Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616

Available online at

ScienceDirect
www.sciencedirect.com

Original article

The effect of coronal decompensation on the biomechanical

parameters in lower limbs in adolescent idiopathic scoliosis
István Márkus a,∗ , Ádám Tibor Schlégl a , Máté Burkus a,b , Kristóf József a , Bálint Niklai a ,
Péter Than a , Miklós Tunyogi-Csapó a
a
University of Pécs Medical School, Department of Orthopaedics, Akác street 1, 7632 Pécs, Hungary
b
Department of Traumatology and Hand Surgery, PetzAladár County Teaching Hospital, VasváriPál street 2-4, 9023 Győr, Hungary

a r t i c l e i n f o a b s t r a c t

Article history: Study hypothesis: We hypothesized that altered coronal balance in adolescent scoliosis leads to asym-
Received 9 February 2018 metric stress on the lower limbs, with subsequent effects on bone maturation and later morphology. We
Accepted 12 June 2018 aimed to assess the correlation between the biomechanical parameters of the lower limbs and coronal
balance in idiopathic scoliosis.
Keywords: Materials and methods: In this study, EOS images of 280 patients and 56 controls were randomly selected
Idiopathic scoliosis from our clinics’ database. The average age of AIS patients was 14.5 years and average Cobb angle 33.48◦ .
EOS 2D/3D
Three D reconstructions of the pelvis and lower limbs were performed and coronal balance assessed. Reli-
Lower limb
Coronal balance
ability of measurements was ensured by intra- and inter-observer agreement. During statistical analysis
Biomechanical parameter the Kolmogorov-Smirnov test, t-test and linear regression analysis were performed. A p value < 0.05 was
considered significant.
Results: Of the 15 examined lower limb parameters, a significant difference between sides was found in
those with AIS for femur length, total length, collodiaphyseal angle, angle between the femoral mechanical
and anatomical axis and tibial torsion. In addition, the tibial length and the mechanical tibiofemoral angle
were significantly higher than those of the controls. The coronal balance was found to be the strongest
predictive factor showing a significant correlation with all of the previous parameters, except tibial tor-
sion. With patients grouped based on coronal balance (compensated, right and left decompensated) the
paired t-test also supported these findings.
Conclusion: The biomechanical parameters of the lower limbs are affected in cases of scoliosis with an
altered coronal balance. It was thought that a shift in balance in the coronal plane accounted for the small
but significant changes seen in our study, with the lower limb on the side of decompensation becoming
shorter in comparison to its’ counterpart, with a lower collodiaphyseal angle and increased varus at the
knee joint.
Level of evidence: III, case-control study.
© 2018 Published by Elsevier Masson SAS.

1. Introduction
Adolescent idiopathic scoliosis is the most common orthopaedic
disease affecting the spine during adolescence. Numerous stud-
ies deal with the aetiology of the disease, X-ray morphology and
classification of curvatures and how the disorder disrupts the
biomechanical balance of the body [1–3]. This is probably best indi-
cated by the changes of the sagittal and coronal balance of the spine
[4,5], however the disorder is a complex three-dimensional defor-
∗ Corresponding author.
E-mail address: markus.istvan@pte.hu (I. Márkus).
mity that affects the body as a whole, and these effects must not be
overlooked. Yagi et al., for example, noted changes in posture and
walking pattern associated with scoliosis, [4], Kotwiczki et al. found
significant asymmetries in the rotational motion between the two
hips, despite no change in range of motion, in patients with scoliosis
[5]. Such studies suggest that the altered biomechanical equilib-
rium, as a result of changes due to adolescent idiopathic scoliosis,
may affect the bone anatomy and biomechanics of the lower limb.
However, in the literature we reviewed, only one publication eval-
uating the bony anatomy of the lower limbs in scoliosis was found.
Saji et al. evaluated the collodiaphyseal angle with conventional X-
rays, and found a significantly increased angle in the scoliosis group
[6].

