Risk factors for recurrent shoulder dystocia, Washington state, 1987-2004

Results 

There were 1,378,099 singleton births in the state of Washington from Jan. 1, 1987, to Dec. 31, 2004; and of those, 1,126,593 were vaginal deliveries. Shoulder dystocia complicated 26,208 of the vaginal deliveries, resulting in a primary shoulder dystocia annual incidence rate of 2.3 per 100 singleton vaginal deliveries. Among the 26,208 women with vaginal deliveries with a primary shoulder dystocia, 8991 had a subsequent delivery of which 1060 women (11.8%) had 2 births complicated by shoulder dystocia, 6759 women had 1 birth complicated by shoulder dystocia without recurrence in a subsequent vaginal delivery (75.2%), and 1172 had cesarean deliveries (13.0%). The overall annual incidence of shoulder dystocia recurrence for the 18 years studied was 13.5 of 100 subsequent vaginal deliveries, with a temporal trend toward increasing annual incidence from 1987 to 2004 (Figure 2). The incidence of severe shoulder dystocia was 0.4 per 100 vaginal births in the index pregnancy and 1.2 per 100 vaginal births in the subsequent vaginal delivery.

View Large Image Download to PowerPoint

FIGURE 2. The annual incidence of primary and recurrent shoulder dystocia at vaginal delivery in Washington state The annual incidence of primary and recurrent shoulder dystocia at vaginal delivery, Washington state, 1987-2004. x axis, birth years; y axis, incidence per 1000 vaginal births. Moore. Risk factors for recurrent shoulder dystocia, Washington state, 1987-2004. Am J Obstet Gynecol 2008.

Among the 1172 women who had cesarean deliveries following their index shoulder dystocia, 153 (13%) had repeat cesarean, no labor; 60 (5%) repeat cesarean with trial of labor; 210 (18%) primary cesarean for placenta previa, breech, cord prolapse, or genital herpes; 82 (7%) had a trial of labor with fetal distress; 176 (18%) had failed induction of labor; and 39% had cesarean deliveries for other indications. Maternal characteristics at the time of index shoulder dystocia of those women who had subsequent cesarean and vaginal deliveries were compared (Table 1). Women who had subsequent cesarean deliveries were more likely to be older, to have had a premature delivery, and to have had gestational diabetes, compared with women who had subsequent vaginal deliveries (P ≤ .05). Women who had cesarean deliveries were more likely to have had gestational diabetes in the index pregnancy (6.8%) than women who had subsequent vaginal deliveries (3.9%). Furthermore, women who had subsequent cesarean deliveries were twice as likely to have had newborns with a severe shoulder dystocia in the index pregnancy, compared with women with a subsequent vaginal delivery, 17.8% vs 8.1% (Table 1).

We compared maternal demographic characteristics and obstetric history at the time of subsequent vaginal delivery between the cases (n = 1060) and controls (n = 4238) (Table 2). We found the cases to be significantly older with greater gravidity (data not shown) and parity at the subsequent delivery, compared with controls (P ≤ .05). Maternal race, marital status, mother's education, maternal smoking, and gestational age at delivery were not significantly different between cases and controls.

TABLE 2. Characteristics at the time of subsequent vaginal delivery, among women with shoulder dystocia in the index delivery with and without recurrence in a subsequent delivery, Washington state 1987-2004

