Journal of Perinatology (2009) 29, 96–100 r 2009 Nature Publishing Group All rights reserved. 0743-8346/09 $32 www.nature.com/jp ORIGINAL ARTICLE Cesarean risk after successful external cephalic version: a matched, retrospective analysis

C Clock1, J Kurtzman2,3, J White3 and JH Chung3 1Department of and Gynecology, Oregon Health Sciences University, Portland, OR, USA; 2Department of Obstetrics and Gynecology, Saddleback Memorial Medical Center, Laguna Hills, CA, USA and 3Department of Obstetrics and Gynecology, Irvine Medical Center, University of California, Orange, CA, USA

has been associated with a higher neonatal morbidity than elective Objective: To determine the odds of cesarean, operative vaginal delivery cesarean delivery. 1,2 In an attempt to minimize the risks associated and vaginal after cesarean after successful external cephalic version with either breech vaginal delivery or elective cesarean delivery, the (ECV) compared with singleton eligible for a trial of labor. American College of Obstetricians and Gynecologists recommended Study Design: A matched case–control study was performed using the application of external cephalic version (ECV) to reduce breech 3 Memorial Care OBStat Database from 1 January 1998 to 31 July 2006. We presentations in singleton . ECV is a safe and effective identified 197 participants who underwent a successful ECV (study group) method that consists of rotating the fetus from breech to cephalic and compared them with the next two women presenting for labor presentation by external manipulation of the mother’s abdomen, 2,4 management, matched for parity, , delivery history with the success rate ranging from 44 to 77%. (previous cesarean delivery) and type of labor (spontaneous or induced). Although the intent of ECV is to decrease the overall cesarean delivery rate for breech presentation, there is controversy as to Result: There was no significant difference in the adjusted matched odds whether this is ultimately achieved. Some studies have shown an of cesarean delivery between the study group and control group overall increased rate of cesarean delivery following successful ECV, (16.8 vs 11.9%; odds ratio (OR) 1.70; 95% confidence interval (CI) 0.98 to whereas others have not.5–12 Part of the difficulty in drawing a 2.97), even when subanalyzed according to parity. There was also no conclusion from these studies is that differing control groups and significant difference in adjusted matched odds of operative vaginal matching characteristics have been used. In addition, small sample delivery for the study group and control group, 15.9 vs 8.9% (OR 1.06; sizes may have been a limiting factor in some studies. Further, 95% CI 0.32 to 3.51). Among patients with a prior cesarean, those who many have excluded participants with prior cesarean delivery, underwent successful ECV had a cesarean delivery rate of 11.1% compared making information on this subset of individuals limited. with 16.7% in the matched control group (OR 0.59; 95% CI 0.47 to 7.43). Therefore, the goals of this investigation were to settle the questions Conclusion: Cesarean delivery and operative vaginal delivery rates as to whether successful ECV truly lowers the rate of cesarean following successful ECV are not increased in our data set compared with delivery in patients with non-cephalic presentation and whether matched controls, even in patients with a prior cesarean delivery. This women undergoing successful ECV have a higher rate of cesarean information may be useful when counseling patients who are delivery as compared with those presenting for labor management contemplating an ECV attempt due to non-cephalic presentation at term. at term with spontaneous cephalic presentations. Journal of Perinatology (2009) 29, 96–100; doi:10.1038/jp.2008.227; published online 8 January 2009

Keywords: external cephalic version; cesarean delivery; breech presentation Methods This is a matched case–control study that was performed at Introduction Saddleback Memorial Medical Center, an academically affiliated, community-based hospital. Before study initiation, institutional Breech presentation complicates 3 to 4% of term pregnancies, and review board approval for human participant research was obtained planned vaginal delivery for a singleton, term, breech presentation from the Memorial Services Office of Research Administration. Correspondence: Dr C Clock, Division of Maternal-Fetal Medicine, Department of Obstetrics Cases were identified in a retrospective fashion using a and Gynecology, Oregon Health Sciences University, 3181 SW Sam Jackson Park Rd, Mail contemporaneously collected, quality assurance perinatal database Code L548, Portland, OR 97239, USA. known as OBStat. Cases were defined as participants with a E-mail: [email protected] Received 1 June 2008; revised 21 November 2008; accepted 23 November 2008; published online singleton who underwent a successful ECV for 8 January 2009 malpresentation during the study period from 1 January 1998 to 31 Cesarean risk after successful cephalic version C Clock et al 97

