88.
8. Zhao D, Li H, Zhao L, et al. Prognosticating factors of relapse in
clubfoot management by ponseti method. J Pediatr Orthop. 2018;38:
514–520.
9. Cosma DI, Corbu A, Nistor DV, et al. Joint hyperlaxity prevents
relapses in clubfeet treated by Ponseti method-preliminary results. Int
Orthop. 2018;42:2437–2442.

68 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
10. Zionts LE, Ebramzadeh E, Morgan RD, et al. Sixty Years on: ponseti
method for clubfoot treatment produces high satisfaction despite
inherent tendency to relapse. J Bone Joint Surg Am. 2018;100:721–728.
11. Dobbs MB, Rudzki JR, Purcell DB, et al. Factors predictive of
outcome after use of the Ponseti method for the treatment of
idiopathic clubfeet. J Bone Joint Surg Am. 2004;86:22–27.
12. Bhaskar A, Patni P. Classification of relapse pattern in clubfoot
treated with Ponseti technique. Indian J Orthop. 2013;47:370–376.
13. Laaveg SJ, Ponseti IV. Long-term results of treatment of congenital
club foot. J Bone Joint Surg Am. 1980;62:23–31.
14. Dobbs MB, Corley CL, Morcuende JA, et al. Late recurrence of
clubfoot deformity: a 45-year followup. Clin Orthop Relat Res.
2003;411:188–192.
15. Grin L, van der Steen MC, Wijnands SDN, et al. Forefoot adduction
and forefoot supination as kinematic indicators of relapse clubfoot.
Gait Posture. 2021;90:415–421.
16. Brown HP, R Prescott. Applied Mixed Models in Medicine, 3rd ed.
John Wiley & Sons, Inc.; 2015. 2000-2022.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Gait Analysis Characteristics in Relapsed Clubfoot
17. Thomas HM, Sangiorgio SN, Ebramzadeh E, et al. Relapse rates in
patients with clubfoot treated using the ponseti method increase with
time: a systematic review. JBJS Rev. 2019;7:e6.
18. Loof E, Andriesse H, Brostrom E. The gait pattern might predict
children at risk for late relapse of clubfoot. Gait Posture. 2019;73:
179–180.
19. Graf A, Hassani S, Krzak J, et al. Long-term outcome evaluation in
young adults following clubfoot surgical release. J Pediatr Orthop.
2010;30:379–385.
20. Alkjaer T, Pedersen EN, Simonsen EB. Evaluation of the walking
pattern in clubfoot patients who received early intensive treatment.
J Pediatr Orthop. 2000;20:642–647.
21. Graf AN, Kuo KN, Kurapati NT, et al. A long-term follow-up of
young adults with idiopathic clubfoot: Does foot morphology relate
to pain? J Pediatr Orthop. 2019;39:527–533.
22. Church C, Coplan JA, Poljak D, et al. A comprehensive outcome
comparison of surgical and Ponseti clubfoot treatments with
reference to pediatric norms. J Child Orthop. 2012;6:51–59.
www.pedorthopaedics.com | 69
 ORIGINAL ARTICLE

Is the Integration Problem in the Sensoriomotor System

the Cause of Adolescent Idiopathic Scoliosis?
Ahmet Payas, PhD,* Sabri Batın, MD,† Erdal Kurtoğlu, PhD,‡ Mustafa Arık, MD,§
Turgut Seber, MD,∥ İlyas Uçar, PhD,‡ and Erdoğan Unur, PhD‡

Purpose: The reason behind the balance control disorder seen in
adolescent idiopathic scoliosis (AIS) has been suggested as a cen-
tral nervous system dysfunction, yet it has not been investigated in
detail whether this problem originates from sensory, motor, or
from both systems. This study aimed to reveal the differences in the
pathways that provide proprioceptive sense, motor control, and
coordination between these 2 systems in female individuals
with AIS.
Methods: Brain Diffusion Tensor Imaging was applied to 30
healthy individuals and 30 Lenke type 1 AIS patients. All of the
individuals included in the study were predominantly right-
handed and aged between 10 and 18. Diffusion tensor imaging of
both groups were performed bilateral tractography on the cor-
ticospinal tract (CS tr), medial lemniscus (ML), superior longi-
tudinal fasciculus (SLF), and inferior longitudinal fasciculus
(ILF) tracts using DSI Studio software.
Results: Significant differences in the parameters of CS tr, ML,
SLF, ILF pathways were found between the AIS and the control
groups. In the AIS group, significant differences were found in
the fiber count and fiber ratio of the ML that carries the pro-
prioceptive sense and CS tr, which is responsible for the soma-
tomotor system. There were also significant differences between
the left and right CS tr, ML, SLF, and ILF pathways of the AIS
group (P < 0.05).
Conclusions: Differences in the CS tr, ML, SLF, and ILF pathways
may trigger muscular asymmetry and cause postural instability and
thus spinal deformity in AIS.
Key Words: idiopathic scoliosis, tractography, sensorimotor
system, anomaly
(J Pediatr Orthop 2023;43:e111–e119)
A dolescent idiopathic scoliosis (AIS) is a 3-dimensional
spinal deformity of unknown etiology. It is known that
this phenomenon is more common in girls aged 11 to
18 years compared with boys.1 In addition to spinal de-
formity, many factors such as postural asymmetry, pro-
prioceptive sensation, and postural instability are seen in
individuals with AIS.2 One of the common problems in
individuals with AIS is postural control problem. To pro-
vide a postural control, the center of gravity must be kept
within the boundaries of the support surface.3 For postural
control, the stimuli coming through the sensory, visual,

TABLE 1. Descriptive Statistics of the Participants

Variable Participants With AIS Asymptomatic Participants P
Age 14.7 ± 1.9 14.9 ± 2.0 0.95
Body mass index (kg/m2) 20.4 ± 1.3 20.0 ± 2.1 0.78
Female, n (%) 29 (100) 29 (100) 1.00
Mean angle of the major curve 46.6 ± 4.4 — —
Risser sign 1.7 ± 1.4 1.7 ± 1.3 0.63
Lenke curve type 1, n (%) 29 (100) — —
Dominant side upper (R), n (%) 29 (100) 29 (100) 1.00
Dominant side lower (R), n (%) 29 (100) 29 (100) 1.00
Independent samples t test or χ2 test.
AIS indicates adolescent idiopathic scoliosis; R, right.

From the *Vocational College of Sungurlu, Hitit University, Çorum; Departments of †Orthopedics and Traumatology; ∥Radiology, Kayseri City
Education and Training Hospital; ‡Department of Anatomy, Faculty of Medicine, Erciyes University, Kayseri; and §Orthopedics and Traumatology
Department, Aksaray Ortaköy Public Hospital, Aksaray, Turkey.
A.P.: study design, performed measurements, and manuscript preparation; S.B. and M.A.: evaluation of patients and manuscript preparation; E.K.:
statistical analysis and manuscript preparation; M.A.: evaluation of patients and manuscript preparation; T.S.: performed measurements and
manuscript preparation; İ.U. and E.U.: study design and manuscript preparation.
The authors declare no conflicts of interest.
Reprints: Ahmet Payas, PhD, Vocational School, Hitit University Sungurlu, 12027, Street No: 3 Sungurlu, Corum, Turkey. E-mail: ahmetpayas@hitit.edu.tr.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002300

J Pediatr Orthop  Volume 43, Number 2, February 2023 www.pedorthopaedics.com | e111

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Payas et al J Pediatr Orthop  Volume 43, Number 2, February 2023

auditory, and proprioceptive systems are evaluated in the

TABLE 2. Comparison of Tractography Values of
Corticospinal Tract central nervous system and form the motor responses nec-
essary for appropriate postural control.4,5 Disturbance in
Control Group (Data of Corticospinal Tract)
any of these systems can cause postural deficiencies and
Left Right Sig. (P) thus scoliosis. In the central nervous system, the commu-
CS tr nication between sensory, visual, auditory, and proprio-
Fiber count 5522.74 ± 311.86 5291.03 ± 303.94 0.285
Mean fiber length (mm) 119.61 ± 1.29 122.04 ± 1.00 0.057 ceptive systems is provided by the white matter (WM). WM
FA 0.52 ± 0.01 0.52 ± 0.01 0.299 consists of axons surrounded by a myelin sheath. Axons are
Fiber ratio 6.46 ± 0.36 6.19 ± 0.35 0.073 responsible for providing fast and efficient connections
AIS group (data of corticospinal tract)
Left Right Sig. (P)
between the cortex and subcortical regions.
CS tr Magnetic resonance imaging (MRI) is a noninvasive
Fiber count 5057.16 ± 278.71 4413.13 ± 340.24 *0.028 imaging technology that produces detailed 3-dimensional
Mean fiber length (mm) 118.33 ± 1.11 121.28 ± 1.16 **0.002 anatomic images using magnets and radio frequency
FA 0.52 ± 0.01 0.53 ± 0.01 0.103
Fiber ratio 5.91 ± 0.32 5.16 ± 0.39 *0.028
waves.6 Diffusion tensor imaging (DTI)—a technique
Between control and AIS groups (data of corticospinal tract) developed from the traditional MRI method—provides
Control AIS Sig. (P) information about the tissue structure by measuring dif-
Left CS tr fusion rates and paths of water molecules. Thereby, it
Fiber count 5522.74 ± 311.86 5057.16 ± 278.71 0.269
Mean fiber length (mm) 119.61 ± 1.29 118.33 ± 1.11 0.457
noninvasively reconstructs the WM fiber tracts in the
FA 0.52 ± 0.01 0.52 ± 0.01 0.299 human brain in vivo.7 DTI defines the geometric proper-
Fiber ratio 6.46 ± 0.36 5.91 ± 0.32 0.264 ties of WM tracts such as fiber length, fiber volume, and
Right CS tr fiber count and also the morphologic features of re-
Fiber count 5291.03 ± 303.94 4413.33 ± 340.24 *0.042
Mean fiber length (mm) 122.04 ± 1.00 121.28 ± 1.16 0.628 constructed fiber paths found in the WM in general.8 DTI
FA 0.53 (0.49-0.56 0.52 80.48-0.59) 0.350 has been used to identify changes in fibers found in the
Fiber ratio 6.19 ± 0.35 5.16 ± 0.39 *0.043 WM and has become the method of choice in brain
Parametric data were presented as mean ± SEM. Paired comparison tests were research.9 In this study, researchers theorized that a
performed with the paired samples t test. defective sensory input or an abnormal sensorimotor
Nonparametric data were shown as median (minimum-maximum) and pairwise integration may cause a change in postural tone in in-
comparisons were made with the Will-Coxon test.
AIS indicates adolescent idiopathic scoliosis; CS tr, corticospinal tract; FA, dividuals with AIS. Researchers predicted that a change in
fractional anisotropy. postural tone will trigger muscular asymmetry and result
*The difference between groups is statistically significant at the 95% CI (P < 0.05).
**The difference between groups was statistically significant at the 99% CI (P < 0.01).
in spinal deformity. No comprehensive tractography study
***The difference between groups was statistically significant at the 99.9% CI examining the projection pathways and the pathways that
(P < 0.001). connect them in individuals with AIS has been encoun-
tered in the literature.

FIGURE 1. Diffusion tensor images of corticospinal tracts of control and AIS group. A, Display of fiber density of CS trs in the control
group (more fiber density, less symmetry between left and right). B, Display of fiber density of CS trs in AIS group (lower fiber density,
more symmetry between left and right). AIS indicates adolescent idiopathic scoliosis; CS tr, corticospinal tracts; L, left; R, right.

e112 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Sensorimotor System in Idiopathic Scoliosis

patients were the candidates for correction surgery and

TABLE 3. Comparison of Tractography Values of Medial
Lemniscus had a mean age of 14.910–17 and a mean angle of the major
curve was 46.6 ± 4.4 degrees. Control group was examined
Control Group (Data of Medial Lemniscus)
with a scoliometer and underwent postural analysis to
Left Right Sig. (P) exclude scoliotic deformity. History of head injury, neu-
ML rological disorder, headache, back injury, weakness or
Fiber count 2801 (1521-8026) 3560 (2419-8684) **0.008
Mean fiber length (mm) 135.68 ± 1.18 131.50 ± 1.47 ***0.000 numbness in any of the extremities, and space occupying
FA 0.46 ± 0.01 0.49 ± 0.01 ***0.000 lesions in the brain MRI were considered as exclusion
*0.010
Fiber ratio 3.88 ± 0.31 4.60 ± 0.28 criteria.
AIS group (data of medial lemniscus)
Left Right Sig. (P)
ML
Fiber count 2014 (600-4173) 3014 (970-7999) **0.002 Study Design
Mean fiber length (mm) 132.37 ± 1.17 138.96 ± 0.89 *0.025 This is a cross-sectional study compares DTI trac-
***0.000
FA 0.45 ± 0.02 0.48 ± 0.01
**0.001
tography images of 4 central nervous system pathways of
Fiber ratio 2.47 ± 0.23 3.72 ± 0.30 AIS patients with those of healthy individuals. These ways
Between control and AIS groups (data of medial lemniscus)
Control AIS Sig. (P) are as follows: (1) corticospinal tracts (CS tr): pathways
Left ML responsible for voluntary motor movements. (2) Medial
**0.002
Fiber count 2801 (1521-8026) 2014 (600-4173) lemniscus (ML): The proprioception and somatosensation
Mean fiber length (mm) 135.68 ± 1.18 132.37 ± 1.17 *0.039
pathway from the skin, joints, and ligaments. (3) Superior
FA 0.46 ± 0.02 0.45 ± 0.02 0.158
Fiber ratio 3.88 ± 0.31 2.47 ± 0.23 **0.001 longitudinal fasciculus (SLF) tracts connecting the sensory
Right ML and motor areas of the same hemisphere. (4) Inferior
Fiber count 3560 (2419-8684) 3014 (970-7999) **0.025 longitudinal fasciculus (ILF) tracts connecting the occipital
Mean fiber length (mm) 131.50 ± 1.47 129.96 ± 0.89 0.365 and temporal lobes.
FA 0.49 ± 0.01 0.48 ± 0.01 0.223
**0.023 During the tractography procedure, the total fiber
Fiber ratio 4.14 (2.87-9.97) 3.58 (1.15-9.35)
number of the tracts, the average fiber length in milli-
Parametric data were presented as mean ± SEM. Paired comparison tests were meters, the fiber ratio in the whole brain of the same in-
performed with the paired samples t test.
Nonparametric data were shown as median (minimum-maximum) and pairwise
dividual, and the fractional anisotropy values were
comparisons were made with the Will-Coxon Test. calculated. These data were compared as right and left
AIS indicates adolescent idiopathic scoliosis; FA, fractional anisotropy; ML, within the group and between the groups on the same side
medial lemniscus.
*The difference between groups is statistically significant at the 95% CI of the pathways.
(P < 0.05).
**The difference between groups was statistically significant at the 99% CI
(P < 0.01).
***The difference between groups was statistically significant at the 99.9% CI Data Acquisition
(P < 0.001). The MRI procedures were performed in the Radiology
Unit of Kayseri City Hospital with a 3T (Tesla) MRI device
(Siemens Magnetom Skyra, the Netherlands). DTI allows us
In this study, researchers examined anomalies in to study the brain structure, especially WM, by measuring
the descending, ascending, and interhemispheric path- the random movements of water molecules in the brain. To
ways in individuals with AIS by using magnetic calculate the diffusion tensor, it is necessary to take a dif-
resonance–based DTI tractography and aimed to shed fusion-weighted image in at least 6 different directions and
light on the pathogenesis of AIS. an additional reference image without diffusion magnetic
field change (b = 0). In practice, more than 6 directions of
METHODS diffusion-weighted images are obtained and the signals are
This single-center, randomized control, prospective, calculated tensor by mathematical operations. With this
and cross-sectional cohort study was conducted at Kayseri technique, in which directions and how much restricted the
Education and Research Hospital, Department of Ortho- diffusion of free water protons in the WM of the brain is
pedics and Traumatology in accordance with the principles determined. In this way, it gives important information
of the Declaration of Helsinki. Written informed consent about the microstructural organization such as the integrity
was obtained from each patient and/or his/her legal and orientation of the WM tracts. DTI sequence was a twice-
guardian. And also, this study was approved by the Kayseri refocused spin-echo sequence based on single-shot echo-
Local Ethics Committee with decision # 2020/161. planar acquisition. Diffusion sensitizing gradients were
applied along 20 orthogonal directions using 2 b values (0
Patient Selection and 1000 s/mm2) and other DTI parameters were TR =
This clinical diagnostic research was designed on 30 4900 ms, TE = 95 ms, number-of-slice = 36, flip angle = 90
female Lenke type 1 (right thoracic) AIS patients and 30 degrees, FOV = 230 × 230 mm2, matrix = 128 × 128, and
healthy female individuals of the same age population. slice thickness = 3.5 mm (voxel size 1.8 × 1.8 × 3.5 mm). The
Individuals in both the AIS and control groups used their acquisition time per data set was ~6 minutes. The original
right hand as the dominant hand. Family members of all raw data were anonymized and transferred from the scanner
individuals included in the study did not have AIS. AIS to the DICOM format.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e113

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Payas et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 2. Diffusion tensor images of medial lemniscus of control and AIS group. A, Display of fiber density of MLs in the control
group (more fiber density, less symmetry between left and right). B, Display of fiber density of MLs in AIS group (lower fiber density,
much more symmetry between left and right). AIS indicates adolescent idiopathic scoliosis; ML, medial lemniscus; L, left; R, right.

Data Processing nonparametric comparisons. Statistical analysis of this

The tractography process was made by the DSI
Studio software, which was downloaded from http://dsi-
studio.labsolver.org/. Before starting the tractography
process, in the “Fiber tracking” tab, following adjustments
were applied: “Threshold” 0.20, “Angular Threshold” 70
degrees, “Smoothing” 0.50, “shortest track” 10 mm,
“longest track” 1000 mm, “terminate if” 100,000 fibers.
These parameters were adjusted again before starting the
tractography process on each new image. In our study,
bilateral CS tr, ML, SLF, ILF, and PCI pathways were
examined. The fibers in the whole brain were counted to
calculate the ratio of the fibers in these pathways. To
minimize the margin of error in the tractography process,
the relevant regions were selected from the “Tractography
Atlas” in the DSI Studio software. Tractography proce-
dures were calculated separately by a radiologist (S.T.)
and 2 anatomists (P.A. and K.E.), and the mean of 3
values was used in statistical analysis.
Statistical Analysis
In this study, the normal distribution analysis of
continuous data was performed using 5 parameters
(Skewness and Kurtosis, Histogram, MEAN/STD, Q-Q
plots, Shapiro-Wilk test). Normally distributed continuous
data were shown as mean ± SEM and independent sam-
ples t test was used for comparisons between pairs. Con-
tinuous data that were not normally distributed were
shown as Median (minimum-maximum) and comparisons
between pairs were made with Mann-Whitney U test.
Paired samples t test was used for parametric variables in
comparison between right and left hemispheres of the
same individuals and Wilcoxon sign test was used for
study was done with IBM SPSS 23.0 software.
RESULTS
A total of 60 participants with (n = 30) and without
AIS (n = 30) with a major thoracic curve (41 to 50 degrees)
were included in the study. Descriptive characteristics of
the participants are presented in Table 1. There was no
significant difference between the 2 groups in terms of
baseline characteristics (P > 0.05).
Tractography Data of Tractus Corticospinalis
Values of the left CS tr were found to be high in both
groups. The fiber count and the fiber ratio between left
and right CS trs were 4.2% and 4.2%, respectively, in the
control group. The difference between the left and right
CS trs of the AIS group was approximately three times of
the control group (12.8% and 12.7%, respectively)
(Table 2).
When CS trs of the groups are compared, the fiber
count and the fiber ratio values of left and right CS trs of
the control group were higher than those of the AIS group.
The difference in fiber count and fiber ratio between the
control and AIS groups was 8.4% and 8.5% on the left,
and 16.6% and 16.2% on the right, respectively (Fig. 1).
However, only the fiber count and fiber ratio data of the
right CS tr were statistically significant among the groups
(P < 0.05).
Tractography Data of ML
Right ML values were found to be large in both the
groups. The ratios between the fiber count and fiber ratio
of the right and left ML were 21.3% and 19.7% in the

e114 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Sensorimotor System in Idiopathic Scoliosis

When the same-party SLFs between groups were

TABLE 4. Comparison of Tractography Values of Superior
Longitudinal Fasciculus compared, left SLF was greater in the AIS group, and
right SLF was greater in the control group. Although the
Control Group (Data of Superior Longitudinal Fasciculus)
difference in fiber count and fiber ratio between the groups
Left Right Sig. (P) was 7.3% and 7.2% in the left SLF, it was calculated as
SLF 15.9% and 17.8% in the right SLF (Fig. 3). Only the fiber
Fiber count 883.18 ± 79.21 981.59 ± 63.89 0.320
Mean fiber length 525.45 ± 17.81 447.98 ± 19.27 **0.002 count and fiber ratio data of the right SLF were found to
(mm) be significant among the groups (P < 0.05).
FA 0.37 ± 0.01 0.35 ± 0.01 *0.048
Fiber ratio 1.03 ± 0.09 1.14 ± 0.07 0.335 Tractography Data of ILF
AIS group (data of superior longitudinal fasciculus) Right ILF values were found to be large in both the
SOL SAĞ Sig. (P) groups. Although the ratio between fiber count and fiber
SLF
Fiber count 953.56 ± 74.23 825.83 ± 53.73 0.145 ratio of ILFs was 19.9% and 19.7% in the control group, it
Mean fiber length 522.41 ± 17.39 436.08 ± 16.22 ***0 was calculated as 35.0% and 35.9% in the AIS group
(mm) (Table 5).
FA 0.35 ± 0.01 0.35 ± 0.01 0.273 Both left and right ILF values were higher in the AIS
Fiber ratio 1.11 ± 0.08 0.96 ± 0.06 0.141
Between control and AIS groups (data of superior longitudinal
group. The difference in fiber count and fiber ratio be-
fasciculus) tween the groups was 8.4% and 7.1% in left ILF, whereas
Control AIS Sig. (P) 25.7% and 25.9% in right ILF, respectively (Fig. 4). Only
Left SLF the mean fiber length of the right ILF was significant
Fiber count 883.18 ± 79.21 953.56 ± 74.23 0.519 between the groups.
Mean fiber length 525.45 (341.88- 522.41 (348.09- 0.835
(mm) 678.33) 668.94)
FA 0.37 ± 0.01 0.35 ± 0.01 0.124 DISCUSSION
Fiber ratio 1.03 (0.35-2.69) 1.11 (0.35-2.24) 0.420 It is known that the curvature with AIS is caused by
Right SLF
Fiber count 981.59 ± 63.89 825.83 ± 53.73 *0.046 the asymmetry between the trunk muscles.18 Our hy-
Mean fiber length 447.98 (240.86- 436.08 (8214.5- 0.811 pothesis is that this muscle asymmetry in individuals with
(mm) 760.92) 594.3) AIS is caused by the inability to provide the required
FA 0.35 ± 0.01 0.35 ± 0.01 0.437 muscle tone by the somatomotor system due to sensory
Fiber ratio 1.14 ± 0.07 0.96 ± 0.06 *0.046
input or sensorimotor integration problems. Visual, au-
Parametric data were presented as mean ± SEM. Paired comparison tests were ditory, and proprioceptive senses are very important for
performed with the paired samples t test. maintaining balance and posture. In-line with the in-
Nonparametric data were shown as median (minimum-maximum) and pairwise
comparisons were made with the Will-Coxon test. formation coming from these senses, a suitable motor re-
AIS indicates adolescent idiopathic scoliosis; FA, fractional anisotropy; SLF, sponse is formed to ensure posture and balance. That is
superior longitudinal fasciculus.
*The difference between groups is statistically significant at the 95% CI
why the connections between the sensory and motor cen-
(P < 0.05). ters of the brain are so important.6,18 Although there are
**The difference between groups was statistically significant at the 99% CI multiple pathways connecting the sensory and motor re-
(P < 0.01).
***The difference between groups was statistically significant at the 99.9% CI gions of the brain, the most important ones are SLF and
(P < 0.001). ILF.10 Although the proprioceptive sense is carried to the
brain by the ML, the motor response is delivered to the
muscle related to the CS tr.11 SLF connects the frontal,
control group, whereas it was 33.2% and 33.6% in the AIS parietal, temporal, and occipital lobes of the same side.
group (Table 3). Therefore, it has an important role in providing commu-
When comparison was made between the groups, it nication between the visual, auditory, and proprioceptive
was seen that both left and right ML values were greater in senses and the motor regions. Most of the ILF connects
the control group. Although the difference between the the occipital and temporal lobes in the ipsilateral hemi-
groups considering the fiber count and fiber ratio was spherium cerebri, and in some fibers, it extends to the
28.1% and 36.4% of the left ML, and 15.3% and 13.5% in frontal lobe. In this way, it acts as a bridge between the
the right ML, respectively (Fig. 2). Differences attributed visual and auditory senses and the motor region. ML
to left and right ML values were found to be significant detects proprioceptive, vibration, contact, and pressure
between the groups (P < 0.05). senses related to our position from the muscle, tendon,
ligament, and joint capsule and carries them to the gyrus
postcentralis. Most of the fibers that make up CS tr leave
Tractography Data of SLF the motor area of the brain and extend along the entire
Although the right SLF values were greater in the spinal cord. Then they leave the spinal cord from the
control group, the left SLF values were greater in the AIS anterior roots and carry voluntary motor commands to
group. The difference between fiber count and fiber ratio of the skeletal muscles.12 The volume and fiber number of
SLFs was 10.0% and 9.6% in the control group, whereas it these pathways in WM increases significantly during
was 13.4% and 13.5% in the AIS group, respectively childhood and adolescence for attention, motor skills,
(Table 4). cognitive ability, and memory. However, this typical de-

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e115

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Payas et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 3. Diffusion tensor images of superior longitudinal fasciculus of control and AIS group. A, Display of fiber density of SFLs in
control group (more fiber density, less symmetry between left and right). B, Display of fiber density of SFLs in AIS group (lower fiber
density, more symmetry between left and right). AIS indicates adolescent idiopathic scoliosis; L, left; R, right; SFL, superior
longitudinal fasciculus.

velopmental trajectory may change in individuals with
developmental, cognitive, and behavioral disorders.13
As far as we know, there is no comprehensive trac-
tography study that examines the projection of pathways
and their connections in AIS patients. For this reason, the
ML that carries proprioceptive sense, SLF, and ILF
pathways, which provide integration between CS tr and
these pathways, were analyzed by tractography in in-
dividuals with AIS and in healthy individuals. In our
study, we detected differences in CS tr, ML, SLF, and ILF
pathways between the control group and the AIS group.
In addition, there were differences between the left and
right pathways of the AIS group.
In the previous studies, it has been reported that the
number of fibers in the brain pathways decreases because
of certain diseases or with aging and these situations affect
brain functions.14,15 Studies have shown that the number
of fibers in the pathways of different brain regions is
closely related to the quality of the connection between the
brain and peripheral structures.16 The decrease in reflexes
and motor movements in elderly individuals is due to
changes in the WM structure of the brain.17 Therefore, we
can say that the number of fibers in the pathways forming
the WM is directly related to motor activity.
Geissele et al19 reported that there was asymmetry in
the ventral part of the pons in the CS tr of individuals with
AIS, Goldberg et al20 reported that individuals with AIS
were more lateralized in their CS tr than healthy in-
dividuals, whereas Shi et al21 reported that the WM
structure of the capsula interna was weaker in individuals
with AIS. Domenech et al22 found that cortico-cortical
inhibition was greatly reduced on the concave side of in-
dividuals with AIS. It is known that individuals with
scoliosis have bilateral muscle weakness compared with
healthy individuals.23 In this study, left and right CS tr
values were found to be lower in AIS than in the control
group. We think that this difference we detected in CS trs
will play an important role in explaining the muscle
weakness seen in scoliosis.
In the AIS group, it was determined that there was a
difference between the left and right CS tr. It was de-
termined that there was no difference between the CS tr of
the control group. In addition, it was observed that the
difference between the fiber count and fiber ratio values of
the CS tr of the AIS group was more than 3 times the
difference in the control group. On the basis of these data,
we think that the asymmetry between the left and right CS
tr in individuals with AIS causes muscular imbalance and
that spinal deformity may have developed because of this
situation.
They reported that muscle fibers were larger on the
convex side of the curve in individuals with AIS, but in-
dividuals with AIS had lower paraspinal muscle fibers on
both sides than healthy individuals.24 In our study, there
was no significant difference between the left CS tr (pro-
viding motor control of the convex side of the curve) of
individuals with AIS and the left CS tr of the control
group. However, the fiber count and fiber ratio of the right

e116 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Sensorimotor System in Idiopathic Scoliosis

for the convex side of the curvature, were lower than the
TABLE 5. Comparison of Tractography Values of Inferior
Longitudinal Fasciculus right ML in individuals with AIS.
Considering the values of the ML in the presented
Control Group (Data of Inferior Longitudinal Fasciculus)
study, we can say that the proprioceptive sense is mostly
Left Right Sig. (P) reduced on the convex side of the curve in individuals with
ILF AIS. In addition, it is seen that both right and left pro-
Fiber count 2446 (581-6959) 3054 (1077-7561) *0.016
Mean fiber length 873.22 ± 17.68 935.28 ± 7.90 **0.002 prioceptive senses are decreased in individuals with AIS
(mm) compared with healthy individuals.
FA 0.41 ± 0.01 0.43 ± 0.01 **0.001
Domenech et al27 reported that cortico-cortical in-
Fiber ratio 2.90 (0.67-8.25) 3.61 (1.25-9.11) *0.016
hibition on the concave side of individuals with AIS was
AIS group (data of inferior longitudinal fasciculus)
Left Right Sig. (P)
reduced compared with the convex side. Xue et al28 reported
ILF that individuals with AIS have changes in WM micro-
Fiber count 2671 (774-6142) 4111.50 (792- *0.013 structure as a result of abnormal brain maturation. They
7790) claimed that somatosensory dysfunction may have devel-
Mean fiber length 911.24 ± 17.25 978.76 ± 11.05 ***0.000 oped by affecting the connection between the motor and
(mm)
FA 0.42 ± 0.01 0.44 ± 0.01 ***0.000 sensory regions of these changes in the WM. In our study, it
Fiber ratio 3.125 (0.89-7.29) 4.87 (0.92-9.14) *0.015 was observed that the fiber count and fiber ratio of the right
Between control and AIS groups (data of inferior longitudinal fasciculus) SLF responsible for the concave side of the curvature of
Control AIS Sig. (P) individuals with AIS were lower than the control group. On
Left ILF
Fiber count 2657 (581-6959) 3071 (774-6142) 0.243
the basis of these data, we think that there may be a dys-
Mean fiber length 873.22 ± 17.68 911.24 ± 17.25 0.130 function in adjusting the tone in the concave side muscles
(mm) because of the lack of sensorimotor integration between
FA 0.41 ± 0.01 0.42 ± 0.01 0.398 the sensory and motor regions in the right hemispherium
Fiber ratio 3.11 (0.67-8.25) 3.58 (0.89-7.29) 0.247 cerebri.
Right ILF
Fiber count 3470.25 ± 319.67 4221.63 ± 327.36 0.108 Visual inputs play an important role in providing a
Mean fiber length 935.28 ± 7.90 978.76 ± 11.05 **0.002 postural control in people with proprioceptive sensory
(mm) loss.29 The increase in postural sway rate when eyes are
FA 0.43 ± 0.01 0.44 ± 0.01 0.215 closed in individuals with AIS indicates insufficient ves-
Fiber ratio 4.07 ± 0.38 4.93 ± 0.38 0.117
tibular and proprioceptive inputs.30 In our study, although
Parametric data were presented as mean ± SEM. Paired comparison tests were the values of ML responsible for proprioceptive sense were
performed with the paired samples t test.
Nonparametric data were shown as median (minimum-maximum) and pairwise
lower in the AIS group, the fiber number and fiber ratio of
comparisons were made with the Will-Coxon test. ILF were found to be higher than the control group. This
AIS indicates adolescent idiopathic scoliosis; FA, fractional anisotropy; ILF, increase in ILF pathways in AIS may be an adaptation to
inferior longitudinal fasciculus.
*The difference between groups is statistically significant at the 95% CI (P < 0.05).
use the visual and vestibular systems more actively and to
**The difference between groups was statistically significant at the 99% CI (P < 0.01). boost the connection between them to compensate for the
***The difference between groups was statistically significant at the 99.9% CI lack of proprioceptive sense, which is necessary for the
(P < 0.001).
postural balance.
The small sample size and the lack of genetic testing
are among the limitations of our study. The AIS patients
CS tr (providing motor control of the concave side of the included in the study had a thoracal major curvature with a
curvature) of individuals with AIS were found to be lower mean angle of 46.6 ± 4.4 degrees (Lenke type 1 scoliosis). If
and significant than the right CS tr of the control group. there were individuals with different degrees of curvature in
These findings indicate that muscle asymmetry between our study, the correlation between WM and major curva-
the convex and concave sides will be related to the number ture could be examined. Because this study was designed for
of CS tr fibers. These findings are supported by studies. Lenke type 1 scoliosis patients for the homogeneity of the
Lao et al11 found that gait parameters on the convex data. Other types of scoliosis will need a different study
side of the curve were higher in individuals with AIS. design.
Yekutiel et al25 found that some individuals with AIS had
a weaker joint position sense at the right elbow. Ford et al
(1984) claimed that individuals with AIS have a marked
reduction in muscle spindle receptors located in their CONCLUSIONS
paraspinal muscles.26 Dabrowska et al5 reported that in- Differences in CS tr, ML, SLF, and ILF pathways
dividuals with AIS have balance disorder associated with between the healthy individuals and the AIS’ group were
inadequate functioning of the proprioceptive system. detected in this study. In addition, there were differences
In our study, it was observed that the fiber number, between the left and right pathways of the AIS group.
average fiber length, and fiber ratio values of the ML re- Muscle asymmetry in individuals with AIS occurs because
sponsible for proprioceptive sensation were lower in the of the inability to provide appropriate postural tone as a
AIS group than in the control group. In addition, it was result of defective proprioceptive sensory input or abnormal
observed that all values of the left ML, which is responsible sensorimotor integration.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e117

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Payas et al
FIGURE 4. Diffusion tensor images of inferior longitudinal fasciculus of control and AIS group. A, Display of fiber density of IFLs in
control group (more fiber density, less symmetry between left and right). B, Display of fiber density of IFLs in AIS group (more fiber
density, more symmetry between left and right). AIS indicates adolescent idiopathic scoliosis; IFL, inferior longitudinal fasciculus; L,
left, R, right.
The low values of the ML, which carries the pro-
prioceptive sense, indicate that the postural balance
problems seen in AIS are caused by proprioceptive sensory
dysfunctions in addition to muscle asymmetry.
We think that proprioceptive sense, sensorimotor in-
tegration, or somatomotor system disorders may be the
basis of spinal deformity and postural control problems
seen in individuals with AIS. In future studies, the alter-
ations of the WM structure of the brain in other Lenke
types of scoliosis (double major, triple major, etc.) can be
examined. In addition, some studies are needed to see more
clearly whether this change in the brain WM is the cause or
the result of scoliosis. For this, these pathways should be
examined in different degrees of major curvature of the
same individuals. It can be said more clearly that scoliosis
develops secondarily, if the degree of major curvature in-
creases but the anomalies in the brain pathways do not
change.
REFERENCES
1. Jada A, Mackel CE, Hwang SW, et al. Evaluation and management
of adolescent idiopathic scoliosis: a review. Neurosurg Focus.
2017;43:E2.
2. Peng Y, Wang SR, Qiu GX, et al. Research progress on the etiology
and pathogenesis of adolescent idiopathic scoliosis. Chin Med J
(Engl). 2020;133:483–493.
3. Arriaga MA. Practical Management of the Dizzy Patient (Second
Edition). Otol Neurotol. 2010;31:1.
4. Fino PC, Peterka RJ, Hullar TE, et al. Assessment and rehabilitation
of central sensory impairments for balance in mTBI using auditory
biofeedback: a randomized clinical trial. BMC Neurol. 2017;17:41.
5. Dabrowska A, Olszewska-Karaban M, Permoda-Białozorczyk A,
et al. The postural control indexes during unipodal support in patients
with idiopathic scoliosis. Biomed Res Int. 2020;2020:7936095.
6. Qiu A, Mori S, Miller MI. Diffusion tensor imaging for under-
standing brain development in early life. Annu Rev Psychol. 2015;66:
853–876.
e118 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
7. Pujol S, Wells W, Pierpaoli C, et al. The DTI challenge: toward
standardized evaluation of diffusion tensor imaging tractography for
neurosurgery. J Neuroimaging. 2015;25:875–882.
8. Jang SH, Kwon YH, Lee MY, et al. Aging of the cingulum in the
human brain: Preliminary study of a diffusion tensor imaging study.
Neurosci Lett. 2016;610:213–217.
9. Cheng H, Wang Y, Sheng J, et al. Characteristics and variability of
structural networks derived from diffusion tensor imaging. Neuro-
image. 2012;61:1153–1164.
10. Davis LE. An anatomic study of the inferıor longitudinal fasciculus.
Arch NeurPsych. 1921;5:370–381.
11. Lao ML, Chow DH, Guo X, et al. Impaired dynamic balance control
in adolescents with idiopathic scoliosis and abnormal somatosensory
evoked potentials. J Pediatr Orthop. 2008;28:846–849.
12. Gray H, Williams P, Bannister L. Gray’s anatomy. Churchill
Livingstone. 1996.
13. Barnea-Goraly N, Menon V, Eckert M, et al. White matter
development during childhood and adolescence: a cross-sectional
diffusion tensor imaging study. Cereb Cortex . 1991;15:1848–1854.
14. Wang D, Shi L, Chu WC, et al. A comparison of morphometric
techniques for studying the shape of the corpus callosum in
adolescent idiopathic scoliosis. Neuroimage. 2009;45:738–748.
15. Burzynska AZ, Preuschhof C, Bäckman L, et al. Age-related
differences in white matter microstructure: region-specific patterns
of diffusivity. Neuroimage. 2010;49:2104–2112.
16. Gao J, Cheung RT, Lee TM, et al. Possible retrogenesis observed
with fiber tracking: an anteroposterior pattern of white matter
disintegrity in normal aging and Alzheimer’s disease. J Alzheimers
Dis. 2011;26:47–58.
17. Higgen FL, Braaß H, Backhaus W, et al. Reduced frontal white
matter microstructure in healthy older adults with low tactile
recognition performance. Sci Rep. 2021;11:11689.
18. Fidler MW, Jowett RL. Muscle imbalance in the aetiology of
scoliosis. J Bone Joint Surg Br. 1976;58:200–201.
19. Geissele AE, Kransdorf MJ, Geyer CA, et al. Magnetic resonance
imaging of the brain stem in adolescent idiopathic scoliosis. Spine
(Phila Pa 1976). 1991;16:761–763.
20. Goldberg CJ, Dowling FE, Fogarty EE, et al. Adolescent
idiopathic scoliosis and cerebral asymmetry. An examination of a
nonspinal perceptual system. Spine (Phila Pa 1976). 1995;20:
1685–1691.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
21. Shi L, Wang D, Chu WC, et al. Volume-based morphometry of brain
MR images in adolescent idiopathic scoliosis and healthy control
subjects. AJNR Am J Neuroradiol. 2009;30:1302–1307.
22. Domenech J, García-Martí G, Martí-Bonmatí L, et al. Abnormal
activation of the motor cortical network in idiopathic scoliosis
demonstrated by functional MRI. Eur Spine J. 2011;20:1069–1078.
23. Kocaman H, Bek N, Kaya MH, et al. The effectiveness of two different
exercise approaches in adolescent idiopathic scoliosis: a single-blind,
randomized-controlled trial. PLoS One. 2021;16:e0249492.
24. Shahidi B, Yoo A, Farnsworth C, et al. Paraspinal muscle
morphology and composition in adolescent idiopathic scoliosis:
a histological analysis. JOR Spine. 2021;4:e1169.
25. Yekutiel M, Robin GC, Yarom R. Proprioceptive function in
children with adolescent idiopathic scoliosis. Spine (Phila Pa 1976).
1981;6:560–566.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Sensorimotor System in Idiopathic Scoliosis
26. Ford DM, Bagnall KM, McFadden KD, et al. Paraspinal muscle
imbalance in adolescent idiopathic scoliosis. Spine (Phila Pa 1976).
1984;9:373–376.
27. Domenech J, Tormos JM, Barrios C, et al. Motor cortical
hyperexcitability in idiopathic scoliosis: could focal dystonia be a
subclinical etiological factor? Eur Spine J. 2010;19:223–230.
28. Xue C, Shi L, Hui SCN, et al. Altered white matter microstructure in
the corpus callosum and its cerebral interhemispheric tracts in
adolescent idiopathic scoliosis: diffusion tensor imaging analysis.
AJNR Am J Neuroradiol. 2018;39:1177–1184.
29. Schmidt RA, Lee TD, Winstein C, et al Motor control and learning:
a behavioral emphasis (8th ed.). Human Kinetics; 2018.
30. Haumont T, Gauchard GC, Lascombes P, et al. Postural instability
in early-stage idiopathic scoliosis in adolescent girls. Spine (Phila Pa
1976). 2011;36:E847–E854.
www.pedorthopaedics.com | e119
 ORIGINAL ARTICLE
Reliability of Preoperative Supine Versus Bending
Radiographs in Estimating the Structural Nature
of Curves in EOS
Kira F. Skaggs, MD,* Nicole M. Bainton, NP,† Afrain Z. Boby, MS,* Christina C. Rymond, BA,*
Michael W. Fields, MD,‡ Benjamin D. Roye, MD, MPH,*† and Michael G. Vitale, MD, MPH*†
Background: The study sought to evaluate the utility of a single
supine radiograph in determining curve flexibility in early-onset
scoliosis (EOS) patients.
Methods: EOS patients with upright (standing/seated), supine,
and side-bending radiographs who underwent spinal deformity
surgery were identified. Coronal parameters included: proximal
thoracic (PT) curve, main thoracic (MT) curve, and thor-
acolumbar/lumbar (TL/L) curve. Each radiograph was measured
twice by 2 different raters. Correlation coefficients were utilized
to investigate associations between the different radiographs.
From the *Division of Pediatric Orthopaedic Surgery, Department of
Orthopaedic Surgery, Columbia University Irving Medical Center;
†Pediatric Orthopaedic Surgery, New York-Presbyterian Morgan
Stanley Children’s Hospital, New York, NY; and ‡Department of
Orthopaedic Surgery, Columbia University Irving Medical Center.
No funding was received to assist with the preparation of this manuscript.
This study was approved by the Columbia University Institutional Re-
view Board (Protocol AAAS9597) and was performed in accordance
with the ethical standards of the 1964 Declaration of Helsinki and its
later amendments or comparable ethical standards.
This study qualifies for a waiver of consent because it is a retrospective
chart review and does not require patient participation since all data
has already been collected during routine clinical care. There is no
potential to adversely affect the rights or welfare of subjects since this
is a chart review, with no intervention being tested. The data are
already existent in the patient’s charts, and the collection of this in-
formation does not affect previous or future care by providers.
B.D.R. has received grants from the Pediatric Orthopaedic Society of
North America and the Orthopedic Science Research Foundation.
M.G.V. has received grants from the Pediatric Orthopaedic Society of
North America, Orthopedic Science Research Foundation, Pediatric
Spine Foundation, and Setting Scoliosis Straight Foundation. He has
received royalties from Biomet. He is a paid consultant for Stryker,
Biomet, and NuVasive. M.G.V. is on the Board of Directors of the
Pediatric Spine Foundation, Pediatric Spine Study Group, and C4K.
He is the former President of the Pediatric Orthopaedic Society of
North America and is a Board Member, Chair Emeritus of the In-
ternational Pediatric Orthopaedic Symposium. The remaining au-
thors declare no conflicts of interest.
Reprints: Christina C. Rymond, BA, Department of Orthopaedic Surgery,
Morgan Stanley Children’s Hospital of New York-Presbyterian, Columbia
University Irving Medical Center, Christina Rymond, 3959 Broadway,
Chony 8-N, New York, NY 10032-3784. E-mail: christinarymond@gmail.
com.
Supplemental Digital Content is available for this article. Direct URL
citations appear in the printed text and are provided in the HTML
and PDF versions of this article on the journal’s website, www.
pedorthopaedics.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002305
70 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Interrater Correlation Coefficient (ICC) assessed intrarater and
interrater reliability.
Results: Thirty-seven EOS patients were identified (age at diag-
nosis: 7.0 ± 2.9 y, preoperative age: 13.0 ± 2.9 y; 73% female;
etiologies: 54% idiopathic, 30% syndromic, and 16% neuro-
muscular). Supine PT and MT curve measurements were highly
associated with corresponding side-bending measurements (PT:
r = 0.75, P < 0.001; MT: r = 0.80, P < 0.001), and TL/L curves
were very highly associated (TL/L: r = 0.92, P < 0.001). The
mean absolute differences between supine and side-bending
measurements were PT: 11.3 ± 7.8 degrees, MT: 14.8 ± 8.3 de-
grees, and TL/L: 16.2 ± 7.6 degrees, where the side-bending was
on average smaller than the supine measurement. The intrarater
reliabilities were excellent, with an ICC ranging from 0.93 to 0.96
for side-bending films and 0.94 to 0.97 for supine films. The
interrater reliability was excellent with ICC value of 0.88 for side-
bending films and 0.93 for supine films.
Conclusions: A single, preoperative supine radiograph was highly
predictive of side-bending radiographs in patients with EOS.
Supine curves measured an average of 15 degrees larger than
bending curves in the MT and TL/L region. A single supine film
may eliminate the need for effort-related, dual side-bending
radiographs.
Level of Evidence: Level II—retrospective study.
Key Words: scoliosis (EOS), flexibility, supine, side-bending,
radiographs
(J Pediatr Orthop 2023;43:70–75)
D uring surgical planning for patients with early-onset
scoliosis (EOS), radiographs are utilized to assess curve
flexibility and select appropriate instrumentation levels.
Flexibility is commonly evaluated with supine side-bending
radiographs. Variability exists among radiographs; how-
ever, because they are dependent on technologist experience
and both patient cooperation and ability. Young children
may be particularly less likely to comply with side-bending
radiographs.
Previous studies have reported on the utility of a
single supine radiograph in determining curve flexibility in
adolescent idiopathic scoliosis (AIS) patients. Cheh et al1
suggested that a single preoperative supine radiograph is
highly predictive of side-bending radiographs and may
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
replace side-bending films to predict curve flexibility in
AIS patients. A more recent study by Ramchandran et al2
reproduced these findings. They defined structural curves
on supine radiographs as those with an MT curve ≥ 30
degrees and a TL/L curve ≥ 35 degrees.2
Extension of these findings to the EOS population
may eliminate the need for side-bending radiographs and
reduce radiation exposure in these young children. It may
also provide a more reliable, reproducible, technologist-
independent and patient-independent method of measur-
ing curve flexibility.
The primary purpose of this study was to evaluate
the utility of a single supine radiograph in determining
curve flexibility in EOS patients. Curve flexibility was
compared using upright, supine, and side-bending radio-
graphs to examine the relationship between these tech-
niques for operative EOS patients. If a highly predictive
relationship exists between them, a single supine film may
eliminate the need for effort-related, dual side-bending
radiographs. The secondary purpose of this study was to
determine the intrarater and interrater reliability of
measuring both supine and bending radiographs.
METHODS
This was a single-center retrospective cohort study.
Investigational review board approval was obtained be-
fore study initiation. We identified 283 patients with a
diagnosis of EOS who had undergone an index spinal
fusion or growth friendly surgery. Patients with congenital
scoliosis or prior spine surgery were excluded from eval-
uation. No patients were braced, and 1 patient was casted
before surgery. Standard of care at this hospital dictated
that all patients receive preoperative upright (standing or
seated), supine, and side-bending (left and right) ante-
roposterior radiographs. Patients did not receive all the 3
images if precluded by a component of their medical his-
tory, such as an inability to stand or sit. Of this initial
cohort, 37 patients met inclusion criteria and received all
three radiographs.
Data collection included patient demographics (age
at diagnosis and surgery, sex, body mass index, and
TABLE 1. Demographic Information
Sex, n (%)
Female 27 (73)
Male 10 (27)
Body mass index, n (%)
Underweight 6 (16)
Normal 27 (73)
Overweight 2 (5)
Obese 2 (5)
Etiology of scoliosis, n (%)
Neuromuscular 6 (16)
Syndromic 11 (30)
Idiopathic 20 (54)
This table indicates the sex, body mass index, and scoliosis etiology for the
37-patient cohort in this study. Patients with congenital scoliosis were not included
in the cohort, per the exclusion criteria. The first value displayed is the number of
patients within that specific demographic category, and the second value is the
relative percentage of the cohort represented by that demographic category.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Curve Flexibility Preoperative vs Supine Radiographs
scoliosis etiology) and radiographic measurements. Prox-
imal thoracic (PT) curves, main thoracic (MT) curves, and
thoracolumbar/lumbar (TL/L) curves were measured on
preoperative anteroposterior radiographs.
Radiographs were each measured on 2 separate oc-
casions and by 2 different raters. Each is familiar and
trained in the measurement of curves using the Cobb
technique. Rater 1 is a Pediatric Nurse Practitioner and
Instructor of Clinical Nursing with over a decade of ex-
perience with pediatric spinal deformity patients. Rater 2
was a Medical Student and is now a Resident Physician in
Orthopaedic Surgery. A minimum of 2 weeks elapsed
between each set of measurements to reduce the effect of
bias and memory of prior measurements. The same
imaging software was used for each measurement.
Statistical Analysis
All analyses were performed using SPSS version
28.0.0.0 (SPSS Inc., Chicago, IL). Correlation coefficients
were utilized to investigate correlation among PT, MT,
and TL/L curves on supine, upright, and side-bending
films. The mean absolute difference was used to evaluate
the differences among methods. Statistical significance was
defined as P < 0.05.
Intrarater correlations were assessed with Pearson r
linear regression coefficient, paired sample test, and in-
terrater correlation coefficient (ICC).3 ICC was utilized to
assess intrarater and interrater reproducibility. A 2-way
random model on absolute agreement was used to analyze
measurement reliability.4 The ICC represents the ratio of
the variance between subjects to the total variance (sub-
jects, raters, and error). The values of the ICC can range
from 0 to 1, with a higher value indicating better reli-
ability. ICC <0.40 was regarded as poor; 0.40 to 0.59 as
fair; 0.60 to 0.74 as good, and 0.75 to 1.00 as excellent. To
evaluate reliability and agreement among raters and the
different imaging methods, the first measurement of the 2
measurements in each method from each rater was applied
to the analysis.
RESULTS
The study group consisted of 37 patients who un-
derwent spinal fusion or growth friendly surgery and re-
ceived preoperative upright, supine, and side-bending
anteroposterior radiographs. The group comprised of 27
females (73%) and 10 males (27%), with a mean age of
7.0 ± 2.9 years at diagnosis (range: birth to 9 y) and a
mean age of 13.0 ± 2.9 years at time of surgery (range: 4 to
19 y). The distribution of curve etiology was neuro-
muscular (n = 6, 16%), syndromic (n = 11, 30%), and idi-
opathic (n = 20, 54%). Table 1 summarizes the
demographic information. Table 2 summarizes curve
measurements for all patients with descriptive statistics.
Our primary analysis demonstrated that (1) supine
PT and MT curve measurements were highly associated
with their corresponding side-bending measurements (PT:
r = 0.75, P < 0.001; MT: r = 0.80, P < 0.001) (Table 3); (2)
supine TL/L curve measurements were very highly
associated with their corresponding side-bending
www.pedorthopaedics.com | 71
Skaggs et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 2. Descriptive Statistics of Curves for Preoperative Radiographs

Minimum Maximum Mean SD
Type of Film Curve N (Degrees) (Degrees) (Degrees) (Degrees)
PT 32 13.0 52.3 30.0 10.3
Preoperative Upright MT 37 23.8 90.8 57.2 14.9
TL/L 37 8.7 106.8 40.5 23.6
PT 33 0.2 40.5 15.5 11.3
Preoperative Side-Bending MT 37 11.7 68.2 33.6 13.2
TL/L 37 0.0 71.7 15.4 17.4
PT 33 5.2 50.8 26.8 10.6
Preoperative Supine MT 36 18.1 83.3 48.9 13.2
TL/L 37 8.0 76.6 31.6 18.6
Preoperative upright, side-bending, and supine radiographs were obtained for all 37 patients. The curve measurements for the PT, MT, and TL/L curves for each
radiograph type are included with their respective statistics, including the average curve measurement with SD and both the minimum and maximum values. N represents the
number of patients who had the curve type with the corresponding type of radiograph. All the curve measurements were completed by both Rater 1 and Rater 2 using the
Cobb technique. Structural curves were defined as those that did not bend out to < 25 degrees. All analyses were performed using SPSS version 28.0.0.0 (SPSS Inc.,
Chicago, IL).
MT indicates main thoracic; PT, proximal thoracic; TL/L, thoracolumbar/lumbar.

measurements (TL/L: r = 0.92, P < 0.001) (Table 3); (3)
upright PT and MT curve measurements were highly
associated with their corresponding side-bending
measurements (PT: r = 0.74, P < 0.001; MT: r = 0.65,
P < 0.001) (Table 4); (4) upright TL/L curve
measurements were very highly associated with their
corresponding side-bending measurements (TL/L:
r = 0.85, P < 0.001) (Table 4); and (5) supine PT, MT,
and TL/L curve measurements were all very highly
associated with their corresponding upright
measurements (PT: r = 0.87, P < 0.001; MT: r = 0.88,
P < 0.001; TL/L: r = 0.94, P < 0.001) (Table 3).
The mean absolute differences between supine and
side-bending measurements were PT: 11.3 ± 7.8 degrees,
MT: 14.8 ± 8.3 degrees, and TL/L: 16.2 ± 7.6 degrees
(Table 3). The mean absolute differences between upright
and side-bending were PT: 14.4 ± 8.0 degrees, MT:
23.6 ± 11.8 degrees, and TL/L: 25.1 ± 12.7 degrees
(Table 4).
Intrarater and Interrater Variability
Only structural curves (defined as curves that do not
bend out to <25 degrees) were included in intrarater and
interrater reliability analyses and included 50 curves for
measurement (25 patients with a single structural curve, 11
patients with 2 structural curves, and 1 patient with 3
structural curves).
The intrarater reliability of Rater 1 is summarized in
Supplemental Table 1A, Supplemental Digital Content 1,
http://links.lww.com/BPO/A556. The absolute intrarater
difference for Rater 1 was 0.1 ± 4.9 degrees for side-
bending films and 0.4 ± 4.5 degrees for supine films. The
intrarater reliability of Rater 2 is summarized in Supple-
mental Table 1B, Supplemental Digital Content 1, http://
links.lww.com/BPO/A556. The absolute intrarater differ-
ence for Rater 2 was 0.3 ± 3.3 degrees for side-bending
films and 1.0 ± 3.0 degrees for supine films. The absolute
mean intrarater difference was 0.2 ± 4.2 degrees for side-
bending films and 0.3 ± 3.9 degrees for supine films. The
difference in mean values between supine measurements
for Rater 2 was significant, but all other differences were
insignificant.
The interrater reliability was excellent as the ICC
value of 0.88 and 0.93 for side-bending and supine films,
respectively (Table 5). The average absolute difference was
2.1 ± 5.4° and 1.1 ± 4.0° for side-bending and supine films,
respectively. The difference in mean values between raters
was significant for bending measurements, whereas the
difference between supine measurements was insignificant.
DISCUSSION
Upright and supine side-bending radiographs re-
main the classic method of assessing curve severity and

TABLE 3. Correlations Between Supine Radiographs and Side-Bending Radiographs and Between Supine Radiographs and Upright
Radiographs
Side-Bending Upright
Supine Mean Absolute Difference Mean Absolute Difference
Predictor Associations (Degrees) Associations (Degrees)
PT r = 0.75, P < 0.001 11.3 ± 7.8 r = 0.87, P < 0.001 2.9 ± 5.4
MT r = 0.80, P < 0.001 14.8 ± 8.3 r = 0.88, P < 0.001 9.2 ± 6.7
TL/L r = 0.92, P < 0.001 16.2 ± 7.6 r = 0.94, P < 0.001 8.8 ± 9.0
Pearson r linear regression coefficients were calculated for both supine and side bending and supine and upright films. A correlation coefficient with rZ0.75 was regarded
as excellent. The mean absolute difference (measured in degrees) represented the average absolute degree difference in curve measurements between the different techniques
for the PT, MT, and TL/L curves. Statistical significance was defined as P < 0.05. All analyses were performed using SPSS version 28.0.0.0 (SPSS Inc., Chicago, IL).
MT indicates main thoracic; PT, proximal thoracic; TL/L, thoracolumbar/lumbar.

72 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
TABLE 4. Correlations Between Upright and Side-Bending
Radiographs
Side Bending
Upright Mean Absolute Difference
Predictor Associations (Degrees)
PT r = 0.74, P < 0.001 14.4 ± 8.0
MT r = 0.65, P < 0.001 23.6 ± 11.8
TL/L r = 0.85, P < 0.001 25.1 ± 12.7
Pearson r linear regression coefficients were calculated for upright and side-
bending films. A correlation coefficient with rZ0.75 was regarded as excellent. The
mean absolute difference (measured in degrees) represented the average absolute
degree difference in curve measurements between the different techniques for the
PT, MT, and TL/L curves. Statistical significance was defined as P < 0.05. All
analyses were performed using SPSS version 28.0.0.0 (SPSS Inc., Chicago, IL).
MT indicates main thoracic; PT, proximal thoracic; TL/L, thoracolumbar/
lumbar.
flexibility and are often obtained for surgical planning in
patients with EOS. The importance of side-bending films
is rooted in the Lenke classification system,5 which relies
on side-bending radiographs in differentiating structural
and nonstructural curves in AIS. Ultimately, this differ-
entiation allows surgeons to plan which curves and seg-
ments will need to be included in the instrumented
construct and to predict surgical correction of a curve.
Side-bending radiographs are obtained by having
the patient use maximal effort to bend into and away from
the curve. Therefore, inherent variability exists as these
radiographs are dependent on technologist experience and
patient cooperation. Our cohort exhibited a discrepancy
between age of diagnosis (7 y) and age of surgery (13 y),
which could mitigate these concerns for our specific pa-
tients. We hope to expand the findings of this study to the
EOS population. Given their younger average age, EOS
patients are more likely to have difficulty with under-
standing and providing full effort when obtaining side-
bending radiographs.
Side-bending radiographs consist of right and left
bending supine films. An average anteroposterior radio-
graph accumulates a dose of 0.92 milligray (mGy) radia-
tion, for a total of 3.68 mGy radiation for upright, supine,
and 2 side-bending films compared with the 1.84 mGy for
1 upright and 1 supine film.6–10 These values are only for
preoperative evaluation and do not consider any other
radiographs taken during a patient’s lifetime. An average
EOS patient receives radiographs every 6 months with an
average of 25 images over a lifetime but may be as high as
50.8 Four landmark studies prospectively examined
whether increased radiation burden secondary to scoliosis
treatment and monitoring was associated with an in-
creased relative risk ratio of cancer incidence and mor-
tality. They used a multidecade follow-up period and all 4
groups reported significant increases in the standardized
incidence ratio and standardized mortality ratio for gen-
eral oncologic burden or specific cancer types.11–15 Oakley
et al6 globally reviewed radiation burden association with
scoliosis treatment and postulated increased cancer risk
was because of the scoliosis disease process itself rather
than radiation burden. Support was drawn from papers
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Curve Flexibility Preoperative vs Supine Radiographs
TABLE 5. Interrater Reliability Between Supine and Bending
Radiographs
Supine Side Bending
ICC 95% CI ICC 95% CI
Rater 1 vs. Rater 2 0.93* 0.86-0.96 0.88* 0.79-0.94
Interrater correlations were assessed using the interrater correlation coefficient
(ICC). The ICC represents the ratio of the variance between subjects to the total
variance. Values for the ICC ranged from 0 to 1, and a higher value indicated better
reliability. ICC <0.40 was regarded as poor; 0.40 to 0.59 as fair; 0.60 to 0.74 as
good; 0.75 to 1.00 as excellent. The first measurement of the 2 measurements in
each method from each rater was applied to the analysis. For each ICC for supine
and side-bending films, 95% CIs were also calculated. Statistical significance was
defined as P < 0.05. All analyses were performed using SPSS version 28.0.0.0
(SPSS Inc., Chicago, IL).
ICC indicates interrater correlation coefficient.
*P < 0.01
examining survivors of nuclear attacks, which showed that
radiation levels above 1000 mGy were required to accrue
carcinogenic potential.6,16,17 Of note, 46% of patients in
our cohort have nonidiopathic scoliosis (neuromuscular or
syndromic), indicating more complex disease etiology.
There is no clear study indicating how radiation burden
associated with scoliosis combined with underlying disease
process can affect a patient’s risk for developing cancer.
Given the documented cancer risk among scoliosis pa-
tients, we believe that minimizing radiation using supine
radiographs should be a treatment goal, especially for the
young and vulnerable EOS population.
Other methods of measuring curve flexibility have
been proposed in the adolescent and adult population, in-
cluding fulcrum,18,19 traction under general anesthesia
(UGA),20,21 push-prone,22 and lateral pressure films. The
fulcrum bending radiographs were the most effective
method for predicting flexibility in moderate curves (40 to
65 degrees) when compared with side-bending and traction
UGA.23 However, this technique was not effective at re-
ducing upper thoracic curves, in assessing overall spinal
balance, or accurately predicting final postoperative curve
correction.24 Traction UGA is often utilized in large curves
and/or neuromuscular patients. This method showed good
correction for severe curves > 60 degrees,25 but using this
technique to select distal fusion levels has been associated
with poor results.21 Recent studies comparing traction
UGA and side-bending radiographs indicated that traction
UGA was superior for patients with curves > 65 degrees,
patients less than 15 years old, and with curve apices be-
tween T4 and T8/T9.26 Push-prone films were accurate at
predicting the effect that correction of the primary curve has
on the unfused curves above and below the fusion
segment.22 However, when compared with traction UGA
and side-bending films, it underperformed in ability to
predict curve flexibility.27 The same comparison study
concluded that traction UGA and side-bending radio-
graphs showed similar flexibility predictions for MT and
TL/L structural curves, though some severe MT curves
demonstrated better flexibility with traction UGA.27 The
authors recommended preoperative imaging to include su-
www.pedorthopaedics.com | 73
Skaggs et al
pine side-bending radiographs with traction UGA
intraoperatively.27 Considering the findings of Cheh et al,1
they promulgated a strong consideration for replacing
preoperative side-bending radiographs with a single supine
radiograph.27 Unfortunately, all of these techniques involve
inherent variability as they are also dependent on technol-
ogist experience and patient effort.1 In addition, none of
these methods reduce radiation exposure.
A single supine film avoids many of the problems
with the aforementioned techniques of assessing curve
flexibility. Cheh et al1 studied the ability of the supine film
to predict curve flexibility in AIS. Their retrospective re-
view of 675 patients suggested that a single preoperative
supine radiograph was highly predictive of side-bending
radiographs and could be used as an adjunct to predict
curve flexibility. A more recent study by Ramchandran
et al2 reproduced these findings, investigating the role of
supine radiographs in determining flexibility of thoracic
and thoracolumbar curves in AIS. This retrospective
review of 90 operative patients found that a single pre-
operative supine radiograph was highly predictive of side-
bending radiographs. They suggested new cutoff values
for structural curves based on supine radiographs: ≥ 30
degrees for MT curves and ≥ 35 degrees for TL/L curves.2
Our study is similar in methodology, and we at-
tempted to extrapolate the findings to the EOS population.
We successfully demonstrated a strong correlation be-
tween upright, supine, and side-bending radiographs in
EOS patients. Supine radiographs had a smaller mean
curve when compared with upright radiographs and a
larger mean curve when compared with bending radio-
graphs. Currently, there is no single standard method for
assessing the flexibility of a curve in EOS, but this study
showed that supine radiographs may be able to replace
side-bending radiographs in evaluating curve flexibility.
We additionally evaluated the intrarater and inter-
rater reliability for both supine and bending films. Both
Rater 1 and 2 had excellent intrarater reliability between
the first and second set of measurements. In addition,
excellent interrater reliability was observed. On the basis
of these findings, we can conclude that both supine and
bending radiographs are reliable.
This study had several limitations. This study did not
investigate whether supine radiographs should replace
side-bending films in surgical planning. Further studies
will be required before such a recommendation and was
not within the scope of the current paper. The novelty of
this study involved using a single supine radiograph in-
stead of dual side-bending radiographs in the EOS pop-
ulation. We therefore endeavored to first identify whether
this principle was clinically applicable. Bowker et al28
published a novel article examining the correlation be-
tween preoperative flexibility and postoperative outcomes
after growth friendly surgery in EOS patients. They con-
cluded that less flexible preoperative curves were asso-
ciated with less postoperative correction and higher risk of
postoperative complications. Although not an aim of this
study, we hypothesized that similar trends would likely be
noted for our patient cohort.
74 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
The reliability and validity of lowest instrumented
vertebra selection in EOS patients using supine radio-
graphs is unknown and was not evaluated in this study.
Another limitation of this study was the relatively limited
sample size. Finally, our study did not compare supine
films to other techniques that could be used to predict
curve flexibility, such as traction or push-prone films.
Trends in the literature indicate the general efficacy of
side-bending radiographs but all reflect patients with
AIS.23,26,27 More studies are certainly needed in the EOS
population to delineate true efficacy of side-bending and
supine radiographs in predicting curve flexibility.
A single, preoperative supine radiograph was
highly predictive of side-bending radiographs in patients
with EOS. We observed that supine curves measured an
average of 15 degrees larger than bending curves in the
MT and TL/L regions. Therefore, a supine radiograph
could be used as an adjunct to predict curve flexibility.
Further studies are needed using a larger patient cohort,
but this study provides evidence that side-bending ra-
diographs may not provide enough information to ne-
cessitate their use for routine preoperative evaluation.
When combined with the general limitations of side-
bending radiographs in the EOS population, we con-
cluded that a single supine film may eliminate the need
for effort-related, dual side-bending radiographs, and
their associated radiation.
REFERENCES
1. Cheh G, Lenke LG, Lehman RA, et al. The reliability of
preoperative supine radiographs to predict the amount of curve
flexibility in adolescent idiopathic scoliosis. Spine (Phila Pa 1976).
2007;32:2668–2672.
2. Ramchandran S, Monsour A, Mihas A, et al. Impact of supine
radiographs to assess curve flexibility in the treatment of adolescent
idiopathic scoliosis. Glob Spine J. 2022;12:1731–1735.
3. Loder RT, Urquhart A, Steen H, et al. Variability in Cobb angle
measurements in children with congenital scoliosis. J Bone Joint Surg
Br. 1995;77:768–770.
4. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater
reliability. Psychol Bull. 1979;86:420–428.
5. Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis:
a new classification to determine extent of spinal arthrodesis. J bone
Jt Surg. 2001;83-A:1169–1181.
6. Oakley PA, Ehsani NN, Harrison DE. The scoliosis quandary: are
radiation exposures from repeated x-rays harmful. Dose-Response.
2019;17:1–10.
7. Damet J, Fournier P, Monnin P, et al. Occupational and patient
exposure as well as image quality for full spine examinations with the
EOS imaging system. Med Phys. 2014;41:063901-1–063901-12.
8. Knott P, Pappo E, Cameron M, et al. SOSORT 2012 consensus
paper: reducing x-ray exposure in pediatric patients with scoliosis.
Scoliosis. 2014;9:4.
9. Kalifa G, Charpak Y, Maccia C, et al. Evaluation of a new low-dose
digital X-ray device: first dosimetric and clinical results in children.
Pediatr Radiol. 1998;28:557–561.
10. Melhem E, Assi A, El Rachkidi R, et al. EOS(®) biplanar X-ray
imaging: concept, developments, benefits, and limitations. J Child
Orthop. 2016;10:1–14.
11. Ronckers CM, Doody MM, Lonstein JE, et al. Multiple diagnostic
X-rays for spine deformities and risk of breast cancer. Cancer
Epidemiol Biomarkers Prev. 2008;17:605–613.
12. Simony A, Hansen EJ, Christensen SB, et al. Incidence of cancer in
adolescent idiopathic scoliosis patients treated 25 years previously.
Eur Spine J. 2016;25:3366–3370.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
13. Levy AR, Goldberg MS, Mayo NE, et al. Reducing the lifetime risk
of cancer from spinal radiographs among people with adolescent
idiopathic scoliosis. Spine. 1996;21:1540–1548.
14. Hoffman DA, Lonstein JE, Morin MM, et al. Breast cancer in
women with scoliosis exposed to multiple diagnostic x rays. J Natl
Cancer Inst. 1989;81:1307–1312.
15. Doody MM, Lonstein JE, Stovall M, et al. Breast cancer mortality
after diagnostic radiography: findings from the U.S. Scoliosis Cohort
Study. Spine (Phila Pa 1976). 2000;25:2052–2063.
16. Cuttler MJ, Welsh SJ. Leukemia and ionizing radiation revisited.
J Leuk. 2015;03:1–2.
17. Cuttler JM. Leukemia incidence of 96,000 Hiroshima atomic bomb
survivors is compelling evidence that the LNT model is wrong. Arch
Toxicol. 2014;88:847–848.
18. Cheung KMC, Luk KDK. Prediction of correction of scoliosis with
use of the fulcrum bending radiograph. J Bone Jt Surg - Ser A.
1997;79:1144–1150.
19. Luk KDK, Cheung KMC, Lu DS, et al. Assessment of scoliosis
correction in relation to flexibility using the fulcrum bending
correction index. Spine (Phila Pa 1976). 1998;23:2303–2307.
20. Polly DW Jr, Sturm PF. Traction versus supine side-bending: which
technique best determines curve flexibility? Spine (Phila Pa 1976).
1998;23:804–808.
21. Vaughan JJ, Winter RB, Lonstein JE. Comparison of the use of supine
bending and traction radiographs in the selection of the fusion area in
adolescent idiopathic scoliosis. Spine (Phila Pa 1976). 1996;21:2469–2473.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Curve Flexibility Preoperative vs Supine Radiographs
22. Vedantam R, Lenke LG, Bridwell KH, et al. Comparison of push-
prone and lateral-bending radiographs for predicting postoperative
coronal alignment in thoracolumbar and lumbar scoliotic curves.
Spine (Phila Pa 1976). 2000;25:76.
23. Hamzaoglu A, Talu U, Tezer M, et al. Assessment of curve flexibility
in adolescent idiopathic scoliosis. Spine (Phila Pa 1976). 2005;30:
1637–1642.
24. Klepps SJ, Lenke LG, Bridwell KH, et al. Prospective comparison of
flexibility radiographs in adolescent idiopathic scoliosis. Spine (Phila
Pa 1976). 2001;26:74–79.
25. Davis BJ, Gadgil A, Trivedi J, et al. Traction radiography performed
under general anesthetic: a new technique for assessing idiopathic
scoliosis curves. Spine (Phila Pa 1976). 2004;29:2466–2470.
26. Watanabe K, Kawakami N, Nishiwaki Y, et al. Traction versus
supine side-bending radiographs in determining flexibility: what
factors influence these techniques. Spine (Phila Pa 1976). 2007;32:
2604–2609.
27. Liu RW, Teng AL, Armstrong DG, et al. Comparison of supine
bending, push-prone, and traction under general anesthesia radio-
graphs in predicting curve flexibility and postoperative correction in
adolescent idiopathic scoliosis. Spine (Phila Pa 1976). 2010;35:
416–422.
28. Bowker R, Morash K, Mishreky A, et al. Scoliosis flexibility
correlates with post-operative outcomes following growth friendly
surgery. Spine Deform. 2022;10:933–941.
www.pedorthopaedics.com | 75
 ORIGINAL ARTICLE
A Comprehensive Assessment of Psychosocial Well-being
Among Growing Rod Graduates: A Preliminary
Investigation
Mevhibe İrem Yıldız, MD,* Barlas Goker, MD,† Talat Demirsöz, PhD,* Cihan Aslan, MD,‡
Halil Gokhan Demirkiran, MD,† Sevilay Karahan, PhD,§ Mumin Kazim Yazici, MD,*
and Müharrem Yazıcı, MD∥
Background: Traditional growing rods (TGR) encompass a long
process, in which patients experience physical and psychosocial
difficulties. However, the effect of repeating surgeries on the overall
psychological functioning of graduated patients has not been
thoroughly investigated in the literature. The aim of this study is to
evaluate the psychological well-being of graduated idiopathic
early-onset scoliosis patients in terms of psychopathology,
neurocognition, and psychosocial functioning, and determine the
accuracy of scoliosis outcome questionnaires in these regards.
Methods: TGR graduates with idiopathic early-onset scoliosis
without known intellectual disabilities or neuromuscular im-
pairments were included. Patients were thoroughly evaluated
using psychological instruments [Wechsler Adult Intelligence
Scale, Auditory Consonant Trigram Test, Verbal Fluency Test,
Beck Depression Inventory, Beck Anxiety Inventory, Rosenberg
Self-Esteem Scale, Symptom Checklist-90, Post-Traumatic
Growth Inventory, Strengths and Difficulties Questionnaire,
Spinal Appearance Questionnaire, Scoliosis Research Society
22-item questionnaire (SRS-22)]. Results were compared with
normative data when available. Spearman correlations were
performed between the results of these tests, the total treatment
duration, and the number of spinal surgeries.
Results: Of the 15 patients included in the study, 9 were females,
and the mean age was 18.73 (16 to 23). The mean age at index
From the *Department of Psychiatry; †Department of Orthopaedics and
Traumatology; ‡Department of Child and Adolescent Psychiatry;
§Department of Biostatistics, Hacettepe University Faculty of Med-
icine; and ∥Department of Orthopaedic Surgery, Cankaya Hospital,
Ankara, Turkey.
Institutional review board approval was obtained from the Hacettepe
University Non-interventional Clinical Research Ethics Board (Date:
February 23, 2021 Number:2021/10-39).
No funds were received in support of this work. No relevant financial
activities outside the submitted work. None of the authors received
financial support for this study.
The authors declare no conflicts of interest.
Reprints: Muharrem Yazici, MD, Department of Orthopaedic
Surgery, Cinnah Cd. 112/2 06690 Çankaya, Ankara, Turkey. E-mail:
mimyazici@gmail.com.
Supplemental Digital Content is available for this article. Direct URL
citations appear in the printed text and are provided in the HTML
and PDF versions of this article on the journal’s website, www.
pedorthopaedics.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002298
76 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
surgery was 6.38 (3 to 10) whereas that of graduation was 14.00
(12 to 16). The average number of spinal surgeries was 14.28 (7 to
20). Two patients performed below the range of adult intellectual
functioning. Auditory Consonant Trigram Test showed normal
verbal working memory and attention control. Six patients had
abnormal Verbal Fluency Test performance. Eight patients had
abnormal ratings on at least one of the assessment scales of
psychopathology (Symptom Checklist-90, Beck Depression In-
ventory, and Beck Anxiety Inventory). Eight patients had low-
to-moderate self-esteem (Rosenberg self-esteem scale). The me-
dian spinal appearance questionnaire and SRS-22 scores were 34
and 4.18, respectively. Pain and function subdomains of SRS-22
scored higher than self-image and mental health. No correlation
was found between the treatment duration and number of sur-
geries and test scores. SRS-22 showed correlations with multiple
psychological tests.
Conclusion: Completed TGR treatment yields acceptable
correction of deformities and surgical outcomes, however, may
fail to improve psychological well-being. This is the first study to
find various psychosocial abnormalities in two-thirds of cases.
Level of Evidence: Level IV; cross-sectional study.
Key Words: early-onset scoliosis, growing spine, growing rod,
traditional growing rod, psychosocial, quality of life, health-related
quality of life, scoliosis, posttraumatic growth, postoperative cog-
nitive dysfunction
(J Pediatr Orthop 2023;43:76–82)
S ince their advent more than 2 decades ago, lengthening
surgeries with traditional dual-growing rods (TGR)
have been a staple in the treatment of early-onset scoliosis
(EOS).1,2 Even though magnetically-controlled growing
rods (MCGR) have been popularized in recent years,
TGR still retains a significant role in the current ortho-
paedic surgery practice. TGR lengthenings have mostly
been considered safe procedures with acceptable compli-
cation rates.2,3 As more children undergo the process of
TGR lengthenings every day, previously neglected areas
such as the psychosocial well-being of these patients have
been gaining prominence.4–7
Repeating cycles of hospital admissions and exposure
to operating theaters could potentially cause detrimental
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
effects on the psychological well-being of EOS patients. In
addition to other psychiatric abnormalities, deterioration of
cognitive function is another possible risk of childhood
surgeries.8 The effect of general anesthesia on the develop-
ing brain of children has been an intriguing area of research.
So far, the literature mostly indicates that a single procedure
under general anesthesia does not significantly lead to
clinically notable deterioration in the development of the
central nervous system.8 However, the effect of multiple
surgeries on the neurodevelopment of children is still a
matter of controversy. Some studies have shown an in-
creased risk of attention deficit and hyperactivity disorder
and learning disabilities, and postoperative cognitive
dysfunction (POCD), a decline in several neurocognitive
functions (impairment in concentration and memory,
executive dysfunction, and weakened social skills), after
repeating surgeries.9–11
Improvements in physical health and quality of life
obtained with numerous surgeries during childhood might
have reflections on the psychological well-being of adoles-
cents. Scoliosis is also presumed to negatively affect body
image perception, an essential psychological construct af-
fecting self-esteem.12 The bodily component of self-esteem
(perception of appearance) is of great importance in de-
veloping ones’ self-identity and has a significant driving
effect on psychosocial functioning. One of the aims of
surgery is to have a positive contribution to body image by
deformity correction.11 Although decreased scoliosis major
curve angles have been associated with better self-image in
adolescent idiopathic scoliosis, the effect of correction on
graduated EOS patients remains unclear.13
As the number of TGR graduates increases, more
studies have focused on the functional outcomes and
quality of life measures of the “graduated” EOS patients
who have reached maturity.14–16 The prolonged duration of
treatment (several years in most cases) and multiple sur-
geries throughout childhood may have lasting effects on the
psychological well-being of patients even after graduation.
Although prior studies showed an increased likelihood of
psychiatric diagnoses (generalized anxiety disorder, de-
pression, attention deficit and hyperactivity disorder) in
children with ongoing lengthenings and a correlation be-
tween the number of surgeries and behavioral difficulties,5
there are no reports that demonstrate whether these findings
remain valid in graduated adolescents or investigate other
domains of psychological well-being such as neuro-
cognition and psychosocial functioning. The aim of this
study is to evaluate the impact of TGR on the psycho-
logical well-being of graduated idiopathic EOS patients in
terms of psychopathology, neurocognition, and psychoso-
cial functioning, and determine the accuracy of scoliosis
outcome questionnaires in these regards.
METHODS
Institutional review board approval was obtained
from our institutions’ Non-interventional Clinical
Research Ethics Board (Date: February 23, 2021 Number:
2021/10-39).
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Psychosocial Well-being of Graduated Patients
A cross-sectional study was performed at a tertiary
referral center’s orthopaedic surgery and adult psychiatry
departments. EOS patients initially treated and graduated
from TGR between January 2004 and December 2020
were included in the study. Inclusion criteria for the study
were as follows: (1) diagnosis of idiopathic EOS, (2)
completion of treatment with surgical TGR lengthenings
(“graduation”), (3) at least 1 year of elapsed time since
graduation, (4) a total number of spinal surgeries > 5, (5)
no known moderate-to-severe intellectual disabilities, and
(6) no neuromuscular disease involvements which cause
psychosocial impairments. Patients who did not give
consent were excluded from the study.
Data Collection
Study participants were appointed to an in-person
interview with an experienced clinical neuropsychologist.
A neurocognitive test battery was administered to the
participants including Wechsler Adult Intelligence Scale
(WAIS), Auditory Consonant Trigram Test (ACT), and
Verbal Fluency Test (VFT) focusing on memory, execu-
tive functions, and verbal fluency accompanied with gen-
eral intellectual functioning (Table 1). A 20-minute rest
was allowed between the tests.
After the in-person interview, patients were given
psychological instruments assessing their psychological well-
being, which included the Beck Depression Inventory (BDI),
Beck Anxiety Inventory (BAI), Rosenberg Self-Esteem Scale
(RSES), Symptom Checklist-90 (SCL-90), Post-Traumatic
Growth Inventory (PGI) and posttraumatic growth in-
ventory, and psychosocial functioning was evaluated by the
Strengths and Difficulties Questionnaire (SDQ). Patients
were also given scoliosis outcome questionnaires such as the
Scoliosis Research Society 22-item questionnaire (SRS-22)
and the Spinal Appearance Questionnaire (SAQ). All of the
given questionnaires were validated in Turkish (Table 1).
Demographic data (age, sex, educational back-
ground, and socioeconomic status), information regarding
functional capacity and well-being, and other variables
related to the psychiatric history of patients (prior diag-
noses, treatments, and family history) were collected by
surveys. Surgical data such as the age at index surgery,
graduation, and evaluation, number of surgeries, and total
duration of TGR treatment were recorded from the hos-
pital database. Scoliosis angles of the major curves were
measured from preoperative and final radiographs.
Data Analysis
IBM SPSS Statistics 26 was used for the statistical
analysis. Sociodemographic and clinical variables and
surgical, psychological, and cognitive test parameters were
analyzed with descriptive statistics. The relationships be-
tween the psychological instruments and treatment dura-
tion and number of surgeries were investigated with
Spearman correlations. Furthermore, SAQ and SRS-22
results were correlated with the other psychological in-
struments. As there was no control group in the study, the
results of the surveys were compared with normative data
when available.
www.pedorthopaedics.com | 77
Yildiz et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 1. The psychological Instruments Used in the Evaluation of Study Participants

BDI17 A 21-item self-reported questionnaire that measures the emotional, vegetative, cognitive, and motivational signs of depression in
individuals 13 y and older.
BAI18 A 21-item self-reported questionnaire that measures recent anxiety levels designed for individuals 17 y and older.
ACT19 A test of verbal working memory and attention control, as participants are given 3 consonant letters and are asked to repeat them after
counting backwards. A consonant trigram is administered to the participant at a rate of 1 letter per second. Then, a 2 or 3-digit random
number is delivered immediately after this trigram. Then, participants are asked to repeat these letters immediately or count out loud
backwards from this given number by threes for the random interval delays of 3, 9, or 18 sec. At the end of the trial, the participant is
asked to remember the given trigram which is presented at the beginning. The total number of letters correctly recalled is the dependent
variable/measure in each interval delay period. Totally, there are 20 trials. Higher scores indicate better performance. Test scores for
ACT are based on the normative data stratified by age and education. The level of impairment of the performance of participants are
determined based on the normative data.
VFT20 A test that assesses 2 types of fluency: category (semantic) fluency and letter (phonemic) fluency. Participants are expected to produce as
many words from a category in a given time. Category fluency was assessed in 3 trials: participants were asked to produce names of
people in the names of people category, then they were asked to produce names of animals in the animal category, and lastly, they were
asked to produce names of people and animal names by changing the categories one after another. Participants were asked to produce
as many words as possible within 60 sec in each trial according to the desired criteria. The letter fluency was assessed in 3 trials:
participants were asked to produce words starting with the letter S, A, and Z. In these trials, participants were asked to refrain from
producing names of people, places, or numbers. Participants’ correct words were counted in each trial, and participants’ scores were
then calculated. Test scores for VFT are based on the normative data stratified by age, sex and education. The level of impairment of
the performance of participants are determined based on the normative data.
WAIS21 An assessment tool designed to measure intelligence which comprises 5 verbal subtests (information, similarities, arithmetic, digit span,
and vocabulary) and 4 performance subtests (picture completion, picture arrangement, block design, and digit-symbol).
SDQ22 A tool used to screen psychological problems in children and adolescents, which comprises 5 domains: behavioral difficulties,
hyperactivity and concentration difficulties, emotional distress, peer relationship problems, and prosocial behaviors.
RSES23 A self-reported scale of 10 items that measures an individual’s sense of self-worth.
SCL-90-R24 A self-reported psychiatric checklist designed to evaluate a broad range of psychological problems and psychopathology.
PGI25 A 21-item Likert-scale, which measures the changes in perception, personal development, and coping mechanisms after an event of
trauma.
SRS-2226 A 22-item self-reported outcome measure that assesses the quality of life of scoliosis patients. It consists of 4 subdomains: pain, function,
self-image, and mental health.
SAQ27 A 33-item self-assessment that assesses the self-perception and satisfaction of a patient’s spinal deformity. It consists of 2 parts:
appearance and expectation.
ACT indicates Auditory Consonant Trigram Test; BAI, Beck Anxiety Inventory; BDI, Beck Depression Inventory; PGI, Posttraumatic Growth Inventory; RSES,
Rosenberg Self-Esteem Scale; SAQ, Spinal Appearance Questionnaire; SDQ, Strengths and Difficulties Questionnaire; SRS-22, Scoliosis Research Society 22-item ques-
tionnaire; VFT, Verbal Fluency Test; WAIS, Wechsler Adult Intelligence Scale.