https://doi.org/10.1016/j.otsr.2018.06.002
1877-0568/© 2018 Published by Elsevier Masson SAS.
610 I. Márkus et al. / Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616
No publications were found assessing the connection between
sagittal or coronal balance and the biomechanical parameters of
the lower extremity, although Chen et al. reported a change in
bone mineral density with coronal balance in AIS [7]. The paucity
of publications may be due to the lack appropriate testing modali-
ties. Assessment of the biomechanical parameters of the lower limb
and the pelvis is possible with physical examination although it’s
rather complicated and inaccurate [8]. Furthermore, although both
conventional MR and CT devices can be used, the devices available
in most clinical settings are not suitable for imaging the upright
patient, with the skeleton under stress, which is required for inves-
tigation of biomechanical equilibrium.
The EOS 2D/3D scanner (Eos Imaging, France) allows the simul-
taneous anteroposterior and lateral imaging of the entire body in
a standing position with ultra-low radiation [9]. Using the asso-
ciated SterEOS (TM) software, 3D surface reconstruction of the
spine, pelvis and lower extremities is possible. The quality of 3D
models is significantly greater than conventional X-ray imaging
or that of a typical scout CT, with significantly lower radiation
exposure [10,11] and numerous studies, including many from
our institute, have been published about its clinical usefulness
and accuracy [11–13]. Three-dimensional mapping of the spine
provides additional, useful information about the curvature and
facilitates surgical planning.
We hypothesized that an altered coronal balance leads to
uneven stress of the lower limbs, which then affects the biome-
chanical parameters of the lower limb. The aim of our study was
to use 3D imaging methods, in a large population, to investigate
first if there is a change in the bone anatomy of the lower limb in
adolescents with AIS. Secondarily, which features of the deformity
have greatest effect on the lower limb biomechanical parameters
in AIS (i.e. severity, localization, coronal balance)?
2. Materials and methods
The EOS 2D/3D scanner has been in use at our clinic since
2007, and it has become a part of daily routine diagnostics. In
the present study, 2579 images were reviewed, the records of all
patients scanned between 2007 and 2012 at our outpatient clinic
aged between 10 and 18 years old. All images, without exception,
were made with orthopaedic indication, and a written agreement
from the parents was acquired to use the images for future research
purposes. Based on the adolescent scoliosis criteria adopted by SRS,
patients were admitted if they were between the age of 10 to 18
years, had a Cobb angle greater than 15 degrees and less than 90
degrees, exhibited a curvature of structural origin and in which
any secondary causes could be excluded. Two hundred and eighty
were then randomly selected, although in 10 cases high quality
reconstruction was not possible due to technical problems (such as
patient motion during examination, missing landmarks or overlap-
ping of important structures, etc.), and so these cases were removed
and new patients were enrolled.
Finally, 3D reconstruction of the lower limb was successful in
280 adolescents (218 girls, 62 boys, average age 14.51 years old),
with a gender distribution corresponding to the distribution of
scoliosis in our overall population [14]. In order to ensure a homo-
geneous group of patients, efforts were made to choose the same
number of cases from each age group during the selection phase.
The cases were grouped by Lenke classification, and after deter-
mination of severity of spinal deformity based on Cobb angle they
were also divided into 4 groups according to the SOSORT guidelines
[15,16].
We estimated bone age using the “Risser +” classification. A con-
trol group of 56 cases, analogous in gender and age, was selected
from our clinical database. These individuals lacked any spinal
deformity and had no previous surgery on the musculoskeletal
system. Examinations of these individuals had been performed pre-
dominantly for low back pain and knee complaints, though finally
no orthopaedic disorders were found. The distribution of the test
and control groups is shown in Tables 1 and 2.
Following patient selection, operator intra- and inter-observer
reliability with the SterEOS 3D reconstruction (EOS Imaging,
Paris, France, version: 1.4.4.5297) was assessed using intraclass
correlation coefficient. Three ‘operators’ (physicians trained and
experienced in the use of the software) completed the recon-
struction of 30 randomly selected cases, 3 times, on 3 different
days. For evaluation, the criteria set by Winer were used: between
0–0.24 was regarded as weak, 0.25–0.49 as low, 0.50–0.69 medium,
0.70–0.89 good, and 90–100 as excellent reliability [17].
After excellent reliability results, we performed full SterEOS 3D
reconstruction of both lower limbs in all 280 cases and in the 56
controls. The software calculated the value of 15 parameters of the
lower limb, and the pelvic tilt, automatically (Fig. 1).
In order to examine the laterality of any differences between
lower extremity parameters, the right limb results were subtracted
from those of the left limb in each case, so not only an absolute value
was obtained but, depending on the sign of the given number, its’
side. The coronal balance was determined on the EOS anteropos-
terior recordings with the help of the 2D SterEOSTM software. The
distance between a vertical line from the midline of the cervical C7
vertebra and central sacral vertebral line (CSVL) was determined
with millimeter accuracy (Fig. 2). The resulting value, indicating
the displacement of the coronal balance, is marked by a negative
sign if the upper body is to the right of the centre of the sacrum
plateau, and positive if it is to the left.
Three groups of the examined patients were established based
on the measured coronal balance. A coronal balance shift below
1.0 cm was taken as compensated, and all cases where this value
was greater than 1 cm were considered decompensated. Decom-
pensated cases were divided into two further groups depending on
whether the patients’ upper body was to the left or to the right
during the examinations.
IBM SPSSTM (IBM Corp., Armonk, NY, USA), version 22, and
Microsoft Office Professional Plus v14.0.6112.5000 (Microsoft
Corp., Redmond, WA, US) were used for statistical processing.
During the statistical analysis the normal distribution of the data
was evaluated by the Kolmogorov-Smirnov test. The correlation
between the coronal balance, the severity of deformity, the pelvic
tilt and the parameters of the lower limb were examined by linear
regression analysis. The difference between the two lower limbs
was examined by paired t-test. The difference between the AIS and
control groups as well the difference between genders were tested
by independent t-test. An ANOVA test was used to compare the
groups based on Lenke classification and also on coronal balance. In
all cases, p < 0.05 values were considered significant for our results.
3. Results
Intra- and inter-observer reliability assessments gave values
above 0.9, regarded as ‘excellent’, for all operators and all param-
eters. Based on the Kolmogorov-Smirnov test, all parameters
examined were normally distributed.
Of the total 280 patients, 118 were considered ‘compensated’ in
the coronal plane based on our criteria, meaning the midline of C7
was less than 1 cm displaced from the sacral midline. Of the 162
patients whose upper body was ‘decompensated’, 90 cases were
displaced to the left of the sacral midline and 72 cases to the right
(Table 3).
In patients with AIS, a significant difference between sides
was found in femoral length, total lower limb length, CD angle,
 I. Márkus et al. / Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616 611

