A Comparison of Shoulder Dystocia-Associated
Transient and Permanent Brachial Plexus Palsies
Robert B. Gherman, MD, Joseph G. Ouzounian, MD, Andrew J. Satin, MD,
T. Murphy Goodwin, MD, and Jeffrey P. Phelan, MD, JD
OBJECTIVE: To estimate differences between shoulder dys-
tocia-associated transient and permanent brachial plexus
palsies.
METHODS: We performed a retrospective case-control anal-
ysis from national birth injury and shoulder dystocia data-
bases. Study patients had permanent brachial plexus palsy
and had been entered into a national birth injury registry.
Cases of Erb or Klumpke palsy with documented neonatal
neuromuscular deficits persisting beyond at least 1 year of
life were classified as permanent. Cases of transient bra-
chial plexus palsy were obtained from a shoulder dystocia
database. Non-shoulder dystocia–related cases of brachial
plexus palsy were excluded from analysis. Cases of perma-
nent brachial plexus palsy (n ⴝ 49) were matched 1:1 with
cases of transient brachial plexus palsy.
RESULTS: Transient brachial plexus palsy cases had a
higher incidence of diabetes mellitus than those with per-
manent brachial plexus palsy (34.7% versus 10.2%, odds
ratio [OR] 4.68, 95% confidence interval [CI] 1.42, 16.32).
Patients with permanent brachial plexus palsies had a
higher mean birth weight (4519 ⴞ 94.3 g versus 4143.6 ⴞ
56.5 g, P < .001) and a greater frequency of birth weight
greater than 4500 grams (38.8% versus 16.3%, OR, 0.31,
95% CI 0.11, 0.87). There were, however, no statistically
significant differences between the two groups with respect
to multiple antepartum, intrapartum, and delivery out-
come measures.
CONCLUSION: Transient and permanent brachial plexus
palsies are not associated with significant differences for
most antepartum and intrapartum characteristics.
From the Department of OB/GYN, Division of Maternal/Fetal Medicine, Na-
tional Naval Medical Center, Bethesda, Maryland; Department of OB/GYN,
Division of Maternal/Fetal Medicine, Kaiser Permanente, Baldwin Park, Cali-
fornia; Uniformed Services University of the Health Sciences, Department of
OB/GYN, Bethesda, Maryland; Department of OB/GYN, Division of Mater-
nal/Fetal Medicine, Women’s and Children’s Hospital, University of Southern
California School of Medicine, Los Angeles, California; and Childbirth Injury
Prevention Foundation, Pasadena, California.
The views expressed in this article are those of the authors and do not reflect the
official policy or position of the Department of the Navy, the Department of Defense,
or the United States Government.
544 VOL. 102, NO. 3, SEPTEMBER 2003
© 2003 by The American College of Obstetricians and Gynecologists. Published by Elsevier.
(Obstet Gynecol 2003;102:544 – 8. © 2003 by The Ameri-
can College of Obstetricians and Gynecologists.)
Brachial plexus injury has been reported to complicate
up to 21% of all deliveries complicated by shoulder
dystocia.1–3 The majority of these palsies will resolve
with conservative therapy, resulting in a 1.6% rate of
permanent brachial plexus injury associated with shoul-
der dystocia.4 To date, most of the studies dealing with
brachial plexus palsy have not specifically commented
upon permanent brachial plexus palsy. Important ques-
tions concerning the natural history and pathogenesis of
shoulder-dystocia related brachial plexus injury there-
fore continue to remain unanswered. We sought to
estimate maternal or fetal characteristics that could be
used to differentiate between transient and permanent
shoulder dystocia-associated brachial plexus injuries.
MATERIALS AND METHODS
We performed a retrospective case-control analysis from
computer-stored databases. Study patients had perma-
nent brachial plexus palsy and were part of a national
registry of children with neonatal brachial plexus inju-
ry.5 This database included litigated cases that had oc-
curred in community and teaching hospitals throughout
the United States. Under the supervision of two of this
study’s authors (JPP and JGO), data were abstracted
from the referred legal cases. Obstetric residents, attend-
ing obstetricians, certified nurse midwives, and family
practice physicians had performed the deliveries. Only
cases of brachial plexus palsy that occurred in conjunc-
tion with shoulder dystocia were included. Shoulder
dystocia was defined as the need for the performance of
ancillary obstetric maneuvers after initial attempts at
gentle downward traction were unsuccessful in deliver-
ing the fetal head.