RESULTS returned to their normal functioning levels after revision

Demographics and Surgical Data
After the exclusion of neuromuscular and syndromic
scoliosis patients, 28 TGR graduates with idiopathic sco-
liosis were selected from the hospital database. One pa-
tient was excluded due to a need for revision surgery and 3
patients had neurological deficits that caused impairment
in function due to spinal cord pathologies. Twenty-four of
these idiopathic EOS patients who graduated from TGR
met the inclusion criteria. Among these, 5 patients de-
clined to give consent to the study and we were unable to
contact 4 patients. A total of 15 patients (6 males and 9
females) participated in the study. All participants were
graduates of TGR. The mean age was 19 years (range, 16
to 23 y) at the time of evaluation. The mean age at index
surgery was 6.3 (range, 2 to 11 y), and was 14 (range, 11 to
16 y) at graduation. The mean total duration of treatment
was 7.3 years (range, 3 to 11 y). An average of 5 years
(range, 1 to 10 y) elapsed since the last surgery.
The average scoliosis major curve angle of patients
was 69 degrees (59 to 85 degrees) before index surgery and
42 degrees (23 to 73 degrees) at the final evaluation. The
mean number of surgeries per patient was 14.8 (7 to 20).
Two patients underwent revision surgery after graduation
for postfusion adding-on development. Both patients
surgeries and enough time had passed for them to meet the
inclusion criteria.
All of the patients were physically active and par-
ticipated in daily activities. Nine of them stated that they
had no shortcomings/difficulties in activities of daily life.
Seven patients stated they had no limitations of physical
capabilities including exercise and recreational sports.
Psychological Outcomes
The Intelligence Quotient (IQ) scores of 2 patients
were below the average range of adult intellectual func-
tioning according to the WAIS test (79 and 58, indicating
borderline and mild intellectual disability, respectively).
The median verbal IQ (M = 105), performance IQ
(M = 99), and full-scale IQ scores (M = 101) were con-
sistent with the average range of adult intellectual ability
(Tables 2, 3). Two patients were unable to participate in
the in-person interview and were evaluated online through
a video conference session. As a result, the intellectual
functioning of these patients could not be assessed by
WAIS. However, they were clinically evaluated by an
online interview, were found to perform above average,
and both were in university education.
The 2 patients who had IQ scores below the average
range of adult intellectual functioning were both literate

78 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Psychosocial Well-being of Graduated Patients

TABLE 2. Median and Range Values of Neuropsychological and Psychosocial Test Parameters
Range Median (min-max)
SAQ-appearance 10-50 21 (10-35)
(higher scores indicate worse outcomes)
SAQ-expectations 4-20 13 (4-20)
(higher scores indicate worse outcomes)
SAQ-total 14-70 34 (14-55)
(higher scores indicate worse outcomes)
SAQ-32 (self-assessment) 1-5 3 (2.0-5)
(higher scores indicate worse outcomes)
SRS-22-pain 1-5 4.2 (2.8-5.0)
SRS-22-function 1-5 4.50 (2.25-5.0)
SRS-22-self-image 1-5 3.60 (1.40-5.0)
SRS-22-mental health 1-5 3.80 (2.8-5.0)
SRS-22-subtotal 1-5 4.10 (2.5-5.0)
SRS-22-satisfaction 1-5 5.0 (2.5-5.0)
SRS-22-total 1-5 4.18 (2.59-5.0)
WAIS-total 90-110: average range of adult intellectual functioning 101 (58-122)
BDI 0-63 10 (2-32)
10-16 mild
17-23 moderate
24-63 severe
BAI 0-63 2 (0-26)
8-15 mild
16-25 moderate
26-65 severe
RSES 0-6 2.0 (0-6)
0-1 high
2-4 medium
5-6 low
SCL-90 Global severity index 0.44 (0.03-2.16)
< 1: no distress
SDQ-total 0-40 7 (2-24)
emotional + conduct + hyperactivity + peer problem scale
(lower scores indicate fewer difficulties)
SDQ-emotional symptoms 0-10 2 (0-9)
(lower scores indicate fewer difficulties)
SDQ-conduct problems 0-10 1 (0-3)
(lower scores indicate fewer difficulties)
SDQ-hyperactivity/inattention 0-10 2 (0-7)
(lower scores indicate fewer difficulties)
SDQ-peer relationship problems 0-10 2 (0-6)
(lower scores indicate fewer difficulties)
SDQ-prosocial behavior 0-10 9 (4-10)
(higher scores indicate fewer difficulties)
PGI 0-105 62 (18-94)
(higher points indicate more growth)
BAI indicates Beck Anxiety Inventory; BDI, Beck Depression Inventory; PGI, Posttraumatic Growth Inventory; RSES, Rosenberg Self-Esteem Scale; SAQ, Spinal
Appearance Questionnaire; SCL-90, Symptom Checklist-90; SDQ, Strengths and Difficulties Questionnaire; SRS-22, Scoliosis Research Society 22-item questionnaire;
WAIS, Wechsler Adult Intelligence Scale.

with ongoing education. Their intelligence levels were
deemed sufficient for taking the psychological well-being
instruments after being interviewed by a psychiatrist and a
clinical psychologist, as IQ scores are not considered eli-
gibility criteria. However, they were not evaluated with
neuropsychological testing due to the intelligence re-
quirements of ACT and VFT, and the other 11 patients
underwent these tests. All of the eligible participants per-
formed in the normal ranges of ACT; however, 6 patients
demonstrated below-normal levels of neurocognitive
performance in VFT(Table 4).
Eight patients had abnormal ratings on at least one of
the assessment scales, which evaluate psychopathology
(SCL-90, BDI, and BAI)(Table 3). SCL-90 indicated
significant psychological distress in 3 patients. BDI had a
mean value of 12.26, and the results suggested moderate-
to-severe depression in 4 and mild depression in 4 patients.
BAI showed that 2 patients had moderate-to-severe anxiety.
RSES revealed that 8 patients had low-to-moderate
self-esteem. All but 1 patient with low self-esteem had
moderate-to-severe depression and/or anxiety.
The median SAQ-Total score was 34. SAQ-appearance
scored worse than the SAQ-expectation subdomain. SRS-22
had a median score of 4.18. The pain and function sub-
domains of the SRS-22 questionnaire scored higher than self-
image and mental health.
Spearman correlation revealed no relationship
between the number of surgeries and the test results

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 79

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Yildiz et al J Pediatr Orthop  Volume 43, Number 2, February 2023

investigation was performed after the period of surgeries

TABLE 3. Study Participants and Their Respective Scores on
the Outcome Questionnaires and Psychological Instruments and visits to the operating theaters had been finished, and
Related to Psychopathology (BDI, BAI, SCL-90), Intellectual the group mainly consisted of older adolescents and young
Functioning (WAIS), and Self-esteem (RSES) adults. This is the first study in which a TGR-treated EOS
WAIS- WAIS- WAIS- SCL- population was surveyed by a multidimensional ques-
# verbal performance total BDI BAI RSES 90 tionnaire such as SCL-90, which screens for various psy-
chological symptoms and psychopathological features.
1 NA NA NA 6 2 0 0.27
2 113 100 111 8 2 2 0.16
The results of SCL-90 indicated significant psychological
3 118 125 122 6 1 1 0.45 distress in one-fifth of the study population. When this is
4 104 98 101 26 26 6 2.16 considered together with the findings of BDI, BAI, and
5 86 100 91 5 2 0 0.04 RSES, it could be concluded that the TGR lengthenings
6 110 93 103 15 2 1 1.1 failed to provide psychological well-being in a consid-
7 60 60 58 32 16 4 1.35
8 96 89 93 2 0 0 0.03 erable number of patients.
9 106 103 105 10 7 3 0.47 As each hospital admission causes a psychosocial
10 107 99 103 18 3 6 0.66 insult for the growing child, TGR lengthenings have pre-
11 105 83 96 12 1 2 0.08 viously been associated with internalizing disorders such as
12 74 88 79 15 6 2 0.44
13 93 99 95 21 7 4 0.87
increased anxiety and depression.5,6 In contrast, these
14 119 120 121 6 0 0 0.07 children may also adapt coping skills with repeating sur-
15 NA NA NA 2 2 0 0.13 geries and demonstrate better social and physical
Bold characters indicate abnormal values.
functioning.4,5 Although this study confirmed the pre-
NA indicates Unable to participate in the in-person interview. The intellectual viously demonstrated findings of increased anxiety and
functioning of these patients could not be assessed by WAIS. depression, a correlation between the number of surgeries
BAI indicates Beck Anxiety Inventory; BDI, Beck Depression Inventory;
RSES, Rosenberg Self-Esteem Scale; SCL-90, Symptom Checklist-90; WAIS,
and psychological parameters could not be found, so the
Wechsler Adult Intelligence Scale. previously reported coping mechanisms were not illus-
trated in the EOS graduates. In contrast, improvements in
the pain and self-image components of SRS-22 were as-
(Appendix-1, Supplemental Digital Content 1, http://links. sociated with the posttraumatic growth (PTG) of partici-
lww.com/BPO/A562). No correlations were found be- pants. PTG is the positive psychological change that results
tween the test results and the individual treatment from a struggle through a life-altering experience: a greater
duration of the patients. appreciation for life, meaningful interpersonal relation-
SRS-22 showed multiple correlations with many ships, a sense of personal strength, changes in priorities,
other test parameters(Appendix-1, Supplemental Digital and enhancement in spiritual development.28 Coping
Content 1, http://links.lww.com/BPO/A562). Self-image strategies used to manage stressors are significant in the
and mental health subdomains of SRS-22 showed strong development of PTG. The improvement in appearance and
correlations with BDI and BAI. No significant correla- lessening of the burden of disease with repetitive surgical
tions were found between SAQ-Total and other tests ex- interventions may have allowed the emergence of PTG.
cept RSES. SAQ-expectations showed correlations with In this study, WAIS results were mostly within the
BDI, RSES, SCL-90, SDQ-emotional symptoms, and normal range. Prior studies indicate that major surgeries,
SDQ-total. Posttraumatic growth inventory showed including interventions such as cardiovascular, orthopae-
correlations with SRS-pain and SRS-self-image. dic, and urologic procedures, would carry an increased
risk of POCD.29 Although TGR lengthening procedures
inflict relatively lower amounts of trauma, initial in-
DISCUSSION strumentation, and final fusion of EOS patients, as well as
Growing rods have been around for decades, and as additional surgeries performed for complications, con-
more EOS patients graduate from this years-spanning stitute prolonged anesthesia exposures and surgical in-
treatment, concerns about its effect on the well-being and vasiveness. A previous study on adolescent idiopathic
quality of life of patients are getting more attention. This scoliosis patients showed intensive stress responses due to
study investigated the psychological profile of the idio- surgery.9 There may be an association between stress re-
pathic EOS graduates in the authors’ institution. Although sponses and the decline in social skills. Apart from this, in
the surgical outcomes were successful in terms of de- the context of neurocognitive functions affected by major
formity correction and function based on SRS-22 and surgeries, it is known that some cognitive functions, such
SAQ, mental well-being was found to be less than ideal, as as verbal fluency and planning skills, might take longer to
almost two-thirds of patients demonstrated abnormalities mature until age 12 or later.30 Therefore, adolescents are
in at least one of the psychological domains. This denotes accepted to have an elevated risk for POCD after ortho-
a suboptimal outlook of the psychological profiles of paedic surgeries and a similar effect may occur after TGR
patients in contrast to expectations. graduation. According to the neurocognitive findings of
Prior studies on the psychology of EOS patients this study, although the function of verbal working
were conducted on children who were undergoing memory seemed to be intact, the verbal functioning and
TGR lengthenings.4–7 In this first-of-its-kind study, the executive function of some of the participants seemed to

80 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Psychosocial Well-being of Graduated Patients

TABLE 4. Age, Educational Status, VFT [Letter Fluency (S, A, and Z) and Category Fluency (Human, Animal, and Alteration)] and
ACT Scores (Delay Intervals of 0, 3, 9, and 18 Seconds)
VFT ACT
Patient no. Age Education S A Z Human Names Animal Names Alternation 0 3 9 18
1 23 University graduate NA NA NA NA NA NA NA NA NA NA
2 20 University undergraduate 11 8* 8 21* 20 17* 15 14 10 10
3 21 University undergraduate 13 20 13 29 26 23 15 15 13 14
4 20 University undergraduate 8* 6* 5 18† 10‡ 13† 15 15 10 12
5 21 University undergraduate 11 10* 4* 18 23† 16* 15 10 9 11
6 22 University undergraduate 8† 6† 6† 15* 12* 11‡ 15 12 13 14
7 18 High school student NA NA NA NA NA NA NA NA NA NA
8 16 High school student NA NA NA NA NA NA NA NA NA NA
9 16 High school student 11 12 10 29 22 21 15 14 11 10
10 16 High school student 16 13 8 32 23 23 15 14 14 10
11 17 High school student 7† 2† 6 20 18 24 15 15 14 6
12 18 High school student 12 9 4 20 16 14* 15 15 11 10
13 18 High school student 11 13 8 25 22 22 15 12 12 8
14 19 University undergraduate 11 12 6* 29 21 20 15 14 13 15
15 16 High school student NA NA NA NA NA NA NA NA NA NA
*Mildly impaired (between 1 and 2 SD below mean).
†Moderately impaired (between 2 and 3 SD below mean).
‡Severely impaired (< 3 SD below mean).
NA indicates data not available.
ACT indicates Auditory Consonant Trigram Test; VFT, Verbal Fluency Test.

be impaired. These may represent the deleterious effects of
TGR on neurocognition, and further studies are needed to
explore the underlying mechanisms.
Other notable findings of this study are the sig-
nificant negative correlations between the SRS-22 and
the BDI, BAI, RSES, and SCL-90 scores; which indicate
a similar trend between these psychological instruments
and SRS-22 in the favorability of outcomes. Fur-
thermore, the emotional and conduct problems domains
and total scores of SDQ were found to be negatively
correlated with multiple subdomains of SRS-22. Strong
correlations suggest that issues in physical functioning
according to SRS-22 could be closely associated with
conduct problems. Similarly, the mental health domain
of SRS-22 was strongly correlated with the emotional
problems and total domains of SDQ. Moreover, these
domains showed strong correlations with the self-image
and satisfaction subdomains of SRS-22. These findings
are suggestive of the consistency between the 2 scales in
the assessment of similar problems and the impact of
perceived ongoing orthopaedic difficulties of graduates’
psychosocial functioning. The psychological instruments
showed correlations with at least one domain of SRS-22.
Therefore, the findings of this study suggest that SRS-22
may be a useful tool for monitoring the mental well-being
of this population. Further studies with larger sample
sizes are needed to determine the utility of this ques-
tionnaire for the purpose of monitoring psychological
well-being.
In contrast to SRS-22, the SAQ-Total scores only had
significant correlations with RSES. The appearance domain
of SAQ showed no significant correlations, whereas the
expectations domain was positively correlated with BDI,
SDQ (emotional symptoms/total), RSES, and SCL-90.
These results might indicate an association between higher
expectations of spinal deformity correction after TGR
graduation and a predisposition to psychopathology.
The strengths of this study include the homogeneity
of a rare disease population, as the patients were all idi-
opathic EOS patients and TGR graduates. This prevented
confounding by the potential effects of other scoliosis
etiologies on neurocognitive development and function.
As the patients were thoroughly investigated by both
clinical interviews and psychological surveys, the obtained
psychological data was reliable for detailed evaluation.
Another strength of the study is that the elapsed time
between the final surgeries and the study assessments was
set to at least 1 year. This allowed for the fears regarding
the disease and surgical trauma to subside, and thereby
negated the acute effects.
The limitations of this study include the variation
between the ages of patients, as some were in the young
adulthood period and others were still in adolescence. The
cohort did not receive a similar survey before index sur-
geries; therefore, it was not possible to compare the pro-
files to before the TGR instrumentation. As it is virtually
impossible to create a control group of idiopathic EOS
with similar characteristics who did not receive surgical
treatment, the findings of this study were compared with
the normative data of the general population readily
available in the literature, and the outcomes of treatment
could not be clearly delineated over the effects of the
disease itself. As this study does not include any MCGR
graduates, the authors recommend performing similar
studies for MCGR, as these results cannot be extrapolated
for MCGR graduates. Furthermore, the small sample size
of this cohort necessitates future studies with larger
sample sizes.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 81

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Yildiz et al
CONCLUSION
Completion of TGR lengthenings can provide
acceptable correction of deformities with good surgical
outcomes. However, the psychological health and well-
being of graduated patients may not always demonstrate
similar improvements. This is the first study to find various
psychosocial abnormalities in two-thirds of cases. Because
of the limited sample size in this study, future studies with
larger populations are needed. EOS should not be
considered solely as a skeletal system disorder, and the
authors recommend referral to psychological evaluation
before and after the treatment process.
ACKNOWLEDGMENTS
The authors thank Dr Mehmet Burke, Dr Erdem
Ertekin, and Dr Taha Aksoy for their contributions to the
data collection by handing the surveys when one of the
authors was abroad and unavailable.
REFERENCES
1. Akbarnia BA. Instrumentation with limited arthrodesis for the
treatment of progressive early-onset scoliosis. Spine: State of the Art
Rev. 2000;14:181–189.
2. Akbarnia BA, Marks DS, Boachie-Adjei O, et al. Dual growing rod
technique for the treatment of progressive early-onset scoliosis: a
multicenter study. Spine (Phila Pa1976). 2005;30(suppl_17):
S46–S57.
3. Akgül T, Dikici F, Sar C, et al. Growing rod instrumentation in the
treatment of early onset scoliosis. Acta Orthop Belg. 2014;80:457–463.
4. Aslan C, Olgun ZD, Ayik G, et al. Does decreased surgical stress really
improve the psychosocial health of early-onset scoliosis patients?: a
comparison of traditional growing rods and magnetically-controlled
growing rods patients reveals disappointing results. Spine (Phila Pa
1976). 2019;44:E656–E663.
5. Aslan C, Olgun ZD, Ertas ES, et al. Psychological profile of children
who require repetitive surgical procedures for early onset scoliosis: is
a poorer quality of life the cost of a straighter spine? Spine Deform.
2017;5:334–341.
6. Flynn JM, Matsumoto H, Torres F, et al. Psychological dysfunction
in children who require repetitive surgery for early onset scoliosis.
J Pediatr Orthop. 2012;32:594–599.
7. Matsumoto H, Williams BA, Corona J, et al. Psychosocial effects of
repetitive surgeries in children with early-onset scoliosis: are we
putting them at risk? J Pediatr Orthop. 2014;34:172–178.
8. Grabowski J, Goldin A, Arthur LG, et al. The effects of early
anesthesia on neurodevelopment: a systematic review. J Pediatr Surg.
2021;56:851–861.
9. Zheng XU, Zhengliang M, Gu X. Plasma levels of tumor necrosis
factor-α in adolescent idiopathic scoliosis patients serve as a predictor
for the incidence of early postoperative cognitive dysfunction
following orthopedic surgery. Exp Ther Med. 2015;9:1443–1447.
10. Hu D, Flick RP, Zaccariello MJ, et al. Association between exposure
of young children to procedures requiring general anesthesia and
learning and behavioral outcomes in a population-based birth
cohort. Anesthesiology. 2017;127:227–240.
11. Wilder RT, Flick RP, Sprung J, et al. Early exposure to anesthesia
and learning disabilities in a population-based birth cohort.
Anesthesiology. 2009;110:796–804.
82 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
12. Bucchianeri MM, Arikian AJ, Hannan PJ, et al. Body dissatisfaction
from adolescence to young adulthood: findings from a 10-year
longitudinal study. Body Image. 2013;10:1–7.
13. Mimura T, Ikegami S, Kuraishi S, et al. Residual thoracolumbar/
lumbar curve is related to self-image after posterior spinal fusion for
Lenke 1 and 2 curves in adolescent idiopathic scoliosis patients.
J Neurosurg Pediatr. 2020;26:211–216.
14. Celebioglu E, Yataganbaba A, Bekmez S, et al. Growing-rod
graduates with idiopathic early-onset scoliosis have comparable
exercise tolerance to patients with surgically treated adolescent
idiopathic scoliosis. J Pediatr Orthop. 2020;40:e734–e739.
15. Flynn JM, Tomlinson LA, Pawelek J, et al. Growing-rod graduates:
lessons learned from ninety-nine patients who completed length-
ening. J Bone Joint Surg Am. 2013;95:1745–1750.
16. Helenius IJ, Sponseller PD, McClung A, et al. Surgical and health-
related quality-of-life outcomes of growing rod “graduates” with
severe versus moderate early-onset scoliosis. Spine (Phila Pa 1976).
2019;44:698–706.
17. Hisli N. Beck Depresyon Envanterinin universite ogrencileri icin
gecerligi, guvenirligi [The reliability and validity of the Beck
Depression Inventory in university students]. Psikoloji Dergisi.
1989;7:3–13.
18. Ulusoy M, Sahin NH, Erkmen H. Turkish version of the Beck
anxiety inventory: psychometric properties. J Cogn Psychother.
1998;12:163–172.
19. Anil AE, Kivircik B, Batur S, et al. The Turkish version of the
auditory consonant trigram test as a measure of working memory: a
normative study. Clin Neuropsychol. 2003;17:159–169.
20. Aki OE, Alkan B, Demirsoz T, et al. Yas, Cinsiyet ve Egitim
Duzeyinin Fonemik ve Semantik Sozel Akicilik Testi Puanlari
Uzerinde Etkisi. [Effects of Age, Gender and Education on Phonemic
and Semantic Verbal Fluency]. Turk Psikiyatri Derg. 2022;33:53–64.
21. Epir S, İskit U. Wechsler Yetiskinler Zeka Olcegi Turkce Cevirisinin
On Analizi ve Universite Danismanlik Merkezlerindeki Uygulama
Potansiyeli. [The preliminary analysis of the Turkish translation of
the Wechsler Adult Intelligence Scale and the potential for its
implementation in University Counseling Centers]. Hacettepe Sosyal
ve Beseri Bilimler Dergisi. 1972;4:198–205.
22. Guvenir T, Ozbek A, Baykara B, et al. Psychometric properties of
the Turkish version of the strengths and difficulties questionnaire
(SDQ). Turk J Child Adolescent Mental Health. 2008;15:65–74.
23. Rosenberg M. Society and the adolescent self-image, 1st edition.
Princeton UniversityPres; 1965.
24. Dag I. Belirti Tarama Listesi (Scl-90-R)‘nin universite ogrencileri icin
guvenirligi ve gecerligi. [Symptom Checklist (SCL-90-R); A Reli-
ability and Validity Study]. Turk Psikiyatri Derg. 1991;2:5–12.
25. Kagan M, Gulec M, Boysan M, et al. Hierarchical factor structure of
the Turkish version of the posttraumatic growth inventory in a
normal population. TAF Prev Med Bull. 2012;11:617–624.
26. Alanay A, Cil A, Berk H, et al. Reliability and validity of adapted
Turkish version of scoliosis research society-22 (SRS-22) question-
naire. Spine (Phila Pa 1976). 2005;30:2464–2468.
27. Yapar A, Yapar D, Ergisi Y, et al. Reliability and validity of the
adapted Turkish version of the spinal appearance questionnaire.
Spine Deform. 2020;9:57–66.
28. Tedeschi RG, Calhoun LG. Trauma and transformation: growing in
the aftermath of suffering, 1st edition. SAGE Publications CA; 1995.
29. Mandal S, Basu M, Kirtania J, et al. Impact of general versus epidural
anesthesia on early post-operative cognitive dysfunction following hip
and knee surgery. J Emerg Trauma Shock. 2011;4:23–28.
30. Welsh MC, Labbe EE, Delayney D. Cognitive strategies and
personality variables in adherence to exercise. Psychol Rep.
1991;68:1327–1335.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Medium-term and Long-term Follow-up Surgical
Outcomes of the 1-stage Posterior-only Lumbosacral
Hemivertebra Resection With Short-segment Fusion
in Children
Rongxuan Gao, MD, Xuejun Zhang, MD, Dong Guo, MD, Jun Cao, MD, Ziming Yao, MD,
Yunsong Bai, MD, and Baosheng Sun, MD
Background The objective of this study was to evaluate the
medium-term and long-term surgical outcomes of the 1-stage
posterior-only lumbosacral hemivertebra resection with short-
segment fusion in children.
Methods: This retrospective chart review included 21 children
with congenital scoliosis due to lumbosacral hemivertebra who
received 1-stage posterior-only hemivertebra resection with
short-segment fusion from 2012 to 2016 with at least 5 years of
follow-up. Standing anteroposterior and lateral radiographs of
the spine were compared preoperatively, postoperatively, and at
last follow-up. Radiographic evaluation included measured
changes in segmental scoliosis and lordosis, compensatory sco-
liosis, thoracic kyphosis, lumbar lordosis, trunk shift, and sagittal
spinopelvic alignment.
Results: There were 12 boys and 9 girls with a mean age of
6.5 ± 3.2 years. The mean follow-up period was 6.7 ± 1.3 years.
The mean fusion level was 2.7 ± 0.9 segments. The mean seg-
mental scoliosis was 29 ± 6 degrees preoperatively, 9 ± 3 degrees
(correction rate of 71%) postoperatively (P < 0.05), and 7 ± 3
degrees (correction rate of 76%) at the latest follow-up. The
compensatory curve of 26 ± 12 degrees was spontaneously cor-
rected to 14 ± 8 degrees (correction rate of 47%) at last follow-up
(P < 0.05). Trunk shift was significantly improved on both co-
ronal (53%) and sagittal (56%) planes after surgery (P<0.05) and
stable at follow-up. The sagittal spinopelvic alignment was bal-
anced in all cases. There were no neurological or infectious
complications.
Conclusions: It is safe and effective to perform 1-stage posterior-
only lumbosacral hemivertebra resection with short-segment fu-
sion, which can significantly correct the segmental scoliosis,
prevent the compensatory curve progress and improve the trunk
From the Department of Orthopaedics, Beijing Children’s Hospital,
Capital Medical University, National Center for Children’s Health,
Beijing, PR China.
R.G. is considered as the first author.
None of the authors received financial support for this study.
The authors declare no conflicts of interest.
Reprints: Xuejun Zhang, MD, Department of Orthopaedics, Beijing
Children’s Hospital, Capital Medical University, National Center for
Children’s Health, No. 56, Nalishi Road, Beijing 100045, PR China.
E-mail: zhang-x-j04@163.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002263
e120 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
shift. This strategy also can save motion segments and avoid long
lumbar fusion. Medium-term and long-term follow-up outcomes
are satisfactory.
Level of Evidence: Level III.
Key Words: congenital scoliosis, lumbosacral hemivertebra,
posterior approach, short-segment fusion
(J Pediatr Orthop 2023;43:e120–e126)
C ongenital scoliosis is a spinal deformity caused by the
imbalance of longitudinal growth in coronal and/or
sagittal planes due to the disorders of vertebral body
formation or segmentation.1 Hemivertebra (HV) is one of
the important causes of congenital scoliosis, and it can be
located in the cervical, thoracic, or lumbar spine.2 Lum-
bosacral HV usually refers to HV at the lumbosacral
junction. The L5 wedge shape is also included in the
lumbosacral HV because it can produce clinical features
similar to lumbosacral HV deformity.3 Compared with
thoracolumbar HV, lumbosacral HV is relatively rare.
However, the lack of compensation from the relatively
fixed sacrum below the HV results in a severe proximal
lumbar spine tilt, which manifests as the “takeoff” sign
and causes a compensatory curve extending to the thor-
acolumbar segment or even the thoracic spine, in addition
to severe trunk shift.4 Therefore, HV resection and
short-segment fusion is the recommended treatment, and
satisfactory short-term results have been achieved.5,6
However, there are few reports on the 5 years or longer
outcomes including sagittal spinopelvic alignment, espe-
cially in the treatment of patients with sacral slanting. This
retrospective chart review evaluated the medium-term and
long-term surgical outcomes of 1-stage posterior-only
lumbosacral HV resection with short-segment fusion in
children.
METHODS
Participants
This study was approved by the institutional review
board of our hospital. Participants were 25 patients with
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023 Lumbosacral Hemivertebra Resection

lumbosacral HV who received surgical treatment at our Radiographic Data

hospital from 2012 to 2016. Inclusion criteria included a
diagnosis of congenital scoliosis due to lumbosacral HV,
treatment with 1-stage posterior-only HV resection with
short-segment fusion, fusion level of 4 or less segments,
and follow-up data of > 5 years. Exclusion criteria were
previous lumbosacral surgery, fusion level of > 4
segments, and incomplete clinical and imaging data.
A total of 21 children were included in the study,
including 12 boys and 9 girls. There were 4 patients were
excluded from this study because of fusion levels of >
4 segments and no patients were excluded due to <5 years
of follow-up. The characteristics of the study population
are described in Table 1. There were 14 patients with fully
segmented HV and 7 patients with incomplete segmented
HV. Eleven patients had left-sided HV, and 10 patients
had right-sided HV. The HV position was located at L5 in
8 patients, L5–S1 in 12 patients, and L6–S1 in 1 patient.
There were 5 patients with concomitant thoracic HV and 3
patients with other lumbar HV. One patient had car-
diovascular malformation (Tetralogy of Fallot), 1 patient
had urinary malformation (absence of right kidney), and
6 patients had nervous system malformation, including 3
patients with tethered spinal cord, 1 patient each with
tethered spinal cord combined with syringomyelia (L1–
L2), diastematomyelia combined with tethered spinal
cord, diastematomyelia combined with syringomyelia
(T8–T11). Neurological abnormalities do not affect HV
treatment in this study.
Standing posteroanterior and lateral radiographs were
taken before surgery, 3 months after surgery, and at the last
follow-up. All patients in our hospital were needed to have a
computed tomography (CT) scan and 3-dimensional (D)
reconstruction before surgery to evaluate the location and
segmentation of the HV. Preoperative magnetic resonance
imaging was performed to detect the presence of spinal cord
or other abnormalities.
Coronal and sagittal parameters were measured ac-
cording to the method described by Bollini et al.7 Seg-
mental scoliosis was measured between the superior
endplate of the 2 vertebrae adjacent to the HV. The
compensatory curve was the maximal curve angle of the
proximal spine. Coronal sacral slant (CSS) angle refers to
the angle between the line connecting the highest points of
bilateral iliac crests and the superior endplate of S1
(Fig. 1). Another name used for CSS is oblique takeoff.
CSS was defined as a CSS angle of > 5 degrees, and severe
CSS was defined as > 10 degrees. The upper instrumented
vertebra-pelvis angle (UPA) refers to the angle between
the upper endplate of the proximal fusion vertebra and the
highest points of bilateral iliac crests (Fig. 1). The gravity
trunk shift (GTS) was the distance between a plumb line
drawn from the middle of C7 body and the middle of
sacrum. When GTS was ≥ 20 mm, the coronal plane was
considered imbalanced. On the lateral view, segmental
kyphosis was measured as the angle between the superior
plate of the 2 vertebrae adjacent to the HV. Thoracic

TABLE 1. Basic Data and Surgical Information of 21 Children With Lumbosacral Hemivertebra
Associated Additional Operation Blood
Case Age HV Other Congenital Intraspinal Fusion Time Loss Follow-
No. Sex (y) Position Side Segmentation HV Comorbidities Anomalies UIV DIV Level (min) (mL) up (y)
1 F 5.75 L5–S1 L Fully — — — L5 S1 2 130 400 6.8
2 M 11.25 L5–S1 R Fully — — — L5 S1 2 300 1000 5.0
3 M 3.42 L5–S1 L Semi — — — L5 S1 2 120 200 5.3
4 M 4.50 L5–S1 R Fully — — — L5 S1 2 180 150 5.0
5 F 4.67 L5–S1 L Semi T7–T8 — — L4 S1 3 140 300 8.3
6 M 7.50 L5 R Fully T6, — — L4 S1 2 210 400 7.3
T10
7 F 4.75 L5 R Fully — — Tethered cord L4 S1 2 130 240 9.0
8 M 9.50 L5–S1 R Semi — — — L4 S1 3 200 410 6.7
9 M 2.17 L5–S1 R Semi T2– Absence of right — L4 S1 3 120 200 8.8
T3, kidney
T3–L4
10 M 2.92 L5–S1 L Fully L1–L2 — Tethered cord L5 S1 2 160 200 7.1
11 F 11.33 L6–S1 L Semi — — — L4 S1 4 150 800 6.7
12 M 4.08 L5–S1 L Fully L2 TOF — L4 S2 4 190 400 5.5
13 M 3.33 L5 L Semi L2–L3 — — L4 S1 2 150 100 5.8
14 F 5.08 L5 R Fully — — Syringomyelia L3 S2 4 150 300 8.8
15 F 10.25 L5–S1 L Fully T3 — — L4 S2 4 130 100 5.0
16 M 10.08 L5 L Fully — — — L4 S1 2 210 500 6.8
17 M 7.25 L5–S1 R Semi — — — L4 S1 3 170 600 7.3
18 F 3.33 L5 R Fully — — Diastematomyelia L4 S1 2 120 300 5.9
tethered cord
19 F 5.33 L5 L Fully — — Tethered cord L4 S1 2 240 100 7.9
20 M 13.00 L5–S1 L Fully T10 Diastematomyelia L5 S1 2 200 600 5.0
syringomyelia
21 F 7.08 L5 R Fully — — — L3 S1 3 140 450 6.7
DIV indicates distal instrumented vertebra; F, female; M, male; L, left; R, right; TOF, Tetralogy of Fallo; UIV, upper instrumented vertebra.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e121

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Gao et al
FIGURE 1. Line “A” runs parallel to the upper endplate of
proximal fusion vertebrae. Line “B” connects the highest
points of the iliac crests. Line “C” runs parallel to the superior
endplate of S1. Upper instrumented vertebra-pelvis angle: the
angle between the line “A” and “B.” Coronal sacral slant angle:
angle between the line “B” and “C.”
kyphosis (TK) and lumbar lordosis (LL) were also
measured and recorded. Sagittal vertebral axis (SVA)
was measured as the distance between C7 plumb line and
the posterior superior corner of S1. Sagittal spinopelvic
parameters, including pelvic incidence (PI), pelvic tilt
(PT), sacral slope (SS), were measured and recorded. Due
to the obvious coronal plane sacral tilt in some of the
patients, the upper endplate of S1 may appear as
2 projections on the lateral radiograph. In these cases,
we used the middle parallel line between the 2 projections
as the upper endplate of S1 to reduce measurement error.
Surgical Technique
All patients received 1-stage posterior-only HV re-
section, short-segment fixation with pedicle screws, and
bone graft fusion. Our indication for surgery is that the
curve to progress in children aged 4 to 5 years or older.
Most of our patients choose 2 segments for fusion, but for
those with large curvature or S1 dysplasia, we may choose
3-level or 4-level fusion. The patient was placed in the prone
position, and a standard back median incision was made.
The spinous process and lamina of the HV and the sched-
uled fusion segment were exposed by subperiosteal dis-
section. The pedicle screw was inserted under fluoroscopy.
If the S1 screw is unstable, it can be fixed to the ilium with
the sacral-alar-iliac screw. The corresponding lamina and
transverse process of the HV were then removed, and the
lateral part of the HV was stripped along the pedicle sub-
periosteally to protect the spinal cord. The HV was re-
moved completely under direct visualization, and the
residual bone of the HV was curetted. The cancellous bone
was preserved for grafting. The upper and lower disks, in-
cluding the cartilage endplate, were completely removed
e122 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
until the bleeding bone was reached. The opposite disk was
then released. Partial laminectomy of the upper and lower
vertebrae was performed. A precontoured rod was
connected to the screws on the convex side. Gradual
compression was applied, while leaving the concave rod
unlocked until the gap was closed. The correction was
carried out alternately. Fluoroscopy was used to ensure that
the balance of the coronal plane of the spine and the upper
and lower endplates of the osteotomy space were parallel.
The HV with sacral slanting and the segmental scoliosis
should be corrected according to UPA, and the fixed ver-
tebrae should be placed in the horizontal neutral position as
far as possible. The lamina of the vertebral body in the fixed
segment was decorticated. Autogenous and/or allogeneic
bone was implanted for the anterior and posterolateral
fusion. Spinal somatosensory evoked potential and motor
evoked potential were used in all cases. After surgery, the
patient was immobilized in bed for 4 weeks. After 4 weeks,
the patient started to walk with braces, and the braces were
used for at least 3 months.
Follow-up
The patients were seen every 3 months for the first
year after surgery and every other year thereafter. Full
spinal anteroposterior and lateral radiographs were taken
after the surgery and at every follow-up. Detailed physical
and neurological examinations were also performed.
Statistical Analysis
Data were analyzed using IBM SPSS 19.0 statistical
software, IBM Corp., Armonk, NY, USA. All measure-
ment data are expressed as mean ± SD. Paired t test was
used to compare the imaging results of patients pre-
operatively, postoperatively, and at last follow-up. A
2-sided value of P value <0.05 was considered to indicate
statistical significance.
RESULTS
A total of 21 children were enrolled in this study. The
age at the time of surgery was a mean age of 6.5 ± 3.2 years
(range: 2.2 to 13.0 y). The mean follow-up was
6.7 ± 1.3 years (range: 5.0 to 9.0 y). The mean fusion level
was 2.7 ± 0.9 segments (range: 2 to 4 segments). The aver-
age operation time was 168.6 ± 46.2 minutes (range: 120 to
300 min). The average estimated intraoperative blood loss
was 368.0 ± 234.6 mL (range: 100 to 1000 mL). Intra-
operative blood salvage was used in all patients resulting in
an average of 188.2 ± 112.9 mL (range: 69 to 500 mL) of
autologous blood transfused. Only 1 case received 200 mL
of fresh frozen plasma in addition to 500 mL of autologous
blood because of 1000 mL of bleeding (Table 1).
The imaging parameters of all the patients
preoperatively, postoperatively, and at last follow-up are
shown in Table 2. There were statistically significant
differences in coronal plane segmental scoliosis, com-
pensatory curve, SSA, UPA, GTS, and SVA after surgery
and at the last follow-up compared with before surgery
(P < 0.05). However, there were no significant differences in
coronal plane segmental scoliosis, proximal compensatory
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Lumbosacral Hemivertebra Resection

vs. Last Follow-up)

CSS indicates coronal sacral slant; GTS, gravity trunk shift; LL, lumbar lordosis; PI, pelvic incidence; PT, pelvic tilt, SS, sacral slope; SVA, sagittal vertebral axis; TK, thoracic kyphosis; UPA, upper instrumented
curvature, SSA, UPA, GTS, and SVA after surgery
P (Postoperative

compared with the last follow-up (P > 0.05), indicating no

> 0.05

> 0.05

> 0.05
> 0.05
> 0.05

> 0.05

> 0.05
> 0.05
> 0.05

> 0.05
> 0.05
> 0.05
significant loss of correction (Fig. 2). The TS and SVA were
significantly improved after surgery (P < 0.05) and were
stable at follow-up. Coronal imbalance occurred in 6 cases
(29%), and sagittal imbalance occurred in 1 case (5%) before
surgery. The correction rates of GTS and SVA were 43%
Correction

and 51% postoperatively and 53% and 56% at the last

Rate (%)

75.8

46.9

43.8
65.7
53.1

13.5

11.3
12.9
56.3

4.8
9.8
3.2
follow-up, respectively. There were 16 patients with CSS
(76%) preoperatively. Among them, 9 cases (43%) had
severe CSS. There were no significant differences in sagittal
plane segmental kyphosis, LL, TK, and sagittal spinopelvic
P (Preoperative Last Follow-up [Mean ± SD

8.6 ± 7.5 (−15.1 to 25.5) parameters (PI, PT, SS) postoperatively and at the last
follow-up compared with before surgery (P > 0.05).
36.3 ± 11.1 (21.4-59.2)

45.5 ± 8.9 (32.1-62.8)

32.7 ± 6.2 (22.7-45.3)

7.0 ± 3.4 (0.4-12.3)

13.7 ± 8.3 (2.2-29.6)

6.8 ± 4.5 (0.4-14.7)

6.8 ± 4.8 (1.5-19.3)

18.5 ± 7.5 (3.3-32.8)

18.4 ± 9.8 (4.0-34.9)

13.4 ± 5.7 (2.9-26.4) There was no infection, pseudarthrosis, correction

5.0 ± 3.8 (0-13.8)
(Range)]

loss, or instrumentation-related complication noted during

follow-up. Two patients had transient lower limb pain
after surgery, which was considered to be related to L5
nerve root traction. After conservative treatment, both
patients’ pain resolved 1 week after surgery.
vs. Postoperative

DISCUSSION
< 0.05

< 0.05

< 0.05
< 0.05
< 0.05

> 0.05

> 0.05
> 0.05
< 0.05

> 0.05
> 0.05
> 0.05

Lumbosacral HV is more difficult to treat surgically

than lumbar or thoracic HV.8 McMaster and Singh2 re-
ported that segmental scoliosis caused by a single, fully
segmented lumbosacral HV can worsen by 1.5 degrees per
Correction

year, while the secondary proximal compensatory curve

Rate (%)

70.6

45.7

41.6
69.7
43.4

10.4

16.6
51.3

11.5
9.2

3.5

1.3

can progress at a rate of 3 degrees per year becoming

structural scoliosis.9 Removing the HV and the upper and
lower ring apophyses and disks, especially for children
with coronal plane imbalance, which can achieve imme-
Postoperative [Mean ± SD

diate good correction.10,11 HV resection is usually per-

9.6 ± 8.2 (−16.1 to 26.8)

formed by staging anterior and posterior approach in the

35.6 ± 10.3 (19.6-52.8)

44.9 ± 9.8 (31.1-62.9)

31.3 ± 6.0 (23.1-44.1)

8.5 ± 3.0 (0.5-14.0)

14.0 ± 7.3 (2.5-30.5)

5.2 ± 3.4 (0.5-14.0)

6.0 ± 4.2 (0.6-13.4)
8.2 ± 5.7 (1.8-20.4)

18.0 ± 7.2 (3.1-32.4)

19.0 ± 8.6 (3.3-32.7)

13.6 ± 6.0 (2.7-25.8)

early stages.12 However, there is a relatively high surgical

(Range)]

risk due to the complex anatomy found in the lumbosacral

anterior approach, and it is easy to cause vascular and
nerve injury.7,13 Thus, 1-stage posterior-only HV resection
has gradually replaced the combined anterior and poste-
TABLE 2. Radiographic Characteristics of the Patients

rior approach and is the first line for the treatment of

lumbosacral HV.5,6 Zhuang et al5 described 14 children
with lumbosacral HV who underwent 1-stage posterior-
Preoperative [Mean ± SD

19.7 ± 15.4 (−34.2 to 67.1)

only HV resection and short-segment fusion with a follow-

up of at least 2 years. The mean fusion level was 3.2
32.6 ± 11.4 (11.4-51.7)

43.4 ± 10.3 (28.4-66.9)

28.9 ± 6.3 (19.9-42.7)

25.8 ± 11.7 (6.5-48.0)

31.7 ± 6.4 (21.1-47.3)

14.5 ± 12.5 (2.2-48.4)
8.9 ± 5.4 (0.2-21.2)
19.8 ± 6.5 (4.6-30.7)

16.3 ± 8.0 (0.2-31.4)

16.3 ± 9.8 (2.3-31.3)

12.2 ± 7.6 (1.0-30.3)

segments. At the last follow-up, the correction rate of

(Range)]

segmental scoliosis was 87%, the correction rate of the

compensatory coronal curve was 57%, and the correction
Sagittal spinopelvic parameters (deg.)

rate of GTS was 58%. Yaszay et al14 found that lumbo-

sacral HV resection can significantly improve the lumbo-
sacral balance and restore the coronal plane balance in
children. In our study, 21 children with lumbosacral HV
were followed up for at least 5 years. The mean fusion
Segmental scoliosis

kyphosis (deg.)

vertebra-pelvis angle.

level was 2.7 ± 0.9 segments. The correction rate of seg-

Compensatory
curve (deg.)

mental scoliosis and compensatory curve were 76% and

Coronal plane

UPA (deg.)

Sagittal plane
GTS (mm)

SVA (mm)
CSS (deg.)

Segmental

TK (deg.)
LL (deg.)

47% at the last follow-up, respectively. The correction rate

(deg.)

of GTS and SVA were 53% and 56% at the last follow-up,
respectively. These results show high success of this pro-
PT
SS
PI

cedure and is consistent with previous findings.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e123

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Gao et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 2. A case with a single left fully segmented L5–S1 hemivertebrae. A, Preoperative posteroanterior standing radiographs
showed lumbosacral scoliosis and compensatory curve. B, Preoperative lateral radiograph. C and D, Preoperative computed
tomography scan and reconstruction. E and F, Postoperative posteroanterior and lateral standing radiograph showed that seg-
mental scoliosis and compensatory curve were well corrected. G and H, Postoperative 6 years of follow-up posteroanterior and
lateral radiographs of spinal column.

Radiographic parameters such as CSS and UPA to
evaluate PT and upper instrumented vertebra (UIV) hori-
zontalization has also been reported in similar articles and
can be used to evaluate surgical results.15 The incidence of
CSS in children with congenital scoliosis is 61.9%, and in
children with L3 and lower HV, the incidence is as high as
73.7%.16 As the base of the spine, the sacrum plays an
important role in maintaining the balance of the spine. All
21 cases in our study had lumbosacral HV with HV posi-
tion below L3. Sixteen cases (76%) were accompanied with
CSS. Nine cases (43%) were accompanied by severe CSS.
The influence of CSS must be considered during surgery.
Closing the osteotomy space after HV resection can ach-
ieve the maximum segmental correction. However, the
CSS can lead to tilting of the upper endplate of the UIV,
which will lead to proximal compensatory scoliosis.
Therefore, when correcting the lumbosacral HV, segmental
scoliosis should be corrected properly according to the
degree of CSS. The HV bone block, interbody fusion cage,
or mesh cage was used to maintain the upper endplate of

e124 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
FIGURE 3. A case using a sacral-alar-iliac screw. Preoperatively the anteroposterior radiographs (A) and computed tomography (B)
showed a single right fully segmented L5–S1 hemivertebrae. After hemivertebrae resection and distal fixation of S1 and the ilia, the
anteroposterior radiographs of spinal column showed satisfactory correction and instrumentation 1 week (C) and 5 years (D) after
surgery.
the UIV in a horizontal neutral position to provide a
balanced base for the proximal spine and avoid post-
operative compensatory scoliosis and coronal plane
imbalance. In this study, the postoperative UPA was sig-
nificantly lower than that before surgery, which ensured
that the upper endplate of the UIV was horizontal to
balance the coronal plane.
Due to the lack of sacral compensation, the lum-
bosacral HV often causes obvious proximal compensatory
scoliosis and trunk shift.5–14 Since the position of lumbo-
sacral HV is low, its influence on the overall sagittal plane
is relatively small, and the sagittal plane balance of most
lumbosacral HV patients is normal.4 In our study, there
were no significant differences in the sagittal plane seg-
ment kyphosis, LL, and TK preoperatively, post-
operatively, and at the last follow-up, and all
measurements were within the normal range. However,
there are few studies on the influence of lumbosacral HV
on sagittal spinopelvic alignment before and after surgery.
PI, SS, and PT were used as sagittal spinopelvic parame-
ters in this study. In children and adolescents, PI increases
with age and will remain unchanged when bone matures.17
The PI value of normal adults is about 51 to 55
degrees,18,19 and Fei et al20 reported that the PI of normal
young adults in China is about 47.2 ± 8.8 degrees. The
reference range of PI of children and adolescents in China
has not been reported. The PI of the 21 patients in this
study before surgery was 43 ± 10 degrees, which was
slightly lower than the normal adult value in China but
within the normal range. Meanwhile, there was no sig-
nificant differences in PI, PT, and SS postoperatively and
at the last follow-up compared with before surgery
(P > 0.05), which suggested that the lumbosacral HV
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Lumbosacral Hemivertebra Resection
would not cause obvious imbalance on sagittal spinopelvic
alignment. Yang et al21 reported that lumbosacral HV
with other vertebral deformities can lead to pelvic ante-
version or retroversion. In our study, there were 8 patients
with additional HV, including 5 with thoracic HV and 3
with other lumbar HV, which did not affect the sagittal
pelvic balance. There are 3 possible reasons for this. First,
the cases with other HV did not have complex lumbo-
sacral deformities and thus had little influence on sagittal
spinopelvic alignment. Second, all cases in this study in-
volved short-segment fixation and fusion, and the max-
imum possible number of lumbar moving segments were
preserved to minimize the influence of surgery on sagittal
spinopelvic alignment. Third, most of the patients had
mild scoliosis, so there was no obvious compensatory
pelvic rotation and pelvic imbalance.
Due to the young age of children with lumbosacral
HV receiving surgery, the surgical field is small, and the
pedicle is weaker, especially S1 pedicle. Therefore, there is
a high risk of vascular and nerve injury during surgical
exposure, pedicle screw placement, and HV resection.22
Thus, the segmentation of HV and the diameter and length
of pedicle should be fully evaluated with spinal CT scan
and 3D reconstruction and magnetic resonance imaging
before surgery. A detailed operative plan should be made
to improve the accuracy of intraoperative pedicle screw
placement and reduce neurological complications. Due to
the characteristics of the sacrum in children, it is important
ensure the stability of the sacral fixation. Our experience is
to increase the length of the screw as much as possible by
inclining the pedicle screw caudally and placing it along the
maximum meridian of the sagittal position of S1, which
can increase the stability of the screw while determining the
www.pedorthopaedics.com | e125
Gao et al
diameter of the screw. If the S1 screw is unstable, it can be
fixed to the ilium with the sacral-alar-iliac screw (Fig. 3). In
recent years, we used 3D modeling based on the
preoperative CT data to improve the efficiency and
accuracy of pedicle screw placement. This was
supplemented using the intraoperative navigation robot
in 2019. In addition, somatosensory evoked potential and
motor evoked potential were monitored intraoperatively
for all enrolled patients to reduce the risk of spinal cord
injury.23 In this study, there were no complications of
severe nerve injury. Only 2 cases presented with transient
lower extremity pain on the HV side, which resolved with
conservative treatment. This pain was likely secondary to
intraoperative nerve root traction and inadequate
decompression of the nerve root. In addition, there was
no infection, pseudarthrosis, correction loss, or
instrumentation-related complications during the follow-
up. However, because the lumbosacral segment has a large
range of motion, special biomechanical characteristics,
poor bone strength in children, and high technical
requirements for screw implantation and is located in a
concentrated stress area between the lumbar vertebrae and
the sacrum, there is a higher incidence of postoperative
internal fixation failure than for other spinal positions.6
Therefore, bone graft fusion after fixation is particularly
important, especially the intervertebral bone graft. The
resected HV is bitten to a bone block and implanted into
the vertebral space to promote the fusion of the anterior
column. If the bone block is too small, an interbody fusion
or mesh cage can be used. At the same time, a bone block,
interbody fusion cage, mesh cage can also reconstruct LL.
Last, postoperative care is also very important. All patients
wore protective braces for 3 months after surgery to reduce
the incidence of postoperative instrumentation-related
complications.
This study has several limitations. First, the sample size
is small. Second, most of the patients are still in the growth
stage, so continued follow-up will be needed to assess out-
comes after reaching bone maturity. Third, our hospital has
started to use the 24-Item Early-Onset Scoliosis Ques-
tionnaire (EOSQ-24) scale, but this group of cases lacked the
preoperative quality of life assessment, which could not be
statistically analyzed. We will add personal assessment
quality of life evaluations in future studies.
In summary, it is safe and effective to perform
1-stage posterior-only lumbosacral HV resection with
short-segment fusion, which can significantly correct seg-
mental scoliosis, prevent the compensatory curve progress,
and improve the trunk shift. This strategy also can save as
many motion segments as possible and avoid long lumbar
fusion. Medium-term and long-term follow-up outcomes
are overall satisfactory.
REFERENCES
1. Daniel H, John E. Congenital scoliosis: a review and update.
J Pediatr Orthop. 2007;27:106–116.
e126 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
2. McMaster MJ, Singh H. Natural history of congenital kyphosis and
kyphoscoliosis. A study of one hundred and twelve patients. J Bone
Joint Surg Am. 1999;81:1367–1383.
3. Bollini G, Docquier PL, Viehweger E, et al. Lumbosacral hemi-
vertebrae resection by combined approach: medium- and long-term
follow-up. Spine. 2006;31:1232–1239.
4. Slabaugh PB, Winter RB, Lonstein JE, et al. Lumbosacral hemi-
vertebrae. A review of twenty-four patients, with excision in eight.
Spine. 1980;5:234–244.
5. Zhuang Q, Zhang J, Li S, et al. One-stage posterior-only lumbosacral
hemivertebra resection with short segmental fusion: a more than
2-year follow-up. Eur Spine. 2016;25:1567–1574.
6. Lyu Q, Hu B, Zhou C, et al. The efficacy of posterior hemivertebra
resection with lumbosacral fixation and fusion in the treatment of
congenital scoliosis: a more than 2-year follow-up study. Clin Neurol
Neurosurg. 2018;164:154–159.
7. Bollini G, Docquier PL, Viehweger E, et al. Lumbar hemivertebra
resection. J Bone Joint Surg Am. 2006;88:1043–1052.
8. Li Y, Wang G, Jiang Z, et al. One-stage posterior excision of
lumbosacral hemivertebrae: retrospective study of case series and
literature review. Medicine. 2017;96:e8393.
9. McMaster MJ, David CV. Hemivertebra as a cause of scoliosis. A
study of 104 patients. J Bone Joint Surg Br. 1986;68:588–595.
10. Holte DC, Winter RB, Lonstein JE, et al. Excision of hemivertebrae
and wedge resection in the treatment of congenital scoliosis. J Bone
Joint Surg Am. 1995;77:159–171.
11. Ruf M, Jensen R, Letko L, et al. Hemivertebra resection and
osteotomies in congenital spine deformity. Spine. 2009;34:1791–1799.
12. Bollini G, Docquier PL, Viehweger E, et al. Thoracolumbar
hemivertebrae resection by double approach in a single procedure:
long-term follow-up. Spine. 2006;31:1745–1757.
13. Nakamura H, Matsuda H, Konishi S, et al. Single-stage excision of
hemivertebrae via the posterior approach alone for congenital spine
deformity: follow-up period longer than ten years. Spine (Phila Pa
1976). 2002;27:110–115.
14. Yaszay B, O’Brien M, Shufflebarger HL, et al. Efficacy of
hemivertebra resection for congenital scoliosis: a multicenter retro-
spective comparison of three surgical techniques. Spine. 2011;36:
2052–2060.
15. Wang Y, Shi B, Liu Z, et al. The upper instrumented vertebra
horizontalization: an essential factor predicting the spontaneous
correction of compensatory curve after lumbosacral hemivertebra
resection and short fusion. Spine. 2020;45:E1272–E1278.
16. Yanbin Z, Jiangguo Z, Guixing Q, et al. Surgical strategy for sacral
slanting in early-onset congenital scoliosis with lumbar hemivertbra.
Chin J Bone Joint Surg. 2018;11:192–196.
17. Mac-Thiong JM, Berthonnaud E, Dimar JR, et al. Sagittal align-
ment of the spine and pelvis during growth. Spine. 2004;29:
1642–1647.
18. Labelle H, Roussouly P, Berthonnaud E, et al. Spondylolisthesis,
pelvic incidence, and spinopelvic balance: a correlation study. Spine.
2004;29:2049–2054.
19. Boulay C, Tardieu C, Hecquet J, et al. Sagittal alignment of spine
and pelvis regulated by pelvic incidence: standard values and
prediction of lordosis. Eur Spine. 2006;15:415–422.
20. Fei H, Li W-S, Sun Z-R, et al. Analysis of spino-pelvic sagittal
alignment in young chinese patients with lumbar disc herniation.
Orthop Surg. 2017;9:271–276.
21. Yang K, Tao H, Li H, et al. The influence of congenital lumbosacral
deformity on pelvic balance. Chin J Spine Spinal Cord. 2018;28:
1089–1094.
22. Crostelli M, Mazza O, Mariani M. Posterior approach lumbar and
thoracolumbar hemivertebra resection in congenital scoliosis in
children under 10 years of age: results with 3 years mean follow up.
Eur Spine. 2014;23:209–215.
23. Chang DG, Kim JH, Ha KY, et al. Posterior hemivertebra resection
and short segment fusion with pedicle screw fixation for congenital
scoliosis in children younger than 10 years: greater than 7-year
follow-up. Spine. 2015;40:E484–E491.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Treating Pediatric Irreducible Atlantoaxial Rotatory Fixation
(IAARF) by Unlocking Facet Joint Through Transoral
Approach and Fixing With Slim-TARP Plate
(15 Cases Series)
Jianhua Wang, MD,*†‡ Hong Xia, MD,*† Xiang yang Ma, MD,*† SuoChao Fu, MD,*†
Qiang Tu, MD,*† Junjie Xu, MD,*† HongLei Yi, MD,*† and Changrong Zhu, MD*†
Background: Irreducible atlantoaxial rotatory fixation (IAARF)
often requires surgical treatment. Transoral unlocking the facet
joints is a key measure for the treatment of IAARF. We investigate
a novel method for treating pediatric IAARF by unlocking facet
joint through transoral appraoch and fixed with slim-tarp plate in
same stage with same approach.
Objective: The objective of this study is to investigate the method
and efficacy of a unique transoral approach to unlock facet joints
and fixation with slim-shaped transoral anterior reduction plate
(slim-TARP) plate in the treatment of IAARF.
Methods: Fifteen patients diagnosed with AARF were transferred
to our hospital. After 1 week of bidirectional cervical cranial
traction, they were diagnosed with irreducible AARF that, and
then underwent transoral release and fixation with slim-TARP
plate procedures. Postoperative computed tomography and mag-
netic resonance were used to evaluate the reduction effect, bone
fusion, and fusion time. Japanese orthopaedic association scores
were used to compare the recovery of spinal cord function in
patients before and after surgery. Complications such as wound
infection, neurovascular injury, and loosening of internal fixation
were evaluated too.
Results: All 15 patients underwent transoral unlocking facet joint
and fixation with slim-TARP procedures smoothly. The operation
time were 129.2 ± 11.9 minutes, blood loose were 83 ± 23 mL.
There were no neurological injury, wound infections, verified or
suspected vertebral artery injury, etc. All patients were followed up
for a mean of 17.8 ± 6.6 months (range: 12 to 36 mo). Bony fusion
was achieved in all patients. Mean fusion time was 3.6 ± 1.2 months
(range: 3 to 6 mo). Complete correction of torticollis was achieved
in all 15 cases. Preoperative symptoms of neck pain and limitation
of neck movement were effectively alleviated at 3 months after
From the *Department of Orthopedics, General Hospital of Southern
Theater Command of PLA; †Institute of Traumatic Orthopaedics of
People’s Liberation Army, Guangzhou, China; and ‡First School of
Clinical Medicine, Southern Medical University. Guangzhou, China.
No funds were received in support of this work.
The authors declare no conflicts of interest.
Reprints: Jianhua Wang, MD, Department of Orthopedics, General
Hospital of Southern Theater Command of PLA, No.111 Liuhua
Road, Yuexiu District, Guangzhou, Guangdong Province, 510010,
China. E-mail: jianhuawangddrr@163.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002307
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
surgery. The 3 patients with preoperative neurological deficits had
significant relief after surgery, and their latest follow-up results
showed that their Japanese orthopaedic association scores
increased from 13.0 ± 1.0 to 16.3 ± 0.6.
Conclusions: Transoral release and fixation with slim-TARP
plate by transoral approach is a feasible and safe method for
treating pediatric irreducible atlantoaxial rotatory fixation.
Key Words: irreducible atlantoaxial rotatory fixation (IAARF),
transoral atlantoaxial reduction and fixation plate (TARP), unlock
facet joint
(J Pediatr Orthop 2023;43:83–90)
tlantoaxial rotatory fixation (AARF) is pathologic
A rotation of the atlas around the axis with atlantoaxial
subluxation or dislocation. Corner1 offered the first de-
scription of this disorder in 1907. Fielding and Hawkins2
described AARF in 1977 as persistent rotatory subluxation
and dysfunction causing torticollis. Most AARF occurs in
pediatric patients, which may cause torticollis and neck
pain because of dislocation or subluxation of the atlan-
toaxial joint. AARF is often caused by an upper airway
infection such as laryngopharyngeal inflammation or by a
mild trauma, and congenital malformation in the cranio-
vertebrae junction(CVJ) also have some relation with it.
Fielding and Hawkins classified AARF into 4 types based
on the characters of images (Fig. 1). In this widely accepted
classification, types I and II have a normal atlantoaxial
interval (< 5 mm). However, type III displays bilateral
anterior facet displacement with an atlantoaxial interval
> 5 mm due to a deficiency of the transverse ligament. Type
IV is an unusual type, with a deficient odontoid and
posterior dislocation of the atlas. Unlike types I and II,
which are rotatory displacements around an intact
odontoid articulation, types III and IV have planar
instability, which allows translation between the atlas and
the axis particularly in the sagittal plane and can result in
narrowing of the space available for the cord at the
atlantoaxial level. In type I or II AARF, the atlantoaxial
stabilizing structures, particularly the transverse ligament
or the odontoid, remain intact. In type III, the intrinsic
www.pedorthopaedics.com | 83
Wang et al
atlantoaxial structures (including the transverse ligament
and alar ligaments) have been affected by the pathologic
process and atlantoaxial stability is thus compromised.
Type IV cases with a deficient odontoid also result in
atlantoaxial instability. These 2 latter conditions present as
torticollis with an unstable atlantoaxial joint in the sagittal
plane and have very different biomechanical features
compared with types I and II. Wang et al3 referred to
clinical torticollis secondary to atlantoaxial rotation,
together with an unstable atlantoaxial joint as
atlantoaxial rotatory fixed dislocations. They proposed
surgery treatment for fielding type III and type IV AARF
(or atlantoaxial rotatory fixed dislocations), because the
existing instability between C1 and C2 must be
immobilization and fusion.
According to the time of diagnose, the AARF can be
divided into acute and chronic AARF,4–6 acute AARF-
referred patient was diagnosed within 2 weeks, and
Chronic AARF, also known as the long-standing fixation,
is typically defined as a rotatory subluxation with dura-
tion > 2 to 3 months. The pathophysiology of chronic
AARF remains unclear despite the numerous previous
studies regarding AARF. In chronic cases, closed reduc-
tion and its maintenance are often unsuccessful, requiring
surgical treatments for such patients with chronic irredu-
cible or recurrent unstable AARF.
Early diagnosis and treatment is important for
AARF, most patients can be successfully treated and
cured in early stage with conservative method such as
traction or cervical brace immobilization,7–11 but a delay
in diagnosis and early treatment of AARF may lead to
recurrence and irreducible AARF, which need surgery in
the end. Goel et al12 noted that a locking of the atlan-
toaxial facet joints mechanical may existed in the hinding
for reduction of atlantoaxial rotatory dislocation, treat-
ment of irreducible AARF with a locked C1–C2 lateral
facet is more troublesome as the locked lateral facet must
be distracted and released first. He performed surgery
by direct physical manipulation of the facets to effect
realignment. Qi et al13 also noticed irreducible AARF
with a locked C1–C2 lateral facet and propose facet joint
release by retropharyngeal approach combined with
FIGURE 1. Fielding classification of atlantoaxial rotational dislocation. A, Type I: the lateral mass of the atlas is displaced anteriorly
on one side, while the other side has not moved and acts as the rotation axis (ADI < 3 mm). B, Type II: the lateral mass atlas is
displaced anteriorly on one side, whereas the other side has not moved and acts as the rotation axis (ADI 3 to 5 mm). C, Type III:
the atlas is rotated with the lateral mass displaced anteriorly on both sides (ADI > 5 mm). D, Type IV: the atlas is rotated and
displaced posteriorly. ADI indicates atalas-dens indice.
84 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
posterior instrumentation for treatment. Wang3 et al
recommended treating the irreducible AARF by transoral
release combined with posterior fixation and fusion techs.
We consider that the combined anterior and posterior
procedures are complicate and challenge techs with higher
morbidity and longer operation time. If we can unlock the
facet joint and realize the reduction, fixation by only 1
approach in the same stage, it will be an ideal choice for
children. The aim of this study is to present our methods
for treating pediatric irreducible atlantoaxial rotatory
fixation (IAARF) by anterior release and fixation techs
through transoral approach.
METHODS
General Material
From October 2016 and June 2019, 15 consecutive
patients with chronic AARF and a locked unilateral C1–C2
lateral facet were referred to our hospital (Table 1). There
were 5 males and 10 females, with a mean age of
9.1 ± 2.0 years (range: 6.2 to 13 y). The mean interval
from the onset of symptoms to admission to our hospital
was 4.5 ± 1.3 months (range: 3.2 to 8.0 mo). And all patients
had neck pain and torticollis, with varying extent limited
cervical range of motion for at least 3months. The causes of
the AARF and resultant torticollis were different for each
patient: 5 patients had a minimal wound in the cervical or
head, 4 patients had acute laryngitis, 3 patients had a
history of Down syndrome, 1 patient had a history of neck
surgery, and 2 patients had unknown reasons, except 3
patients of AARF have OS odeium, the other 12 patients
show no congential malformation in the CVJ; 3 patient had
fielding type II atlantoaxial rotatory subluxation, whereas
12 patients had fielding type III. All the patients underwent
transoral releasing and fixation techs with slim-shaped
transoral anterior reduction plate (slim-TARP) procedures.
Preoperative assessment included computed tomography
(CT), computed tomography angiography (CTA), and
magnetic resonance imaging (MRI). We can assess
variations of the vertebral arteries in the CVJ by CT
angiography to avoid vertebral artery injury during surgery,
and understand cervical spinal cord compression by cervical
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Treating Pediatric IAARF by Transoral Unlocking Facet Joint

TABLE 1. Treatment of pediatric irreducible Atlantoaxial rotatory fixation (IAARF) by transoral facet joint (15 cases)
Myelopathy
Duration Fielding C2 Anterior Operation Blood Follow-up Bone Fusion
Case Age Sex Etiology (mo) Types Pre Pos Fixation Pedicle Screw Time Loss (mo) Time (mo)
1 6.3 F AL 4.2 II N N Slim- 0 125 75 12 3
TARP
2 8.2 M MW 5.3 III 13 16 Slim- 2 135 85 13 3
TARP
3 8.5 F AL 5.1 III N N Slim- 1 138 65 16 3
TARP
4 9.5 F AL 6.4 III N N Slim- 2 145 68 11 3
TARP
5 11.3 M MW 3.3 III 12 16 Slim- 1 128 60 14 3
TARP
6 10.4 F MW 5.4 II N N Slim- 2 109 65 24 6
TARP
7 7.1 F MW 4.0 III N N Slim- 1 121 85 12 3
TARP
8 6.2 M DS 3.4 III N N Slim- 1 132 88 36 3
TARP
9 8.5 F AL 3.4 III N N Slim- 2 129 145 15.5 3
TARP
10 10.3 F DS 4.3 III 14 17 Slim- 1 112 105 13 3
TARP
11 12.3 M HNS 4.2 III N N Slim- 2 137 82 18 3
TARP
12 13 F MW 8.0 II N N Slim- 1 139 78 17 6
TARP
13 9.5 M DS 3.3 III N N Slim- 1 145 50 20 3
TARP
14 7.6 F UR 3.2 III N N Slim- 1 109 105 22 6
TARP
15 8.3 F UR 4.3 III N N Slim- 2 135 95 24 3
TARP
AL indicates acute laryngitis; DS, Down syndrome; HNS, history of neck surgery; JOA, Japanese orthopaedic association scores; MW, minimal wound in the cervical or
head; Pos, postsurgery; Pre, presurgery; slim-TARP, slim-shaped transoral anterior reduction plate, Weigao, China, patented by Jian and Yin.

MR images. The height and length of the C1 lateral mass,
C2 pedicle widths, or C2 vertebral lengths were measured
on the CT images. Atlantodental rotation was assessed on
axial CT, and dynamic radiographs (flexion and extension)
were utilized to identify atlantoaxial instability and confirm
the fixation of rotatory atlanaxial dislocation. This report
has been approved by our Institutional Review Board.
Preoperative Preparation and Management
After admission, all the 15 patients underwent bi-
directional cervical traction (Fig. 2) for 1 week (vertical
traction of 2 to 3 kg and longitudinal traction of 15 to 20%
body weight). They were subjected to radiography every
2 days to estimate the effectiveness of the traction. If
traction had failed to obtain reduction, then we decided
surgery and all the patients who reduced with traction
were excluded from this report. Preoperative oral cavity
cleaning was conducted with chlorhexidine 3 to 4 days
before the surgery, and dental cleaning was performed by
dentist.
Surgical Procedures
All the patients underwent transoral approach for
AARF reduction and fixation with plate procedures. After
the induction of general anesthesia with trans nasal
endotracheal intubation and administration of muscle re-
laxants, the patient was placed supine with the head in a
slightly extended position with skull traction (6 to 8 kg). A
midline incisions of ~4 to 5 cm were made along the
posterior pharyngeal wall. The long us capitals and long
us colli muscles were detached bilaterally to expose the
anterior arch of the atlas, the C2 vertebral body, and both
lateral atlantoaxial joints. After cleaning the entrapped
scar or contracture ligamentous and capsular tissue in the
lateral mass joints from the rotatory fixation, a osteotome
was inserted into the facets joint gap for distracting and
unlocking the locked joint effectively (Fig. 3A). After that,
the cartilage of lateral joints was removed by high-speed
drill, and 2 pieces of bone block harvested from iliac crest
were implanted into both sides of the atlantoaxial lateral
mass joint space (Fig. 4A). To get a further reduction and
solid fixation, a suitable size slim transoral atlantoaxial
reduction and fixation plate (slim-TARP, Weigao
Company, ZL201420087055.6, patented by Wang and
Ying ) was choosed and fixed with C1 by 2 screws inserting
into the lateral masses of C1, a TARP retractor (Weigao
Medical Systems) was installed according to the detailed
description by Yin et al.14 The complex of the atlas and
the plate were pushed back for further reduction by
turning the nut on the top of the retractor until the

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 85

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Wang et al J Pediatr Orthop  Volume 43, Number 2, February 2023

The Japanese orthopaedic association (JOA) score was used

to evaluate the improvement of myelopath postsurgery.

Statistical Analysis
The data were accumulated and imputed into
SPSS20.0 software (IBM Corp, Armonk, NY) for Sta-
tistical Analysis. Paired t tests were used to compare JOA
scores changes in parametric values before and after sur-
gery, which were considered significant if the P-value was
<0.05.