Table 1
Distribution of population by age, gender and Lenke classification.

Lenke main group 1 2 3 4 5 6 Sum Control

Lenke lumbar mod. A B A A B C A C A B C A B C

Age (years)

10 Male 1 1 2 14
Female 7 2 2 1 12 2
11 Male 2 2 14
Female 2 2 1 1 3 2 1 12 2
12 Male 2 1 1 4 34 1
Female 12 2 3 2 2 1 4 1 1 2 30 6
13 Male 3 2 1 6 38 1
Female 13 1 2 2 4 2 6 2 32 7
14 Male 3 2 1 3 2 11 39 2
Female 7 3 1 2 4 1 2 2 6 28 6
15 Male 5 1 2 3 11 37 2
Female 10 1 2 1 2 1 8 1 26 5
16 Male 4 2 3 1 10 38 2
Female 5 2 1 3 9 6 2 28 6
17 Male 3 1 1 1 4 10 40 3
Female 8 1 1 4 2 1 6 5 2 30 6
19 Male 1 1 2 4 26 1
Female 6 2 1 3 7 3 22 4
Sum 92 7 15 6 18 19 1 1 4 40 59 1 1 16 280 56

Bold numbers are the summarized value of the previous raws.

Table 2
Distribution of the population grouped by age and ‘Risser +’ stage, severity clusters formed based on 2016 SOSORT guidelines.

Age (year) Severity Risser stage Av. Cobb angle

0- 0 1 2 3 4 5 Sum

10 Low (10–20◦ ) 9 3 12 21.38

Moderate (21–40◦ ) 1 1 2
Severe (41–55◦ )
Very severe (56◦ –)
11 Low (10–20◦ ) 10 1 11 23.38
Moderate (21–40◦ ) 1 1 1 3
Severe (41–55◦ )
Very severe (56◦ –)
12 Low (10–20◦ ) 1 6 2 4 3 16 37.7
Moderate (21–40◦ ) 5 2 2 1 2 12
Severe (41–55◦ ) 2 1 2 5
Very severe (56◦ –) 1 1
13 Low (10–20◦ ) 1 3 3 5 12 38.6
Moderate (21–40◦ ) 2 2 3 2 2 3 14
Severe (41–55◦ ) 3 3 6
Very severe (56◦ –) 1 3 1 1 6
14 Low (10–20◦ ) 2 3 3 2 1 11 41.62
Moderate (21–40◦ ) 1 2 7 10
Severe (41–55◦ ) 1 3 5 9
Very severe (56◦ –) 2 1 5 1 9
15 Low (10–20◦ ) 1 1 9 1 12 31.71
Moderate (21–40◦ ) 3 1 2 6 1 13
Severe (41–55◦ ) 2 1 2 1 6
Very severe (56◦ –) 1 2 3
16 Low (10–20◦ ) 1 1 7 5 14 34.49
Moderate (21–40◦ ) 2 1 12 15
Severe (41–55◦ ) 6 6
Very severe (56◦ –) 1 1 1 3
17 Low (10–20◦ ) 1 9 9 19 32.41
Moderate (21–40◦ ) 9 3 12
Severe (41–55◦ ) 3 3 6
Very severe (56◦ –) 2 1 3
18 Low (10–20◦ ) 5 6 11 31.04
Moderate (21–40◦ ) 3 11 14
Severe (41–55◦ ) 1 3 4
Very severe (56◦ –) 0
Sum Low (10–20◦ ) 10 20 10 6 10 40 22 118 33.48
Moderate (21–40◦ ) 9 4 11 7 7 42 15 95
Severe (41–55◦ ) 2 3 1 7 22 7 42
Very severe (56◦ –) 2 5 2 1 11 4 25
Total 19 28 29 16 25 115 48 280

Bold numbers are the summarized value of the previous raws.