Transient brachial plexus palsies, obtained from a
previously established shoulder dystocia database6,7
served as the control group. This database was created
0029-7844/03/$30.00
doi:10.1016/S0029-7844(03)00660-4
by two of the study’s authors (RBG and TMG) by Table 1. Comparison of Select Antepartum Characteristics
reviewing the maternal and neonatal charts of cases of in Mothers of Neonates With Transient Versus
shoulder dystocia that occurred between January 1, Permanent Brachial Plexus Palsy
1991, and July 1, 1996, at Los Angeles County/Univer- Transient Permanent P, OR
sity of Southern California Women’s Hospital. The 49 (n ⫽ 49) (n ⫽ 49) (95% CI)
cases of transient brachial plexus palsy were drawn from Maternal age (y) 27.6 ⫾ 0.9 29.1 ⫾ 0.8 .67
a cohort that included 291 cases of shoulder dystocia, Gravidity 3.53 ⫾ 0.3 3.45 ⫾ 0.3 .86
yielding a transient injury rate of 16.8%. The consecutive Parity
Nulliparous 8 (16.3) 15 (30.6) .10, 0.44
cases of transient brachial plexus palsy were matched 1:1 Multiparous 41 (83.7) 34 (69.4) (0.15, 1.28)
by the closest date of delivery, with cases from the study Maternal weight 179.9 ⫾ 4.8 206.1 ⫾ 6.7 .54
group. This study had previously been approved by the at delivery
investigational review board. Neither of the two data- (lb)
bases used in this study were funded by the government Prior shoulder 2 (4.1) 4 (8.2) .68
dystocia
or by a professional liability organization. Estimated 39.6 ⫾ 0.3 39.2 ⫾ 0.2 .23
Cases of Erb or Klumpke palsy with documented gestational
neonatal neuromuscular deficits persisting beyond at age (wk)
least 1 year of life were classified as permanent. This time ⬍ 37 2 (4.1) 3 (6.1)
frame was chosen because it represented a period during 37–40 19 (38.8) 11 (22.5)
⬎ 40 28 (57.1) 35 (71.4)
which most infants are able to make an acceptable func- Diabetes 17 (34.7) 5 (10.2) .004, 4.68
tional recovery.8,9 Cases of brachial plexus palsy that (1.42, 16.32)
occurred in the absence of shoulder dystocia were ex- OR ⫽ odds ratio; CI ⫽ confidence interval.
cluded from this analysis. We likewise excluded from Data are mean ⫾ standard deviation or n (%).
review any case in which we were unable to establish
neonatal follow-up for at least 1 year after delivery. estimate of time was noted as the time required to
The maternal medical records that we reviewed in- alleviate the shoulder dystocia. These records were also
cluded the prenatal care data and labor record. Neonatal reviewed to determine whether fundal pressure had
information was gathered both from the hospital postde- been applied to alleviate the shoulder dystocia. Neonatal
livery neonatal records and long-term follow-up. For the data included birth weight, the presence of bone fracture
latter, we reviewed pediatric, physical rehabilitation, (either clavicular or humeral fracture), other nerve in-
neurologic, and neurosurgical records. Maternal out- jury, and the location of the brachial plexus palsy.
comes assessed included maternal age, gravidity, parity, Statistical methods used included Student t test and
maternal weight at delivery, estimated gestational age as analysis of variance for continuous variables with nor-
determined by the last menstrual period or ultrasound, mal or near-normal distributions. The Mann-Whitney U
and the presence of diabetes (either insulin dependent or test was used as a nonparametric test. For proportional
non-insulin dependent). We coded charts for the intra- data, ␹2 (with Yates correction) and Fisher tests were
partum variables of oxytocin administration, epidural used, where appropriate, with calculation of odds ratio
anesthesia, length of the active phase, length of the [OR] and Cornfield 95% confidence limits.11 All analy-
second stage of labor, need for operative vaginal delivery ses were two sided, with a P value less than 0.05 consid-
(either vacuum or forceps), and the number of ancillary ered statistically significant.