FIGURE 2. Diagrammatic representation of bidirectional trac-
tion. The patient is put in a supine position and a horizontal force
used to apply traction to the neck while a perpendicular force is
applied to the neck. The combined effect achieves atlantoaxial
reduction.
rotation of atlas was fully reduced (Fig. 3B). After that, 2
screws of C2 were inserted to finished the plate fixation
(Figs. 3C, 4A, B). In the end, the muscle and mucoma
were sutured layer by layer to close the wound.
Postoperative Treatment
All patients were sent to intensive care unit for
observation within 24 hours after surgery. The nasal tracheal
cannula was maintained for 1 to 2 days and then removed
after the oral mucosal swelling decrease. Nasogastric nu-
trition was kept for 5 to 7 days after the operation. On ac-
count of that the oral incision is a relatively bacterial
environment, we used antibiotics for 5 to 6 days to prevent
wound infection in our transoral plate implantation proce-
dures. All the 15 patients can walk in the second day after
surgery. All patients were fitted with a head-neck-chest brace
for 3 to 6 months until the bone fusion was confirmed by CT
scan. The patients were followed up from 12 to 36 months
and subjected to CT scan, radiography, and MR to evaluate
the reduction and bone fusion. In the following up, all the
patient underwent CT scan and x-rays in the third, sixth, and
12th month to observe and evaluate the bone fusion time.
And the MRI was performed in the first week, sixth month,
and 12 month to evaluate the spinal cord decompression.
RESULTS
All 15 patients underwent the TARP procedures
smoothly. The locked facet joint was released and fixed with
slim-TARP by transoral approach. The mean blood loss
and operating times were 83 ± 23 mL (range: 60 to 145 mL)
and 129.2 ± 11.9 minutes (range: 109 to 145 min), re-
spectively. There were no neurological injury or wound
infections. In our 15 patients, there are 20 C2 anterior
pedicle screws were used in 14 patients (Table 1).
Postoperative CT scans indicated that 58 screws were in
ideal positions and that no screw was inserted into the
spinal canal except for 2 C2 reverse screw that partially
violated into the cervical vertebral artery foramen. No
special treatment was administered to this patient because
there were no symptoms. Complicate correction of
torticollis was achieved in all 15 cases. Preoperative
symptoms of neck pain and limitation of neck movement
were effectively alleviated at 3 months after surgery. The 3
patients with preoperative neurological deficits had
significant relief after surgery, and their latest follow-up
results showed that their JOA scores increased from
13.0 ± 1.0 to 16.3 ± 0.6 (P < 0.05). All patients were
followed up for a mean of 17.8 ± 6.6 months (range: 12 to
36 mo). Bony fusion was achieved in all patients (Figs. 5, 6).
Mean fusion time was 3.6 ± 1.2 months (range: 3.0 to
6.0 mo).
DISCUSSION
AARF is a disorder involved in CVJ, which pri-
marily occurs in pediatric patients because of the immature
bone and ligaments structures. The clinical symptoms
usually include torticollis, cervical pain, and limited
motion of the neck. The most common causes are trauma,

FIGURE 3. A, An osteotome was inserted into the facets joint gap for distracting and unlocking the locked joint by prying and
rotating it vertically. B, A suitable size transoral atlantoaxial reduction and fixation plate with slim shape (slim-TARP) was fixed with
C1 by 2 screws inserting into the lateral masses of C1, then a TARP retractor was engaged with the plate and temporary screws in
axis, by turning the nut on the top of the retractor to push lateral mass of atlas for further reduction. C, After that, the atlantoaxial
complex were fixed with slim-shaped transoral anterior reduction plate (Fig. 4B).

86 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Treating Pediatric IAARF by Transoral Unlocking Facet Joint

FIGURE 4. Two pieces of bone block harvested from iliac crest were implanted into both sides of the atlantoaxial lateral mass joint
space (A, arrow), and the atlantoaxial complex were fixed with slim-shaped transoral anterior reduction plate (B).

inflammation-related ligament loosening, cervical muscle
spasms, or congenital malformation in the CVJ. In general,
the acute AARF can be cured with a cervical collar or
traction and usually receive a good result;15–17 however,
delaying diagnosis and treatment always lead to recurrence
and reduction failure, the patients may need surgery fi-
nally. All the 15 patients in our series had an improper
initial treatment in local hospital and recurrent with tor-
ticollis or neck pain for more than 3 months, then was
transferred to our hospital for further treatment.
Fernandez et al18 reported that patients with
AARF that remain unreduced for longer than 3 weeks
are at a higher risk of recurrence or permanent de-
formity because of the development of chronic changes
in the transverse and alar ligaments. AARF would be
more likely to require surgery. Pang and Li19,20 recom-
mended skull traction for patients who failed treatment
for acute AARF, and notice that a delay of treatment
longer than 2 to 3 months leads to a more severe form of
AARF who may need surgery treatment. Therefore,

FIGURE 5. Case 5, a 11-year-old boy patient, the computed tomography scan show a pathologic rotation of the atlas around the
axis with facet joint locked in the left side (A). The atlanto-dens gap is asymmetrical in the coronal plane (B), The ADI enlarged
obviously over than 5 mm (C). The patient underwent transoral release by unlocking the facet joint and fixation with slim-shaped
transoral anterior reduction plate. The postsurgery computed tomography in the 12th month show solid fusion occurs in the lateral
facet joints and atlanto-dens gap (D and E). The postsurgery magnetic resonance (G) showed the compression of brain brain-stem
relieved compared with the presurgery magnetic resonance (F).

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 87

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Wang et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 6. Case 6, a 10-year-old girl with irreducible atlantoaxial rotatory fixation (IAARF), the computed tomography (CT) scan
shows a pathologic rotation of the atlas around the axis with facet joint locked in the right side (A). The atlanto-dens gap is
asymmetrical in the coronal plane (B), the 3-dimentional reconstruction CT images showed the atlas rotate nearly 90 degrees around
the axis and form a locked dislocation (C and D). The girl underwent transoral release and fixation with slim-shaped transoral anterior
reduction plate by transoral approach, the postsurgery CT showed sound bone fusion occurred in the 12th month after surgery (E and
F), and the 3-dimentional CT images showed that the IAARF has been reduced and fixed with slim-shaped transoral anterior reduction
plate appropriately (G and H).

surgery treatment is an important consideration for
chronic irreducible AARF of conservative failure and
recurrence patients.
In our series, all the patients underwent conservative
treatment with traction or cervical collar immobilization in
local hospital for more than 3 months. After admission in
our hospital, they underwent bidirectional skull traction for
1 week again, and cannot achieve reduction, then surgery
was decided. During surgery, we found that the scar or callus
between the long time locked facet joint play a role in
hinding the reduction of AARF, hence the transoral release
to unlock the joint is necessary and critical step. Here, we use
unique transoral-only techs for treating pediatric AARD
without the need for posterior manipulation, which can not
only release and unlock the facet joint for reduction, but also
fix the atlantoaxial joint with the same approach. The
transoral release and fixation with plate procedures (TARP)
were first developed and reported by Yin14 in 2005 for
treating adult irreducible atlantoaxial dislocation. And then
widely confirmed to be effective and safe for the treatment of
IAAD or upper cervical revision surgery in adult
patients.21,22 To using it for treating pediatric AARF, we
designed a slim-shaped transoral plate (slim-TARP, Weigao
Company, China, patented by Wang and Yin) to fit the
anatomic structure of children’s bone and oral cavity. The
plate was slim and thinner than the adult TARP, which has a
anatomic contour to fit the children’s bone structures. These
15 patients were treated by this technique and received a
good result.
To treat irreducible AARF, unlocking the C1–C2
facet joint is a key step. Many authors used different ap-
proaches in manipulating locked C1–C2 facet joints:
Schmidek et al23 used a transoral route and Crockard and
Rogers24 used an extreme lateral approach to remove ob-
structing ligamentous or bone structures within the locked
facet joint. Goto et al25 released the atlantoaxial lateral
facet joint via a transoral route and subsequently performed
posterior fixation using interlaminar clamps. Goel and
colleagues used the posterior-only approach to unlock facet
joint and realize fixation by posterior screw-rod techs. The
advantages of Goel tech is only need 1 approach to unlock
the C1–C2 facet joint and fulfill fixation same time. How-
ever, when performing this technique, intraoperative
bleeding from the venous plexus around the C1/2 facet and
C2 nerve root irritation from the screw may occur, which
made it a high skill-demanding techs in treating pediatric
patients. Although the combined transoral release and
posterior instrumentation is another method used by many
authors for treating irreducible AAD, however, it can bring
more risk for spinal medulla injury during the change of
position and need longer operation time, which was some-
times regarded as a more complicate and challenge techs
with higher morbidity.
We found that there are several advantages in our
method. First, there are no worry about venous plexus
in the posterior side of atlantoaial joints, the bleeding
was very little in our case series during the procedures of
exposuring and manipulating the facet joints. And

88 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
FIGURE 7. There are 2 ways for C2 screws in the anterior approach: anterior pedicle screw and vertebrae screw. A shows the
anterior pedicle screw has a longer screw path with a good purchase and B shows the vertebrae screw was inserted into vertebrae
body of C2 with less chance of damaging the vertebrae artery.
second, the transoral approach provide us a directly
vision for releasing and distracting the facet joints than
posterior approaches. We can easily debride the scar and
callus between C1 and C2 lateral mass joint and insert a
reamer into the facet joint gap to distract and release
them. No need of cutting off the C2 ganglion to add
exposure like the study by Goel and colleagues. Because
the venous plexus and C2 nerve ganglion usually cover
the facet joint gap in the posterior side and make it
difficult and dangerous in manipulating and distracting
lateral mass joint. In addition, by transoral approach,
we can distract and decartilage the facet joint
conveniently and implant enough allograft bone for
good fusion. In our series, all of the 15 patients show
solid bone fusion in the third to sixth month after
operation.
Because there are high possibility of a small pedicle
and vertebral artery anomalies in pediatric patients, the
thin-slice CT scans with multilane reconstruction and
vertebrae artery CT angiograms of the craniocervical
junction are necessary for preoperation works. By using of
PACS image system, the size and direction of the C2
pedicle and C1 lateral mass can be measured, and the path
of the vertebral artery should be determined before sur-
gery. There are 2 ways for C2 screws in the anterior ap-
proach: anterior pedicle screw and vertebrae screw
(Figs. 7A, B). The anterior pedicle screw, which has a
longer screw path with a good purchase, and the vertebrae
screw looks like the screw used in the lower cervical, which
is inserted into the vertebrae body of C2, with less chance
of damaging the vertebrae artery. If the pedicle diameter
were > 4 mm, we choose the anterior pedicle screw and if
<4 mm, the vertebrae screw can be used. Our
postoperative CT scans indicated that 58 screws were in
ideal positions and that no screw was inserted into the
spinal canal except for 2 C2 anterior pedicle screw that
partially violated into the cervical vertebral artery hole.
Therefore, if an anomalous vertebral artery and a small
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Treating Pediatric IAARF by Transoral Unlocking Facet Joint
size pedicle was found in patient, the anterior pedicle
screw should be avoided and vertebral screws were
recommended with priority to prevent nerve and
vertebral artery injury.
One major risk associated with the transoral sur-
gical approach is the potential surgical site infection.
According to the literature, the surgical site infection rate
can range from 0% to 3.5% in adult patients.26 We found
that is not fully accordance with pediatric patients, in our
series, there are no infection happen. To minimize the
chance of children’s oral wound infection, before sur-
gery, rigorous examination of the child’s oral cavity
should be performed. For those with the complications
of tonsillitis, chronic pharyngitis, and dental caries, an-
terior internal fixation is strictly prohibited, and we use a
specially designed slim-TARP for anterior fixation,
which can fit perfectly with the children’s anatomic
character, and decrease the suture tension greatly, hence
benefit for wound healing and decreasing the risk of in-
fection theoretically, which should be confirmed by our
further research works.
The irreducible AARF with locked facet joint is a
difficult disorder for spine surgeons clinically. Although
there are many ways for dealing with this special entities,
the direct manipulating and releasing unlock the joint is
still a critical steps. Reliable fixation and effective bone
implantation are also imperative for the success of the
final treatment. By our series patients’ observation, the
transoral release and fixation with slim-TARP proved
preliminary, a feasible and relative convenient techs for
treating pediatric chronic AARF with locked facet joint.
However, there are still some limitations in our research
works. As the pediatric AARF patients requiring surgery
is rare, the sample size was relatively small yet. There-
fore, it is necessary to analyze more cases and conduct a
prospective study to determine whether our results are
universal. Despite its limitations, this study describes
a viable means of treating irreducible pediatric AARF.
www.pedorthopaedics.com | 89
Wang et al
REFERENCES
1. Corner EM. Rotary dislocations of the Atlas. Anm Surg. 1907;45:9–26.
2. Fielding JW, Hawkins RJ. Atlanto-axial rotatory fixation. (Fixed
rotatory subluxation of the atlanto-axial joint). JBone Joint Surg Am.
1977;59:37–44.
3. Wang S, Yan M, Passias PG, et al. Atlantoaxial rotatory fixed dislocation
report on a series of 32 pediatric cases. Spin. 2015;41:E725–E732.
4. Hsu PT, Chung HY, Wang JL, et al. Successful conservative
treatment of chronic atlantoaxial rotatory fixation in a child with
torticollis. Am J Phys Med Rehabil. 2010;89:776–778.
5. Neal KM, Mohamed AS. Atlantoaxial rotatory subluxation in
children. J Am Acad Orthop Surg. 2015;23:382–392.
6. Chazono M, Saito S, Liu K, et al. Continuous skull traction followed
by closed reduction in chronic pediatric atlantoaxial rotatory
fixation. Acta Neurochir (Wien). 2011;153:1443–1445.
7. Chechik O, Wientroub S, Danino B, et al. Successful conservative
treatment for neglected rotatory atlantoaxial dislocation. J Pediatr
Orthop. 2013;33:389–392.
8. Park SW, Cho KH, Shin YS, et al. Successful reduction for a
pediatric chronic atlantoaxial rotatory fixation (Grisel syndrome)
with long-term halter traction: case report. Spine (Phila Pa 1976).
2005;30:E444–E449.
9. Sae-Huang M, Borg A, Hill CS. Systematic review of the nonsurgical
management of atlantoaxial rotatory fixation in childhood.
J Neurosurg Pediatr. 2020;9:1–12.
10. Ishii K, Chiba K, Maruiwa H, et al. Pathognomonic radiological
signs for predicting prognosis in patients with chronic atlantoaxial
rotatory fixation. Neurosurg Spine. 2006;5:385–391.
11. Ishii K, Toyama Y, Nakamura M, et al. Management of chronic
atlantoaxial rotatory fixation. Spine (Pbila Pa 1976). 2012;37:
E278–E285.
12. Goel A, Shah A. Atlantoaxial facet locking: treatment by facet
manipulation and fixation. Experience in 14 cases. J Neurosurg
Sbime. 2011;14:3–9.
13. Qi C, Xi Y, Chunguang Z, et al. Anterior retropharyngeal reduction
and sequential posterior fusion for atlantoaxial rotatory fixation with
locked C1-C2 lateral facet. Spine. 2015;40:1121–1127.
14. Yin Q, Ai F, Zhang K, et al. Irreduc-ible anterior atlantoaxial
dislocation: one-stage treatment with a transoral atlantoaxial
reduction plate fixation and fusion. Report of 5 cases and review
of the literature. Spine (Phila Pa 1976). 2005;30:E375–E381.
90 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
15. Subach BR, McLaughlin MR, Albright AL, et al. Current manage-
ment of pediatric atlantoaxial rotatory subluxation. Spine (Phila
Pa1976). 1998;23:2174–2179.
16. Mihara H, Onari K, Hachiya M, et al. Follow-up study of conservative
treatment for atlantoaxial rotatory displacement. J Spinal Disord.
2001;14:494–499.
17. Landi A, Pietrantonio A, Marotta N, et al. Atlantoaxial rotatory
dislocation(AARD) in pediatric age: MRI study on conservative
treatment with Philadelphia collar-experience of nine consecutive
cases. Eur Spine J. 2012;21(suppl 1):S94–S99.
18. Fernandez Cornejo VJ, Martinez-Lage JF, Pigueras C, et al.
Inflammatory atlanto-axial subluxation (Grisel’s syndrome) in
children: clinical diagnosis and management. Childs Nerv Syst.
2003;19:342–347.
19. Pang D. Atlantoaxial rotatory fixation. Neurosurgery. 2010;66(suppl 3):
161–183.
20. Pang D, Li V. Atlantoaxial rotatory fixation: part 2-new diagnostic
paradigm and a new classification based on motion analysis using
computed tomographic imaging. Neurosurgery. 2005;57:941–953;
discussion-53.
21. Xia H, Yin Q, Ai F, et al. Treatment of basilar invagination with
atlantoaxial dislocation: atlantoaxial joint distraction and fixation
with transoral atlantoaxial reduction plate (TARP) without odon-
toidectomy. Eur Spine J. 2014;23:1648–1655.
22. Yin QS, Li XS, Bai ZH, et al. An 11-year review of the TARP
procedure in the treatment of atlantoaxial dislocation. Spine (Phila
Pa 1976). 2016;41:E1151–E1158.
23. Schmidek HH, Smith DA, Sofferman RA, et al. Transoral unilateral
facetectomy in the management of unilateral anterior rotatory
atlantoaxial fracture/dislocation: a case report. Neurosurgery.
1986;18:645–652.
24. Crockard HA, Rogers MA. Open reduction of traumatic atlanto-axial
rotatory dislocation with use of the extreme lateral approach. A report
of two cases. JBone Joint Surg Am. 1996;78:431–436.
25. Goto S, Mochizuki M, Kita T, et al. Transoral joint release of the
dislocated atlanto-axial joints combined with posterior reduction and
fusion for a late infantile atlantoaxial rotatory fixation. A case
report. Spine. 1998;23:1485–1489.
26. Yin Q, Xia H, Wu Z, et al. Surgical site infections following the
transoral approach: a review of 172 consecutive cases. Clin Spine
Surg. 2016;29:E502–E508.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Hip Displacement Does Not Change After Pelvic Obliquity
Correction During Spinal Fusion in Children With
Cerebral Palsy
Ali Asma, MD,* Mutlu Cobanoglu, MD,*† Armagan Can Ulusaloglu, MD,*
Kenneth J. Rogers, PhD, ATC,* Freeman Miller, MD,* Jason J. Howard, MD,*
Suken A. Shah, MD,* and M. Wade Shrader, MD*
Background: Children with cerebral palsy (CP) frequently
develop both neuromuscular hip dysplasia and scoliosis, and
occasionally, the timing of the worsening of both of these path-
ologies is concurrent. The question as to whether the hip or spine
should be addressed first in CP remains controversial, with the
majority of evidence being “expert opinion.” The purpose of this
project was to determine the impact of posterior spinal fusion
(PSF) on the change in hip displacement for children with CP
without previous reconstructive hip surgery.
Methods: This was an Institutional Review Board-approved
study that observed 67 patients from 2004 to 2018. Inclusion
criteria included children with CP, 18 years of age and younger,
Gross Motor Function Classification System IV and V, under-
going PSF at a single tertiary care children’s hospital with a
minimum 2-year follow-up. The primary outcome was the
change in hip displacement as quantified by the migration per-
centage (MP). The hip with the highest MP (worst hip) at the
spine preoperative analysis were included for analysis. Triradiate
cartilage (TRC) status and pelvic obliquity correction were
analyzed with multivariate analysis.
Results: Sixty-seven patients were included for analysis, with a
mean age of 12.5 ± 2.3 years. The mean major curve angle of the
major curve was 77 ± 23 degrees and the mean preoperative
pelvic obliquity was 21 ± 12 degrees. There was no statistically
significant change in MP after PSF from a mean preoperative
value of 41 ± 27%, to a mean postoperative value of 41 ± 29% at
the last follow-up, (P = 0.76) The mean follow-up time was
4.1 ± 2.7 years. TRC status (P = 0.52) and the severity of pelvic
obliquity (P = 0.10) did not statistically impact the change in MP
after PSF.
Conclusion: PSF did not influence—either negatively or
positively—the progression of hip displacement in children with
CP, regardless of pelvic obliquity correction or TRC status. The
From the *Nemours Children’s Health, Delaware, Wilmington, DE;
and †Medicana International, Izmir, Izmir, Turkey.
No funding was received in support of this work.
The authors declare no conflicts of interest.
Reprints: M. Wade Shrader, MD, Nemours Children’s Health, Delaware,
Department of Orthopaedics, 1600 Rockland Road, Wilmington, DE
19803. E-mail: Wade.Shrader@nemours.org.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002292
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
lack of deterioration in hip displacement post-PSF, however,
may suggest a protective effect of spine surgery.
Level of Evidence: Level III—retrospective cohort study.
Key Words: nonambulatory, cerebral palsy, hip displacement,
spinal fusion
(J Pediatr Orthop 2023;43:e127–e131)
C hildren with cerebral palsy (CP) with Gross Motor
Function Classification System (GMFCS) level IV
and V function have an increased risk for hip dysplasia
and scoliosis. Progressive hip dysplasia requiring re-
construction typically occurs between the ages of 6 and
12 years, whereas progressive spine deformity requiring
posterior spinal fusion (PSF) typically occurs after age
10 years. Therefore, for most children, the timing of hip
reconstruction typically occurs before the need for PSF.1
However, occasionally, the progression of these patholo-
gies, and the need for treatment decisions, are concurrent.2
The question as to whether the hip or spine should be
addressed first in CP remains controversial, with the
majority of evidence being “expert opinion.”3
The arguments for performing hip reconstruction—
soft-tissue releases and osteotomies—before scoliosis cor-
rection are 3-fold. First, the hip displacement may prog-
ress further with increased femoral head deformity,
necessitating more salvage rather than reconstructive
procedures.4 Second, associated fixed hip adductor and
flexor contractures (with associated pelvic obliquity) can
make positioning for spine surgery difficult, and may
negatively affect pelvic obliquity correction. Hence, some
spine surgeons feel it is important to correct obliquity
from an infrapelvic cause with hip procedures before ad-
dressing supra-pelvic obliquity with PSF. Finally, the use
of spinopelvic fixation can potentially interfere with the
execution of acetabular osteotomies.5
Proponents supporting scoliosis correction before hip
reconstruction may feel it is better to obtain a level pelvis
through supra-pelvic obliquity correction, possibly improv-
ing femoral head coverage and allowing for ease of sub-
sequent hip reconstruction. In addition, PSF is reported to
have positive impacts on medical comorbidities, mortality,
www.pedorthopaedics.com | e127
Asma et al J Pediatr Orthop  Volume 43, Number 2, February 2023

and longevity.6 Also, postoperative pain after hip osteotomy and soft-tissue lengthening) was a progressive
reconstruction is reportedly higher than for scoliosis correc- MP > 50%.
tion, which may cause families to delay spinal surgery given
the difficulties endured.7 This delay may result in further RESULTS
increases in curve magnitude and stiffness, increasing
We identified 67 patients (40 females) who met the
perioperative risk.8
inclusion criteria, with a mean age of 12.5 ± 2.3 years at
The purpose of this project was to determine the
the time of scoliosis correction. The magnitude of the
impact of PSF on the change in hip displacement for
major curve, as measured by the Cobb method, was
children with CP without previous hip surgery. Under-
77 ± 23 degrees, and the mean preoperative pelvic obliq-
standing the impact of scoliosis correction on femoral
uity was 20.7 ± 12 degrees. Ten patients had a windblown
head coverage will allow for more informed decision-
pelvis before spinal fusion. The mean preoperative worst
making when determining the timing of hip reconstruction
hip MP was 41 ± 27%. Thirty of these hips (45%) were on
when both deformities are present.
the low side of the pelvis and 37 (55%) were high side. The
mean follow-up time was 4.1 years ( ± 2.7 y) postscoliosis
correction. None of the patients had reconstructive or soft-
METHODS tissue (adductor +/− iliopsoas release) surgeries before
spinal fusion. Twenty-two patients had hip surgery after
The study design was a retrospective review, con-
spinal fusion, involving 18 reconstructive and 4 soft-tissue
ducted at an academic tertiary-level children’s hospital.
surgeries.
Ethics approval was provided by our institution’s Office of
There was no statistically significant change in MP
Human Subjects Protection Institutional Review Board.
after PSF; MP changed from a mean preoperative value of
Patients with CP aged 18 years and younger; presenting
41 ± 27% to a mean postoperative value of 41 ± 29% at the
from 2004 until 2018; GMFCS IV and V; spastic, dystonic,
last follow-up (P = 0.76) (Fig. 1). No correlation was
and mixed motor types; previous PSF; and minimum
found between preoperative MP and MP change
2 years of follow-up were included. A typical PSF event
(P = 0.67).
included a posterior approach with instrumentation from
TRC status (P = 0.52) and the severity of pelvic
the upper thoracic spine (T2 or T3) to the pelvis. Exclusion
obliquity (P = 0.06) were not significant factors contrib-
criteria were prior reconstructive hip surgery, ambulatory
uting to the change of MP after PSF. The mean post-
patients (GMFCS I, II, and III), and a nonhypertonic
operative major curve angle was 18 ± 11 degrees, for a
motor type.
mean correction of 77%. The mean final pelvic obliquity
The primary outcome was the change in hip dis-
(until hip surgery or last follow-up) was 4 ± 3 degrees), for
placement as quantified by the migration percentage
a mean correction of 94 ± 27%.
(MP).9 The change in MP from before PSF and at the last
We also performed a subgroup analysis for patients
postoperative follow-up was measured. For those patients
with pre-PSF MP between 30% and 60%, where 28 pa-
undergoing hip surgery after scoliosis correction, the last
tients (42%) from the overall main cohort were included.
hip radiograph before reconstruction was utilized. The hip
For this subgroup, at a mean of 46 ± 33 months post-PSF
with the highest MP (worst hip) before scoliosis correction
follow-up, preoperative MP did not significantly change at
was utilized for analysis. Pelvic obliquity was measured as
the last follow-up (41 ± 10% vs 38 ± 18%; P = 0.244).
the angle between a line drawn perpendicular to the
For the entire cohort, 9 hips (13%) progressed to
middle of the vertical T1-S1 line and the line paralleling
MP ≥ 50%, with at least 10% progression after PSF
the top of the iliac crest.10 The status of triradiate cartilage
(Table 1; Figs. 2 and 3). There was no significant
(TRC) closure was determined according to the Modified
difference in pelvic obliquity correction for those
Oxford hip score,11 with a TRC score of 3 defined as
patients with MP progression ≥ 50% versus without
closed, and <3 defined as open.TRC
(mean, 93 ± 20% vs 94 ± 28%; P = 0.89). The mean MP
The Kolmogorov-Smirnov test and Shapiro-Wilk test
before spinal fusion for progressed hips versus
were used to assess the normality of numeric variables of
nonprogressed hips was not significantly different
age and MP. For the normally distributed numeric varia-
(44 ± 19% vs 41 ± 28%, respectively; P = 0.32).
bles, the student t test assuming equal variances was used.
With respect to the influence of pelvic obliquity on
Descriptive statistics [mean, SD ( ± ), variance, range] was
MP, the high side pelvis preoperative MP was unchanged
performed. Linear regression analysis was utilized for the
post-PSF (mean, 34 ± 27% vs 38 ± 28%; P = 0.1). Similarly,
pelvic obliquity correction relationship with MP change.
the low side pelvis preoperative MP was also unchanged
The Spearman correlation test was used to assess for cor-
post-PSF (mean, 33 ± 27% vs 34 ± 30%; P = 0.9).
relation between preoperative MP and MP change after
spinal fusion. A P-value below 0.05 was considered statis-
tically significant. SPSS v.27 (IBM Corp., Armonk, NY) DISCUSSION
was used for all statistical analyses. Hips with at least 10% For most patients with CP, the progression of hip
progression and reaching to ≥ 50% MP after spinal fusion displacement usually precedes the progression of scoliosis,
were defined as hips that progressed. Indication for hip with hip reconstruction typically occurring before age
reconstruction (varus derotational osteotomy +/− pelvic 10 years. Specifically, the first peak of hip subluxation

e128 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Hip Displacement After CP Spinal Fusion

FIGURE 1. Images of an 11-year-old patient with Gross Motor Function Classification System (GMFCS) V quadriplegic cerebral
palsy. A, Prespine hip radiograph, migration percentage (MP) at right hip: 10%, left hip: 50%. B, 10 years later, postposterior spinal
fusion (PSF) hip radiograph showing MP at right hip: 21, left hip: 15. No clinically relevant progress was seen after PSF at 10 years
follow-up.

usually occurs around 3 to 5 years of age, and the second
peak occurs around age 6 years of age.1,12,13 By contrast,
most patients with CP do not develop clinically relevant
scoliosis before age 8 years, with PSF typically recom-
mended later in the preadolescent growth period.14 Occa-
sionally, both hip and spine deformities become clinically
significant concurrently, and the decision about which to
treat first has to be considered. To date, this important
clinical question has had only expert opinion in the literature
to guide families and orthopaedic surgeons. Therefore, the
results from this retrospective analysis can be directly ap-
plied to help guide shared decision-making in these patients.
Helenius et al3 wrote the most direct paper to date
on this topic with an expert opinion review. The results of
their study suggest that when hip dislocation and scoliosis
present at the same time, the most symptomatic deformity
should be addressed first. Additional guidance is that if
pelvic obliquity is <10 degrees, hip reconstruction should
be performed first if the hips are symptomatic. If pelvic
obliquity is ≥ 10 degrees, then they suggest scoliosis sur-
gery should be addressed first. Crawford et al15 claim that
leveling the pelvis with spinal fusion can cause new hip
instability, and therefore suggest cautious and close hip
surveillance after spine fusion.
However, no previous clinical series with data have
directly addressed the question, of which deformity should be
addressed first if presenting concurrently. There is some
concern that performing PSF, with substantial pelvic obliq-
uity correction, may worsen preexisting hip displacement and
make subsequent hip reconstruction (where needed) more
difficult. Our typical protocol for patients who present with
operative concurrent hip and spine deformities is to treat
scoliosis first, followed by hip reconstruction ~6 months after
PSF. This approach was supported by our results, as there
was no change in post-PSF MP compared with pre-PSF MP.
Overall, PSF did not significantly worsen or improve the MP
in a group of patients with CP with hips at risk.
In the current study, neither the severity of the pre-
operative pelvic obliquity nor the amount of pelvic obliq-
uity correction influenced change in MP post-PSF. This is
also reported by other authors.16–19 Abel et al17 show a
weak correlation between supra-pelvic obliquity and hip
displacement over several years. They did not show in-
creased hip displacement opposite of the convex side of
scoliosis (high side) as proposed by Letts et al20 Lonstein
et al19 show no relationship between high or low side pelvic
obliquity and hip displacement and conclude that muscle
imbalance has a larger role. In their study of “windblown

TABLE 1. Progressed Hips Preoperative and Last Follow-up Information

Case no. GMFCS Level Spinal Fusion Age (y) Prespine MP (%) Last Follow-up MP (%) Future Hip Surgery Hip Surgery Age (y)
1 V 12.2 55 70 VDRO+Dega 13.0
2 V 11.7 78 100 VDRO+Dega 14.0
3 V 14.5 25 72 VDRO+Dega 16.3
4 V 15.4 17 52 Soft tissue 19.7
5 V 12.2 33 50 VDRO+Dega 14.8
6 V 9 22 70 VDRO+Dega 11.7
7 V 8.5 47 66 VDRO+Dega 9.1
8 V 12.1 58 68 No surgery –
9 V 11.8 50 74 VDRO+Dega 13.8
GMFCS indicates Gross Motor Function Classification System; MP, migration percentage; VDRO, varus derotation osteotomy.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e129

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Asma et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 2. Images of a 9-year-old patient with GMFCS V. A, Prespine bilaterally located hips with possible abduction contracture.
B, Postspine 1.5-year follow-up, hips are located. C, Postspine 3.5-year-old patient who was lost follow-up presented with left hip
subluxation with 70% MP.

hips,” Madigan et al18 report that there was no reliable
pattern of hip abduction-adduction by direction of scoliosis
apex (ie, toward the convexity or concavity). These results
show the inconsistent relationship between hip displace-
ment, pelvic obliquity, and scoliosis direction.
Although the low side hip may seem relatively cov-
ered compared with the high side hip, in many cases the
pelvic obliquity is causing apparent hip coverage. Not un-
commonly, the low side hip may have occult displacement
that becomes apparent after PSF and pelvic obliquity

FIGURE 3. Images of a 14.5-year-old boy with GMFCS V. A, Prespine bilaterally located hips with possible abduction contracture.
Patient spinal fusion was complicated with spinal infection, requiring subsequent irrigation and debridement. B, Postspine
9 months right hip started to dislocate with windblown pelvis appearance. C, Postspine 1.8-year follow-up presented with right hip
subluxation with 72% MP. D, Post bilateral varus derotation osteotomy—right Dega surgery 6-month follow-up with the union.

e130 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
correction. This is more obvious if the cause of the pelvic
obliquity is supra-pelvic. Crawford et al15 report more hip
reconstructive surgery directed at the low side of the pelvis
than the high side. As stated above, we found no relation-
ship between pelvic obliquity high or low side and change in
MP. That said, in our experience, patients with substantial
pelvic obliquity ( > 20 degrees) typically will have a high
failure rate of hip reconstruction if it is done before the
pelvic obliquity is corrected.
In our study, no patient had such severe lower
extremity contractures that positioning for spinal fusion
could not be adequately and safely performed. In the au-
thors’ experience, patience in altering positioning, includ-
ing alteration of pelvic pad placement, lowering the leg
sling for severe flexion contractures, or adding platforms
for severe extension contractures will typically allow for
appropriate positioning on a standard spine table. Fur-
thermore, no patient had a spinopelvic fixation that neg-
atively affected the ability for future hip reconstruction.
Thoughtful consideration of the placement of the spino-
pelvic fixation can pay off making later hip and pelvic
surgery easier. Specifically, our practice is to use sacroiliac
screws for pelvic fixation and to use a length of 65 mm to
allow for later pelvic osteotomy.
The strengths of this study are the relatively large
number of patients with concurrent hip and spine pathology
at a tertiary care center specializing in the treatment of
children with CP. Our primary outcome variable of change
in MP rather than the prevalence of hip reconstruction was
chosen to eliminate any biases secondary to surgical deci-
sion-making. The other strength of our study is utilizing
patients who did not have prior hip surgical reconstruction.
This allowed us to eliminate a potential protective effect of
previous reconstructive osteotomy on hip displacement
after spinal fusion. The major weakness of the study was the
retrospective nature of this case series. An analysis of
change in MP across the group obviously does not account
for individual patients that may have worsened, which
clearly occurred. Although some patients had worsening
hip displacement after PSF, logistic regression analysis for
possible risk factors was not possible given the small
number of patients in this subgroup. Inherent biases of
retrospective studies, such as selection or misclassification,
cannot be avoided. Future studies are planned with pro-
spective data collection with a more stringent measurement
of outcomes following our typical protocols.
In conclusion, PSF did not influence—either neg-
atively or positively—the progression of hip displacement
in children with CP, regardless of pelvic obliquity cor-
rection or TRC status. The lack of deterioration in hip
displacement post-PSF, however, may suggest a protective
effect of spine surgery. This information will be helpful in
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Hip Displacement After CP Spinal Fusion
shared decision-making with families of children with CP
who have to choose which pathology to treat first when
presented concurrently.
REFERENCES
1. Pruszczynski B, Sees J, Miller F. Risk factors for hip displacement in
children with cerebral palsy: systematic review. J Pediatr Orthop.
2016;36:829–833.
2. Cobanoglu M, Chen BP, Perotti L, et al. The impact of spinal fusion
on hip displacement in cerebral palsy. Indian J Orthop. 2021;55:
176–182.
3. Helenius IJ, Viehweger E, Castelein RM. Cerebral palsy with
dislocated hip and scoliosis: what to deal with first? J Child Orthop.
2020;14:24–29.
4. Kolman SE, Ruzbarsky JJ, Spiegel DA, et al. Salvage options in the
cerebral palsy hip: a systematic review. J Pediatr Orthop. 2016;36:
645–650.
5. Jain A, Brooks JT, Kebaish KM, et al. Sacral alar iliac fixation for
spine deformity. JBJS Essent Surg Tech. 2016;6:e10.
6. Miyanji F, Nasto LA, Sponseller PD, et al. Assessing the risk-benefit
ratio of scoliosis surgery in cerebral palsy: surgery is worth it. J Bone
Joint Surg Am. 2018;100:556–563.
7. Shrader MW, Jones J, Falk MN, et al. Hip reconstruction is more
painful than spine fusion in children with cerebral palsy. J Child
Orthop. 2015;9:221–225.
8. Hollenbeck SM, Yaszay B, Sponseller PD, et al. The pros and cons
of operating early versus late in the progression of cerebral palsy
scoliosis. Spine Deform. 2019;7:489–493.
9. Reimers J. The stability of the hip in children. A radiological study of
the results of muscle surgery in cerebral palsy. Acta Orthop Scand
Suppl. 1980;184:1–100.
10. Shrader MW, Andrisevic EM, Belthur MV, et al. Inter and
intraobserver reliability of pelvic obliquity measurement methods
in patients with cerebral palsy. Spine Deform. 2018;6:257–262.
11. Popejoy D, Emara K, Birch J. Prediction of contralateral slipped
capital femoral epiphysis using the modified Oxford bone age score.
J Pediatr Orthop. 2012;32:290–294.
12. Elkamil AI, Andersen GL, Hägglund G, et al. Prevalence of hip
dislocation among children with cerebral palsy in regions with and
without a surveillance programme: a cross sectional study in Sweden
and Norway. BMC Musculoskelet Disord. 2011;12:284.
13. Graham HK. Painful hip dislocation in cerebral palsy. Lancet.
2002;359:907–908.
14. Takeuchi R, Mutsuzaki H, Mataki Y, et al. Progressive age and
other factors affecting scoliosis severity in cerebral palsy patients.
J Rural Med. 2020;15:164–169.
15. Crawford L, Herrera-Soto J, Ruder JA, et al. The fate of the
neuromuscular hip after spinal fusion. J Pediatr Orthop. 2017;37:
403–408.
16. Garg S, Engelman G, Yoshihara H, et al. The relationship of Gross
Motor Functional Classification Scale level and hip dysplasia on the
pattern and progression of scoliosis in children with cerebral palsy.
Spine Deform. 2013;1:266–271.
17. Abel MF, Blanco JS, Pavlovich L, et al. Asymmetric hip deformity
and subluxation in cerebral palsy: an analysis of surgical treatment.
J Pediatr Orthop. 1999;19:479–485.
18. Madigan RR, Wallace SL. Scoliosis in the institutionalized cerebral
palsy population. Spine (Phila Pa 1976). 1981;6:583–590.
19. Lonstein JE, Beck K. Hip dislocation and subluxation in cerebral
palsy. J Pediatr Orthop. 1986;6:521–526.
20. Letts M, Shapiro L, Mulder K, et al. The windblown hip syndrome
in total body cerebral palsy. J Pediatr Orthop. 1984;4:55–62.
www.pedorthopaedics.com | e131
 ORIGINAL ARTICLE
Surgical Management of Severe Equinus Deformity
in Ambulatory Children With Cerebral Palsy
Khadija Soufi, BS,*† Anita Bagley, PhD, MPH,*‡ Sean A. Brown, MA, BS,*§
David E. Westberry, MD,∥ Vedant A. Kulkarni, MD,*‡ Prabhav Saraswat, PhD,∥
and Jon R. Davids, MD*‡
Background: Tendo Achilles lengthening (TAL) for the manage-
ment of equinus contractures in ambulatory children with cerebral
palsy (CP) is generally not recommended due to concerns of over-
lengthening, resulting in weakness and plantar flexor insufficiency.
However, in some cases, surgical correction of severe equinus
deformities can only be achieved by TAL. The goal of this study is
to assess the outcomes following TAL in these cases.
Methods: A retrospective cohort study of children with CP with
severe equinus contractures (ankle dorsiflexion with the knee
extended of -20 degrees or worse) who underwent TAL as part of
a single event multilevel surgery, with preoperative and post-
operative gait analysis studies. Continuous data were analyzed
by paired t test, and categorical data by McNemar Test.
Results: There were 60 subjects: 42 unilateral, 18 bilateral CP; 41
GMFCS II, 17 GMFCS I; mean age at surgery was 10.6 years,
mean follow-up was 1.3 years. Ankle dorsiflexion with the knee
extended improved from −28 to 5 degrees (P < 0.001). The ankle
Gait Variable Score improved from 34.4 to 8.6 (P < 0.001). The
ankle moment in terminal stance improved from 0.43 to
0.97 Nm/kg (P < 0.001). Significant improvements (P < 0.001)
were seen in radiographic measures of foot alignment following
surgery. There were few significant differences in the outcome
parameters between subjects with unilateral versus bilateral CP
(eg, only the bilateral group showed improved but persistent
increased knee flexion in mid-stance).
Conclusions: The outcomes following TAL for the management
of severe equinus deformity in ambulatory children with CP were
favorable 1 year after surgery, with significant improvements in
all domains measured.
From the *Department of Orthopaedic Surgery, Shriners Hospital for
Children; †University of California School of Medicine; ‡Department
of Orthopaedic Surgery, University of California Davis Health,
Sacramento, CA; §Department of Kinesiology, Recreation, and Sport
Studies, The University of Tennessee, Knoxville, TN; and ∥Depart-
ment of Orthopaedic Surgery, Shriners Hospital for Children,
Greenville, SC.
J.R.D. is on the Editorial Board of the Journal of Pediatric Orthopaedics,
has been a consultant of OrthoPediatrics Corp., and is on the Board
of the Foundation for Advancing Pediatric Orthopaedics, all of which
are outside the scope of the submitted work. The remaining authors
declare no conflict of interest.
Reprints: Jon R. Davids, MD, Department of Orthopaedic Surgery,
Shriners Hospital for Children, 2425 Stockton Blvd, Sacramento, CA
95817. E-mail: jdavids@shrinenet.org.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002310
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Significance: This study does not advocate for the widespread use
of TAL to correct equinus deformity in children with CP.
However, it does show that good short-term outcomes following
TAL are possible in properly selected subjects with severe con-
tractures when the dosing of the surgery is optimal (correction of
contracture to between 0 and 5 degrees of dorsiflexion with the
knee extended) and the procedure is performed in the setting of
single event multilevel surgery with subsequent proper orthotic
management and rehabilitation.
Key Words: cerebral palsy, equinus gait, tendo achilles lengthening
(J Pediatr Orthop 2023;43:91–98)
E quinus gait, characterized by excessive ankle plantar-
flexion in the stance and swing phases of gait, is a
common gait disruption pattern seen in children with
cerebral palsy (CP).1,2 The 2 most common causes of
equinus gait are dynamic overactivity or fixed shortening
(myostatic deformity) of the gastrocsoleus complex
(GSC). Myostatic deformity of the GSC causing gait
disruption in children with CP is typically managed by
surgical lengthening of the GSC muscle-tendon unit
(MTU). There are a variety of surgical procedures, applied
at different locations along the MTU, described in the
literature, reflecting the range of surgical dosing available
to address this problem.3–5
The surgical anatomy of the GSC consists of 3
zones, or sites, for potential lengthening.6 In zone 1, which
is the most proximal in the MTU, selective lengthening of
only the gastrocnemius muscle is possible. In zone 2, lo-
cated at the myotendinous junction of the MTU, selective
lengthening of both the gastrocnemius and soleus muscles
is possible. Zone 3, the most distal site, is located at the
tendon level of the MTU and results in nonselective
lengthening of the gastrocnemius and soleus muscles.
The GSC plays a crucial role in normal gait, and
weakness of this muscle group can cause dramatic gait
deviations, particularly in children with CP (eg, crouch
gait).7,8 The greater the magnitude of the myostatic de-
formity of the GSC, the greater the degree of surgical
lengthening required to restore the functional range of
motion. The degree of lengthening achievable with zone 1
surgical techniques is relatively minimal, with progressively
increased lengthening possible with zone 2 and 3 surgical
www.pedorthopaedics.com | 91
Soufi et al
techniques.6,9,10 Weakness following surgical lengthening
of the MTU may be a consequence of excessive length-
ening (most likely to occur with nonselective lengthening
at zone 3), or due to direct damage to the contractile
components (most likely to occur at zones 1 and 2).11–13
Clinical studies of the lengthening of the GSC in
children with CP have identified risk factors for both re-
currence of deformity (eg, age at the time of surgery) and
over-lengthening resulting in excessive weakness (eg,. zone
3 surgical lengthening techniques).3–5,14–19 Many of the
cases considered in these studies are from an era when
surgeons had a poor understanding of the pathophysio-
logy of gait disruption in children with CP, limited access
to quantitative gait analysis (QGA), lack of experience
with performing single-event multilevel surgery (SEMLS),
and sub-optimal orthotic management and therapy fol-
lowing surgery. In cases with severe myostatic equinus
deformities, adequate surgical correction can only be
achieved by zone 3 lengthening (Tendo Achilles length-
ening, TAL). Justifiable concerns for over-lengthening and
excessive weakness based on historical literature can
complicate surgical decision-making, potentially resulting
in insufficient correction and persistent equinus gait pat-
terns following surgery. The goal of this study is to
document outcomes following TAL performed to correct
severe equinus deformities in children with CP, where
clinical decision-making and surgical dosing for SEMLS
were informed by QGA, and postoperative rehabilitation
included standardized orthotics management and physical
therapy.
METHODS
The study design was a retrospective cohort, result-
ing in level 4 evidence. The research protocol was reviewed
and approved by the Institutional Research Board for the
parent hospital system for the 2 institutions contributing
cases and data to this study. This study covered the time
period between January 1, 2012 and February 16, 2021.
A clinical patient database, maintained in the Motion
Analysis Centers (MAC) of each institution, was used to
identify all patients with CP, from 4 to 18 years of age,
whose physical examination at the index MAC study
(preoperative evaluation) revealed on goniometric-based
physical examination 20 degrees or greater fixed ankle
plantar flexion with the knee extended on at least 1 side.
This criteria for inclusion was selected as we thought it
was the most likely situation in which the surgeons would
consistently select TAL as the treatment option, regardless
of the findings on the examination with the knee flexed.
Indeed, all of the children so identified had TAL surgery.
We did not identify any cases with ankle dorsiflexion of 20
degrees or greater with the knee extended who did not
have limited ankle dorsiflexion with the knee flexed (in-
dicating concomitant myostatic deformity of the soleus).
A more nuanced assessment of the degree of myostatic
deformity of the soleus was not part of the clinical
decision-making process for selecting TAL for surgical
correction in these cases.
92 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
Following surgery, the subjects were placed into a
short leg cast (if between -10 degrees and full knee ex-
tension was achieved following the SEMLS surgery) with
a knee immobilizer, or a long leg cast (if greater than -10
degrees knee extension was present following the SEMLS
surgery, in anticipation of weekly serial stretch casting to
correct the residual knee flexion deformity), with the ankle
placed in a plantigrade (ie, 0 degrees for SEMLS surgery
dorsiflexion or plantarflexion) position. The patients were
non–weight-bearing in a wheelchair (with the hips flexed
to 90 degrees, that is, seated upright, for as long as tol-
erated each day) for 4 (if the SEMLS included only soft
tissue surgery) to 6 weeks (if the SEMLS included soft
tissue and skeletal surgeries).
The study included both unilateral (hemiplegic) and
bilateral (diplegic) types of CP, and all subjects functioned
at GMFCS levels I, II, or III. Only 1 side per subject was
included in the study cohort. In unilateral subjects, the
affected side was included in the analysis. For subjects
with bilateral CP, if only 1 side met the physical exami-
nation criteria, it was included. If both sides met the in-
clusion criteria, then the side with the greater magnitude
of equinus deformity was selected. If both sides had
identical degrees of equinus deformity, then the right side
was selected for inclusion in the study.
All subjects had undergone TAL (by Hoke triple
hemi-tenotomy or Z-lengthening), usually in the setting of
SEMLS, based upon recommendations from the index
MAC studies, and had follow-up MAC studies to access
outcomes following recovery from surgery. The surgeries
were performed by 3 surgeons, all with a special interest in
caring for children with CP, training in the use of quanti-
tative gait analysis for clinical decision-making and out-
come assessment, and experience in performing SEMLS
and guiding post-surgical orthotic management and re-
habilitation. Regardless of the technique of TAL utilized,
all surgeons dosed the GSC lengthening to achieve between
0 and 5 degrees of ankle dorsiflexion with the knee ex-
tended at the time of surgery. Intraoperative assessment of
ankle dorsiflexion was performed with the knee extended
using a thermoplastic platform with a T handle to provide
appropriate dorsiflexion stress to the plantar aspect of the
foot. Ankle alignment was assessed visually, and the op-
erative notes did not consistently describe the degrees of
dorsiflexion obtained. None of the subjects in the study
group had previous orthopaedic surgery to improve am-
bulation, and none had received botulinum toxin injections
within 6 months of the index MAC evaluation.
Patient clinical history, demographic, physical
examination, kinematic, kinetic, pedobarographic, and ra-
diographic data were derived from the subjects’ institutional
electronic health records and MAC data files. All data were
extracted, measured, and interpreted by a single observer
(K.S.), after appropriate training and under the supervision
of the senior authors (J.R.D. and A.B.). Quantitative data
points (kinematics and kinetics) were extracted by in-
stitution-specific software. Pedobarography data were ana-
lyzed based on the progression (location and duration) of
the center of pressure.20 The selected parameters were the
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
location of the center of pressure (CoP) at Initial Contact
(normal is in the hindfoot), and the duration of the CoP in
the segments of the foot (hindfoot, midfoot, and forefoot)
throughout Stance. Radiographic measurements of hind-
foot alignment were performed from standing ante-
roposterior and lateral views.21 Three radiographic angle
measurements were selected: calcaneal pitch, tibiocalcaneal
angle, and lateral talocalcaneal angle.
Bilateral 3-dimensional kinematic and kinetic data
were collected with the use of a 12-camera motion meas-
urement system (Motion Analysis Corporation, Rohnert
Park, CA, for 1 site and Vicon 512, Oxford Metrics
Group, Oxford, England, for the other) and 2 force plat-
forms (Kistler, Winterthur, Switzerland for 1 site and
Advanced Mechanical Technology, Watertown, Massa-
chusetts for the other). Subjects were instrumented with
passive reflective markers consistent with the Newington
model for gait analysis.22 Subjects made several passes
through the laboratory measurement volume walking, and
over the force platforms, at a self-selected speed. Kine-
matic data were calculated through the Newington model
and combined with force platform data (inverse dynamics)
to calculate kinetic data. Kinematic data collected from
typically developing children, in the same manner, were
used to compute the Gait Variable Score (GVS).23 MAC
clinicians selected 1 walking trial that was most repre-
sentative of the subjects’ walk for clinical interpretation.
That walk was used to extract Peak Ankle Dorsiflexion,
Peak Ankle Moment, and Mean Knee Flexion in Single
Support as well as calculate Ankle GVS.
Pedobarographic data were collected with RSscan
system (Beringen, Belgium). Subjects made several passes
over the pedobarographic mat at a self-selected speed.
MAC clinicians selected a representative walk for clinical
interpretation. That walk was used to extract the location
of the COP at the Initial Contact of the foot with the floor
and data with respect to the duration of time spent in hind,
mid, and forefoot sections.
Statistical analysis was done through SPSS
Statistics for Windows, version 27 (SPSS Inc., Chicago,
IL,). Continuous data were analyzed by paired t test, and
categorical data by paired McNemar Test. Comparisons
between unilateral and bilateral patients utilized
unpaired t test. Results were considered statistically
significant for P < 0.05.
RESULTS
Patient Demographics
Sixty patients met the inclusion criteria and were
included in the study group. There were 23 female, 37
male patients, and the racial and ethnic distribution in-
cluded 27 White, 14 Hispanic, 8 Black, 8 Other, and 3
Asian patients. Preoperatively, 28% (17/60) were GMFCS
level I, 68% (41/60) were GMFCS level II, and 3% (2/60)
were GMFCS level III. Seventy percent (42 subjects) of
the study group were classified as having unilateral CP
(hemiplegia); the remaining 30% (18 subjects) were clas-
sified as having bilateral CP (diplegia). The mean age at
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Surgical Management of Severe Equinus in CP
operation was 10.6 years (median of 10 y, range 4 to 18 y).
The mean time from surgery to postoperative gait study
was 1.3 years (median 1.1 y, range 1.0 to 6.1 y). All sub-
jects underwent TAL, which was performed as an isolated
procedure in 6 cases. The remaining 54 subjects had a
mean of 3 additional procedures (range 1 to 6 additional
procedures). There were 48 additional soft tissue surgeries
and 11 skeletal surgeries about the foot, 22 additional soft
tissue and 5 skeletal surgeries about the ankle, 45 addi-
tional soft tissue surgeries and 1 skeletal surgery about the
knee, and 14 additional skeletal surgeries about the hip.
Physical Examination
The physical examination results are presented in
Table 1.21 Statistically significant improvements in ankle
dorsiflexion range of motion, with the knee, flexed and
extended, were noted for the entire study group following
TAL surgery. There were no significant differences between
unilateral and bilateral subjects for any of the preoperative or
postoperative physical examination measures. Postoperative
ankle dorsiflexion with the knee extended was 4.7 degrees for
the entire cohort, 3.3 degrees for the unilateral group, and 7.8
degrees for the bilateral group, indicating that surgical
dosing of the TAL to achieve between 0 and 5 degrees of
dorsiflexion with the knee extended avoided over lengthening
at follow-up.
Kinematics and Kinetics
The kinematic and kinetic data are presented in
Table 2. Ankle kinematic plots are shown in Fig. 1.
Statistically significant improvements in Peak Ankle
Dorsiflexion in Stance and the Ankle GVS were noted
for the entire study group following TAL surgery. There
were no significant differences between unilateral and
bilateral subjects for any of the preoperative or
postoperative ankle kinematic measures. Significant
improvements in Peak Ankle Dorsiflexion in Stance
(−28.4 degrees to 13.1 degrees, within the normal range),
and Ankle GVS (34.4 degrees to 8.6 degrees) indicate that
surgical dosing of the TAL to achieve between 0 and 5
degrees of dorsiflexion with the knee extended avoided
over lengthening at follow up.
Ankle kinetic plots are shown in Fig. 2. Due use of a
walker, short stride, and inability to strike the force plate,
kinetic data were successfully collected on 46 subjects (35
with unilateral and 11 with bilateral CP). Ankle kinetic
analysis showed statistically significant improvements in the
Peak Internal Plantar Flexor Moment in Terminal Stance
for the entire, unilateral, and bilateral groups, indicating
improved strength and lever arm alignment following TAL
surgery. There were no significant differences in the kinetic
measures for the unilateral and bilateral groups, with the
exception of the magnitude of the postoperative Peak
Internal Plantar Flexor Moment in Terminal Stance, which
was significantly greater for the bilateral group than the
unilateral group (1.12MN/kg and 0.93, respectively,
P = 0.007). Kinetic evidence for the first rocker was absent
in all cases before surgery. Following surgery, a kinetic first
www.pedorthopaedics.com | 93
Soufi et al
TABLE 1. Physical Examination
Ankle Dorsiflexion (Degrees)
Knee Flexed: Mean (SD)
(Typical = 26.9 +/− 6.6)*
Pre
All (n = 60) −19 (11.4)
Unilateral (Hemiplegic, n = 42), n (%) −20.7 (11.3)
Bilateral (Diplegic, n = 18), n (%) −15.3 (10.8)
*typical values from Mudge et al.24
rocker was present in 35% of the entire group, 31% of the
unilateral group, and 50% of the bilateral group.
Kinematic analysis at the knee showed statistically
significant improvements in the Mean Knee Flexion in
Single Support for the entire, unilateral, and bilateral
groups, indicating no evidence of crouch gait due to
weakness following TAL surgery (knee kinematic plots are
shown in Fig. 3). Preoperative Mean Knee Flexion in Single
Support was significantly greater for the bilateral group
than the unilateral group (24.6 degrees and 19.7 degrees,
respectively, P = 0.016). Postoperative Mean Knee Flexion
in Single Support was significantly greater for the bilateral
group than the unilateral group (19.4 degrees and 10
degrees, respectively, P = 0.008). Postoperative Mean Knee
Flexion in Single Support was normalized in the unilateral
group but remained slightly increased in the bilateral group.
Six of the 18 subjects with bilateral CP had ankle
plantar flexion contractures on both lower extremities that
met study inclusion criteria, and in each case, the side with
the greater plantar flexion contracture was selected for in-
clusion in the study. In 5 of these 6 cases, qualitative analysis
of the physical examination, kinematic, and kinetic (when
available) outcomes for the opposite side (ie, not included in
the study) were similar to those from the index (ie, included
in the study) side. Specifically, these 5 subjects with bilateral
CP and bilateral ankle plantarflexion contractures of at least
20 degrees with the knee extended, who were treated with
bilateral TALs, showed no kinematic or kinetic evidence of
crouch gait following surgery. One of these 6 patients,
functioning at the GMFCS III level, did show kinematic
evidence at the ankle and knee of mild crouch gait when
ambulating barefoot with a rear rolling walker, 1 year and
7 months following surgery. In this case, the increased ankle
dorsiflexion and knee flexion in the stance phase were cor-
rected with the use of bilateral solid ankle foot orthoses.
Dynamic Pedobarography
Pedobarographic data were successfully collected
preoperatively and postoperatively on 44 subjects (32
with unilateral and 12 with bilateral CP; due to the use of
walker, short stride, and the inability to strike the pe-
dobarography plate). Table 3 shows data describing the
location of the CoP at Initial Contact. Of the 44 patients
with pedobarography data, the CoP was located in the
forefoot in the majority of cases (30 cases, 68%) before
surgery. In the remaining cases, the CoP at Initial
Contact was located in the midfoot. In no cases was it
94 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
Knee Extended: Mean (SD)
(Typical = 21.3 +/− 5.4)*
Post Pre Post P
9.9 (7.5) −28 (10.4) 4.7 (8.2) < 0.001
9.2 (7.0) −28.8 (10.9) 3.3 (7.9) < 0.001
11.4 (8.5) −26.4 (9.0) 7.8 (8.6) < 0.001
located in the hindfoot. Following surgery, the CoP at
Initial Contact was located in the hindfoot in 25 cases
(57%). The change in the location of the CoP at Initial
Contact was significant (P < 0.001). A similar significant
distribution, preoperatively and postoperatively, was
noted for the group with unilateral CP. In the bilateral
CP group, the majority of cases showed the location of
the CoP at Initial Contact to be in the forefoot before
surgery. Following surgery, the location was in the
midfoot in the majority of cases (these changes were not
statistically significant, P = 0.063; however, the small
sample size should be noted).
Table 4 shows data describing the duration of the
CoP in the segments of the foot (hindfoot, midfoot,
and forefoot) throughout Stance.20 Preoperative to
postoperative changes in the duration of the CoP in the
segments of the foot during Stance were statistically
significant for the entire, unilateral, and bilateral groups
(P < 0.001, = 0.003, = 0.001, respectively). For the entire
group, the CoP spent the majority of the stance phase
located in the forefoot (83.7%) before surgery. Following
surgery, the distribution of the duration of the CoP in
the segments of the foot was improved (20.7% in the
hindfoot, 37.1% in the midfoot, and 41.3% in the
forefoot). Similar changes were documented for both
the unilateral and bilateral CP groups, and there were no
significant differences between these 2 groups. These data
show improved foot loading, with no evidence of ankle
plantar flexor weakness (ie, prolonged duration of the
CoP in the hindfoot, decreased duration in the forefoot)
following TAL surgery.
Radiographic Imaging
Radiographic data were successfully collected on 59
subjects (41 with unilateral and 18 with bilateral CP).
Table 5 shows the data for the 3 angles selected to measure
hindfoot alignment.21 Statistically significant improvements
in all 3 measures were seen for the entire group, with
normalization of the calcaneal pitch following surgery.
Similar changes were documented for both the unilateral
and bilateral CP groups, though the magnitude of the
change in the talocalcaneal angle was significantly greater
for the unilateral than the bilateral group. Otherwise, there
were no significant differences between these 2 groups. No
measure showed values (eg, increased calcaneal pitch) that
would be indicative of over lengthening of the GSC
following TAL.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Surgical Management of Severe Equinus in CP

< 0.001
< 0.001

< 0.001
Mean Knee Flexion in Single Support (Degrees):

P-
Mean (SD) (Typical = 10.84+/−5.67)*

12.8 (10.4)

19.4 (12.6)
10.0 (7.9)
Post
Knee Kinematics (Flexion is “+”)

21.7 (11.8)
19.7 (12.8)

26.4 (7.8)
Pre

(Hemiplegic,

(Diplegic,
< 0.001 All (n = 60)
< 0.001 Unilateral

n = 42)

n = 18)
< 0.001 Bilateral
(Nm/kg): Mean (SD) (Typical = 1.27 +/−0.27)*
Peak Internal PF Moment in Terminal Stance

FIGURE 1. Sagittal plane ankle kinematics plots. Mean data

P

(+/− 2 SDs) from typically developing subjects are shown by

the gray bands. The preoperative plot shows increased plan-
0.97 (0.26)
0.93 (0.26)

1.12 (0.17)

tarflexion in both Stance and Swing Phases, with disruption of

Post

all 3 rockers in the Stance Phase. The postoperative plot shows

improved alignment in both Stance and Swing Phases, with
the restoration of the second and third rockers in the
Stance Phase.
0.43 (0.40)
0.45 (0.44)

0.38 (0.21)
Pre

DISCUSSION
The GSC plays an essential role in normal gait,
generating the largest moment of any single muscle group
Ankle Kinetics

(Hemiplegic,

at any point of the gait cycle.7 Ankle equinus (defined as

(Diplegic,
All (n = 46)
Unilateral

increased ankle plantarflexion during stance and swing

n = 35)

n = 11)
Bilateral

phases of the gait cycle) is the most common gait deviation

seen in children with CP who are ambulatory, and
Ankle Movement Analysis Profile

< 0.001
< 0.001

< 0.001
(Degrees): Mean (SD) (Typical

P
= 8.21+/−1.7)*

*typical values from site-specific Motion Analysis Center normative data set.
8.6 (3.0)
8.3 (2.3)

9.3 (4.1)
Post
Ankle Kinematics (Dorsiflexion is “+”, Plantarflexion is “-“).

34.4 (16.2)
35.7 (16.7)

31.4 (15.0)
Pre
13.1 (5.9)
12.4 (5.2)

14.8 (8.5)
(Typical = 13.1 ± 4.0)*
(Degrees): Mean (SD)

Post
Peak DF in Stance
TABLE 2. Kinematics and Kinetics

−28.4 (19.7)
−30.7 (20.0)

−23.1 (10.8)
Pre

FIGURE 2. Sagittal plane ankle kinetics (Internal Moments).

Mean data (+/− 2 SDs) from typically developing subjects are
shown by the gray bands. The preoperative plot shows a
n = 42), n (%)

n = 18), n (%)
(Hemiplegic,

double bump moment pattern, with the first Internal Dorsi-

(Diplegic,
All (n = 60)

flexor Moment peak greater than the second peak. The

Unilateral

Bilateral

postoperative plot shows restoration of the single bump

pattern but the persistent loss of the Loading Response
Internal Plantarflexor Moment.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 95

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Soufi et al
FIGURE 3. Sagittal plane knee kinematic plots. Mean data (+/−
2 SDs) from typically developing subjects are shown by the
gray bands. The preoperative plot shows increased flexion for
all of the Stance Phase and at the end of the Swing Phase. The
postoperative plot shows improved extension in both Stance
and Swing Phases.
correction of dynamic /myostatic deformity of the GSC by
lengthening is the most common surgery performed to
improve gait in these children.1–4,25 It is recognized that
there are multiple variables that determine the gait out-
comes following surgical lengthening of the GSC. It has
been postulated that insufficient lengthening can result in
the persistence or recurrence of the equinus gait deviation
over time in the growing child.3,4,15,18 Excessive length-
ening can result in weakness that contributes to disruption
of the ankle plantar flexion knee extension couple, leading
to the development of progressive crouch gait in the
growing child.3,4,8,14,16 Progressive crouch gait may sig-
nificantly compromise or curtail ambulatory function and
can be difficult to correct, even with aggressive surgery
and orthotic management.26 For these reasons, it has
justifiably been recommended that the risk of over-
lengthening be minimized by avoiding the most aggressive
surgical lengthening techniques (eg, zone 3 or TAL
lengthening).
There is extensive literature considering the outcomes
following lengthening of the GSC in children with CP, the
goals of which are to identify risk factors for both under and
over-lengthening. Two systematic reviews found that poor
study quality; a lack of consistent definitions of gait
TABLE 3. Pedobarography
Location of Center of Pressure at Initial Contact
Forefoot Count, (%)
Pre Post
Total (n = 44) 30 (68) 10 (23)
Unilateral (Hemiplegic, n = 32) 23 (72) 8 (25)
Bilateral (Diplegic, n = 12) 7 (58) 2 (17)
96 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
deviations and outcomes; and marked variability in study
cohorts, surgical indications, surgical techniques, and or-
thotic and therapy management after surgery made com-
parisons among and between studies difficult.3,4 In children
with unilateral CP, recurrent equinus ranged from 0% to
38%, and calcaneal gait from 4% to 30%. In children with
bilateral CP, recurrent equinus ranged from 0% to 35%, and
calcaneal gait from 0% to 40%.3 Recurrent equinus ranged
from 0% to 18% in studies with less than 4 years of follow-
up and 16% to 38% in studies with more than four years of
follow-up.3 For calcaneus, studies with less than 4 years of
follow-up reported a rate of 0% to 30%, while studies with
more than 4 years of follow-up reported a rate of 3% to
40%.3 Age at index surgery (younger at greater risk), sex,
topographic type of CP (subjects with unilateral CP are at
greater risk for recurrent equinus, and those with bilateral
CP are at greater risk for crouch gait), level of functional
motor impairment (ie, GMFCS level), surgical technique,
single versus multilevel surgery, magnitude of concomitant
hamstring contracture, use of orthoses and physical therapy
following surgery, access to QGA for pre-operative clinical
decision making, ankle kinematics at Initial Contact (the
greater the plantarflexion the greater the risk of recurrence),
and duration of follow-up have all been considered as risk
factors for either under or over correction.3–5,17,27
In cases with severe myostatic equinus deformities,
adequate surgical correction (ie, achieving plantigrade
foot alignment) can only be achieved by TAL. This is the
first study to document outcomes following TAL per-
formed to correct severe equinus deformities in children
with CP, where clinical decision-making and surgical
dosing for SEMLS was informed by QGA, and post-
operative rehabilitation included standardized orthotics
management and physical therapy. At the time of short-
term follow-up (mean 1.3 y, median 1.1 y, range 1.0 to
6.1 y), the subjects showed good results with respect to
improved ankle range of motion on physical examination,
improved sagittal plane ankle kinematics (significant im-
provements in Peak Ankle Dorsiflexion in Stance and
Ankle GVS), ankle kinetics (significant improvements in
the Peak Internal Plantar Flexor Moment in Terminal
Stance), knee kinematics (significant improvement in
Mean Knee Flexion in Single Support), pedobarography
(significant improvement in the location of the CoP at
Initial Contact, and duration of the CoP in the segments
of the foot during Stance), and radiography (improve-
ments, the majority of which were significant, in 3 mea-
sures of hindfoot alignment). These data outcomes do not
show evidence for over-lengthening or excessive weakness
Midfoot Count, (%) Hindfoot Count, (%)
Pre Post Pre Post P
14 (32) 9 (20) 0 (0) 25 (57) < 0.001
9 (28) 9 (28) 0 (0) 15 (47) < 0.001
5 (42) 0 (0) 0 (0) 10 (83) 0.063
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Surgical Management of Severe Equinus in CP

TABLE 4. Pedobarography: Duration of Center of Pressure

Duration of Center of Pressure (% Stance Phase)
Forefoot: Mean (SD) Midfoot: Mean (SD) Hindfoot: Mean (SD)
(Typical = 53.3 +/− 10.4)* (Typical = 14.5 +/− 6.6)* (Typical = 32.2 +/− 7.9)*
Pre Post Pre Post Pre Post P
Total (n = 44) 83.7 (30.1) 41.3 (16.7) 15.9 (29.9) 37.1 (17.7) 0.1 (0.7) 20.7 (19.7) < 0.001
Unilateral 82.1 (33.4) 42.0 (17.7) 17.6 (33.1) 38.0 (18.1) 0.1 (0.7) 19.6 (20.2) 0.003
(Hemiplegic, n = 32)
Bilateral (Diplegic, 88.3 (19.2) 39.7 (14.2) 11.3 (19.1) 34.8 (17.6) 0.2 (0.6) 23.5 (17.0) 0.001
n = 12)
*normative data from Jamison et al.20

following TAL performed within the clinical decision-
making and management paradigm embraced by both
institutions that provided treatment for these subjects.
There were no significant differences between unilateral
and bilateral subjects for any of the preoperative or
postoperative physical examination measures, ankle kin-
ematic measures, ankle kinetic measures (with the ex-
ception of the magnitude of the postoperative Peak
Internal Plantar Flexor Moment in Terminal Stance,
which was greater for the bilateral group), dynamic pe-
dobarography, and radiographic imaging of the foot.
Improved knee kinematic were seen in both unilateral and
bilateral groups, though preoperative and postoperative
Mean Knee Flexion in Single Support was significantly
greater for the bilateral group; and postoperative Mean
Knee Flexion in Single Support was normalized in the
unilateral group but remained slightly increased in the
bilateral group.
There were ~1100 SEMLS surgeries performed at
the 2 institutions during the duration of the study, with
only 60 cases (5%) with severe equinus requiring TAL to
achieve plantigrade ankle/foot alignment at the time of
surgery. Only a small fraction of patients with ambula-
tory CP and equinus in our institutions have been treated
with TAL, with the vast majority managed with zone 1
or 2 lengthenings. Only 2 patients who functioned at
the GMFCS III level were identified as undergoing
TAL. With the indication criteria we have defined,
the desired intraoperative range of dorsiflexion (0 to
5 degrees) and our postoperative orthotic and PT pro-
tocol, TAL seems to be a safe and effective intervention
in the short term.
This study indirectly supports the recommendations
in the literature that the vast majority of patients with CP
who are ambulatory and have equinus contractures that
have failed nonoperative treatments and require surgery
can be managed with either zone 1 or 2 surgical length-
ening of the GSC.5,17,19 The results of this study are
consistent with previous literature that suggest that TAL
in a subject with unilateral CP seems to be a safe proce-
dure relative to the risk of over-lengthening and sub-
sequent weakness affecting gait.14,15,18 However, the
results in this study for the subjects with bilateral CP
should be interpreted with caution. The literature clearly
identifies bilateral CP and longer follow-up as risk factors
for crouch gait following TAL in this group of
patients.3,4,8 The short time to follow-up in the current
study, the failure to fully correct the Mean Knee Flexion
in Single Support in the subjects with bilateral CP, and the
single case of a subject with bilateral CP who underwent
bilateral TALs and was found to have mild crouch gait at
the 1 year and 7 months follow-up are all causes for
concern. Surgeons should proceed with great caution, use
QGA for surgical decision-making, use orthotics and
physical therapy aggressively following surgery, and
maintain frequent follow-up, in patients with bilateral CP
and severe equinus contractures that are managed by
TAL. The strengths of this study are the consistent sur-
gical decision-making and intraoperative surgical dosing
for SEMLS utilizing QGA, and postoperative re-
habilitation included standardized orthotics management
and physical therapy; the comprehensive data collection
before and after surgery; and the high volume of cases
treated at each institution. The limitations are related to

TABLE 5. Radiography
Radiographic Measurements (Degrees)
Calcaneal Pitch: Mean (SD) Tibiocalcaneal Angle: Mean (SD) Lateral Talocalcaneal Angle:
(Typical = 17 +/− 6)* (Typical = 69 +/− 8.4)* Mean (SD) (Typical = 49 +/− 6.9)*
Pre Post P Pre Post P Pre Post P
Total (n = 59), n (%) 3.2 (22.4) 18.6 (6.3) < 0.001 77.7 (34.5) 57.4 (24.7) < 0.001 42.8 (10.5) 50.2 (9.4) < 0.001
Unilateral (n = 41), n (%) 3.4 (24.2) 19.4 (6.3) < 0.001 81.3 (34.8) 58.4 (24.7) < 0.001 42.8 (10.6) 51.9 (9.5) < 0.001
Bilateral (n = 18), n (%) 2.7 (18.2) 16.3 (5.6) = 0.005 69.5 (33.3) 54.9 (24.6) = 0.003 42.9 (10.7) 46.2 (8.3) = 0.152
*typical data from Davids, et al.21