612 I. Márkus et al. / Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616
Fig. 1. Measured parameters. FMAn: femoral mechanical angle; FM-FS: femoral mechanical axis-femoral shaft angle; mTFA: mechanical tibiofemoral angle; TMAn: tibial
mechanical angle.
FM-FS and tibial torsion. No significant difference was found
between sides in any parameters in the control group. In addition
to previous parameters, significantly higher values for tibial length
and mTFA were observed in AIS when compared with the control
(Table 4).
Differences between genders were found in the sTFA (mean
difference = 1.57◦ , p = 0.18) and femorotibial rotation (mean differ-
ence = 1.86◦ , p = 0.034), however the magnitude of the difference
was not affected by age nor the curve location as per Lenke classifi-
cation (data divided by gender, age and Lenke groups can be found
in E-supplementary material 1).
Results from three other factors that may cause biomechanical
changes in the lower limb – Cobb angle, pelvic tilt and coronal bal-
ance – are reported in Table 4. The Cobb angle showed a significant
correlation with femur length, total limb length, CD angle, FM-FS
and femorotibial rotation. Significant correlation was between the
pelvic tilt and femur length, tibia length, total length and CD angle.
Coronal balance was found to have a significant correlation with
femur length, tibial length, total length, CD angle and mTFA, show-
ing a negative correlation i.e. lower values on the side to which C7
was displaced. A significant positive correlation (higher values on
the side to which C7 deviated) was found with the femoral offset
and FM-FS.
When compensated and decompensated cases were examined
separately, similar results were seen (Table 5, Fig. 3). In those
parameters where a significant correlation was found (femoral
length, total limb length, collodiaphyseal angle and mTFA) the
paired t-test also showed significant correlation in decompensated
cases, while in compensated cases no correlation was found. The
direction of the changes also agreed with the direction of beta
coefficient calculated from regression analysis. The only exceptions
were femoral offset, where, although the trend of the data matched
expectations, there was a high standard deviation for cases decom-
pensated to the left (SD = 4.72◦ ), and the correlation did not reach
statistical significance (p = 0.069). A significant difference was also
seen between sides in patients with compensated coronal balance,
namely in the cases of FM-FS (diff = 0.21◦ , p = 0.01) and tibial torsion
(diff = 1.75◦ , p = 0.003).
4. Discussion
Our findings supported our hypothesis that the disruption
of biomechanical balance seen in idiopathic scoliosis affects the
bone anatomy of the lower limb. We found significant differences
between sides in 5 of the investigated 15 lower limb parameters
(femur length, total lower limb length, CD angle, FM-FS and tibial
torsion), moreover considerable differences between sides were
seen in mTFA and tibial length compared to the control group. It
should be noted that while a clear trend was seen, the differences
are small, which likely do not have clinical relevance in adolescent
scoliosis.
The main limitation of our study was its’ cross-sectional design.
We did not have information about the exact time of appearance
of each patients’ deformity, so we cannot see if lower limb bony
growth is affected specifically at certain ages. It is not clear whether
the lower limb asymmetry contributed to the scoliosis or appeared
after the deformation of the spine. Furthermore, as we could not
directly measure each lower limb’s load we can not state with
100% certainty that the coronal asymmetry also caused asymmetric
stress.
After analysis, neither the Lenke curve type nor the age of the
individual were found to have an effect on the magnitude of the
differences between sides. A gender difference in mean values was
only seen in the sagittal femorotibial angle (mean 2.2◦ higher in
girls) and femorotibial rotation (mean 1.2◦ higher in girls), but
the difference was not significant. This gender difference may be
explained by the fact that these parameters are influenced by the
ligamental stability of the knee which generally have more laxity
in girls.
Three factors that may also cause biomechanical changes in the
lower limbs were also investigated: coronal balance, pelvic tilt and
Cobb angle. Coronal balance appeared to be the strongest influ-
encing factor, as it showed significant correlation with 7 of the 15
parameters. The severity of the curvature correlated with only 5
parameters, and with a lower beta coefficient. The pelvic tilt had a
strong correlation with the length parameters and CD angle which
can explained by the geometrical connection between them.
 I. Márkus et al. / Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616 613