obstetric maneuvers used to alleviate the shoulder dys-
tocia. The maneuvers included McRoberts maneuver,
suprapubic pressure, proctoepisiotomy, Woods cork- RESULTS
screw maneuver, Rubin maneuver, posterior arm extrac- In the permanent palsy group (n ⫽ 49) almost all of the
tion, or Zavanelli maneuver. The active phase of labor brachial plexus palsies evaluated were of the Erb or
was determined to have begun when the examination Duchenne type, with only three cases of Klumpke palsy
found the cervix to be 4 cm dilated. The second stage of present. In the transient palsy cohort (n ⫽ 49), there were
labor of labor was considered prolonged as defined by 47 cases of unilateral Erb palsy, one case of bilateral Erb
the American College of Obstetrician and Gynecolo- palsy, and one case of Klumpke palsy. Among the tran-
gists.10 sient brachial plexus palsies, an additional 17 cases were
Medical and nursing records were reviewed to deter- associated with spontaneous vaginal deliveries. An addi-
mine the time difference between the delivery of the fetal tional four cases of permanent brachial plexus palsy were
head and the delivery of the fetal body. If this informa- excluded because they did not occur in the setting of
tion was not available, then the delivering provider’s shoulder dystocia.

VOL. 102, NO. 3, SEPTEMBER 2003 Gherman et al Dystocia and Brachial Plexus Palsies 545
Table 2. Comparison of Select Intrapartum Characteristics
Transient Permanent
(n ⫽ 49) (n ⫽ 49) P, OR (95% CI)
Oxytocin administration 31 (63.3) 28 (57.1) .32, 1.29 (0.53, 3.15)
Epidural anesthesia 13 (26.5) 19 (38.8) .20, 0.57 (0.22, 1.46)
Length of active phase (min) 304 ⫾ 36.8 282 ⫾ 18.2 .64
Length of second stage 87 ⫾ 11.1 59 ⫾ 6.7 .58
⬎2h 14 (28.6) 7 (14.3) .08, 2.40 (0.79, 7.47)
⬍ 15 min 4 (8.2) 8 (16.3) .22, 0.46 (0.11, 1.85)
Operative delivery 8 (16.3) 13 (26.5) .22, 0.54 (0.18, 1.60)
Number of maneuvers to alleviate shoulder dystocia 2.6 ⫾ 0.6 2.9 ⫾ 0.4 .71
Time to alleviate shoulder dystocia (min) 1.8 ⫾ 0.3 2.1 ⫾ 0.4 .56
Abbreviations as in Table 1.
Data are mean ⫾ standard deviation or n (%).

As listed in Table 1, there were no statistically signifi-
cant differences between transient and permanent bra-
chial plexus palsies with respect to mean maternal age,
gravidity, parity, maternal weight at delivery, history of
prior shoulder dystocia, and estimated gestational age at
delivery. There were also no differences found between
the two groups with respect to the number of cases that
had estimated gestational ages greater than 40 weeks.
Cases of transient brachial plexus palsy were associated
with an increased incidence of diabetes mellitus. There
was, however, no statistically significant difference with
respect to the need for insulin among the diabetic pa-
tients (ten of 17 [transient] versus four of five [perma-
nent], P ⫽ .38).
No statistically significant differences were found be-
tween the two groups with respect to several intrapartum
variables, including oxytocin administration (either for
labor induction or augmentation), epidural anesthesia,
the mean length of the active phase of labor, the mean
length of the second stage of labor, and the need for
operative vaginal delivery. An estimate of the time inter-
val between delivery of the fetal head and body was
obtained in most of the transient (n ⫽ 36, 73.5%) and
permanent palsies (n ⫽ 46, 93.9%). The two cohorts also
did not differ with respect to the number of maneuvers
or the amount of time needed to alleviate the shoulder
dystocia (Table 2). Although fundal pressure was used
more often in cases of shoulder dystocia associated with
permanent brachial plexus palsy (one of 49 versus six of
49), this difference did not achieve statistical significance
(P ⫽ .11, OR 0.15, 95% confidence interval [CI] 0.01,
1.34).