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 97

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Soufi et al
the relatively short-term follow-up, the small sizes of the
unilateral and bilateral patient cohorts, the small number
of patients functioning at GMFCS III level, and the lack
of functional and patient/parent-reported outcomes.
This study does not advocate for the widespread use of
TAL to correct equinus deformity in children with CP.
However, it does show that good short-term outcomes fol-
lowing TAL are possible in properly selected subjects with
severe contractures when the dosing of the surgery is
standardized (correction of contracture to between 0 and 5
degrees of dorsiflexion with the knee extended), and the
procedure is performed in the setting of QGA-guided
SEMLS with consistent postoperative orthotic management
and rehabilitation.
REFERENCES
1. Rethlefsen SA, Blumstein G, Kay RM, et al. Prevalence of specific
gait abnormalities in children with cerebral palsy revisited: influence
of age, prior surgery, and Gross Motor Function Classification
System level. Dev Med Child Neurol. 2017;59:79–88.
2. Wren TAL, Rethlefsen S, Kay RM. Prevalence of specific gait
abnormalities in children with cerebral palsy: influence of cerebral palsy
subtype, age, and previous surgery. J Pediatr Orthop. 2005;25:79–83.
3. Ma N, Sclavos N, Passmore E, et al. Three-dimensional gait analysis in
children undergoing gastrocsoleus lengthening for equinus secondary
to cerebral palsy. Medicina (Kaunas). 2021;57:98.
4. Shore BJ, White N, Kerr GH. Surgical correction of equinus
deformity in children with cerebral palsy: a systematic review. J Child
Orthop. 2010;4:277–290.
5. Steinwender G, Saraph V, Zwick EB, et al. Fixed and dynamic
equinus in cerebral palsy: evaluation of ankle function after
multilevel surgery. J Pediatr Orthop. 2001;21:102–107.
6. Firth GB, McMullan M, Chin T, et al. Lengthening of the
gastrocnemius-soleus complex: an anatomical and biomechanical
study in human cadavers. J Bone Joint Surg Am. 2013;95:1489–1496.
7. Sutherland D, Cooper L, Daniel D. The role of the ankle plantar
flexors in normal walking. J Bone Joint Surg Am. 1980;62:354–363.
8. Vuillermin C, Rodda J, Rutz E, et al. Severe crouch gait in spastic
diplegia can be prevented: a population-based study. J Bone Joint
Surg Br. 2011;93:1670–1675.
9. Gendy S, ElGebeily M, El-Sobky TA, et al. Current practice and
preferences to management of equinus in children with ambulatory
cerebral palsy: a survey of orthopedic surgeons. SICOT J. 2019;5:3.
10. Rutz E, McCarthy J, Shore BJ, et al. Indications for gastrocsoleus
lengthening in ambulatory children with cerebral palsy: a Delphi
consensus study. J Child Orthop. 2020;14:405–414.
11. Brunner R, Jaspers RT, Pel JJ, et al. Acute and long-term effects on
muscle force after intramuscular aponeurotic lengthening. Clin
Orthop Relat Res. 2000;378:264–273.
98 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
12. Delp SL, Statler K, Carroll NC. Preserving plantar flexion
strength after surgical treatment for contracture of the triceps
surae: a computer simulation study. J Orthop Res. 1995;13:
96–104.
13. Rivares C, Brunner R, Pel JJ, et al. Remodeling of rat m.
gastrocnemius medialis during recovery from aponeurotomy. Front
Physiol. 2020:1218.
14. Borton DC, Walker K, Pirpiris M, et al. Isolated calf lengthening in
cerebral palsy. Outcome analysis of risk factors. J Bone Joint Surg
Br. 2001;83:364–370.
15. Chung CY, Sung KH, Lee KM, et al. Recurrence of equinus foot
deformity after tendo-achilles lengthening in patients with cerebral
palsy. J Pediatr Orthop. 2015;35:419–425.
16. Dietz FR, Albright JC, Dolan L. Medium-term follow-up of Achilles
tendon lengthening in the treatment of ankle equinus in cerebral
palsy. Iowa Orthop J. 2006;26:27–32.
17. Dreher T, Buccoliero T, Wolf SI, et al. Long-term results after
gastrocnemius-soleus intramuscular aponeurotic recession as a part
of multilevel surgery in spastic diplegic cerebral palsy. J Bone Joint
Surg Am. 2012;94:627–637.
18. Joo SY, Knowtharapu DN, Rogers KJ, et al. Recurrence after
surgery for equinus foot deformity in children with cerebral palsy:
assessment of predisposing factors for recurrence in a long-term
follow-up study. J Child Orthop. 2011;5:289–296.
19. Kay RM, Rethlefsen SA, Ryan JA, et al. Outcome of gastro-
cnemius recession and tendo-achilles lengthening in ambu-
latory children with cerebral palsy. J Pediatr Orthop B. 2004;13:
92–98.
20. Jamison EG, Davids JR, Anderson JP, et al. Dynamic pedobarog-
raphy for children: use of the center of pressure progression. J Pediatr
Orthop. 2008;28:254–258.
21. Davids JR, Gibson TW, Pugh LI. Quantitative segmental analysis of
weight-bearing radiographs of the foot and ankle for children:
normal alignment. J Pediatr Orthop. 2005;25:769–776.
22. Davis RB, Õunpuu S, Tyburski D, et al. A gait analysis data
collection and reduction technique. Hum Mov Sci. 1991;10:
575–587.
23. Baker R, McGinley JL, Schwartz MH, et al. The gait profile
score and movement analysis profile. Gait Posture. 2009;30:
265–269.
24. Mudge AJ, Bau KV, Purcell LN, et al. Normative reference values
for lower limb joint range, bone torsion, and alignment in children
aged 4–16 years. J Pediatr Orthop Part B.. 2014;23:15–25.
25. Horsch A, Götze M, Geisbüsch A, et al. Prevalence and classification
of equinus foot in bilateral spastic cerebral palsy. World J Pediatr.
2019;15:276–280.
26. Rodda J, Graham HK, Nattrass G, et al. Correction of severe
crouch gait in patients with spastic diplegia with use of multi-
level orthopaedic surgery. J Bone Joint Surg Am. 2006;88:
2653–2664.
27. Lofterød B, Terjesen T. Local and distant effects of isolated calf
muscle lengthening in children with cerebral palsy and equinus gait.
J Child Orthop. 2008;2:55–61.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
The Diagnosis of Developmental Dysplasia of the Hip
From Hip Ultrasonography Images With Deep Learning
Methods
Hakan Atalar, MD,* Kemal Üreten, MD, PhD,†‡ Gül Tokdemir, PhD,§ Tolga Tolunay, MD,*
Murat Çiçeklidağ, MD,* and Osman Şahap Atik, MD∥
Background: Hip ultrasonography is very important in the early
diagnosis of developmental dysplasia of the hip. The application
of deep learning-based medical image analysis to computer-aided
diagnosis has the potential to provide decision-making support
to clinicians and improve the accuracy and efficiency of various
diagnostic and treatment processes. This has encouraged new
research and development efforts in computer-aided diagnosis.
The aim of this study was to evaluate hip sonograms using
computer-assisted deep-learning methods.
Methods: The study included 376 sonograms evaluated as normal
according to the Graf method, 541 images with dysplasia and
365 images with incorrect probe position. To classify the devel-
opmental hip dysplasia ultrasound images, transfer learning was
applied with pretrained VGG-16, ResNet-101, MobileNetV2
and GoogLeNet networks. The performances of the networks
were evaluated with the performance parameters of accuracy,
sensitivity, specificity, precision, F1 score, and AUC (area under
the ROC curve).
Results: The accuracy, sensitivity, specificity, precision, F1 score,
and AUC results obtained by testing the VGG-16, ResNet-101,
MobileNetV2, and GoogLeNet models showed performance
> 80%. With the pretrained VGG-19 model, 93%, 93.5%, 96.7%,
92.3%, 92.6%, and 0.99 accuracy, sensitivity, specificity, pre-
cision, F1 score, and AUC results were obtained, respectively.
Conclusion: In this study, in addition to the ultrasonography
images of dysplastic and healthy hips, images were also included
of probe malpositioning, and these images were able to be suc-
cessfully evaluated with deep learning methods. On the sono-
grams, which provided criteria appropriate for evaluation,
successful differentiation could be made of healthy hips and
dysplastic hips.
Level of Evidence: Level-IV; diagnostic studies.
From the *Department of Orthopaedics and Traumatology, Faculty of
Medicine, Gazi University; †Department of Rheumatology, Faculty of
Medicine, Ufuk University; ‡Department of Computer Engineering;
§Department of Computer Engineering, Faculty of Engineering, Çankaya
University; and ∥Turkish Joint Diseases Foundation, Çankaya, Ankara,
Turkey.
No funding has been received to support this study.
The authors declare no conflicts of interest.
Reprints: Hakan Atalar, MD, Department of Orthopaedics and Trau-
matology, Faculty of Medicine, Gazi University, Ankara, 06560,
Turkey. E-mail: atalarhakan@yahoo.comss.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002294
e132 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Key Words: hip, ultrasonography, deep learning
(J Pediatr Orthop 2023;43:e132–e137)
D evelopmental Dysplasia of the Hip (DDH) is an or-
thopaedic condition that should be diagnosed and
treated early in infancy. One of the methods currently
widely used in the evaluation of newborn hips is the Graf
method based on the evaluation of coronal hip sonograms.
This method was developed by Dr Graf at the beginning
of the 1980s and then came into widespread use. However,
there may be differences in Graf method evaluations be-
tween examiners with and without an Advanced Speci-
alized Practice Course in Infantile Hip Ultrasonography.1
Artificial intelligence (AI) methods have been used in
many fields in recent years, including the applications of
virtual assistants, games, smart cars, smartphones, fraud
detection and prevention, smart homes and security sys-
tems, and artificial intelligence studies continue to develop
with increasing momentum. There have been many suc-
cessful studies recently, which have used the deep learning
method of convolutional neural networks (CNN), especially
in the field of image processing.2,3 The layers in the CNN
hidden layer are mainly convolution layers, pooling layers,
and a fully connected layer. Feature extraction is performed
in the convolution layers of the image imported to the net-
work through the input layer. In the pooling layer, the di-
mension of the inputs is reduced. The fully connected layer
is often used as a classifier, and a softmax classifier is added
to the output layer to calculate the prediction probability.4–6
There have been many recent studies that have
successfully used CNN on magnetic resonance7 images, plain
radiographs,8 computed tomography images,9 and ultra-
sonography images.8,10 Thousands of images are required to
train a CNN model from the start. However, it is difficult to
find enough labeled data in the medical field. There are cur-
rently some pretrained networks, which have been developed
from the Imagenet data set, and transfer learning is used with
these models for the new task. Some of these are AlexNet,11
VGG-16,12 GoogLeNet,13 ResNet-50,14 DenseNet,15 and
MobileNet.16 Another method applied to increase the per-
formance of the model in deep learning methods is data
augmentation, which is the artificial increasing of the number
of data by cropping, scaling, flipping, and rotating the images.
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
The aim of this study was to develop a deep learning
method for the diagnosis of DDH using hip ultra-
sonography images obtained with the Graf method.
Images of neonatal dysplasia of the hip, normal hip im-
ages, and ultrasound probe malposition images were used
in the study. Deep learning methods were applied with
pretrained VGG-19, ResNet-101, MobileNetV2, and
GoogLeNet networks.
METHODS
Data Set
Approval for this retrospective study was granted by
the Institutional Review Board (Number: E2-22-1655).
The images were obtained from the personal archive of an
orthopaedic specialist (H.A.) with 20 years of experience
on the subject of hip ultrasonography and were evaluated
for typing and consistency with the Graf method by the
same specialist. Ultrasonography was performed in all
infants with a 7.5 MHz linear transducer (Logiq E; GE
Healthcare Co., Ltd, China). The images in this study
were evaluated by separating them into 3 groups of nor-
mal hip sonogram (Type 1), abnormal hip sonogram, and
sonograms that did not meet the standard criteria for
evaluation. The study was performed on 376 sonograms
from 52 hips evaluated as normal (Type 1) according to
the Graf method, 541 abnormal sonograms (dysplasia
class) from 64 hips (Type 2a, b; 303 sonograms, Type 2c;
88 sonograms, Type D: 14 sonograms, Type 3; 107
sonograms, Type 4; 29 sonograms) and 365 sonograms
that did not meet standard evaluation criteria.
Data Processing Environment
This study was carried out on a computer with a
GeForce RTX2060 graphics processing unit and in
MATLAB and Image Processing Toolbox. The data were
analyzed by Kemal Üreten and Gül Tokdemir.
Data Preprocessing
The ultrasound images used in this study had the
dimensions of 819 × 590 pixels. Noise interference on the
images, such as patient name, hospital name, date, and
FIGURE 1. Hip ultrasonography original images.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Diagnosis of DDH
some numbers, which are not necessary for training and
may negatively affect the results, were automatically
cropped using the main image coordinates with the help of
the “for loop” function (a software term, each image is
processed 1 by 1 automatically). Figure 1 shows examples
of the original hip ultrasonography images, and Figure 2
shows examples of preprocessed hip ultrasonography
images. The dimensions of the cropped images were
235 × 300 pixels, and these images were resized to
224 × 224 and 300 × 300 pixels. The input image size
should be 300 × 300 pixels for the pretrained ResNet
network, and 224 × 224 pixels for VGG-16, MobileNet,
and GoogLeNet networks. For testing the models, 100
dysplasia images were obtained from 16 hips, 70 normal
images were obtained from 12 hips, and 73 incorrect probe
position images were used. These images were used only
for testing and were not used during training and
validation. The numbers of images used for training,
validation, and testing are shown in Table 1.
Transfer Learning, Data Augmentation
In this study, which was conducted to classify
developmental hip dysplasia ultrasound images, transfer
learning was applied with pretrained VGG-16, ResNet-
101, MobileNetV2, and GoogLeNet networks. These
models were tested with test data that were not sub-
sequently used for training or validation. Rotation (−20,
20 degrees), translation (−30, 30 pixels) and scaling (0.9,
1.1 range) were applied to the images for data augmenta-
tion. The training parameters were set as follows; opti-
mizer: sgdm (stochastic gradient descent with momentum),
mini-batch size:16, initial learning rate: 3e-4, L2 regulari-
zation: 0.004, validation frequency: 16, and number of
epochs: 20.
Statistical Analysis
As a result of the classification, performances of the
networks were evaluated with the performance parameters
of accuracy, sensitivity, specificity, precision, F1 score,
and AUC (area under the ROC curve). These perfor-
mance metrics were calculated using macro averages from
the confusion matrix obtained during the testing of the
www.pedorthopaedics.com | e133
Atalar et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 2. Preprocessed hip ultrasonography images.

models. (TP indicates true positive; FP, false positive; TN, with these 4 groups (dysplasia, borderline, probe incorrect
true negative; FN, false negative). position, and normal classes) with the pretrained VGG-16
TP + TN network, 85.2%, 85.5%, 95.3%, 83.8%, 83.5%, and 0.97
accuracy = , accuracy, sensitivity, specificity, precision, F1 score, and
TP + TN + FP + FN AUC results were obtained, respectively.
TP
sensitivity(recall) = ,
TP + FN
TN DISCUSSION
specificity = , The ultrasound images used in this study for DDH
TN + FP
diagnosis were applied with transfer learning with pre-
TP trained VGG-16, ResNet-101, MobileNetV2, and Goo-
precision = , gLeNet networks. To prevent overfitting, the methods of
TP + FP
data augmentation, dropout, and L2 regularization were
precision × recall performed. No overfitting was observed in the test results
F1score = 2 × . of the models or in the training graphs. The study results
precision + recall
showed that the pretrained VGG-16 model outperformed
the other pretrained networks. This may be due to dif-
ferences in kernel size (receptive-field filter size) or the
RESULTS architecture of the networks and we observed similar re-
Table 2 shows the results of accuracy, sensitivity, sults in our previous studies with plain radiographs.
specificity, precision, F1 score, and AUC obtained by The development of the bone and cartilage of the
testing the VGG-16, ResNet-101, MobileNetV2, and acetabulum is evaluated with alpha and beta angles in the
GoogLeNet models. Figure 3 shows the confusion Graf method.17,18 In a Graf type 1 mature hip, the alpha
matrices of the models obtained by testing, and Figure 4 angle is > 60 degrees and does not deteriorate over time.
shows the prediction results of 6 randomly selected images Hips with an alpha angle <50 degrees are evaluated as
during testing with the VGG-16 model. dysplastic hips (Type 2c, Type 2d, Type 3, and Type 4). In a
The dysplasia class consisted of 238 dysplastic hip child younger than 3 months, hips with an alpha angle of 50
images obtained from 24 hips and 303 borderline hip to 60 degrees are physiologically immature (Type 2a), but
images obtained from 40 hips. A new group (group 4) was these hips are considered pathologically immature if the
formed by separating the borderline ultrasound images child is older than 3 months. The interobserver agreement
(Types 2a, b) within the dysplasia class. In the training of this method in the differentiation of normal and patho-
logic hips is substantial and has been reported as a moderate
TABLE 1. Number of Images Used for Training, Validation, and
agreement in the differentiation of Graf types.19,20 Infants
Testing at risk for DDH may need radiologic follow-up.21 To
evaluate a hip sonogram applied with the Graf method
Training Validation Test Total
defined on the checklist as mature hip (alpha angle > 60
Dysplasia class 353 88 100 541 degrees), dysplastic hip (alpha angle <50 degrees), or bor-
Incorrect probe position 232 56 73 365 derline hip (Types 2a, b: alpha angle 50 to 60 degrees), it is
Normal class 245 61 70 376
important to avoid incorrect clinical application.

e134 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

 J Pediatr Orthop  Volume 43, Number 2, February 2023 Diagnosis of DDH

attributed this to the incorrect application of the Graf

TABLE 2. Performance Results Obtained by Testing
the Models method, and have emphasized the importance of obtain-
ing a standard sonogram appropriate to the evaluation
VGG- ResNet-
criteria of hip sonograms.24,25
16 (%) 101 (%) MobileNetV2 (%) GoogLeNet (%)
Unlike the current study, previous studies in the
Accuracy 93 89.3 82.3 82.3 literature have focused on the differentiation of normal
Sensitivity 93.5 90.2 83.6 84.4
Specificity 96.7 94.8 91.2 91.6
and pathologic hips.8,10,26 Oelen D et al27 showed that the
Precision 92.3 88.4 81.6 80.8 accuracy of physicians in measuring angles in hip ultra-
F1 score 92.6 88.9 81.9 81.8 sonography is lower than deep learning-based algorithms
AUC 0.99 0.97 0.95 0.95 for determining angles on newborn hip ultrasound. In the
AUC indicates area under the ROC curve. current study, in addition to the differentiation of standard
and nonstandard sonograms (ultrasound probe malposi-
tion), successful differentiation was made of mature hips
The iliac wing, acetabulum contours, and tissues and dysplastic hips on standard sonograms. Although
such as the labrum and the femoral osteochondral angle measurements are used by the clinician in the clas-
junction, which form the basis of the Graf method, were sification of hip ultrasonography images, it is interesting
successfully evaluated in this study. The differentiation of that the deep learning methods used in the study suc-
standard and nonstandard sonograms conforming to the cessfully evaluated the ultrasonography images without
evaluations made according to the checklist defined by measuring the angle. In a study of 321 sonograms by Lee
Graf could be made successfully. The ability to make this and colleagues agreement of the clinician and AI was
differentiation is important in terms of preventing mis- found to be at a moderate level (80.9%) in the determi-
diagnosis of normal or pathologic hips by evaluating nation of hips that met and did not meet the criteria, with
nonstandard sonograms. Cases have been reported in the which measurements could be made according to the Graf
literature, which has been diagnosed as Graf type 1 and method. Of the hips that were appropriate for measure-
have then developed dysplasia.22,23 Other studies have ment, the AI system was found to differentiate normal and

FIGURE 3. Confusion matrices of the pretrained models; VGG-16 model (A), ResNet-101 model (B), GoogLeNet model (C),
MobileNetV2 model (D).

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e135

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Atalar et al
FIGURE 4. A normal hip (A), a type 2a hip (B), a type 2c hip (C), a type D hip (D), a type 4 hip (E), and a hip with incorrect probe
position (F).
abnormal hip ultrasonography images with an agreement
rate of 84.37% with the clinician.26 Hareendranathan
et al10 successfully evaluated 3D ultrasound images using
a semiautomatic segmentation technique to differentiate
normal and pathologic hips. In another study by Quader
N. et al,8 2D sonograms were evaluated using an auto-
matic near-real-time method, and the evaluation was
successful in determining whether or not the sonograms
met the standard slice definition according to Graf. A
moderate level of interobserver agreement has been re-
ported for the Graf method in differentiating pathologic
hip types.19,20 In the current study, differentiation within
the pathologic Graf types was not evaluated. As these hips
usually require treatment, this differentiation was not
considered to be of clinical importance.
Limitations of this study could be said to be the
amount of data and that a single ultrasonography device
was used. To be able to generalize these results,
further studies are needed with a greater number of
e136 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
ultrasonography images. Real-time application allows
for use in clinical practice.
The experience of examiners who perform hip ul-
trasonography and interpret images is important in the
management of the Graf method, and the evaluation re-
sults of those who are experienced in this method and
those who are not may differ.1 According to Graf, ig-
noring the defined checklist leads to misdiagnoses and
images related to probe incorrect position should not be
taken into account.24 Deep learning-assisted hip ultra-
sonography applications can reduce the risk of making
mistakes for those who are not experienced in this field
and accelerate their learning processes. In conclusion,
successful differentiation of normal hips, pathologic hips,
and nonstandard sonograms that may lead to mis-
diagnosis with the use of deep learning methods, and early
diagnosis and treatment of DDH with hip ultra-
sonography supported by deep learning methods may be
possible.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
REFERENCES
1. Ömeroğlu H. Significant assessment discrepancies occur between the
examiners having and not having further special hands-on course
trainings in infantile hip ultrasonography by the Graf method. Indian
J Orthop. 2021;55:1523–1528.
2. Hamet P, Tremblay J. Artificial intelligence in medicine. Metabo-
lism. 2017;69S:S36–S40.
3. Hessler G, Baringhaus KH. Artificial intelligence in drug design.
Molecules. 2018;23:2520.
4. Mintz Y, Brodie R. Introduction to artificial intelligence in medicine.
Minim Invasive Ther Allied Technol. 2019;28:73–81.
5. Nichols JA, Herbert Chan HW, Baker MAB. Machine learning:
applications of artificial intelligence to imaging and diagnosis.
Biophys Rev. 2019;11:111–118.
6. Sakib S, Ahmed N, Kabir AJ, et al. An overview of convolutional
neural network: its architecture and applications. Preprints. 2018
2018110546. Available at: https://www.preprints.org/manuscript/
201811.0546/v4.
7. Rai HM, Chatterjee K. Detection of brain abnormality by a novel
Lu-Net deep neural CNN model from MR images. Machine
Learning with Applications. 2020;2:100004.
8. Quader N, Hodgson AJ, Mulpuri K, et al. Automatic evaluation of
scan adequacy and dysplasia metrics in 2D ultrasound Images of the
neonatal hip. Ultrasound Med Biol. 2017;43:1252–1262.
9. Liu F, Guan B, Zhou Z, et al. Fully automated diagnosis of anterior
cruciate ligament tears on knee MR images by using deep learning.
Radiol Artif Intell. 2019;1:180091.
10. Hareendranathan AR, Mabee M, Punithakumar K, et al. A
technique for semiautomatic segmentation of echogenic structures
in 3D ultrasound, applied to infant hip dysplasia. Int J Comput Assist
Radiol Surg. 2016;11:31–42.
11. Krizhevsky A, Sutskever I, Hinton GE. Imagenet classication with
deep convolutional neural networks. In Advances 37 in Neural
Information Processing Systems. 2012:1097–1105. Available at:
https://www.cv-foundation.org/openaccess/content_cvpr_2015/html/
Szegedy_Going_Deeper_With_2015_CVPR_paper.html.
12. Simonyan K, Zisserman A. Very deep convolutional networks for
large-scale image recognition. arXiv preprint arXiv. 2014;1409:
1556.
13. Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions. In
Proceedings of the IEEE conference on computer vision and pattern
recognition. 2015:1–9.
14. He K, Zhang X, Ren S, et al. Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR). 2016:770–778.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Diagnosis of DDH
Available at: https://openaccess.thecvf.com/content_cvpr_2016/
papers/He_Deep_Residual_Learning_CVPR_2016_paper.pdf.
15. Huang G, Liu Z, Maaten L v d, et al. Weinberger; Proceedings of the
IEEE Conference on Computer Vision and Pattern Recognition
(CVPR). J Software Eng Appl. 2017;12:4700–4708.
16. Howard AG, Zhu M, Chen B, et al. Mobilenets: efficient convolu-
tional neural networks for mobile vision applications. arXiv preprint
arXiv. 2017;1704:04861.
17. Graf R. The diagnosis of congenital hip-joint dislocation by the ultrasonic
Combound treatment. Arch Orthop Trauma Surg. 1980;97:117–133.
18. Morris WZ, Mayfield L, Beckwith T, et al. Late hip dysplasia after
normal ultrasound in breech babies: implications on surveillance
recommendations. J Pediatr Orthop. 2021;41:e304–e308.
19. Roovers EA, Boere-Boonekamp MM, Geertsma TS, et al. Ultra-
sonographic screening for developmental dysplasia of the hip in
infants. Reproducibility of assessments made by radiographers.
J Bone Joint Surg Br. 2003;85:726–730.
20. Simon EA, Saur F, Buerge M, et al. Interobserver agreement of
ultrasonographic measurement of alpha and beta angles and the final
type classification based on the Graf method. Swiss Med Wkly.
2004;134:671–677.
21. Atalar H, Arıkan M, Tolunay T, et al. The infants who have mature
hip on ultrasonography but have risk factors of developmental
dysplasia of the hip are required radiographic examination. Jt Dis
Relat Surg. 2021;32:598–604.
22. Rafique A, Set P, Berman L. Late presentation of developmental
dysplasia of the hip following normal ultrasound examination. Clin
Radiol. 2007;62:181–184.
23. Gwynne Jones DP, Dunbar JD, Theis JC. Late presenting
dislocation of sonographically stable hips. J Pediatr Orthop B.
2006;15:257–261.
24. Graf R, Mohajer M, Plattner F. Hip sonography update. Quality-
management, catastrophes tips and tricks. Med Ultrason. 2013;15:
299–303.
25. Gunay C, Atalar H, Kaptan AY, et al. Can a hip diagnosed as Graf
Type 1according to Graf checklist deteriorate over time? a case series
and evaluation of the Graf method. J Orthop Case Rep. 2019;9:7–10.
26. Lee SW, Ye HU, Lee KJ, et al. Accuracy of new deep learning
model-based segmentation and key-point multi-detection method for
ultrasonographic developmental dysplasia of the hip (DDH) screen-
ing. Diagnostics (Basel). 2021;11:1174.
27. Oelen D, Kaiser P, Baumann T, et al. Accuracy of trained physicians
is inferior to deep learning-based algorithm for determining angles in
ultrasound of the newborn hip. Ultraschall Med. 2020;43:49–55.
www.pedorthopaedics.com | e137
 ORIGINAL ARTICLE
Application of Accepted Use Criteria for the Treatment
of Developmental Dysplasia of the Hip Decreases
the Number of Infants Treated With a Pavlik Harness
Margaret S. Murphy-Zane, MD,*† Patrick M. Carry, PhD,*† Reba L. Salton, BS,†
Nancy Hadley Miller, MD,*† Kaley Holmes, BA,† Tyler Freeman, MD,* Matthew Belton, MD,*
Brian Kohuth, PA-C,† Deborah Burke, PA-C,† and Gaia Georgopoulos, MD*†
Background: This analysis examined how the application of the
American Academy of Orthopedic Surgeons appropriate use
criteria (AUC) for developmental dysplasia of the hip in infants
would change treatment patterns and outcomes for Graf IIA hips
at a single quaternary pediatric hospital.
Methods: After Institutional Review Board approval, patient
medical records were reviewed and data were collected. Graf IIa
hips were defined as alpha angle (AA) 50 to 59 degrees. AA and
femoral head coverage (FHC) were measured from initial and
6-month ultrasounds and acetabular index (AI) was measured
from radiographs at 6 months of age. Instability (positive Or-
tolani and Barlow tests) was noted. On the basis of the American
Academy of Orthopedic Surgeons AUC for managing devel-
opmental dysplasia of the hip, hips were further categorized as
normal (FHC ≥ 45%), borderline (FHC 35% to 44%), or dys-
plastic (FHC <35%).
Results: Overall, 13% (49/371) of Graf IIa hips (AA 50 to 59
degrees) were dysplastic (FHC <35%). Total 24% (89/371) were
clinically unstable. Total 42% (37/89) of unstable Graf IIa hips
were dysplastic. Only 4% of stable Graf IIa hips were dysplastic
(12/282). Out of 371 Graf IIa hips, 256 were treated with Pavlik
harness (n = 250) or Rhino brace (n = 6). Among stable, non-
dysplastic (SND) hips (those with normal and borderline
FHC ≥ 35%), 33% (52/158) were treated because of a more severe
contralateral side. If the AUC had been applied, 67% (106/158)
of SND Graf IIa hips would not have been treated. Among the
n = 162 hips that returned for a 6-month radiograph, there was
no difference in AI in the 115 treated and 47 untreated SND hips
(mean difference treatment vs. no treatment: −1.5, 95% CI, −3.1
to 0.2, P = 0.0808).
Conclusions: Using AUC recommendations, our center could
reduce the number of SND Graf IIa hips we treat by 67%. Al-
though 24% of Graf IIa hips were clinically unstable and 13%
From the *University of Colorado School of Medicine; and †Children’s
Hospital Colorado, Orthopedics Institute, Aurora, Colorado.
No funding was secured for this study.
The authors declare no conflicts of interest.
Reprints: Margaret S. Murphy-Zane, MD, Department of Pediatric Ortho-
pedic Surgery, Children’s Hospital Colorado, 13123 East 16th Avenue,
Aurora, CO 80045. E-mail: margaret.murphy-zane@childrenscolorado.
org.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002295
e138 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
were dysplastic based on FHC, most Graf IIa hips had normal or
borderline FHC per the AUC and may do well with observation
and follow-up ultrasound at 12 weeks old.
Level of Evidence: Level III—diagnostic study.
Key Words: DDH, developmental dysplasia of the hip, accepted
use criteria, AUC, Pavlik, Graf
(J Pediatr Orthop 2023;43:e138–e143)
D evelopmental dysplasia of the hip (DDH) is a com-
mon developmental orthopaedic problem that repre-
sents a spectrum from mild acetabular dysplasia to
dislocation. Early diagnosis and treatment of DDH can
prevent later functional limitations and osteoarthritis.1
Selective screening of newborns rather than universal
testing for hip dysplasia is currently recommended. Di-
agnostic ultrasound (US) is typically suggested for patients
with risk factors that include breech presentation, family
history and/or clinical hip instability, or other clinical
findings like asymmetric thigh folds, gluteal folds, limited
hip range of motion, or persistent hip click.2–10 Hip in-
stability at birth is common, noticed in about 1% to 2% of
infants. However, almost 90% of patients with mild in-
stability at birth stabilize spontaneously within the first 8
weeks of life.11 US abnormalities leading to intervention
have been reported in 5% to 7% of all newborns, though
more than 90% identified by US seem to resolve naturally
in early infancy.2,3
Hips on US are often described by 2 parameters, the
alpha angle and the percent femoral head coverage
(FHC). Graf characterized acetabular development with
the alpha angle (Table 1), which is the intersection of lines
drawn along the pelvis superior to the acetabulum and
along the bony roof of the acetabulum.12 The percent
FHC is the ratio of acetabular depth to diameter of the
femoral head (d/D).13,14
Normal hips have been characterized as alpha angle
≥ 60 degrees (Graf I) and FHC ≥ 50%.15–22 Our center has
historically used that definition for normal hips in the
protocols we have developed for the treatment of hip
dysplasia. In our institution’s DDH protocol, Pavlik har-
ness is started with hips that are Graf IIc or worse
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
TABLE 1. Graf Classification From Graf R
Graf Type* Alpha Angle
Graf I (normal) ≥ 60 degrees
Graf IIa 50-59 degrees, less than 12 wk old
Graf IIb 50-59 degrees, more than 12 wk old
Graf IIc 43-49 degrees, any age
Graf IId 43-40 degrees, decentering
Graf III < 43 degrees, eccentric hip
Graf IV < 43 degrees, everted labrum
Classification of hip joint dysplasia by means of sonography.
*Based on Graf.12
(alpha angle ≤ 49 degrees) or if they are clinically un-
stable, regardless of age. However, many infants less than
12 weeks old will have alpha angles from 50 to 59 degrees,
classified as Graf IIa. The majority of these will mature
without intervention by 12 weeks old, and thus in our
protocol are observed and recommended to follow-up
with US around 12 weeks of age. However, we observed
that sometimes practitioners in our group will choose to
treat Graf IIa hips with a Pavlik harness, particularly
when FHC is <50%
In light of the 2018 American Academy of Ortho-
paedic Surgeons “Appropriate Use Criteria (AUC) for the
Management of DDH in Infants up to 6 Months of Age,”
we hypothesized that our institution was overtreating Graf
IIa hips.23,24 The AUC stratifies hips into 3 groups: nor-
mal (alpha angle ≥ 60 degrees and FHC ≥ 45%), border-
line (alpha angle 50 to 59 degrees or FHC 35% to 44%), or
dysplastic (alpha angle <50 degrees or FHC <35%). Per
the AUC, all Graf IIa hips are at least borderline, but only
the hips with FHC <35% are truly dysplastic and should
be treated immediately. In addition, although our protocol
called for the immediate treatment of all infants found to
have hip instability, the AUC supports either immediate
or delayed (until 2 to 9 wk of age) brace treatment for
neonatal hips with a positive instability exam. It has been
noted that neonatal instability will often resolve without
intervention in the first 8 weeks of life.11
METHODS
After institutional review board approval, Interna-
tional Classification of Diseases and Current Procedural
Terminology codes were used to identify all patients who
were referred for the evaluation of DDH at a single pe-
diatric institution between 2009 and 2018. Included in this
study were hips that were classified as Graf IIa (n = 397
hips). Graf IIa hips were defined as having US alpha an-
gles of 50 to 59 degrees in infants less than 12 weeks old.
Patients were removed from the study if they had in-
complete documentation or treatment at an outside in-
stitution, an underlying syndrome, congenital birth defect,
or neurological or neuromuscular condition, leaving
371 Graf IIa hips (267 patients) in the study. Percent FHC
and acetabular alpha angle measurements were obtained
from the initial US evaluation and, when available,
from the 6-month follow-up US. On the basis of the
2018 AUC, hips were further categorized into the
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Application of AUC in Treatment of Infant Hip Dysplasia
following FHC severity categories: normal (FHC ≥ 45%),
borderline (FHC 35% to 44%), or dysplastic (FHC <35%).
Stable hips with FHC ≥ 35 degrees, part of the borderline
and normal strata, were grouped as stable, nondysplastic
Graf IIa hips (SND).
Demographics and clinical characteristics including
sex, age at presentation, Ortolani test result, Barlow test
result, family history of DDH, breech birth, age at ini-
tiation of treatment, age at initial US, and type of treat-
ment were collected from all hips. Hips were defined as
unstable if the Barlow and/or Ortolani test was positive.
Imaging parameters were collected from initial US, and
US and radiograph at 6 months of age. Acetabular index
(AI) values were measured on plain AP pelvis radiographs
obtained at 6 months of age, with residual DDH defined
as an AI value of > 30 degrees.
Statistical Methods
Descriptive statistics were used to summarize the
demographics and treatment characteristics among the IIa
hips included in the study population. To compare out-
comes among IIa stable hips that underwent Pavlik har-
ness treatment relative to IIa stable hips who did not
undergo Pavlik harness treatment, the analysis was limited
to patients who were between 5 and 9 months of age at
their 6-month radiograph visit (n = 161 hips). In hips that
were treated because of a more severe contralateral side,
only the Graf IIa hip was included in the analysis. The
more severe hip was not included, despite being the cause
of treatment, because it was not a Graf IIa. Among stable
hips included in the descriptive analysis (n = 270), hips that
underwent Rhino bracing rather than Pavlik harness
treatment (n = 3) and hips that did not return for a
6-month radiograph and US evaluation between 5 and
9 months of age (n = 105) were excluded from the treat-
ment analysis. Among stable hips, the impact of treatment
was compared within strata based on FHC, FHC ≥ 45%
(n = 112 hips) and FHC 35% to 44% (n = 50 hips). Gen-
eralized logistic regression models were used to compare
demographics and clinical characteristics among treated
hips versus untreated hips. Generalized estimating equa-
tions were used to account for correlation because of in-
clusion of multiple limbs per patient. Linear mixed model
regression analyses were used to compare AI values and
alpha angles at 6 months among hips that were treated
compared with hips that were not treated. Random in-
tercepts were used to account for correlation because of
the inclusion of multiple hips per patient. Finally, gener-
alized logistic regression models were used to estimate the
incidence of treatment failure, defined as progression to
Graf IIb at 90 days of age and/or presence of DDH (AI
value > 30 at 6 months) in treated versus untreated hips.
Because of the fact that this was a retrospective study, a
power analysis was not performed before the initiation of this
study, as we used all available data at the time of data
analysis. Furthermore, we did not perform a power analysis
after analyzing the data. Post hoc power analyses provide no
additional information beyond the information captured by
the P-value and 95% CI. Hoenig and Heisy25 as well Dorey26
www.pedorthopaedics.com | e139
Murphy-Zane et al
provide an excellent overview of the appropriateness of
power analyses in clinical research.
RESULTS
We identified n = 371 type IIa hips (n = 267 unique
patients) who met the inclusion criteria (Table 2). Overall,
250 Graf IIa hips were treated with the Pavlik harness. We
initially evaluated our results by first dividing by stability
(Fig. 1). Of 371 Graf IIa hips, 24% were clinically unstable.
Total 42% (37/89) of unstable Graf IIa hips were also
dysplastic (FHC <35%). Of stable Graf IIa hips, only 4%
(12/282) were dysplastic. We then evaluated the same
data set by first dividing by FHC ≥ 35% or FHC <35%
(dysplastic per the AUC) (Fig. 2). Thirteen percentage (49/
371) of the Graf IIa hips were considered dysplastic. Total
270 stable hips that had either normal FHC (group 1: FHC
> 45%) or borderline FHC (group 2: FHC 35% to 44%)
would have been initially observed per the AUC. Group 1
and group 2 combined comprised the SND group.
The reasons for treatment of the 158 SND Graf IIa
hips were evaluated (Table 3). Thirty-three percentage were
treated in the Pavlik harness because of a more severe
contralateral hip. These contralateral hips were not a Graf
IIa, so they were not included in this study. The associated,
less severe Graf IIa was included in this analysis. As this
phenomena occurs frequently in the treatment of these hips,
the authors feel that including these IIa hips is critical for a
comprehensive analysis of Graf IIa hips. This allowed us to
compare them to IIa hips that did not undergo treatment. If
we adhered to the AUC recommendations, we would have
not treated 106 hips, which equals a 67% reduction in the
number of SND Graf IIa hips treated.
To evaluate whether there would be poorer
outcomes if we did not immediately treat SND hips, we
TABLE 2. Demographic and Clinical Characteristics of 371
Graf IIa hips (267 Patients)
AUC dysplasia category
Borderline (alpha angle 50-60 and % femoral
head coverage ≥ 35%)
Dysplastic (% femoral head coverage <35%)
Female
Breech delivery
Family history
Ortolani positive
Barlow positive
Unstable
Bilateral
Treatment (within 16 wk of age)
Pavlik
Brace
Observation
Treatment after 90 d (IIb hip), n %
Age presentation [Days], median (range)
Age baseline US [days], median (range)
Femoral head coverage [%], median (range)
Alpha angle baseline, median (range)
AUC indicates appropriate use criteria; US, ultrasound.
e140 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
compared the acetabular indices on 6-month pelvis x-rays
in the treated and untreated SND hips (which both should
not have been treated per the AUC). Median duration of
treatment among all SND hips was 84 days. The differ-
ence in treatment length was not statistically significant
(P = 0.7170) between SND subgroups (Group 1 and
Group 2). Overall, after adjusting for age at the 6-month
visit, sex, and laterality (bilateral vs. unilateral DDH), AI
values at 6 months were higher in untreated hips relative
to treated hips (mean difference treatment vs. no treat-
ment: −1.5, 95% CI, −3.1 to 0.2, P = 0.0808). Alpha angle
values at 6 months were lower in untreated versus treated
hips (mean difference treatment vs. no treatment: 1.6, 95%
CI, −0.2 to 3.3, P = 0.0834).
After adjusting for age at the 6-month visit, sex, the
incidence of DDH was 8.3% (95% CI, 2.3% to 25.9%)
among treated hips compared with 10.5% (95% CI, 3.2%
to 29.0%) in untreated hips. Overall, the odds of treatment
failure (progression to IIb or AI > 30 at 6 months) among
treated hips were 0.5 times (95% CI, 0.2 to 1.4, P = 0.1917)
the odds of treatment failure among untreated hips.
DISCUSSION
Graf IIa hips (alpha angle 50 to 59 degrees in infants
less than 12 wk old) can often be observed with the ex-
pectation that they will resolve with normal development.
Using the 2018 American Academy of Orthopedic Sur-
geons AUC for DDH, 73% (270/371) of the Graf IIa hips
studied were SND and may do well with observation and
follow-up US around 12 weeks old. Among stable hips in-
cluded in this analysis (n = 270), only 6 hips had not
normalized by the time they returned at 12 week/90 day
follow-up—progressing to a IIb—and underwent bracing.
In applying the stratified FHC parameters outlined in the
AUC, we would reduce the number of SND Graf IIa hips
we treat by 67%. However, this does not mean that no Graf
IIa hips need to be treated. Importantly, this study also
Median, found that 24% of Graf IIa hips were clinically unstable.
Range N (%) Our institution and the AUC agree that unstable hips
should be treated, especially when instability persists past
322 (87) the perinatal period. All of the unstable Graf IIa hips in this
study were treated. Thirteen percentage of the Graf IIa hips
49 (13) were dysplastic (FHC <35%) and were treated 96% (47/49)
294 (79) of the time, in line with American Academy of Orthopedic
147 (40)
61 (16) Surgeons AUC recommendations. Total 27% of Graf IIa
24 (6) hips (101/371) should be treated, either for FHC <35%
62 (17) (N = 12) or instability (N = 52) or both (n = 37).
89 (24) Although our institution has an established protocol
280 (75)
for infant hip dysplasia, we found that there was consid-
250 (67) erable practice variability throughout our large system.
6 (2) Our institution’s protocol differs from the AUC in that we
115 (31) do not specify diagnostic ranges for FHC. Hips with FHC
6 (2) below 50% are considered abnormal (not borderline), with
16 (2-89)
22 (0-84) normal defined as alpha angle ≥ 60 degrees and FHC ≥
47 (1.8-67.7) 50%. Anything below those benchmarks is considered
55 (50-59.9) abnormal per our institution. Therefore, there are some
hips with FHC between 45% and 50% that are considered
“normal” per the AUC that received treatment at our
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Application of AUC in Treatment of Infant Hip Dysplasia

FIGURE 1. Evaluation first by stability. Graf IIa hips were divided first by clinical stability versus instability. All unstable hips were
treated with a Pavlik harness. Stable hips were then divided by FHC: FHC ≥ 35% (borderline or normal per the developmental
dysplasia of the hip appropriate use criteria ) and FHC <35% (dysplastic). Only 4% of stable Graf IIa hips were considered
dysplastic. FHC indicates femoral head coverage.

institution. Although there seemed to be consistency with
regard to Graf IIa hips and the Graf classification, this
study identified uncertainty surrounding the threshold for
concern with regard to FHC at our institution, causing
differences in treatment between providers. Some pro-
viders chose to treat “borderline” hips with FHC between
35% and 50%, where as some practitioners chose to repeat
US and exam after a period of observation for these
borderline hips. Providers tended to be consistent within
their chosen method of treatment.
In addition to this inconsistency, we saw many pa-
tients with risk factors like breech presentation, thus for

FIGURE 2. Evaluation first by FHC. Graf IIa hips were first divided by FHC: FHC ≥ 35% or FHC <35% (dysplastic per the appropriate
use criteria ). Thirteen percentage of the Graf IIa hips were considered dysplastic. Total 164 treated stable hips that had either
normal or borderline FHC (group 1 and group 2 combined) would not have been treated per the appropriate use criteria. FHC
indicates femoral head coverage.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e141

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Murphy-Zane et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 3. Reason for Treatment in Stable Nondysplastic Graf TABLE 4. Comparison at Presentation of Treated and
IIa Hips (SND): 33% Were Treated in the Pavlik Because of a Untreated Stable Nondysplastic (SND) Graf Type IIa Hips That
More Severe Contralateral Hip. Returned for Evaluation at 6 Months of Age.
N (%) Treatment No Treatment
Group 1: IIa + stable + ≥ 45% FHC + treated (n = 101) Mean (Range), Mean (Range),
Contralateral hip 35 (34.7) N (%) N (%) P*
Provider preference 66 (65.3)
Group 2: IIa + stable + 35%-44% FHC + treated (n = 57) Group 1: stable FHC ≥ 45% | treatment n = 73, no treatment n = 39
Contralateral hip 17 (29.8) Female, n (%) 62 (84.9) 32 (82.1) 0.9697
Provider preference 40 (70.2) Breech, n (%) 30 (41.1) 23 (59.0) 0.1741
Combined: stable and ≥ 35% FHC + treated (n = 158) Family history, n (%) 12 (16.4) 11 (28.2) 0.3209
Contralateral hip 52 (32.9) Bilateral, n (%) 63 (86.3) 24 (61.5) 0.0005
Provider preference 106 (67.1) Age at initial 31.6 (23.9) 39.5 (24.9) 0.1165
presentation [days],
If we adhered to the AUC recommendations, we would have not treated 106 mean (SD)
hips, which equals a 67% reduction in the number of stable nondysplastic Graf IIa Age at initial US [days], 32.1 (19.6) 32.8 (16.7) 0.5588
hips treated. mean (SD)
FHC indicates femoral head coverage; US, ultrasound. Percent femoral head 51.6 (4.9) 51.6 (4.1) 0.7346
coverage, mean (SD)
Alpha angle, mean (SD) 55.5 (2.6) 55.9 (2.8) 0.2536
this study we were able to have sufficiently sized groups of Group 2: stable FHC 35%-44% | treatment n = 42, no treatment n = 8
SND Graf IIa hips that had been either treated or not Female, n (%) 35 (83.3) 7 (87.5) 0.8533
Breech, n (%) 21 (50.0) 2 (25.0) 0.4553
treated with a Pavlik harness. These groups allow us to Family history, n (%) 6 (14.3) 2 (25.0) 0.4874
compare hip radiographs and USs at 6 months of age. Bilateral, n (%) 38 (90.5) 6 (75.0) 0.2536
Among the 162 hips that returned for a 6-month US, 47 Age at initial 22.9 (18.9) 19.8 (13.4) 0.7573
SND Graf IIa hips were untreated. Thirteen of them had presentation [days],
mean (SD)
no risk factors for DDH and normalized on US before Age at initial US [days], 24.7 (14.5) 19.9 (12.8) 0.6405
6 months old. The remaining 34 hips were either breech mean (SD)
presentation and/or had significant family history of DDH Percent femoral head 40.8 (2.6) 40.7 (2.4) 0.9606
and, per our institution’s protocol, returned for evaluation coverage, mean (SD)
at 6 months old with an AP pelvis x-ray and/or a hip US. Alpha angle, mean (SD) 53.5 (2.4) 54.5 (2.8) 0.2103
Combined: treatment n = 115, no treatment n = 47
Risk factors and presenting alpha angles were similar Female, n (%) 97 (84.3) 39.0 (83.0) 0.9320
between the groups (Table 4). Breech, n (%) 51 (44.3) 25.0 (53.2) 0.3055
At 6 months, there was no difference between hips Family history, n (%) 18 (15.7) 13.0 (27.7) 0.3997
treated with a Pavlik harness versus untreated hips Bilateral, n (%) 101 (87.8) 30.0 (63.8) < 0.0001
Age at initial 28.5 (22.5) 36.1 (24.4) 0.0853
(Table 5). Among SND hips, the pairwise difference in AI presentation [days],
between treated versus untreated hips −1.5 (95% CI, −3.1 to mean (SD)
0.2), indicating that on average, AI values among treated Age at Initial US [days], 29.4 (18.2) 30.6 (16.7) 0.4169
hips were 1.5 degrees lower compared with untreated hips mean (SD)
(95% CI, 3.1 degrees lower to 0.2 degrees higher). On the Percent femoral head 47.7 (6.7) 49.7 (5.6) 0.4153
coverage, mean (SD)
basis of the lower bound of the 95% CI, we can exclude the Alpha angle, mean (SD) 54.8 (2.7) 55.6 (2.8) 0.1555
possibility that treatment achieves greater than a 3.1-degree
improvement in AI relative to no treatment. This effect is *From GEE model accounting for clustering of multiple hips within patients.
Group 1 had FHC ≥ 45%, considered normal per the AUC. Group 2 had FHC
less than the intraobserver measurement error for AI (3.6, 35% to 44%, considered borderline and should be observed per the AUC.
Spatz et al27) and thus, unlikely to change clinical practice.
Follow-up showed no clinically or statistically significant
difference between stable Graf IIa hips with FHC ≥ 35% 59 degrees after 12 wk of age) and even Graf IIc (alpha
(SND hips) regardless of treatment with a Pavlik harness angle 43 to 49 degrees, any age) hips do not need treat-
(Table 5). Our sample size of untreated borderline hips ment, though they do not specifically address FHC.28 Our
(FHC 35 to 45 degrees) was small (n = 8), and therefore practice was so consistent in treating unstable hips with a
insufficient to detect differences between treated and Pavlik harness, regardless of age, that this study is unable
untreated borderline hips. to comment on whether unstable hips do well without
In regard to hip instability, our institution’s protocol treatment. At minimum, unstable Graf IIa hips need
called for the immediate treatment of all infants found to careful consideration and follow-up.
have unstable hips. However, the AUC supports both Protocols and AUCs are established to improve
immediate and delayed (2 to 9 wk) brace treatment for patient care, and periodic review is important. In stepping
neonatal hips with a positive instability exam (positive back to see how well our institution was aligned with an
Barlow or positive Ortolani), given that neonatal in- established AUC, we were able to identify that better
stability will often resolve without intervention in the first defining what constitutes pathologic FHC would be
8 weeks of life. There has been a recent small prospective helpful. This may reduce the number of SND Graf IIa
study that suggests that stable Graf IIb (alpha angle 50 to hips we treat with a Pavlik harness by 67%. Because we

e142 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Application of AUC in Treatment of Infant Hip Dysplasia

should all dislocated or dislocatable hips be treated? J Pediatr

TABLE 5. Radiograph and Ultrasound Outcomes Among Orthop. 2002;22:613–621.
Treated Versus Untreated Hips by Femoral Head Categories 10. Groarke PJ, McLoughlin L, Whitla L, et al. Retrospective multi-
Treatment No Treatment center analysis of the accuracy of clinical examination by community
physicians in diagnosing developmental dysplasia of the hip.
Mean* LCL UCL Mean* LCL UCL J Pediatr. 2017;181:163–166.e161.
Group 1: stable FHC ≥ 45% | treatment n = 73, no treatment n = 39 11. Alves C, Truong WH, Thompson MV, et al. Diagnostic and
Acetabular 26.2 24.6 27.7 27.7 26.0 29.3 treatment preferences for developmental dysplasia of the hip:
index a survey of EPOS and POSNA members. J Child Orthop. 2018;12:
Alpha angle 25.4 22.6 28.3 25.1 20.2 29.9 236–244.
Group 2: stable FHC 35-44% | treatment n = 42, no treatment n = 8 12. Graf R. Classification of hip joint dysplasia by means of sonography.
Acetabular 72.1 70.5 73.7 70.3 68.4 72.1 Arch Orthop Trauma Surg (1978). 1984;102:248–255.
index 13. Morin C, Harcke HT, MacEwen GD. The infant hip: real-time
Alpha angle 71.2 68.4 74.0 69.7 65.6 73.7 US assessment of acetabular development. Radiology. 1985;157:
Combined: treatment n = 115, no treatment n = 47 673–677.
Acetabular 25.9 24.6 27.1 27.3 25.8 28.8 14. Morin C, Zouaoui S, Delvalle-Fayada A, et al. Ultrasound assess-
index ment of the acetabulum in the infant hip. Acta Orthop Belg. 1999;65:
Alpha angle 71.8 70.5 73.2 70.3 68.6 72.0 261–265.
15. Terjesen T, Bredland T, Berg V. Ultrasound for hip assessment in the
*Adjusted for age at the 6-month visit, sex, and laterality (bilateral vs. unilat- newborn. J Bone Joint Surg Br. 1989;71:767–773.
eral DDH. 16. Rosendahl K, Toma P. Ultrasound in the diagnosis of developmental
FHC indicates femoral head coverage, LCL; Lower 95% Confidence Interval, dysplasia of the hip in newborns. The European approach. A review
UCL; Upper 95% Confidence Interval. of methods, accuracy and clinical validity. Eur Radiol. 2007;17:
1960–1967.
17. Patel H. Preventive health care, 2001 update: screening and
did have considerable practice variation, we were able to management of developmental dysplasia of the hip in newborns.
show that stable Graf IIa hips with FHC ≥ 35 degrees do Cmaj. 2001;164:1669–1677.
not benefit from Pavlik harness treatment. Most Graf IIa 18. Harcke HT, Pruszczynski B. Hip ultrasound for developmental
hips are considered borderline by the AUC criteria, and dysplasia: the 50% rule. Pediatr Radiol. 2017;47:817–821.
19. Chen HW, Chang CH, Tsai ST, et al. Natural progression of
should be initially observed rather than braced. However, hip dysplasia in newborns: a reflection of hip ultrasonographic
this study cautions that 24% of Graf IIa hips are clinically screenings in newborn nurseries. J Pediatr Orthop B. 2010;19:
unstable and 13% have FHC <35% (dysplastic per the 418–423.
AUC), both warranting more immediate treatment. 20. Castelein RM, Sauter AJ, de Vlieger M, et al. Natural history
of ultrasound hip abnormalities in clinically normal newborns.
J Pediatr Orthop. 1992;12:423–427.
REFERENCES 21. Terjesen T, Holen KJ, Tegnander A. Hip abnormalities detected by
1. Murphy SB, Ganz R, Müller ME. The prognosis in untreated ultrasound in clinically normal newborn infants. J Bone Joint Surg
dysplasia of the hip. A study of radiographic factors that predict the Br. 1996;78:636–640.
outcome. JBJS. 1995;77:985–989. 22. Gunay C, Atalar H, Dogruel H, et al. Correlation of femoral head
2. Engesæter IØ, Lie SA, Lehmann TG, et al. Neonatal hip instability coverage and Graf α angle in infants being screened for devel-
and risk of total hip replacement in young adulthood. Acta opmental dysplasia of the hip. Int Orthop. 2009;33:761–764.
Orthopaedica. 2008;79:321–326. 23. Samora J, Quinn RH, Murray J, et al. Management of devel-
3. Barlow TG. Early diagnosis and treatment of congenital dislocation opmental dysplasia of the hip in infants up to six months of age:
of the hip. Proc R Soc Med. 1963;56:804–806. intended for use by orthopaedic specialists. J Am Acad Orthop Surg.
4. Rosendahl K, Markestad T, Lie RT. Congenital dislocation of the 2019;27:e360–e363.
hip: a prospective study comparing ultrasound and clinical exami- 24. Mulpuri K, Song KM. AAOS clinical practice guideline: detection
nation. Acta Paediatr. 1992;81:177–181. and nonoperative management of pediatric developmental dysplasia
5. Myers J, Hadlow S, Lynskey T. The effectiveness of a programme for of the hip in infants up to six months of age. J Am Acad Orthop Surg.
neonatal hip screening over a period of 40 years: a follow-up of the 2015;23:206–207.
New Plymouth experience. J Bone Joint Surg Br. 2009;91:245–248. 25. Hoenig JM, Heisey DM. The Abuse of Power: The Pervasive Fallacy
6. Wenger D, Düppe H, Tiderius CJ. Acetabular dysplasia at the age of of Power Calculations for Data Analysis. Am Stat. 2001;55:6.
1 year in children with neonatal instability of the hip. Acta Orthop. 26. Dorey FJ. In Brief: Statistics in Brief: Statistical Power: What Is It
2013;84:483–488. and When Should It Be Used? Clin Orthop Relat Res. 2011;469:
7. Vaquero-Picado A, González-Morán G, Garay EG, et al. Devel- 619–620.
opmental dysplasia of the hip: update of management. EFORT Open 27. Spatz DK, Reiger M, Klaumann M, et al. Measurement of
Rev. 2019;4:548–556. acetabular index intraobserver and interobserver variation. J Pediatr
8. Barlow TG. Early diagnosis and treatment of congenital dislocation Orthop. 1997;17:174–175.
of the hip. J Bone Joint Surg . 1962;44-B:292–3012. 28. Pollet V, Castelein RM, van de Sande M, et al. Abduction treatment
9. Lorente Moltó FJ, Gregori AM, Casas LM, et al. Three-year in stable hip dysplasia does not alter the acetabular growth: results of
prospective study of developmental dysplasia of the hip at birth: a randomized clinical trial. Sci Rep. 2020;10:9647.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e143

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

 ORIGINAL ARTICLE
Prediction of Late Dysplasia Based on Ultrasound and Plain
X-Ray at 6 Months
Reba L. Salton, BS,* Patrick M. Carry, PhD,*† Nancy Hadley-Miller, MD,†‡
Margaret Siobhan Murphy-Zane, MD,†‡ Christopher Brazell, DO,* Eduardo Novais, MD,§∥
and Gaia Georgopoulos, MD†‡
Background: Developmental dysplasia of the hip represents a
spectrum of deformity. Residual dysplasia at 2 years of age is
associated with an increased risk for osteoarthritis and functional
limitations. We compared the prognostic value of 6-month
imaging modalities and aimed to identify optimal diagnostic
metrics for the prediction of residual dysplasia.
Methods: After IRB approval, patients who underwent Pavlik
treatment between 2009 and 2018 with 2-year follow-up were
identified. Sonographs [ultrasound (US)] and radiographs (x-ray)
were obtained at 6-month and 2-year-old visits. Dysplasia at
2 years was defined as an acetabular index (AI) > 24 degrees.
Receiver operating characteristic curves were constructed to
quantitatively compare the prognostic ability of US and x-ray-
based measures at 6 months. Youden’s index [(YI) (values range
from 0 (poor test) to 1 (perfect test)] was used to evaluate existing
cutoffs at 6 months of age (normal measurements: alpha angle
(AA) ≥ 60 degrees, femoral head coverage (FHC) ≥ 50%, and
AI <30 degrees) relative to newly proposed limits.
Results: Fifty-nine patients were included, of which 28.8% of
patients (95% CI: 17.3 to 40.4%) had acetabular dysplasia at
2 years. After adjusting for sex, AA [Area under the Curve
(AUC): 80] and AI (AUC: 79) at 6 months of age were better
tests than FHC (AUC: 0.77). Current diagnostic cutoffs for AA
(YI: 0.08), AI (YI: 0.0), and FHC (YI: 0.06) at 6 months had
poor ability to predict dysplasia at 2 years. A composite test of
all measures based on proposed cutoffs (AA ≥ 73 degrees, FHC
> 62% and AI ≤ 24 degrees) was a better predictor of dysplasia
at 2 years (Youden’s index (YI): 0.63) than any single metric.
Conclusions: The rate of residual dysplasia remains concerning.
The 6-month x-ray and US both play a role in the ongoing
management of the developmental dysplasia of the hip. The
prediction of dysplasia is maximized when all metrics are con-
sidered collectively. Existing parameters were not accurate; We
From the *Musculoskeletal Research Center; ‡Orthopedics Institute,
Children’s Hospital Colorado; †Department of Orthopedics, Uni-
versity of Colorado School of Medicine, Anschutz Medical Campus,
Aurora, CO; §Department of Orthopaedic Surgery, Harvard Medical
School; and ∥Department of Orthopedics, Boston Children’s Hospi-
tal, Boston, MA.
The authors declare no conflicts of interest.
Reprints: Gaia Georgopoulos, MD, Orthopedics Institute, Children’s
Hospital Colorado, 13123 East 17th Avenue, B060, Aurora, CO
80045. E-mail: Gaia.Georgopoulos@childrenscolorado.org.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002301
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
recommend the following cutoffs: AA ≥ 73 degrees, FHC >
62%, and AI ≤ 24 degrees. These cutoffs must be validated.
Level of Evidence: Prognostic Level II.
Key Words: DDH, developmental dysplasia of the hip, prog-
nosis, late dysplasia, 6-month ultrasound, 6-month radiograph
(J Pediatr Orthop 2023;43:99–104)
D evelopmental dysplasia of the hip (DDH) describes a
wide spectrum of pediatric disorders that are charac-
terized by a disruption in the normal relationship between
the femoral head and the acetabulum. On the basis of the
physical exam alone, the reported incidence of DDH per
1000 births is estimated at 8.6 when a pediatrician per-
forms the exam versus 11.5 when an orthopedist performs
the exam.1 Because dysplastic hips may have a normal
clinical exam, screening with ultrasound (US) has been
recommended for infants with risk factors.2 When the US
is utilized as a screening tool, the incidence of DDH is
estimated at 25 per 1000 births.1 This imaging technique is
helpful in ensuring accurate diagnoses and proper reduc-
tion of the hip.3 Early diagnosis and accurate treatment are
associated with the highest rate of successful outcomes in
infants with DDH, with the goal of conservative treatment
being to maintain a concentric, stable reduction of the hip
and prevent long-term complications.4–6
Because DDH is developmental in nature, predicting
recurrent or late-presenting dysplasia can be difficult.5 The
incidence of acetabular dysplasia at 2 years of age following
successful treatment is cited between 3% to 33%.7–12 Radio-
graphs are also used to monitor the hips of patients during
treatment but are of little value during the first 6 months of
life when the femoral head and labrum are primarily carti-
laginous and cannot be visualized on the radiograph.2 At
6 months of age, radiographs and USs are often used in
tandem for evaluation and ongoing diagnosis as the femoral
heads begin to ossify.2,3 Studies at this age are done to de-
termine if prior treatment has been successful and if addi-
tional treatment is needed. They have not historically been
used to predict late dysplasia. From a resource management
perspective, it would be beneficial to determine if only 1 or
both of these imaging modalities are capable of accurate
prognosis. This study aimed to compare both imaging tech-
niques and their ability to predict recurrent dysplasia. We
www.pedorthopaedics.com | 99
Salton et al
hypothesized the 6-month US is a better predictor of re-
current dysplasia at 2 years compared with plain x-rays due
to the visibility of the hip cartilage on US.
METHODS
After IRB approval, ICD-9 and 10 codes were used
to identify all patients with a primary diagnosis of idio-
pathic DDH who underwent successful Pavlik treatment
between 2009 and 2018, were followed for a minimum of 2
consecutive years and did not require additional treatment
at 6 months of age. Patients were excluded from any
treatment or follow-up at another institution or an un-
derlying neurological/teratologic condition. A successful
Pavlik treatment course was defined as a normal exam and
US after 12 weeks of Pavlik bracing and normal imaging
at 6 months of age. Normal imaging was dictated by alpha
angle (AA) > 60 degrees and femoral head coverage
greater than 50% on US, and acetabular index (AI) less
than 30 degrees on plain film at 6 months. Subject
enrollment is described in Figure 1.
US and radiographic images were obtained from all
subjects at their 6-month-old clinic visit. Measurements at
6 months (range: 4.5 to 7.5 mo) included alpha angle and
percent femoral head coverage from the US and acetabular
index measured on plain film. Congruency of Shenton’s line
and acetabular index measures were also obtained at their
2-year-old clinic visit (range: 20 to 28 mo). All measurements
were performed by the senior author to limit interobserver
variability. AA and AI are reliable measures that are routinely
used in clinical practice. Previous studies have reported mod-
erate to high levels of inter-rater and intra-rater reliability for
AA (intraclass correlation coefficient (ICC): 0.72) and AI
FIGURE 1. Enrollment. Patient enrollment numbers and exclusion
criteria.
100 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
(ICC: 0.9).13–15 The presence of late acetabular dysplasia was
defined as an acetabular index value greater than 24 degrees at
2 years, per our institution’s treatment protocol for DDH.16
The predictive value of currently accepted cutoffs for
defining a dysplastic hip at the 6-month visit was eval-
uated relative to the definitive AI measurement obtained
at the 2-year follow-up visit, with an AI > 24 degrees in-
dicating a dysplastic hip at that time point. At the 6-month
visit, a normal hip presents with the following measure-
ments on imaging: an alpha angle ≥ 60 degrees, percent
femoral head coverage ≥ 50%, and an AI <30 degrees.
Statistical Methods
Descriptive statistics were used to characterize all
subjects included in the study cohort. The sensitivity,
specificity, and area under the curve (AUC) values for
existing cutoff values at the 6-month visit were estimated
based on the presence of acetabular dysplasia at 2 years
(AI > 24). Among patients with bilateral dysplasia, 1 hip
was randomly selected for inclusion in the analysis. This
methodology was used to avoid underestimating varia-
bility in radiographic and ultrasonographic variables due
to the clustering of hips within subjects. Receiver operat-
ing characteristic curves (ROC) were constructed for all
parameters for the purpose of exploring alternative cutoff
values for predicting the presence of dysplasia at 2 years.
Leave one out cross-validation was used to assess the
stability of the predictive models.
RESULTS
Study Population
The final study population consisted of 59 patients
with developmental dysplasia of the hip who underwent
successful treatment. The median duration of Pavlik
bracing was 2.8 months (interquartile range: 2.4 to
2.9 mo). The most common Graf classification at the in-
itiation of treatment was IIC hips (50.8%, 30/59), followed
by IIA (40.7%, 25/59), III (5.1%, 3/59), IIB (1.7%, 1/59),
and IV (1.7%, 1/59). The mean age of the subjects at the
6-month US, 6-month x-ray, and 2-year x-ray visits was
0.5 years ( ± 0.03), 0.5 years ( ± 0.04), and 2.1 years
( ± 0.11), respectively.
Predictors of Acetabular Dysplasia at Two Years
The prevalence of acetabular dysplasia in the pop-
ulation was determined based on the acetabular index
values at the 2-year visit. The prevalence of dysplasia at
24 months was 28.8% (95% CI: 17.3 to 40.4%). The dis-
tribution of the radiographic parameters and hip dysplasia
risk factors is outlined in Table 1.
We used receiver operating characteristic curves to
compare the predictive ability of US-based parameters at
6 months to radiographic parameters at 6 months. After
adjusting for sex, the AA at 6 months and AI were asso-
ciated with the highest AUC values (Table 2). Leave one
out cross-validation was used to assess the stability of the
predictive models. As expected, given the relatively small
sample size, the AUC values for the cross-validation
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
TABLE 1. Comparison of Hips Classified as Dysplastic Versus
Nondysplastic at Two Years
Dysplastic N = 17 Nondysplastic N = 42
N | Mean (% | SD) N | Mean (% | SD)
Bilateral, n (%) 17 (100.0) 42 (100.0)
Family History, n (%) 3 (17.6) 5 (11.9)
Female, n (%) 12 (70.6) 37 (88.1)
Unstable*, n (%) 9 (52.9) 21 (50.0)
Acetabular Index, 29.6 (3.4) 26.0 (3.9)
mean (SD)
Alpha Angle, mean 68.6 (2.8) 73.2 (4.7)
(SD)
Percent Femoral 59.8 (6.7) 65.1 (5.8)
Coverage, mean (SD)
*Unstable defined as positive Barlow/Ortolani maneuver on clinical exam or
instability noted on ultrasound.
models were lower than the original observed models.
However, 95% confidence for all models excluded 0.50,
indicating the models are significantly better than random
chance at predicting dysplasia at 2 years.
Alternative Cutoffs for Ultrasound Parameters
ROC plots (Figs. 2–4) were used to determine
alternative cutoffs for the US and radiographic
parameters at the 6-month visit. Optimal cutoffs for each
parameter were determined based on Youden’s index
(values range from 0, the test is worthless, to 1, the test is
perfect). Sensitivity, specificity, and Youden’s index for the
alternative cutoffs were compared with existing values. New
cutoffs for US and radiographic parameters are as follows:
AA ≥ 73 degrees, FHC ≥ 62% and AI ≤ 24 degrees. These
values are also presented in Table 3. On the basis of
Youden’s index metric, a composite test based on all 3
measures represents the best test for differentiating between
hips that will versus will not develop acetabular dysplasia at
2 years (see the composite test in Table 3).
DISCUSSION
This study compares the ability of the 6-month US and
radiograph to predict acetabular dysplasia at 2 years of age
after the successful Pavlik harness treatment for Devel-
opmental Dysplasia of the Hip (DDH). Current diagnostic
metrics for AA, percentage femoral head coverage (FHC), AI
were assessed and deemed to be inadequate for predicting late
acetabular dysplasia at the 2-year follow-up for patients who
did not require subsequent treatment at 6 months. AA on US
was identified to be the test with the strongest prognostic value
at the 6-month follow-up, as indicated by our AUC values.
Not surprisingly, patients who were ultimately not
dysplastic (n = 42) at 2 years of age (24 mo) had notably
higher AA and FHC values, and lower AI values on
6-month images relative to those who had recurrent ace-
tabular dysplasia (n = 17) noted at their 2-year follow-up. At
6-month follow-up, this nondysplastic group had average
values of: AI = 26.0, AA = 73.2 degrees, and FHC = 65.1%
on US and radiograph. These average values are similar to
our data-driven 6-month diagnostic cutoffs (Table 3).
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Prognostic Value of 6mo. Images for Late Dysplasia
Interestingly, the distribution of patients with female sex
(∼80% of patients), family history (∼14% of patients), or
instability (∼50% of patients) was relatively similar between
both the nondysplastic and dysplastic hips.
All currently utilized 6-month cut-off values were
identified to be insufficient for predicting recurrent dys-
plasia in this population and, when used independently,
result in high rates of either over or undertreatment. AA
has a very high specificity (100%) with very low sensitivity
(0.0%), risking the undertreatment of patients. FHC and
AI present the opposite problem. These measures have an
extremely high sensitivity (100%) and will correctly iden-
tify every patient that will go on to develop acetabular
dysplasia at the risk of a very high rate of unnecessary
treatment. Based upon the sensitivity and specificity of
each test, ROC plots were constructed, and Youden’s in-
dex was used to identify more appropriate diagnostic
cutoffs. Employing these new cutoffs in clinical practice
may help to accurately distinguish the patients who will
and will not go on to develop late dysplasia.
At 6 months of age, these optimal values were identi-
fied as follows: AI < 24 degrees, AA > 73 degrees, and FHC
> 62%. Sensitivity and specificity values for these cutoffs are
presented in Table 3. A combination of all diagnostic tests at
this study’s newly designated values proved to be the best
predictor of residual dysplasia at 2 years. Relative to the
values currently used in clinical practice at 6 months of age
(AI <30, AA > 60 degrees, FHC > 50%), the new values are
more restrictive and may have drastic implications in the
treatment of patients currently considered to have normal
hips at this time point.
While screening with both US and x-ray at 6 months of
age is the ideal approach, we acknowledge this may not always
be realistic. Utilization of both imaging techniques increases the
cost of care. Plain films result in direct radiation exposure to the
patient, but they also tend to have a lower out-of-pocket ex-
pense to the patient than the point-of-care US. Plain films offer
the benefit of being able to observe the ossification of the
femoral heads, but US provides better visibility of hip cartilage.
In rare cases, plain films may be able to identify a subluxed or
dislocated hip that was missed on US.17 One study only con-
sidered hips that screened positive on both types of imaging as
truly dysplastic.18 At our institution, USs require advanced
scheduling and are technician-dependent, whereas radiographs
can be performed point-of-care. At many institutions, the
treating MD or DO obtains US images. Our institution rou-
tinely gets both images at 6 months of age. Generally speaking,
it is up to the discretion of the treating provider and the patient
as to what is clinically relevant and in the best interests of the
patient and the patient’s family. Both imaging techniques per-
form well in practice and are clinically valuable tools for the
evaluation and prediction of DDH at 6 months of age.
Whenever possible, assessment using both methods may offer
the greatest predictive value of dysplasia at 2 years of age.
Regardless of whether 1 or 2 imaging techniques are used at
6 months of age, the current definition of what constitutes a
‘normal’ hip at this age needs to be examined, as existing
cutoffs are not effective at differentiating between hips that are
high versus low risk for developing dysplasia at 2 years.
www.pedorthopaedics.com | 101
Salton et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 2. Predictive value of Ultrasound and Radiographic

Parameters at 6 Months
Leave One Out
Original Sample Cross-Validation
Lower Upper
Variable AUC* (95% CI) (95% CI) AUC* LCL UCL
Acetabular Index 0.79 0.67 0.92 0.76 0.62 0.89
6 mo
Alpha Angle 0.80 0.69 0.92 0.75 0.62 0.88
Percent Femoral 0.77 0.63 0.91 0.70 0.54 0.86
Coverage
*AUC indicates area under the curve

There is extensive literature discussing the strengths of

both US and plain films for the diagnosis of DDH, as well as
the incidence of late-presenting dysplasia in various
cohorts.11,12,19–21 Barrera et al22 discuss the various imaging
techniques that are utilized over the course of treatment for
DDH. They describe the strengths of US for discerning true
dysplastic hips from those who merely have a concerning
clinical exam or risk factors.22 The addition of US before
4 months of age has been shown to minimize delays in diag-
nosis and rates of the surgical intervention relative to only
clinical examination.23 The authors go on to acknowledge that
after the secondary ossification centers begin to form, plain
x-ray is also an appropriate diagnostic tool for visualizing hip
subluxation or full dislocation.22 Barrera and his co-authors
support the idea that both US and radiographs are beneficial
FIGURE 2. ROC Curve for Acetabular Index. Receiver operating
characteristic curve constructed for acetabular index based on
radiograph at 6 months. The true positive rate (sensitivity) is
plotted over the false positive rate (1-specificity), and illustrates the
performance (in this case, diagnostic ability) of a binary system/
classifier (in this case, over or under our threshold values).
FIGURE 3. ROC Curve for Alpha Angle. Receiver operating
characteristic curve constructed for alpha angle based on ultra-
sound at 6 months. The true positive rate (sensitivity) is plotted
over the false positive rate (1-specificity), and illustrates the per-
formance (in this case, diagnostic ability) of a binary system/
classifier (in this case, over or under our threshold values).
for patient evaluation. When used at the appropriate time
points, these isolated imaging techniques are able to provide a
solid appraisal of hip morphology for diagnostic purposes.
Mulrain and colleagues (2021) studied the reliability of
initial US screening due to risk factors for DDH, such as a
breech presentation or family history, to predict acetabular
dysplasia at 6 months. Their results indicate that about 8% of
patients who screen normally (Graf I or Graf IIa that nor-
malizes) on an initial US have abnormal findings on the ra-
diograph at 6 months.17 This rate of late diagnosis implies the
need for further screening at 6 months for the individuals who
are referred for risk factors, regardless of a negative US, early
in life. Mulrain and colleagues emphasize the need for con-
tinued assessment of hips with DDH at and after 6 months of
age, utilizing both the US and radiograph. Neither Barrera nor
Mulrain discusses the prognostic value of these imaging mo-
dalities in the same manner that the current study attempts to
describe. Both studies do support the need for ongoing eval-
uation and the importance of both US and radiograph as tools
for diagnosis. In our study, 28.8% of patients had recurrent
dysplasia at 2 years of age, based upon an acetabular index
greater than 24 degrees, after successful Pavlik harness treat-
ment and no subsequent bracing. To our knowledge, this is the
first study attempting to use diagnostic metrics to address the
predictive value of US and radiograph at 6 months of age.
In a study done by Kitay et al18 in 2018, US
was deemed to be a more appropriate imaging technique
for diagnosis relative to the 6-month x-ray. Among 31
patients ranging from 5 to 7 months, US proved more

