British Journal of Obstetrics and Gynaecology

September 1996, Vol. 103, pp. 868-872

Risk factors and fetal outcome in cases of

shoulder dystocia compared with normal
deliveries of a similar birthweight
Ahmed Mohammed Bahar Assistant Professor and Consultant
Department of Obstetrics and Gynaecology, University ofKuwait and Kuwait Maternity Hospital

Objectives To compare risk factors and fetal morbidity in deliveries complicated by shoulder dystocia
with deliveries of similar infant birthweights but not complicated by shoulder dystocia.
Design A retrospectivecase-controlled study.
Setting Kuwait Maternity Hospital.
Participants Sixty-ninecases of true shoulder dystocia and 138 controls matched for exact infant’s birth-
weight.
Methods Demographic data and data regarding history of previous shoulder dystocia, diabetes mellitus,
labour course, method of delivery and newborns’ condition were collected from patients and case
notes following delivery. The mothers’ height and weight were measured. Oral glucose tolerance test
were performed on patients who were not known as diabetics. The infants’ head and chest circum-
ferences and bisacromial diameter were measured.
Results There were no significant differences between cases and controls when mean age, parity, height,
weight and gestational ages were compared. The cases demonstrated a higher incidence of previous
shoulder dystocia (P < O.Ol), diabetes mellitus (P < 0.001), use of oxytocin for acceleration of labour
(P < 0.01) and operative vaginal deliveries (P < 0.01). Differences between cases and controls in
their newborn infants’ head and chest circumferences were not significant,but the newborns of cases
have a longer mean bisacromial diameter and a shorter head circumference:bisacromial diameter
ratio (P < 0.001 and P < 0.001, respectively). Thirty-seven infants (53.6%) from cases and two from
controls (1 -4%)sustained birth injuries. There were two stillbirths among the cases.
Conclusions Although fetal macrosomia is the principal risk factor for shoulder dystocia, other important
risk factors include diabetes mellitus, previous history of shoulder dystocia, prolonged labour, delay
in the second stage of labour and fetal shoulder width which appear to be independent of fetal
weight.