More patients in the transient palsy group had a
second stage of labor longer than 2 hours, but this
difference did not reach statistical significance. There
was no significant difference in the rate of prolonged
second stage between the transient and permanent palsy
groups (four of 49 versus three of 49, P ⫽ .41, OR 1.36,
95% CI 0.24, 8.23). Infants with permanent brachial
plexus palsy had higher mean birth weight (4519 ⫾

Table 3. Comparison of Select Neonatal Characteristics

Transient Permanent
(n ⫽ 49) (n ⫽ 49) P, OR (95% CI)
Birth weight (g) 4143 ⫾ 56.5 4519 ⫾ 94.3 ⬍.001
⬍ 3500 3 (6.1) 1 (2)
3501–4000 12 (24.5) 11 (22.5)
4000–4500 26 (53.1) 18 (36.7)
4500–5000 8 (16.3) 8 (16.3)
⬎ 5000 0 (0) 11 (22.5)
Birth weight ⬎ 4500 g 8 (16.3) 19 (38.8) .01, 0.31 (0.11, 0.87)
Birth weight ⬎ 4000 g 34 (69.4) 37 (75.5) .36, 0.74, (0.27, 1.96)
Bone fracture 8 (16.3) 12 (24.5) .32, 0.60 (0.20, 1.81)
Clavicle 7 10
Humerus 1 2
Other nerve injury 0 (0) 0 (0) ⬎.99
Location
Right 31 (63.3) 26 (53.1) .31, 1.52 (0.63, 3.70)
Left 18 (36.7) 23 (46.9)
Abbreviations as in Table 1.
Data are mean ⫾ standard deviation or n (%).

546 Gherman et al Dystocia and Brachial Plexus Palsies OBSTETRICS & GYNECOLOGY
94.3 g versus 4143 ⫾ 56.5 g, P ⫽ .001, Table 3). A large
proportion of cases of brachial plexus palsy in both
groups had a birth weight greater than 4000 g. Although
no differences were found with respect to this birth
weight level, there were more cases with birth weights
greater than 4500 grams in the permanent brachial
plexus palsy group (38.8% versus 16.3%, P ⫽ .01, OR
0.31, 95% CI 0.11, 0.87). The most common comorbid-
ity associated with brachial plexus palsy was clavicular
fracture, which occurred in 14.3% (seven of 49) of the
transient group and 20.4% (ten of 49) of the permanent
group.
DISCUSSION
In this retrospective case-control study, we found that
there was an increased prevalence of diabetes mellitus
among mothers whose infants had transient brachial
plexus injury. Infants with permanent brachial plexus
palsies had a higher mean birth weight and a higher
percentage of cases with birth weight greater than
4500 g. There were, however, no statistically significant
differences between the two groups with respect to mul-
tiple antepartum, intrapartum, and delivery outcomes.
Which cases of transient brachial plexus palsy will re-
solve is an important clinical question that is commonly
posed. We performed a MEDLINE search of the English
language literature between 1980 and 2002 using the
search terms “brachial plexus palsy” and “shoulder dys-
tocia” and found no studies that specifically compared
neonates with transient and permanent brachial plexus
palsy.
Essentially all of the previous studies of brachial
plexus palsy have follow-up limited to the immediate
postpartum period, with the data coming from medical
record searches, database queries, or birth certificate
investigations.12–15 Among 63 infants with Erb palsy,
Ouzounian et al5 found that most of their mothers were
not diabetic (89%), not obese (76%), had normal labor
(91%), and did not have a midpelvic operative delivery
(79%). In addition, 59% of the cases occurred in infants
with a birth weight less than 4500 g and 19% occurred in
infants with a birth weight less than 4000 g. Using
multivariable analysis that included birth weight and
infant sex determination, Wolf et al16 found that risk
factors for 16 cases of nonrecovered brachial plexus
palsy were similar to 56 control cases.