Fig. 2. Coronal balance measurement. Displacement was calculated as the distance between the midline of the cervical C7 and the sacral S1 vertebrae, in the horizontal plane
(right displacement–negative, left displacement–positive). Left: 17 years old, 2.9 cm coronal decompensation, 28 ◦ Cobb angle, CD angle difference: 11.2◦ . Total limb length
difference: 1.7 cm, mTFA difference: 3.5◦ ; centre: 16 years old, 0.2 cm coronal decompensation, 62 ◦ Cobb angle, CD angle difference: 1.2◦ ; total limb length difference: 0.4 cm,
mTFA difference: 0.6◦ ; right: 18 years old, 2.1 cm coronal decompensation, 40 ◦ Cobb angle, CD angle difference: 6.2◦ . Total limb length difference: 1.4 cm, mTFA difference:
2.4◦ .

Five parameters showed a decrease on the side of decompensa-
tion based on the coronal balance: femur length, tibial length, total
length, CD angle and mTFA (mTFA saw a varus deviation on the
side of decompensation). Two parameters increased in the side of
the decompensation: femoral offset and FM-FS. Taken together, a
picture emerges in which the side of decompensation (the side to
which the center of gravity is shifted) sees changes in the femoral
neck and knee joint, a decrease in the length of the femur, and the
length of the whole lower limb, while the femoral offset and the
FM-FS increase.
We believe that one of the key aspects of the change in the
femur is the decreasing CD angle and its’ geometric correlation to
the other parameters. If the femoral parameters are imagined as a
triangle enclosed by the femoral neck, anatomical axis and mechan-
ical axis of the femur, it is obvious that the decrease of the CD angle
(the angle formed between the femoral neck and anatomical axis
of the femur) would be accompanied by the increase of the FM-FS
(angle between the anatomical and mechanical axes of the femur).
Furthermore, as the femoral neck length remained unchanged, a
decreasing CD angle would also be associated with a decrease in
614 I. Márkus et al. / Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616

Table 3
Distribution of population by age and coronal balance.

Age (year) Left decompensated Compensated Right decompensated

n Decomp. (cm) Cobb angle Pelvic tilt (mm) n Cobb angle Pelvic tilt (mm) n Decomp. (cm) Cobb angle Pelvic tilt (mm)

10 1 1.50 15.08 4.36 7 21.61 ± 7.00 4.89 ± 3.29 6 2.18 ± 1.10 22.18 ± 4.10 5.26 ± 5.90
11 4 1.75 ± 0.76 20.41 ± 7.88 3.24 ± 4.04 4 19.38 ± 4.69 4.20 ± 2.35 6 1.67 ± 0.65 28.04 ± 13.78 5.03 ± 6.09
12 6 2.36 ± 0.75 25.94 ± 9.63 4.40 ± 2.87 19 28.69 ± 18.09 9.10 ± 6.63 9 1.93 ± 0.98 40.29 ± 22.68 4.95 ± 4.69
13 11 2.27 ± 0.64 28.56 ± 12.02 7.40 ± 3.63 17 41.06 ± 18.15 7.44 ± 3.81 10 2.03 ± 0.66 43.92 ± 20.62 6.82 ± 4.95
14 15 2.36 ± 0.87 45.63 ± 20.16 7.82 ± 5.02 17 36.67 ± 21.87 6.16 ± 5.39 7 2.54 ± 1.10 45.05 ± 21.72 5.82 ± 4.57
15 13 1.97 ± 0.63 33.11 ± 13.69 8.42 ± 6.52 14 27.62 ± 19.72 6.30 ± 4.33 10 1.72 ± 0.45 36.71 ± 19.58 4.28 ± 4.40
16 17 1.61 ± 0.49 39.02 ± 15.90 5.56 ± 4.08 14 25.98 ± 9.51 6.70 ± 4.15 7 1.67 ± 0.67 40.48 ± 23.09 4.37 ± 4.30
17 13 1.80 ± 0.54 33.15 ± 12.77 7.04 ± 4.15 16 32.81 ± 19.92 6.36 ± 4.52 11 1.80 ± 0.55 30.93 ± 16.27 2.92 ± 2.78
18 10 1.86 ± 0.65 36.25 ± 14.31 5.78 ± 4.53 10 30.97 ± 15.37 3.42 ± 4.19 6 2.31 ± 1.07 22.20 ± 11.98 6.89 ± 4.52

The table shows the distribution of patients by age group, the Cobb angle’s and Pelvic tilt’s average value in each group, distribution of compensated, right and left
decompensated cases within a given group and the average degree of decompensation and its deviation. Bold numbers are the summarized value of the previous raws.

Table 4
Summary of statistical analyses.