102 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
FIGURE 4. ROC Curve for Femoral Head Coverage. Receiver op-
erating characteristic curve constructed for femoral head coverage
based on ultrasound at 6 months. The true positive rate (sensitivity)
is plotted over the false positive rate (1-specificity), and illustrates
the performance (in this case, diagnostic ability) of a binary system/
classifier (in this case, over or under our threshold values).
sensitive than x-ray, identifying 13 abnormalities com-
pared with 4 abnormalities on plain film. Further, no
patient who screened negative in the US required addi-
tional treatment during the subsequent 26-month (on
average) follow-up. This would support the use of US as
the primary method of screening up to and including
6 months of age. In 1 prospective clinical trial, seventy-
four 4 to 6-month-old infants were identified as abnormal
in the US among a cohort of 1430 patients.24 These
abnormal hips were then secondarily screened with x-ray;
only 44 were also identified as dysplastic or dislocated on
the radiograph. The authors found a statistically sig-
nificant difference (P < 0.01) in the incidence of DDH in
US (51.75:1000 hips) versus radiograph (30.77:1000 hips)
between 4 and 6 months of age but interpreted only those
hips who screened positive on both images as ‘true’
dysplasia.24 Hips that were positive only on sonograph
TABLE 3. Alternative Cutoff Values for Acetabular Index Alpha Angle, and Percent Femoral Coverage at Six Months of Age
Current Cutoffs
Cutoff Sens (%) Spec (%)
Acetabular Index > 30 90.5 17.6
Alpha Angle ≤ 60 0.0 100.0
Percent Femoral Coverage < 50 100.0 5.9
Composite Test* All + 0.0 100.0
*Test is positive if Acetabular index, alpha angle, and percent femoral head coverage values exceed their respective cutoff values.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Prognostic Value of 6mo. Images for Late Dysplasia
were monitored, and no further instances of dysplasia in
these 30 hips were reported over a 12-month follow-up.
These 2 studies portray the variable interplay between the
6-month US and plain radiograph in the ongoing management
of DDH. Both of these imaging techniques play a vital role in
monitoring and treating this condition. In the current study,
AA on ultrasound was identified as the singular metric with the
best predictive ability for recurrent dysplasia at 2 years of age
(Table 2). However, the best prognostic value came from a
composite test that utilizes all metrics (AA, FHC, and AI).
Both imaging modalities performed well and were shown to be
clinically relevant. The differences between a composite test
versus a test based solely on sonographic or radiographic
measures are unlikely to be clinically meaningful for most
patients. Similar to many things in medicine, the ultimate
decision as to what images are necessary for a patient falls to
the discretion of the treating provider. Other elements of
treatment, such as family history and other risk factors, original
diagnosis severity, and response to bracing, must be considered.
Two successive studies by Dornacher et al25,26 (2010,
2013) identified high rates of residual dysplasia after suc-
cessful Pavlik harness or abduction bracing treatment. They
initially studied 90 patients (180 hips) from 2004 to 2006.
These patients were referred within the German healthcare
system after mandatory ultrasound screening between 4 and 6
weeks of life indicated abnormality.25 All patients underwent
successful treatment and achieved normal Graf I US findings.
Patients then received an anteroposterior (AP) radiograph at
the onset of ambulation, averaging 14.8 months of age.
Acetabular indices (AI) were measured and classified as de-
scribed by Tönnis.27 One hundred twelve hips (62.2%)
showed residual dysplasia on this radiograph ranging from
mild to severe.25 Their secondary study tried to quantify the
rate of persistent dysplasia after an initial positive radio-
graphic screen. Follow-up radiographs were reviewed for 72
infants (144 hips) at a mean age of 31.3 months who screened
positive for residual dysplasia on their initial radiograph de-
scribed above.26 Sixty-two hips (43%) now had a normal AI
on x-ray. Of the remaining patients, 56 hips (38.9%) pre-
sented with mild residual dysplasia and 26 hips (18.1%) had
severe residual dysplasia, according to Tonnis.26,27
Limitations
This study is limited by the single institution design and
a small number of patients who completed a 2-year follow-up
over the course of the study. Our results are only general-
izable to hips that do not require subsequent treatment at
Revised Cutoffs
Youden’s Index Cutoff Sens (%) Spec (%) Youden’s Index
0.08 > 24 59.5 94.1 0.54
0.00 ≤ 73 66.7 94.1 0.61
0.06 ≤ 62 81.0 70.6 0.52
0.00 All + 92.9% 70.6 0.63
www.pedorthopaedics.com | 103
Salton et al
6 months of age, as patients who received subsequent treat-
ment at that time were removed from our analysis. Patient
charts were reviewed by 3 pediatric orthopaedic surgeons, 2
orthopaedic residents, and 1 research assistant. All images
were Measured and Verified by the senior author, limiting
interobserver variability. Finally, there is a need to validate
our proposed cutoffs in a prospective cohort.
CONCLUSIONS
Both the US and radiograph have strengths in the
diagnosis and treatment of developmental dysplasia of the
hip. We identified that a combination of sonographic and
radiographic parameters at 6 months was the best pre-
dictor of late dysplasia at 2 years of age in patients suc-
cessfully treated with the Pavlik harness. Point-of-care US
and x-ray remain clinically important when utilized in-
dependently or in the union at 6 months. Given that the
rate of late or recurrent dysplasia at 2 years of age remains
high, identifying patients who may go on to develop late
dysplasia is critical to ensure appropriate follow-up care
and minimize risk. Diagnostic values of DDH currently
utilized in patient care for AA, FHC, and AI at 6 months
of age may be inadequate to predict the occurrence of
dysplasia at 2 years of age. This study identified an AA
≤ 73 degrees, FHC ≤ 62%, and AI > 24 degrees at
6 months of age to be predictive of residual dysplasia at
2 years. This data raises multiple questions that are out-
side of the scope of this particular study but that are ab-
solutely necessary to explore further. Specifically, are our
proposed cutoffs reproducible in an independent pop-
ulation? Among patients who do not meet these proposed
thresholds, is further treatment warranted to decrease the
risk of pain and/or long-term functional issues? Do we
need to continue to follow patients who are normal at
6 months and/or two years of age? Although all of these
questions warrant further prospective study, they are
outside of the scope of the current study. This study re-
inforces that both the 6-month of age radiograph and the
US are important images and proposes that cutoffs should
be reexamined if the goal is to capture patients who will
require further treatment and follow-up.
REFERENCES
1. Lehmann HP, Hinton R, Morello P, et al. Developmental dysplasia
of the hip practice guideline: technical report. Pediatrics. 2000;105:
e57–e57.
2. Alsaleem M, Set KK, Saadeh L. Developmental dysplasia of hip.
Clin Pediatr (Phila). 2015;54:921–928.
3. Harding M, Theodore HH, Bowen RJ, et al. Management of
dislocated hips with Pavlik harness treatment and ultrasound
monitoring. J Pediatr Orthop. 1997;17:189–198.
4. Swaroop VT, Mubarak SJ. Difficult-to-treat ortolani-positive hip
improved success with new treatment protocol. J Pediatr Orthop.
2009;29:224–230.
5. Albinana J, Dolan LA, Spratt KF, et al. Acetabular dysplasia after
treatment for developmental dysplasia of the hip. J Bone Joint Surg
Br. 2004;86-B:876–886.
6. Smith W, Badgley C, Orwig J, et al. Correlation of postreduction
roentgenograms and thirty-one year follow up in congenital
dislocation of the hip. JBJS American Volume. 1968;50:1081–1098.
104 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
7. Sarkissian EJ, Sankar WN, Zhu X, et al. Radiographic Follow-up of
DDH in Infants: are x-rays necessary after a normalized ultrasound?
J Pediatr Orthop. 2015;35:551–555.
8. Alexiev VA, Theodore Harcke H, Kumar S. Residual dysplasia
after successful Pavlik harness treatment. J Pediatr Orthop.
2006;26:16–23.
9. Cashman J, Round J, Taylor G, et al. The natural history of
developmental dysplasia of the hip after early supervised treatment in
Pavlik harness. a prospective, longitudinal follow-up. J Bone Joint
Surg Br. 2002;84:418–425.
10. Nakamura J, Kamegaya M, Saisu T, et al. Treatment for
developmental dysplasia of the hip using Pavlik harness. J Bone
Joint Surg Br. 2007;89-B:230–235.
11. Tucci J, Kumar S, Guille J, et al. Late acetabular dysplasia following
early successful pavlik harness treatment of cngenital dislocation of
the hip. J Pediatr Orthop. 1991;11:502–505.
12. Shaw KA, Moreland CM, Olszewski D, et al. Late acetabular
dysplasia after successful treatment for developmental dysplasia of
the hip using the Pavlik method: a systematic literature review.
J Orthop. 2019;16:5–10.
13. Copuroglu C, Ozcan M, Aykac B, et al. Reliability of ultrasono-
graphic measurements in suspected patients of developmental
dysplasia of the hip and correlation with the acetabular index.
Indian J Orthop. 2011;45:553–557.
14. Upasani VV, Bomar JD, Parikh G, et al. Reliability of plain
radiographic parameters for developmental dysplasia of the hip in
children. J Child Orthop. 2012;6:173–176.
15. Davila-Parrilla AD, Wylie J, O’Donnell C, et al. Reliability of and
correlation between measurements of acetabular morphology.
Orthopedics. 2018;41:e629–e635.
16. Novais EN, Pan Z, Autruong PT, et al. Normal percentile reference
curves and correlation of acetabular index and acetabular depth ratio
in children. J Pediatr Orthop. 2018;38:163–169.
17. Mulrain J, Hennebry J, Dicker P, et al. A normal screening
ultrasound does not provide complete reassurance in infants at risk of
hip dysplasia; further follow-up is required. Ir J Med Sci. 2021;190:
233–238.
18. Kitay A, Widmann RF, Doyle SM, et al. Ultrasound Is an
alternative to x-ray for diagnosing developmental dysplasia of the
hips in 6-month-old children. HSS J. 2019;15:153–158.
19. Bin K, Laville JM, Salmeron F. Developmental dysplasia of the hip
in neonates: evolution of acetabular dysplasia after hip stabilization
by brief Pavlik harness treatment. Orthop Traumatol Surg Res.
2014;100:357–361.
20. Modaressi K, Erschbamer M, Exner GU. Dysplasia of the hip in
adolescent patients successfully treated for developmental dysplasia
of the hip. J Child Orthop. 2011;5:261–266.
21. Morris WZ, Mayfield L, Beckwith T, et al. Late hip dysplasia after
normal ultrasound in breech babies: Implications on surveillance
recommendations. J Pediatr Orthop. 2021;41:e304–e308.
22. Barrera CA, Cohen SA, Sankar WN, et al. Imaging of devel-
opmental dysplasia of the hip: ultrasound, radiography and magnetic
resonance imaging. Pediatr Radiol. 2019;49:1652–1668.
23. Paton RW, Hinduja K, Thomas CD. The significance of at-risk
factors in ultrasound surveillance of developmental dysplasia of the
hip: a Ten Year Prospective Study. J Bone Joint Surg Br. 2005;87-B:
1264–1266.
24. Tudor A, Sestan B, Rakovac I, et al. The rational strategies for
detecting developmental dysplasia of the hip at the age of
4-6 months old infants: a prospective study. Coll Anthropol.
2007;31:475–481.
25. Dornacher D, Cakir B, Reichel H, et al. Early radiological outcome
of ultrasound monitoring in infants with developmental dysplasia of
the hips. J Pediatr Orthop B. 2010;19:27–31.
26. Dornacher D, Lippacher S, Reichel H, et al. Mid-term results after
ultrasound-monitored treatment of developmental dysplasia of the
hips: to what extent can a physiological development be expected?
J Pediatr Orthop B. 2013;22:30–35.
27. Tonnis D. Normal values of the hip joint for the evaluation of x-rays
in children and adults. Clin Orthop Relat Res. 1976;119:39–47.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Caregiver Experiences Using Orthotic Treatment Options
for Developmental Dysplasia of the Hip in Children
Georgia Grzybowski, BSc,* Emily Bliven, MEng,* Luke Wu, BSc,† Emily K. Schaeffer, PhD,‡§
Marissa Gibbard, BSc,† Bryn O. Zomar, PhD,‡§ Alice Casagrande Cesconetto, MEng,*
Claire Mundy, BSc,∥ Kishore Mulpuri, MBBS, MS, MHSc, FRCSC,‡§
and International Hip Dysplasia Registry Knowledge Translation Advisory Board
Background: Developmental dysplasia of the hip (DDH) is a
common condition affecting 5 in 1000 newborns. The standard
first line of treatment is the use of an orthotic, which has gen-
erally high success rates, but can pose substantial difficulties and
put undue burden on caregivers. The general experience of
caregivers using these orthotics has not been well documented on
an orthotic-specific basis. The purpose of this study was to in-
vestigate caregiver experience using prescribed DDH orthotics to
identify challenges, differences between treatment options, and
areas of improvement.
Methods: A survey assessing treatment prescription, respondent
demographics, and caregiver experience was distributed online to
caregivers whose child/children were treated for DDH with an
From the *University of British Columbia; †BC Children’s Hospital
Research Institute; ‡Department of Orthopaedics, University of
British Columbia; ∥University of British Columbia; and §Department
of Orthopaedic Surgery, BC Children’s Hospital, Vancouver, BC,
Canada.
IHDR Knowledge Translation Advisory Board members: Sarah Two-
mey, Nancy Muir, Bridget Watts, Loredana Guetg-Wyatt, Natalie
Trice, Charles Price, Regina Wilken, and Jennifer Farr. We’d also like
to acknowledge the collaborating organizations: Healthy Hips Aus-
tralia, Miles4Hips, Spica Life, Steps Charity Worldwide, DDH UK,
International Hip Dysplasia Institute, and I’m A HIPpy.
There was no external funding for this manuscript. None of the authors
received financial support for this study. The authors have no finan-
cial relationships.
K.M. receives funding support from the I’m a HIPpy Foundation, the
Peterson Fund for Global Hip Health, IHDI, BC Children’s Hospital
Foundation, the Canadian Orthopaedic Foundation, POSNA, Divi’s
Foundation for Gifted Children, Johnson & Johnson, Pega Medical
and OrthoPediatrics. The remaining authors declare no conflicts of
interest.
Reprints: Kishore Mulpuri, MBBS, MS, MHSC, FRCSC, Department of
Orthopaedic Surgery, BC Children’s Hospital, 1D.66 – 4480 Oak Street,
Vancouver, BC, Canada V6H 3V4. E-mail: kmulpuri@cw.bc.ca.
Supplemental Digital Content is available for this article. Direct URL
citations appear in the printed text and are provided in the HTML
and PDF versions of this article on the journal’s website, www.
pedorthopaedics.com.
Copyright © 2022 The Author(s). Published by Wolters Kluwer Health,
Inc. This is an open access article distributed under the terms of the
Creative Commons Attribution-Non Commercial-No Derivatives
License 4.0 (CCBY-NC-ND), where it is permissible to download and
share the work provided it is properly cited. The work cannot be
changed in any way or used commercially without permission from
the journal.
DOI: 10.1097/BPO.0000000000002312
J Pediatr Orthop  Volume 43, Number 2, February 2023
orthotic. Seven-point positively phrased Likert scale statements
and open-ended questions were included to assess caregiver ex-
perience. The results were analyzed using summary statistics and
orthotics with more than 30 responses were selected for more in-
depth analysis.
Results: A total of 530 survey responses were collected with 63%
(334/530) of respondents using a Pavlik harness, 45% (236/530) a
Rhino brace, and 13% (67/530) a Denis Browne Bar. The overall
weighted average score across all Likert Scale statements was
positive for the Pavlik harness, Rhino brace, and Denis Browne
Bar at 4.19 (95% CI, 3.83 to 4.54), 4.63 (95% CI, 4.27 to 4.99) and
4.91 (95% CI, 4.58 to 5.24), respectively. In the open-ended re-
sponses, all 3 orthotics were perceived as easy to use and not
hindering child-caregiver bonding, but raised concerns of dis-
comfort and skin irritation, as well as preventing the ability to
cuddle their child the way they desired. The Pavlik harness re-
spondents consistently brought up concerns regarding cleanability.
Conclusions: The results show that the DDH orthotics analyzed
are generally easy to use and perceived positively by caregivers,
but have orthotic-specific challenges that should be a focus of
future improvement work.
Clinical Relevance: This study evaluated opinions and attitudes
of caregivers for children being treated with DDH orthotics,
revealing experiences, concerns, and challenges associated with
the use of commonly prescribed options.
Key Words: DDH, orthotics, caregiver experience, pavlik harness,
rhino brace, denis browne bar
(J Pediatr Orthop 2023;43:105–110)
D evelopmental dysplasia of the hip (DDH) is a com-
mon congenital condition reported to affect up to 5 in
1000 newborns.1 DDH includes hips that range from un-
stable to completely dislocated.2 Without effective treat-
ment, this condition can lead to hip pain, osteoarthritis,
degenerative hip disease, and surgical intervention in se-
vere cases.3,4 The primary mode of treatment for DDH is
the use of an orthotic device or harness, which positions
hips into proper alignment in the joint to provide security
and stability during development.
There is a wide range of orthotic devices available
for nonsurgical treatment of infantile DDH, from soft
www.pedorthopaedics.com | 105
Grzybowski et al
fabric harnesses to plaster casting to stiff braces of various
shapes and sizes. Options commonly prescribed include
the Pavlik harness, a soft harness with adjustable feet-to-
shoulder straps (ages 0 to 8 mo); the Rhino brace, a rigid
brace on the hip and waist area that allows crawling and
walking (ages 6 wk and older); and the Denis Browne
(DB) bar, 2 thigh cuffs spanned by a stiff bar (newborn
and older). DDH can also be treated by a multimode
approach with various braces used at staggered phases of
recovery.
Clinical practice varies widely in terms of method,
timing, and duration of orthotic use, even among surgeons
practicing within the same country.5 Devices like the
Pavlik harness are often prescribed for continued at-home
use for up to 24 hours per day and up to 24 weeks. These
options further rely on caregiver compliance for ensuring
treatment efficacy. Studies have linked caregiver non-
compliance to both increased duration of treatment and
treatment failure.6,7
Caregiver attitudes toward specific orthotics remain
relatively unknown, outside of anecdotal evidence gath-
ered on a site-specific8,9 or surgeon-specific10,11 basis. As-
sessing the opinions and experiences of caregivers in their
use of DDH orthotics can identify treatment gaps, barriers
to administering proper treatment, and ways to improve
compliance and usability. Previous investigations have
demonstrated caregiver noncompliance due to orthotic-
related complaints,12 so applying information on such
trends could increase compliance and overall experience
using treatment devices.
To determine the current experience of caregivers
of children in prescribed DDH nonsurgical orthotics, a
survey was developed and distributed online. The pur-
pose of this study was to report current attitudes of
caregivers toward orthotic harnesses and braces pre-
scribed for DDH, identify common themes among sim-
ilar options, and summarize areas for innovation and
improvement in orthotics for DDH.
METHODS
We created a survey inquiring about DDH treat-
ment details, respondent demographics, and specific
questions about the different orthotics (Appendix 1, Sup-
plemental Digital Content 1, http://links.lww.com/BPO/
A558). The survey was distributed online through the In-
ternational Hip Dysplasia Registry (IHDR) and collabo-
rating organizations to caregivers of pediatric patients
who were treated for DDH with an orthotic harness
or brace. More specifically, survey distribution was
TABLE 1. Treatment History Indicated by Respondents, Reported as Mean (95% CI)
Category Pavlik Harness (95% CI)
Age of DDH diagnosis (wk) NA
Age DDH treatment began (wk) 7.3 (5.8–8.7)
Orthotic treatment length (wk) 7.7 (7.1–8.2)
Orthotic treatment length (h/d) 23.5 (23.3–23.7)
NA indicates not asked.
106 | www.pedorthopaedics.com Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.
J Pediatr Orthop  Volume 43, Number 2, February 2023
conducted online via IHDR mailing lists and social media
groups, and in-person during clinic visits at the authors’
institutions. Study data were collected and managed using
Research Electronic Data Capture hosted at our
institution.13,14 Research Electronic Data Capture (Van-
derbilt University, Nashville, TN) is a secure web-based
software platform designed to support data capture for
research studies. All respondents were required to agree to
a consent and privacy statement that detailed the pro-
tection of their data before proceeding to complete the
survey. This research did not undergo Institutional Review
Board approval, as it was a Quality Assessment and
Quality Improvement Project. The survey was available
online for 2 months (May 11 to July 14, 2020).
Survey Structure
The survey consisted of 5 question sections: (1) demo-
graphics and treatment history, (2) experience with the Pavlik
harness, Rhino brace, or “other” orthotics, and (3) compar-
ing the Pavlik harness and Rhino brace. Respondents were
instructed to select all applicable orthotic options. The survey
questions were chosen by the authors based on interest,
perceived concerns, and important factors for orthotics based
on clinical experience. The survey consisted of both closed-
ended and open-ended questions and positively framed
statements with responses ranging from “strongly agree”
to “strongly disagree” (Likert scale) and scored from a
maximum of 7 to a minimum of 1, respectively.
Survey Analysis
Surveys were analyzed on a per-question basis,
meaning that for each question only the respondents who
answered that question were included in the respective
analysis. Orthotics with more than 30 responses were
chosen for more in-depth analysis. Treatment history was
analyzed using summary statistics and 95% CI. The data
were first manually cleaned for values outside the allowable
range, such as responses with > 24 hours per day. For as-
sessment of the Likert statements, after responses marked
“not applicable” were excluded, weighted averages were
calculated with 95% CIs. Open-ended questions were ana-
lyzed in NVivo 12 (QSR International, Melbourne, Aus-
tralia).15 For every open-ended question, 20% of the
responses were randomly sampled for review to categorize
responses and identify themes. Two independent reviewers
analyzed the responses and subsequently categorized them
into prevalent themes.
Rhino Brace (95% CI) Denis Browne bar All Respondents (95% CI)
NA NA 14.6 (12.1–17.2)
38.6 (32.9–44.3) 21.1 (15.9–26.3) 17.9 (15.1–20.7)
15.5 (12.7–18.3) 12.3 (10.3–14.2) NA
20.3 (19.5–21.0) 22.4 (21.6–23.2) NA
J Pediatr Orthop  Volume 43, Number 2, February 2023 Caregiver Experiences Using Orthotic Treatments for DDH

FIGURE 1. Survey structure overview and breakdown of orthotics indicated as recommended for developmental dysplasia of the
hip treatment. Respondents were instructed to select all applicable options.

RESULTS http://links.lww.com/BPO/A559, http://links.lww.com/

A total of 530 survey responses were collected in
the first section, with treatment summary statistics shown
in Table 1. The survey included participants from 20
countries, such as Australia (46%, 242/524), the United
States (26%, 136/524), the United Kingdom (13%, 66/
524), and Canada (9%, 49/524). Among all respondents,
46% (245/530) indicated that more than 1 treatment
option was prescribed, 63% (334/530) reported using a
Pavlik harness, 45% (236/530) a Rhino brace, and 13%
(67/530) a DB Bar (Fig. 1).
Likert Statements
The overall weighted average score (out of 7) across
all statements was 4.19 (95% CI, 3.83 to 4.54) for the Pavlik
harness, 4.63 (95% CI, 4.27 to 4.99) for the Rhino brace,
and 4.91 (95% CI, 4.58 to 5.24) DB Bar. These scores
demonstrate a significant difference between the DB Bar
and Pavlik harness groups, but no significant differences
between the Pavlik harness and Rhino groups or Rhino and
DB Bar groups. The weighted average distribution sum-
marizing all Likert scale responses for the Pavlik harness,
Rhino brace, and DB bar are presented in Supplemental
Figures 2, 3, and 4, Supplemental Digital Contents 2-4,
BPO/A560, http://links.lww.com/BPO/A561, respectively.
Overall, the Pavlik harness had 5 Likert statements that
scored negatively (score < 4), Rhino brace had 2, and DB
Bar had 1. The highest (most positive) and lowest (most
negative) scoring Likert scale statements were the same
across each orthotic option analyze.
Open-ended Questions
Analysis of the open-ended questions revealed
a number of unifying themes shared across the assessed
orthotics (Table 2).
Child comfort was a consistent concern across all the
3 orthotics, with respondents calling for softer materials
and increased padding. Improper fit was also raised as a
concern, particularly for the Rhino brace. Skin irritation
was identified as a major concern for both the Pavlik
harness and Rhino brace, arising in the form of rashes and
chafing because of material type or improper fit. Re-
spondents indicated that all the 3 orthotics were easy to
use, despite the expressing apprehensiveness in an ability
to properly adjust the Pavlik harness and DB Bar. They
called for more intuitive straps to relieve the anxiety sur-
rounding orthotic adjustments.

Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. www.pedorthopaedics.com | 107
Grzybowski et al
TABLE 2. Key Themes Found in Open-ended Questions With (+) Denoting a Positive Response and (−) Denoting a Negative
Response
Key Theme Pavlik Harness
Comfort (+) “Softer than other harnesses”
and (−) “Wearing the harness 24 h a day can cause
irritation some uncomfort to the child such as irritation of (−) “It also gave my child several rashes and dry
skin.”
Ease of use (+) “It is easy to understand how it works and it is (+) “It was easy to use and put on as a parent”
pretty easy overall”
(+) “It is a simple but effective and non-invasive
medical treatment”
(−) “It was challenging to make it adjusted
correctly at all times”
(−) “That it would be too loose or tight and not
effective”
Cleanliness (−) “My daughter had reflux and it became so
gross and dirty”
(−) “very difficult to keep clean, we ended up
buying 2 so that we could swap them each week
to wash one when the physio changed it over”
Impediment (−) “Caring for our daughter in the way we
of daily intended was interrupted because we had to pick ways to manage shopping trolleys, etc. was
activities her up, hold her, and change her diaper all
significantly different”
Hygiene of both the child and their orthotic seemed
to be important to caregivers. Respondents praised the
material of the Rhino brace for being easy to clean and be
able to remove for bathing, whereas the Pavlik harness
was criticized for the difficulty in cleaning the harness and
the child while wearing it. Both DB Bar and Pavlik har-
ness respondents suggested removable covers to facilitate
easier cleaning of the orthotics.
All orthotics analyzed were perceived as a burden in
terms of impeding daily activities: the Rhino brace was
described as bulky and incompatible with sleeping, car
seats, and baby carriages (prams), and the Pavlik harness
as interfering with feeding and diaper changes, as well as
incompatible with standard clothing. All the 3 orthotics
were reported by caregivers to impede cuddling and
holding of their child.
Pavlik Versus Rhino
When asked to compare the Pavlik harness with the
Rhino brace, respondents indicated that they preferred the
softer, more flexible material of the Pavlik harness over
the Rhino brace. Generally, respondents found the Pavlik
harness more difficult to use than the Rhino brace, spe-
cifically finding the straps more challenging. Respondents
enjoyed that the Rhino brace was removable for certain
activities such as baths, diaper changes, or tummy time as
opposed to the Pavlik harness. There were no standout
negative features reported for the Rhino brace.
DISCUSSION
This study evaluated caregiver experiences and atti-
tudes toward orthotic harnesses and braces prescribed for
DDH through a survey consisting of open-ended ques-
108 | www.pedorthopaedics.com
J Pediatr Orthop  Volume 43, Number 2, February 2023
Rhino Brace Denis Browne bar
(−) “I was concerned about her being (−) “My daughter was not
uncomfortable as a brand new baby” comfortable in it.”
skin from rubbing”
(−) “Leg foam was also ill-fitting and hot”
(+) “It seems much easier and less
(+) “It’s easier than the pavlik because you can’t invasive than many other
mess up the fit” options.”
(+) “It was adjustable depending
on how my daughter was
growing.”
(−) “It was difficult to know how
tight the straps should be when
replacing them with clean
ones.”
(+) “It’s easy to clean” (−) “[An improvement could be
(+) “Easy to keep child clean.” a] Removable cover over bar to
help keep clean”
(−) “Cleaning my child [was a
concern].
(−) “finding car seats/prams/carriers/high chairs/ (−) “Not being able to cuddle
her”
really hard. Especially when I needed a double
pram to accommodate toddler too.”
tions and Likert scale statements. In terms of weighted
average scores across all Likert scale statements, all the 3
analyzed options scored above neutral, suggesting positive
perceptions of use. The Pavlik harness scored the lowest
(4.19) compared with the Rhino brace (4.64) and DB Bar
(4.91). The Pavlik harness also had a greater number of
statements that elicited an average negatively scored re-
sponse; at least 2-fold that of the Rhino brace or DB Bar.
These trends could indicate that caregivers perceive the
Pavlik harness more negatively or as more difficult to use
than the other 2 orthotics included in our study.
Respondents voiced strong concerns about the
treatment being uncomfortable for their baby and/or ir-
ritating their child’s skin. However, the statement: “My
child is happy and/or appears comfortable while wearing
the recommended treatment option” scored positively
( > 4) in all the 3 orthotics. Worries about skin irritation
and discomfort from the prescribed use of a DDH orthotic
have been consistently reported in other studies, with a
majority of respondents voicing similar concerns.8–10,16
These concerns may be the unfortunate reality of wearing
any orthotic harness or brace for a prolonged time, but
remain a documented adverse experience of orthotic use
for the treatment of DDH.
Respondents consistently provided the lowest scores
to the statement “I can cuddle my child the way I wish
while they are wearing the recommended orthotics.” Past
research has demonstrated that DDH orthotics present an
emotional and physical barrier to cuddling and are per-
ceived to hinder the quality of attachment between parent
and child.8,10,16 Our study also revealed in both Likert
scale statements and open-ended questions that there were
difficulties experienced with cuddling and holding during
Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.
J Pediatr Orthop  Volume 43, Number 2, February 2023
use of orthotics. Interestingly, negative scores were not
found for child attachment or bonding, as the statement
“I feel like I’ve been able to bond well with my child
throughout their treatment” demonstrated the highest
score for all the 3 orthotics analyzed. This is in contrast
with previous literature, which described the psychosocial
consequences for caregivers when bathing is difficult or
forbidden when using a certain orthotic, resulting in
feelings of loss of parental autonomy and missing out on
a chance to bond with their child.11 However, the
discrepancies found may be because of the difficulty
distinguishing between the meanings behind the keywords
“cuddling,” “holding,” and “bonding.”
Open-ended responses indicated that respondents
felt that all the 3 orthotics were easy to use and take on
and off, a sentiment further supported by high-scored
statements on this subject. These results align with those of
a previous study, which found a majority of parents did
not report the harness/brace prescribed for their infant’s
DDH as difficult to manage.17 Despite these trends, re-
spondents using the Pavlik harness and DB Bar showed
concern in the open-ended questions that the adjustment
straps or fit were not correct, either initially or after they
took the orthotic on or off by themselves. Rhino brace
users were also concerned that the fit was not correct for
their child and worried this improper fit resulted in dis-
comfort. Respondents liked that they were allowed to take
off the Rhino brace for periods during the day, which is
indicative of the treatment stage and nature of prescribed
treatment rather than reflective of the design of the device
itself. General discrepancies in reported values for time
out of brace between orthotic options may be an issue in
physician instruction rather than an aspect inherent to the
orthotic itself.
Feeding and breastfeeding their child was found to
be an interrupted aspect of daily life for respondents using
the Pavlik harness. However, Pavlik respondents scored
the statement pertaining to feeding their child an almost-
neutral score of 4.2. Feeding and breastfeeding were not
frequently mentioned concerns or interruptions of daily
life for the Rhino brace or DB Bar, which may be because
the children being treated with these braces were typically
older, on average 38.6 and 21.1 weeks, respectively,
compared with 7.3 weeks for Pavlik harness (Table 1). A
Swedish study reported significantly lower breastfeeding
frequency in infants prescribed a von Rosen splint for
DDH when compared with a control group of healthy,
nonsplinted infants.16 Although researchers acknowledged
that these differences in feeding may be because of other
factors, parents often described issues attributed to the
orthotic similar to those mentioned by respondents in our
study, such as skin irritation, poor contact, and practical
problems like clothing compatibility.16
Harness cleanliness is important for infant hygiene
and caregiver experience; a case study of a mother of an
infant diagnosed with DDH described similar concerns of
keeping the orthotic clean while changing diapers.10 Con-
cerns of cleanliness were apparent in the survey results for
the Pavlik harness and are especially relevant for the Pavlik
Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.
Caregiver Experiences Using Orthotic Treatments for DDH
harness, which is made of a soft and flexible textile that is
not designed to be water/stain resistant, unlike the Rhino
brace, which is made from a stiff plastic shell likely to repel
liquids and absorb fewer materials and odors. Many
caregivers suggested improving Pavlik harness cleanability
by adding removable covers to the straps, allowing the
dirty covers to be washed and replaced with clean ones
without interrupting the prescribed treatment. Another
potential solution is for clinicians to prescribe (and insurers
to subsidize) an additional Pavlik harness for patients,
enabling uninterrupted wear while the first harness is being
washed. Correlating the concerns of cleanliness and
hygiene documented in this study with medical conditions
like irritation and infection were out of the scope of the
study. However, this research lays the groundwork for
studying this in the future.
The insights gained from this research can inform
future innovation of DDH orthotics based on the specific
concerns reported by caregivers. Many of the limitations
of this work are inherent to the collection of survey data,
including recall bias and selection bias. We used positively
phrased Likert statements to improve reliability and con-
struct validity. However, this may have increased acqui-
escence bias.18 We did not control for the independent
effect of prescribed treatment plan variation among par-
ticipants and between orthotics. Some of the trends found
between orthotics could be because of the differences in
typical orthotic treatment prescription or infant age,
rather than differences between orthotic design. Infant age
at time of prescription, or whether the infant was a first-
born child or younger sibling, may have influenced parent
responses to many of the questions across the survey.
Newer parents still becoming comfortable with caring for
their newborn may find dealing with brace wear more
stressful than more experienced parents. Consequently,
patient age and birth order should be taken in consid-
eration for future research. Finally, the scope of our study
prevents conclusions from being drawn beyond between-
group comparisons; it could be that some of the concerns
raised are innate to an infant wearing any prescribed
medical device, and we are unable to comment further on
this topic without the use of a control comparison.
CONCLUSIONS
Generally, the 3 DDH orthotics analyzed (Pavlik
harness, Rhino brace, and DB bar) were positively scored
by respondents and perceived as easy to use and not hin-
dering child-caregiver bonding. However, a closer look at
the individual responses and trends within and between
groups demonstrated areas for improvement. Caregivers
across the 3 options analyzed reported consistent concerns
with the discomfort and skin irritation due to wearing of
the orthotic, as well as an inability to cuddle their child in
the way they desired. Despite being described as easy to
use, the Pavlik harness and Rhino brace were perceived as
problematic in terms of fit, and specifically in the way of
strap adjustments for the Pavlik harness. The cleanability
www.pedorthopaedics.com | 109
Grzybowski et al
of the Pavlik harness was also repeatedly voiced as
worrisome to caregivers.
The presented study paints a well-rounded picture of
caregivers’ attitudes toward DDH orthotics on a device-
specific basis, allowing for the pros and cons of various
options to be identified. This work fills a gap in the current
literature for a population that often lacks consideration
in orthotic design: the families and caregivers of patients
that use these orthotic devices. Our findings elucidate what
should be prioritized in future DDH orthotic design and
development to optimize caregiver experience, potentially
increasing user compliance and positive implications for
the clinical setting.
ACKNOWLEDGMENTS
The authors thank all members of the IHDR Knowledge
Translation Advisory Board, as well as collaborating organ-
izations: International Hip Dysplasia Institute, I’m A HIPpy,
and the UBC Engineers in Scrubs program. The authors also
thank Wendy Krishnaswamy at BC Children’s Hospital for
generously offering expertise in designing this study.
REFERENCES
1. Woodacre T, Ball T, Cox P. Epidemiology of developmental
dysplasia of the hip within the UK: refining the risk factors. J Child
Orthop. 2016;10:633–642.
2. Aronsson DD, Goldberg MJ, Kling TF, et al. Developmental
dysplasia of the hip. Pediatrics. 1994;94:201–209.
3. Malvitz TA, Weinstein SL. Closed reduction for congenital dysplasia
of the hip. Functional and radiographic results after an average of
thirty years. J Bone Joint Surg Am. 1994;76:1777–1792.
4. Muller GM, Seddon HJ. Late results of treatment of congenital
dislocation of the hip. J Bone Joint Surg Br. 1953;35(B(3)):342–362.
5. Heeres RHM, Witbreuk MMEH, van der Sluijs JA. Diagnosis and
treatment of developmental dysplasia of the hip in the Netherlands:
national questionnaire of paediatric orthopaedic surgeons on current
practice in children less than 1 year old. J Child Orthop. 2011;5:267–271.
110 | www.pedorthopaedics.com Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.
J Pediatr Orthop  Volume 43, Number 2, February 2023
6. Mubarak S, Garfin S, Vance R, et al. Pitfalls in the use of the Pavlik
harness for treatment of congenital hip dysplasia, subluxation, and
dislocation of the hip. J Bone Joint Surg Am. 1981;63:1239–1248.
7. McHale K, Corbett D. Parental noncompliance with Pavlik harness
treatment of infantile hip problems. J Pediatr Orthop. 1989;9:
649–652.
8. Hassan FA. Compliance of parents with regard to Pavlik harness
treatment in developmental dysplasia of the hip. J Pediatr Ortho B.
2009;18:111–115.
9. Chao M, Chiang VC. Impact on and coping behaviours of a Chinese
mother with a child suffering from developmental dysplasia of the
hip. J Orthop Nurs. 2003;7:176–183.
10. Elander G. Breast feeding of infants diagnosed as having congenital
hip joint dislocation and treated in the von Rosen splint. Midwifery.
1986;2:147–151.
11. Poole C. exploring the experiences of parents caring for their infant
with developmental dysplasia of the hip (DDH): an interpretative
phenomenological analysis doctoral dissertation. Edinburgh Napier
University; 2019.
12. Jennings HJ, Gooney M, O’Beirne J, et al. Exploring the experiences
of parents caring for infants with developmental dysplasia of the hip
attending a dedicated clinic. Int J Orthop Trauma Nurs. 2017;25:
48–53.
13. Harris PA, Taylor R, Thielke R, et al. Research electronic data
capture (REDCap)—a metadata-driven methodology and workflow
process for providing translational research informatics support.
J Biomed Inform. 2009;42:377–381.
14. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium:
building an international community of software partners. J Biomed
Inform. 2019;95:103208.
15. NVivo Qualitative Data Analysis Software [computer program]
Version 12. QSR International Pty Ltd; 2018.
16. Wakely L, Easey P, Leys J, et al. Exploring the lived experience of
parenting a child with developmental dysplasia of the hip. Phys
Occup Ther Pediatr. 2021;8:1–12.
17. Gardner F, Dezateux C, Elbourne D, et al. The hip trial:
psychosocial consequences for mothers of using ultrasound to
manage infants with developmental hip dysplasia. Arch Dis Child
Fetal Neonatal Ed. 2005;90:F17–F24.
18. Chyung SY, Barkin JR, Shamsy JA. Evidence-based survey design:
the use of negatively worded items in surveys. Perform Improv.
2018;57:16–25.
 ORIGINAL ARTICLE

Does Prolonged Weight Relief Increase the Chances

of a Favourable Outcome After Containment
for Perthes Disease?
Hitesh Shah, MS Orth, Kumar Amerendra Singh, MS Orth,
and Benjamin Joseph, MS Orth, MCh Orth, FRCS Ed

Key Words: Legg-Calvé-Perthes disease, containment, weight

Abstract: We have been treating children with Legg-Calvé-Perthes relief, Sphericity Deviation score
disease (LCPD) with a femoral varus osteotomy (PFVO) and
weight relief till the disease evolves to the latter part of the stage of (J Pediatr Orthop 2023;43:e144–e150)
reconstitution (Stage IIIb). This entails weight relief for 18 to
24 months. We undertook this case-control study to test if a shorter
period of weight relief would compromise the chance of retaining
the spherical shape of the femoral head when the disease healed.
Forty-one children diagnosed in the early stages of LCPD (Stages
T he role of weight relief in the treatment of Legg-
Calvé-Perthes disease (LCPD) is unclear. Currently,
most surgeons do not advocate weight relief beyond the
Ia, Ib, and IIa), were treated by PFVO and non–weight-bearing for
few weeks needed for bones to heal after a femoral or
a period of 6 months following surgery (6m group). Eighty-two
acetabular procedure done to achieve containment.
children with LCPD matched for age, sex, and stage at surgery,
However, a few studies have included weight relief as an
who resumed weight-bearing only once they reached Stage IIIb,
integral part of the treatment of LCPD; they include the
served as the control group (3b group). Both groups were followed
odd report of using weight relief alone as a form of
up till the disease healed. The sphericity deviation score was cal-
definitive treatment and those that combine weight relief
culated, and the height and width of the epiphysis were measured
with operative or nonoperative containment.1–5
on the first radiograph designated as Stage IV. The median
At this center, for the past 30 years, we have been
sphericity deviation score value at healing was 3 in the 3b group
treating children with LCPD with a proximal femoral
and 11 in the 6m group (P < 0.001). The frequency of spherical
varus osteotomy (PFVO) combined with weight relief till
heads was 76% in the 3b group and 49% in the 6m group
the disease has evolved to the latter part of the stage of
(P < 0.003). The Odds Ratio of the disease healing with an as-
reconstitution (Stage IIIb of the modified Waldenström
pherical head in 6-month group was 3.05 (CI: 1.28 to 7.22) com-
classification6).5 The rationale for delaying weight-bearing
pared with the 3b group. The percentage increase in width of the
till Stage IIIb is based on the assumption that earlier in the
femoral epiphysis at healing was greater in the 6 group
disease, the epiphyseal bone is not strong enough to resist
(111.5 ± 8.5% vs. 106.5 ± 7.2%; P < 0.001). The study confirms that
deforming stresses of weight-bearing. Based on the cu-
containment by PFVO performed early in the course of LCPD
mulative duration of Stages Ia to IIIa7,8 the period of
combined with weight relief till the disease has evolved to Stage
weight relief could be between 18 to 24 months.
IIIb is likely to result in spherical hips in 75% of children. Reducing
Realizing the need to justify such a prolonged period
the period of weight relief to 6 months may yield significantly
of crutch walking, we undertook this study to test if a
poorer results with only 49% spherical femoral heads.
shorter period of weight relief following a PFVO would
jeopardize the chance of retaining the spherical shape of
the femoral head when the disease healed.
From the Department of Paediatric Orthopaedics, Kasturba Medical
College, Manipal Academy of Higher Education, Manipal, 576104,
Karnataka, India. METHODS
H.S.: Study design, data collection, data analysis, and writing the We embarked on this retrospective case-control
manuscript. K.A.S.: Data collection, data analysis, and writing study after obtaining clearance from the Institutional
the manuscript. B.J.: Study design, data analysis, and writing the
manuscript.
Ethical Committee.
Study was done at the Department of Paediatric Orthopaedics, Kasturba The protocol of treatment of all the children in this
Medical College, Manipal, India. study involved achieving surgical containment by per-
The authors declare no conflicts of interest. forming an open-wedge PFVO at the sub-trochanteric
Reprints: Benjamin Joseph, MS Orth, MCh Orth, FRCS Ed, level and fixing the osteotomy with a plate prebent to 20
18 HIG HUDCO Colony, Manipal, Karnataka 576104, India.
E-mail: bjosephortho@yahoo.co.in. degrees. The surgery was performed on all the children in
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. an identical manner, either by 1 of the authors or by
DOI: 10.1097/BPO.0000000000002302 trainees under the direct supervision of 1 of the authors.

e144 | www.pedorthopaedics.com J Pediatr Orthop  Volume 43, Number 2, February 2023

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
The technique of surgery has not changed over the last 3
decades.5 Children below the age of 8 years at onset of the
disease with extrusion of the femoral head on the AP ra-
diograph and children aged 8 years and above at disease
onset, irrespective of the presence or absence of extrusion,
were treated with PFVO provided they were in Stage Ia, Ib
or IIa of the disease according to the modified Walden-
strom classification and if more than half the epiphysis
was sclerotic or if the sub-chondral fracture line of Salter
and Thompson extended over half the width of the
epiphysis.9 The surgery was followed by a prolonged pe-
riod of complete non–weight bearing with the aid of ax-
illary crutches, as outlined below. The implants were
removed after the disease healed.
Forty-one children (36 boys and 6 girls) with LCPD
(mean age at onset: 7.88 ± 1.07 y), diagnosed in the early
stages of the disease (Stages Ia, Ib, and IIa) between 2013
and 2018, were treated by PFVO and non–weight-bearing
with axillary crutches for a period of 6 months after sur-
gery (designated as the 6m group). This cohort included 24
children who were part of a separate study with totally
different objectives.10 We deviated from our practice of
avoiding weight-bearing till Stage IIIb in this cohort of
children to fulfill the criteria for their inclusion in an in-
ternational multi-center study.
From a prospectively collected database of 340
children with LCPD treated between 2007 and 2018 who
are being currently followed up till skeletal maturity at our
center, we identified 82 children matched with children in
the 6m group for age, sex, and stage at surgery based on
the modified Waldenstrom classification. They served as
the control group (designated as the 3b group), giving a
case: control ratio of 1: 2.
All the children were followed up prospectively in a
dedicated LCPD clinic; they were seen once every
4 months till the disease progressed to Stage IIIb and
thereafter, once every 6 months till the disease healed; no
child was excluded from the study. The disease was
deemed to have healed when there were no sclerotic areas
in the femoral epiphysis in the AP and lateral view ra-
diographs.
The children in the 3b group had been treated and
followed up in exactly the same way as those in the 6m
group with 1 exception; they resumed weight-bearing only
once they reached Stage IIIb. This entailed a period of
weight relief of 25.77 ± 7.16 months (range: 12 to 46 mo).
Sequential radiographs of each child were studied;
the Catterall grouping and Herring classification were
applied to radiographs in the fragmentation stage. The
classification of Catterall was used to estimate the extent
of hypovascularity of the femoral epiphysis rather than on
a perfusion MRI scan, as only the 6m group had MRI
scans. The extrusion of the femoral head was measured on
AP radiographs of the pelvis. The duration of the disease
from onset to the commencement of Stage IIIb and from
onset to healing was calculated. The Sphericity Deviation
Score (SDS)9,10 was calculated once the disease healed.
The frequency of hips with SDS values of 10 or under was
calculated for each group (the lower the SDS value, the
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Prolonged weight-relief in Perthes disease
better the shape of the femoral head, and an SDS value
below 10 implies that the head is spherical).11,12 The
height and width of the epiphysis at healing were mea-
sured and expressed as a percentage ratio of the corre-
sponding dimensions of the contralateral normal hip. All
radiographic measurements were performed by 2 of the
authors, who are pediatric orthopaedic surgeons with 10
and 20 years of experience, respectively, in treating LCPD.
Statistical Methods
Reproducibility of radiographic measurement was
estimated by 2 authors measuring 30 radiographs on 2
separate occasions, 2 weeks apart, and computing the
Intra-class Correlation Coefficients (ICC). The authors
were blinded to the group to which the child belonged
while making these measurements. The Mann-Whitney
test was applied to compare the SDS values between the 2
groups and the χ2 test was used to compare the frequency
of SDS values below 10 between the 2 groups. The Odds
Ratio and 95% Confidence Intervals (CI) of a poor out-
come in children in the 6m group were computed.
RESULTS
The reproducibility of the measurement of SDS and
all other measurements was excellent (Table 1). The
distribution of Catterall groups was comparable in the 3b
and 6m groups; the vast majority of children of both
groups were classified as Catterall III whereas no child had
Catterall I involvement in both the groups (Table 2). The
extent of the epiphyseal collapse was also comparable in
both 3b and 6m groups; the majority of children were
classified as Herring B.
The median SDS values at healing were distinctly
lower in the 3b group (median: 3; IQR: 1, 11.25) as
compared with the 6m group (median: 11; IQR: 5.5, 18.5;
P < 0.001 – Fig. 1) and the frequency of spherical heads
(SDS < 10) was 76% among the children in the 3b group
whereas it was 49% in the 6m group.
The Odds Ratio (OR) of the disease healing with an
aspherical head in 6m group was 3.05 (CI: 1.28 to 7.22)
compared with the 3b group.
The percentage increase in width of the femoral
epiphysis at healing was greater in the 6m group
(111.5 ± 8.5% vs. 106.5 ± 7.2; P < 0.001) and the height of
TABLE 1. Reproducibility of Radiographic Measurements
Inter-observer Intra-observer
Reliability ICC Reliability ICC
Variable (95% CI) (95% CI)
Extrusion 0.88 (0.82-0.94) 0.92 (0.88-0.96)
Stage of disease 0.93 (0.87-0.98) 0.96 (0.9-0.98)
SDS 0.89 (0.83-0.94) 0.92 (0.89-0.96)
Width ratio of the 0.92 (0.84-0.96) 0.94 (0.89-0.99)
epiphysis
Height ratio of the 0.91 (0.82-0.96) 0.93 (0.88-0.98)
epiphysis
ICC indicates intra-class/inter-class correlation coefficients; SDS, sphericity
deviation score
www.pedorthopaedics.com | e145
Shah et al
TABLE 2. Characteristics of the 6m and 3b Groups
Variable 6m Group (n = 41)
Age at onset 7.88 ± 1.07 y
(Range: 6-10.3 y)
Sex Boys: 35
Stage at the commencement of treatment Stage Ia Stage Ib
2 36
Extrusion 17.02 ± 8.20%
(Range: 0-34)
Age at healing 11.18 ± 1.6 y
(Range: 8.08-14.75)
Duration of disease 39.63 ± 12.92 mo
(Range: 16-72)
the epiphysis was less in the 6m group, but the difference
was not statistically significant (Table 3).
DISCUSSION
Weight relief and Containment
The aim of the treatment of LCPD is to prevent the
femoral head from getting deformed and the strategies for
the treatment revolve around 2 concepts; containment and
weight relief. Weight relief was popular 50 years ago13,14
but in more recent times, it is seldom practiced largely
because of the concern regarding the possibility of adverse
psychological effects on the child. The current trend is to
rely solely on containment, ignoring the role of weight
relief.
However, experimental studies have shown that
weight relief does have a definite protective effect on the
shape of the femoral head following ischemic necrosis.15
Similarly, a few clinical studies have endorsed this
view.3,4,16,17 The impression that weight-bearing in the
early stages of LCPD predisposes to deformation of the
femoral head has been recently supported by the findings
in a novel MRI study.18 The study demonstrated flat-
tening and widening of the epiphysis as soon as children
with LCPD who were early in the course of the disease
bore weight on the limb. This study, for the first time,
furnished objective evidence of the association between
weight-bearing and femoral head deformation in LCPD,
which till now had been assumed by protagonists of
weight relief. The observations we report in the present
study further justify an approach of combining contain-
ment and weight relief as both these elements of treatment
have an additive effect on the outcome.
Duration of Weight relief
If weight relief is incorporated as part of treatment,
how long should the duration of weight relief be? The ra-
tional answer would be, “till the propensity for deformation
of the femoral head persists”. We have been avoiding
weight-bearing till the disease has passed through the stages
of fragmentation and early reconstitution (Stage IIa, IIb,
and IIIa) because we postulate that resorption of the ne-
crotic bone is at its peak in the Stage II rendering the bone
weak and woven bone being laid down in Stage IIIa
(Fig. 2A, B) is also not strong enough to resist deformation
e146 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
3b Group (n = 82) Significance
7.84 + 1.1 y NS
(Range: 6-10.3) P = 0.843
Girls: 6 Boys: 70 Girls:12 Not applicable
Stage IIa Stage Ia Stage Ib Stage IIa Not applicable
3 4 72 6
18.05 ± 8.00% NS
(Range: 0-37) P = 0.509
10.9 ± 1.57 y NS
(Range: 7.75-16 y) P = 0.401
37.20 ± 13.74 mo NS
(Range: 16-76) P = 0.348
if subjected to weight-bearing stress. Interestingly, 50 years
ago, Lauritzen14 pursued weight relief “till healing was so
far advanced that …the epiphyseal contours were fairly
continuous”, a description of what seems to be Stage IIIb
(Fig. 2C, D). The time required to reach this point in his
patients was around 24 months14 and similar periods of
weight relief have been reported by other authors.3,4 The
mean duration of weight relief in the 3b group was also
24 months.
Is there a real need to persist with weight relief for
such a long period of time; would not a shorter period
suffice? The present study addressed this question. The
results of weight relief for 6 months following containment
with a proximal femoral varus osteotomy (PFVO) were
far inferior to the results in those who did not bear weight
till Stage IIIb (49% spherical heads vs. 76%, P = 0.003).
These results justify our approach of advocating pro-
longed weight relief as an integral part of our treatment
protocol.
The Effect of Weight-bearing and Weight relief
on the Avascular Femoral Epiphysis
In a recent study, the authors suggested that weight-
bearing stresses on an avascular femoral epiphysis may
result either in compaction of the bony tissue or mush-
rooming; compaction results in the loss of epiphyseal
height while mushrooming results in an increase in epi-
physeal width in conjunction with a reduction in height18
The MRI study of Aarvold et al19 demonstrated that
weight-bearing causes mushrooming of the avascular ep-
iphysis in LCPD. Though the deformation was transient
in their study, it could become permanent over time. In-
creased width of the epiphysis noted in the 6m group
suggests that the epiphysis has undergone mushrooming
and early weight-bearing may have contributed to this.
Effect of Prolonged Weight relief on the
Outcome of Treatment
This study focused on the radiologic outcome with
specific reference to the shape of the femoral head mea-
sured by the SDS. Our results indicate that the disease
heals with spherical femoral heads in a high proportion of
children treated by prolonged weight relief combined with
containment Figure 3A–D). Direct comparison of our
results with those of the other studies is difficult for 2
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Prolonged weight-relief in Perthes disease

TABLE 3. Comparison of Results of 6m and 3b Groups

Groups
6m (n = 41), 3b (n = 82),
Variable n (%) n (%) Significance
Sphericity Deviation Score
≤ 10 20 (49) 62 (76) P < 0.003
> 10 21 (51) 20 (24)
Width of the femoral 111.5 ± 8.5 106.5 ± 7.2 P < 0.001
epiphysis at healing (Range: (Range:
94-129) 92-122)
Height of the femoral 76.78 ± 13.62 79.44 ± 16.01 NS
epiphysis at healing (Range: 49- (Range: 54- P = 0.361
102) 104)

indicating that nonoperative containment when combined

with prolonged weight relief may be equally effective as
our approach of surgical containment combined with
prolonged weight relief.3,4

Limitations and Strengths of this Study

A prospective randomized trial (RCT) is un-
doubtedly the most robust study design to compare 2
different approaches to treatment, and the level of evi-
dence of the present case-control study is lower than an
RCT. However, by matching for age, sex, and the stage
of the disease at treatment, we adjusted for 3 important

FIGURE 1. Scatter plot (A) and box plots (B) depicting the
SDS values at the healing of children in 6m and 3b groups.
The median values are shown as dark lines in the boxes.

reasons. Firstly, many authors did not measure the SDS,

which is a relatively new quantitative measure of the shape
of the femoral head11,12 but reported the frequency of
Stulberg classes.20 Secondly, most studies were case series
without controls. Nevertheless, there appears to be a trend
towards better results in studies that combined
containment with prolonged weight relief 3,4,17 than
when containment alone was performed.21–30 Two
reports of children treated in abduction, non–weight-
bearing braces worn for 18 months show that a high
proportion had spherical hips when the disease healed,
FIGURE 2. In stage IIIa of the disease, the new-bone laid down
on the periphery of the epiphysis is not of normal texture
(arrows in 2a, 2b). Weight-bearing is not permitted at this
stage as the woven bone may not resist these forces. Once the
disease has progressed to Stage IIIb, the bone formed on the
periphery of the epiphysis is of normal texture (arrows in 2c,
2d). This is the point at which weight-bearing is resumed.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e147

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Shah et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 3. A and B,: AP and frog-lateral radiographs of an 8-year-old child with LCPD in stage I of the disease; the sub-chondral
fracture line extends across more than half the femoral epiphysis. PFVO was performed, and weight-bearing was avoided till the
disease evolved to Stage IIIb. C, D: The disease healed with a spherical femoral head (SDS: 9).

potentially confounding variables. Furthermore, 2 more
potential confounding variables of profound prognostic
significance, the extent of epiphyseal avascularity and
the extent of epiphyseal collapse, were comparable in
both the groups. The treatment (other than the period
of weight relief) was also identical in both the groups
enabling us to make a meaningful comparison of
outcomes.
Another concern that the varus osteotomy of the
femur and the hypoplasia of the limb, secondary to not
bearing weight on the limb, will result in an unacceptable
degree of permanent shortening, which needs to be ad-
dressed. In 2 previous studies from this center of children
treated in a manner identical to that of this study, the
mean shortening at skeletal maturity was 0.44cm and
0.5cm, respectively.31,32 We did not do limb-length
measurements in this study as the end-point of our study
was healing of the disease, and the potential for sponta-
neous equalization of limb lengths was present till skeletal
maturity.
It may be argued that since we have been practicing
prolonged weight relief, there is likely to be an element of
bias towards prolonged weight relief while undertaking the
study. However, for some time, we have been asking
ourselves whether we do need to persist with non–weight-
bearing for several months, and the inclusion of children
in a multi-center study that mandated a 6-month period of
weight relief gave us the opportunity to answer this
question. We would have been very pleased if the study
did show that a shorter period of weight relief was as
effective as it would have greatly benefitted the children
and their families.
One question that remains to be answered is how
compliant the children were with non–weight-bearing.
Unfortunately, there is no reliable way of monitoring
compliance without fixing sensors in the crutches. Never-
theless, over the years, we have observed our patients
closely and we noted compliance to be far better among the
children who underwent surgery than children treated
nonoperatively (Fig. 4). This appears to be because both the
children and their parents are afraid that failure to adhere to
our recommendations may lead to complications and
failure of surgery. Singh et al10 stated that hypoplasia of
the calf and foot, wearing out of the axillary pads and
bushes of the crutches, and the ease with which the child
walked with crutches to the radiology department and back
suggest that compliance is good. We noted these features in
this study too, and based on them, we could not identify any
child who was obviously noncompliant.
Our primary outcome measure was a quantitative
radiographic estimate of the shape of the femoral head.
We did not assess the functional outcome. However, a
clear association between functional outcomes and the
shape of the femoral head has been demonstrated.33
Finally, we did not evaluate the functional and
psychological effects of a prolonged period of weight re-
lief, but we are currently collecting data on these issues. In
the meanwhile, recent reports suggest that prolonged
weight relief may not have adverse psychological and
functional effects.1,34

e148 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
FIGURE 4. Non–weight-bearing with the aid of axillary
crutches.
CONCLUSION
The study confirms that containment by PFVO per-
formed early in the course of LCPD combined with weight
relief till the disease has evolved to Stage IIIb is likely to
result in spherical hips in 75% of children. Reducing the
period of weight relief to 6 months may yield significantly
poorer results, with only half the hips retaining their
sphericity.
ACKNOWLEDGMENTS
The authors thank Dr Vasudev Guddettu for the
statistical analysis of the data.
REFERENCES
1. Do DH, McGuire MF, Jo CH, et al. Weightbearing and activity
restriction treatments and quality of life in patients with perthes
disease. Clin Orthop Relat Res. 2021;479:1360–1370.
2. Mintowt-Czyz W, Tayton K. Indication for weight relief and containment
in the treatment of Perthes’ disease. Acta Orthop Scand. 1983;54:439–445.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Prolonged weight-relief in Perthes disease
3. Iwamoto M, Nakashima Y, Nakamura T, et al. Clinical outcomes of
conservative treatment with a non-weight-bearing abduction brace
for Legg-Calvé-Perthes disease. J Orthop Sci. 2018;23:156–160;
PMID: 28982606.
4. Kim WC, Hosokawa M, Tsuchida Y, et al. Outcomes of new pogo-
stick brace for Legg-Calvé-Perthes’ disease. J Pediatr Orthop B.
2006;15:98–103; PMID: 16436943.
5. Joseph B, Srinivas G, Thomas R. Management of Perthes disease of
late onset in southern India. The evaluation of a surgical method. J
Bone Joint Surg Br. 1996;78:625–630; PMID: 8682832.
6. Hyman JE, Trupia EP, Wright ML, et al. International Perthes
Study Group Members. Interobserver and intraobserver reliability of
the modified Waldenström classification system for staging of Legg-
Calvé-Perthes disease. J Bone Joint Surg Am. 2015;97:643–650.
7. Joseph B, Varghese G, Mulpuri K, et al. Natural evolution of Perthes
disease: a study of 610 children under 12 years of age at disease onset.
J Pediatr Orthop. 2003;23:590–600.
8. Ailabouni R, Zomar BO, Slobogean BL, et al. The Natural History
of Non-operatively Managed Legg-Calvé-Perthes’ Disease. Indian J
Orthop. 2022;56:867–873.
9. Salter RB, Thompson GH. Legg-Calvé-Perthes disease. The
prognostic significance of the subchondral fracture and a two-group
classification of the femoral head involvement. J Bone Joint Surg
Am. 1984;66:479–489.
10. Singh KA, Shah H, Joseph B, et al. Evolution of Legg-Calvé-Perthes
disease following proximal femoral varus osteotomy performed in
the avascular necrosis stage:a prospective study. J Child Orthop.
2020;14:58–67.
11. Shah H, Siddesh ND, Pai H, et al. Quantitative measures for
evaluating the radiographic outcome of Legg-Calvé-Perthes disease.
J Bone Joint Surg Am. 2013;95:354–361.
12. Siddesh ND, Shah H, Tercier S, et al. The sphericity deviation score:
a quantitative radiologic outcome measure of Legg-Calvé Perthes
disease applicable at the stage of healing and at skeletal maturity.
J Pediatr Orthop. 2014;34:522–528.
13. Lauritzen J. Legg-Calvé-Perthes disease. Acta Orthop Scand.
1971;42:456.
14. Lauritzen J. Legg-Calvé-Perthes disease – A comprehensive study.
Acta Orthop Scand. 1975;(suppl 159):26–117.
15. Kim HK, Aruwajoye O, Stetler J, et al. Effects of non-weight-
bearing on the immature femoral head following ischemic osteonec-
rosis: an experimental investigation in immature pigs. J Bone Joint
Surg Am. 2012;94:2228–2237.
16. Wagenaar FB, Maathuis CG, van Erve RH. Treatment outcome in
the most severely affected Legg-Perthes patients, comparing pro-
longed traction in abduction with femoral varus derotation treat-
ment. J Child Orthop. 2011;5:89–95.
17. Peck JB, Greenhill DA, Morris WZ, et al. Prolonged non-
weightbearing treatment decreases femoral head deformity compared
to symptomatic treatment in the initial stage of Legg-Calvé-Perthes
disease. J Pediatr Orthop B. 2022;31:209–215.
18. Shah H, Singh KA, Joseph B. The “Discoid Epiphysis”-An
Uncommon Presentation of Legg-Calvé-Perthes Disease. J Pediatr
Orthop. 2022;42:e570–e576.
19. Aarvold A, Lohre R, Chhina H, et al. Dynamic deformation of the
femoral head occurs on weightbearing in Legg-Calves-Perthes
disease: a translational pilot study. Bone Jt Open. 2020;1:364–369.
20. Stulberg SD, Cooperman DR, Wallensten R. The natural history of
Legg-Calvé-Perthes disease. J Bone Joint Surg Am. 1981;63:
1095–1108.
21. Kim S, Oh HS, Lim JH, et al. Results of early proximal femoral
osteotomy at skeletal maturity in legg-Calvé-Perthes disease:
Implication for the bypass of fragmentation stage. J Pediatr Orthop.
2021;41:e768–e773.
22. Mosow N, Vettorazzi E, Breyer S, et al. Outcome after combined
pelvic and femoral osteotomies in patients with Legg-Calvé-Perthes
Disease. J Bone Joint Surg Am. 2017;99:207–213.
23. Citlak A. Long-term follow-up results of femoral varus osteotomy in
the treatment of Perthes disease, and comparison of open-wedge and
closed-wedge osteotomy techniques: A retrospective observational
study. Medicine (Baltimore). 2020;99:e19041.
www.pedorthopaedics.com | e149
Shah et al
24. Kamegaya M, Saisu T, Ochiai N, et al. A paired study of Perthes’
disease comparing conservative and surgical treatment. J Bone Joint
Surg Br. 2004;86:1176–1181.
25. Saran N, Varghese R, Mulpuri K. Do femoral or salter innominate
osteotomies improve femoral head sphericity in Legg-Calvé-Perthes
disease? A meta-analysis. Clin Orthop Relat Res. 2012;470:2383–2393.
26. Castañeda P, Haynes R, Mijares J, et al. Varus-producing osteotomy
for patients with lateral pillar type B and C Legg-Calvé-Perthes disease
followed to skeletal maturity. J Child Orthop. 2008;2:373–379.
27. Carsi B, Judd J, Clarke NM. Shelf acetabuloplasty for containment
in the early stages of Legg-Calve-Perthes disease. J Pediatr Orthop.
2015;35:151–156.
28. Yoo WJ, Choi IH, Cho TJ, et al. Shelf acetabuloplasty for children with
Perthes’ disease and reducible subluxation of the hip: prognostic factors
related to hip remodelling. J Bone Joint Surg Br. 2009;91:1383–1387.
29. Sponseller PD, Desai SS, Millis MB. Comparison of femoral and
innominate osteotomies for the treatment of Legg-Calvé-Perthes
disease. J Bone Joint Surg Am. 1988;70:1131–1139.
e150 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
30. Shohat N, Copeliovitch L, Smorgick Y, et al. The long-term outcome
after varus derotational osteotomy for Legg-Calvé-Perthes Disease:
A mean follow-up of 42 Years. J Bone Joint Surg Am. 2016;98:
1277–1285.
31. Shah H, Siddesh ND, Joseph B, et al. Effect of prophylactic
trochanteric epiphyseodesis in older children with Perthes’ disease.
J Pediatr Orthop. 2009;29:889–895.
32. Tercier S, Shah H, Siddesh ND, et al. Does proximal femoral
varus osteotomy in Legg-Calvé-Perthes disease predispose
to angular mal-alignment of the knee? A clinical and radio-
graphic study at skeletal maturity. J Child Orthop. 2013;7:
205–211.
33. Hailer YD, Haag AC, Nilsson O. Legg-Calvé-perthes disease:
quality of life, physical activity, and behavior pattern. J Pediatr
Orthop. 2014;34:514–521.
34. Roposch A. CORR Insights: weightbearing and activity restriction
treatments and quality of life in patients with Perthes Disease. Clin
Orthop Relat Res. 2021;479:1371–1372.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Compression Screw Fixation for Pediatric Olecranon
Fractures
Frederick Mun, BS, Krishna V. Suresh, MD, Brooke Hayashi, DO, Adam Margalit, MD,
Paul D. Sponseller, MD, and R. Jay Lee, MD
Background: Pediatric olecranon fractures can be treated with
several methods of fixation. Though postoperative outcomes of
various fixation techniques, including cannulated intramedullary
screws, have been described in adults, functional and radiographic
outcomes of screw fixation in pediatric patients are unclear. In this
study, we assessed clinical, radiographic, functional, and patient-
reported outcomes of pediatric olecranon fractures treated with
compression screw fixation.
Methods: We retrospectively identified 37 patients aged 16 years
or younger with a total of 40 olecranon fractures treated with
screw fixation at our level-1 trauma center between April 2005
and April 2022. From medical records, we extracted data on
demographic characteristics, time to radiographic union, range
of elbow motion at final follow-up, and complications during the
follow-up period. Patient-reported outcomes were evaluated us-
ing the Quick Disabilities of the Arm, Shoulder, and Hand and
Patient-Reported Outcomes Measurement Information System
Pediatric Upper Extremity Short Form 8a measures.
Results: There were no malunions or nonunions at the final mean
follow-up of 140 days (range, 26 to 614 d). Four patients had
implant failure (11%), of whom 3 experienced fracture union with
no loss of fixation or need for revision surgery. One patient un-
derwent a revision for fracture malreduction. Screw prominence
was documented in 1 patient. Instrumentation was removed at our
institution for 33 of 40 fractures. Mean time to radiographic union
was 53 days (range, 20 to 168 d). Postoperative range of motion at
the most recent follow-up visit showed a mean extension deficit of
6 degrees (range, 0–30 degrees) and mean flexion of 134 degrees
(range, 60–150 degrees). At the final follow-up, the mean ( ± SD)
Quick Disabilities of the Arm, Shoulder, and Hand score was
4.2 ± 8.0, and the mean Patient-Reported Outcomes Measurement
Information System score was 37 ± 1.5, indicating good function
and patient satisfaction.
Conclusions: All 37 patients in our series had excellent radio-
graphic, functional, and patient-reported outcomes after screw
fixation. We observed no cases of nonunion or malunion, growth
disturbance, or refracture. These results suggest that screw fixation
is a safe and effective option for pediatric olecranon fractures.
From the Department of Orthopaedic Surgery, The Johns Hopkins
University, Baltimore, MD.
The authors declare no conflicts of interest.
Reprints: R. Jay Lee, MD, Department of Orthopaedic Surgery, The
Johns Hopkins Hospital, 1800 Orleans Street, Baltimore, MD 21287.
E-mail: editorialservices@jhmi.edu.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002284
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Level of Evidence: Level IV, case series.
Key Words: case series, compression screw fixation, pediatric
olecranon fracture, Quick Disabilities of the Arm, Shoulder, and
Hand, Patient-Reported Outcomes Msasurement Information
System
(J Pediatr Orthop 2023;43:e151–e156)
A lthough olecranon fractures account for only 5% of all
pediatric elbow fractures, they can cause major func-
tional impairment if managed inappropriately.1–3 Most of
these fractures are minimally displaced, with <2 mm of
translation, and can be treated nonoperatively with closed
reduction and casting. However, fractures with > 2 mm of
displacement may require surgical intervention to prevent
ulnohumeral joint stiffness, ulnar nerve injury, or persis-
tent deformity.4 Therefore, the identification of safe and
effective methods of surgical fixation is needed.
Surgical treatment of pediatric olecranon fractures
has been reported to have poor long-term outcomes in
7.6% of patients (including decreased range of motion
[ROM], chronic pain, muscular atrophy, neurological
impairment, and joint degeneration).4 Surgical fixation
techniques include tension band technique using wire or
suture fixation, intramedullary screws, and plate
fixation.5–8 Though outcomes after compression screw
fixation (herein, “screw fixation”) for olecranon fractures
have been well described in adults, much less is known
about such outcomes in the pediatric population.9–11
The purpose of this study was to determine clinical,
radiographic, functional, and patient-reported outcomes of
pediatric olecranon fractures treated with screw fixation.
METHODS
Patient Selection
This study received institutional review board
approval. We used Current Procedural Terminology code
24675 to identify olecranon fracture fixation performed at
our level-1 pediatric trauma center between April 2005
and April 2022. The medical records and radiographs of
all operatively treated olecranon fractures were reviewed.
All fractures were treated by 1 of 8 fellowship-trained
pediatric orthopaedic surgeons. Mechanism of injury was
categorized as direct trauma (eg, fall onto elbow, motor
www.pedorthopaedics.com | e151
Mun et al J Pediatr Orthop  Volume 43, Number 2, February 2023

vehicle accident, gunshot wound, direct contact with an

TABLE 2. Surgical Characteristics of 40 Pediatric Olecranon
object), indirect trauma (eg, eccentric contraction of tri- Fractures Treated With Screw Fixation at a Single Center
ceps muscle), or unknown. We excluded patients with Between April 2005 and April 2022
extra-articular ulnar fractures, fractures treated with ten-
Characteristic N (%)
sion band wiring or plate fixation, patients older than
16 years, and patients lost to follow-up before fracture Method of reduction
Closed/percutaneous 15 (37)
union. We reviewed medical records for 70 patients, Open 25 (62)
of whom 38 were treated with screw fixation. One Implant size, mm
patient was lost to follow-up, leaving 37 patients with 4.0 8 (20)
40 olecranon fractures for analysis. 4.5 24 (60)
6.5 8 (20)
Study Participants Fixation type
Intramedullary screw 36 (90)
In the 37 patients in our cohort, olecranon fractures Bicortical screw 4 (10)
were bilateral in 3 patients, for a total of 40 fractures (24 Instrumentation type
in the left arm and 16 in the right). Eleven patients were Partially threaded screws 32 (80)
female and 26 were male, with a mean age of 9 years Fully threaded screws 4 (10)
Multiple screws 1 (2.5)
(range, 3 to 16 y). Only 2 patients had been diagnosed Tension band wire + screw 1 (2.5)
with osteogenesis imperfecta, 1 of whom had bilateral K-wire + screw 1 (2.5)
fractures. Of the 40 fractures, 21 were type IIA meta- Mini plate + screw 1 (2.5)
physeal fractures, 8 were type IA metaphyseal fractures, Screw with washer 8 (20)
Intraoperative arthrography 10 (25)
7 were type IIIA metaphyseal fractures, 3 were apo-
physeal sleeve fractures, and 1 was a ballistic apophyseal
fracture with comminution. Thirty-nine fractures were Designation A refers to the absence of comminution, and
caused by direct trauma, and the mechanism of injury B refers to the presence of comminution. Type I represents
was unknown for the remaining fracture (Table 1). Of nondisplaced fractures; type II represents displaced frac-
the 8 fractures classified as type IA, 1 was treated as a tures with a stable ulnohumeral joint, and type III repre-
symptomatic nonunion. Seven patients had an associated sents displaced fractures with an unstable ulnohumeral
radial neck fracture, and 16 patients had associated joint. Typically, types IIA, IIB, IIIA, and IIIB require
elbow injuries at the time of fracture fixation. Table 2 operative fixation.
summarizes the method of reduction, implant sizes,
instrumentation type, and use of washers. The mean time
from the index procedure to the final follow-up was Surgical Technique
140 days (range, 26 to 614 d). All patients were placed supine on the operating table.
Closed reduction was achieved through manual manipu-
Radiographic Review lation or percutaneous reduction. If percutaneous reduction
A pediatric orthopaedic surgery clinical fellow re- was inadequate, open reduction was performed. A cannu-
viewed preoperative radiographs of all included patients. lated screw was then placed to provide compression and
Fractures were categorized by location (apophyseal avul- adequate fixation along the intramedullary canal or bi-
sions or metaphyseal intra-articular fractures) and by the cortically across the fracture site. Operative decision-
Mayo classification for olecranon fractures.12 In this making, including a selection of implants and operative
classification system, olecranon fractures are categorized techniques, was at the discretion of the surgeons. However,
as type I, II, or III, along with an A or B designation.12 it is the preference of the senior author to use larger-
diameter intramedullary screws for transverse fractures
TABLE 1. Demographic and Injury Characteristics of 37 (Fig. 1). Screw diameter and length are selected to achieve
Pediatric Patients With 40 Olecranon Fractures who sufficient intramedullary thread purchase (Fig. 2). Rotation
Underwent Screw Fixation at a Single Center Between April during screw insertion can be prevented by using a second
2005 and April 2022 derotation wire. Intramedullary screws can be used with
Characteristic N (%) oblique fractures, but care must be taken that final
compression across the oblique fracture does not also
Female sex 11 (30) cause translation and loss of reduction. It is the preference
Age, y 8.9 ± 4.2*
Injury location of the senior author to use smaller-diameter bicortical
Left 24 (60)† screws for oblique fractures, in particular where its dorsal to
Right 13 (32)† volar trajectory may position the screw close to
Bilateral 3 (7.5)† perpendicular to the fracture line (Fig. 3). Care must be
Mechanism of injury
Direct trauma 39 (97)†
taken to avoid violation of the anterior neurovascular
Unknown 1 (2.5)† structures with bicortical guide pin placement and drilling
(Fig. 4). Most patients (36 of 40) were treated with
*Presented as mean ± SD.
†Calculated using a denominator of 40 fractures. compression screw fixation alone, but 4 patients required
supplemental fixation to strengthen construct stability.

e152 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
FIGURE 1. Anterior (A) and lateral (B) radiographs of a 4-year-old boy’s left elbow showing a type IIA metaphyseal intra-articular
olecranon fracture.
Postoperative Management
After surgery, the arm was immobilized in a splint or
cast for ~4 to 5 weeks. Postoperative radiographs were
obtained until adequate fracture healing was seen. After
immobilization was removed, ROM was allowed as tol-
erated. We recommended that all patients have in-
strumentation removed after the fracture healed to
minimize growth disturbance and discomfort.
Outcome Measures
Postoperative Radiographic and Functional Outcomes
The presence of nonunion or malunion was de-
termined by the senior author, a fellowship-trained pe-
diatric orthopaedic surgeon, for each patient during the
follow-up period. We defined time to fracture union as the
time between operative fixation and radiographic presence
of tricortical fracture bridging without lucency.13 Non-
union was defined as the absence of radiographic union by
6 months. Malunion was defined as malalignment of
subchondral bone (“step-off”) or a gap in concentricity
(contour defect) of the sigmoid cavity and center of the
trochlea.14 At the final follow-up for each patient, the
range of elbow extension and flexion was recorded.
FIGURE 2. Anterior (A) and lateral (B) radiographs of the same 4-year-old boy as shown in Figure 1 at 4 months after compression
screw fixation with a 6.5-mm screw, showing evidence of healing.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Screw Fixation for Pediatric Olecranon Fractures
Patient-Reported Outcomes
Patient-reported outcomes were evaluated using the
Quick Disabilities of the Arm, Shoulder, and Hand
(qDASH) score and Patient-Reported Outcomes Meas-
urement Information System (PROMIS) Pediatric Upper
Extremity Short Form 8a score. qDASH scores range
from 0 (no disability) to 100 (most severe disability). The
PROMIS Pediatric Upper Extremity Short Form 8a score
ranges from 8 (1 point for each activity of daily living the
patient is unable to do) to 38 (able to do the activity of
daily living with no difficulty). Surveys were administered
through telephone. If patients or their respective care-
givers did not answer the telephone, we made 3 additional
attempts to contact them to administer the surveys.
Rates of Instrumentation Removal and Complications
For each patient, we recorded whether instrumentation
removal was performed. Time to instrumentation removal
was calculated as the difference between the dates of the
index procedure and instrumentation removal. Complica-
tions were recorded during the follow-up period, including
revision surgery, screw prominence, screw backout, surgical
site or instrumentation infection, implant irritation, and
www.pedorthopaedics.com | e153
Mun et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 3. Anterior (A) and lateral (B) radiographs of a 6-year-old boy’s left elbow showing a type IIA metaphyseal intra-articular
olecranon fracture.

postoperative manipulation under anesthesia. Sixty-two
percent of fractures required open reduction.
Statistical Analysis
Data were analyzed using Excel, version 2020, soft-
ware (Microsoft Corp). Descriptive statistics were calculated
for demographic characteristics, ROM, complications, and
patient-reported outcome scores.
RESULTS
Clinical Outcomes
There were no malunions or nonunions at the final
mean follow-up of 140 days (range, 26 to 614 d).
Instrumentation was removed at our institution for 33 of
40 fractures (82%). Four patients had implant-related
complications (11%). Of those 4 patients, 1 had residual
intra-articular displacement on radiographs taken 8 days
after the index procedure. The patient underwent revision
surgery to place another intramedullary screw and experi-
enced radiographic union 42 days after revision surgery. The
remaining 3 patients had instrumentation-related compli-
cations; 1 patient had a screw “backout,” and 2 patients had
screw heads penetrate into the proximal fracture fragment.
These complications did not affect fracture reduction or
healing and did not result in surgical revision. Another pa-
tient experienced instrumentation irritation from the tension
band and intramedullary screw construct, which was re-
moved. The patient had no skin or soft-tissue breakdown
due to instrumentation. Two patients underwent manipu-
lation under anesthesia at the time of screw removal. Both
patients achieved a functional range of elbow motion at the
final follow-up. We observed no postoperative infections or
refractures during the follow-up period.