		Cases (N = 1,060) na (%)	Controls (N = 4,238) na (%)
	Maternal Age (years)b
	<20	21(2.0)	158(3.7)
	20-24	206(19.4)	979(23.1)
	25-29	351(33.1)	1319(31.1)
	30-34	321(30.3)	1152(27.2)
	35+	161(15.2)	640(14.9)
	Maternal Race
	White	800(75.5)	3232(76.3)
	Black	22(2.1)	120(2.8)
	Hispanic	120(11.3)	443(10.5)
	Asian/Pacific Islander	51(4.8)	222(5.2)
	Native American	25(2.4)	79(1.9)
	Marital Status
	Married	852(80.4)	3369(79.5)
	Unmarried	207(19.5)	866(20.4)
	Mother's Education (years)
	<12	178(16.8)	640(15.1)
	12	276(26.0)	1115(26.3)
	>12	606(57.2)	2483(59.6)
	Parityb
	1	450(42.4)	2691(63.5)
	2	345(32.6)	858(20.3)
	3	131(12.4)	341(8.1)
	4	49(4.6)	125(3.0)
	5+	76(7.2)	170(4.0)
	Gestation Age at Delivery (weeks)
	≤37	44(4.2)	250(5.9)
	38-40	812(76.6)	3228(76.2)
	41+	199(18.8)	743(17.5)

Moore. Risk factors for recurrent shoulder dystocia, Washington state, 1987-2004. Am J Obstet Gynecol 2008.

a Numbers may not sum to total due to missing data. b Significant difference between groups, p ≤ 0.05.

We evaluated the association between specific risk factors in the index pregnancy and recurrent shoulder dystocia (Table 3). Birthweight in the index pregnancy was associated with an increased risk of subsequent shoulder dystocia with a significant trend of increasing magnitude of risk with increasing birthweight in the index pregnancy (P < .001). Fracture of the newborn clavicle, presence of brachial plexus injury, and Apgar score of 5 or less at 5 minutes were all associated with an increased risk of subsequent vaginal birth with recurrent shoulder dystocia. We did not find a significant association between gestational age, gestational diabetes, operative vaginal delivery, induction of labor, maternal BMI, or maternal weight gain in the index pregnancy and shoulder dystocia recurrence. Vacuum delivery and induction of labor approached significance.

TABLE 3. Risk factors of recurrent shoulder dystocia among women with a shoulder dystocia in a previous delivery with and without recurrence in a subsequent delivery, Washington state 1987-2004

		Cases N = 1,060 na	Controls N = 4,238 na	Odds ratio (95% confidence interval)
	Risk factors at the index delivery
	Birth weight (grams)b,c
	<2500	0	67	-
	2500-2999	15	294	0.4(0.2,0.7)
	3000-3499	145	1166	1.0
	3500-3999	413	1682	2.0(1.6,2.4)
	4000-4499	373	805	3.7(3.0,4.6)
	4500-4999	94	197	3.8(2.8,5.2)
	5000+	20	27	5.9(3.2,11.0)
	Gestational age (weeks)
	≤37	41	277	0.6(0.4,0.8)
	38-40	703	2630	1.0
	41+	214	923	0.9(0.7,1.0)
	Gestational diabetes
	No	1006	4043	1.0
	Yes	49	177	1.1(0.8,1.5)
	Operative delivery
	No operative assistance	722	2995	1.0
	Forceps	74	341	0.9(0.7,1.2)
	Vacuum	264	897	1.2(1.0,1.4)
	Both	0	5	-
	Induction of labor
	No	712	2950	1.0
	Yes	303	1094	1.2(1.0,1.3)
	BMI (kg/m2)
	<18.5	11	77	0.6(0.3,1.1)
	18.5-24.9	290	1148	1.0
	25-29.9	114	483	0.9(0.7,1.2)
	30-34.9	52	180	1.1(0.8,1.6)
	35-39.9	14	67	0.8(0.5,1.5)
	40+	7	29	1.0(0.4,2.2)
	Weight gain (pounds)
	<25	120	560	0.9(0.7,1.1)
	25-35	263	1068	1.0
	36-50	254	1004	1.0(0.8,1.2)
	51+	423	1604	1.1(0.9,1.3)
	Risk factors at the index delivery
	Severity of shoulder dystociab
	Fracture of clavicle
	No	1009	4134	1.0
	Yes	51	104	2.0(1.4,2.8)
	Brachial plexus injury
	No	1015	4185	1.0
	Yes	45	53	3.5(2.3,5.2)
	Apgar ≤5 at 5 minutes
	No	1039	4189	1.0
	Yes	21	41	2.1(1.2,3.5)
	Any severe shoulder dystociad
	No	947	4046	1.0
	Yes	113	192	2.5(2.0,3.2)
	Risk factors at the subsequent delivery
	Gestational age (weeks)
	≤37	44	250	0.7(0.5,1.0)
	38-40	812	3228	1.0
	41+	199	743	1.1(1.0,1.3)
	Gestational diabetesb
	No	898	3685	1.0
	Yes	94	253	1.5(1.2,2.0)
	Operative delivery
	None	938	3693	1.0
	Forceps	15	57	1.0(0.6,1.8)
	Vacuum	107	483	0.9(0.7,1.1)
	Both	0	5
	Induction of labor
	No	629	2683	1.0
	Yes	431	1555	1.2(1.0,1.4)
	BMI (kg/m2)
	<18.5	12	71	0.7(0.4,1.3)
	18.5-25.0	310	1284	1.0
	25.1-30.0	190	702	1.1(0.9,1.4)
	30.1-35.0	102	386	1.1(0.9,1.4)
	35.1-40.0	47	157	1.2(0.9,1.8)
	≥41.1	35	159	0.9(0.6,1.3)
	Risk factors at the subsequent delivery
	Weight gain (pounds)
	<25	197	849	0.9(0,1.2)
	25-29	113	450	1.0
	30-39	255	1038	1.0(0.8,1.3)
	40-49	157	538	1.1(0.8,1.4)
	50+	91	336	1.1(0.8,1.5)