INTRODUCTION assess the effects of factors other than fetal weight

Shoulder dystocia is associated with high fetal mor-
bidity’,*and is mainly due to fetal macro~ornia~,~.The
incidence in unselected vaginal deliveries is variously
reported to range from 0.37% to l.l%275, and it
increases with increasing birthweights6. Reports
suggest an increase in the incidence of fetal macro-
somia worldwide7-I0. Other reported risk factors are
maternal diabetes, obesity, postdatism, multiparity,
protracted labour and disproportionate fetal
Most of these factors are associated with
fetal macrosomia. However, shoulder dystocia also
occurs in the average size infant6J1J2,and not all
macrosomic infants develop shoulder dystocia. To
Correspondence:Dr A. M. Bahar, King Saud University, College of
Medicine, PO Box 64 1, Abha, Saudi Arabia.
868
per se, the present matched controlled study was
carried out in Kuwait Maternity Hospital (Arabian
gulf) where risk factors were compared between
deliveries complicated by shoulder dystocia and
deliveries of infants with similar birthweights not
complicated by shoulder dystocia. The study also
evaluated the extent of fetal morbidity.
METHODS
Of 13,756 singleton vertex deliveries in the period
January 1989 to December 1989, one hundred and
sixty (1* 16%) were coded for shoulder dystocia.
Sixty-nine of these were the subject of this study.
Patients were only included in the study if the follow-
ing conditions were met:
0 RCOG 1996 British Journal of Obstetrics and Gynaecology
 RISK FACTORS A N D FETAL OUTCOME I N SHOULDER DYSTOCIA
1. They developed true shoulder dystocia (defined as
delivery requiring, in addition to downward trac-
tion and episiotomy, manoeuvres to deliver the
shoulder^'^).
2. Matched controls could be found for them.
3. All information was complete and collected
within 24 h of delivery.
In the first 24 h following delivery data were collected
on a special sheet and later entered into a computer.
The data consisted of the woman’s age, parity, gesta-
tional age, history of previous shoulder dystocia,
labour summary and whether the woman was a frank
diabetic or had impaired glucose tolerance as defined
by the WHO interpretation of a 75 g oral glucose
tolerance testI4. All women whose glucose tolerance
status was not known at time of delivery had an oral
glucose tolerance test performed the next day follow-
ing delivery. Height and weight of mothers were
recorded, as pre-pregnancy weight was not available.
The infant’s 5 min Apgar score following delivery
was recorded. If the score was 7 or above, the condi-
tion was coded as good. If the score was 5 to 6 it was
coded as fair. If the score was below 5 , it was coded
as bad. The infant’s weight in grammes was recorded,
and anthropometric measurements of the infant were
made. These consisted of head circumference, chest
circumference and shoulder width; all measurements
were made by the author. The head circumference
was taken as the standard occipitofiontal circum-
ference using a disposable measuring tape read to the
nearest millimetre. Three measurements were carried
out, and the average was taken. The same procedure
was carried out for the chest circumference taken at
the level of the nipples. The shoulder width (bisacro-
mial diameter) was measured using an orthopaedic
anthropometer while the infant was lying on its back
and its arms lying to the sides of the trunk. The inside
edges of the anthropometer’s arms were placed under
the outside edges of the acromial processes, and the
counter was then read to the nearest millimetre. The
mean of three measurements was recorded. Birth in-
juries for the infant were recorded from neonatology
case records.
For each case an immediate daily search for two
controls matched for the infant’s birthweight was
undertaken within the following two weeks. This
short time interval was given so that cases and con-
trols should have as similar an environment at the
time of delivery as possible. The controls were
women who were delivered vaginally but did not
develop shoulder dystocia and were taken consecu-
tively. The same procedures regarding data collection
and measurements were carried out for controls.
There was difficulty in finding controls for 21 of the
0 RCOG 1996 Br J Obstet Gynaecol 103, 868-872
869
Table 1. Maternal characteristics. All statistics performed using
xz
Student’s t test except gestational age, determined by test. P not
significant for all characteristics.BMI = body mass index.
Characteristics Cases (n = 69) Controls (n = 138)
Age (years): mean (SD) 29.7 (5.1) 28.6 (4.8)
Parity:mean (SD) 3 (2.4) 2.9 (2.3)
Weight (kg): mean (SD) 75.5 (12) 77.5 (13.3)
Height (cm): mean (SD) 162.3 (5.9) 163 (4.8)
BMI: mean (SD)* 28.7 (4.4) 29 (4.6)
Gestational age (weeks): n [%]
37 2 [2.9] 4 [2.9]
3741 52 [73-91 100 [72.5]
> 41 15 [21.7] 33 [23.9]
*Weight in kg I height in cm2.
very large infants (weight > 4.5 kg) because most of
these were delivered by caesarean section. Failure to
collect information soon after delivery occurred in 29
cases; controls with an exact match for birthweight
could not be found within the allotted two weeks for
41 cases. At the end there were 69 cases and 138
controls which satisfied the criteria for inclusion in
the study.
Statistical analysis was camed out using the
Statistical Package for Social Science (SPSS version
6.1). Student’s t test was used to analyse continuous
variables and the x2 test was used for discrete
variables. The level of significance was set to 0.05.
RESULTS
There were no significant differences in the mean
age, parity, weight, height, body mass index and ges-
tational ages between cases and controls (Table 1).
Table 2 shows the incidence of previous shoulder
dystocia, diabetes, use of oxytocin for acceleration of
labour, operative vaginal deliveries and past dates.
History of previous shoulder dystocia was six times
more in cases than in controls (P< 0.01, OR 6.75;
95% CI 1.76-25.8). The incidence of diabetes was
also significantly more in cases than in controls
(P < 0.001, OR 4.3; 95% CI 2.2-8*3), as was the
incidence of those with impaired glucose tolerance
(gestational diabetes). There was a significant differ-
ence in the use of oxytocin for the acceleration of the
first stage of labour and in the incidence of operative
vaginal deliveries (P < 0-01, OR 3.52; 95% CI
1.32-9-37, and P < 0.01, OR 3.4; 95% CI 1.58-7.32,
respectively). All the operative vaginal deliveries
were by vacuum extractor.
The infants’ mean birthweight was 4-25 (SD
0.34) kg (range 3 . 4 4 9 kg) and the incidence of
infants weighing 4 kg or more was 855%. Table 3
shows the distribution of infants’ birthweights. There
870 A. M. BAHAR