July 2006. The obstetric and perinatal information of this study confidence interval (CI) for the primary outcome of cesarean population was retrieved from the Memorial Care OBStat Database. delivery were then calculated using conditional logistic regression. The information in this database is entered by a trained clinical Covariates in the multivariate model included diabetes mellitus, nurse and is checked weekly for accuracy. pre-eclampsia, cardiac disease and gestational age at delivery. All During the study period, the clinical protocol of the department analyses were performed using Stata, version 9.0 (StataCorp LP, was to offer ECV to women with a breech presentation at 36 weeks’ College Station, TX, USA). gestation or beyond. ECV was not offered or performed if the patient had any of the following contraindications: multifetal , previa, third trimester bleeding, oligohydramnios (amniotic fluid index <5), premature rupture of membranes, Results intrauterine growth restriction, previous uterine surgery other than During the study period, 914 participants presented with breech prior low-transverse cesarean section, gross fetal anomalies, non- presentation at >36 weeks’ gestation. ECV was attempted on 420 reassuring fetal heart tracing or uterine anomaly. Breech vaginal (46.0%) participants, of which 197 were successful (47.5%). We delivery for the term, singleton gestation was not routinely offered identified 394 matched controls for a total sample of 591 at our hospital during the study period. participants. Though delivery volume remained relatively steady in All ECV procedures in our hospital are performed by the our medical center during the study period (approximately 250 per Maternal–Fetal Medicine faculty members in a similar fashion. A month), the average number of ECV attempts during the study standard ECV order set was used by all practitioners. Before the period peaked at 3 per month in 2002 and had been decreasing procedure, the following clinical parameters are routinely since that time till it reached less than 1 per month. documented: confirmation by ultrasound that the fetus remains in There were no significant differences between the two groups a non-vertex presentation, adequacy of the amniotic fluid index, with respect to maternal age or medical complications including the absence of active labor and the presence of a reassuring fetal insulin-dependent diabetes, cardiac disease, gestational diabetes heart rate pattern by external fetal monitoring. Informed consent mellitus, chronic hypertension or pre-eclampsia (Table 1). Other for the procedure is then obtained. Following the establishment of demographic factors that may potentially impact the cesarean intravenous access, terbutaline 0.25 mg is administered delivery rate, such as ethnicity and epidural use, were not subcutaneously to all women immediately before the procedure. statistically significantly different between the groups. In contrast, Epidural anesthesia is not used. The external cephalic procedure is participants who underwent successful ECV had a higher then performed by a single physician with the assistance of a proportion of deliveries at 41 weeks’ gestation or later, though not nurse. The assistant monitors fetal presentation and heart rate by significant. ultrasound during the procedure. Before discharge, post-procedure Among those participants requiring induction of labor, the fetal monitoring is performed for a minimum of 1 h. The care of mean interval from ECV to induction was 13.6 days (range 0 to 38 women after ECV was the same as in those pregnancies with days). As induction status was one of the matching criteria, there spontaneous cephalic presentation. was no difference between the two groups (Table 1). Twenty Before study initiation, an a priori, two-sided, sample size patients (10.2%) underwent induction of labor within 24 h calculation was performed for our primary outcome of cesarean following successful ECV. Within this group, indications for delivery, utilizing prior studies in this area to estimate incidence induction included elective (15%), abnormal post-procedure non- rates.4–10,13 With an a error of 0.05, we determined that 197 stress tests (10%), post-dates (10%), suspected macrosomia (5%) patients and 394 controls, matched in a 2:1 fashion, would yield and pre-eclampsia (5%). The remaining indications for induction 80% power to detect a 10% (25 vs 35%) difference in cesarean were unspecified. delivery, assuming a cesarean delivery rate in the control group of Of the women undergoing successful ECV, 83.2% were approximately 25%. Controls were identified in a retrospective subsequently delivered vaginally. Despite a slightly higher overall fashion using a quality assurance database of prospectively rate of cesarean delivery in the ECV group vs controls (16.8 vs collected data as described earlier and were matched for parity, 11.9%), the matched odds ratio (OR) for the primary outcome of history of prior cesarean delivery, gestational age at delivery and cesarean delivery was not statistically different between the two induction status. Following identification of a case patient, the groups, based on a 95% CI that crosses one (OR 1.70; 95% CI 0.98 subsequent two participants who met the matching criteria were to 2.97). There was also no statistical difference in the indication chosen as controls. for cesarean delivery between the ECV cases and controls (data not Demographic data were compared by ECV status, using shown). When the population was subdivided based on parity, the conditional logistic regression for categorical variables and the cesarean rate was higher among nulliparous women, but the paired t-test for continuous variables, to account for the matched matched odds of cesarean delivery showed no statistical difference study design. The unadjusted and adjusted odds and 95% between the ECV group and controls (Table 2).