Models adjusted for age, parity and birth year. Moore. Risk factors for recurrent shoulder dystocia, Washington state, 1987-2004. Am J Obstet Gynecol 2008.

a Numbers may not sum to total due to missing data. b Significant difference between groups, p ≤ 0.05. c χ2 test for trend p < 0.0001. d Presence of fracture of clavicle, brachial plexus injury, and/or Apgar score ≤5 at five minutes.

We also evaluated the association between specific risk factors in the subsequent pregnancy, which are known prior to delivery, and the risk of recurrent shoulder dystocia (Table 3). A significantly increased risk of shoulder dystocia recurrence was associated with gestational diabetes and induction of labor. Women with gestational diabetes were more likely to be induced than women without gestational diabetes (53% vs 38%), but the risk of shoulder dystocia associated with induction of labor did not vary significantly between women with and without gestational diabetes. We did not find significant relationships with gestational age, prepregnancy BMI, maternal weight gain, forceps or vacuum use in the subsequent delivery and shoulder dystocia recurrence.

We included in our multivariable logistic regression model all the variables that were significant in the univariable analysis, adjusting for maternal age and parity. Increased birthweight, 3500 g or greater in the index pregnancy, remained associated with risk of recurrence (Figure 3). Vacuum delivery in the index pregnancy was associated with a 40% increased risk of recurrence. Forceps use in the index pregnancy, however, was not associated with recurrence. A severe dystocia in the index pregnancy as measured by the composite severity variable was associated with a 2-fold increased risk of shoulder dystocia. Gestational diabetes and induction of labor in the subsequent pregnancy were not significantly associated with repeat shoulder dystocia.

View Large Image Download to PowerPoint

FIGURE 3. Risk factors of recurrent shoulder dystocia Risk factors of recurrent shoulder dystocia among women with a shoulder dystocia in a previous delivery, with and without recurrence in a subsequent delivery, Washington state, 1987-2004. Multivariable model is adjusted for all variables listed (birthweight at index pregnancy, operative vaginal delivery at index pregnancy, and severe shoulder dystocia at index pregnancy) as well as maternal age, parity, and year of birth. Moore. Risk factors for recurrent shoulder dystocia, Washington state, 1987-2004. Am J Obstet Gynecol 2008.