We found it notable that clavicular fracture occurred
in 14.3% and 20.4% of transient and permanent palsies,
respectively. Among studies that carefully evaluated
neonatal shoulder dystocia–associated morbidity, the
rate of clavicular fracture (1.7–9.5%) appears to be half
that of transient brachial plexus palsy.2 Conversely,
VOL. 102, NO. 3, SEPTEMBER 2003
approximately one third of brachial plexus palsies are
associated with concomitant bone fracture, most com-
monly the clavicle (94%).7 The pathogenesis for clavic-
ular fracture associated with brachial plexus palsy is
currently unknown. An excessive amount of pressure
might be placed on the clavicle by the overlying symphy-
sis pubis; the application of suprapubic pressure can also
result in transverse or oblique fractures.17 It is likewise
possible that a complex relationship exists between cla-
vicular fracture, fetal size, and pelvic angle orientation.18
Our study found a surprisingly high percentage (12%)
of permanent brachial plexus palsies in which fundal
pressure had been used. Although it is not entirely clear
whether fundal pressure alone is associated with an
increased risk of permanent neurologic injury, its appli-
cation has been associated with a high incidence of
orthopedic and neurologic damage, lower thoracic spinal
cord injury, and increased intrauterine pressure.19,20
We acknowledge that this study would have been
stronger if it had been designed as a prospective cohort
study rather than as a case-control study. The former
type of research effort would have involved detailed
neonatal follow-up for at least 1 year on a large cohort of
neonates with brachial plexus injury. Many of the details
are lost as these children are followed up in specialty
clinics or move with their families. In addition, the actual
incidence of permanent brachial plexus injury associated
with shoulder dystocia is exceedingly low. For example,
among 250 reported cases of shoulder dystocia only
three infants had persistent palsy after 1 year of pediatric
neurologic follow-up.6 Recent studies have reported that
shoulder dystocia complicates approximately 1% of all
vaginal deliveries, that the rate of transient shoulder
dystocia–related brachial plexus injury is 15%, and that
the rate of permanent neurologic injury is 5%.2,4 Conse-
quently, a researcher would have to gain access to the
data from 1.3 million vaginal deliveries in order to
describe 100 permanent brachial plexus injuries.
Our study had approximately 80% power (95% con-
fidence level) for the outcome measure of an increased
rate of diabetes mellitus among the transiently injured
group. This finding, however, should be interpreted
with caution. We are currently unable to provide the
specific racial breakdown for the cases and controls. We
do acknowledge that the rate of Hispanic patients in the
transient brachial plexus palsy cohort is most likely
higher because this database included patients from Los
Angeles County/University of Southern California
School of Medicine Women’s and Children’s Hospital.
This institution is a large tertiary care referral center with
a primarily indigent Hispanic patient population. More-
over, patients of Hispanic descent are at increased risk
for the development of gestational diabetes. Interest-
Gherman et al Dystocia and Brachial Plexus Palsies 547
ingly, however, that the transient nerve injuries occurred
in smaller infants who were more likely to be from
diabetic mothers. This may perhaps illustrate a lower
threshold for shoulder dystocia in diabetic patients, a
finding that is consistent with anthropometric data.21
Recent studies found that shoulder dystocia itself may
be a causal factor for the neonate’s brachial plexus inju-
ry.4,22,23 During a shoulder dystocia event, there are
significant compressive forces exerted on the fetal neck
by the symphysis pubis. In Gonik et al’s mathematic
model, clinician-applied traction to the fetal head was
estimated to produce 22.9 kPa of pressure between the
fetal neck and symphysis pubis.24 Uterine and maternal
expulsive efforts, however, resulted in contact pressures
that ranged from 91.1 to 202.5 kPa. Because transient
brachial plexus palsy is rather common with shoulder
dystocia but permanent nerve injury is exceedingly rare,
this implies that each fetus may possess a unique thresh-
old for injury. Further studies, including comparison of
neurosurgical findings with obstetric antecedents and
development of a tool to gauge excessive downward
traction, are urgently needed.
REFERENCES
1. American College of Obstetricians and Gynecologists.
Shoulder dystocia. Washington, DC: American College of
Obstetricians and Gynecologists, ACOG Practice Bulletin
#40, November 2002.
2. Gherman RB. Shoulder dystocia: An evidence-based eval-
uation of the obstetric nightmare. Clin Obstet Gynecol
2002;45:345–62.
3. Pollack RN, Buchman AS, Yaffe H, Divon MY. Obstetri-
cal brachial plexus palsy: Pathogenesis, risk factors, and
prevention. Clin Obstet Gynecol 2000;43:236–46.
4. Gherman RB, Ouzounian JG, Goodwin TM. Brachial
plexus palsy: An in utero injury? Am J Obstet Gynecol
1999;180:1303–7.