Table 2. Number (percent) of previous shoulder dystocia, diabetes mellitus, acceleration of labour and operative vaginal deliveriesin cases and
controls.

Variables Cases ( n = 69) Controls ( n = 138) P Odds ratio 95% CI

History of previous shoulder dystocia 9 (13.0) 3 (2.2) < 0.01 6.75 1.76-25.8
All diabetics 31 (44.9) 22 (1 5.9) < 0.001 4.3 2.2-8.3
Impaired glucose tolerance * 13/51 (25.5) 14/130 (10.8) < 0.05 2.8 1.2-6.56
Frank diabetic 18 (26.1) 8 (5.8) < 0-001 5.7 2,3614
Use of oxytocin for acceralation of labour ** 11/52(21.2) 8/113 (7.1) < 0.01 3.52 1.32-937
Operativevaginal deliveries 19 (27.9) 14 (10.2) < 0.01 3.4 1.58-7.32

*Frank diabetics are excluded.

**Womenwhose labour was induced are excluded.

was no difference in the distribution of infants' sex Table 3. Distribution of infants according to birthweights.Values are
given as n (%).
between cases and controls. The fetal biometry is ~~ ~

shown in Table 4. There was no difference in the mea-
surements of head and chest circumferences between
cases and controls; but a significant difference was
found in the measurement of the bisacromial diame-
ter and in the ratio between head circumference and
bisacromial diameter (P < 0.001, mean difference
0 5 5 [95% CI 0.264841 and P < 0.001, mean differ-
ence -0.08 [%% CI -0.126 to 0.0381, respectively).
Table 5 shows the immediate general condition of
the newborns. The Apgar scores were less than seven
in 52% of cases and in only 8.7% of controls
(P< 0.001, OR 11.45; 95% CI 5.07-26.34). Fetal
injuries are shown in Table 6. Among the cases, 39
infants (56.5%) sustained injuries of one form or
another. Only two infants from controls (1.4%) sus-
tained injuries in the form of fracture of the clavicle.
The weights of these two infants were 3.9 kg. and
4.4 kg, respectively. Both were born spontaneously
by the vertex, and there was no record of difficulties
during delivery, although in one infant the general
condition at delivery was recorded as bad. No long
term follow up for the infants was carried out. The
perinatal mortality was 29AOOO. Maternal injuries
included one case with a cervical tear and two
controls with vaginal tears.
DISCUSSION
Due to the high incidence of shoulder dystocia in this
large maternity unit with an annual delivery rate of
over 16,000 it was possible to complete the study in
one year. This is the first controlled study in Kuwait
to address this problem. To get as accurate infor-
mation as possible and reduce bias in collecting data
from hospital case notes, all information was
collected soon after delivery. This short interval was
also chosen for consistency in fetal measurements,
maternal weights and blood sugars. Unlike previous
studies, the effect of birthweight was controlled in
order to nullify its effect when comparing the
Birthweight (g) Cases Controls
(n = 69) (n = 138)
3000-3999 10 (14.5) 20 (14.5)
400M499 41 (59.5) 82 (59-5)
2 4500 18 (26.0) 36 (26.0)
incidences of other risk factors. These strict criteria
imposed by the study design contributed to the reduc-
tion in the number of cases enrolled for the study.
However, it was the object of this study to look for
factors other than fetal macrosomia since shoulder
dystocia can occur in an infant of average weight. In
the present study 145% of infants had birthweights
of less than 4 kg. Since there were no differences in
parity, obesity and postdatism between cases and
controls in this study, it seems that these factors are
not primary in the aetiology of shoulder dystocia.
However, they may operate via their positive influ-
ence on birthweight.
Previous history of shoulder dystocia emerged as a
highly significant factor. This is mentioned in the lit-
erature; one paper which quantified this risk is that of
Smith et U Z . ' ~ in which they reported a recurrence rate
of 9.8%, compared with a recurrence rate of 0.58% in
the total population (RR= 16.8, 95% CI 7.23-39-1).
In the present study two women had previously had
shoulder dystocia twice and in both instances and in
the index pregnancies the infants' weights were
below 4 kg, indicating that this factor may operate
independently of fetal macrosomia.
Although the study showed a high incidence of
diabetes in both cases and controls (44.9% and 15*9%,
respectively), the incidence in cases was nearly three
times that in controls, which is statistically highly
significant. This confirms previous studies that
maternal diabetes is an important risk fa~tor'-~J.It
has been reported that infants of diabetic mothers
have disproportionate growth between head and
trunk7J6J7. Performing postnatal fetal biometry,