Side difference (p) Diff. from control (p) Cobb angle Coronal balance Pelvic tilt

Parameter beta p beta p beta p

Femur length (cm) 0.039 0.001 −0.219 < 0.001 −0.261 < 0.001 −0.296 < 0.001
Tibia length (cm) 0.057 0.041 −0.087 0.145 −0.128 0.031 −0.328 < 0.001
Total length (cm) 0.049 < 0.001 −0.199 0.001 −0.231 < 0.001 −0.405 < 0.001
Fem. head diam. (mm) 0.869 0.652 0.034 0.570 −0.081 0.173 −0.046 0.443
Femoral offset (mm) 0.194 0.245 0.154 0.090 0.128 0.032 0.045 0.452
Neck length (mm) 0.178 0.056 0.680 0.255 0.450 0.455 −0.023 0.697
CD angle (◦ ) 0.029 0.002 −0.197 0.001 −0.210 < 0.001 −0.151 0.011
mTFA (◦ ) 0.442 0.003 −0.024 0.682 −0.192 0.002 0.049 0.412
sTFA (◦ ) 0.053 0.073 −0.108 0.071 −0.108 0.070 −0.049 0.412
Fem. mech .angle (◦ ) 0.555 0.895 −0.027 0.651 −0.115 0.053 0.042 0.411
Tibial mech. angle (◦ ) 0.065 0.473 0.156 0.080 0.040 0.502 0.073 0.222
FM-FS (◦ ) < 0.001 0.003 0.121 0.042 0.165 0.005 −0.019 0.752
Femoral torsion (◦ ) 0.954 0.031 0.123 0.059 0.081 0.174 −0.042 0.481
Tibial torsion (◦ ) < 0.001 0.004 −0.104 0.821 −0.084 0.158 0.004 0.946
Femorotibial rot. (◦ ) 0.671 1.895 0.146 0.014 0.440 0.464 −0.030 0.610

Side difference presents the results of the paired-samples t-tests, where the left side was compared to the right side. The difference from control demonstrates the results of
the independent samples t-test comparing the side difference of the AIS and control group. The Cobb angle, Coronal balance and Pelvic tilt represents its’ correlation with the
side difference of the parameter examined by linear regression analysis. (p < 0.05 are highlighted, Diff: difference, Fem: femur/femoral, mech: mechanical, diam: diameter,
rot: rotation). Bold numbers are the summarized value of the previous raws.

Table 5
Summary of side differences of lower limb parameters.

Left decomp. Compensated Right decomp. Control

n 90 118 72 56

Parameter Mean S.D. p Mean S.D. p Mean S.D. p Mean S.D. p

Femur length (cm) −0.16 0.40 0.013 0.07 0.38 0.088 0.16 0.38 0.009 0.02 0.18 0.858
Tibia length (cm) −0.10 0.31 0.046 0.00 0.28 0.923 0.07 0.28 0.049 0.01 0.25 0.795
Total length (cm) −0.15 0.61 0.010 0.07 0.56 0.205 0.18 0.53 0.014 0.05 0.88 0.295
Fem. head diam. (mm) −0.11 1.33 0.399 0.01 1.10 0.935 0.16 1.25 0.423 −0.03 2.15 0.896
Femoral offset (mm) 0.40 4.72 0.062 −0.52 4.04 0.191 −0.99 4.26 0.064 0.18 3.28 0.436
Neck length (mm) 0.12 4.19 0.162 −0.51 3.79 0.234 −0.56 3.75 0.768 0.18 5.12 0.564
CD angle (◦ ) −0.45 4.04 0.020 0.36 3.46 0.294 1.19 4.15 0.038 −0.19 2.58 0.191
mTFA (◦ ) −0.46 2.06 0.031 −0.17 1.85 0.338 0.52 1.95 0.024 0.26 3.59 0.177
sTFA (◦ ) −1.40 4.81 0.356 −0.03 3.96 0.977 −0.51 5.27 0.128 2.15 3.96 0.069
Fem. mech .angle (◦ ) −0.79 1.59 0.162 −0.67 1.49 0.058 −0.39 1.78 0.091 0.02 2.60 0.295
Tibial mech. angle (◦ ) 0.86 2.47 0.073 0.49 2.44 0.071 0.87 2.24 0.065 0.48 3.59 0.292
FM-FS (◦ ) 0.48 1.11 0.009 0.21 1.03 0.006 −0.22 1.22 0.021 0.02 1.56 0.851
Femoral torsion (◦ ) 1.40 8.07 0.973 −0.16 7.81 0.201 −1.06 9.56 0.675 0.98 8.89 0.592
Tibial torsion (◦ ) −1.93 6.28 0.399 −1.75 6.04 0.028 −1.25 7.05 0.064 0.48 7.56 0.292
Femorotibial rot. (◦ ) 0.42 6.09 0.499 −0.50 5.73 0.174 −0.33 6.80 0.361 -0.19 6.25 0.494