FIGURE 4. Anterior (A) and lateral (B) radiographs of the same 6-year-old boy as shown in Figure 3 at 1 month after bicortical
compression screw fixation with a 4.5-mm screw, showing evidence of healing.

e154 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Screw Fixation for Pediatric Olecranon Fractures

Radiographic Outcomes
Fracture healing was first observed, on average, by
53 days (range, 20 to 168 d) after surgery. All fractures
had achieved union at the final follow-up.
Functional Outcomes
At the final follow-up, the elbow ROM showed a
mean ( ± SD) extension deficit of 6 degrees ± 9 degrees and
flexion of 134 degrees ± 21 degrees.
Patient-Reported Outcomes
Of 37 patients, we obtained qDASH and PROMIS
scores for 29. Seven patients did not answer any of the 3
call attempts, and 1 patient declined to participate. Mean
qDASH score was 4.2 ± 8.0, and mean PROMIS Pediatric
Upper Extremity score was 37 ± 1.5 (Table 3).
DISCUSSION
Screw fixation of pediatric olecranon fractures
produced excellent radiographic, functional, and patient-
reported outcomes. The postoperative complication rate
through final follow-up was low, and most patients re-
turned for recommended instrumentation removal. Our
findings suggest that screw fixation is an effective and safe
option for treating these fractures.
Screw fixation may have several advantages over
other operative techniques. Tension band wire and suture
constructs involve open incisions, placement of parallel
K-wires, and steel wires or sutures in a figure-of-8 tension
band configuration to provide compression across the
fracture site with the eccentric triceps function.15,16 Al-
though tension band fixation can provide articular com-
pression and enable early elbow motion, painful implants
and soft-tissue irritation have been reported in up to 88%
of patients.17 In our series, only 1 patient reported soft-
tissue irritation from prominent instrumentation. In ad-
dition, open incisions are needed to create bone tunnels for
tension band wire and suture constructs. In our series, 37%
of patients were treated with percutaneous reduction and
TABLE 3. Outcomes at Latest follow-up for 37 Pediatric
Patients With Olecranon Fractures Treated With Screw Fixation
at a Single Center Between April 2005 and April 2022
Outcome Mean ± SD
Range of elbow motion
Extension, degrees 5.9 ± 8.9
Flexion, degrees 134 ± 21
Complications
Revision surgery — 1
Instrumentation-related
Irritation — 1
Loss of fixation — 1
Screw migration — 1
Screw prominence — 2
Patient-reported outcomes
qDASH score 4.2 ± 8.0
PROMIS score 37 ± 1.5
PROMIS indicates Patient-Reported Outcomes Measurement Information
System Pediatric Upper Extremity-Short Form 8a; qDASH, Quick Disabilities of
the Arm, Shoulder, and Hand.
screw fixation, eliminating the need for extensive incisions.
Another common postoperative complication is elbow
stiffness (typically <20 degrees loss of elbow extension).5
In our study, we noted a mean return of motion of 6
degrees of extension to 134 degrees of flexion.
In our series, the rate of implant loosening or failure
that required subsequent surgical revision was low. A
previous study of 18 pediatric patients with osteogenesis
imperfecta who were treated with screw fixation reported
implant loosening in 5 patients.18 The high failure rate was
likely related to poor bone quality, confirmed by low
Z-scores before fracture. In our series, implant failure
occurred in 4 patients (11%). Two patients had screw
penetration through the proximal apophysis and bony
fragment. Both patients had fixation without washers and
a screw diameter of 6.5 mm, which are factors that could
have contributed to implant penetration. One patient with
screw loosening and migration did achieve bicortical fix-
ation as seen on intraoperative fluoroscopy; however, later
follow-up imaging showed instrumentation migration. All
3 patients achieved union with no loss of fixation or need
for revision surgery. Only 1 patient required revision
surgery because of malreduction 8 days after percutaneous
fixation. The patient subsequently underwent revision
surgery with another intramedullary screw that provided
adequate stabilization and fracture union at 42 days.
Corradin et al5 analyzed data from 12 pediatric pa-
tients with olecranon fractures who did not have osteo-
genesis imperfecta and were treated with either closed or
open reduction and screw fixation. Results were sat-
isfactory, with good fracture reduction and 1 minor
complication of loss of elbow extension that was clinically
inconsequential. They reported no instrumentation irrita-
tion, secondary displacement, refractures, delays in the
union, or growth arrest. Persiani et al18 analyzed data
from 10 pediatric patients with osteogenesis imperfecta
who were treated with screw fixation. At the 3-year follow-
up, patients had a full return of forearm ROM, no elbow
angular deviation, and mean loss of elbow extension of 15
degrees. Although excluded from their study, 2 patients
treated with screw fixation had implant loosening that
required revision screw fixation. Another 2 patients ex-
cluded from their study had refractures with implant mi-
N gration after the loss of screw fixation requiring revision.
We observed no refractures after instrumentation removal
in our series. Persiani et al18 and Gwynne-Jones et al19
—
— reported refracture rates of 17% to 25% after in-
strumentation removal in children with osteogenesis im-
perfecta. However, they noted that none of the patients
without osteogenesis imperfecta sustained a refracture
after instrumentation removal.
Screw fixation in pediatric olecranon fractures may
not cause clinically relevant growth restriction. Grimm
et al20 observed no major growth disturbance during more
— than 20 years of clinical practice, reflecting previous studies
—
that have shown olecranon shortening without functional
deficits in 0% to 11% of cases.6,14 After ~age 9, the
ossification core of the olecranon apophysis contributes a
very small percentage of the total longitudinal growth of

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e155

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Mun et al
the physis.18,21 On the basis of this reasoning, the risk of
growth restriction secondary to fracture or instrument fix-
ation may have been possible in ~half of our patients. (The
mean patient age in our cohort was 8.9 y) However, we saw
no evidence of growth disturbance during the follow-up
period in our patients treated with screw fixation.
This study has limitations inherent to its retro-
spective design. The mean duration of follow-up was
140 days, which may limit the generalizability of our
results. Another limitation was the use of qDASH and
PROMIS scores as measures of function in patients with
preexisting medical conditions. Although Hung et al22
reported that qDASH and PROMIS scores are re-
sponsive to functional changes after upper extremity in-
jury, a major limitation of these measures is that they do
not consider the baseline level of function. Three patients
(2 with osteogenesis imperfecta and 1 with osteochon-
dritis dissecans) reported a complete return to pre-
operative functional level but still had elevated qDASH
and decreased upper extremity PROMIS scores because
of their underlying musculoskeletal conditions. Despite
these limitations, to our knowledge, this is the largest
case series of olecranon fractures treated with screw fix-
ation. A future long-term follow-up study comparing the
ROM between patients who underwent fixation with
different techniques may further elucidate their advan-
tages and disadvantages.
CONCLUSION
Screw fixation of displaced pediatric olecranon
fractures resulted in excellent radiographic, functional, and
patient-reported outcomes. In 40 fractures, we observed no
cases of nonunion or malunion, growth disturbance, or
refracture, and only 1 patient required revision surgery for
malreduction. Range of elbow extension and flexion re-
turned nearly to baseline measures, and patient-reported
outcomes after this procedure were positive. These results
suggest that screw fixation is a safe and effective treatment
option for pediatric olecranon fractures.
REFERENCES
1. Grantham SA, Kiernan HA. Displaced olecranon fracture in
children. J Trauma. 1975;15:197–204.
2. Matthews JG. Fractures of the olecranon in children. Injury.
1980;12:207–212.
3. Papavasiliou VA, Beslikas TA, Nenopoulos S. Isolated fractures of
the olecranon in children. Injury. 1987;18:100–102.
e156 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
4. Caterini R, Farsetti P, D’Arrigo C, et al. Fractures of the olecranon
in children. Long-term follow-up of 39 cases. J Pediatr Orthop B.
2002;11:320–328.
5. Corradin M, Marengo L, Andreacchio A, et al. Outcome of isolated
olecranon fractures in skeletally immature patients: comparison of
open reduction and tension band wiring fixation versus closed
reduction and percutaneous screw fixation. Eur J Orthop Surg
Traumatol. 2016;26:469–476.
6. Evans MC, Graham HK. Olecranon fractures in children: Part 1:
a clinical review; Part 2: a new classification and management
algorithm. J Pediatr Orthop. 1999;19:559–569.
7. Gaddy BC, Strecker WB, Schoenecker PL. Surgical treatment of
displaced olecranon fractures in children. J Pediatr Orthop. 1997;
17:321–324.
8. Graves SC, Canale ST. Fractures of the olecranon in children: long-
term follow-up. J Pediatr Orthop. 1993;13:239–241.
9. Ahmed AR, Sweed T, Wanas A. The role of cancellous screw with
tension band fixation in the treatment of displaced olecranon fractures,
a comparative study. Eur J Orthop Surg Traumatol. 2008;18:571.
10. Helm RH, Hornby R, Miller SW. The complications of surgical
treatment of displaced fractures of the olecranon. Injury. 1987;18:48–50.
11. Wadsworth TG. Screw fixation of the olecranon. Lancet. 1976;2:
1118–1119.
12. Cabanela M, Morrey B. Part IV: Acute Trauma. In: Morrey BF,
ed. The Elbow and Its Disorders. Philadelphia, PA: WB Saunders;
1993:417.
13. DiSilvio F, Foyil S, Schiffman B, et al. Long bone union accurately
predicted by cortical bridging within 4 months. JB JS Open Access.
2018;3:e0012.
14. Gicquel PH, De Billy B, Karger CS, et al. Olecranon fractures in 26
children with mean follow-up of 59 months. J Pediatr Orthop. 2001;
21:141–147.
15. Perkins CA, Busch MT, Christino MA, et al. Olecranon fractures in
children and adolescents: outcomes based on fracture fixation.
J Child Orthop. 2018;12:497–501.
16. Wilkerson JA, Rosenwasser MP. Surgical techniques of olecranon
fractures. J Hand Surg Am. 2014;39:1606–1614.
17. Flinterman HJA, Doornberg JN, Guitton TG, et al. Long-term
outcome of displaced, transverse, noncomminuted olecranon frac-
tures. Clin Orthop Relat Res. 2014;472:1955–1961.
18. Persiani P, Ranaldi FM, Graci J, et al. Isolated olecranon fractures
in children affected by osteogenesis imperfecta type I treated with
single screw or tension band wiring system: Outcomes and pitfalls in
relation to bone mineral density. Medicine. 2017;96:e6766.
19. Gwynne-Jones DP. Displaced olecranon apophyseal fractures in
children with osteogenesis imperfecta. J Pediatr Orthop. 2005;25:
154–157.
20. Grimm N, Herman M. Pediatric Olecranon Fractures. In: Abzug
JM, Herman MJ, Kozin S, eds. Pediatric Elbow Fractures: A
Clinical Guide to Management. Cham, Switzerland: Springer
International Publishing; 2018:151–167.
21. Pritchett JW. Growth plate activity in the upper extremity. Clin
Orthop Relat Res. 1991;268:235–242.
22. Hung M, Saltzman CL, Greene T, et al. The responsiveness of the
PROMIS instruments and the qDASH in an upper extremity
population. J Patient Rep Outcomes. 2017;1:12.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Complication Rates in Patients With Classic
and Radiographic Variants of Seymour Fractures
Rami P. Dibbs, BA,*† Thomas W. Mitchell, MD,*‡ Rita E. Baumgartner, MD,*‡
John C. Koshy, MD,*† and Bryce R. Bell, MD*‡
Background: At a tertiary-care, level 1 pediatric trauma center, we
have observed fractures of the distal phalanx involving the physis,
with associated nail bed injuries, that are distinct from the classic
description of the Seymour fracture. We investigated the time to
definitive management and the associated morbidity of these
Seymour fracture variants compared with classically described
Seymour fractures. We hypothesize that these Seymour variants
are similarly problematic in terms of complications and delays to
the definitive treatment and thus warrant increased awareness.
Methods: A retrospective chart review was performed of all
patients with distal phalanx fractures involving the physis and
associated nail bed injuries that were treated with operative
intervention at a single pediatric specialty institution over a
9-year period. Radiographs and clinical photographs were re-
viewed to determine if the patient presented with a classic
Seymour fracture or variant. Primary outcomes included time
from injury to definitive treatment and complication rate.
Results: Of the 66 Seymour fractures identified in the chart re-
view, 36 (55%) were identified as classic Seymour fractures and
30 (45%) were identified as variants. The mean time to operative
intervention in the classic and variant groups was 7.3 versus
12.7 days (P = 0.216). The complication rates in the classic and
variant groups were 11.1% versus 23.3% (P = 0.185), with in-
fections accounting for nearly all complications identified.
Overall infection rates for the classic and variant cohorts were
8.3% and 20.0% (P = 0.169), respectively, with the majority
presenting preoperatively (5.6% vs. 13.3%, P = 0.274).
Conclusions: We found that patients with classic Seymour fractures
or radiographic variants had statistically similar incidence rates,
complication rates, and delays in treatment, with a trend towards
higher complication rates and delayed time to treatment in patients
with variant-type injuries. We propose a minor expansion of the
definition of Seymour fractures to include common variants to
increase awareness of these problematic injuries, minimize delays in
treatment, and decrease complications.
Level of Evidence: Level III; Retrospective Comparative Study.
From the *Michael E. DeBakey Department of Surgery, Baylor College of
Medicine; †Division of Plastic Surgery, Department of Surgery; and
‡Division of Orthopedics, Department of Surgery, Texas Children’s
Hospital, Houston, TX.
The authors declare no conflicts of interest.
Reprints: Bryce R. Bell, MD, 18200 W. Katy Freeway Suite 520,
Houston, TX, 77094. E-mail: brbell@texaschildrens.org.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002306
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Key Words: Seymour fracture, distal phalanx fracture, nail bed
injury
(J Pediatr Orthop 2023;43:e157–e162)
T he hand is the most commonly injured part of the body
in the pediatric population, with a significant fraction
of these injuries occurring between 5 and 14 years of
age.1–3 Originally described by N. Seymour in 1966, Sey-
mour fractures were defined as distal phalangeal fractures
involving or near the physis and included Salter-Harris
Type I fractures, Salter-Harris Type II fractures, and
metaphyseal fractures 1 to 2 mm distal to the epiphyseal
plate.4 It was not until later that nail bed injuries were
observed to often occur in conjunction with the distal
phalanx fracture. The nail bed laceration frequently results
in the proximal nail plate being extruded from the prox-
imal nail fold. Consequently, this is considered to be an
open fracture.5–7
The clinical deformity present in a Seymour fracture
resembles a mallet finger presenting with the typical flex-
ion deformity at the fracture site. This is due to the ana-
tomically distinct locations of the extensor and flexor
tendon insertion points. The extensor tendon inserts into
the proximal epiphysis of the distal phalanx whereas the
flexor digitorum profundus tendon inserts distally onto the
metaphysis.4,5,8 Unlike a mallet injury, however, the ex-
tensor tendon frequently remains intact following the
Seymour fracture.4,9
Acute treatment of a classic Seymour fracture is
recommended to prevent complications such as chronic
osteomyelitis, nail dystrophy, and physeal arrest.7,10,11
Treatment typically involves nail plate removal, irrigation
and debridement, fracture reduction (with the removal of
interposed tissue), closure of the nail bed laceration, and a
period of postoperative immobilization.11–14
Although Seymour fractures are open fractures
with an associated high risk of complications and war-
rant expedient intervention, they remain under-reported
and can go unrecognized.5,12 Early treatment of Sey-
mour fractures demonstrates improved outcomes when
compared with delayed or inadequate treatment, which
can result in higher rates of infection, including early
wound infections and delayed osteomyelitis.13 Con-
sequently, accurate diagnosis and early management of
Seymour fractures can optimize patient outcomes. At
www.pedorthopaedics.com | e157
Dibbs et al
our institution, we have observed several Seymour
fracture variants and believe their atypical features can
contribute to their delayed diagnosis and treatment.
Therefore, we aim to identify and describe the most
common Seymour fracture variants based on radio-
graphic and clinical findings and to investigate the
treatment outcomes of these variants at our institution.
METHODS
After institutional review board approval, a retro-
spective chart review of all distal phalangeal fractures at
a Level 1 pediatric trauma center was conducted from
January 2012 to January 2021. A query for Seymour
fractures was undertaken using Current Procedural
Terminology (CPT) codes 11760, 11762, 26750, 26755,
26756, and 26765. CPT codes included those involving
any treatment of distal phalangeal fractures and repair of
nail bed injuries. Patients under 18 years old who were
treated for distal phalangeal fractures involving the
physis with associated nail bed injuries were included,
and a detailed chart review was performed. Patients were
excluded if they did not have appropriate radiographs, if
they had a distal phalanx fracture that did not involve
the physis (e.g., tuft fractures, skeletal maturity), or if
they were part of a more complex injury pattern (e.g.,
near amputation). Demographic information obtained
included age, gender, race, and ethnicity. Additional
variables included the cause of injury, time from the
injury to initial presentation and definitive treatment,
location of definitive intervention (e.g., bedside, or
operating room), type of intervention, type of antibiotic
administered (if any), and postoperative complications.
In addition, preoperative anteroposterior (AP), oblique
and lateral radiographs, and any clinical photographs
were evaluated.
Patients were grouped into 2 cohorts: the classic
Seymour fracture group and the Seymour fracture var-
iant group. All patients in both the groups had injuries
involving the distal phalanx physis and overlying nail bed
or nail fold injuries, thus making them consistent with
open physeal fractures. The classic Seymour fracture was
defined as a transverse fracture through the metaphyseal
region of the distal phalanx with an associated nail bed
laceration involving the base of the nail and/or the nail
fold and possible bone exposure on clinical exam
(Fig. 1).6 Alternatively, the Seymour fracture variant
group comprised of patients with atypical fracture
patterns with associated nail bed injuries. Patients were
subcategorized into 6 unique radiographic patterns:
epiphyseal dislocation, coronal split, sagittal split, volar
translation, adolescent type, and infantile type (Fig. 2),
the adolescent subtype having a closing or partially
closed physis, and infantile subtype having no radio-
opaque epiphyseal ossification center. The other types
were defined based on the fracture pattern. Time from the
injury to definitive intervention and secondary
complications were evaluated and compared between
classic and variant Seymour fractures.
e158 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
Treatment Protocol
Patients were initially evaluated in the emergency
room by an orthopaedic or plastic surgery resident or
advanced practice provider with bedside irrigation and
attempted closed reduction of the physis, along with in-
travenous or oral antibiotic therapy. Thereafter, patients
were either admitted for urgent operative treatment or sent
home with hand surgery clinic follow-up based on the
quality of bedside debridement and reduction. However,
as our institution is a tertiary referral center, many pa-
tients were initially managed at outside centers and were
directly referred to the hand surgery clinic. These patients
underwent operative fixation as soon as feasible. Surgical
management typically comprised of nail removal, irriga-
tion, debridement, open reduction with or without K-wire
placement, nail bed laceration repair, and perioperative
and postoperative antibiotic administration. This man-
agement was provided irrespective of the Seymour
fracture type.
Statistics
A descriptive statistical analysis of all patient demo-
graphic and clinical variables was undertaken by evaluating
the mean and range for interval data and proportions for
categorical data. Analysis utilizing a t test for interval data
and χ2 test for categorical data was performed. Differences
in secondary complication rates and the time interval be-
tween injury and definitive management were also assessed.
All P values were calculated using the Wilcoxon rank sum
test, with a statistically significant result being defined as
having a P value less than 0.05. All statistical analyses were
conducted using STATA (StataCorp., College Station,
Texas).
RESULTS
After the CPT query and database generation, 100
medical records were screened, and a total of 66 patients
met the inclusion criteria. Of the 66 distal phalanx frac-
tures identified, 36 (55%) were classified as the classic
Seymour fractures whereas 30 (45%) met the criteria as
radiographic variants. Of the 6 variant types, 6 were
adolescent, 3 were coronal split, 3 were epiphyseal dis-
location, 2 were infantile, 11 were sagittal split, and 4 were
volar translation types. The mean age at the time of
management was 9.44 years (range, 0.94 to 15.33 y) and
7.91 years (range, 0.79 to 18.50 y) for the classic and
variant groups, respectively (P = 0.204). (Table 1).
Comparison of the postoperative outcomes, including
total complications and infections between the classic and
variant Seymour fracture cohorts, can be found in Table 2.
The mean time to the final follow-up was 76.1 days, with
67.0 days for the classic group and 87.1 days for the
variant group (P = 0.240) (Table 1).
Among all patients, 57 underwent definitive treat-
ment in the operating room (86.4%), whereas 8 (12.1%)
underwent bedside management in the emergency room,
but required delayed surgical treatment due to inadequate
fracture debridement and reduction. One patient under-
went only bedside treatment in the emergency room
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Variants of Seymour Fractures

FIGURE 1. Clinical and radiographic presentation of a classic Seymour fracture demonstrating a distal phalangeal fracture with an
associated nail bed laceration involving the base of the nail overlying the nail fold with bone exposure.

(1.5%). For patients who underwent urgent or delayed
operative intervention, 53 of 65 (81.5%) underwent open
reduction and Kirschner wire (K-wire) fixation.
The difference in the mean time from injury to de-
finitive treatment was not found to be statistically sig-
nificant between the 2 groups. The mean time from injury
to definitive treatment in the classic Seymour fracture
group was 7.3 days versus 12.7 days in the variant group
(P = 0.216), with a combined average of 9.8 days from
injury to definitive management (Table 1). Overall, the
complication rate was 16.7% (11 of 66 fractures), with the
majority of these being infections at 13.6% (9 of 11,
81.8%). The difference in complication rates was not
found to be statistically significant in the classic Seymour
versus variant group, with complications in 11.1% and
23.3% of cases, respectively (P = 0.185). When comparing

FIGURE 2. Radiographic findings of all variant Seymour fractures. Epiphyseal dislocation (A); coronal split (B); sagittal split (C);
volar translocation (D); adolescent type (E); infantile type (F).

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e159

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Dibbs et al
TABLE 1. Comparison of Demographic Variables, Time From Injury to Management, and Follow-up Time Between the Classic and
Variant Seymour Fracture Cohorts
Classic Seymour Fracture, %
Number of patients
Proportion of males:females
Mean age at presentation (y)
Mean Time from Injury to Presentation (d)
Mean Time from Injury to Definitive Management (d)
Mean Follow-up Time (d)
the infection rates between the 2 cohorts, there was also no
statistically significant difference observed, with 8.3% in
the classic Seymour fracture group and 20.0% in the
variant cohort (P = 0.169) (Table 2). Of the 3 infections in
the classic group, 1 was diagnosed as osteomyelitis based
on intraoperative findings of necrotic bone with deep
purulence; however, of the 6 infections in the variant
group, 3 were osteomyelitis based on similar criteria. Most
of the cases of osteomyelitis were diagnosed preoperatively
due to delayed presentation, however, 1 patient in each
group developed postoperative osteomyelitis that required
a second debridement procedure. Of the 9 patients who
developed an infection, 4 (44.4%) required additional
surgical intervention.
DISCUSSION
Since the original description of the Seymour frac-
tures in 1966,4 an increasing amount of attention has been
given to these uncommon pediatric injuries, given that
they do not always present as obvious open fractures and
can be missed if the examiner does not maintain a high
index of suspicion.5–7,12 Because of this, Seymour fractures
have been associated with high complication rates when
undergoing delayed treatment.5,12,14,15 Samade et al12 re-
ported that 32% of patients with the delayed presentation
were a result of an initially misdiagnosed Seymour frac-
ture, likely due to the variability in the clinical and ra-
diographic appearance.12 Our study demonstrated that
there were statistically similar delays between injury and
operative management between the 2 groups. However,
there was a trend towards an increased delay in treatment
for patients presenting with a Seymour fracture variant.
To our knowledge, the present study is the first to
describe and evaluate Seymour variant fractures in com-
parison with classic Seymour fractures. In each of the var-
iant sub-types included, there is a fracture of the distal
phalanx that involves the physis with associated nail bed
injuries, with fracture patterns, and characteristics that are
TABLE 2. Comparison of Postoperative Outcomes Including Total Complications and Infections Between the Classic and Variant
Seymour Fracture Cohorts
Classic Seymour Fractures
Total Patients Percentage of Patients
Patients Affected Affected, %
All Complications 36 4 11.1
Infection 36 3 8.33
e160 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
Seymour Fracture Variant, % P
36 30
2.60:1 (27.8 female) 1.73:1 (36.7 female) 0.596
9.44 7.91 0.204
3.0 4.9 0.46
7.3 12.7 0.216
67.0 87.1 0.240
distinct from the classic description of the Seymour fracture.
When considering the Seymour variant fractures captured
in this series, the overall complication rate and infection
rate were 23.3% and 20.0%, respectively. More importantly,
most of the infections encountered in the variant cohort
developed preoperatively, and the rate was considerably
higher compared with the classic Seymour cohort (13.3% vs.
5.6%, respectively). As previously stated by other authors,
this finding is likely a result of the inherent difficulty asso-
ciated with diagnosing variant Seymour fractures. Con-
sequently, these fractures may be undertreated initially. In
addition, the trend towards longer time from injury to
management noted for the variant cohort compared with
patients with the classic Seymour fractures is likely a
contributing factor in the increased rate of pretreatment
infections and total complications observed.
The overall complication rate and infection rate in
our study were 16.7% and 13.6%, respectively. For the
classic Seymour fractures, our complication rate was 11.1%
and the infection rate was 8.3%. It should be noted that
these rates are relatively low when compared with pre-
viously reported rates in the literature, though there is ap-
preciable variability in the data ranging from 9.2% to
57.6%.11–14,16 Nevertheless, infection rates vary consid-
erably depending on the timing of presentation and man-
agement of injury. Specifically, the rate of infection is higher
in fractures treated in a delayed fashion.13 Patients treated
at our institution demonstrated lower rates of infection
compared with other studies strictly investigating patients
with delayed presentation of Seymour fractures.12,13 Reyes
and colleagues reported that patients who underwent de-
layed operative management ( > 24 hours from injury) were
significantly more likely to present with infection, including
osteomyelitis, compared with those who underwent acute,
appropriate intervention (P = 0.035). In their study, 45% of
patients treated in a delayed fashion developed the
infection.13 In another investigation, 16.4% of 73 patients
who presented at a mean time of 7 days after injury
developed the infection.12
Seymour Fracture Variants
Total Patients Percentage of Patients
Patients Affected Affected, % P
30 7 23.3 0.185
30 6 20.0 0.169
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
The reduced rates of infection observed in our series
may be a result of the treatment protocol utilized at our
institution. As noted above, nearly all of the patients in
our cohort received antibiotic therapy at the initial pre-
sentation and after the operative management. On aver-
age, our patients were prescribed antibiotics for an 8.4-day
duration, with 66.1% receiving cephalexin. Rask et al
demonstrated that earlier management of Seymour frac-
tures with antibiotics resulted in significantly lower rates
of infection compared with those who received antibiotic
therapy beyond 24 hours after the injury at 6.9% and
76.5%, respectively.14 Several other studies investigate the
importance of administering antibiotics in the manage-
ment of Seymour fractures.9,11,12,17 Perez-Lopez et al
concluded that antibiotic use resulted in lower risk of in-
fection compared with those managed without anti-
biotics.17 However, more recent studies have questioned
the necessity of antibiotics if appropriate debridement and
repair are promptly performed. In a systematic review,
Metcalfe et al18 demonstrated that prophylactic antibiotics
did not result in lower rates of the superficial infection in
patients with open distal phalanx fractures.18 Similarly,
Stevenson et al reported that prophylactic administration
of flucloxacillin for the treatment of open distal phalanx
fractures did not improve postoperative outcomes.19 Thus,
prophylactic antibiotics should be administered at the
discretion of the treating physician.
Regardless of early antibiotic administration, the
operative technique and tendency toward operative inter-
vention may have contributed to the lower infection rate
in this cohort. All but one of our patients underwent in-
tervention in the operating room to properly debride and
stabilize the fracture and to repair the associated nail bed
injury. This management was provided irrespective of the
Seymour fracture type once the injury was properly di-
agnosed. Although Lin et al demonstrated that most
Seymour fractures may be managed in the emergency
department, our institutional practices and resource allo-
cation is such that the treatment in the emergency room is
usually less feasible, especially in the case of delayed
presentation.11 Multiple studies have reported no differ-
ence in complication rates based on the setting of inter-
vention performed; however, most of the patients in these
studies presented acutely.11,14
Our study has several limitations. Although this
cohort does represent nearly the highest number of Sey-
mour fractures in the current literature, it is possible that a
larger number of patients would have led to statistically
significant differences between the two groups.12 Our
methodology of identifying patients using CPTs codes
could lead to underrepresenting patients treated only in
the emergency room. However, as our institution is a
tertiary referral center, many of the patients in our cohort
were not seen at one of our institutional emergency rooms
before clinic referral. Thus, we believe that even patients
who may have not be properly diagnosed and treated at
the initial evaluation eventually followed up in our clinic
and underwent appropriate treatment in the operating
room to be captured by our chart review. In addition,
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Variants of Seymour Fractures
given the retrospective nature of the study design, we were
unable to obtain robust long-term outcomes on most pa-
tients, which would have provided more opportunity to
identify delayed complications, such as malunion, nail
dystrophy, and growth arrest.20,21 Future work will aim to
study these variants in a prospective manner to confirm
our findings presented here and obtain longer-term
outcome data.
We propose an ideal treatment protocol for
Seymour fractures that involves bedside irrigation and
attempted reduction by either hand surgery-trained
providers or emergency providers with prompt referral to
the hand surgery specialist. Antibiotics should be given
on presentation, and we believe a course of oral anti-
biotics should be prescribed if the patient will be un-
dergoing definitive fixation after the initial 24-hour
period. Operative fixation, when required, should involve
debridement of nail bed material from physis, anatomic
reduction with or without K-wire fixation, and nail bed
repair. Postoperative antibiotics may be optional based
on the provider’s preference. K-wires are typically re-
moved at 3 to 4 weeks postoperatively. There is a limited
role for blood work for routine fractures, but may be
indicated if osteomyelitis or soft tissue infection is
suspected.
In conclusion, patients with the variant Seymour
fractures had similar time to treatment and complication
rates as patients with the classic Seymour fractures, with a
trend toward higher complication and increased time to
management. As such, we propose a minor expansion of
the definition of Seymour fractures to include variants to
optimize the treatment of the injuries. Continued educa-
tion of practitioners who encounter these injuries in the
emergency department, urgent care, and clinic settings
helps to increase the index of suspicion for an open
physeal fracture associated with nail bed injury leading to
decreased complication rates.
REFERENCES
1. Vadivelu R, Dias JJ, Burke FD, et al. Hand injuries in children:
a prospective study. J Pediatr Orthop. 2006;26:29–35.
2. Nellans KW, Chung KC. Pediatric hand fractures. Hand Clin.
2013;29:569–578.
3. Abzug JM, Dua K, Bauer AS, et al. Pediatric phalanx fractures.
J Am Acad Orthop Surg. 2016;24:e174–e183.
4. Seymour N. Juxta-epiphysial fracture of the terminal phalanx of the
finger. J Bone Joint Surg Br. 1966;48:347–349.
5. Al-Qattan MM. Extra-articular transverse fractures of the base of
the distal phalanx (Seymour’s fracture) in children and adults.
J Hand Surg Am. 2001;26 B:201–206.
6. Ganayem M, Edelson G. Base of distal phalanx fracture in children:
a mallet finger mimic. J Pediatr Orthop. 2005;25:487–489.
7. Abzug JM, Kozin SH. Seymour fractures. J Hand Surg Am. 2013;38:
2267–2270.
8. Wolfe S, Pederson WC, Kozin S, et al. Green’s Operative Hand
Surgery, 7th Edition. Philadelphia: Elsevier; 2016;1.
9. Krusche-Mandl I, Köttstorfer J, Thalhammer G, et al. Seymour
fractures: Retrospective analysis and therapeutic considerations.
J Hand Surg Am. 2013;38:258–264.
10. Banerjee A. Irreducible distal phalangeal epiphyseal injuries. J Hand
Surg Am. 1992;17:337–338.
11. Lin JS, Popp JE, Balch Samora J. Treatment of acute Seymour
fractures. J Pediatr Orthop. 2019;39:e23–e27.
www.pedorthopaedics.com | e161
Dibbs et al
12. Samade R, Lin JS, Popp JE, et al. Delayed presentation of Seymour
fractures: a single institution experience and management recom-
mendations. Hand. 2021;16:686–693.
13. Reyes BA, Ho CA. The high risk of infection with delayed treatment
of open Seymour fractures: Salter-Harris I/II or juxta-epiphyseal
fractures of the distal phalanx with associated nailbed laceration.
J Pediatr Orthop. 2017;37:247–253.
14. Rask DM, Wingfield J, Elrick B, et al. Seymour fractures:
a retrospective review of infection rates, treatment and timing of
antibiotic administration. Pediatr Emerg Care. 2020;37:e1299–e1302.
15. Venkatesh A, Khajuria A, Greig A. Management of pediatric distal
fingertip injuries: a systematic literature review. Plast Reconstr Surg -
Glob Open. 2020;8:e2595.
16. Baker CE, Leafblad N, Larson AN. Pediatric Seymour fractures of
the toe. J Pediatr Orthop. 2021;41:e55–e59.
e162 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
17. Perez-Lopez LM, Parada-Avendaño I, Cabrera-Gonzalez M, et al.
Seymour fracture: better do not underestimate it. Jt Dis Relat Surg.
2021;32:569–574.
18. Metcalfe D, Aquilina AL, Hedley HM. Prophylactic antibiotics in
open distal phalanx fractures: systematic review and meta-analysis.
J Hand Surg Eur Vol. 2016;41:423–430.
19. Stevenson J, McNaughton G, Riley J. The use of prophylactic
flucloxacillin in treatment of open fractures of the distal phalanx within
an accident and emergency department: a double-blind randomized
placebo-controlled trial. J Hand Surg Br. 2003;28:388–394.
20. Pandhi D, Verma P. Nail avulsion: Indications and methods (surgical
nail avulsion). Indian J Dermatol Venereol Leprol. 2012;78:299–308.
21. Lai W, Tang W, Loo S, et al. Clinical characteristics and treatment
outcomes of patients undergoing nail avulsion surgery for dystrophic
nails. Hong Kong Med J. 2011;17:127–131.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Correction of Pediatric Cubitus Varus by Centralization
of the Distal Humeral Fragment: A Surgical Technique
to Avoid Lateral Condylar Prominence
Yong Liu, MD, PhD,* Lisheng Kan, MD, PhD,† Jun Sun, MD,* and Xiangjun Chu, MD*
Background: Cubitus varus is a well-recognized late complication
of supracondylar humerus fractures in children. Various osteot-
omies have been described to correct this deformity, but each has
disadvantages. The purpose of this study was to investigate the
outcomes of a surgical technique, lateral closing wedge osteot-
omy combined with centralization of the distal humeral frag-
ment, in cubitus varus. Meanwhile, the occurrence of
postoperative lateral condylar prominence was observed.
Methods: From January 2015 to December 2019, 36 pediatric
cubitus varus deformity patients who were treated with lateral
closing wedge osteotomy combined with centralization of the
distal fragment in our institution were included. The corrective
osteotomies were performed through a standard lateral approach
and fixed with crossed Kirschner wires. The preoperative and
postoperative full-length anteroposterior and lateral radiographs
of bilateral upper limbs were assessed. At the same time, clinical
and radiologic parameters were reviewed. Lateral condylar
prominence index (LCPI) and any other complications were
evaluated.
Results: The mean preoperative humerus-elbow-wrist angle
(HEWA) on the affected side was 20.9 degrees of varus, which
was significantly improved to 9.2 degrees of valgus post-
operatively. The mean postoperative value of LCPI was −0.047.
Postoperative LCPI and HEWA was compared with the normal
side, and there was no significant difference. All of the patients
had excellent clinical and radiographic alignment. No surgical
complications and limitation of range of motion were noted. No
patient complained of lateral bony prominence.
From the *Department of Pediatric Orthopedics, The Affiliated Provin-
cial Pediatric Hospital of Anhui Medical University, Hefei, China;
and †Military Hospital of Chinese PLA, Dalian, China.
Ethical approval was waived by Clinical Ethics Committee of the Af-
filiated Provincial Pediatric Hospital of Anhui Medical University in
view of the retrospective nature of the study and all the procedures
being performed were part of the routine care.
Y.L.: data recruitment, study design, manuscript preparation. L.K.: data
analysis, manuscript preparation. J.S.: study design, manuscript
preparation. X.C.: data analysis, manuscript preparation. Y.L. and J.S.
are co-corresponding authors.
Informed consent: Informed consent was obtained from all individual
participants included in the study.
The authors declare no conflicts of interest.
Reprints: Yong Liu, MD, PhD, № 39, Wangjiang Road, Hefei 230051,
China. E-mail: dazguoly@163.com
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002286
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Conclusion: Lateral closing wedge osteotomy by centralization of
distal fragment is a safe and highly effective method and prevents
lateral prominence with a minimal complication rate. We rec-
ommend this technique for the treatment of cubitus varus in
children.
Level of Evidence: Level IV Case series.
Key Words: pediatric cubitus varus, lateral closing wedge os-
teotomy, centralization of distal humeral fragment, lateral con-
dylar prominence, kirschner wires
(J Pediatr Orthop 2023;43:111–116)
P ediatric cubitus varus deformity, the most common
late complication following displaced supracondylar
humerus fractures, is a typical triplanar malalignment.
Although the limitation of forearm movement arising
from which is rare, the poorly tolerated appearance of the
deformity is the main reason for the child’s parents asking
for correction surgery.1–3 Up to now, multiple corrective
osteotomies have been reported in the treatment of cubitus
varus deformity, although each of them has its own ad-
vantages and disadvantages. The lateral closing wedge
osteotomy is currently the most popular procedure be-
cause of its ease and simplicity2,4–8 but criticized for un-
sightly scar and lateral condylar prominence.9 Literatures
have proposed that the prominence of lateral condyle
could be avoided by the medial shift of the distal fragment
of the humerus;1,10 however, to the best of our knowledge,
no consensus on the optimal surgical protocol has been
reached for this complex pathology.
Thus, in the present study, we evaluated the out-
comes of a procedure (lateral closing wedge osteotomy
combined with centralization of the distal humeral frag-
ment) in the treatment of pediatric cubitus varus, which is
designed to correct the 3-dimensional deformity while
avoiding the complications of traditional techniques.
Meanwhile, the occurrence of postoperative lateral con-
dylar prominence was observed.
METHODS
Between January 2015 and December 2019, 36 cu-
bitus varus patients who underwent lateral closing wedge
osteotomy combined with centralization of distal humeral
fragment in our institution were included. The median
www.pedorthopaedics.com | 111
Liu et al J Pediatr Orthop  Volume 43, Number 2, February 2023

follow-up time was 33 months (range 24 to 61). There

were 24 males and 12 females. Of all the patients, 24
(66.7%) had a left-sided deformity, and 12 (33.3%) had a
right-sided deformity. All cases had a history of unilateral
supracondylar humeral fracture. The mean age at the
fracture and the osteotomy was 5.6 years (range 2.8 to 8.4)
and 6.9 years (range 3.8 to 12.0), respectively. The study
was approved by the Clinical Ethics Committee of the
Affiliated Provincial Pediatric Hospital of Anhui Medical
University.

Operative Technique
The osteotomy was performed under general anes-
thesia through a standard lateral approach in the supine
position. After tourniquet control, the operated arm was
placed on a fluoroscopy arm. A distal cut was performed
parallel to the articular surface and 0.5 to 1.0 cm superior
to the olecranon fossa. The proximal fragment cut was
made oblique to the shaft of the humerus, correlating to
the difference in the humerus-elbow-wrist angle (HEWA,
Fig. 1) between the ipsilateral and contralateral side, as
measured preoperatively. As Joysticks, the preplaced
Kirschner wires (A wires) were drilled across the
osteotomy to engage the medial cortex. After the
removal of the wedge, we performed a centralization of
distal fragment reduction maneuver of hyperflexion,
medialization, and derotation, ensuring that the anterior
humeral line (AHL) passing through the middle one-third
of capitulum in the sagittal plane, and central humeral
axis (CHA) through the middle point of line connecting
the 2 epicondyles (Fig. 2CD line) in the coral plane. If
necessary, the anterior cortex of the proximal humerus FIGURE 1. Humeral-elbow-wrist angle is calculated from the
was removed. The osteotomy site was 3 cross-fixed with angle of intersection of the forearm and humeral axis.
1.5-or 2.0 mm Kirschner wires (B wires), inserted
percutaneously. All surgical steps were performed under
the periosteum. To avoid ulnar nerve injury, the surgeon’s
thumb was placed over the medial epicondyle with the
elbow extension and swept posteriorly over the cubital
tunnel. Then, the third smooth Kirschner wire should
make almost immediate contact with the bone/cartilage of
the medial epicondyle once it is introduced through the
skin. When placement is complete, this pin should engage
the opposite far cortex. We do not recommend a small
incision to identify and protect the ulnar nerve in our
series. The stability of fixation in maximum flexion and
extension was tested under fluoroscopy. After A wires
were removed, B wires were cut and left outside the skin.
Postoperatively, a drain was used for 24 hours, and a
split cast was applied with the arm in 90 degrees of flexion. FIGURE 2. A osteotomy technique. (1) Draw a CD line (a line
In all patients, the casts and pins were removed post- connecting the 2 epicondyles of the distal fragment). The ini-
operatively 4 to 6 weeks in outpatient settings after callus tial osteotomy was performed 0.5 to 1.0 cm superior to the
formation was confirmed by the radiograph, and exercise olecranon fossa and parallel to the articular surface (BO),
and proximal cut oblique to the shaft of the humerus (AO). The
of the elbow was required without physical therapy. Four degree of correction is ∠AOB according to the preoperative
weeks after the follow-up, all elbows were evaluated for plan. (2) Intraoperative centralization of distal fragment re-
the range of motion, and 2 (5.6%) elbow movements were duction ensuring that central humeral axis meeting point
found limited. After 3 times a week of physical therapy E (the middle point of the CD) in the coral plane, (3) and
treatment for 1 month, the 2 elbows movement returned to anterior humeral line passing through the middle one-third of
normal. capitulum in the sagittal plane.

112 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
Data Collection and Outcome Evaluation
In all patients, the full-length anteroposterior and
lateral radiographs of bilateral upper limbs taken before
the operation and at the last follow-up were evaluated for
the HEWA. Valgus was recorded as a positive value
wheras varus was recorded as a negative value. The
comparison of preoperative HEWA was performed be-
tween the affected and normal sides to determine the de-
sired angle of correction.11 Differences of HEWA more
than 20 degrees and deformity persisted for more than
1 year were indications for the corrective osteotomy. Pa-
tients’ cosmetic outcomes were evaluated according to the
criteria of Barrett,7 and final clinical outcomes were clas-
sified as the scale of Oppenheim.11 The active range of
motion (ROM) was clinically determined before the op-
eration and at the final follow-up. Preoperative and the
last follow-up radiographs were measured according to the
method described by Wong et al8 for the lateral condylar
prominence index (LCPI), which was recorded as a mild
negative value in the coronal plane commonly.
Statistical Analysis
Statistical analyses were performed using SPSS sta-
tistical analysis software (SPSS Statistics 26 Inc.). All
values were presented as mean ± SD. Comparisons within
or between groups were made using a t test. P value less
than 0.05 was considered statistically significant.
RESULTS
Clinical Outcomes
At the last follow-up, 32 (88.9%) patients were in
excellent condition, and 4 (11.1%) patients were in good
condition when evaluated according to the criteria of
Barrett. Thirty-one (86.1%) and 5 (13.9%) patients ach-
ieved excellent and good results, respectively, when eval-
uating according to the Oppenheim scale, and no poor
results were found. AHL intersected the capitellum in 16
cases and fell anterior to the capitellum in 20 cases, to
correct the hyperextension deformity, in which the ante-
rior cortex of the proximal humerus was removed. The
mean preoperative and postoperative ROM of the affected
side was 134.9 ± 13.8 degrees (range 100-165 degrees) and
135.1 ± 11.5 degrees (range 115 to 155 degrees), re-
spectively, and no significant difference was found when
compared using a paired t test (t = −0.093, P = 0.926).
Preoperative and posterative clinical and functional
characteristics are summarized in Table 1.
Radiographic Outcomes
The mean preoperative HEWA of the affected side
was 20.9 degrees of varus (range −10 to −33 degrees),
which was significantly improved to 9.2 degrees of valgus
(range 4-15 degrees) postoperatively (t = −20.862 P =
0.000), and the corresponding data of the normal side was
10.6 degrees of valgus (range 4-18 degrees). At the last
follow-up, no statistically significant difference was found
between the HEWA of both sides (t = −1.866 P = 0.066)
(Fig. 3).
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Correction of Pediatric Cubitus Varus
TABLE 1. Preoperative and Postoperative Clinical and
Functional Characteristics of the Study Participants
N
Mean age at surgery (y) 6.9 (3.8 to 12.0)
Sex (female/male) 12/24
Side (left/right) 24/12
Mean surgical time (min) 54.5 ± 14.44(40 to 90)
Oppenheim’s criteria
Excellent (n) 31
Good (n) 5
Poor (n) none
Barrett’s criteria
Excellent (n) 32
Good (n) 4
Postoperative range of motion (degree) 135.1 ± 11.5 (range 115 to 155)
Pronation (degree) 85.4 ± 3.5（80 to 90）
Supination (degree) 82.4 ± 3.0（75 to 85）
Median follow-up time (mo) 33 (24 to 61)
Before the operation, the mean LCPI of the operated
elbows was −0.074 (range −0.320 to 0.130), which was not
different from the postoperative value (-0.047, range
−0.170 to −0.010 P = 0.131). Also, no significant differ-
ence was found in terms of the mean LCPI between both
sides at the final follow-up (P = 0.297).
At the last follow-up, none of the patients com-
plained of actual complications such as lateral bulging or
obvious scar, limited elbow movement, and no pin-site
infections or late loss of fixation occurred.
DISCUSSION
Cubitus varus deformity is the most common late
sequelae of supracondylar fracture in children, especially
in patients who received nonoperative management orig-
inally. Since the elbow joint function is not often impaired
greatly, the appearance following deformity is ignored by
many caregivers, which may lead to some long-term
consequences, including posterolateral rotatory instability,
tardy ulnar nerve palsy, secondary fractures, and internal
rotational malalignment,12,13 which should be noted. At
present, cosmesis is the primary consideration for correc-
tion. However, the optimal timing of osteotomy is still
unclear. More and more authors advocate early correction
rather than till late skeletal maturity because cubitus varus
is a 3-dimensional deformity as a result of malunion with
less remodeling capacity.2,7,14
To date, the surgical generally preferred standard for
the correction of 3-dimensional deformity of the elbow has
not been well-established. The reported osteotomies have
their own advantages and disadvantages.1,6,9,15–17 The
technique described by Skaggs et al18 can only correct
coronal and sagittal deformities but cannot correct rota-
tional deformities. Furthermore, it puts forward better
requirements for osteotomy accuracy and is technically
challenging. Although the oblique lateral closing wedge
osteotomy described by Greenhill et al19 is safe and re-
producible, the additional cuts still mandate uniplanar
osteotomies and are not suitable for younger children
( ≤ 5 y), and it is difficult to fix with the crossed Kirschner
www.pedorthopaedics.com | 113
Liu et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 3. A and B, AP and lateral radiographs of a cubitus varus in a 5-year-old girl. Humeral-elbow-wrist angle and LCPI in the AP
view (A) is −15 degrees and −0.167, and AHL fell anterior to the capitellum in the lateral view (B). C and D, Postoperative
demonstration of cross K-wire fixation following lateral closing wedge osteotomies by centralization of the distal fragment. passes
through the middle one-third of capitulum in the sagittal plane (C), and central humeral axis through the middle point of the line
connecting the 2 epicondyles in the coral plane (D). E and F, Postoperative AP and lateral elbow radiographs at 4-year follow-up.
LCPI is −0.077. AP indicates anteroposterior; LCPI, lateral condylar prominence index.

wire which can provide maximum mechanical stability. prominence while enhancing inherent stability. However,
The closed wedge counter shift osteotomy described by it was performed through a standard posterior approach
Abdelmotaal et al20 theoretically minimizes lateral in the lateral decubitus position and demands more

114 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
extensive dissection and higher surgical technology.
What’s more, they cannot correct severe internal rotation
and hyperextension deformities. The lateral closing wedge
osteotomy through a standard lateral approach in the
supine position is currently the most popular procedure
because of its simplicity and validity,21 but is widely
criticized for lateral condylar prominence. Cho et al9 re-
ported that remodeling of a lateral prominence has a
higher potential for a better outcome before puberty;
however, the incidence of radiographic lateral condylar
prominence was up to 57.1% at the final follow-up. Raney
et al5 also reported a high rate of 62%. Wong et al8 noted a
significantly increased condylar prominence following the
osteotemy in patients over 12 years of age. Thus,
investigators1,10 proposed the distal fragment be displaced
medially to prevent deformity. However, there has been
no consensus achieved on the most effective surgical
technique for reducing the prominence of lateral condyle
up to now. In the technique described here, all patients in
our series underwent lateral closing wedge osteotomy
combined with centralization of the distal humeral frag-
ment. When AHL passed through the middle one-third of
capitulum in the sagittal plane and CHA through the
middle point of the line connecting the 2 epicondyles (CD
line) in the coronal plane, the proximal and distal of the
osteotomy would match exactly. In this study, AHL fell
anterior to the capitellum in 20 cases. To correct the hy-
perextension deformity, the anterior cortex of the prox-
imal humerus was excised. Study has confirmed that AHL
is a reliable radiographic indicator for quantifying the
anatomic reduction in the sagittal plane.22 It is easy to
distinguish the AHL and CHA under fluoroscopy. In the
present study, we used a CHA through the midpoint of
distal osteotomy as the standard of coronal reduction
which is described for the first time. At the last follow-up,
favorable outcomes were achieved in all patients, and no
lateral condylar prominence and other complications were
observed. The present study proposed a method for the
first time to match exactly the proximal and distal of the
osteotomy to correct triplanar malalignment. Our study
demonstrated that the combination of lateral closing
wedge osteotomy and centralization of the distal humeral
fragment yielded satisfied 3-dimensional reduction, and no
reported postoperative complications were observed.
Various methods of fixation for cubitus varus correc-
tion have been well described. Persiani et al4 recommended
a locking angular plate in posttraumatic cubitus varus for its
advantages over pin fixation in allowing early mobilization,
better outcomes concerning the correction of HEWA and
LCP, and the patient’s satisfaction. Nonetheless, it demands
more extensive dissection, secondary hardware removal,
and higher surgical technology, and it also has a risk of
delayed union. Comparative study23 demonstrated that ex-
ternal fixation has advantages over internal fixation with
regard to easier preoperative planning, shorter operation
time, and easier postoperative fixation removal. However, a
high rate of pin-site infection has been reported. Meanwhile,
the expensive and special frame has higher technical re-
quirements compared with Kirschner wires. Furthermore,
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Correction of Pediatric Cubitus Varus
external fixation, such as Ilizarov method, may have a risk
of nerve injuries.24 Although complications, including loss
of reduction and pin infection, have been reported in the
literature,4,5 Kirschner wires fixation has been proven to be
a practical, effective, reliable, safe, and simple method in the
treatment of pediatric cubitus varus.6,15,16,20,25 As is known
that the tenant of maximizing the spread of k-wire fixation
for construct stability would be beneficial to reinforce the
biomechanical principle and minimize the potential for
complications. Nerve injury is perhaps the most feared
complication of humeral osteotomies, which occurred in
2.53% of cases26 and is related to surgical approaches, es-
pecially to the posterior, triceps splitting approach.5,25,26
Besides, the potential iatrogenic injury to the ulnar nerve
occurs when a medial approach is used.10,27 Our results
demonstrated that the crossed Kirschner wire fixation was
associated with enhanced mechanical stability, simplicity of
operation, less invasion, and no need for a second operation
for removal. Furthermore, loss of correction, nerve injuries,
infection, and the obvious scar was not observed within the
follow-up period. The reason may be that: (1) lateral closing
wedge osteotomy by centralization of the distal humeral
fragment is a simple and highly effective method, which can
lead to minimal soft-tissue damage; (2) the corresponding
procedure was performed under the periosteum to avoid
iatrogenic nerve injury; (3) the osteotomy site was 3 cross-
fixed percutaneously with Kirschner wires to maximize the
spread fixation for construct stability, meanwhile, minimiz-
ing the potential complications; (4) postoperatively, a drain
was routinely used for 24 hours in our series.
This study has several limitations as (1) the retro-
spective nature and no comparison with other established
techniques; (2) the lack of validated patient outcome
measurements; (3) subjective bias was potentially present
due to the single-center study.
In conclusion, the surgical technique can offer a
satisfying triplanar reduction and achieve good outcomes
without the problems of lateral condylar prominence and
other complications. Thus we advocate this technique for
the treatment of appropriate cubitus varus patients.
REFERENCES
1. Vashisht S, Sudesh P, Gopinathan NR, et al. Results of the modified
reverse step-cut osteotomy in paediatric cubitus varus. Int Orthop.
2020;44:1417–1426.
2. Pankaj A, Dua A, Malhotra R, et al. Dome osteotomy for
posttraumatic cubitus varus: a surgical technique to avoid lateral
condylar prominence. J Pediatr Orthop. 2006;26:61–66.
3. Singh P, Krishna A, Arora S, et al. Shortening dome osteotomy for
correction of severe cubitus varus secondary to malunited supra-
condylar fractures in children. Arch Orthop Trauma Surg. 2022.
Available at: https://doi.org/10.1007/s00402-021-04288-y.
4. Persiani P, Noia G, de Cristo C, et al. Lateral wedge osteotomy for
pediatric post-traumatic cubitus varus: Kirschner-wires or locking
angular plate? J Pediatr Orthop B. 2017;26:405–411.
5. Raney EM, Thielen Z, Gregory S, et al. Complications of
supracondylar osteotomies for cubitus varus. J Pediatr Orthop. 2012;
32:232–240.
6. Su Y, Nan G. Lateral closing isosceles triangular osteotomy for the
treatment of a post-traumatic cubitus varus deformity in children.
Bone Joint J. 2016;98-B:1521–1525.
www.pedorthopaedics.com | 115
Liu et al
7. Barrett IR, Bellemore MC, Kwon YM. Cosmetic results of supra-
condylar osteotomy for correction of cubitus varus. J Pediatr Orthop.
1998;18:445–447.
8. Wong HK, Lee EH, Balasubramaniam P. The lateral condylar
prominence. A complication of supracondylar osteotomy for cubitus
varus. J Bone Joint Surg Br. 1990;72:859–861.
9. Cho CH, Song KS, Min BW, et al. Long-term results of remodel-
ing of lateral condylar prominence after lateral closed-wedge
osteotomy for cubitus varus. J Shoulder Elbow Surg. 2009;18:
478–483.
10. Hui JH, Torode IP, Chatterjee A. Medial approach for corrective
osteotomy of cubitus varus: a cosmetic incision. J Pediatr Orthop.
2004;24:477–81.
11. Oppenheim WL, Clader TJ, Smith C, et al. Supracondylar humeral
osteotomy for traumatic childhood cubitus varus deformity. Clin
Orthop Relat Res. 1984;188:34–39.
12. Ho CA. Cubitus Varus-It’s More Than Just a Crooked Arm!
J Pediatr Orthop. 2017;37(Suppl 2):S37–S41.
13. O’Driscoll SW, Spinner RJ, McKee MD, et al. Tardy posterolateral
rotatory instability of the elbow due to cubitus varus. J Bone Joint
Surg Am. 2001;83:1358–1369.
14. Cha SM, Shin HD, Ahn JS. Relationship of cubitus varus and ulnar
varus deformity in supracondylar humeral fractures according to the
age at injury. J Shoulder Elbow Surg. 2016;25:289–296.
15. Eamsobhana P, Kaewpornsawan K. Double dome osteotomy for the
treatment of cubitus varus in children. Int Orthop. 2013;37:641–646.
16. Liddell TR, Hennrikus WL, Bresnahan JJ. A Simple Dual-Planar
Osteotomy to Correct Cubitus Varus Deformity in Children.
J Orthop Trauma. 2016;30:e340–e345.
17. Takagi T, Takayama S, Nakamura T, et al. Supracondylar
osteotomy of the humerus to correct cubitus varus: do both internal
116 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
rotation and extension deformities need to be corrected? J Bone Joint
Surg Am. 2010;92:1619–1626.
18. Skaggs DL, Glassman D, Weiss JM, et al. A new surgical technique
for the treatment of supracondylar humerus fracture malunions in
children. J Child Orthop. 2011;5:305–312.
19. Greenhill DA, Kozin SH, Kwon M, et al. Oblique Lateral Closing-
Wedge Osteotomy for Cubitus Varus in Skeletally Immature
Patients. JBJS Essent Surg Tech. 2019;9:1–8.
20. Abdelmotaal HM, Abutalib RA, Khoshhal KI. The closed wedge
counter shift osteotomy for the correction of post-traumatic cubitus
varus. J Pediatr Orthop B. 2013;22:376–382.
21. Griffin PP. Supracondylar fractures of the humerus. Treatment and
complications. Pediatr Clin North Am. 1975;22:477–486.
22. Sato K, Mimata Y, Takahashi G, et al. Validity of the distance between
the anterior humeral line and capitellum as a quantitative measure of
supracondylar humeral fracture in children. Injury. 2020;51:1321–1325.
23. Tang X, Wang J, Slongo T, et al. Comparison of internal fixation vs.
external fixation after corrective osteotomy in children with cubitus
varus. J Shoulder Elbow Surg. 2020;29:845–52.
24. Piskin A, Tomak Y, Sen C, et al. The management of cubitus varus
and valgus using the Ilizarov method. J Bone Joint Surg Br. 2007;
89:1615–1619.
25. Takagi T, Seki A, Takayama S, et al. Modified step-cut osteotomy
for correction of post-traumatic cubitus varus deformity: a report of
19 cases. J Pediatr Orthop B. 2016;25:424–428.
26. Solfelt DA, Hill BW, Anderson CP, et al. Supracondylar osteotomy
for the treatment of cubitus varus in children: a systematic review.
Bone Joint J. 2014;96-B:691–700.
27. Millis MB, Singer IJ, Hall JE. Supracondylar fracture of the humerus
in children. Further experience with a study in orthopaedic decision-
making. Clin Orthop Relat Res. 1984;188:90–97.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
The Effectiveness of Serial Casting in the Treatment
of Recurrent Equinovarus in Children With Arthrogryposis
Chris Church, MPT, Christina Bourantas, BA, Stephanie Butler, DPT, Jose J. Salazar-Torres, PhD,
John Henley, PhD, Maureen Donohoe, PT, DPT, Freeman Miller, MD,
and L. Reid Boyce Nichols, MD
Background: Arthrogryposis multiplex congenita (AMC) is
characterized by joint contractures in 2 or more body areas, often
resulting in clubfoot deformities that are typically stiffer than
those seen in idiopathic clubfoot deformities. While surgery is
routinely used to treat clubfoot in AMC, it has a high rate of
recurrence and complications. Current literature suggests serial
casting (SC) could be useful in treating clubfoot in AMC, though
evidence of its effectiveness is limited.
Methods: Passive range of motion (PROM), dynamic foot pres-
sure, parent-reported Pediatric Outcomes Data Collection In-
strument, brace tolerance, and the need for post-casting surgery
were evaluated retrospectively in children with AMC treated
with SC to address clubfoot deformities. Analysis of variance or
paired t tests were used as appropriate on pre-casting, short-term
(within 6 mo after SC) and/or longer-term (6 to 18 mo after SC)
parameters to determine the effectiveness of SC. Brace tolerance
before and after SC was analyzed using the Global Test for
Symmetry, and medical records were reviewed to determine the
need for surgery post-SC.
Results: Forty-six children (6.1 ± 3.1 y old) were cast an average
of 2.5 ± 1.9 times, resulting in 206 SC episodes. PROM showed
improvement in ankle dorsiflexion and forefoot abduction in the
short term (P < 0.05), returning to baseline measurements in the
long term (P = 0.09). Brace tolerance improved after casting
(P < 0.05). Only 15% of feet required surgery at follow-up at
10.3 ± 5.5 years. There were no significant changes in dynamic
foot pressure or Pediatric Outcomes Data Collection Instrument
results after SC, except for an increase in the pain subtest
(P < 0.05).
Conclusions: Serial casting in children with AMC can be effective
in temporarily improving PROM and improving brace tolerance,
but it does not impact dynamic barefoot position. Positive im-
pact of conservative management in children with AMC can
potentially delay or reduce the need for invasive surgical inter-
vention by improving PROM and brace tolerance.
Level of Evidence: Level III, Retrospective Comparative Study.
From the Department of Orthopaedics, Nemours Children’s Health,
Delaware Valley, Wilmington, DE.
The authors declare no conflicts of interest.
Reprints: Chris Church, MPT, Nemours Children’s Health, Delaware
Valley, Gait Analysis Laboratory, 1600 Rockland Rd, Wilmington,
DE 19803. E-mail: cchurch@nemours.org.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002309
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Key Words: clubfoot, arthrogryposis multiplex congenita, serial
casting, recurrence
(J Pediatr Orthop 2023;43:117–122)
A rthrogryposis multiplex congenita (AMC) is a term
used to describe a group of congenital conditions
characterized by joint contractures in 2 or more body
areas, including frequent foot deformities and limitations
in functional mobility.1 Arthrogryposis occurs in 1:3000 to
1:5100 live births,1,2 with fetal akinesia being the suggested
primary cause.3 Amyoplasia (involvement of all the 4
limbs) and distal arthrogryposis (lower extremity in-
volvement only) are the first and second most common
types of arthrogryposis, respectively, and comprise at least
50% of children diagnosed.1 With regular specialized
treatment, 85% of children can walk independently by the
age of 5 years.4
The most common type of foot deformity in ar-
throgryposis is clubfoot.5,6 Traditionally, arthrogrypotic
clubfeet has been treated with invasive bony surgical
procedures due to the frequency of recurrence and severity
and rigidity of the deformities.7–10 In contrast, idiopathic
clubfoot is routinely managed with the Ponseti method,
which uses a series of casts, orthotics, and physical therapy
to convert the equinovarus clubfoot into a plantigrade
foot, starting in infancy.11–13 The effectiveness of the
Ponseti method has been well-demonstrated in the setting
of idiopathic clubfoot and is widely accepted.14–17 Re-
cently, conservative management has been encouraged to
treat the arthrogrypotic clubfoot as well, despite little
supporting evidence. Previous research focuses on the
correction of clubfoot in infants and young children with
arthrogryposis based on static measurements such as Pi-
rani scores, the Dimeglio grading system, and ankle dor-
siflexion passive range of motion (PROM).11,13,18 These
studies suggest the Ponseti method could be useful in
treating arthrogrypotic clubfeet, but there is limited evi-
dence-based information regarding dynamic outcomes,
recurrence rates, or functional implications after serial
casting treatment in this population. Despite initial con-
servative management, Matar et al11 reported 35% of
children with AMC required foot surgery with a review of
5.8-year follow-up from the initial presentation at 5 weeks
www.pedorthopaedics.com | 117
Church et al J Pediatr Orthop  Volume 43, Number 2, February 2023

of age, and Morcuende et al13 reported 31% of children breakdown with the regular use of braces, fair tolerance
with AMC required foot surgery with a review of follow- referred to a child with some pain or redness with regular
up to age 4.6 years. While not reported in the literature, use, and poor tolerance referred to a child with pain or
centers that have focused on surgical intervention for skin breakdown that limited the use of the braces.
AMC-related clubfoot likely have surgery rates much
higher than these reports.
Therefore, the aim of this study was to evaluate the Post-Casting Surgery
effectiveness of serial casting (SC) in treating clubfoot Electronic medical records were reviewed for all
deformities in ambulatory children with arthrogryposis by children from the time of the initial SC episode until April
measuring short- and long-term outcomes of PROM, dy- 2021, looking for bony procedures or posterior medial
namic foot pressure, brace tolerance, parent-reported clubfoot releases performed after SC related to persistent
outcome scores, and need for post-casting surgery. clubfoot. The amount of time between the final SC episode
and this surgery was recorded.
METHODS
This retrospective study was approved by our in- Outcome Questionnaire
stitutional review board. Children with a diagnosis of
The PODCI was completed by the parent before and
AMC between October 2014 and April 2021 were identi-
after a casting episode.19 This questionnaire evaluates
fied from the hospital database. Ambulatory children with
physical function in children with chronic orthopaedic
AMC who were 2 years of age or older, had at least 1
conditions. The survey contains 5 subsections, including
episode of SC for an equinovarus foot and had pre-SC and
upper extremity and physical functioning, transfer and
post-SC assessments were included. Children who had
basic mobility, sports and physical functioning, pain/
bony foot surgery or a clubfoot release procedure before
comfort, and happiness, and a PODCI global function
SC were excluded. Outcome measures included PROM,
score.19
foot pressure measurements, and Pediatric Outcomes
Data Collection Instrument (PODCI) scores.19 Brace tol-
erance before and after SC and the need for bony surgery Pedobarograph
after SC were also recorded. Dynamic foot pressure readings were collected using
Serial Casting a pedobarograph (Tekscan, Boston, MA) for children who
Indications for SC included progressive equinovarus could walk barefoot at a self-selected pace. Walking aids,
leading to limited brace tolerance, pain, or functional such as walkers, canes, and crutches or hand-held assis-
limitation. Children were serially casted with short leg tance, were used as needed. Three contacts on the pedo-
hybrid casts (plaster base for ideal molding, then fiberglass barograph for each foot were required, then data were
on top for strength to allow walking) weekly or biweekly averaged to obtain final foot pressure data. Heel pressures,
with the duration of casting episode based on PROM medial and lateral midfoot and forefoot pressures, and
response.15–17 Casting was completed by the primary or- coronal plane pressure index (CPPI) were studied
thopedic surgeon, physician assistant, or physical thera- (Fig. 1).21 The CPPI is the ratio of medial-to-lateral
pist, all of whom had more than 10 years of experience plantar pressure distribution with a scale of -100 (severe
casting children with arthrogryposis. A serial casting epi- varus) to +100 (severe valgus).
sode is defined as the amount of casting required to ach-
ieve a braceable foot with the forefoot in neutral and the
hindfoot near plantigrade. Children who had a recurrence
of foot deformity later in childhood underwent repeat SC,
with each episode counted separately in this study.
Physical Examination
Passive range of motion measurements were re-
corded by the same clinician during the child’s clinic visits
and included ankle dorsiflexion, ankle plantarflexion,
forefoot abduction, and forefoot adduction.20 Measure-
ments were recorded before the initial cast application, at
cast changes, and after the final cast removal in the short
term (within 6 mo after cast removal) and long term (6 to
14 mo after removal).
Brace Tolerance FIGURE 1. Division of plantar pressure areas into medial and
Brace tolerance was categorized as poor, fair, or lateral forefoot, medial and lateral midfoot, and heel. H in-
good based on clinic notes pre-SC and post-SC episodes. dicates heel; LF, lateral forefoot; MF, medial forefoot; MM,
Good tolerance referred to a child with no pain or skin medial midfoot.

118 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 The Effectiveness of Serial Casting

TABLE 1. Comparison of Passive Range of Motion Data Between Pre-casting (0.4 ± 1.8 mo) and Short-term Post-casting
(0.3 ± 1.0 mo)
Variable N Pre-casting Mean (degrees) Short-term Post-casting Mean (degrees) P
Ankle dorsiflexion 206 −10 ± 11 −3 ± 8 < 0.0001*
Ankle plantarflexion 143 36 ± 13 34 ± 12 0.11
Forefoot abduction 149 6 ± 17 17 ± 11 < 0.0001*
Forefoot adduction 139 31 ± 16 29 ± 10 0.06
*Indicates statistical significance. N, number of casting events with associated pre-measurements and short-term post measurements.

Statistics the onset of serial casting, with 4 children being excluded

Evaluations completed before SC were retro-
spectively compared with data at short term and long
term. Data include results from pedobarograph, PROM,
PODCI, brace tolerance, and need for post-casting sur-
gery. Pre-casting and short-term outcomes for all paired
assessments were compared using a paired t test analysis;
pre-casting and short-term and long-term outcomes were
compared using analysis of variance with Bonferroni
post-hoc assessments. Changes in brace tolerance before
and after SC were analyzed using the Global Test for
Symmetry (McNemar-Bowker Test).22 This test is a
method for performing multiple tests of paired pro-
portions and can be considered a repeated measures ver-
sion of the Chi-square test of independence. To determine
the factors associated with positive outcomes, correlation
coefficients were calculated between changes in outcome
variables pre-SC to post-SC (dynamic foot position/CPPI,
dynamic heel impulse, long-term change in dorsiflexion
PROM) and pre-casting factors (age, Gross Motor
Function Measure-dimension D, number of previous ep-
isodes of casting, pre-casting dorsiflexion PROM, pre-
casting CPPI). All statistical analyses were performed
using R statistical package.23
RESULTS
A total of 46 children met the inclusion criteria
(6.1 ± 3.1 y old). Fifty-nine percent of the children had a
diagnosis of amyoplasia and 41%, had distal arthrogry-
posis. Ninety-three percent (43/46) of the children were
ambulatory. All children included in this study had the
same treatment before serial casting, which included
Achilles tenotomy and early Ponseti casting before the age
of 2 years. In addition, none of the 46 children included in
the study had any bony surgery or clubfeet release before
due to prior bony foot surgery.
Serial Casting
There were a total of 206 SC episodes, with an
average of 2.5 ± 1.9 casting episodes per child. Some of the
46 children had more than 1 episode of casting between
2014 and 2021, and not all were cast bilaterally during
each SC episode.
Physical Examination
Foot and ankle PROM were collected 0.4 ± 1.8 months
before casting (n = 206 measurements), 0.3 ± 1.0 months after
casting (short term; n = 206 measurements), and
8.5 ± 2.7 months after casting (long term; n = 124 feet; range
6 to 14 mo). All children had at least partial short-term
PROM measurements available; however, some had in-
complete data collected during the clinic visit resulting in
different numbers for each measure. Passive ankle dorsi-
flexion and forefoot abduction range had a statistically sig-
nificant improvement after casting in the short term
(P ≤ 0.05) whereas passive ankle plantarflexion and forefoot
adduction range did not change (Table 1). Some children did
not return for long-term follow-up measurements, or
incomplete data were collected during the clinic visit;
therefore, 124 SC episodes had available pre-casting and
long-term PROM data. Comparing pre-casting with long-
term data show that forefoot abduction and ankle
dorsiflexion PROM measurements were no longer
significantly different, indicating recurrence of the baseline
deformity (P = 0.1) (Table 2). There was no change in ankle
plantarflexion and forefoot adduction measurements from
pre-casting to long-term evaluation.
Brace Tolerance
Before casting, 68% of children used ankle-foot
orthoses, 28% used knee-ankle-foot orthoses, 1% used

TABLE 2. Comparison of Passive Range of Motion Data Between Pre-casting (0.4 ± 1.8 mo), Short-term Post-casting
(0.3 ± 1.0 mo), and Long-term Post-casting (9.0 ± 3.0 mo)
Short-term P (Pre-casting to Long-term Post- P (Pre-casting to
Pre-casting Mean Post-casting Mean Short-term casting Mean Long-term
Variable N (degrees) (degrees) Post-casting) (degrees) Post-casting)
Ankle dorsiflexion 120 −11 ± 12 −3 ± 11 1.81E-11* −9 ± 11 0.09
Ankle plantarflexion 81 38 ± 14 33 ± 13 0.04 37 ± 11 1
Forefoot abduction 89 8 ± 17 19 ± 11 2.23E-06* 7 ± 15 1
Forefoot adduction 67 33 ± 17 28 ± 10 0.04 33 ± 16 1
*Indicates statistical significance; N, number of casting episodes with associated pre-measurements and short-term and long-term post measurements.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 119

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Church et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 3. Comparison of Pediatric Outcomes Data Collection

Relationship Between Outcome Measures and
Instrument Data Between Pre-casting and Post-casting (n = 40 Pre-casting Factors
children) More severe pre-casting varus deformity correlated
Pre-casting Post-casting with a greater post-casting improvement in CPPI
Variable Mean Mean P (R = −0.27; P = 0.024; Table 5), and more severe pre-casting
equinus contractures correlated with a greater improvement
Extremity and physical 69 ± 21 70 ± 23 0.28
functioning in long-term dorsiflexion PROM (R = −0.426; P = 0.0012;
Transfer and basic mobility 80 ± 17 80 ± 17 0.63 Table 5). Fewer previous episodes of casting correlated with
Sports and physical 56 ± 19 56 ± 17 0.2 greater improvement in heel contact after casting
functioning (R = −0.238; P = 0.047; Table 5).
Pain/comfort 82 ± 17 72 ± 20 0.01*
Happiness 83 ± 17 84 ± 22 0.96
Global functioning 72 ± 13 70 ± 14 0.12 DISCUSSION
*Indicates statistical significance. Recent research and success with using the Ponseti
method to treat idiopathic clubfoot14–17 have led re-
searchers to believe this method might be effective in
supramalleolar orthoses, and 3% did not use orthotics. treating clubfoot in children with arthrogryposis, which is
These values were unchanged following casting. Brace traditionally treated with bony surgery. Nevertheless, very
tolerance per child improved after casting (pre: good 30%, few studies have tested the effectiveness of conservative
fair 22%, poor 48%; post: good 79%, fair 17%, poor 4%; management in children with arthrogryposis, and no
P < 0.05). studies have tested the effect of SC in this population.
Post-Casting Surgery The PROM improvements seen in our short-term
results are consistent with the results that have been pre-
With the follow-up to 10.3 ± 5.5 years of age, only
viously reported using casting to treat children with
10/46 children (15% of feet) required bony surgery or
arthrogryposis.11,13,18,24,25 Boehm et al’s18 study demon-
posterior medial clubfoot release. Indications for surgery
strate short-term improvement following casting, with a
were determined during a routine physical examination
27% recurrence rate by an average of 6 months post-
when equinovarus deformities were fixed, and feet were no
casting. Similarly, our data showed improvements in ankle
longer braceable. These children had their first SC episode
dorsiflexion and forefoot abduction PROM in the short
at 4.6 ± 2.1 years of age, had 5.6 ± 3.2 total episodes of SC,
term, with a recurrence of deformity by 9 months after
and required surgery at age 9.7 ± 3.0 years (5.1 y later).
casting.
Outcomes Questionnaire Our results also showed brace tolerance improved
Forty children had both pre-casting and post-casting for the children in our study. Improving PROM likely
PODCI scores recorded. These PODCI results were col- improved brace fit and made them more tolerable. The
lected before casting and 7.0 ± 4.0 months after casting. percent of children who demonstrated good brace toler-
Parent-reported outcomes did not demonstrate any sig- ance before SC more than doubled after casting. In ad-
nificant change in the subtest scores, except that the pain dition, the percent of children reporting poor brace
subtest scores worsened from 82 ± 17 to 72 ± 20 points tolerance went from nearly half of the children to almost
(P = 0.01) (Table 3). none (4%, or 2 children). Many children with arthrogry-
posis required braces or orthotics to ambulate (68% of the
Pedobarography children in this study); therefore, improving brace toler-
Foot pressure results were collected before casting ance will enhance function and ability to participate in the
and 7.0 ± 4.0 months after casting. Seventy-seven feet had community, positively affecting the overall quality of life.
both pre-casting and post-casting foot pressure data. Foot Outcome goals for treating clubfoot in children
pressure analysis did not demonstrate a dynamic reduction with arthrogryposis are vastly different from those for
in the equinovarus foot deformity following casting, with children with idiopathic clubfoot. Goals for children with
63 feet remaining in varus position during walking arthrogryposis include increasing functional mobility by
(Table 4). creating a plantigrade, braceable foot, and delaying

TABLE 4. Comparison of Foot Pressure Data Between Pre-casting (1.0 ± 2.0 mo) and Post-casting (7.0 ± 4.0 mo) (n = 77 feet)
Variable Pre-casting Mean Post-casting Mean P
Heel % of total impulse 13 ± 19 13 ± 19 0.94
Lateral midfoot % of total impulse 53 ± 24 54 ± 24 0.58
Medial midfoot % of total impulse 3±8 3±8 0.99
Lateral forefoot % of total impulse 21 ± 16 21 ± 16 0.98
Medial forefoot % of total impulse 9 ± 12 8 ± 10 0.27
Time of heel rise, % gait cycle 36 ± 31 33 ± 39 0.35
Coronal plane pressure index −75 ± 31 −75 ± 29 0.89
Foot progression, degrees −2 ± 21 −4 ± 21 0.11

120 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
TABLE 5. Correlation R (P) Between Change in Outcome Parameters and Pre-casting Measurements
Change Pre-casting Age Pre-casting GMFM-D
Pain −0.0743 (0.46) 0.0610 (0.39)
CPPI −0.1376 (0.38) 0.2281 (0.49)
Heel −0.0276 (0.6) 0.1089 (0.96)
LT DF 0.1440 (0.26) 0.0021 (0.32)
*Indicates statistical significance; CPPI, coronal plane pressure index; DF, dorsiflexion; GMFM-D, Gross Motor Function Measure-dimension D; LT, long-term.
surgery. Idiopathic clubfoot tends to occur without ad-
ditional pathology, with higher-level functional goals
expected as a result. The children in our study not only
had clubfeet but also joint contractures and bony de-
formities in other body segments, often including knees
and hips, all of which contributed to their functional
limitations. This was likely a reason why the change in
individual PODCI domain scores was limited. The
PODCI does not seem to be sensitive enough to reflect
the functional changes made by improving foot and an-
kle PROM and brace tolerance. However, one cannot
ignore the significant increase in reported pain in the
PODCI assessment for which we have no explanation.
Of the 46 children included in this study, only 10
children (15% of feet) required surgery after casting, with
surgical intervention occurring 4.7 ± 3.2 years after the
initiation of casting. In our practice, casting was initiated
due to progressive foot deformity and to improve brace
tolerance; however, these were often indications for sur-
gery elsewhere. Because casting was the treatment of
choice at our institution, we did not have a control group.
However, if more than 15% of the arthrogrypotic clubfeet
in a pediatric practice required surgical correction before
the age of 10 years, SC could be considered. Other studies
using similar conservative methods find the need for sur-
gery to be a bit higher than ours (Matar et al: 35%11 and
Morcuende et al: 31%13). Fewer surgical procedures re-
duce the cost of care, risks from anesthesia, and other
operative complications, as well as minimize scar tissue
formation due to surgical procedures at a younger age.
Given the severe restrictions in passive flexibility in these
children, minimization of scarring is particularly
important.
With variation in response to casting, it is important to
consider which factors contribute to successful conservative
intervention. Feet with more severe varus deformity before
SC had greater improvements afterwards. Similarly, the feet
that had the most severe equinus contractures before casting
had the greatest change in long-term dorsiflexion PROM
after casting. This inverse relationship indicates that those
with the most severe deformities have better outcomes, and
this may result in delaying or decreasing the need for sur-
gical correction. Even if surgery is inevitable, our results
showed SC can at least temporarily reduce deformity and
improve brace tolerance so the period before surgery can be
more tolerable for the patient.
Although our study is the first of its kind to provide a
long-term evaluation of children with arthrogryposis
treated with SC, the retrospective nature and relatively
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
The Effectiveness of Serial Casting
Cast Number Pre-casting DF Pre-casting CPPI Time From Casting
0.0329 (0.93) −0.129 (0.47) −0.083 (0.62) 0.055 (0.76)
−0.0885 (0.72) −0.069 (0.59) −0.27 (0.024)* 0.029 (0.82)
−0.2398 (0.047)* 0.167 (0.19) 0.095 (0.43) −0.013 (0.9)
−0.0379 (0.86) −0.426 (0.0012)* −0.158 (0.25) −0.222 (0.1)
short follow-up of the study has its limitations. The data
collected were part of routine clinical care and as such, did
not follow a strict protocol. Not all measures were made at
each instance. The PODCI data were collected over a pe-
riod of rapid growth and development, and some changes
could be attributed to natural evolution and the multi-
faceted nature of patients with arthrogryposis. Our sample
size was small and was quite diverse in terms of the severity
or type of arthrogryposis seen in these children. It is im-
portant to consider that only 60% of children returned for
the long-term follow-up visit, and more research is needed
to evaluate the recurrence rates and long-term outcomes
beyond 8.5 months after cast removal.
Serial casting in children with clubfeet related to ar-
throgryposis is effective in temporarily improving foot and
ankle PROM, brace tolerance, and possibly delaying the
need for surgical intervention, especially in the more severe
cases. Delaying surgery will reduce costs and risks asso-
ciated with surgery, including minimizing the formation of
scar tissue, which can further limit passive flexibility. Im-
proving brace tolerance can improve a child’s ability to
participate in the community and therefore improve quality
of life.
REFERENCES
1. Dahan-Oliel N, Cachecho S, Barnes D, et al. International multi-
disciplinary collaboration toward an annotated definition of
arthrogryposis multiplex congenita. Am J Med Genet C Semin Med
Genet. 2019;181:288–299.
2. Darin N, Kimber E, Kroksmark AK, et al. Multiple congenital
contractures: birth prevalence, etiology, and outcome. J Pediatr.
2002;140:61–67.
3. Hall JG. Arthrogryposis (multiple congenital contractures): diag-
nostic approach to etiology, classification, genetics, and general
principles. Eur J Med Genet. 2014;57:464–472.
4. Sells JM, Jaffe KM, Hall JG. Amyoplasia, the most common type of
arthrogryposis: the potential for good outcome. Pediatrics. 1996;97:
225–231.
5. Zimbler S, Craig CL. The arthrogrypotic foot plan of management
and results of treatment. Foot Ankle. 1983;3:211–219.
6. van Bosse HJ. Syndromic feet: arthrogryposis and myelomeningo-
cele. Foot Ankle Clin. 2015;20:619–644.
7. Drummond DS, Cruess RL. The management of the foot and ankle in
arthrogryposis multiplex congenita. J Bone Joint Surg Br. 1978;60:96–99.
8. Widmann RF, Do TT, Burke SW. Radical soft-tissue release of the
arthrogrypotic clubfoot. J Pediatr Orthop B. 2005;14:111–115.
9. Cassis N, Capdevila R. Talectomy for clubfoot in arthrogryposis.
J Pediatr Orthop. 2000;20:652–655.
10. Legaspi J, Li YH, Chow W, et al. Talectomy in patients with
recurrent deformity in club foot. A long-term follow-up study. J Bone
Joint Surg Br. 2001;83:384–387.
11. Matar HE, Beirne P, Garg N. The effectiveness of the Ponseti
method for treating clubfoot associated with arthrogryposis: up to
8 years follow-up. J Child Orthop. 2016;10:15–18.
www.pedorthopaedics.com | 121
Church et al
12. Lloyd-Roberts GC, Lettin AWF. Arthrogryposis multiplex congenita.
J Bone Joint Surg Br. 1970;52-B:494–508.
13. Morcuende JA, Dobbs MB, Frick SL. Results of the Ponseti method
in patients with clubfoot associated with arthrogryposis. Iowa Orthop
J. 2008;28:22–26.
14. Morcuende JA, Dolan LA, Dietz FR, et al. Radical reduction in the
rate of extensive corrective surgery for clubfoot using the Ponseti
method. Pediatrics. 2004;113:376–380.
15. Ponseti IV. Congenital Clubfoot: Fundamentals of Treatment.
Oxford: Oxford University Press; 1996.
16. Ponseti IV, Smoley EN. Congenital clubfoot: the results of
treatment. J Bone Joint Surg Am. 1963;45:261–344.
17. Ponseti IV, Zhivkov M, Davis N, et al. Treatment of the complex
idiopathic clubfoot. Clin Orthop Relat Res. 2006;451:171–176.
18. Boehm S, Limpaphayom N, Alaee F, et al. Early results of the
Ponseti method for the treatment of clubfoot in distal arthrogryposis.
J Bone Joint Surg Am. 2008;90:1501–1507.
19. Norkin CC, White DJ. Measurement of Joint Motion: A Guide to
Goniometry. Philadelphia, PA: FA Davis Company; 2016.
122 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
20. Gates PE, Campbell SR. Effects of age, sex, and comorbidities on the
pediatric outcomes data collection instrument (PODCI) in the
general population. J Pediatr Orthop. 2015;35:203–209.
21. Chang CH, Miller F, Schuyler J. Dynamic pedobarograph in
evaluation of varus and valgus foot deformities. J Pediatr Orthop.
2002;22:813–818.
22. Westfall PH, Troendle JF, Pennello G. Multiple McNemar tests.
Biometrics. 2010;66:1185–1191.
23. R Core Team. R: a language and environment for statistical
computing. 2014. R Foundation for Statistical Computing, Vienna,
Austria. Available at: https://www.R.project.org/. Accessed June 15,
2022.
24. Kowalczyk B, Lejman T. Short-term experience with Ponseti
casting and the Achilles tenotomy method for clubfeet treatment in
arthrogryposis multiplex congenita. J Child Orthop. 2008;2:
365–371.
25. van Bosse HJ, Marangoz S, Lehman WB, et al. Correction of
arthrogrypotic clubfoot with a modified Ponseti technique. Clin
Orthop Relat Res. 2009;467:1283–1293.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE
Comparison of Procalcitonin With Commonly Used
Biomarkers and Algorithms for Evaluating Suspected
Pediatric Musculoskeletal Infection in the Emergency
Department
Lyndsey van der Laan, MD, MPH,*† Nakia Gaines, MD,*† Ngoc Van Horn, MD,*†
Chanhee Jo, PhD,‡ Yuhan Ma, MS,‡ and Lawson A. Copley, MD, MBA, FAAOS*†‡
Introduction: It is difficult to distinguish between children with
infectious versus noninfectious conditions of the musculoskeletal
system during initial evaluation. Clinical predictive algorithms
potentially support this effort but not without limitations. Pro-
calcitonin (PCT) has been proposed as a biomarker to help dif-
ferentiate infection from noninfection. This study evaluates the
adoption and utility of PCT during initial infection evaluations
and assesses test characteristics of commonly used parameters
and algorithms.
Methods: PCT was introduced for initial laboratory evaluation of
the suspected musculoskeletal infection. Prospective enrollment
occurred from July 2020 to November 2021 with 3 cohorts es-
tablished after a retrospective review of final diagnoses at the end
of treatment: 1) deep infection, 2) superficial infection, and 3)
noninfection. Univariate and multivariate logistic regression
analysis of parameters and diagnoses was performed. Test
characteristics of individual and aggregated parameters were
assessed.
Results: Among 258 children evaluated, 188 (72.9%) had PCT
drawn during the evaluation. An increase of PCT acquisition from
67.8% to 82.4% occurred over the study timeframe. Eighty-five
children were prospectively studied, including those with deep in-
fection (n = 21); superficial infection (n = 10), and noninfection
(n = 54). Test characteristics of parameters showed accuracy
ranging from 48.2% to 85.9%. PCT > 0.1 ng/mL independently
predicted deep infection in 84.7% of cases, outperforming white
blood cell, C-reactive protein (CRP), and absolute neutrophil
count. Using study thresholds for CRP, erythrocyte sedimentation
rate, PCT, and Temp improved accuracy to 89.4%.
Conclusions: PCT is a potentially useful biomarker during the initial
assessment of children suspected to have a musculoskeletal in-
fection. Systematic evaluation using a combination of parameters
improves the accuracy of assessment and assists predictive judgment
under uncertainty. PCT <0.1 ng/mL, erythrocyte sedimentation rate
From the *Children’s Health System of Texas, Dallas, TX; †University of
Texas Southwestern; and ‡Texas Scottish Rite Hospital for Children,
Dallas, TX.
The authors declare no conflicts of interest.
Reprints: Lawson A. Copley, MD, MBA, FAAOS, Children’s Medical
Center Dallas, Dallas 75235, TX. E-mail: lawson.copley@childrens.com
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002303
e168 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
<18 mm/hr, CRP <3.3 mg/dL, and temperature <37.8°C should
reasonably reassure clinicians that deep musculoskeletal infection is
less likely, given the high negative predictive value and collective
accuracy of these parameters.
Level of Evidence: Level III – Retrospective cohort comparison
Key Words: Procalcitonin, biomarkers, musculoskeletal infection,
pediatric
(J Pediatr Orthop 2023;43:e168–e173)
C hildren with signs and symptoms concerning in-
fectious, inflammatory, or reactive conditions in-
volving the musculoskeletal system create a diagnostic
dilemma. It is challenging to distinguish between these
conditions to establish accurate diagnoses during a single
encounter due to the similarity of history, physical
findings, and laboratory results. Although unrealistic in
deriving final diagnoses from the initial assessment, there
is value in rapidly determining, which children should be
(1) admitted for further assessment; (2) scheduled for
subspecialty evaluation, or (3) allowed to follow-up with
primary care physicians or only as needed.
No single laboratory parameter consistently dis-
tinguishes infection from similar conditions. This can lead
to inaccurate diagnosis and delay in the treatment at one
extreme or unnecessary intervention at the other.1 Clinical
prediction algorithms using combinations of risk factors to
determine relative probabilities of serious musculoskeletal
infections have the potential for error.2–8 Nonetheless, the
advantage of using discrete criteria to establish the relative
risk of infection is appealing to simplify early decisions
concerning the appropriateness of additional inpatient
assessment. There is increasing evidence that decision al-
gorithms reduce noise in healthcare, particularly when
uncertainty and diagnostic complexity are encountered
because humans have a tendency to be noisy (highly var-
iable) in our evaluations and judgments.9 However, there
are also disadvantages of decision algorithms due to the
potential to be misleading when applied to populations
that differ from those initially modeled or when applied
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
to conditions other than that for which they were
intended.9–11
Procalcitonin (PCT), a precursor of the calcitonin
peptide, is produced in the presence of bacterial endo-
toxins, tumor necrosis factor (TNF)-alpha, and inter-
leukin (IL)-6. Evidence suggests that PCT is minimally
produced in viral, reactive, or inflammatory conditions.12
A previous meta-analysis showed PCT was more accurate
than C-reactive protein (CRP) in diagnosing systemic
bacterial infections.13 Despite the widespread use of PCT
in adult infections, its utility is less established for
children.14 The purpose of this study is to introduce PCT
in the evaluation of children at a tertiary pediatric center
and assess adoption among providers. A secondary aim is
to evaluate the relative merit and accuracy of clinical
prediction algorithms and laboratory parameters, includ-
ing PCT, during the initial assessment of children with
acute presentations concerning musculoskeletal infection.
Although this study seeks to define general thresholds of
laboratory parameters that may be useful to distinguish
infection from noninfection, it is not intended to add yet
another prediction algorithm for this purpose.
METHODS
Following the Institutional Review Board (IRB)
approval, children with initial concern for musculoskeletal
infection who presented to the institution from July 2020
to November 2021 were prospectively enrolled by in-
formed consent and retrospectively studied after the fol-
low-up. The target for prospective enrollment was ~200, as
determined by the average rate of musculoskeletal in-
fection consultations at this institution of 350 to 400 per
year. Because PCT technology was just being introduced
to the institution, there was no preliminary data for power
analysis.
Because of growing interest in PCT for the work-up
of sepsis, an institutional decision was made to procure the
laboratory capability in April 2020. An order set, includ-
ing PCT, among other commonly ordered infection labs,
was created for use at the ED provider’s discretion.
Children entered the system through the ED, outpatient
clinic, or inpatient admission (inclusive of hospital trans-
fers and direct admission by community pediatricians who
contacted the admitting service through the hospital access
center). Follow-up evaluation occurred in the orthopaedic
outpatient clinic or, when necessary, by telephone contact.
After the follow-up, 3 study cohorts were established: 1)
deep infection (osteomyelitis, septic arthritis or pyomyo-
sitis); 2) superficial or skin structure infection (cellulitis or
abscess), or 3) noninfection. Data were retrospectively
gathered from the chart review, including history, vital
signs, laboratory values, and diagnoses. Temperature re-
cordings performed in the ED or inpatient hospital were
reviewed for the entire period of observation (up to
24 hours) to capture the maximum recorded temperature
for study purposes. All temperature measurements were
done by a temporal artery thermostat or temporal scan-
ner. Children were excluded for symptom duration
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Procalcitonin as Deep Infection Biomarker
≥ 28 days, antibiotic treatment, insufficient follow-up, or if
no PCT was obtained.
Statistical analysis of continuous variables was ac-
complished with ANOVA and Mann-Whitney test. Tukey
test was conducted for 2 group comparisons. χ2 was used
for discrete variables with Fisher exact test for small
sample sizes (≤ 5). Statistical significance was established
at P < 0.05. Multivariate logistic regression analysis and
receiver operating characteristics (ROC) identified pa-
rameters and thresholds for the risk of deep infection. Test
characteristics of independent variables and common
combinations of parameters with historically established
thresholds were determined for deep infection versus
superficial infection or noninfection.
RESULTS
Study Population
During the study timeframe, 258 children were
evaluated for suspected musculoskeletal infection. Among
these, 200 (77.5%) were initially assessed in the ED, 36
(14.0%) in the clinic, and 22 (8.5%) after direct admission
to the hospital. Of 129 children prospectively enrolled,
after exclusion criteria were applied, 85 were categorized
into cohorts of 1) deep infection (n = 21); 2) superficial
infection (n = 10), and 3) noninfection (n = 54). A chart
review was performed for 80 children who had hospital
admission or subsequent encounters in the outpatient
clinic. Telephone contact was necessary for 5 nonadmitted
children who did not follow-up in the clinic. These families
each confirmed that their child’s symptoms, which
prompted the concern, had completely resolved; hence,
they did not elect to keep the follow-up appointment.
Final diagnoses were determined as viral or reactive ar-
thritis (n = 21), trauma (n = 16), and self-limited pain
(n = 12) for noninfection; osteomyelitis (n = 17) and septic
arthritis (n = 4) for deep infection; and cellulitis, abscesses,
and septic bursitis for superficial infection (n = 10). Hos-
pital admission occurred for 36 of 85 (42.4%) children,
including 14 (25.9%) for noninfection, 3 (30.0%) for su-
perficial infection, and 19 (90.5%) for deep infection.
Children with deep infection had a significantly higher
admission rate than that of the noninfection cohort
(P < 0.001) and superficial infection cohort (P = 0.0013).
PCT Adoption
Among 258 children, 188 (72.9%) had PCT drawn
during their evaluation. There was a progressive increase
in PCT acquisition throughout the study period, with PCT
acquired for 40 of 59 (67.8%) children during the first
quarter. By the final quarter, PCT was acquired for 28 of
34 (82.4%) children (Fig. 1). Systematic acquisition of all
parameters, including complete blood count with
differential, erythrocyte sedimentation rate (ESR), CRP,
and PCT occurred during 187 (72.5%) evaluations.
Cohort Comparison
There were no significant differences in sex, insurance,
ethnicity/race, trauma history, or viral symptoms. Children
www.pedorthopaedics.com | e169
Laan et al J Pediatr Orthop  Volume 43, Number 2, February 2023

deep infection (mean of 0.4 ± 0.5 ng/mL) from noninfection

(mean of 0.1 ± 0.1 ng/mL) (P = 0.0002) (Table 2).
Multivariate Analysis
Logistic regression modeling identified significant
contributions of maximum recorded temperature and
CRP when differentiating cohorts (Table 3). Area under
the curve (AUC) was the highest for fever, ESR, CRP, and
PCT to differentiate deep infection from the others
(Table 4). PCT had a 90.5% sensitivity to identify deep
infection.