Lastly, because of the importance of brachial plexus injury, we evaluated whether brachial plexus injury in the index shoulder dystocia pregnancy predicted risk for recurrent shoulder dystocia. We found that 4.2% (45 of 1060) of women who had a recurrent shoulder dystocia delivery had a brachial plexus injury at the index shoulder dystocia delivery and that only 1.3% (53 of 4238) of the women without a recurrent shoulder dystocia delivery had a brachial plexus injury at the time of the index shoulder dystocia delivery. Brachial plexus injury, as an independent risk factor for recurrent shoulder dystocia, increased the risk for recurrent shoulder dystocia 2- to 3-fold, odds ratio of 2.6 (95% confidence interval, 1.7 to 4.0) adjusted for age, parity, year of birth, Apgar less than 5 at 5 minutes, clavicle fracture, birthweight, operative delivery.

Comment 

With 26,208 vaginal deliveries complicated by primary shoulder dystocia, to our knowledge, our study is the largest to date to evaluate factors that might predict recurrent shoulder dystocia in women who attempt a subsequent vaginal delivery. We found an increasing magnitude of risk with increasing birthweight of 3500 g or greater in the index pregnancy. In addition, vacuum delivery and a severe shoulder dystocia in the index pregnancy were associated with increased risk. We found no significant association between gestational diabetes or operative delivery in the subsequent pregnancy with risk of recurrent shoulder dystocia. These findings should be interpreted with the knowledge that women with GDM or those who had a severe shoulder dystocia were twice as likely to deliver subsequent pregnancies by cesarean birth. Therefore, our study and any study of this nature are limited by these inherent biases.

Others have found a lower annual incidence rate of primary shoulder dystocia than the 2.3 per 100 singleton vaginal deliveries that we observed.1, 2, 3, 4, 5, 6, 7, 8, 11 However, the overall recurrence rate of shoulder dystocia in subsequent vaginal births of 13.5 per 100 vaginal deliveries per year found in our study was similar to that observed previously, 11.9 to 16.7 per 100 births.11, 29, 30, 31 The trend toward an increasing risk of recurrence over time (1987-2004) has not been measured in other studies.

The 3 other published studies to specifically examine the risk of recurrent shoulder dystocia in subsequent vaginal births supported our finding of an increased risk of dystocia with increasing birthweight.29, 30, 31 In the only other published population-based study, among 203 women with primary shoulder dystocia deliveries, 42 women had subsequent vaginal deliveries and 5 were complicated by a recurrent shoulder dystocia. Increased birthweight in the index pregnancy increased the risk of recurrence, but this was not statistically significant, most likely because of a small sample size.29

Two other studies from academic tertiary care centers reported risk of shoulder dystocia recurrence. At the Louisiana State University Medical Center, 747 women with primary shoulder dystocia deliveries were identified and 17 of 123 (13.8%) had subsequent vaginal deliveries complicated by shoulder dystocia with a significant association with fetal macrosomia.30 A study at Northwestern Medical Center reported 602 primary shoulder dystocia deliveries and 11 of 66 (16.7%) subsequent vaginal deliveries were complicated by shoulder dystocia. Birthweight in the index pregnancy was significantly associated with recurrence.31 Although these studies support an increased risk with macrosomic infants, ours was the only study to suggest a clear delineation of increased risk for recurrent shoulder dystocia at 3500 g or greater.

The association that we observed between vacuum delivery in the index delivery and shoulder dystocia recurrence is similar to that found in studies of primary shoulder dystocia34; however, studies of recurrent shoulder dystocia have not examined this risk factor. This finding could have various interpretations, but vacuum delivery may be a marker of increased risk of cephalopelvic disproportion that could predispose to subsequent cephalopelvic disproportion in subsequent birth with associated shoulder dystocia delivery. However, we did not find an increased risk with forceps delivery.