Data is presented as mean and standard deviation. For evaluating difference, the right limb parameter was subtracted from the left one in all cases (positive value = left value
was higher, negative = right was higher). Significance between sides was evaluated with paired t-test (p values are presented in the table, significant values are highlighted
in bold). (Fem: femur/femoral, mech: mechanical, diam: diameter, rot: rotation). Bold numbers are the summarized value of the previous raws.

the femoral length (aka the mechanical axis) and an increase in
femoral offset, all of which were seen in our results.
The concept of limb changes due to altered balance is further
supported by the fact that the patient group with compensated
coronal balance showed no significant difference between left
and right lower limbs. The femoral offset was the only excep-
tion, as in those cases which were decompensated to the right
the correlation was not found to be significant, although the direc-
tion of the difference was in line with expectations, albeit with a
large standard deviation. In the case of FM-FS and tibial torsion,
 I. Márkus et al. / Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616 615

Fig. 3. The average side difference of the parameters grouped by coronal decompensation (mean ± standard error). *: significant side difference based on paired samples
t-test. For evaluating difference, the right limb parameter was subtracted from the left one in all cases (positive value = left value was higher, negative = right was higher).

a difference between the two sides was observed in the compen-
sated group, but we could not explain this based on the data our
study.
We could not compare these results with those of other authors
as data for these parameters have not been reported elsewhere.
Contrary to our expectations, we did not find any significant cor-
relation with torsional parameters although the trend of alteration
was in line with that expected. This was due to a large standard
deviation of the data (7.81–9.56◦ ), which was noted in earlier pub-
lications.
616 I. Márkus et al. / Orthopaedics & Traumatology: Surgery & Research 104 (2018) 609–616
The changes observed in our study support our hypothesis that
the asymmetry of stress on the lower limbs due to displaced coro-
nal balance affects their biomechanical parameters. We presume
that the reason behind this is that the biomechanical changes of
the lower limbs predominantly occur before the age of 10 [18,19],
whereas the onset of AIS, and hence the possibility to elicit changes
of a greater magnitude, occurs after that age. As such it would be
assumed that far more significant differences would be detected in
early onset scoliosis, where the age of onset of the disease is much
closer to the time of early bone development. Our results are consis-
tent with the findings of Chen et al., who also found a correlation
between coronal balance and bone mineral density difference in
the proximal femoral region in patients with AIS [7].
According to our knowledge, the connection between scoliosis
and the anatomy of the lower limbs is mentioned only in one arti-
cle published in 1995 by Saji et al. The authors reported an increase
in the value of the CD angle and its’ asymmetry in adolescent sco-
liosis as assessed with bidirectional, conventional X-ray images.
[6] In our present study, the aggregated results contradict these.
The average CD angle values, found by our research team, tend
to show a decrease compared to our previously published study
on healthy population (128.16◦ vs. 129.18◦ ) [19]. However, when
curve direction is investigated, they too saw an increase in CD angle
asymmetry, which was not highlighted in their study. Patients with
right thoracic curves exhibited lower CD angles on the right (mean
difference 1.1◦ ± 2.2◦ ), whereas in left lumbar curves, the CD angle
was lower on the left (mean difference 2.7◦ ± 2.7◦ ).
The limitations of our study include the fact that the patients
are distributed throughout a group where the development of the
lower limbs is still ongoing, although growth is not as dynamic as in
younger ages. To remedy this, efforts were made to involve patients
in a similar number across all age groups. Also subjects were not
equally distributed by gender, due to the natural presentation of
the disease, so we felt our population was not suitable for detecting
significant gender differences. Due to the retrospective nature, BMI
data of patients were not available which may also have affected
the load of the lower limbs. It should also be mentioned that some
deviations might occur, due to the nature of the EOS scanner posi-
tioning, as the patients step forward slightly with one foot, and the
tests are performed in that position.
The findings provided in the present study showed that patients
with decompensated adolescent idiopathic scoliosis have alter-