Journal of Perinatology Cesarean risk after successful cephalic version C Clock et al 98

Table 1 Demographic data by ECV status

ECV Unadjusted matched OR (95% CI)a

No (N ¼ 394) Yes (N ¼ 197)

Maternal age X35 years 110 (27.9%) 63 (32.0%) 1.23 (0.84–1.80) <35 years 284 (72.1%) 134 (68.0%) 1.0

Parity Nulliparous 150 (38.1%) 75 (38.1%) F Multiparous 244 (61.9%) 122 (61.9%) F

Gestational age at delivery X41 weeks 45 (11.4%) 34 (17.3%) 1.96 (1.10–3.52) <41 weeks 349 (88.6%) 163 (82.7%) 1.0

Prior cesarean 18 (4.6%) 9 (4.6%) F Induction 168 (42.6%) 84 (42.6%) F Insulin-dependent diabetes mellitus 2 (0.5%) 2 (1.0%) 2.0 (0.28–14.2) Gestational diabetes mellitus 7 (1.8%) 4 (2.0%) 1.14 (0.33–3.90) Pre-eclampsia 7 (1.8%) 7 (3.4%) 2.0 (0.70–5.70)

Abbreviations: CI, confidence interval; ECV, external cephalic version; OR, odds ratio. aMatched ORs and 95% CIs not presented for matching variables.

Table 2 Primary outcome of cesarean delivery as a function of ECV status

Cesarean delivery rate Unadjusted matched ORs of cesarean delivery Adjusted matched ORs of cesarean delivery OR (95% CI) OR (95% CI)a

All subjects ECV 33/197 (16.8%) 1.60 (0.94–2.72) 1.70 (0.98–2.97) No ECV 47/394 (11.9%) 1.0 1.0

Among nulliparous ECV 24/75 (32.0%) 1.26 (0.65–2.47) 1.36 (0.67–2.78) No ECV 39/150 (26.0%) 1.0 1.0

Among multiparous ECV 9/122 (7.4%) 2.39 (0.88–6.49) 1.88 (0.49–7.22) No ECV 8/244 (3.3%) 1.0 1.0

Among VBACS ECV 1/9 (11.1%) 0.59 (0.47–7.43) F No ECV 3/18 (16.7%) 1.0 F

Abbreviations: CI, confidence interval; ECV, external cephalic version; OR, odds ratio. aCovariates in the multivariate model included diabetes mellitus, pre-eclampsia, cardiac disease and gestational age at delivery. Multivariate analysis among subjects with a history of a prior cesarean delivery unable to be performed secondary to the small sample size.

No significant statistical difference was found in the rate of OR 1.06; 95% CI 0.32 to 3.51). There was no significant increase in operative vaginal delivery between the ECV group and the control fetal or maternal morbidity or mortality in participants following group, when the adjusted matched OR was calculated (15.9 vs 8.9%; an ECV attempt (whether or not it was successful). There were no

Journal of Perinatology Cesarean risk after successful cephalic version C Clock et al 99