0 RCOG 1996 Br J Obstet Gynaecol 103, 868-872

 RISK FACTORS AND FETAL OUTCOME I N SHOULDER DYSTOCIA 871

Table 4. Fetal biometry. Values are given as mean (SD).

Measurement(cm) Cases Controls P Difference

(n = 69) (n = 138) Mean (95%CI)
~~~~~~ ~ ~ ~

Head circumference 35.96 (1.06) 35.9 (1.09) NS

Chest circumference 38.1 (1.3) 38 (1.1) NS
Bisacromial diameter 15.16 (1.12) 14.61 (0.94) < 0.001 0.55 (0261 to 0.84)
Head circumferenceibisacromialdiameter 2.38 (0.17) 2.46 (0-14) < 0.00 1 0.08 (0.12 to -0.038)

Table 5. Immediate condition of the newborn as assessed by the 5 Table 6. Fetal injuries. Values are shown as n (%). P < 0-001.
min Apgar scores. 1 versus (2&3): '
x = 48.8, P 5 0.001; OR = 11.45;
95% CI = 5-07to 26.34. Values are given as n ("36). Injuries Cases Controls
(n = 69) (n = 138)
Condition Cases Controls
(n = 69) (n = 138) Fracture
~~~~~~~~~~

Clavicle 2 (2.9) 2(1.4)