ations in the biomechanical parameters of the lower limbs. The
femoral neck and knee joint changes, and the length of the femur
and whole lower limb decreased on the side to which the coronal
balance was shifted, although differences were small and likely do
not impact the clinical picture of the deformity.
Disclosure of interest
The authors declare that they have no competing interest.
Funding sources
This work was supported by the GINOP 2.3.2-15-2016-00022
grant of the Hungarian Government.
The present scientific contribution is dedicated to the 650th
anniversary of the foundation of the University of Pécs, Hungary.
Authors’ contribution
István Márkus: performed the analysis of all patient related
data, take part in the 3D EOS reconstructions, and coronal balance
measurement, performed the statistical analyses, put together the
final manuscript.
Ádám Tibor Schlégl: helped to built the study design, 3D EOS
reconstructions, reviewed the manuscript, verifyed the statistical
analyses.
Máté Burkus, Kristóf József, Bálint Niklai: analyzed datas and
performed the most of the 3D EOS reconstructions, reviewed the
manuscript.
Miklós Tunyogi-Csapó and Péter Than: designed the study,
supervised all the analysis lower limb parameters, managed
patient-related administrative care and institutional requirements.
All authors have read and approved the final submitted version.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.otsr.2018.06.002.
References
[1] Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis: a new clas-
sification to determine extent of spinal arthrodesis. J Bone Joint Surg Am
2001;83:1169–81.
[2] Wajchenberg M, Astur N, Kanas M, et al. Adolescent idiopathic scoliosis: cur-
rent concepts on neurological and muscular etiologies. Scoliosis Spinal Disord
2016;11:4.
[3] Donzelli S, Poma S, Balzarini L, et al. State of the art of current 3-D scoliosis clas-
sifications: a systematic review from a clinical perspective. J Neuroeng Rehabil
2015;16:12–91.
[4] Yagi M, Ohne H, Konomi T, et al. Walking balance and compensatory gait
mechanisms in surgically treated patients with adult spinal deformity. Spine J
2017;17:409–17.
[5] Kotwicki T, Walczak A, Szulc A. Trunk rotation and hip joint range of rota-
tion in adolescent girls with idiopathic scoliosis: does the “dinner plate” turn
asymmetrically? Scoliosis 2008;19:1–3.
[6] Saji MJ, Upadhyay SS, Leong JC. Increased femoral neck-shaft angles in adoles-
cent idiopathic scoliosis. Spine 1995;20:303–11.
[7] Chen RQ, Watanabe K, Hosogane N, et al. Spinal coronal profiles and proxi-
mal femur bone mineral density in adolescent idiopathic scoliosis. Eur Spine J
2013;22:2433–7.
[8] Kraus VB, Vail TP, Worrell T, et al. A comparative assessment of alignment
angle of the knee by radiographic and physical examination methods. Arthritis
Rheum 2005;52:1730–5.
[9] Charpak G. Prospects for the use in medicine of new detectors of ionizing
radiation. Bull Acad Natl Med 1996;180:161–8 [discussion 8–9].
[10] Hui SC, Pialasse JP, Wong JY, et al. Radiation dose of digital radiography (DR)
versus micro-dose x-ray (EOS) on patients with adolescent idiopathic scolio-
sis: 2016 SOSORT- IRSSD “John Sevastic Award” Winner in Imaging Research.
Scoliosis Spinal Disord 2016;11:46.
[11] Melhem E, Assi A, El Rachkidi R, et al. EOS((R)) biplanar X-ray imaging:
concept, developments, benefits, and limitations. J Child Orthop 2016;10:
1–14.
[12] Schlegl AT, Szuper K, Somoskeoy S, et al. Three dimensional radiological
imaging of normal lower-limb alignment in children. Int Orthop 2015;39:
2073–80.
[13] Hirsch C, Ilharreborde B, Mazda K. Flexibility analysis in adolescent idiopathic
scoliosis on side-bending images using the EOS imaging system. Orthop Trau-
matol Surg Res 2016;102:495–500.
[14] Rogala EJ, Drummond DS, Gurr J. Scoliosis: incidence and natural history. A
prospective epidemiological study. J Bone Joint Surg Am 1978;60:173–6.
[15] Lenke LG, Edwards 2nd CC, Bridwell KH. The Lenke classification of adolescent
idiopathic scoliosis: how it organizes curve patterns as a template to perform
selective fusions of the spine. Spine 2003;28:199–207.
[16] Negrini S, Hresko TM, O’Brien JP, et al. Recommendations for research studies
on treatment of idiopathic scoliosis: Consensus 2014 between SOSORT and SRS
non-operative management committee. Scoliosis 2015;10:8.
[17] Winer BJ, Brown DR, Michels KM. Statistical principles in experimental design.
3rd ed New York: McGraw-Hill; 1991. p. 1057.
[18] Szuper K, Schlégl Á, Leidecker E, et al. Three-dimensional quantitative anal-
ysis of the proximal femur and the pelvis in children and adolescents using
an upright biplanar slot scanning X-ray system. Pediatr Radiol 2015;45:
411–21.
[19] Schlégl Á, Szuper K, Somoskeöy S, et al. Three dimensional radiolog-
ical imaging of normal lower-limb alignment in children. Int Orthop
2015;39:7.