Table 3 Neonatal outcomes increase in cesarean deliveries in the ECV group. This finding raises the possibility that residual confounding differences ECV remained between the two groups, which we could not identify and No Yes Unadjusted matched thus control the possibility that our study was underpowered. (N ¼ 394) (N ¼ 197) OR (95% CI) Although our success rate of 47% for ECV in our institution was somewhat low, it is consistent with reported rates of 44 to 77%.2,4 One notable and somewhat surprising finding of our investigation >4000 g 59 (15.0%) 28 (14.2%) 0.95 (0.14–6.4) 2500–4000 g 331 (84.0%) 167 (84.8%) 1.01 (0.16–6.5) was the low cesarean delivery rates in both the ECV and the control <2500 g 4 (1.0%) 2 (1.0%) 1.0 groups. This is in contrast to the current overall cesarean delivery rate in our hospital of approximately 25% during the study period. Gender This difference may be explained by the relatively low risk profile of Female 187 (47.5%) 100 (50.8%) 0.89 (0.63–1.23) our study population, with few concomitant medical problems, such as pregestational and gestational diabetes, chronic hypertension and One-minute Apgar <7 other pregnancy complications including pre-eclampsia. This Yes 17 (4.3%) 10 (5.1%) 1.17 (0.54–2.57) apparent low level of maternal risk factors in our study population may be the result of selection bias. Specifically, it is possible that Five-minute Apgar <9 participants who are offered and subsequently undergo ECV are better Yes 17 (4.3%) 8 (4.1%) 0.94 (0.39–2.25) candidates for the ECV procedure and subsequent vaginal delivery because they are a generally healthy population. The selected Admission to neonatal intensive care Yes 9 (2.3%) 3 (1.5%) 0.66 (0.18–2.46) controls were thus similarly healthy and relatively of low risk, due to the matching process. Abbreviations: CI, confidence interval; ECV, external cephalic version; OR, odds ratio. Though the lower-than anticipated cesarean delivery rate was a favorable outcome for both groups, it also limited the power of our statistically significant differences in birth weight, gender, 1- and study, and thus we cannot exclude a type II error. A post hoc, two- 5-min Apgar scores, or in the rates of admission to the neonatal sided, sample size calculation was performed showing that our intensive care unit between the two groups (Table 3). study had 51% power to show the difference that we found (17 vs Among the women with a history of prior cesarean delivery, the 12%) and >90% power to show a 50% difference (24 vs 12%), with median parity for both the ECV and control groups was 2 (range 1 an a error of 0.05. to 3). Of the nine ECV participants with a prior cesarean delivery, Although we had a sample size too small for separate 1 (11.1%) required cesarean delivery compared with 3 (16.7%) in evaluation, our study demonstrated a high rate of successful ECV the control group (OR 0.59; 95% CI 0.47 to 7.43). The cesarean and subsequent vaginal delivery (88.9%) in participants with a delivery in the ECV group was performed in a primaparous woman prior cesarean delivery. Many institutions treat prior cesarean as an because of failure to progress. There was only one uterine rupture exclusion criteria for ECV attempt. In three individual studies, de among all the women with a prior cesarean delivery, which Meeus et al.14, Flamm15 and Schachter et al.13 reported successful occurred in a control patient who presented in spontaneous labor. ECV in a total of 105 women with prior cesarean delivery, with rates ranging from 65 to 100%.14,15,13 Of these, 54 to 76% had a successful vaginal delivery, and there was no reported increase in Discussion the rate of uterine rupture or dehiscence. In our own study, the Some argue that even if ECV is successful, cesarean delivery is higher median parity likely increased the success of subsequent more likely due to malpresentation or due to higher rates of vaginal delivery in both the groups. Despite the limited sample size, subsequent non-reassuring fetal heart rate testing in labor. these data suggest that an appropriate selection of ECV candidates However, the existing literature as to whether ECV is associated with can result in successful vaginal birth after cesarean without an an increased cesarean delivery rate at term is divided.5–12 These increased rate of complications. varying conclusions may be related to differences in control groups An interesting observation during our study period was the and matching characteristics. In addition, many studies are limited fluctuation in the number of attempted ECVs each year. As noted, by small sample sizes. despite a relatively steady delivery volume during the study period, Our study is one of the largest controlled studies to evaluate the the average number of ECV procedures peaked in 2002 at 3 per mode of delivery following a successful ECV. In our multivariate month and has subsequently decreased to 1.0, 1.3 and 0.6 per model, we did not find a statistically significant increase in the month in 2004, 2005 and 2006, respectively. Possible explanations matched odds of cesarean delivery, although the point estimate was include the lower physician threshold for cesarean delivery over consistent with a clinically small, but potentially interesting, time and increased patient demand for cesarean delivery.