1. Good (Apgar score ? 7) 33 (48)
2. Fair (Apgar scores 5 4 ) 18 (26)
3. Bad (Apgar scores < 5 ) 18 (26)
Madanlou et ~ 1found . ~ that infants of diabetic moth-
ers had significantly greater shoulder-head and
chest-head size differences than did infants of nondi-
abetic mothers of comparable birthweight. Elliot et
~ 1 . carried
'~ out ultrasound fetal measurements in dia-
betic patients within 21 days of delivery and found a
chest-biparietal diameter of 1-4cm or greater in 20/23
infants (87%) weighing greater than 4000 g and only
1.3 cm or less in 34/37 infants (92%) weighing less
than 4000 g. This suggests that disproportionate
growth increased with macrosomia. In this study a
number of large babies with shoulder dystocia
excluded due to a lack of controls. However, there
were no significant differences in head and chest
circumference between cases and controls, although
there was a significant difference in bisacromial
diameter and head circumference:bisacromial dia-
meter ratio. This was also found by Modanlou et uL7
in their comparison of 10 newborn infants who devel-
oped shoulder dystocia with 130 newborn infants of
nearly similar birthweight who did not develop
shoulder dystocia. The finding of greater mean shoul-
der width and greater head-shoulder disproportion in
the cases in this study may be due to the fact that
there were significantly more diabetics among cases
than among controls. Head-shoulder disproportion
may be instrumental in the development of shoulder
dystocia. These measurements will only be useful in
the prediction of shoulder dystocia if they are
performed antenatally on the fetus. Ultrasound mea-
surement can be performed for the fetal head and
chest but is not feasible for the shoulders. Fetal
weight estimation by ultrasound is also possible, but
not accurate, especially when the fetus is large18J9.
Since the shoulders are the first to get stuck when
126 (91.3) Humerus 2 (2.9)
11 (8) Brachial palsy
l(0.7) Mild 14 (20.3)
Moderate 12 (17.4)
Severe 4 (5.8)
Brachial palsy and
hypoxic encephalopathy 2 (2-9)
Hypoxic encephalopathy l(1.4)
Stillbirth 2 (2.9)
TOTAL, 39 (56.5) 2 (1.4)
shoulder dystocia occurs, it is logical to assume that
measurement of shoulder width will be a better indi-
cator for the possible development of shoulder dysto-
cia than the measurement of chest circumference.
Measurement of fetal shoulder width using computed
tomography appears promising20,21. Recently Kastler
et uLzl were able to measure fetal shoulder width
using MRI. They found that shoulder width measure-
ment by MRI correlated significantly with postnatal
orthopaedic calliper measurements (r = 0.955, P =
0-00001)and with birthweight (r = 0.63, P = 0.0005).
Because of its excellent soft tissue contrast, MRI
offers the potential for measuring both fetal shoulder
width and maternal pelvic dimensions. This may be a
useful research tool for studies designed to investi-
gate this problem. Results of such studies may
improve our ability to predict shoulder dystocia.
Prolonged labour, oxytocin augmentation, pro-
longed second stage and operative vaginal delivery
have been documented as risk f a ~ t o r s ' , ~ ~In~the
J'~~~.
present study both the incidence of use of oxytocin
for acceleration of labour and the higher incidence of
operative vaginal delivery were significant in cases,
compared with controls. No analysis was done for the
actual length of first and second stages of labour
because the nurses' notes were not accurate in this
respect. Nevertheless, both the use of oxytocin and
operative vaginal delivery give an indirect reflection
as to prolongation of first and second stage of labour,
respectively. Practically all cases in this study