Severity of shoulder dystocia and risk of recurrence of dystocia in a subsequent vaginal delivery has not been previously described. We were surprised that our data set contained a significant number of women who elected to proceed with a vaginal delivery following a delivery complicated by a severe shoulder dystocia. The presence of this group of high-risk women may be explained in several ways. It could be that clinicians were unaware of a previous shoulder dystocia if care was delivered at another institution. Or it is possible that clinicians recognize, as outlined by the American College of Obstetricians and Gynecologists guidelines8, that strong scientific evidence is lacking to support universal cesarean delivery in a subsequent pregnancy following a severe shoulder dystocia. Our study also assessed the relationship between shoulder dystocia and brachial plexus injury and found that brachial plexus injury, as an independent risk factor for recurrent shoulder dystocia, increased the risk for recurrent shoulder dystocia 2- to 3-fold.

We evaluated maternal BMI and weight gain and did not observe a significant association with maternal obesity or weight gain in either the index or subsequent pregnancies. This finding was in contrast to that of Lewis et al,30 who found that recurrence was significantly associated with maternal obesity, high maternal weight gain in pregnancy, and prolonged second stage of labor in the subsequent pregnancy. Our findings could be attributed to missing data on prepregnancy weight and weight gain in pregnancy, but only if the cases with missing data were more likely to be obese and to have a higher weight gain during pregnancy as compared with the controls who had missing data on these parameters. This reporting bias is possible but unlikely. More probable is a true lack of association because we described 1060 women from a population-based registry, whereas Lewis et al30 described 17 women from a tertiary care center.

The strength of our study is that it is a large, population-based study of a rare occurrence, namely recurrent shoulder dystocia. The selection of risk factors was confined to those known to be reliably reported in the Washington state birth record combined with the hospital discharge codes.35, 36 Our study did have some limitations similar to the 3 other published studies.29, 30, 31 All 4 studies were observational and were able to evaluate the risk of a recurrent shoulder dystocia delivery only among those women with a subsequent pregnancy delivered vaginally. All studies suffered from a potential loss to follow-up, but there is no reason to believe that in our study, the risk factors for women who were lost to follow-up differed from those who were not lost to follow-up. As noted in previous studies, the diagnosis of shoulder dystocia is affected by the degree to which the shoulder dystocia was documented by the clinician and ascertained by ICD-9 coding. Lastly, Washington state has fewer black and Hispanic women, compared with the US population.37 Therefore, our results may not be generalizable to all populations in the United States.

In summary, to the best of our knowledge, this is the largest population-based study to evaluate recurrent shoulder dystocia. We observed that birthweight of 3500 g or greater, a severe shoulder dystocia, and the use of vacuum extraction in the index pregnancy were strong independent risk factors for recurrence of shoulder dystocia. These risk factors exist despite an inherent clinical selection of high-risk women for subsequent cesarean delivery. Shoulder dystocia is a serious concern to obstetricians because of its unpredictability and high risk of morbidity and mortality.

In the context of increasing primary shoulder dystocia incidence, there is an urgent need for establishing evidence on which to base clinical decision making for women with a history of shoulder dystocia desiring a subsequent vaginal delivery. Further investigation on the risks of interventions such as induction and augmentation of labor is warranted. Until additional large, population-based studies are performed to confirm our findings, it is not unreasonable for women to choose vaginal delivery after a mild to moderate shoulder dystocia in a spontaneous vaginal delivery of an infant weighing less than 3500 g.

Given our findings and in the absence of other informative studies, women with a previous delivery complicated by shoulder dystocia with an infant weighing 3500 g or greater or with a previous delivery complicated by a severe shoulder dystocia should carefully weigh the risks and benefits of a subsequent vaginal delivery vs cesarean delivery with their provider. Because the women in our study who had gestational diabetes and a previous delivery complicated by shoulder dystocia were likely to subsequently deliver by cesarean section, our finding of no increased risk of shoulder dystocia in subsequent vaginal deliveries among women with GDM, as compared with women without GDM, should be interpreted with caution.