FIGURE 1. Trend of increasing acquisition of PCT assessed on a Test Characteristics

quarterly basis during the study period. PCT acquisition im-
proved from 67.8% to 82.4% institutionally over a 17-month
timeframe.
with deep infection were significantly differentiated from
those with superficial infection by fever, maximum
temperature, admission, CRP, ESR, and number of
Kocher ± Caird criteria (Tables 1 and 2). Similarly, children
with deep infection were differentiated from those with
noninfection by fever, tachycardia, white blood cell (WBC)
count, absolute neutrophil count, CRP, ESR, number of
Kocher ± Caird criteria, and PCT. Children without
infection were differentiated from children with superficial
infection by age and WBC. PCT significantly differentiated
The accuracy of parameters ranged from 48.2% (in-
ability to bear weight) to 85.9% (Kocher+Caird ≥ 3 risk
factors). PCT > 0.1 ng/mL independently predicted deep
infection in 84.7% of cases, outperforming WBC, CRP, and
absolute neutrophil count (Table 5). The false positive rate
for identification of deep infection was 11.8% when a cutoff
of 0.1 ng/mL was used. When using receiver operating
characteristics cutoff values for CRP (3.3 mg/dL), ESR (18
mm/hr), PCT (0.1 ng/mL), and temperature (37.8° C),
accuracy improved to 89.4%.
DISCUSSION
A multi-center study recently reported that 10 per-
cent of pediatric orthopaedic consultations were for
musculoskeletal infection and that culture-positive

TABLE 1. Discrete Data Comparison Between Children with Non-Infection, Superficial Infection and Deep Infection
Superficial Deep
Noninfection Infection Infection Overall Noninfection vs. Superficial vs.
Variable Category N (%) N (%) N (%) P Superficial Deep Deep Infection
Sex Female 21 (38.9) 5 (50.0) 6 (28.6) 0.491 0.728 0.438 0.423
Male 33 (61.1) 5 (50.0) 15 (71.4) — — — —
Insurance Class CHIP 2 (3.8) 0 (0.0) 0 (0.0) 0.290 0.330 0.418 0.215
Commercial 14 (26.5) 3 (30.0) 7 (33.3) — — — —
Medicaid 35 (66.0) 5 (50.0) 14 (66.7) — — — —
Self Pay 2 (3.8) 2 (20.0) 0 (0.0) — — — —
Ethnicity; Race Hispanic; American Indian 1 (1.9) 0 (0.0) 0 (0.0) 0.809 0.711 0.835 0.379
Hispanic; Other 2 (3.8) 0 (0.0) 1 (4.8) — — — —
Hispanic; White or 14 (26.4) 3 (30.0) 6 (28.6) — — — —
Caucasian
Non-Hisp; Asian 2 (3.8) 0 (0.0) 0 (0.0) — — — —
NonHisp; Black/African 12 (22.6) 2 (20.0) 3 (14.3) — — — —
American
Non-Hisp; Other 3 (5.7) 1 (10.0) 0 (0.0) — — — —
Non-Hisp; Unknown 1 (1.9) 1 (10.0) 0 (0.0) — — — —
Non-hisp; White/Caucasian 18 (34.0) 3 (30.0) 11 (52.4) — — — —
Hx Trauma 15 (27.8) 5 (50.0) 4 (19.0) 0.200 0.264 0.560 0.105
Hx Viral Symptoms 12 (22.2) 1 (10.0) 2 (9.5) 0.344 0.672 0.325 1.000
Inability to bear 32 (59.3) 6 (60.0) 16 (76.2) 0.628 1.000 0.483 0.417
weight
Fever ( ≥ 38C) in ED 9 (16.7) 1 (10.0) 17 (81.0) < 0.001 1.000 < 0.001 < 0.001
Tachycardia 10 (20.8) 3 (30.0) 9 (50.0) 0.069 0.678 0.032 0.434
WBC > 12.0 10 (18.5) 6 (60.0) 11 (52.4) 0.002 0.012 0.008 1.000
ESR ≥ 40 4 (7.4) 0 (0.0) 13 (61.9) < 0.001 1.000 < 0.001 0.001
CRP ≥ 2 13 (24.1) 6 (60.0) 19 (90.5) < 0.001 0.053 < 0.001 0.067
Hospital Admission 14 (25.9) 3 (30.0) 19 (90.5) < 0.001 1.000 < 0.001 0.001
Discrete data comparison between cohorts using χ2 and Fisher Exact Test.
C indicates Centigrade; CHIP, Children’s Health Insurance Program; CRP, C-Reactive Protein; ESR, Erythrocyte Sedimentation Rate; Hx, History; Non-hisp, Non-
Hispanic; PCT, Procalcitonin; WBC, White Blood Cell Count.

e170 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Procalcitonin as Deep Infection Biomarker

TABLE 2. Continuous Data Comparison Between Children with Non-Infection, Superficial Infection, and Deep Infection
Superficial
NonInfection Superficial Infection Deep Infection Overall Noninfection vs. Infection vs.
Superficial Deep Deep
Variable n Mean SD Min Max n Mean SD Min Max n Mean SD Min Max Anova P Infection Infection Infection
Max 46 37.3 0.6 36.4 39.0 10 37.2 0.5 36.4 38.2 18 38.9 0.9 37.1 40.2 < 0.001 0.977 < 0.001 < 0.001
Temp
WBC 54 9.7 4.1 4.1 25.3 10 14.3 5.2 8.5 23.8 21 13.5 5.9 6.0 26.4 0.002 0.019 0.009 0.897
ANC 54 4.9 3.7 1.1 18.8 10 7.9 2.2 4.7 11.4 21 9.4 5.1 3.4 21.4 < 0.001 0.074 < 0.001 0.611
ESR 54 15 13 3 58 10 9 6 1 17 21 44 23 9 92 < 0.001 0.489 < 0.001 < 0.001
CRP 52 1.6 2.8 0.4 14.6 10 4.4 6.9 0.4 23.5 21 9.2 7.8 0.4 28.1 < 0.001 0.238 < 0.001 0.040
PCT 54 0.1 0.1 0.0 0.9 10 0.2 0.4 0.0 1.3 21 0.4 0.5 0.0 2.0 < 0.001 0.617 < 0.001 0.119
Kocher 54 1.0 0.9 0.0 3.0 10 1.2 0.8 0.0 2.0 21 2.6 0.7 1.0 4.0 < 0.001 0.687 < 0.001 < 0.001
Criteria
Kocher 54 1.2 1.1 0.0 4.0 10 1.8 0.8 1.0 3.0 21 3.5 0.9 1.0 5.0 < 0.001 0.235 < 0.001 < 0.001
+Caird
Continuous variable cohort comparison with Analysis of Variance (Anova) and Tukey analysis between groups.
ANC indicates Absolute Neutrophil Count; C, Centigrade; CRP, C-Reactive Protein; ED, Emergency Department; ESR, Erythrocyte Sedimentation Rate; max,
maximum; min, minimum; PCT, Procalcitonin; Temp, Temperature; WBC, White Blood Cell Count.

confirmation of infection occurred in only 37% of cases.15 The findings of this study emphasize that parameters
Accurate diagnosis of children with signs and symptoms of and threshold values commonly used for this purpose,
musculoskeletal infection is challenging due to the tre- individually or in aggregate, all have limitations. It is not
mendous overlap of symptoms, physical findings, and in- surprising that the accuracy was low, ranging from 48.2%
flammatory markers between infectious and noninfectious to 84.7%. The greatest contribution of the parameters
conditions. Daniel Kahneman recently explored the extent determined in this study was their negative predictive
to which judgment under uncertainty, particularly pre- value. As such, providers should generally trust negative
dictive judgment, is subject to noise, bias, and objective results and be reassured that ongoing conservative ob-
ignorance in modern healthcare.9 Within the past 2 dec- servation is reasonable in the presence of normal results or
ades, there has been a diligent search for strategies to whenever the values are well below the cutoff levels, which
systematically reduce these errors through the formulation this study identified.
of guidelines and decision algorithms.2–11 Investigators This study also found that the accuracy of parame-
have attempted to apply these algorithms to differentiate ters varied based on cutoff values and when multiple pa-
infection from other conditions.2–11 This is necessary not rameters were used in combination. However, even with
only to reduce diagnostic variability but also to assist multiple risk factors, the overall accuracy did not exceed
human judgment, particularly for providers with less ex- 90%. It is, therefore, not the intention of this study to
perience. It is a cautionary tale, however, that guidelines propose yet another algorithm with new thresholds and
and algorithms are potentially misleading and may in- probabilities. Rather, the purpose is to demonstrate the
crease the risk of unnecessary hospitalizations and in- facility by which PCT was introduced into the systematic
vasive procedures on 1 extreme or delay in diagnosis and work-up for musculoskeletal infection at a tertiary pe-
progression of infection on the other.10,11 diatric medical center and its relative merit as a biomarker
for infection. PCT appears to value during musculoskel-
etal infection evaluations, but this study demonstrates its
TABLE 3. Logistic Regression Modeling of Deep Infection
Versus Non-Infection or Superficial Infection TABLE 4. ROC Analysis of Parameters to Predict Deep Infection
Logistic Regression Modeling ROC Analysis
Deep Infection versus Superficial Infection or Noninfection Outcome: Deep Infection (Deep Infection vs. Superficial Infection or
Odds Ratio 95% Wald CI Noninfection)
Variable Estimates of OR P N Sensitivity Specificity Cutoff AUC
Max Recorded Temp 0.16 0.46 0.06 0.00 Max Temp 74 0.89 0.86 37.8 0.934
in ED in ED
ESR 0.99 1.03 0.96 0.74 ESR 85 0.90 0.78 18.0 0.901
CRP 0.88 0.99 0.79 0.03 CRP 83 0.76 0.84 3.3 0.866
PCT 0.20 1.05 0.04 0.06 PCT 85 0.90 0.81 0.1 0.852
CI indicates Confidence Interval; CRP, C-Reactive Protein; ED, Emergency AUC indicates Area Under the Curve; CRP, C-Reactive Protein; ED, Emer-
Department; ESR, Erythrocyte Sedimentation Rate; Max, Maximum; OR, Odds gency Department; Max, Maximum; PCT, Procalcitonin; ROC, Receiver Operator
Ratio; PCT, Procalcitonin; Temp, Temperature. Characteristics; Temp, Temperature.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e171

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Laan et al J Pediatr Orthop  Volume 43, Number 2, February 2023

TABLE 5. Test Characteristics of Commonly Assessed Parameters to Evaluate Children for Deep Infection
Test Characteristics of Commonly Assessed Parameters, Individually and Combined
Variable Sensitivity, % Specificity, % PPV, % NPV, % Accuracy, %
Inability to bear weight 76.2 39.1 29.1 83.3 48.2
Fever ≥ 38.5C 81.0 82.8 60.7 93.0 82.4
WBC ≥ 12.0 52.4 73.4 39.3 82.5 68.2
ESR ≥ 40 75.0 93.8 75.0 87.0 84.7
Kocher Criteria ≥ 2 95.2 73.4 54.1 97.9 78.8
CRP ≥ 2.0 85.7 71.9 50.0 93.9 75.3
Kocher+Caird Criteria ≥ 3 90.5 84.4 65.5 96.4 85.9
PCT > 0.10 90.4 84.4 64.3 94.7 84.7
Admission 90.5 73.4 52.8 95.9 77.6
Temp,CRP > 3.3,ESR > 18,PCT > 0.1 ( > 2) 81.0 92.2 77.3 93.7 89.4
CRP indicates C-Reactive Protein indicates; ESR, Erythrocyte Sedimentation Rate; NPV, Negative Predictive Value; PPV, Positive Predictive Value; PCT, Procalcitonin;
ROC, Receiver Operator Characteristics; Temp, (temperature > 37.8C); WBC, White Blood Cell Count.

limitations, which are similar to that of other commonly
used parameters.
PCT has been utilized to diagnose serious bacterial
infections in neonates, children, and adults with sepsis and
pneumonia.13,14,16,17 One study showed a PCT cutoff
value of 0.2 ng/mL with the sensitivity of 100% and
specificity of 94.4% in diagnosing septic arthritis.18 In our
study, PCT of > 0.1 ng/mL had the sensitivity of 90.4%,
specificity of 84.4%, and AUC of 0.852 for deep infection.
These findings are similar to another study reporting an
AUC of 0.72 with cutoff values of PCT > 0.1 ng/mL,
ESR > 19.5 mm/hr, and temperature > 37.2°C being
twice as likely to identify musculoskeletal infection.19
Given limited evidence available to support the use of
PCT to assess pediatric infections, the Pediatric Infectious
Disease Society guideline for osteomyelitis did not rec-
ommend its routine use due to insufficient evidence in its
support.20 Data from our study contribute to the ongoing
assessment of PCT for these evaluations. In agreement
with the guidelines, our study confirms that more data is
needed.
Experience aggregated at this center suggests that a
systematic approach is useful to guide decisions during
musculoskeletal infection assessments. Providers should
conduct a careful history and physical examination and
acquire the full panel of initial labs, including complete
blood count with differential, CRP, ESR, PCT, and blood
culture. While it may seem trivial to mention history and
physical examination as part of this systematic approach,
the ability to rapidly recognize certain conditions using
history and physical findings should not be discounted.
This approach minimizes the tendency of parameter-based
decision algorithms to overly focus on numerical values or
thresholds with less attention to the bigger picture fol-
lowing a comprehensive evaluation of the child.
Laboratory results should be reviewed with mind-
fulness of the lowest reportable lab value and range of
each parameter that might be anticipated in healthy chil-
dren. These values may differ from 1 reference lab to
another, but at our center, they are CRP <0.4 mg/dL;
ESR ~ 4 to 8 mm/hr; and PCT <0.04 ng/mL. Next, rea-
sonable cutoff or threshold values should be considered
to establish a level of concern regarding the relative
magnitude of elevation of the child’s labs. With this ac-
complished, the provider should have an informed in-
tuition regarding the possibility of deep infection. This
should guide the decisions for admission and discharge
from the ED with planned follow-up in subspecialty
clinics, with the primary care physician, or only as needed.
The 72.5% rate of acquisition of all desired labo-
ratory studies and 77.6% rate of admission accuracy are
indications that more work is needed to improve these
processes at this institution. Providers demonstrate varia-
tion in tendencies to order and review a variety of pa-
rameters to help with the judgment of infection cases.
Given that there are over 40 ED staff at this institution
and recognizing that not all providers evaluate patients the
same way, the rate of PCT acquisition and admission are
enlightening as to the potential challenges to the adoption
of these principles at any center.
This study has several limitations. The sample size
was smaller than intended during the study design, with
enrollment during the peak of the COVID-19 outbreak
when ED volumes of viral and bacterial infections were
impacted by societal measures of hygiene and social-
distancing. This lowered the musculoskeletal infection
consultation rate to less than half of the historical average.
Another limitation was initial slow enrollment when PCT
ordering was not the standard practice. As the adoption of
PCT increased at our center, pediatric hospitalists and in-
tensivists have found value in trending PCT in cases of deep
infection. A declining PCT enables the recognition of the
effectiveness of therapy before the decline of CRP. This is
consistent with the findings of other investigators who have
reported the benefit of PCT due to its rapid decline in the
presence of effective antibiotic treatment.14,21,22
CONCLUSION
PCT is a potentially useful inflammatory marker
during the evaluation of children with suspected
musculoskeletal infection. A combination of parameters
gathered during the systematic assessment of the child ap-
pears more helpful in supporting the decision-making and
predictive judgment under uncertainty. PCT <0.1 ng/mL,
ESR <18 mm/hr, CRP <3.3 mg/dL, and temperature

e172 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023
<37.8°C should reasonably reassure clinicians that deep
musculoskeletal infection is less likely, given the high neg-
ative predictive value and accuracy of these parameters at
the proposed thresholds. For children with a low risk of
deep infection, it is more appropriate to consider 1) the
outpatient follow-up with a subspecialist for a second look
or 2) the follow-up with the primary care physician or as
needed. This strategy has been employed at our institution
for over 10 years while practicing under guidelines. Annual
stakeholder updates are given to ED providers to empha-
size that orthopaedic clinic follow-up is appropriate for
children sent out from the ED when there is a preliminary
musculoskeletal concern but insufficient to warrant admis-
sion. The purpose is to allow for additional assessment until
a more definitive diagnosis can be determined or, alter-
natively, spontaneous resolution is reached. An essential
lesson of this study is that trending lab values, either in the
inpatient or outpatient setting, improve diagnostic accuracy
and decision-making over time. Such a practice extends the
process of evaluation over hours or days to help differ-
entiate infection from noninfection.
REFERENCES
1. Pierrie SN, Scannell BP, Brighton BK, et al. Characteristics of
Pyogenic musculoskeletal infections in older children and adoles-
cents. Orthopedics. 2020;43:e291–e298.
2. Kocher MS, Zurakowski D, Kasser JR. Differentiating between
septic arthritis and transient synovitis of the hip in children: an
evidence-based clinical prediction algorithm. J Bone Joint Surg Am.
1999;81:1662–1670.
3. Kocher MS, Mandiga R, Zurakowski D, et al. Validation of a
clinical prediction rule for the differentiation between septic arthritis
and transient synovitis of the hip in children. J Bone Joint Surg Am.
2004;86:1629–1635.
4. Caird MS, Flynn JM, Leung YL, et al. Factors distinguishing septic
arthritis from transient synovitis of the hip in children. a prospective
study. J Bone Joint Surg Am. 2006;88:1251–1257.
5. Herman MJ, Martinek M. The limping child. Pediatr Rev. 2015;36:
184–195; quiz 196-7.
6. Hwang C. Calculated decisions: Kocher criteria for septic arthritis.
Pediatr Emerg Med Pract. 2019;16:Cd1–cd2.
7. Ryan DD. Differentiating transient synovitis of the hip from more
urgent conditions. Pediatr Ann. 2016;45:e209–e213.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Procalcitonin as Deep Infection Biomarker
8. Mo M, Guilak F, Elward A, et al. The use of biomarkers in the early
diagnosis of septic arthritis and osteomyelitis-A pilot study. J Pediatr
Orthop. 2022;42:e526–e532.
9. Kahneman DSOSCR, Noise: a flaw in human judgement. 2022.
10. Luhmann SJ, Jones A, Schootman M, et al. Differentiation between
septic arthritis and transient synovitis of the hip in children with
clinical prediction algorithms. J Bone Joint Surg Am. 2004;86:
956–962.
11. Good JJ, Rabener MJ, Fisher GW. Using a decision tool to evaluate
for osteomyelitis in children. Jaapa. 2021;34:29–32.
12. Balog A, Ocsovszki I, Mándi Y. Flow cytometric analysis of
procalcitonin expression in human monocytes and granulocytes.
Immunol Lett. 2002;84:199–203.
13. Simon L, Gauvin F, Amre DK, et al. Serum procalcitonin and
C-reactive protein levels as markers of bacterial infection: a
systematic review and meta-analysis. Clin Infect Dis. 2004;39:
206–217.
14. Assicot M, Gendrel D, Carsin H, et al. High serum procalcitonin
concentrations in patients with sepsis and infection. Lancet.
1993;341:515–518.
15. Koehler RJ, Shore BJ, Heyworth BE, et al. Defining the volume of
consultations for musculoskeletal infection encountered by pediatric
orthopaedic services in the United States. PLoS One. 2020;15:
e0234055.
16. Kamat IS, Ramachandran V, Eswaran H, et al. Procalcitonin to
distinguish viral from bacterial pneumonia: a systematic review and
meta-analysis. Clin Infect Dis. 2019;70:538–542.
17. Omar J, Isa S, Ismail TST, et al. Procalcitonin as an early laboratory
marker of sepsis in neonates: variation in diagnostic performance and
discrimination value. Malays J Med Sci. 2019;26:61–69.
18. Fottner A, Birkenmaier C, Pellengahr C, et al. Can serum
procalcitonin help to differentiate between septic and nonseptic
arthritis? Arthroscopy. 2008;24:229–233.
19. McMichael BS, Nickel AJ, Christensen EW, et al. Discriminative
accuracy of procalcitonin and traditional biomarkers in pediatric
acute musculoskeletal infection. Pediatr Emerg Care. 2021;37:
e1220–e1226.
20. Woods CR, Bradley JS, Chatterjee A, et al. Clinical practice
guideline by the pediatric infectious diseases society and the
infectious diseases society of America: 2021 guideline on diagnosis
and management of acute hematogenous osteomyelitis in pediatrics.
J Pediatr Infect Dis Soc. 2021;10:801–844.
21. Schuetz P, Christ-Crain M, Müller B. Procalcitonin and other
biomarkers to improve assessment and antibiotic stewardship in
infections–hope for hype? Swiss Med Wkly. 2009;139:318–326.
22. Schuetz P, Albrich W, Mueller B. Procalcitonin for diagnosis of
infection and guide to antibiotic decisions: past, present and future.
BMC Med. 2011;9:107.
www.pedorthopaedics.com | e173
 ORIGINAL ARTICLE
Bethlem Myopathy (Collagen VI-Related Dystrophies):
A Retrospective Cohort Study on Musculoskeletal
Pathologies and Clinical Course
Rachel S. Silverstein, MD, MBS,*† Daniel D. Wang, MD,* Lee S. Haruno, MD,‡
Timothy E. Lotze, MD,*† Allison C. Scott, MD,§ and Scott B. Rosenfeld, MD*†
Background: Collagen VI-related myopathies with pathologic
COL6A1, COL6A2, and COL6A3 variants manifest as a pheno-
typic continuum of rare disorders, including Bethlem myopathy
(BM), characterized by early onset muscle weakness, proximal joint
contractures, and distal joint laxity. Herein we discuss the con-
comitant orthopedic manifestations of BM, potential management
strategies, and patient outcomes.
Methods: An IRB-approved retrospective cohort study (n = 23)
from 2 pediatric institutions with a confirmed diagnosis of BM.
Charts were reviewed for demographic data, age of disease pre-
sentation and diagnosis, COL6 genotype, diagnosis method,
ambulation status, need for assistance, musculoskeletal abnor-
malities, other systemic comorbidities, advanced imaging
and screening diagnostics, previous surgical interventions, and
progression of the disease.
Results: The mean age was 11.65 years (range 3 to 19 y). Mean
age at initial presentation with symptoms was 4.18 years old,
whereas diagnosis was delayed until 8.22 years old on average.
Muscle weakness was the most common presenting symptom
(65.2%), and 73.9% of patients required some use of assistive or
mobility devices. Overall, 30.4% of patients were diagnosed with
scoliosis; 57.1% required operative intervention for their sco-
liosis; 43.5% of patients had acetabular dysplasia; 10% required
open reduction of a dislocated hip; 10% required closed reduc-
tion with hip spica application; 10% required bilateral peri-
acetabular osteotomies for instability; 91.3% of patients
developed foot and ankle deformities; 33.3% of patients under-
went posteromedial-lateral equinovarus releases; 28.6% required
an Achilles tendon lengthening, and 86.9% of patients had
From the *Department of Orthopaedic Surgery, Baylor College of
Medicine; †Department of Pediatric Orthopaedic Surgery, Texas
Children’s Hospital; §Department of Pediatric Orthopedic Surgery,
Shriners Children’s Texas, Galveston TX; and ‡Department of Or-
thopaedic Surgery, Cedar-Siani Medical Center, Los Angeles, CA.
All authors declare no conflicts of interest or conflicts of copyright. Each
author certifies that his or her institution has approved the animal and
human protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research. No in-
formed consent was required for this study.
The authors declare no conflicts of interest.
Reprints: Rachel S. Silverstein, MD, MBS, Texas Children’s Hospital,
Department of Pediatric Orthopaedic Surgery, 6701 Fannin Street,
Houston TX 77030. E-mail: rssilverstein@gmail.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002283
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
muscle tendon contractures, the most common locations being
the ankle (55%) and elbow (40%).
Conclusion: Although often less severe than other more common
neuropathies and myopathies like Charcot-Marie-Tooth disease
and Duchenne muscular dystrophy, BM does lead to progressive
musculoskeletal deformity and disability. Its relative rarity
makes it less familiar to providers and likely contributes to delays
in diagnosis. Scoliosis, hip dysplasia, and equinus and varus
ankle deformities are the most common musculoskeletal
deformities. Physicians and surgeons should appropriately
counsel patients and families about the clinical course of this
disorder and the potential need for mobility assistance or surgical
procedures.
Level of Evidence: III, Prognostic. study.
Key Words: bethlem myopathy, collagen VI myopathy, myopathy,
pediatrics
(J Pediatr Orthop 2023;43:e163–e167)
C ollagen VI-related myopathies with a mutation in the
COL6 genes manifest as a phenotypic continuum of
disorders ranging from mild Bethlem myopathy to severe
Ulrich congenital muscular dystrophy.1,2 Bethlem myo-
pathy (BM) is a rare connective tissue disorder, with an
incidence of 0.77 per 100,000 individuals, that was first
described in 3 families in 1976 by Bethlem and van
Wijngaarden.1,3–5 It typically exhibits an autosomal dom-
inant inheritance pattern but has also been reported as an
autosomal recessive or de novo gene mutation.3,4,6 Colla-
gen VI consists of 3 major peptides (COL6A1, COL6A2,
and COL6A3), and mutations in each have been shown to
produce disease.1,7,8 BM is characterized by delayed motor
skills in the first and second decades of life and a number of
musculoskeletal manifestations, including increased joint
laxity, flexion contractures of the ankles, elbows, fingers
and/or wrists, and muscle weakness, most often in the
proximal muscles of the upper extremity.2–4,6,9,10 Most
children tend to exhibit features of neuromuscular differ-
ences, including atypical crawling and delayed motor
milestones by the age of 2, which continue to develop in
severity until the age of 5 or older.11–13 Because of slow
disease progression, previous reports suggest that most
individuals do not have significant ambulatory problems
until the fourth or fifth decades of life.14 However, many
www.pedorthopaedics.com | e163
Silverstein et al
require assisted devices, including walking aids, pelvic
girdles, or shoulder girdles.
Patients with neuromuscular diseases are often seen by
orthopaedic surgeons due to muscle tendon contractures.
Although patients with known neuromuscular conditions are
often seen for consultation, many patients are referred by
their primary care physician without prior recognition of a
neuromuscular etiology for the contractures. Orthopaedic
surgeons familiar with common conditions associated with
contractures, including Duchenne-Becker muscular dys-
trophy (DMD) and Charcot-Marie-Tooth (CMT) neuro-
pathies, might subsequently recommend a more immediate
assessment by a neurologist to assist in the diagnostic eval-
uation of a suspected neuromuscular condition. However,
less common neuromuscular disorders, such as collagen VI-
related myopathies, might not be recognized in the ortho-
paedic evaluation, which can lead to uncertainty in the eti-
ology of worsening or recurring contractures following
surgical repair.
There are around 30 known types of muscular dys-
trophies. The constellation and timing of symptoms seen in
BM are similar to those of more commonly encountered,
well-studied pathologies such as Duchenne muscular dys-
trophy, congenital muscle dystrophy, limb-girdle muscular
dystrophy, and Emery-Dreifuss muscular dystrophy. The
patient presentation could also be similar to other known
muscle neuropathies such as Charcot-Marie-Tooth disease
or known disorders such as arthrogryposis, Ehlers-Danlos
syndrome and Marfans syndrome. However, due to BM’s
relative rarity, clinicians often initiate workups to rule out
more common neuromuscular or musculoskeletal diseases.
The early finding of distal hyperlaxity with low/normal
creatine kinase levels are unique to BM in the setting of
muscle weakness and proximal joint contractures but are
not always present on the initiation of workup.
The diagnosis of BM is commonly made through
genetic testing and/or muscle biopsy. Electromyography is
often ordered as part of the workup for muscular dystro-
phies; however, for BM, it has been shown to have generic
myopathic features.6 Creatine kinase is part of the lab
panel, and unlike some other common muscle dystrophies,
will be either normal or mildly elevated.6 Progress has
been made with regard to understanding BM’s genetic
etiology, but due to the rarity of the disease, there is sparce
literature concerning the spectrum of musculoskeletal
manifestations of patients affected by BM. In majority of
the literature published, the subjects are related groups of
families of all ages and or more genetic based.1–3,7,15
This study involves a large cohort of nonfamilial BM
pediatric patients seen at 2 different academic pediatric
orthopaedic hospitals. The objectives of this report are
to describe the disorder’s clinical course, broaden the
description of musculoskeletal manifestations, and to
describe necessary treatments.
MATERIALS AND METHODS
A retrospective study was conducted in a cohort
of 23 patients from 2 pediatric institutions, all with
e164 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
confirmed diagnoses of BM through genetic testing or
muscle biopsy. Patients included in the study were
18 years or younger at the time of diagnosis. This study
protocol was approved by both hospitals’ internal review
boards. Charts were reviewed for demographic data, age
at initial presentation and diagnosis, method of genetic
diagnosis, COL6 genotype, presence of subjective pain at
presentation, ambulation status, level of independence
with regard to activities of daily living, dependence on
assistive devices for mobility, previous surgical inter-
ventions, musculoskeletal abnormalities, neurological
findings, advanced imaging and screening diagnostics,
other systemic comorbidities, and progression of the
disease.
RESULTS
The mean age of the study group at the time of
review was 11.65 years old, ranging from 3 to 19 years old.
Patients initially presented with symptoms around
4.18 years old, whereas diagnosis was delayed until the
average age of 8.22 years old. The time period between
initial presentation and diagnosis was, on average,
4.09 years. Ten patients were diagnosed by genetic testing
and muscle biopsy, 12 patients were diagnosed by genetic
testing only, and 1 patient was diagnosed only by muscle
biopsy. Presenting symptoms included muscle weakness in
15 (65.2%) patients and hypotonia in 5 (21.7%) patients.
Only 5 (21.7%) patients presented with a chief complaint
of pain. On physical examination, Gower sign was docu-
mented in 5 (21.7%) patients, and muscle atrophy was
found in 5 (21.7%) patients.
Six (26.0%) patients did not require any assistive
devices for ambulation. Seventeen (73.9%) patients re-
quired some use of assistive or mobility devices such as
braces, walkers, or wheelchairs. Of those 17, 4 (23.5%)
patients required bracing. Six (35.3%) patients required a
walker for mobility. Seven (41.2%) patients required a
wheelchair for mobility. (Table 1)
Eighteen (78.3%) patients underwent at least 1 sur-
gical procedure, the most common being muscle biopsy
for diagnosis (73.33%). The most common musculoskel-
etal manifestations included scoliosis, hip dysplasia, and
ankle deformities. Seven (30.4%) patients were diagnosed
with scoliosis, and four (57.1%) required operative inter-
vention with spinal fusion. Ten (43.5%) patients were
found to have acetabular dysplasia, 1 (10%) patient re-
quired closed hip reduction and spica casting, 1 (10%)
required open reduction of a dislocated hip, and 1 (10%)
required bilateral periacetabular osteotomies for in-
stability. Twenty-one (91.3%) patients developed foot and
ankle deformities such as hindfoot varus, equinovarus,
cavus, and pes planus. Seven (33.3%) patients underwent
posteromedial-lateral equinovarus releases, and 6 (28.6%)
required Achilles tendon lengthening.
Twenty (86.9%) patients had muscle contractures on
clinical exam, the most common locations being the ankle
(55%) and the elbow (40%). No patients were found to
have any cardiac or pulmonary abnormalities (Fig. 1).
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Bethlem Myopathy

TABLE 1. Bethlem Myopathy Patient Activity Level, Assistive Devices, and Abnormalities
Age ADL’s/Activity Level Assistive Devices Abnormality
4 Some recreational activity None Foot/ankle
5 Some recreational activity None None
11 Full and/or unrestricted activity None Hip, Foot/ankle
17 Some recreational activity None Foot/ankle
17 ADLS without assistance None Foot/ankle
18 ADLS without assistance None Foot/ankle
8 Some recreational activity Orthoses Foot/ankle
10 Some recreational activity Orthoses Spine, Hip, Foot/ankle
17 Full and/or unrestricted activity Orthoses Foot/ankle
18 ADLs restricted requiring some assistance Orthoses Spine, Foot/ankle
8 ADLS without assistance Walker, orthoses Hip, Foot/ankle
3 ADLs restricted requiring some assistance Walker Foot/ankle
4 ADLs restricted requiring some assistance Walker Hip, Foot/ankle
6 ADLs restricted requiring some assistance Walker Hip, Foot/ankle
12 ADLS without assistance Walker Hip, Foot/ankle
19 Some recreational activity Walker Foot/ankle
9 ADLs restricted requiring some assistance Wheelchair in some settings Hip, Foot/ankle
13 ADLS without assistance Wheelchair in some settings Spine, Foot/ankle
7 Severely restricted/requiring full time assistance Wheelchair in most/all settings Hip, Foot/ankle
11 ADLs restricted requiring some assistance Wheelchair in most/all settings Spine, Hip
14 Severely restricted/requiring full time assistance Wheelchair in most/all settings Spine, Hip, Foot/ankle
18 Severely restricted/requiring full time assistance Wheelchair in most/all settings Spine, Foot/ankle
19 ADLs restricted requiring some assistance Wheelchair in most/all settings Spine, Foot/ankle

DISCUSSION indicate a delay in diagnosis of only 2.5 years.16 The delay

This study sought to characterize the musculoskel-
etal manifestations of BM in a pediatric population of
nonfamilial patients. Making this diagnosis can be chal-
lenging for patient families and clinicians when no history
of this disease is present in the family. This report can
assist the clinician in 2 major ways. First, it can help the
clinician to better recognize the manifestations of this rare
condition to make an early and accurate diagnosis of BM,
especially with patients presenting with muscle weakness,
joint contractures, and distal hyperlaxity. Secondly, it can
shed light on the common musculoskeletal manifestation
of BM and enable the clinician to provide families with
more detailed expectations about their child’s clinical
course and better understand what to screen for as
they age.
Previous reports suggest that children begin to
develop symptoms and signs of BM in their first 2 years of
life, and symptoms peak around 5 years old.3,6 This was
corroborated in our study, with patients initially present-
ing with symptoms around 4 years old. Most presented
with signs of muscle weakness and hypotonia. Our study
found muscle weakness in 65% of the population and
hypotonia in 22%. This later progressed to muscle
contractures, most notably elbow and Achilles tendon
contractures.
Our population had a delay of diagnosis on average
of 4 years from the onset of symptoms. This is likely due to
less provider familiarity with this relatively rare disease.
Although BM has some characteristic findings with
proximal muscle weakness and flexion contractures, if a
patient has a milder presentation, it can mimic other more
common congenital muscular dystrophies. Reports on
other more common myopathies and neuropathies, such
as (DMD, CMT) that have similar presentations to BM,
in diagnosis could also be due to the fact that our pop-
ulation of BM is nonfamilial. Thus, none of our patients
had any known family members with BM to help guide
clinicians to this specific disease.
Part of the workup is musculoskeletal imaging and
muscle biopsy. On imaging, muscle involvement shows
fatty and connective tissue infiltration mostly in proximal
muscles.1,15 However, on muscle biopsy, this has been
found to be variable ranging from normal to myopathic
changes to atrophic muscle fibers with the build-up of fat
and connective tissue.1,6 Although muscle biopsy and
immunohistochemistry may be part of the workup and the
majority of our patients underwent this procedure, it is less
helpful in the diagnosis of this disease, especially in the
younger patient population. Molecular genetic testing,
although more expensive, is more widely accepted in
diagnosing BM.6
BM has a phenotypic spectrum of presentation in
which patients may demonstrate a combination of joint
contractures and hypermobility. Patients with severe dis-
ease will have congenital weakness and hypotonia with
proximal joint contractures and marked distal joint
laxity.17 On the milder end of the spectrum, patients might
not have obvious laxity at all and tend to develop con-
tractures later in life.9 Similar variability is found in the
ambulatory function, with some patients experiencing
deficits in the first/second decades of life, whereas others
only encounter significant issues much later (fourth or fifth
decade).
We found that our study population demonstrated
motor impairment progression requiring assistive devices
such as walkers or wheelchairs even when initially pre-
senting with only mild motor impairment differences.
However, there was no overall trend with regard to age.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e165

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Silverstein et al
FIGURE 1. Bethlem Myopathy Musculoskeletal Manifestations.
Some patients required ambulation aids and had restricted
activity levels as early as 3 years old and other patients at
the age of 18 were able to participate in all activity levels
and required no assisted devices. We found patients pro-
gressed with requiring assistance but found that it was not
age-specific. This is consistent with previous reports that
BM is slowly progressive, with ~60% of patients requiring
aids for ambulation over the age of 50.6,15 Jobis and col-
leagues found that 13/36 patients required a walking aid
such as a walking cane, and 14/36 needed a wheelchair. In
the same study, 73% of patients over the age of 50 required
a wheelchair for mobility.3 Since BM is generally consid-
ered to be a relatively mild form of myopathy, these
findings are important as patients and their families should
be counseled that the early and lifelong use of an assistive
device may be necessary.
In our study, the most common musculoskeletal
manifestations were foot and ankle deformities, followed
by ankle and elbow contractures, acetabular dysplasia,
and scoliosis. These findings are similar to previous re-
ports. An international workshop on BM described mul-
tiple families with Achilles tendon contractures and other
foot deformities such as pes cavus and contractures of the
upper extremity, including the elbows and fingers.15 Jobsis
et al3 found foot contractures such as equinovarus were
commonly present and that the progression of this de-
formity occurred over time, leading to subsequently tight
heel cords. The most common physical manifestations in
our study population that required surgical treatment were
foot and ankle deformities. This was also reported in the
literature, where Achilles tenotomies were required to help
improve ambulation.3
There have been few documented cases of scoliosis
in patients diagnosed with BM.15,18–20 Our study cohort
had the most patients with scoliosis (7, 30%), however,
only 57% of these individuals required posterior in-
strumentation and fusion compared with 100% of the
patients in Dede et al18 cohort of 3 patients.
Hip dysplasia in BM has been described in a few
familial case reports.3,21 Bönnemann found that hip dys-
plasia/dislocations were seen in about 50% of patients with
BM.1 Our population had 10 patients with hip dysplasia, 2
required intervention for dislocated hips, and 1 required
an osteotomy for instability. The remainder of patients
e166 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
with hip dysplasia did not have dislocation. It remains to
be seen whether these patients will, over time, develop
symptoms from hip dysplasia.
Both scoliosis and hip dysplasia should be screened
for all patients with a known diagnosis of BM. Given that
MSK manifestations are common, the authors recom-
mend that patients with BM should undergo a yearly
physical examination by a provider experienced in MSK
examinations. This should include yearly spine physical
examinations, including Adams forward bend. If the
physical examination is suggestive of scoliosis, then spine
radiographs should be obtained. In patients who are
wheelchair bound, we recommend yearly scoliosis radio-
graphs. Since nearly half of our patients with BM had hip
dysplasia, screening for this should also be considered. In
our population, hip dysplasia was diagnosed in the age
range of 4 to 14 years old. Therefore, we recommend in-
itial screening pelvis radiographs upon diagnosis of BM,
regardless of age at the time of diagnosis. After that,
yearly examinations should include evaluation of the hips
and gait, with radiographs obtained as needed based on
physical examination findings. However a small number
required intervention for hip dysplasia in our cohort, and
less than half required surgical intervention.
Families of children diagnosed with BM should
be counseled that their child may require the use of some type
of assistive device for mobility as they age, and if they have
an abnormality of the hip, spine, and or foot/ankle, they are
more likely to require an assisted device. Children with a
spine and or hip abnormality might also be more likely to be
lower functioning with activities and require more help for
ADL’s, however, there is no association of function/activity
or joint abnormality with regards to age. Families can also
be reassured that BM patients do not appear to have an
increased risk for cardiac comorbidities. Surgical inter-
vention for these conditions may decrease symptomatic pain
and improve cosmesis, but their value in preserving or
increasing functional status requires further study.
This study has limitations, the most significant of
which is the retrospective nature of the data collection. In
addition, we were unable to document the age of initiation
of orthotics for each patient. However, our data has a
wide age group of patients requiring orthotics (3 to 19 y
old); unlike the relatively predictable timeline for loss of
ambulation for patients with DMD, there is a less clear
onset of such symptoms in patients with BM and thus
there might not be a common age initiation of orthotics in
these patients. Another limitation of this study is in pa-
tients diagnosed with acetabular dysplasia—we were un-
able to document the date of diagnosis, whether they had
screening radiographs, or the age at which screening ra-
diographs occurred. We were also unable to obtain what
prompted the workup, whether they had any symptoms
leading up to the radiographs and/or workup, or whether
there was a progression of acetabular dysplasia for the
patients who did not require surgical intervention. There
was additionally no documentation of screening radio-
graphs available for the other patients without a diagnosis
of acetabular dysplasia.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
CONCLUSION
This is the largest cohort of nonfamilial BM patients
to be reported in the literature. Although physical mani-
festations are generally less severe than those of the sim-
ilarly presenting but more common neuropathies and
myopathies, including CMT and DMD, BM does lead to
progressive musculoskeletal deformity and disability. Its
relative rarity makes it less familiar to providers and likely
contributes to a delay in diagnosis of 4 years from symp-
tom onset. This study adds to the body of research re-
garding the rare collagen VI disorder, BM. Our cohort of
patients demonstrated wide phenotypic variability but
commonly features significant and early ambulatory im-
pairments. We also found a greater number of patients
with scoliosis and hip dysplasia than previously reported in
the literature. Clinicians must take care to appropriately
counsel patients and families about the possible clinical
course of this disorder and potential needs for mobility
assistance or surgical procedures, while emphasizing the
true spectrum of disease severity. BM should be considered
when establishing a differential diagnosis for an atypical
neuromuscular disorder/myopathy early in life and evalu-
ation or confirmation of diagnosis with genetic testing.
REFERENCES
1. Bonnemann CG. The collagen VI-related myopathies Ullrich
congenital muscular dystrophy and Bethlem myopathy. Handb Clin
Neurol. 2011;101:81–96.
2. Hicks D, Lampe AK, Barresi R, et al. A refined diagnostic algorithm
for Bethlem myopathy. Neurology. 2008;70:1192–1199.
3. Jobsis GJ, Boers JM, Barth PG, et al. Bethlem myopathy: a slowly
progressive congenital muscular dystrophy with contractures. Brain.
1999;122 (Pt 4):649–655.
4. Bethlem J, Wijngaarden GK. Benign myopathy, with autosomal
dominant inheritance. A report on three pedigrees. Brain. 1976;99:91–100.
5. Norwood FL, Harling C, Chinnery PF, et al. Prevalence of genetic
muscle disease in Northern England: in-depth analysis of a muscle
clinic population. Brain. 2009;132(Pt 11):3175–3186.
6. Deconinck N, Stojkovic T. Ullrich congenital dystrophy and bethlem
myopathy: avenues of collagen vi related muscular dystrophies. Cur
Pediatr Rev. 2009;5:28–35.
7. Deconinck N, Richard P, Allamand V, et al. Bethlem myop-
athy: long-term follow-up identifies COL6 mutations predicting
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Bethlem Myopathy
severe clinical evolution. J Neurol Neurosurg Psychiatry. 2015;86:
1337–1346.
8. Lampe AK, Bushby KM. Collagen VI related muscle disorders.
J Med Genet. 2005;42:673–685.
9. de Visser M, de Voogt WG, la Riviere GV. The heart in Becker
muscular dystrophy, facioscapulohumeral dystrophy, and Bethlem
myopathy. Muscle Nerve. 1992;15:591–596.
10. Merlini L, Morandi L, Granata C, et al. Bethlem myopathy: early-
onset benign autosomal dominant myopathy with contractures.
Description of two new families. Neuromuscul Disord. 1994;4:
503–511.
11. Arts WF, Bethlem J, Volkers WS. Further investigations on benign
myopathy with autosomal dominant inheritance. J Neurol. 1978;217:
201–206.
12. Mohire MD, Tandan R, Fries TJ, et al. Early-onset benign
autosomal dominant limb-girdle myopathy with contractures (Beth-
lem myopathy. Neurology. 1988;38:573–580.
13. Somer H, Laulumaa V, Paljarvi L, et al. Benign muscular dystrophy
with autosomal dominant inheritance. Neuromuscul Disord. 1991;1:
267–273.
14. Serratrice G, Pellissier JF. [2 families with benign myopathy
predominantly on the limb girdle with dominant autosomal
heredity]. Rev Neurol (Paris). 1988;144:43–46; Deux familles de
myopathies benignes predominant sur les ceintures d’heredite
autosomique dominante.
15. Pepe G, Lucarini L, Zhang RZ, et al. COL6A1 genomic deletions in
Bethlem myopathy and Ullrich muscular dystrophy. Ann Neurol.
2006;59:190–195.
16. Ciafaloni E, Fox DJ, Pandya S, et al. Delayed diagnosis in duchenne
muscular dystrophy: data from the Muscular Dystrophy Surveil-
lance, Tracking, and Research Network (MD STARnet. J Pediatr.
2009;155:380–385.
17. Bertini E, Pepe G. Collagen type VI and related disorders: Bethlem
myopathy and Ullrich scleroatonic muscular dystrophy. Eur J
Paediatr Neurol. 2002;6:193–198.
18. Dede O, Abdel-Hamid HZ, Deeney VF. Spinal deformity in bethlem
myopathy. Spine Deform. 2014;2:143–151.
19. Li JY, Liu SZ, Zheng DF, et al. Collagen VI-related myopathy with
scoliosis alone: a case report and literature review. World J Clin
Cases. 2021;9:5302–5312.
20. Mirzashahi BKF, Gharakhan-Maleki R, Heshmatifar M. Scoliosis
correction surgery in collagen type VI dysfunction. J Orthop Spine
Trauma. 2018;4:62–64.
21. Mohassel PBC. Limb-girdle Muscular Dystrophies: Collagen VI
Related Dystrophies. In: Basil T, Darras HRJ, Monique MR, De
Vivo DC, eds. Chapter 34. Neuromuscular Disorders of Infancy,
Childhood, and Adolescence A Clinician’s Approach Second ed.
Science Direct: Academic Press; 2015:635–666.
www.pedorthopaedics.com | e167
 REVIEW ARTICLE
What is New in Pediatric Orthopaedic: Basic Science
Brian Lynch, MD,* Daniel Botros, MD,* Matthew Halanski, MD,† and James Barsi, MD*
Background: An understanding of musculoskeletal basic science
underpins most advancements in the field of orthopaedic surgery.
Knowledge of biomechanics, genetics, and molecular pathways is
integral to the understanding of the pathophysiology of disease
and guides novel treatment options to improve patient outcomes.
The purpose of this review is to provide a comprehensive and
current overview of musculoskeletal basic science relevant to
pediatric orthopaedic surgery.
Methods: Comprehensive Pubmed database searches were per-
formed for all English language articles published between Jan-
uary 2016 and November 2021 using the following search terms:
basic science, pediatric orthopaedics, fracture, trauma, spine,
scoliosis, DDH, hip dysplasia, Perthes, Legg-Calve-Perthes,
clubfoot, and sports medicine. Inclusion criteria focused on basic
science studies of pediatric orthopaedic conditions. Clinical
studies or case reports were excluded. A total of 3855 articles
were retrieved. After removing duplicates and those failing to
meet our inclusion criteria, 49 articles were included in the final
review.
Results: A total of 49 papers were selected for review based on
the date of publication and updated findings. Findings are dis-
cussed in the subheadings below. Articles were then sorted into
the following sub-disciplines of pediatric orthopaedics: spine,
trauma, sports medicine, hip, and foot.
Conclusions: With this review, we have identified many exciting
developments in pediatric orthopaedic trauma, spine, hip, foot,
and sports medicine that could potentially lead to changes in
disease management and how we think of these processes.
Level of Evidence: Level V
Key Words: basic science, pediatric orthopaedics
(J Pediatr Orthop 2023;43:e174–e178)
A good understanding of musculoskeletal basic science
underpins most advancements in the field of ortho-
paedic surgery. Knowledge of biomechanics, genetics, and
molecular pathways is integral to the understanding of the
pathophysiology of disease and guides novel treatment
options for improved patient outcomes. The purpose of
this review is to provide a comprehensive and current
From the *Stony Brook University Hospital, Stony Brook, NY; and
†Children’s Hospital and Medical Center, Omaha, NE.
The authors declare no conflicts of interest.
Reprints: James Barsi, MD, Department of Orthopaedic Surgery, HSC 18-
030, Stony Brook, NY 11794. E-mail: james.barsi@stonybrookmedicine.
edu.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002297
e174 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
overview of musculoskeletal basic science relevant to pe-
diatric orthopaedic surgery in the fields of the spine, hip,
foot, sports medicine, and trauma.
The “What is New in…” series of articles in the
Journal of Pediatric Orthopaedics has been endorsed by
the POSNA Presidential Line. Authors have been vetted
by the POSNA Publications Committee to provide experts
in each subspecialty area and to minimize any potential
personal conflicts of interest. All review articles of this
type undergo the full JPO review process to ensure the
highest quality information.
METHODS
Comprehensive Pubmed database searches were
performed for all English language articles published be-
tween January 2016 and November 2021 using the fol-
lowing search terms: basic science, pediatric orthopaedics,
fracture, trauma, spine, scoliosis, DDH, hip dysplasia,
Perthes, Legg-Calve-Perthes, clubfoot, sports medicine.
Inclusion criteria focused on basic science studies of pe-
diatric orthopaedic conditions. Clinical studies or case
reports were excluded. A total of 3855 articles were re-
trieved. After removing duplicates and those meeting our
exclusion criteria, 49 articles were included in the final
review. Articles were then sorted based on the following
sub-disciplines of pediatric orthopaedics: spine, trauma,
sports medicine, hip, and foot.
RESULTS
Spine
Much of the literature relevant to pediatric spine
conditions focused on the etiologic theories behind ado-
lescent idiopathic scoliosis (AIS).
Metabolic
The association of low bone mineral density (BMD)
with AIS development and curve progression continues to
be investigated. It is reported that 19% of AIS patients
undergoing surgery have low spinal BMD and low serum
vitamin D levels; however, no correlation between these
metabolic parameters and major curve magnitude has
been found.1 Another analysis found an increased risk for
AIS associated with a single nucleotide polymorphism in
the vitamin D receptor.2 Vitamin D receptor gene poly-
morphisms also have been linked to decreased bone
marrow density in patients.3
Osteocalcin, which is produced by osteoblasts and
used as a serum marker for bone formation, has been
studied in relation to histologic and mechanical properties
J Pediatr Orthop  Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
of bone in AIS individuals. Osteoblasts progress to the
formation of osteocytes, which are the most abundant
regulatory cell in bone metabolism. Chen and colleagues
found an association between abnormal osteocyte mor-
phology and osteocalcin levels in patients with confirmed
AIS. The relationship between osteocalcin and osteocyte
morphology was performed by harvesting trabecular bone
from the iliac wing during spinal fusion surgery for AIS.
Histologic analysis of these samples demonstrated less
numerous osteocytes and deranged architecture in samples
from patients with AIS. Serum analysis demonstrated a
significant negative correlation between osteocalcin and
major curve magnitude.4 Abnormal bone metabolism
from altered osteoblast and osteocyte function may con-
tribute to curve progression in AIS. Future studies are
needed to examine the use of osteocalcin as a serum
marker to predict curve progression.
Hormonal
Several hormones have been implicated in the
pathogenesis of AIS. The majority of studies focus on
leptin, melatonin, and grehlin. Leptin affects bone me-
tabolism through direct signaling from the brain.5,6 This
hormone acts by decreasing cancellous bone and increasing
cortical bone formation. Application of leptin to control
osteoblasts stimulated proliferation, differentiation, and
mineralization; however, these effects were not observed in
osteoblasts from AIS patients. A meta-analysis investigat-
ing the role of leptin in AIS demonstrated that although
there was no significant difference in serum leptin levels
between AIS patients and controls, there was a significant
increase in the soluble leptin receptor.7 Deficiencies in
leptin action due to abnormal receptor function and not
leptin itself may account for its role in the pathogenesis of
AIS. Another meta-analysis did find a significantly lower
level of leptin in AIS patients indicating that there is a need
for further study and clarification.8
Several studies have found an association between
melatonin and AIS but the exact etiologic underpinnings
have been elusive.3,9,10 Investigation of the relationship
between melatonin deficiency and decreased osteoblastic
function among AIS patients demonstrated that melatonin
in a dose-dependent manner leads to osteoblast apoptosis
through a mitochondrial pathway.11 Dose-dependent de-
struction of osteoblasts in melatonin-deficient individuals
may have a causative effect on dysfunctional bone me-
tabolism in individuals with AIS.
Ghrelin is a hormone that modulates osteoblast
function and differentiation and has a positive effect on
bone metabolism.12,13 It is produced by the enter-
oendocrine cells of the gastrointestinal tract.14 Increased
plasma ghrelin concentration and lower BMD were seen in
AIS patients compared with controls.15 RNA and protein
analysis revealed higher RANKL/OPG levels in AIS pa-
tients with the insensitivity of AIS cells to increasing
ghrelin concentration, indicating that dysregulation in the
ghrelin/RANKL/OPG pathway may lead to the decreased
osteogenic ability of osteoblasts. This downregulation of
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
What is New in Pediatric Orthopaedic: Basic Science
osteoblastic activity is hypothesized to contribute to the
lower bone mass seen in the AIS patient group.
Genetics
The role of fibroblast differentiation and dysfunction
in the pathophysiology of scoliosis remains unclear. Pre-
vious studies have examined several genes in the fibroblast
differentiation pathway as candidate genes for developing
scoliosis.16,17 Examination of the molecular regulatory
mechanism underlying fibroblast differentiation found
upregulation of cGMP dependent protein kinase gene,
PRKG1, which has been identified as a novel negative
regulator of osteoblast differentiation.18
Abnormal myogenesis has also been implicated in
the development of AIS. Muscle asymmetry between the
concavity and convexity of a scoliotic curve is observed
clinically. Characterization of abnormal paraspinal mus-
cle morphology demonstrated increased levels of myofiber
necrosis and oxidative stress in the paraspinal muscles of
AIS patients compared with controls.19,20 The PAX3 gene
is regulatory in the formation of limbs and myogenesis
and mutations of it have been linked to muscular defects
and vertebral column malformation. Investigation of the
expression of PAX3 in the paravertebral muscles found
significantly lower PAX3 expression in AIS patients
compared to controls.21 Within the AIS cohort, expression
of PAX3 on the concave paraspinal muscles was decreased
compared with the convex side. Furthermore, PAX3 ex-
pression was correlated with muscle volume, suggesting
that PAX3 may play a role in the abnormal paravertebral
muscle differences seen in AIS.21
Gene sequencing among AIS patients continues to
identify novel mutations in the hepatocyte growth factor,
estrogen receptor, and noncoding regions of the genome
that implicate osteogenic activity.22–25 These studies fur-
ther support the hypothesis that scoliosis is polygenic and
multifactorial and that further work is needed to identify
all of the genes associated with AIS development and
progression.
Trauma
Although physeal fractures are the most common
cause of physeal bar formation, the etiology of this oc-
currence remains incompletely understood.26 One theory
suggests that vascular disruption between the epiphyseal
and metaphyseal microcirculation may be a factor. Physis
disruption in a rat model was studied and an increase in
physeal bar formation was noted in specimens where the
basement plate of the physis was disrupted. Disruption of
the basement membrane alters the end arteriole anasto-
mosis between the epiphyseal and metaphyseal vessels. In
addition, it is theorized that the basement membrane
provides structural integrity for the remaining damaged
physis, and its disruption results in the increased like-
lihood of physeal arrest. Specimens with bar formation
also had less expression of vascular endothelial growth
factor expression within the resting and proliferative zone
suggesting that lack of vascular ingrowth may play a role
in physeal bar formation.26
www.pedorthopaedics.com | e175
Lynch et al
Sports Medicine
Juvenile osteochondritis dessicans (JOCD) continue
to be an area of research. The pathogenesis of JOCD was
studied using quantitative magnetic resonance imaging
(MRI) of epiphyseal cartilage.27 Three nonanastomosing
vascular networks (2 peripheral and 1 central) were found
to supply the distal femoral epiphysis. Early vascular re-
gression of the central network and delayed ossification
corresponded to sites with a predilection for JOCD sug-
gesting that a vascular perturbation during skeletal ma-
turation may play a role in development. OCD causes are
hypothesized to include genetic causes, issues with ossifi-
cation centers, and endocrine disorders resulting in sub-
chondral ischemia and the development of osteochondritis
dessicans (OCD).
Autologous chondrocyte implantation has become
more common in the treatment of unstable OCD lesions of
the knee. This procedure involves the harvesting of cartilage
along the intercondylar notch or trochlea and is associated
with increased donor site morbidity. Osteochondral loose
bodies are often associated with juvenile OCD and may be a
source of healthy chondrocytes. A cellular analysis of carti-
lage harvested from osteochondral loose bodies compared
with the traditional harvest site along the intercondylar notch
found no difference in viability, yield, potency, or density.17
Cartilage obtained from loose body harvest may be
a viable replacement for chondrocytes in cartilage restora-
tions procedures without the additional donor site morbidity
of an autologous chondrocyte implantation procedure.
Hip
Developmental Hip Dysplasia
Developmental dysplasia of the hip (DDH) is 1 of
the most common orthopaedic disorders of newborns. It is
thought that fetal positioning and joint laxity contribute to
its development. Several genetic factors have been ex-
plored as contributing to DDH. Controversy exists about
whether the pathogenetics lie in the femoral head or in the
acetabulum in DDH. Dysregulation of the GDF5 (growth
differentiation factor 5) gene, which plays an essential role
in cartilage and bone development, was found in femoral
head cartilage among DDH patients undergoing hip
surgery.28 In addition, the downregulation of GDF11 in a
rabbit model was studied and showed impaired acetabular
chondrocyte function and delayed endochondral ossifica-
tion of the acetabulum.29 Further investigation is needed
to delineate the mechanisms that contribute to the devel-
opment of DDH.
Legg-Calve-Perthes
Although it is known that Legg-Calve-Perthes dis-
ease (LCPD) results from a vascular disruption to the
proximal femoral epiphysis, there is continued interest in
anatomic factors that contribute to bony collapse. Ex-
aminers conducted a cadaveric study on pediatric femur
epiphysis. They investigated the anterior, posterior, lateral
(superior), medial (inferior) quadrants of the epiphysis for
differences in structural morphology and vascular supply.
e176 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
The anterior and lateral quadrants had similar trabecular
structures, whereas the posterior aspect of both the lateral
and medial quadrants and increased vasculature supply.
Though both the lateral and anterior quadrants have
similar architecture, the propensity of posterior blood flow
in the epiphysis makes the anterior femoral epiphysis, a
relative watershed area with an increased risk of collapse
after vascular insult.30
Treatment of LCPD and detection of revasculari-
zation after treatment continue to spark research.31 Two
procedures used to stimulate the healing of the necrotic
femoral head are transphyseal head-neck tunneling and
multiple epiphyseal drilling (MED). A comparative study
in a large animal model of LCPD found that multiple
epiphyseal drilling produced a higher bone volume and
stimulated greater bone formation than the transphyseal
head-neck tunneling or nonweight-bearing control groups.
Techniques for detecting early revascularization
continue to be investigated. A new MRI technique has
been developed utilizing susceptibility-weighted imaging
of the cartilage canals and surrounding cartilage matrix
image vascularity in the epiphyseal cartilage without
contrast administration.32 These images can be quanti-
tively mapped to provide a detailed visualization of the
3-dimensional vascular architecture of epiphyseal carti-
lage. In piglet models, femoral head ischemia was inten-
tionally induced by transection of the ligamentum teres
and cerclage of the femoral neck. At 4 weeks the post-
surgical group had MRI-detectable vessel volume, carti-
lage volume, and vessel density in femoral epiphysis
compared with controls. This technique may be useful for
assessing future revascularization treatments for LCPD.
Talipes Equinovarus Deformity
Genetics and Molecular Pathogenesis
Population studies continue to demonstrate a strong
genetic component to the development of talipes equi-
novarus (TEV).33,34 Genes involved in limb development
such as PITX1B-Tbx4, homeobox genes have been asso-
ciated with clubfoot.35–38 In addition to these devel-
opmental genes, mutations in other pathways including
matrix proteins, sulfation genes, GLI3 (a transcription
repressor), N-acetylation genes, and TGF-β signaling
seem possibly contributory.39–42
Lim-domain kinase 1 (LIMK1) is a known regulator
in embryologic actin organization and cell migration. A
mutation and upregulation of the LIMK1 resulted in re-
duced growth of the lateral motor column neurons to the
peroneal musculature in a mouse clubfoot model suggesting
a potential neuromuscular etiology of TEV.42 An alter-
native etiologic pathway has been suggested by inhibition of
chondrogenesis through activation of the SDF-1/CXCR4/
ROCK2 signaling pathway.43 As a single universal gene or
pathway responsible for all clinically observed TEV seems
unlikely, some investigators have begun using an innovative
integrated bioinformatics approach to try and understand
the complex interactions between potential pathways and
the disease manifestation.44,45
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
Differences Between Normal and Talipes Equinovarus
The anatomic and molecular differences between the
tissues of normal and TEV continue to be studied. Mul-
tiple groups have characterized anatomic differences in the
vascular anatomy and noted that the dorsalis pedis artery
was deficient or completely absent in 20% to 40% of TEV
and that this may be related to the severity of the
deformity.46–48 Interestingly, HOX gene expression and
oxidative damage were increased in the tissue of TEV
when compared with tissue obtained in control children.49
Pitx1 expression, a homeobox gene known to be mutated
in some familial clubfoot patients and causes a similar
phenotype in mice, was downregulated from the tibialis
anterior tendon of idiopathic clubfoot patients.35 Among
the multiple etiologies contributing to clubfoot, the pro-
liferation of extracellular matrix proteins is suggested to
contribute to pathogenesis. Mass spectroscopy analysis
identified 11 upregulated proteins, 7 of which were on the
medial side and 4 on the lateral side of cadaver specimens
with clubfoot. Compared with controls, derangement of
extracellular matrix proteins suggests altered tissue quality
and medial to lateral-sided asymmetry gives insight into
the fibroproliferative nature of the disease and possible
therapeutic targets.50
CONCLUSIONS
This review presents an update on the latest basic
science articles relevant to pediatric orthopaedic surgery
with a focus on scoliosis, trauma, sports medicine, hip,
and clubfoot. Improved knowledge of the molecular, ge-
netic, and biomechanical underpinnings of disease will
allow providers/practitioners/surgeons to provide the best
care for patients. Although these findings present exciting
avenues for better patient care and an improved under-
standing of the processes it is clear that there is still much
to learn about these conditions.
REFERENCES
1. Alsiddiky A, Alfadhil R, Al-Aqel M, et al. Assessment of serum
vitamin D levels in surgical adolescent idiopathic scoliosis patients.
BMC pediatrics. 2020;20:202.
2. Dai J, Lv ZT, Huang JM, et al. Association between polymorphisms
in vitamin D receptor gene and adolescent idiopathic scoliosis: a
meta-analysis. Eur Spine J. 2018;27:2175–2183.
3. Bao L, Chen M, Lei Y, et al. Association between vitamin D
receptor BsmI polymorphism and bone mineral density in pediatric
patients: a meta-analysis and systematic review of observational
studies. Medicine. 2017;96:e6718.
4. Chen H, Zhang J, Wang Y, et al. Abnormal lacuno-canalicular
network and negative correlation between serum osteocalcin and
Cobb angle indicate abnormal osteocyte function in adolescent
idiopathic scoliosis. FASEB J. 2019;33:13882–13892.
5. Ducy P, Amling M, Takeda S, et al. Leptin inhibits bone formation
through a hypothalamic relay: a central control of bone mass. Cell.
2000;100:197–207.
6. Hamrick MW, Ferrari SL. Leptin and the sympathetic connection of
fat to bone. Osteoporos Int. 2008;19:905–912.
7. Man GC, Tam EM, Wong YS, et al. Abnormal osteoblastic response
to leptin in patients with adolescent idiopathic scoliosis. Sci Rep.
2019;9:17128.
8. Wang Q, Wang C, Hu W, et al. Disordered leptin and ghrelin
bioactivity in adolescent idiopathic scoliosis (AIS): a systematic
review and meta-analysis. J Orthop Surg Res. 2020;15:502.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
What is New in Pediatric Orthopaedic: Basic Science
9. Girardo M, Bettini N, Dema E, et al. The role of melatonin in the
pathogenesis of adolescent idiopathic scoliosis (AIS). Eur Spine J.
2011;20(Suppl 1):S68–S74.
10. Latalski M, Danielewicz-Bromberek A, Fatyga M, et al. Current
insights into the aetiology of adolescent idiopathic scoliosis. Arch
Orthop Trauma Surg. 2017;137:1327–1333.
11. Qiu S, Tao ZB, Tao L, et al. Melatonin induces mitochondrial
apoptosis in osteoblasts by regulating the STIM1/cytosolic calcium
elevation/ERK pathway. Life Sci. 2020;248:117455.
12. Delhanty PJ, van der Eerden BC, van Leeuwen JP. Ghrelin and
bone. Biofactors. 2014;40:41–48.
13. Nikolopoulos D, Theocharis S, Kouraklis G. Ghrelin, another factor
affecting bone metabolism. Med Sci Monit. 2010;16:Ra147–Ra162.
14. Kojima M, Hosoda H, Date Y, et al. Ghrelin is a growth-
hormone-releasing acylated peptide from stomach. Nature. 1999;
402:656–660.
15. Xiao L, Zhang H, Wang Y, et al. Dysregulation of the ghrelin/
RANKL/OPG pathway in bone mass is related to AIS osteopenia.
Bone. 2020;134:115291.
16. Pérez-Machado G, Berenguer-Pascual E, Bovea-Marco M, et al.
From genetics to epigenetics to unravel the etiology of adolescent
idiopathic scoliosis. Bone. 2020;140:115563.
17. Robinson S, Kramer J, Shelton T, et al. Assessment of cartilage
growth after biopsy of osteochondral loose bodies in adolescent
knees for use in autologous chondrocyte implantation. J Pediatr
Orthop. 2020;40:110–113.
18. Hou CL, Li B, Cheng YJ, et al. Upregulation of cGMP-dependent
protein kinase (PRKG1) in the development of adolescent idiopathic
scoliosis. Orthop Surg. 2020;12:1261–1269.
19. Jennings W, Hou M, Perterson D, et al. Paraspinal muscle ladybird
homeobox 1 (LBX1) in adolescent idiopathic scoliosis: a cross-
sectional study. Spine J. 2019;19:1911–1916.
20. Li J, Tang M, Yang G, et al. Muscle injury associated elevated
oxidative stress and abnormal myogenesis in patients with idiopathic
scoliosis. Int J Biol Sci. 2019;15:2584–2595.
21. Qin X, He Z, Yin R, et al. Abnormal paravertebral muscles
development is associated with abnormal expression of PAX3 in
adolescent idiopathic scoliosis. Eur Spine J. 2020;29:737–743.
22. Meng Y, Ma J, Lin T, et al. Functional variants of hepatocyte
growth factor identified in patients with adolescent idiopathic
scoliosis. J Cell Biochem. 2019;120:18236–18245.
23. Wang J, Cui Y, Xing K, et al. Generation and characterization of a
human iPSC line derived from congenital clubfoot amniotic fluid
cells. Stem Cell Res. 2020;43:101712.
24. Zhuang Q, Ye B, Hui S, et al. Long noncoding RNA lncAIS
downregulation in mesenchymal stem cells is implicated in the
pathogenesis of adolescent idiopathic scoliosis. Cell Death Differ.
2019;26:1700–1715.
25. Zhao L, Roffey DM, Chen S. Association between the estrogen
receptor beta (ESR2) Rs1256120 single nucleotide polymorphism
and adolescent idiopathic scoliosis: a systematic review and meta-
analysis. Spine (Phila Pa 1976). 2017;42:871–878.
26. Gruber HE, Ashraf N, Cox MD, et al. Experimental induction of
physeal injuries by fracture, drill, and ablation techniques: analyses of
immunohistochemical findings. J Pediatr Orthop. 2019;39:479–486.
27. Ellermann JM, Ludwig KD, Nissi MJ, et al. Three-dimensional
quantitative magnetic resonance imaging of epiphyseal cartilage
vascularity using vessel image features: new insights into juvenile
osteochondritis dissecans. JB JS Open Access. 2019;4:1–9.
28. Baghdadi T, Nejadhosseinian M, Shirkoohi R, et al. DNA hyper-
methylation of GDF5 in developmental dysplasia of the hip (DDH).
Mol Genet Genomic Med. 2019;7:e887.
29. Liu X, Deng X, Ding R, et al. Chondrocyte suppression is mediated
by miR-129-5p via GDF11/SMAD3 signaling in developmental
dysplasia of the hip. J Orthop Res. 2020;38:2559–2569.
30. Morris WZ, Liu RW, Chen E, et al. Analysis of trabecular
microstructure and vascular distribution of capital femoral epiphysis
relevant to Legg-Calve-Perthes disease. J Orthop Res. 2019;37:
1784–1789.
31. Aruwajoye OO, Monte F, Kim A, et al. A comparison of transphyseal
neck-head tunneling and multiple epiphyseal drilling on femoral head
www.pedorthopaedics.com | e177
Lynch et al
healing following ischemic osteonecrosis: an experimental investigation
in immature pigs. J Pediatr Orthop. 2020;40:168–175.
32. Johnson CP, Wang L, Tóth F, et al. Quantitative susceptibility
mapping detects neovascularization of the epiphyseal cartilage after
ischemic injury in a piglet model of legg-calvé-perthes disease.
J Magn Reson Imaging. 2019;50:106–113.
33. Beck JJ, Sangiorgio SN, Jew MH, et al. Alteration in hypoplasia of
the hindfoot structures during early growth in clubfeet treated using
the Ponseti method. J Child Orthop. 2017;11:434–439.
34. Burghardt RD, Tettenborn LP, Stücker R. Growth disturbance of
the distal tibia in patients with idiopathic clubfeet: ankle valgus and
anteflexion of the distal tibia. J Pediatr Orthop. 2016;36:343–348.
35. Zhao X, Yang X. Retinoic Acid promotes retinoic acid signaling by
suppression of Pitx1 in tendon cells: a possible mechanism of a
clubfoot-like phenotype induced by retinoic acid. Med Sci Monit.
2019;25:6980–6989.
36. Weymouth KS, Blanton SH, Powell T, et al. Functional assessment
of clubfoot associated HOXA9, TPM1, and TPM2 variants suggests
a potential gene regulation mechanism. Clin Orthop Relat Res.
2016;474:1726–1735.
37. Alvarado DM, McCall K, Hecht JT, et al. Deletions of 5’ HOXC
genes are associated with lower extremity malformations, including
clubfoot and vertical talus. J Med Genet. 2016;53:250–255.
38. Sadler B, Gurnett CA, Dobbs MB. The genetics of isolated and
syndromic clubfoot. J Child Orthop. 2019;13:238–244.
39. Zhao XL, Wang YJ, Wu YL, et al. Role of COL9A1 genetic
polymorphisms in development of congenital talipes equinovarus in a
Chinese population. Genet Mol Res. 2016;15:1–7.
40. Yong BC, Xun FX, Zhao LJ, et al. A systematic review of
association studies of common variants associated with idiopathic
congenital talipes equinovarus (ICTEV) in humans in the past
30 years. Springerplus. 2016;5:896.
e178 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
41. Yang H, Zheng Z, Cai H, et al. Three novel missense mutations in
the filamin B gene are associated with isolated congenital talipes
equinovarus. Hum Genet. 2016;135:1181–1189.
42. Collinson JM, Lindström NO, Neves C, et al. The developmental
and genetic basis of ‘clubfoot’ in the peroneal muscular atrophy
mutant mouse. Development. 2018;145:1–15.
43. Hu QX, Li XD, Xie P, et al. All-trans-retinoic acid activates SDF-1/
CXCR4/ROCK2 signaling pathway to inhibit chondrogenesis. Am J
Transl Res. 2017;9:2296–2305.
44. Cai G, Yang X, Chen T, et al. Integrated bioinformatics analysis of
potential pathway biomarkers using abnormal proteins in clubfoot.
PeerJ. 2020;8:e8422.
45. Ding J, Liang Z, Feng W, et al. Integrated bioinformatics analysis
reveals potential pathway biomarkers and their interactions for
clubfoot. Med Sci Monit. 2020;26:e925249.
46. Munajat I, Yoysefi M, Nik Mahdi NM. Deficient dorsalis pedis flow
in severe idiopathic clubfeet: Does Ponseti casting affect the
outcome? Foot (Edinb). 2017;32:30–34.
47. Shaheen S, Bahar MEH, Mohammed AHA, et al. Arterial
tree anomalies in patients with clubfoot: an investigation
carried out at Soba University Hospital. J Pediatr Orthop B.
2018;27:67–72.
48. Puri AMC, Hughes KP, Stenson KM, et al. Variations
in arterial pedal circulation in idiopathic congenital talipes
equinovarus: a systematic review. J Pediatr Orthop B. 2021;30:
59–65.
49. Wang Y. Relationship between HOX gene and pediatric congenital
clubfoot. Exp Ther Med. 2018;15:4861–4865.
50. Eckhardt A, Novotny T, Doubkova M, et al. Novel contribution to
clubfoot pathogenesis: the possible role of extracellular matrix
proteins. J Orthop Res. 2019;37:769–778.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
 ORIGINAL ARTICLE