0 RCOG 1996 Br J Obstet Gynaecol 103, 868-872

872 A . M. B A H A R
required an operative vaginal delivery for delay in the
second stage of labour. Since both cases and controls
had the same birthweights, these factors should be
looked at independently of fetal weight in the intra-
partum prediction of shoulder dystocia.
The extent of fetal injury in the study compares
well with other s t ~ d i e s ’ >
and
~ >attests
~ ~ to the serious-
ness of the problem. Manoeuvres used to release the
shoulders, such as the McRoberts manoeuvre, the
corkscrew manoeuvre and delivery of posterior
shoulder, are attained with less morbidity than fundal
pressure and traction24, but all these manoeuvres
require experience. When shoulder dystocia occurs,
the patient is usually attended by a nurse or a junior
doctor, and in most cases there will be delay before
senior staff arrives. Therefore a high index of sus-
picion and prediction is essential if morbidity is to be
reduced.
Acknowledgement
I would like to thank the consultants of Kuwait
Maternity Hospital who allowed me to study their
patients. My greatest appreciation to Dr A. AbdAziz
for her valuable assistance in the collection of infor-
mation from patients.
References
Benedetti TJ, Gabbe SG. Shoulder dystocia: a complication of fetal
macrosomia and prolonged second stage of labour with midpelvic
delivery. Obstet Gynecoll978; 52: 526529.
Gross SJ, Shime J, Farine D. Shoulder dystocia: predictors and out-
come. Am JObstet GynecoZl987; 156: 334-336.
Spellacy WN, Miller S, Winegar A, Peterson PQ. Macrosomia-
maternal characteristics and infant complications. Obstet Gynecol
1985; 66: 158-161.
Walker D. The dangers of fetal macrosomia.J Obstet Gynaecoll991;
11: 336-338.
Hopwood HG, Jr. Shoulder dystocia: fifteen years’ experience in a
community hospital.Am JUbstet Gynecoll982; 144: 162-166.
Acker DB, Sachs BP, Friedman EA. Risk factors for shoulder dysto-
cia in the average-weightinfant. Obstet Gynecoll986; 67: 614-618.
7 Modanlou HD, Komatsu G, Dorchester W, Freeman RK, Bosu SK.
Large-for-gestational-age neonates: anthropometric reasons for
shoulder dystocia. Obstet Gynecoll982; 60: 417423.
8 Johar R, Raybum W, Weir D, Eggert L. Birth weights in term infants:
a 50-year perspective. JReprodMed 1988; 33: 813-816.
9 Baker PN, Lever P, Gorton E. An increasing incidence of fetal macro-
somia. Jobstet Gynaecoll992; 12: 281
10 Fan ZF X Z , Xin W, Xiou ZY, Xin W, Wei CY, Lee RV. Neonatal
macrosomia and the obstetric complicationsof macrosomia as mark-
ers of socio-economic change in China: a retrospective study in one
hospital. JObstet Gynaecoll993; 13: 163-166.
11 Keller JD, Lopez-Zeno JA, Dooley SL, Socol ML. Shoulder dystocia
and birth trauma in gestational diabetes: a five-year experience. Am J
Obstet Gynecoll991; 165: 928-930.
12 Momson JC, Sanders JR,Magann EF, Wiser WL. The diagnosis and
management of dystocia of the shoulder. Surg Gynecol Obstet 1992;
175: 515-522.
13 Resnik R. Management of shoulder girdle dystocia in vertex delivery.
Clin Obstet Gynecoll980; 23: 559-564.
14 WHO Expert Committee on Diabetes Mellitus. Second Report. World
Health Organ Tech Rep Ser 1980; 646: 1&11.
15 Smith RB, Lane C, Pearson JF. Shoulder dystocia: what happens at
the next delivery?Br JObstet Gynaecoll994; 101: 713-715.
16 Elliott JP, Garite TJ, Freeman FX,McQuown DS, Pate1 JM.
Ultrasonic prediction of fetal macrosomia in diabetic patients. Ubstet
Gynecoll982; 6 0 159-162.
17 Sacks DA. Fetal macrosomia and gestational diabetes: what’s the
problem? Obstet Gynecoll993; 81: 775-781.
18 Delpapa EH, Mueller-Heubach E. Pregnancy outcome following
ultrasound diagnosis of macrosomia. Obstet Gynecol 1991; 78:
340-343.
19 Sandmiire HE Whither ultrasonic prediction of fetal macrosomia?
Obstet Gynecoll993; 82: 86s862.
20 Kitzmiller J, Mall J, Gin G, Hendricks S, Newman R, Scheerer L.
Measurement of fetal shoulder width with computed tomography in
diabetic women. Obstet Gynecoll987; 70: 941-945.
21 Kastler B, Gangi A, Mathelin C et al. Fetal shoulder measurements
with MRI. JComput Assist Tomogr 1993; 17: 777-780.
22 If€y L, Varadi V, JakobovitsA. Common intrap- denominatorsof
shoulder dystocia related birth injuries. Zentralbl Gynakol 1994; 116:
”,.
11-17
23 B~~~~~ B, ~ ~ ~ ~S, Mohammed
~ ~ N, ~ H. i
i ~ ~ l A,i ~persad ~
Shoulder dystocia: an obstetrical nightmare. West Indian Med J 1992;
41: 15&159.
24 Roberts L, Shoulder dystocia, In: Studd mw, editor. progress in
Obstetrics and Gynaecology. Edinburgh: Churchill Livingstone,
1994 201-216.
Received 21 J u I995
~
Accepted 6 February 1996
0 RCOG 1996 Br J Obstet Gynaecol 103, 868-872