Journal of Perinatology Cesarean risk after successful cephalic version C Clock et al 100

On the basis of the data from our study, the impact of ECV on References the broader rate of cesarean delivery may be theoretically 1 Hannah M, Hannah W, Hewson S, Hodnett E, Saigal S, Willan A. Planned cesarean estimated. If we assume that there are approximately 3.5 million section versus planned vaginal birth for breech presentation at term: a randomized term, singleton in the United States each year, of which 3.5% multicentre trial. The Lancet 2000; 356: 1375–1383. are breech, 122 500 pregnancies would be eligible for ECV.1,16 2 Saling EM. External cephalic version under tocolysis. J Perinatal Med 1975; 3: 115. Using the results from our study, 46% of these pregnancies would 3 American College of Obstetricians and Gynecologists. External Cephalic Version (February 2000). undergo ECV. If at least 47% of ECV procedures are successful and 4 Chan LY, Tang JL, Tsoi KF, Fok WY, Chan LW, Lau TK. Intrapartum cesarean delivery 83.2% of successful ECV patients are delivered vaginally, 22 161 after successful external cephalic version: a meta-analysis. Obstet Gynecol 2004; 104: cesarean deliveries per year could be avoided. Although this may 155–160. seem to be a small number relative to the number of births 5 Chan LKS, Leung TK, Fok WY, Chan LW, Lau TK. High incidence of obstetric annually in the United States, any reduction in the number of interventions after successful external cephalic version. BJOG 2002; 109: patients undergoing primary cesarean delivery will also reduce the 627–631. 6 Lau TK, Lo KWK, Rogers M. Pregnancy outcome after successful external number of probable repeat cesarean deliveries in the future. In cephalic version for breech presentation at term. Am J Obstet Gynecol 1997; 176: addition, avoiding the first cesarean section would reduce the 218–223. number of patients who will have abnormal placentation in a 7 Laros Jr RK, Flanagan TA, Kilpatrick SJ. Management of term breech presentation: a subsequent pregnancy, such as a placenta accreta (leading to protocol of external cephalic version and selective trial of labor. Am J Obstet Gynecol hemorrhage and significant operative morbidity) or abnormal 1995; 172: 1916–1923. 8 Vezina Y, Bujold E, Varin J, Marquette G, Boucher M. Cesarean delivery after successful implantation (which can be associated with stillbirth). external cephalic version of breech presentation at term: a comparative study. On the basis of the sample size of our study, our data support Am J Obstet Gynecol 2004; 190: 763–768. the role of ECV as a safe alternative to cesarean delivery for breech 9 Ben-Haroush A, Perri T, Bar J, Yogev Y, Kaplan B. Mode of delivery following presentation at term. It also demonstrates that once an ECV is successful external cephalic version. Am J Perinatol 2002; 19(7): successfully performed, the patient’s likelihood of a successful 355–360. vaginal delivery is comparable with that of a patient with a 10 Wax JR, Sutula K, Lerer T, Steinfeld JD, Ingardia CJ. Labor and delivery following successful external cephalic version. Am J Perinatol 2000; 17: 183–186. spontaneous cephalic presentation, even when the patient has 11 Siddiqui D, Stiller RJ, Collins J, Laifer SA. Pregnancy outcome after successful external undergone an earlier cesarean delivery. cephalic version. Am J Obstet Gynecol 1999; 181: 1092–1095. 12 Egge T, Schauberger C, Shaper A. Dysfunctional labor after external cephalic version. Financial support Obstet Gynecol 1994; 80: 771–773. 13 Schachter M, Kogan S, Blickstein I. External cephalic version after previous cesarean None. sectionFa clinical dilemma. Int J Gynecol Obstet 1994; 45: 17–20. Synopsis 14 de Meeus JB, Ellia F, Magnin G. External cephalic version after previous cesarean section: a series of 38 cases. Eur J Obstet Gynecol Reprod Biol 1998; 81: Successful external cephalic version does not seem to be associated with increased 65–68. odds of subsequent cesarean delivery compared with matched controls. 15 Flamm BL. External cephalic version after previous cesarean section. Am J Obstet Conflict of interest Gynecol 1991; 165: 370–372. 16 Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Kirmeyer S et al. Birth: The authors of this paper declare no conflict of interest. Final Data for 2005. Natl Vital Stat Rep 2007; 56(6): 1–104.

Journal of Perinatology