5. Ouzounian JG, Korst LM, Phelan JP. Permanent Erb’s
palsy: A lack of relationship with obstetrical risk factors.
Am J Perinatol 1998;15:221–3.
6. Gherman RB, Goodwin TM, Souter I, Neumann K,
Ouzounian JG, Paul RH. The McRoberts’ maneuver for
the alleviation of shoulder dystocia: How successful is it?
Am J Obstet Gynecol 1997;176:656–61.
7. Gherman RB, Ouzounian JG, Goodwin TM. Obstetrical
maneuvers for shoulder dystocia and associated fetal mor-
bidity. Am J Obstet Gynecol 1998;178:1126–30.
8. Eng GD, Binder H, Getson P, O’Donnell R. Obstetrical
brachial plexus palsy (OBPP) outcome with conservative
management. Muscle Nerve 1996;19:884–91.
9. Michelow BJ, Clarke HM, Curtis CG, Zuker RM, Seifu Y,
Andrews DF. The natural history of obstetrical brachial
plexus palsy. Plast Reconstr Surg 1994;93:675–80.
548 Gherman et al Dystocia and Brachial Plexus Palsies
10. American College of Obstetricians and Gynecologists.
Operative vaginal delivery. Washington, DC: College of
Obstetricians and Gynecologists, ACOG Practice Bulletin
#17, June 2000.
11. Visintainer PF, Tejani N. Understanding and using confi-
dence intervals in clinical research. J Matern Fetal Med
1998;7:201–6.
12. Nocon JJ, McKenzie DK, Thomas LJ, Hansell RS. Shoul-
der dystocia: An analysis of risks and obstetric maneuvers.
Am J Obstet Gynecol 1993;168:1732–9.
13. Baskett TF, Allen AC. Perinatal implications of shoulder
dystocia. Obstet Gynecol 1995;86:14–7.
14. Gilbert WM, Nesbitt TS, Danielsen B. Associated factors
in 1611 cases of brachial plexus injury. Obstet Gynecol
1999;93:536–40.
15. Gordon M, Rich H, Deutschberger J, Green M. The
immediate and long-term outcome of obstetric birth
trauma. I. Brachial plexus paralysis. Am J Obstet Gynecol
1973;117:51–6.
16. Wolf H, Hoeksma AF, Oei SL, Bleker OP. Obstetric
brachial plexus injury: Risk factors related to recovery.
Eur J Obstet Gynecol Reprod Biol 2000;88:133–8.
17. Oppenheim WL, Davis A, Growdon WA, Dorey FJ,
Davlin LB. Clavicle fracture in the newborn. Clin Ortho-
pedic Res 1990;250:176–80.
18. Gonik B, Allen R, Sorab J. Objective evaluation of shoul-
der dystocia phenomenon: Effect of maternal pelvic orien-
tation on force reduction. Obstet Gynecol 1989;74:44–7.
19. Hankins GDV. Lower thoracic spinal cord injury—A
severe complication of shoulder dystocia. Am J Perinatol
1998;15:443–4.
20. Buhimschi CS, Buhimschi IA, Malinow AM, Kopelman
JN, Weiner CP. The effect of fundal pressure manoeuvre
on intrauterine pressure in the second stage of labour. Br J
Obstet Gynaecol 2002;109:520–6.
21. Modanlou HD, Komatsu G, Dorchester W, Freeman RK,
Bosu SK. Large-for-gestational age neonates: Anthropo-
metric reasons for shoulder dystocia. Obstet Gynecol
1982;60:414–23.
22. Sandmire HF, DeMott RK. Erb’s palsy: Concepts of cau-
sation. Obstet Gynecol 2000;95:940–942.
23. Sandmire HF, DeMott RK. Erb’s palsy causation: A his-
torical perspective. Birth 2002;29:52–4.
24. Gonik B, Walker A, Grimm M. Mathematic modeling of
forces associated with shoulder dystocia: A comparison of
endogenous and exogenous sources. Am J Obstet Gynecol
2000;182:689–91.
Reprints are not available.
Received February 10, 2003. Received in revised form April 15, 2003.
Accepted June 2, 2003.
OBSTETRICS & GYNECOLOGY