The Ulnar Olecranon-Coronoid Notch Angle Affects

Terminal Elbow Extension in Children and Adolescents
Edward Abraham, MD,*† Ye Lin, MD,* Julio Castillo Tafur, MD,* Nickolas Boroda, MD,*
and Garrett Schwarzman, MD*

Key Words: olecranon-coronoid notch angle, terminal elbow

Background: Limitations to terminal elbow extension (TEE) in extension, elbow range of motion, olecranon-coronoid notch
pediatric populations have been commonly associated with the
degree of ligamentous laxity and not bony factors. Ligamentous (J Pediatr Orthop 2023;43:e179–e187)
laxity, quantified through the Beighton score, is criticized for
unreliably assessing joint mobility. This study aims to show that
the olecranon-coronoid notch angle (OCNA) affects TEE in
healthy children and adolescents.
Methods: A retrospective study of 711 pediatric patients treated
T he main restrictions to normal terminal elbow extension
(TEE) are the joint’s anterior soft tissues and the bony
structure of the olecranon-coronoid notch of the humer-
for upper extremity and shoulder injuries was cross-sectionally
oulnar joint.1,2 Most studies have focused on the soft tissue
studied at 2 tertiary centers from 2014 to 2021. Radiographs
mechanism, referencing generalized hypermobility caused
were used to measure the OCNA, humerocondylar angle, prox-
by ligamentous laxity seen in children3,4 and changes in
imal anterior ulnar angle, and the presence of secondary centers
body mass index and pubertal stage.5 Attempts to quantify
of ossification. A 2-axis goniometer measured clinical TEE to a
the laxity were first reported by Carter and Wilkinson in
firm endpoint. The statistical analysis studied the relationships
19646 and refined by Beighton et al7 in 1973. However,
between OCNA and TEE and the effect that age and sex have on
authors reported a lack of correlation between these in-
these measurements.
struments and clinical conditions such as developmental
Results: Increased TEE was associated with increased OCNA
dysplasia of the hip,8 supracondylar fractures,9 and shoulder
(P < 0.001) when accounting for age and sex. The average OC-
instability.10,11 In addition, these scores inaccurately assess
NA was 30.0 degrees (7.5 degrees), and the average TEE was 5.6
hypermobility in population studies12,13 with differences in
degrees (8.0 degrees). There was a difference in OCNA between
cutoff scores for classifying hypermobility.7,14 Finally, the
subjects who had elbow hypoextension, normal TEE, and elbow
literature review revealed no prior publications associating
hyperextension (P < 0.001). The most common injuries were
elbow hypomobility with ligament integrity.
distal radius fractures (182, 26%), elbow sprains and contusions
The inconsistencies in defining and quantifying soft
(111, 16%), distal both bone forearm fractures (95, 14%), single
tissue restraints on TEE led us to study the elbow bony block
or both bone shaft fractures (77, 11%), and supracondylar
mechanism as a more reliable and quantifiable restraining
fractures (74, 11%).
factor. This mechanism occurs when the ulna’s olecranon
Conclusion: These results show that the orientation of the
apophysis impinges against the posterior olecranon fossa,
opening of the olecranon-coronoid notch influences the arc of
blocking further joint extension.15,16 For example, a baboon
TEE motion in a healthy pediatric population. The notch re-
model found that the earlier bony block caused a mild elbow
strains TEE by activating the bony block mechanism between the
flexion contracture in old-world primates.17
olecranon apophysis and the olecranon fossa. The measurement
We hypothesize that the degree of anterior proximal
of the OCNA can serve as a reproducible and quantitative
tilt of the olecranon-coronoid notch caused by the size of
method to predict hypomobility to hypermobility of TEE
its angle restricts the arc of TEE motion. This study aims
motion.
to show that the olecranon-coronoid notch angle (OCNA)
Level of Evidence: Prognostic study: Level II
is a major and reliable predictor of TEE in healthy chil-
dren and adolescents.
From the *Department of Orthopaedics, University of Illinois at Chi-
cago; and †Division of Orthopaedic Surgery John H. Stroger Jr
Hospital of Cook County, Chicago, IL. METHODS
Partially funded by the National Institute of Health (UL1TR002003) for
statistical support provided by Hajwa Kim, MS, Associate Director This retrospective cross-sectional study received in-
of the University of Illinois at Chicago biostatistics core. stitutional review board approval to enroll patients be-
The authors declare no conflicts of interest. tween 1 month and 17 years old who were treated for
Reprints: Edward Abraham, MD, Department of Orthopaedics, Uni- upper extremity and shoulder injuries between February
versity of Illinois at Chicago, 840 South Wolcott Street, Chicago, IL
60012. E-mail: Eda1215@gmail.com. 2014 and September 2021 at two metropolitan outpatient
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. centers. This population was chosen as these subjects were
DOI: 10.1097/BPO.0000000000002304 likely to have an available lateral radiograph of the elbow

J Pediatr Orthop  Volume 43, Number 2, February 2023 www.pedorthopaedics.com | e179

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Abraham et al
FIGURE 1. A and B, Methodology for measuring the olecra-
non-coronoid notch angle of the proximal ulna. Note, in a
skeletally immature subject, the proximal tip of the notch can
be part of the (A) ulnar metaphysis or (B) olecranon apophysis
centers of ossification.
or forearm and their bilateral elbow range of motion ex-
amined by the principal investigator as part of routine
care. Subjects with generalized joint laxity were not
excluded.13 The Beighton score index was used to diag-
nose generalized idiopathic joint laxity in 14 patients.7 No
patient had added radiographs for this study. Local ex-
clusion criteria included patients with congenital radial
head dislocation, radius and ulna synostosis, forearm de-
ficiencies, earlier trauma, and surgery of the elbow area.
Systemic exclusion criteria included children with known
collagen disorders such as Ehlers Danlos’ syndrome,
Marfan’s syndrome, and osteogenesis imperfecta. Demo-
graphic data were retrieved from the electronic medical
record. Age was rounded down to the nearest integer.
Terminal Elbow Extension Measurement
The neutral-zero measuring method was used to
measure TEE.18 Active TEE was measured to a firm
endpoint. The lack of full elbow extension to 0 degrees was
recorded as a negative number. For infants less than 1 year
of age and young children, elbow joints were passively
extended to a firm endpoint. Infants’ TEE was omitted in
some analyses due to the normal transient perinatal elbow
flexion contractures. The contralateral, uncompromised
elbow was used to measure TEE since patients were en-
rolled during the acute and subacute phases of injury. The
arm was externally rotated and fully supinated. Measure-
ments were taken using a 2-axis goniometer (Goniometer
G360A, AirmatCo Inc, Bryan, OH). The center of the
hinge was positioned over the lateral epicondyle following
the standard measurement technique.18 All measurements
were performed by the principal investigator. We per-
formed a post-hoc analysis of intra-rater reliability for
a representative sample of our measurements. New
e180 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
measurements were performed by the same senior pediatric
orthopaedic surgeon who completed the original meas-
urements. K values show excellent reliability for OCNA
and HCA (0.966 and 0.964, respectively) and good
reliability for UU (0.633).
Radiographic Measurements Technique
Lateral radiographs of the elbow and forearm of the
affected extremity were taken.19 Three radiographic angles
were measured using the Picture Archiving and Commu-
nication System or fluoroscopic images: the OCNA, the
humerocondylar angle of the articular surface of the distal
humerus, and the proximal anterior ulnar angle. Radio-
graphs were obtained within 3 weeks of sprains and within
8 weeks of fractures. The follow-up was not long enough
for any noticeable radiographic changes to occur. The
OCNA was found by drawing a line along the long axis of
the proximal ulnar shaft, which intersected a line drawn
along the anterior tips of the olecranon-coronoid notch
(Fig. 1A). In skeletally immature subjects, the proximal tip
of the olecranon-coronoid notch can be part of the ulnar
metaphysis or olecranon apophysis centers of ossification
(Fig. 1B). The humerocondylar angle was measured by
drawing a line along the distal humeral shaft, which
bisects a line drawn through the capitellum, perpendicular
to its growth plate (Fig. 2, Angle A). In subjects with fused
growth plates, the second line is drawn, bisecting the
capitellum.20 This measurement was not taken in infants
or in subjects with supracondylar or lateral condylar
fractures. The proximal ulnar angle was calculated by
drawing a line along the ulna’s posterior proximal
diaphyseal cortex, which bisects a line drawn along the
same cortex of the proximal metaphysis (Fig. 2, Angle B).
All clinical and radiographic measurements were
taken by the principal investigator and lead author, who is
a senior pediatric orthopaedic surgeon. We could not use
the updated Sauvegrain system since it only determined
FIGURE 2. The measurement method for the humerocondylar
angle (A) and the proximal ulna angle (B) on the lateral
radiographic image of the elbow.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023 Ulnar Olecranon-Coronoid Notch Angle

TABLE 1. Radiographic and Clinical Elbow Characteristics

By Sex
Characteristic Overall, N = 711, n (%) Female, N = 267, n (%) Male, N = 444, n (%) P*
Age (y) 7.9 (4.4) 6.8 (4.2) 8.6 (4.4) < 0.001
Olecranon-coronoid notch angle (deg) 30.0 (7.5) 29.4 (7.1) 30.4 (7.7) 0.072
Terminal elbow extension (deg) 5.6 (8.0) 6.1 (7.9) 5.4 (8.0) 0.2
Humerocondylar angle (deg) 38.9 (8.5) 38.1 (8.4) 39.4 (8.5) 0.016
Proximal anterior ulnar angle(deg) 1.6 (3.1) 1.7 (3.2) 1.5 (3.1) 0.5
Number of ossification centers 0.2
0 379 (53) 155 (58) 224 (50)
1 148 (21) 56 (21) 92 (21)
2 52 (7.3) 16 (6.0) 36 (8.1)
3 3 (0.4) 0 (0) 3 (0.7)
Fused 129 (18) 40 (15) 89 (20)
*Wilcoxon rank sum test and Welch’s t test comparison by sex.
Continuous variables are reported as mean (SD). Categorical variables are reported as frequency (%).

FIGURE 3. A and B, Average TEE and OCNA for subjects by age with standard error bars. (A) Tee peaks at 6 years old, (B) OCNA
peaks at 8 years old OCNA indicates olecranon-coronoid notch angle; TEE, terminal elbow extension.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e181

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Abraham et al J Pediatr Orthop  Volume 43, Number 2, February 2023

FIGURE 4. A–F,. A collage of lateral elbow radiographs of different male patients showed anatomic changes of the olecranon-
coronoid notches and their angles with growth. A, 10 months, 20 degrees. (B) 3 years, 24 degrees. (C) 7 years, 44 degrees. (D)
9 years, 26 degrees. (E). 12 years, 10 degrees. (F) 14 years, 16 degrees. Note the deepening of the olecranon notch caused mainly
by the growth of the distal coronoid processes of the notch and the decrease of the notch angle in the prepuberty patients due to
the maturation of the proximal olecranon-coronoid notch and the olecranon apophyses.

the skeletal age in patients during the pubertal period (2 y)
rather than including children of all ages.21
Olecranon Centers of Ossification
The numbers of secondary centers of ossification of
the olecranon apophysis were counted. Fusion of the
proximal ulnar growth plate was recorded. Images from
multiple subjects were used to describe patterns of growth
and development of the olecranon apophysis. These data
were used to study the effect of ossification centers on
OCNA and TEE.
Statistical Analysis
Continuous variables were reported as mean (SD),
and categorical variables were reported as frequency
(percentage). Subjects were stratified into subgroups by
TEE: hypoextension (TEE <0 degrees), normal extension
as historically described (TEE between 0 and 10 degrees),
and hyperextension (TEE > 10 degrees). Analysis of var-
iance (ANOVA) was used to examine the differences in
OCNA between subgroups. Post-hoc comparison with
Tukey HSD was performed. Pearson’s correlation, linear
regression, and Welch t test analyses were done to study
the relationship between the radiographic measurements,

e182 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Ulnar Olecranon-Coronoid Notch Angle

age, and sex. Significance was set at P = 0.05. All analyses
used R version 3.6.2 (R Foundation for Statistical Com-
puting, Vienna, Austria, 2019).
RESULTS
Demographics
Seven hundred eleven subjects were included in the
study. The average age was 8.4 (4.4) years. There were 444
(62%) males and 267 females. Two hundred eighty-nine
(41%) subjects were identified as Hispanic, and 270 (38%)
were identified as African-American. The most common
injuries were distal radius fractures (182, 26%), elbow
sprains and contusions (111, 16%), distal both bone fore-
arm fractures (95, 14%), single or both bone shaft frac-
tures (77, 11%), supracondylar fractures (74, 11%), lateral
condyle fractures (36, 5%), humerus fractures (25, 3.5%),
and radial head and neck fractures (17, 2.4%). The aver-
age OCNA in patients with supracondylar fractures was
greater than those with distal radius and distal both bone
forearm fractures, 33.0 and 30.5 degrees, respectively
(P = 0.005).
Clinical Terminal Elbow Extension
The total number of patients by age and sex from
1 year to 17 years were divided into 3 groups with Group 1
(1 to 5 y, n = 223): 117 males and 106 females, Group 2
(6 to 12 y, n = 259): 161 males and 98 females, and Group
3 (13 to 17 y, n = 212): 157 males and 55 females. The
overall average TEE for both sexes was 5.6 degrees (7.9)
(Table 1). The average TEE peaked at 6 years old
(Fig. 3A). Univariate linear regression analysis showed a
significant relationship between TEE and age (P < 0,001,
R2 = 0.03). There was no significant difference in the
overall TEE by sex (P = 0.7) or when the subjects were
divided into the above age groups (P = 0.96, P = 0.44, and
P = 0.66, respectively). There were 81 (12%) subjects with
elbow joint hypoextension (< 0 degrees), 470 (67%) with
historically normal TEE (0 to 10 degrees), and 148 (21%)
with elbow joint hyperextension ( > 10 degrees). Overall,
95% of the subjects had a TEE between -7 to 20 degrees
for Group 1, -10 to 23 degrees for Group 2, and -13 to 18
degrees for Group 3, with an overall average range of TEE
motion for the 3 groups being -10 to 20 degrees. The
average TEE in subjects with supracondylar fractures was
numerically greater than TEE in subjects with distal radius
and distal both bone fractures, 7.1 and 5.8 degrees,
respectively (P = 0.075).
Radiographic Olecranon-Coronoid Notch Angle
The mean OCNA was 30.0 degrees (7.5 degrees)
(Table 1). The average OCNA peaked at 8 years old
(Fig. 3B). The average OCNA for infants was 28.0 degrees
(5.7 degrees). Univariate multiple regression found a
relationship between the OCNA and both age
(P < 0.001) and sex (P < 0.001) (R2 = 0.16). Figure 4
shows radiographic changes of the olecranon-coronoid
notches and their angles with growth. The younger
patients had an average increase in the OCNA OCNA
stands for Olecranon Coronoid Notch Angle from 1 to
8 years of age, while the prepubertal patients had a
progressive decrease of the OCNA as the olecranon
apophyses matured to skeletal maturity.
The potential for the differences of the OCNA
measurement between the elbow and forearm radiographs
on the same extremity was studied on a cohort of 100
consecutive patients from the same institution. There were
20 patients with adequate elbow and forearm radiographs.

FIGURE 5. A–C. Clinical examples show a positive correlation between OCNA and TEE. A, Hypoextension: In a 9-year-old boy,
OCNA was 19 degrees, and TEE was -14 degrees. B, Normal* Extension: In a 13-year-old boy, OCNA was 23 degrees and TEE was 0
degrees. C, Hyperextension: In an 8-year-old boy, OCNA was 38 degrees, and TEE was 26 degrees, with a Beighton score of 2/9.
*’Normal’ as historically defined as TEE between 0 degrees and 10 degrees.22 OCNA indicates olecranon-coronoid notch angle;
TEE, terminal elbow extension.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e183

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Abraham et al J Pediatr Orthop  Volume 43, Number 2, February 2023

Paired t test showed no significant difference between the 2
sets of radiographs (t (19) = 1.28, P = 0.215). The intra-
class correlation analysis showed excellent reliability of
OCNA values (ICC = 0.98, P < 0.001).
The Relationship Between Terminal Elbow
Extension and Olecranon-Coronoid Notch Angle
When accounting for age and sex, multiple re-
gression analysis showed that the OCNA was the only
significant predictor of TEE (P < 0.001, R2 = 0.13). The
OCNA differed between subjects who had elbow hypo-
extension, historically normal TEE, and elbow hyper-
extension (P < 0.001). Three clinical cases illustrated this
positive correlation (Fig. 5).
Radiographic Anterior Humerocondylar Angle
and Proximal Anterior Ulnar Angle
The humerocondylar angle was measured in 552
(78%) subjects with an average of 38.9 degrees (8.5 de-
grees). No relationship was found between the humer-
ocondylar angle and OCNA or TEE (P > 0.05). The
proximal anterior ulnar angle was measured in 678 (95%)
subjects. The mean angle was 1.6 degrees (3.1 degrees),
and the median angle was 0 degrees (range: -2 to 15 de-
grees). There was no relationship between the proximal
anterior ulnar angle and TEE (P = 0.8).
Olecranon Apophysis Growth and Development
Four stages of olecranon apophysis and olecranon-
coronoid notch growth are illustrated (Fig. 6). Secondary
centers of ossification were first seen in females at 5 years
old and males at 6 years old. Fusion of the proximal ulnar
growth plates was completed in females by 13 years of age
and males by 15 years of age. (Fig. 7). In subjects aged 10
to 13 years, 1 secondary center of ossification was seen in
87/186 (47%) radiographs, 2 centers in 40/186 (22%)
radiographs, and 3 centers in 1/186 (1%) radiographs.
Subjects with multiple ossification centers had a lower
OCNA than those with 1 ossification center, 25.8 and 30.4
degrees, respectively (P < 0.001).
DISCUSSION
To our knowledge, this is the first quantitative study
to positively correlate radiographic OCNA with clinical
TEE in healthy children and adolescents. We found that
the changes of TEE in the growing children and adoles-
cents were associated with significant simultaneous
changes of the OCNA (Fig. 3). However, there was no
significant difference in TEE between males and females
for all ages. It is generally said in the literature that girls
tend to have greater TEE than boys. It is not clear whether
our patient demographics of 79% Hispanic and African-
American subjects and/or the careful examination of each
patient’s elbow range of motion with a goniometer
contributed to this finding.

FIGURE 6. A–D. Illustration shows typical characteristics for 4 stages of olecranon apophysis and olecranon-coronoid notch growth
and development. A, Stage 1: 0 to 7 years old. Progressive deepening of Shallow olecranon-coronoid notch; cartilaginous apophysis
and rounded end of metaphysis. B, Stage 2: 8 to 9 years old. Appearance of a single posterior center of ossification in the apophysis;
oblique physeal growth; square end to metaphysis. C, Stage 3: 10 to 13 years old. C1: rapid oblique anterior metaphyseal growth
(wandering physeal line)23; C2 and C3: appearance of smaller, anterior ossification centers. D, Stage 4: 14+ years old. Fusion of
ossification centers and closure of the growth plate. D1: only 1 ossification center in the olecranon apophysis which may not
contribute to olecranon-coronoid notch development; D2 and D3: further deepening of the OCN and reduction of the OCNA with
anterior bowing of the proximal ulna. Note: these developmental milestones appear earlier in female subjects. OCNA indicates
olecranon-coronoid notch angle.

e184 | www.pedorthopaedics.com Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Ulnar Olecranon-Coronoid Notch Angle

FIGURE 7. Aand B, Bar graphs show the percentage distribution of centers of ossification of the olecranon apophysis by age and
sex. A, Males. B, Females. All males younger than 6 and all female younger than 5 years of age had 0 centers of ossification. The
subjects older than 15 years of age showed fusion of the proximal ulnar growth plate.

During the early growth period before the age of
9 years, there was a steady increase in TEE and the OC-
NA, which we attributed mainly to a deepening of the
notch caused in part by a rapid increase in the height of
the distal coronoid process. The later growth at the pu-
bertal stage was characterized by a decrease in both TEE
and OCNA in association with growth changes of the
proximal end of the notch and olecranon apophysis. The
cartilaginous portions of these structures underwent en-
dochondral ossification, which ended in the fusion of os-
sification centers and closure of the proximal ulnar growth
plate by age 13 years in girls and 15 years in boys. The
cartilage growth changes may be another independent
factor contributing to the TEE motion.
The clinical use of the OCNA in healthy children
and adolescents can be summarized as the following: First,
the OCNA is another important clinical factor that can be
used in addition to ligamentous laxity to explain all ranges
of normal TEE motion. Second, the notch angle meas-
urement can be useful in explaining the possible cause of
elbow joint hypomobility. Third, TEE motion and OCNA
are inversely affected by the number of centers of ossifi-
cation of the olecranon apophysis.
We compared our results with 2 cross-sectional
studies examining healthy pediatric populations. In the
first study, the mean TEE and age were 4.0 degrees and
9.7 years, respectively.24 In the second study, they were 11
degrees and 8.3 years, respectively.24 Our results were
comparable with that of the first study. This study found
that 81 (12%) subjects had an elbow flexion contracture
less than 0 degrees, whereas none of the 1361 subjects seen
by Barad and colleagues had a flexion contracture. In a
textbook on pediatric musculoskeletal physical diagnosis,
normal TEE was considered between 0 and 10 degrees.22

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e185

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

Abraham et al
Two hundred thirty-five (33%) of our subjects had a TEE
outside of this range. When the outlier patients were ex-
cluded and the remaining 95% of subjects were analyzed,
the average TEE motion ranged from -10 to 20 degrees A
differentiating feature of our study could be that a more
focused examination of joint motion of each subject was
performed by the principal investigator with a goniometer
over the 7-year study period.
This study highlighted the importance of the bony
block mechanism as a primary restraint to TEE (Fig. 8).
The mechanism is analogous to a lever or pivot point to
explain why younger children with increased TEE or
ligamentous laxity are predisposed to sustaining extension
supracondylar fractures.26,27 In this study, we found that
the larger the OCNA, the greater the arc of TEE before
the bony block occurred at the elbow.
Other skeletal structures about the elbow also influ-
ence TEE. The radiographic changes seen during the growth
of the immature olecranon apophysis affected the OCNA.
Subjects with 2 or more ossification centers had a smaller
OCNA than subjects with 1 ossification center. In addition,
different patterns of ossification centers exist during normal
growth and development. In some instances, the OCNA
were measured between 2 metaphyseal landmarks (Fig. 1A).
In others, the centers of ossification of the apophyses
affected the measurement of OCNA (Fig. 1B). A recent
study used the olecranon apophyseal ossification staging
system to predict the timing of peak height velocity and
percentage of final height in adolescents.28 Similar data was
used to predict peak height velocity to manage patients with
scoliosis.29 In our study, the number and location of these
FIGURE 8. A and B, Illustrates the activation of the bony block
mechanism as the tip of the olecranon apophysis impinges in
the olecranon fossa floor of the humerus (arrow). A, Earlier
bony block to TEE motion was most likely due to a shorter arc
of motion caused by a smaller OCNA. B, Later bony block to
TEE motion was most likely caused by a larger OCNA. OCNA
indicates olecranon-coronoid notch angle.
e186 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
olecranon ossification centers played a significant role in
deciding the OCNA.
Radiographic humerocondylar angle measurements
did not significantly affect TEE in this study. A smaller
humerocondylar angle is expected to be associated with
increased elbow extension. A prior study reported the
average humerocondylar angle for children and adoles-
cents to be 41.6 degrees,30 comparable with the values
found in this study, 38.9 degrees.
Two other skeletal findings that may have influenced
TEE were the anterior bow of the proximal ulna and the
presence of a foramen ovale in the olecranon fossa. The
anterior bowing of the proximal ulna measured by a dif-
ferent technique was found to indirectly reduce adult TEE.31
The median anterior ulnar bow measurement was 0 degrees
in our subjects and did not affect TEE. Finally, an adult
study of 166 dry, unpaired humeri bones reported a 10.8%
presence of a foramen ovale, with diameters ranging from
2.7 to 6.5 mm.32 Although a larger foramen ovale may af-
fect TEE, the critical diameter needed has not been pre-
viously reported to the best of the authors’ knowledge.
There are limitations to this study. This was a cross-
sectional study in which data collection was limited to a
single or few clinical visits during routine clinical care. A
longitudinal study with repeated subject examination until
skeletal maturity would give more definitive information
about the natural developmental history of the olecranon-
coronoid notch and olecranon apophysis. Repeat radiation
exposure without clinical benefit to the subjects is difficult to
justify. Second, using the contralateral upper extremity to
measure TEE was necessary to avoid the traumatized limb.
Prior studies did not find significant differences in bilateral
elbow TEE.24,33,34 Third, TEE measurement by a single
investigator may be seen as a limitation. However, data
collected from multiple examiners was less reliable than that
of a single examiner repeating measurements over an ex-
tended period.35 Finally, the OCNA, like the acetabular
index angle used to evaluate hip dysplasia in skeletally im-
mature patients, is based on bony landmarks, which may
not account for the effect of the radiolucent cartilaginous
portions of the olecranon and coronoid process making up
the olecranon-coronoid notch.
In conclusion, the results of this study showed that
the orientation of the opening of the olecranon-coronoid
notch of the proximal ulna can influence the arc of TEE
motion in a healthy pediatric population. In addition, the
growth and development of the olecranon apophysis had a
significant effect on the OCNA. The olecranon-coronoid
notch acts as a restraint to TEE by activating the bony
block mechanism between the olecranon apophysis and
the olecranon fossa. In clinical practice, the measurement
of the OCNA can serve as a reproducible and quantitative
method to predict all ranges of normal TEE.
REFERENCES
1. Gutierrez LS. A Contribution to the study of the limiting factors of
elbow extension. CTO. 1964;56:146–156.
2. Kapandji IA. The physiology of the joints: annotated diagrams of the
mechanics of the human joints. 1996;78:1.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
3. Henrikson B. Supracondylar fracture of the humerus in children. a
late review of end-results with special reference to the cause of
deformity, disability, and complications. Acta Chir Scand Suppl.
1966;369:1–72.
4. Nork SE, Hennrikus WL, Loncarich DP, et al. Relationship between
ligamentous laxity and the site of upper extremity fractures in
children: extension supracondylar fracture versus distal forearm
fracture. J Pediatr Orthop B. 1999;8:90–92.
5. Quatman CE, Ford KR, Myer GD, et al. The effects of gender and
pubertal status on generalized joint laxity in young athletes. J Sci
Med Sport. 2008;11:257–263.
6. Carter C, Wilkinson J. Persistent joint laxity and congenital
dislocation of the hip. J Bone Joint Surg Br. 1964;46:40–45.
7. Beighton P, Solomon L, Soskolne CL. Articular mobility in an
African population. Ann Rheum Dis. 1973;32:413–418.
8. Wynne-Davies R. Acetabular dysplasia and familial joint laxity: two
etiological factors in congenital dislocation of the hip. A review of 589
patients and their families. J Bone Joint Surg Br. 1970;52:704–716.
9. McLauchlan GJ, Walker CR, Cowan B, et al. Extension of the elbow
and supracondylar fractures in children. J Bone Joint Surg Br.
1999;81:402–405.
10. Whitehead NA, Mohammed KD, Fulcher ML. Does the Beighton
Score correlate with specific measures of shoulder joint laxity?
Orthop J Sports Med. 2018;6:2325967118770633.
11. Cahill SV, Sharkey MS, Carter CW. Revisiting the Beighton
Criteria: Does ligamentous laxity testing correlate with shoulder
range of motion norms in a North American, pediatric population?
J Pediatr Orthop. 2020;40:536–542.
12. Lamari NM, Chueire AG, Cordeiro JA. Analysis of joint mobility
patterns among preschool children. Sao Paulo Med J. 2005;123:119–123.
13. Clinch J, Deere K, Sayers A, et al. Epidemiology of generalized joint
laxity (hypermobility) in fourteen-year-old children from the UK:
a population-based evaluation. Arthritis Rheum. 2011;63:2819–2827.
14. Corten L, Ferguson G, Smits-Engelsman B. Does the item ‘hands on
floor’ add value to the Beighton score in identifying joint hyper-
mobility? Eur J Rheumatol. 2020;7:79–83.
15. Abraham E, Powers T, Witt P, et al. Experimental hyperextension
supracondylar fractures in monkeys. Clin Orthop Relat Res.
1982;171:309–318.
16. Rose MD. Another look at the anthropoid elbow. J Hum Evol.
1988;17:193–224.
17. Abraham E, Castillo Tafur J, Lin Y, et al The Effect of the
Olecranon-Coronoid Notch Angle of the Ulna in Terminal Elbow
Extension in Humans and Baboons. Presented as a poster exhibit at
the Orthopaedic Research Society Annual Meeting; February 4-8,
2022; Tampa, FL. Poster no. 1066.
18. Gerhardt JJ, Rondinelli RD. Goniometric techniques for range-
of-motion assessment. Phys Med Rehabil Clin N Am. 2001;12:507–527.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Ulnar Olecranon-Coronoid Notch Angle
19. Frank ED, Long BW, Rollins JH, et al. Merrill’s Atlas of Radiographic
Positioning and Procedures. St. Louis MO: Elsevier Health Sciences;
2013.
20. Goldfarb CA, Patterson JMM, Sutter M, et al. Elbow radiographic
anatomy: measurement techniques and normative data. J Shoulder
Elbow Surg. 2012;21:1236–1246.
21. Diméglio A, Charles YP, Daures JP, et al. Accuracy of the
Sauvegrain method in determining skeletal age during puberty.
J Bone Joint Surg Am. 2005;87:1689–1696. 10.2106/JBJS.D.02418
22. Noonan K, Kocher M. Pediatric Musculoskeletal Physical Diagnosis:
A Video-Enhanced Guide. Philadelphia, PA: Lippincott Williams &
Wilkins; 2020.
23. Silberstein MJ, Brodeur AE, Graviss ER, et al. Some vagaries of the
olecranon. J Bone Joint Surg Am. 1981;63:722–725.
24. Barad JH, Kim RS, Ebramzadeh E, et al. Range of motion of the
healthy pediatric elbow: cross-sectional study of a large population.
J Pediatr Orthop B. 2013;22:117–122.
25. Golden DW, Jhee JT, Gilpin SP, et al. Elbow range of motion and
clinical carrying angle in a healthy pediatric population. J Pediatr
Orthop B. 2007;16:144–149.
26. Wilkins KE. Fractures and dislocations of the elbow region. In:
Rockwood C, ed. Fractures in Children. Philadelphia: JB Lipppin-
cott; 1984:363–575.
27. Cheng JC, Shen WY1993Limb fracture pattern in different pediatric
age groups: a study of 3,350 children. J Orthop Trauma, 7:15–22.
28. Greene JD, Yu K, Li DT, et al. The relationship of olecranon
apophyseal ossification and sanders hand scores with the timing of
peak height velocity in adolescents. J Bone Joint Surg Am. 2021;103:
1543–1551.
29. Sauvegrain J, Nahum H, Bronstein H. Study of bone maturation of
the elbow. Ann Radiol (Paris). 1962;5:542–550.
30. Simanovsky N, Lamdan R, Hiller N, et al. The measurements and
standardization of humerocondylar angle in children. J Pediatr
Orthop. 2008;28:463–465.
31. Rouleau DM, Canet F, Chapleau J, et al. The influence of proximal
ulnar morphology on elbow range of motion. J Shoulder Elbow Surg.
2012;21:384–388.
32. Erdogmus S, Guler M, Eroglu S, et al. The importance of the
supratrochlear foramen of the humerus in humans: an anatomical
study. Med Sci Monit. 2014;20:2643–2650.
33. Boone DC, Azen SP. Normal range of motion of joints in male
subjects. J Bone Joint Surg Am. 1979;61:756–759.
34. Spencer HT, Wong M, Fong Y-J, et al. Prospective longitudinal
evaluation of elbow motion following pediatric supracondylar
humeral fractures. J Bone Joint Surg Am. 2010;92:904–910.
35. Armstrong AD, MacDermid JC, Chinchalkarmethod S, et al.
Reliability of range-of-motion measurement in the elbow and
forearm. J Shoulder Elbow Surg. 1998;7:573–580.
www.pedorthopaedics.com | e187
 ORIGINAL ARTICLE
Clinical Diagnosis of the Infantile Shoulder Subluxation
in Residual Brachial Plexus Birth Injury and Its Correlation
With Ultrasound Diagnosis
Maulin Shah, MBBS, MS, DNB,* Shalin Shah, MBBS, MS, DNB,†
Chinmay Sangole, MBBS, MS,† Sheenam Bansal, MBBS, MS,† Nischal Naik, MS, MCh,‡
and Tejas Patel, BPT, C/NDT, AT(WST)§
Introduction and aims of study: Timely detection of shoulder
subluxation in infants with brachial plexus birth injury (BPBI) is
essential to prevent the progression of glenohumeral deformity.
Shoulder ultrasonography (USG) is routinely used to detect an
infantile subluxation/dislocation, but its use is limited because of
the paucity of expert radiologists in developing countries. The
aim of this study was to determine the clinical examination
predictors to determine shoulder subluxation in patients with
BPBI correlating with ultrasound confirmation.
Methods: We prospectively studied children who presented to our
hospital between 2017 and 2021 diagnosed as brachial plexus
birth injury. In patients developing internal rotation contracture
of the shoulder, we looked for 3 standard clinical signs: reduced
passive external rotation <60 degrees, deep anterior crease
(DAC) and relatively short arm segment. Shoulder subluxation
was defined as USG measurement of alpha angle > 30 degrees
and ossific nuclei of the humerus lying behind the dorsal scapular
line. Sensitivity and specificity were used to assess their efficacy in
clinical diagnosis of shoulder subluxation in different groups.
The predicted probability of shoulder subluxation from each
prediction rule was compared with actual distributions based on
USG confirmation.
Results: Of the 58 BPBI infants who developed PER < 60 degrees
at the shoulder, 41 had USG confirmed shoulder subluxation.
The 2 independent predictors of shoulder subluxation (PER < 45
degrees and DAC) were identified in the current patient pop-
ulation based on data analysis. The presence of short arm seg-
ment is a very specific marker of shoulder subluxation but not
sensitive. The predicted probability of shoulder subluxation from
the prediction rule combining all the 3 markers were similar to
the actual distributions in the current patient population.
From the *Pediatric Orthopedic Departments, Orthokids Clinic; †Pediatric
Orthopedics Department, Orthokids Clinic; ‡Pediatric Orthopedic
Departments, Divyam Hospital; and §Pediatric Orthopedic Depart-
ments, Sparsh Paediatric Rehabilitation Clinic, Ahmedabad, Gujarat,
India.
The authors declare no conflicts of interest.
Reprints: Maulin Shah, MBBS, MS, DNB, Orthokids Clinic 7th Floor,
Golden Icon, Opp. Medilink Hospital, Near Shivranjani Flyover, Sat-
ellite, Ahmedabad 380015, India. E-mail: maulinmshah@gmail.com.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002311
J Pediatr Orthop  Volume 43, Number 2, February 2023
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Conclusions: PER < 45 degrees and presence of deep anterior
crease are clinical markers indicating shoulder dislocation in
patients with BPBI developing reduced external rotation at the
shoulder. On the basis of the proposed clinical diagnosis
algorithm, the above markers along with the selective use of
USG can help in early detection and treatment of infantile
shoulder dislocation.
Key Words: brachial plexus birth palsy, surveillance, ultra-
sonography, BPBI, glenohumeral dysplasia
(J Pediatr Orthop 2023;43:123–127)
R elative early recruitment of shoulder internal rotators in
the upper brachial plexus birth injury can lead to
muscle imbalance across the shoulder joint. Muscular im-
balance coupled with the failure of elongation of dener-
vated muscles can lead to the development of internal
rotation contracture and thereby induces glenohumeral
instability.1–3 In the early 1900s, even before the advent of
modern imaging, Whitman had described the posterior
shoulder subluxation in children with brachial plexus birth
injury.4 Glenohumeral dysplasia can develop early in the
infancy (8%).5 Glenohumeral instability is seen in infancy in
as high as one third of infants with brachial plexus birth
injury.6 Early diagnosis and treatment becomes essential for
the recruited muscles to execute desired motion. Recent
studies have even proposed a universal USG-based sur-
veillance program for the early detection of subluxations.6
Risk criteria exist for the surveillance of hip dislocation, but
no criteria have been defined for shoulder subluxation.7,8
Waters et al1 described diagnostic criteria for gleno-
humeral dysplasia based on axial magnetic resonance
imaging. But the requirement of sedation in an infant and
the cost on health care does not allow us to use it as a
screening tool. Ultrasonography (USG) has been shown by
Hunter in 1996 to be equally effective for identifying pos-
terior shoulder subluxation.9 Recent reports recommend
ultrasound-based screening for surveillance in at risk cases
and even universal surveillance in limited populations.6,10
However, expert musculoskeletal radiologists are not uni-
versally available, especially in developing countries. Thus,
the utility of USG as a routine surveillance tool is not
www.pedorthopaedics.com | 123
Shah et al
practical. Various clinical signs associated with infantile
shoulder subluxation have been described in the literature.
Loss of passive external rotation (PER), fullness in the
posterior aspect of the shoulder, apparent shortening of the
arm, and asymmetry in the soft tissue folds of the upper
arm are some of them.5,11 Bauer and colleagues in their
study found a correlation between reduced PER be-
yond < 60 degrees and presence of shoulder dislocation.
They found it to be a sensitive marker.10
With this study, we aimed to define a set of clinical
examination predictors consistent with ultrasound con-
firmed posterior shoulder subluxation diagnosis. We also
tried to develop an algorithm where these clinical pre-
dictors can help guide early intervention for emergent
cases and further imaging in only borderline cases for use
in clinical practice.
METHODS
This study was carried out at a tertiary referral
center in Ahmedabad, India, over a period of 4 years from
2017 to 2021. Appropriate institutional review board ap-
proval was obtained. Out of the various clinical parame-
ters of shoulder dislocation, reduced PER, Deep anterior
crease (DAC), and relative short arm segments (SASs)
were considered. PER was measured with the infant in
FIGURE 1. A, Measurement of passive external rotation. B, Image of a child with posterior shoulder dislocation, a passive external
rotation of 30 degrees and present deep anterior crease. C, Relative short arm segment in an infant with posterior shoulder
dislocation.
124 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
supine position to prevent scapular compensation. An
adducted arm was externally rotated with the elbow flexed
and the degrees of external rotation was measured with the
help of a hand-held goniometer (Fig. 1A). While
measuring PER, it was made sure that the scapula is not
riding superiorly. DAC was noticed by inspecting anterior
axillary fold along both the shoulders (Fig. 1B). Arm
segment length was measured from the tip of the acromion
to the lateral point of anterior elbow crease (Fig. 1C). A
note of all these clinical markers was made in children
with PER <60 degrees.
All the infants with PER < 60 degrees on prior
clinical examination underwent USG examination by an
expert musculoskeletal radiologist. USG was performed
without sedation, with the child sitting in the mother’s
lap from the posterior aspect of the shoulder joint. Alpha
angle as described by Hunter was measured.9 The
shoulder position in neutral rotation was considered for
alpha angle measurement. Shoulder subluxation was
defined as USG measurement of alpha angle > 30 de-
grees and ossific nuclei of the humerus lying behind the
dorsal scapular line (Fig. 2).9 Analysis of the data
obtained was performed with the help of Microsoft Excel
and SPSS software. Frequency distribution of dislocation
with 3 groups of degrees of Passive ER ( > 45, 20 to 44,
and <20 degrees) was calculated. Sensitivity and
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
FIGURE 2. Ultrasonographic image with ossific neuclei of humeral head behind dorsal scapular line and alpha angle of 46 degrees.
specificity of each clinical marker was measured to
predict the dislocation. Positive and Negative predictive
values were calculated.
RESULTS
Over a period of 4 years from 2017 to 2021, a total
of 143 infants with BPBI were presented at our clinic. Out
of these, 58 children had PER < 60 degrees at shoulder.
There were 34 males and 24 females. Right side was in-
volved in 45 cases and the left side was involved in 13
cases. The average age of infants in our study was
7.4 months (range: 3 to 12 mo).
PER
Forty one out of 58 infants with PER < 60 degrees
had USG confirmation of shoulder subluxation. The
average PER in infantile subluxations was 22 degrees
with a SD of 12.05 degrees. With a 90% of confidence,
we could observe that the cases with subluxation are
confined within the range of 22 ± 19.7 degrees [1.64SD
equivalent to 90% of the distribution, range of degrees 0
to 42 degrees].
In this subset of infants with subluxation, there were
2 subluxations in the 45 to 60 group, 18 in the 21 to 44
group, and 21 in the < 21 group. The distribution of sub-
luxation and nosubluxation in each of these groups has
been provided in Table 1. On the basis of this distribution
of the subluxations, when we reduce the criteria of PER
from 60 to 45 degrees, we could get a sensitivity of 95.1%
and a specificity of 88.2%.
Two infants with subluxation had PER > 45 de-
grees. However, these two patients demonstrated presence
of other two clinical markers which are DAC and SAS.
Deep Anterior Crease and SAS
Deep anterior crease and SAS are markers of infantile
shoulder subluxation. Deep anterior crease was present in
37 infants in our study. Out of these 37 patients, 36 infants
were confirmed to have a shoulder subluxation on USG.
Twenty-five of the 58 infants in our study demonstrated a
presence of relative SAS of the arm. Twenty-four infants
had subluxation on USG. Only 1 child who had sustained a
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Clinical Diagnosis Algorithm of Infantile Shoulder Dislocation
humerus fracture and subsequent varus malunion did not
have a shoulder subluxation in the subset.
Table 2 describes the positive and negative predictive
value of each clinical marker in 2 different PER groups
and the sensitivity and specificity related to these factors.
First group is according to the previously defined criteria
of PER < 60 degrees and combination of Deep anterior
crease and SAS along with them. The second group has
the newer criteria of PER <45 degrees. As shown, the
negative predictive value can be improved when PER < 45
degrees criteria is applied.
DISCUSSION
There is a high prevalence of glenohumeral sub-
luxation in infants with obstetric brachial plexus birth
injury. A reduced glenohumeral joint is essential for the
recovering muscle to execute motion across the joint. Be-
fore the advent of USG, magnetic resonance imaging or
computed tomography arthrogram were routinely used as
imaging modality for diagnosing infantile subluxation/
dislocation.12 However, these investigations required se-
dation or general anesthesia. Hunter et al9 introduced the
utility of ultrasound to detect posterior shoulder dis-
location. Moukoko and colleagues reported USG as a
standard and reliable tool to diagnose infantile shoulder
dislocations without the need of anesthesia in this young
population.5 Bauer and colleagues have discussed the PER
range as a clinical marker of shoulder subluxation and
identified 60 degrees as the cutoff point for suspecting
presence of shoulder subluxation in infants with brachial
plexus birth injury.
Shoulder subluxation and dislocation terms are
being used interchangeably in the literature. Ezaki and
TABLE 1. The Distribution of Dislocators and Nondislocators in
Each Group
PER < 20 (n = 21) 21-44 (n = 20) > 45 (n = 17)
Dislocator 21 18 2
Nondislocator 0 2 15
% 100 90 11.8
www.pedorthopaedics.com | 125
Shah et al
TABLE 2. Positive Predictive Value, Negative Predictive Value, Sensitivity, and Specificity in Various Combination of Clinical Markers
Positive Predictive Value
Group: <60 Degrees Group: <45 Degrees
PER 70.7 95.1
PER or DAC 70.7 90.9
PER or DAC or SAS 70.7 89.13
PER + DAC 97.2 100
DAC indicates deep anterior crease; PER, passive external rotation; SAS, short arm segment.
colleagues coined the term shoulder dysplasia for this
changing pathologic anatomy of the shoulder joint. They
identified clinical markers of shoulder dysplasia including
progressive reduction in the PER range over monthly
clinic visit, DAC or extra humeral fold, relative SAS and
palpable posterior humeral head, or a click of shoulder
rotation.5 They have suggested presence of any one of
them to be an indication for ultrasonographic study of the
shoulder joint. We identified 3 clinical markers, which can
be reliably utilized for diagnosing infantile shoulder sub-
luxation. We found 2 of them to be individual predictors
of dislocation and combining all the3 of them were able to
develop an algorithm to proceed with further intervention
in these children.
Bauer and colleagues with 60 degrees as the criteria,
got a sensitivity of 93.8% and a specificity of 69.4% for
diagnosing shoulder subluxation. However, in our cohort,
we found that a majority of the infants with PER range
between 45 and 60 had their glenohumeral joints located
and had normal values of alpha angle and percentage of
humeral head anterior to the scapular line percentage. By
reducing the diagnostic range to < 45 degrees, we could
increase the specificity without affecting the sensitivity.
Ezaki and colleagues observed that relative SAS is a very
specific clinical marker for diagnosing shoulder subluxation
and it develops as a result of telescoping of posteriorly sub-
luxated humeral head.5 However, no correlation was estab-
lished through ultrasound diagnosis in their study. In our
cohort, we found that although SAS is 94% specific, it is only
FIGURE 3. Infantile shoulder dislocation diagnosis algorithm. ER indicates external rotation; DAC, deep anterior crease; USG,
ultrasonography.
126 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
Negative Predictive Value
Group: <45 Degrees Sensitivity Specificity
88.2 95.1 88.2
92.9 97.6 76.5
100 100 70.6
70.8 82.9 100
58% sensitive. One should also keep other etiologies like con-
genital short limb or previous ipsilateral or contralateral
humeral fracture while using this marker.
Deep anterior crease along with extra humeral fold
develop because of posterior migration of the subluxated
humeral head. We believe that Deep anterior crease ap-
pears earlier in the clinical course of shoulder subluxation
compared with the extra humeral fold, and hence DAC is
used in isolation as a clinical predictor. DAC has a high
sensitivity and specificity in diagnosing shoulder sub-
luxation. Our observation concurs with previously re-
ported studies.5,10 Presence of DAC and PER < 45 degrees
are independent predictors of infantile shoulder dis-
location. In the absence of expertise of USG inves-
tigations, they can be used as surrogate markers.
Although Universal screening of infantile shoulder
dislocation is not practical, a selective USG screening can
be useful to diagnose the outliers of positive clinical pre-
dictors. On the basis of our observation in this study, we
came up with an Infantile Shoulder Subluxation Diagnosis
Algorithm (Fig. 3). This algorithm narrows down the need
of expert shoulder USG to diagnose infantile shoulder
subluxation. We observed that 95% of patients with
infantile shoulder subluxation can be identified only based
on clinical examination markers.
There are some limitations of the study. We included
only those children who had <60 degrees PER for USG
examination, as that cohort has already been identified to
be at high risk for shoulder dysplasia and our intention
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
J Pediatr Orthop  Volume 43, Number 2, February 2023
was to assess the severity of limitation of ER associated
with it. Reliability studies of clinical markers are required
with a greater number of patients.
CONCLUSIONS
Shoulder subluxation/dislocation in infants with
brachial plexus birth injury is not uncommon. PER <45
degrees, deep anterior axillary crease, and SAS are im-
portant clinical predictors for the diagnosis of shoulder
dysplasia. Along with the described clinical parameters,
selective use of ultrasound can help in early diagnosis and
treatment of infantile shoulder subluxation.
ACKNOWLEDGMENTS
The authors thank Dr Dhiren Ganjwala, Dr Hitesh
Shah, and Dr Atul Bhaskar who have helped in reviewing the
manuscript.
REFERENCES
1. Waters PM, Smith GR, Jaramillo D. Glenohumeral deformity
secondary to brachial plexus birth palsy. JBJS. 1998;80:668–677.
2. Pearl ML, Edgerton BW. Glenoid deformity secondary to brachial
plexus birth palsy. JBJS. 1998;80:659–657.
3. Nikolaou S, Peterson E, Kim A, et al. Impaired growth of
denervated muscle contributes to contracture formation following
neonatal brachial plexus injury. JBJS. 2011;93:461–470.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
Clinical Diagnosis Algorithm of Infantile Shoulder Dislocation
4. Whitman R. The treatment of congenital and acquired luxations at
the shoulder in childhood. Ann Surg. 1905;42:110–115.
5. Moukoko D, Ezaki M, Wilkes D. Posterior shoulder dislocation in
infants with neonatal brachial plexus palsy. J Bone Joint SurgAm.
2004;86:787–793.
6. Pöyhiä TH, Lamminen AE, Peltonen JI, et al. Brachial plexus birth
injury: US screening for glenohumeral joint instability. Radiology.
2010;254:253–260.
7. Newborn and infant physical examination (NIPE) screening progra-
mme handbook. GOV.UK. 2019. Available at: https://www.gov.uk/
government/publications/newborn-and-infant-physical-examination-
programme-handbook/newborn-and-infant-physical-examination-
screening-programme-handbook. Accessed December 10, 2021.
8. Biedermann R, Riccabona J, Giesinger JM, et al. Results of
universal ultrasound screening for developmental dysplasia of the
hip: a prospective follow-up of 28 092 consecutive infants. Bone Joint
J. 2018;100:1399–1404.
9. Hunter JD, Franklin K, Hughes PM. The ultrasound diagnosis of
posterior shoulder dislocation associated with Erb’s palsy. Pediatr
Radiol. 1998;28:510–511.
10. Bauer AS, Lucas JF, Heyrani N, et al. Ultrasound screening for
posterior shoulder dislocation in infants with persistent brachial
plexus birth palsy. JBJS. 2017;99:778–783.
11. Chung KC, Yang LJ, McGillicuddy JE. Practical Management of
Pediatric and Adult Brachial Plexus Palsies E-Book. 1st ed. Elsevier
Health Sciences: Saunders, 2011:103-113.
12. White SJ, Blane CE, DiPietro MA, et al. Arthrography in evaluation
of birth injuries of the shoulder. Can AssocRadiol J. 1987;38:
113–115.
www.pedorthopaedics.com | 127
 ART AND PEDIATRIC ORTHOPAEDICS

Back to Carracci
Gleeson Rebello, MD* and Benjamin Joseph, MS Orth, MCh Orth, FRCS Ed†

the artist’s hand-writing. This sketch is believed to be one

Abstract: Annibale Carracci’s sketch of a boy with a hunchback
conveys the boy’s emotions remarkably well. The possibilities of
the underlying cause of the spinal deformity and its neurologic
complications are discussed.
Key Words: spinal deformity, kyphosis, spinal tuberculosis,
congenital kyphosis, neurofibromatosis
(J Pediatr Orthop 2023;43:e188–e189)
THE ARTIST: ANNIBALE CARRACCI (1560–1609)
of Carracci’s early works and depicts life exactly as he saw
it, a trait for which Carracci was greatly respected. The
sketch is currently part of the Devonshire Collection in
Chatsworth House, Devonshire, UK.
RELEVANCE TO PEDIATRIC ORTHOPAEDICS
It is conjectural as to what the cause of the thoraco-
lumbar kyphosis is. Congenital kyphosis, and acquired
kyphosis secondary to spinal tuberculosis come to mind.
The shaded patches on the torso and arm may possibly be

A nnibale Carracci was Italian painter who came from a

family of artists and is regarded as the most talented of
3 distinguished painters. Annibale and his elder brother
Agostino were children of a tailor who were persuaded to
pursue a career in art by an older cousin Lodovico, who
was also an artist of repute. The younger Carracci is re-
puted to have had precocious artistic talents which
blossomed. He was instrumental in recovering the classic
traditions of the Renaissance artists, who focussed on
accurate and realistic depiction of people, from the ten-
dency of Mannerism to experiment with altered pro-
portions, perspectives, and highly stylized poses.1 Carracci
is also credited with reviving interest in the work of the
early 16th-century painter Correggio, who had been ef-
fectively forgotten.
café-au-lait spots and this raises the possibility of kyphosis
secondary to neurofibromatosis. The long kyphotic curve
suggestive of involvement of multiple vertebral segments
may suggest that the cause of the deformity is an under-
lying neurological disorder.
The prominence of the posterior axillary fold, the
latissimus dorsi and triceps gives the impression that
he is propping himself up, taking weight on the right
upper limb.
One wonders whether the sheet drawn up above the
boy’s waist is to hide the fact that the lower limbs are
paralyzed and that he may be incontinent. The look of
hopelessness and the plea for God’s help may imply that

THE WORK OF ART: A HUNCHBACK BOY

The sketch (Fig. 1) shows a boy with a hunchback
looking over his shoulder at the viewer. The sparse hair
and face give the impression that the subject is a young
man but the title of the work indicates that he is, in fact, a
boy. The boy is not wearing any clothes above his waist
and a sheet drawn up to his waist obscures the lower torso
and the lower limbs. The deformity of the back is striking
as is the melancholic expression on the boy’s face. The
inscription on the sketch that reads “Non so se Dio
m’aiuta” and translates to “I do not know if God will help
me” adds to the poignancy so effectively conveyed
through this work of art, though the inscription is not in

From the *Department of Pediatric Orthopedics, Massachusetts General

Hospital, Boston, MA; and †Former Head of Paediatric Orthopaedic
Service, Kasturba Medical College, Manipal, Karnataka, India.
No financial disclosures.
The authors declare no conflicts of interest.
Reprints: Benjamin Joseph, MS Orth, MCh Orth, FRCS Ed, 18 HIG
HUDCO Colony, Manipal, Karnataka 576104, India. E-mail:
bjosephortho@yahoo.co.in.
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BPO.0000000000002117 FIGURE 1. A Hunchback Boy by Annibale Carracci.

e188 | www.pedorthopaedics.com J Pediatr Orthop  Volume 43, Number 2, February 2023

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop  Volume 43, Number 2, February 2023 Art and Pediatric Orthopaedics

the boy had more serious issues other than just a deformity
and that he was probably aware that no effective treat-
ment was available, at the time, for any of the conditions
responsible for his sad situation.
REFERENCE
1. Christiansen K. “Annibale Carracci (1560–1609).” In Heilbrunn
Timeline of Art History. New York, NY: The Metropolitan Museum
of Art, 2000. Available at: http://www.metmuseum.org/toah/hd/carr/
hd_carr.htm. Accessed October, 2003.

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e189

Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.


Impacts of Fracture
Types on Success Rate of
Closed Reduction and
Percutaneous Pinning in
Pediatric Lateral Condyle
Humerus Fractures
Displaced > 4 mm
To the Editor:
We read with interest the recent
publication by Xie and colleagues.1 Xi
et al2 already published data from the
same patients in 2021 but stated in
their current publication1 that all 50
fractures were a Jakob type III, which
is equivalent to a Song stage 5 (dis-
placed and rotated fragment).1,3 We
would like to ask the authors how they
explain the inconsistency of them
having reported that only 30 of 50
fractures were a Song type 51 and
what the Song type 5 rate was for each
Milch type group, since malrotation
might be a contributing factor to the
failed closed reduction rate? We also
would like to ask Xie et al1 why they
compared the Milch with the Song3
classification, despite the latter not
considering Milch type I fracture
patterns, which are the result of ab-
duction forces and Song et al4 them-
selves having abandoned their own
classification already in 2010, when
the authors used the Jakob classi-
fication without referencing their own
classification, indicating that the Song
classification is deficient and not ad-
equate to classify the spectrum of lat-
eral condyle fractures.
Song et al’s3 5 fracture stages do
not differentiate between fractu-
res running through the capitello-
trochlear sulcus or lateral to it (Milch
type I) and those involving the troch-
lea (Milch type II), with their illus-
tration of all 5 stages depicting Milch
type II fractures of increasing severity.
The described severity progression
from stage I to 53 only applies to
avulsion fractures which are caused by
forearm adduction injuries, resulting
in downward displacement of the
e190 | www.pedorthopaedics.com
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
LETTERS TO THE EDITOR
condylar fragment. In contrary to the
latter, Song et al3 claimed that frag-
ments mainly displaced poster-
olaterally and did not consider
compression injuries from longi-
tudinal and transverse forces, which
could also cause a similar appearance
to Song stage 3 to 5 fractures but
would start at the articular impact
site, with fragments being pushed an-
teriorly if the elbow was extended and
posteriorly if the elbow was flexed.
In their previous publication Xi
et al2 reported a mean elbow flexion
of 133.1 degrees, mean extension of
6.7 degrees and mean total range of
movement (ROM) of 140.2 degrees
at a mean follow-up of 13.9 months
for patients who had either closed
reduction or open reduction for
failed closed reduction (no difference
compared with their primary open
reduction group), without providing
data for the uninjured side. McKay
et al5 reported normative ROM data
from normal children 3 to 9 years of
age, with mean elbow flexion of 146
degrees, mean extension of 3 degrees
and mean ROM of 149 degrees. The
latter indicates that Xi et al’s1,2 pa-
tients had a mean loss of 12.9 degrees
flexion and a mean increased ex-
tension of 3.7 degrees. Therefore,
none of the patients had achieved an
excellent Hardacre result but the
authors claimed 94% excellent re-
sults. The Hardacre criteria do not
consider loss of flexion and the as-
sessment of the carrying angle is ar-
bitrary, leaving it up to the assessor
to decide if an angle change is in-
conspicuous or not, with Xi et al1,2
not having assessed the difference
between the injured and uninjured
arm. Xi et al’s1,2 outcomes were only
fair according to the Flynn criteria
(motion reduced by 11 to 15 degrees).
In conclusion, we think that
there is a need for a lateral condyle
fracture classification which consid-
ers the anatomy, displacement, and
malrotation. Xi et al’s1,2 fair Flynn
outcomes with mean loss of flexion of
12.9 degrees for their closed and
failed closed reduction groups and
13.3 degrees for their primary open
J Pediatr Orthop
reduction group indicates a possible
need for the authors to reassess and
change their management strategy.
Areas we are concerned about be-
cause of a potential negative impact
on functional recovery are: accuracy
of intraoperative arthrogram assess-
ment of post-reduction displacement,
acceptance of an articular step of up
to 2 mm, the use of 3 or 4 wires (all of
Xi et al’s1,2 depicted fixations) in
these small children (mean age 5.6 y)
which increases potential damage to
the soft tissues and growth plate,
and leaves wires in place for a me-
an of > 15 weeks following open
reduction.2
Andreas Rehm, MD, FRCS Glasgow,
FRCS Tr&Ortho*
Azeem Thahir, MD†
Luke Granger, MB BS‡
Elizabeth Ashby, MD*
*Department of Paediatric Orthopaedics
†Department of Trauma & Orthopaedics
‡Department of Paediatric Orthopaedics,
Cambridge University Hospitals NHS
Trust, Cambridge, UK
A.R., A.T., L.G., and E.A.: literature review,
manuscript preparation.
The authors did not receive any funding for
this work.
The authors declare no conflicts of interest.
DOI: 10.1097/BPO.0000000000002233
REFERENCES
1. Xie LW, Tan G, Deng ZQ, et al. Impacts of
fracture types on success rate of closed
reduction and percutaneous pinning in pe-
diatric lateral condyle humerus fractures
displaced > 4 mm. J Pediatr Orthop.
2022;42:265–272.
2. Xie LW, Deng ZQ, Zhao RH, et al. Closed
reduction and percutaneous pinning vs
open reduction and internal fixation in
pediatric lateral condylar humerus fractures
displaced by > 4mm: an observational cross-
sectional study. BMC Musculoskel Disord.
2021;22:985.
3. Song KS, Kang CH, Min BW, et al. Closed
reduction and internal fixation of displaced
unstable lateral condylar fractures of the
humerus in children. J Bone Joint Surg Am.
2008;90:2673–2681.
4. Song KS, Shin YW, Wug C, et al. Closed
reduction and internal fixation of completely
displaced and rotated lateral condyle frac-
tures of the humerus in children. J Orthop
Trauma. 2010;24:434–439.
 Volume 43, Number 2, February 2023
J Pediatr Orthop  Volume 43, Number 2, February 2023
5. McKay MJ, Baldwin JN, Ferreira P, et al.
Normative reference values for strength and
flexibility of 1000 children and adults.
Neurology. 2017;88:36–43.
Assessment of Lateral
Rotation Percentage
and Rotational
Deformity of the Elbow
in Type 3 Supracondylar
Humerus Fractures:
A Biomechanical Study
To the Editor:
We read with interest the recent
publication by Berdis et al.1 The
authors1 stated that their lateral rotation
percentages (LRP) do not apply if co-
ronal or sagittal deformity exists but
there are other factors which will also
affect the measurements. Camp et al2
measured that the perceived Baumann
angle increases with internal and de-
creases with external humeral rotation
of the humerus if an anteroposterior
radiograph is not taken as a true ante-
roposterior radiograph, with the per-
ceived angle changing by ± 1.6 degrees
per 10 degrees change of rotation with
the humerus parallel to the collector/
cassette and by ± 5 degrees per 10 de-
grees change of rotation with the hu-
merus flexed 30 degrees. The LRP is
dependent on the lateral radiograph
being a true lateral view of the distal
fragment, with the perceived measure-
ments changing depending on the
amount of rotation of the distal hume-
rus away from the true lateral position.
We could not identify a single study
reproducing Camp et al’s2 study for
lateral elbow radiographs. We also as-
sume that the LRP depends on the
center of rotation and therefore on the
position of the first K-wire at the level of
the fracture and its direction, since this
defines the axis of rotational displace-
ment. We would like to ask Berdis et al1
why they inserted their K-wire with in-
clination in a posterior to anterior di-
rection, stating that this makes their
study as clinically relevant as possible by
replicating standard intraoperative pin
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
placement, despite this being the oppo-
site to the anterior to posterior in-
clination recommended by Skaggs et al3
in accordance with normal anatomy?
We would also like to ask the authors if
they agree that their statement that “it
should not matter if the arm is in in-
ternal or external rotation because it is a
percentage” is a factual error, since on a
true lateral view the perceived postero-
anterior distance of the distal humeral
fragment will be at its smallest and will
increase with arm rotation in either di-
rection in relation to the xray beam. The
perceived width of the proximal frag-
ment will either increase or decrease,
depending on the initial rotational rela-
tionship between the 2 fragments.
Therefore, the LRP will change in at
least half (if not in all) of the rotational
scenarios if the lateral radiograph is not
taken as a true lateral radiograph.
Ryan et al4 reported that the
AHL runs through the middle third of
the capitellum in 100% of children
5 years of age and above but Berdis
et al’s1 AHL runs through the front of
the capitellum, indicating that the used
bone model might not be an anatomic
reproduction of a normal child’s hu-
merus, which raises doubts about the
reliability of the measured percentages.
Bahk et al5 defined malrotation
arbitrarily as a discrepancy of more than
20% between the width of the distal hu-
merus above and below the fracture on
the lateral radiograph, which is the only
documentation of a definition of malro-
tation we could identify in the literature.
The American Academy of Orthopaedic
Surgeons on the other hand does not
consider rotational malalignment in their
Appropriate Use Criteria at all (http\
\www.aaos.org/quality/quality-programs/
pediatric-supracondylar-humerus-
fractures). In contrary to the latter,
O’Hara et al6 recommended K-wire fix-
ation for fractures with any degree of
malrotation but not for those without
malrotation.
We assess intraoperative rota-
tional malalignment by taking internal
and external oblique elbow views,
which show the alignment of the me-
dial and lateral column, with our im-
age intensifier machines not allowing
to perform any measurements. Could
Berdis et al1 explain how their meas-
urements help with intraoperative
Letters to the Editor
decision making, as they stated in their
discussion, since we generally do not
accept malrotation, not using a specific
LRP value as a decision guide?
In conclusion, there are multiple
factors which can affect the LRP, with
Berdis et al’s measured LRPs only ap-
plying to one specific K-wire position
without any coronal or sagittal dis-
placement. Therefore, it is highly likely
that the provided angles do not apply to
the majority of intraoperative and
postoperative scenarios, because we fre-
quently accept some displacement, have
varying wire positions and frequently do
not get true lateral views in the clinic.
Therefore, we do not think that Berdis
et al’s1 data can be used reliably for the
majority of fractures to estimate the true
degree of malrotation, with it being
hitherto unknown if any malrotation/
LRP can be accepted, without affecting
cosmesis and/or function.
Andreas Rehm, MD*
Azeem Thahir, MD†
Elizabeth Ashby, MD*
Pinelopi Linardatou Novak, MBBCh*
Departments of *Paediatric Orthopaedics
†Trauma & Orthopaedics, Cambridge
University Hospitals NHS Trust
Cambridge, UK
The authors declare no conflicts of interest.
DOI: 10.1097/BPO.0000000000002280
REFERENCES
1. Berdis G, Hooper M, Talwalker V, et al.
Assessment of lateral rotation percentage and
rotational deformity of the elbow in type 3
supracondylar humerus fractures: a biomechan-
ical study. J Pediatr Orthop. 2021;41:e605–e609.
2. Camp J, Ishizue K, Gomez M, et al. Alter-
ation of Baumann’s angle by humeral
position: implications for treatment of supra-
condylar humerus fractues. J Pediatr Orthop.
1993;13:521–525.
3. Skaggs DL, Cluck MW, Mostofi A, et al.
Lateral-entry pin fixation in the management
of supracondylar fractures in children. J Bone
Joint Surg Am. 2004;86:702–707.
4. Ryan DD, Lightdale-Miric NR, Joiner ERA,
et al. Variability of the anterior humeral line
in normal pediatric elbows. J Pediatr Orthop.
2016;36:e14–e16.
5. Bahk MS, Srikumaran U, Ain MC, et al.
Patterns of pediatric supracondylar humerus
fractures. J Pediatr Orthop. 2008;28:493–499.
6. O’Hara LJ, Barlow JW, Clarke NMP.
Displaced supracondylar fractures of the
humerus in children. J Bone Joint Surg Br.
2000;82-B:204–210.
www.pedorthopaedics.com | e191
J Pediatr Orthop  Volume 43, Number 2, February 2023
5. McKay MJ, Baldwin JN, Ferreira P, et al.
Normative reference values for strength and
flexibility of 1000 children and adults.
Neurology. 2017;88:36–43.
Assessment of Lateral
Rotation Percentage
and Rotational
Deformity of the Elbow
in Type 3 Supracondylar
Humerus Fractures:
A Biomechanical Study
To the Editor:
We read with interest the recent
publication by Berdis et al.1 The
authors1 stated that their lateral rotation
percentages (LRP) do not apply if co-
ronal or sagittal deformity exists but
there are other factors which will also
affect the measurements. Camp et al2
measured that the perceived Baumann
angle increases with internal and de-
creases with external humeral rotation
of the humerus if an anteroposterior
radiograph is not taken as a true ante-
roposterior radiograph, with the per-
ceived angle changing by ± 1.6 degrees
per 10 degrees change of rotation with
the humerus parallel to the collector/
cassette and by ± 5 degrees per 10 de-
grees change of rotation with the hu-
merus flexed 30 degrees. The LRP is
dependent on the lateral radiograph
being a true lateral view of the distal
fragment, with the perceived measure-
ments changing depending on the
amount of rotation of the distal hume-
rus away from the true lateral position.
We could not identify a single study
reproducing Camp et al’s2 study for
lateral elbow radiographs. We also as-
sume that the LRP depends on the
center of rotation and therefore on the
position of the first K-wire at the level of
the fracture and its direction, since this
defines the axis of rotational displace-
ment. We would like to ask Berdis et al1
why they inserted their K-wire with in-
clination in a posterior to anterior di-
rection, stating that this makes their
study as clinically relevant as possible by
replicating standard intraoperative pin
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2022 Wolters Kluwer Health, Inc. All rights reserved.
placement, despite this being the oppo-
site to the anterior to posterior in-
clination recommended by Skaggs et al3
in accordance with normal anatomy?
We would also like to ask the authors if
they agree that their statement that “it
should not matter if the arm is in in-
ternal or external rotation because it is a
percentage” is a factual error, since on a
true lateral view the perceived postero-
anterior distance of the distal humeral
fragment will be at its smallest and will
increase with arm rotation in either di-
rection in relation to the xray beam. The
perceived width of the proximal frag-
ment will either increase or decrease,
depending on the initial rotational rela-
tionship between the 2 fragments.
Therefore, the LRP will change in at
least half (if not in all) of the rotational
scenarios if the lateral radiograph is not
taken as a true lateral radiograph.
Ryan et al4 reported that the
AHL runs through the middle third of
the capitellum in 100% of children
5 years of age and above but Berdis
et al’s1 AHL runs through the front of
the capitellum, indicating that the used
bone model might not be an anatomic
reproduction of a normal child’s hu-
merus, which raises doubts about the
reliability of the measured percentages.
Bahk et al5 defined malrotation
arbitrarily as a discrepancy of more than
20% between the width of the distal hu-
merus above and below the fracture on
the lateral radiograph, which is the only
documentation of a definition of malro-
tation we could identify in the literature.
The American Academy of Orthopaedic
Surgeons on the other hand does not
consider rotational malalignment in their
Appropriate Use Criteria at all (http\
\www.aaos.org/quality/quality-programs/
pediatric-supracondylar-humerus-
fractures). In contrary to the latter,
O’Hara et al6 recommended K-wire fix-
ation for fractures with any degree of
malrotation but not for those without
malrotation.
We assess intraoperative rota-
tional malalignment by taking internal
and external oblique elbow views,
which show the alignment of the me-
dial and lateral column, with our im-
age intensifier machines not allowing
to perform any measurements. Could
Berdis et al1 explain how their meas-
urements help with intraoperative
Letters to the Editor
decision making, as they stated in their
discussion, since we generally do not
accept malrotation, not using a specific
LRP value as a decision guide?
In conclusion, there are multiple
factors which can affect the LRP, with
Berdis et al’s measured LRPs only ap-
plying to one specific K-wire position
without any coronal or sagittal dis-
placement. Therefore, it is highly likely
that the provided angles do not apply to
the majority of intraoperative and
postoperative scenarios, because we fre-
quently accept some displacement, have
varying wire positions and frequently do
not get true lateral views in the clinic.
Therefore, we do not think that Berdis
et al’s1 data can be used reliably for the
majority of fractures to estimate the true
degree of malrotation, with it being
hitherto unknown if any malrotation/
LRP can be accepted, without affecting
cosmesis and/or function.
Andreas Rehm, MD*
Azeem Thahir, MD†
Elizabeth Ashby, MD*
Pinelopi Linardatou Novak, MBBCh*
Departments of *Paediatric Orthopaedics
†Trauma & Orthopaedics, Cambridge
University Hospitals NHS Trust
Cambridge, UK
The authors declare no conflicts of interest.
DOI: 10.1097/BPO.0000000000002280
REFERENCES
1. Berdis G, Hooper M, Talwalker V, et al.
Assessment of lateral rotation percentage and
rotational deformity of the elbow in type 3
supracondylar humerus fractures: a biomechan-
ical study. J Pediatr Orthop. 2021;41:e605–e609.
2. Camp J, Ishizue K, Gomez M, et al. Alter-
ation of Baumann’s angle by humeral
position: implications for treatment of supra-
condylar humerus fractues. J Pediatr Orthop.
1993;13:521–525.
3. Skaggs DL, Cluck MW, Mostofi A, et al.
Lateral-entry pin fixation in the management
of supracondylar fractures in children. J Bone
Joint Surg Am. 2004;86:702–707.
4. Ryan DD, Lightdale-Miric NR, Joiner ERA,
et al. Variability of the anterior humeral line
in normal pediatric elbows. J Pediatr Orthop.
2016;36:e14–e16.
5. Bahk MS, Srikumaran U, Ain MC, et al.
Patterns of pediatric supracondylar humerus
fractures. J Pediatr Orthop. 2008;28:493–499.
6. O’Hara LJ, Barlow JW, Clarke NMP.
Displaced supracondylar fractures of the
humerus in children. J Bone Joint Surg Br.
2000;82-B:204–210.
www.pedorthopaedics.com | e191