- Email: [email protected]
Predictors of cesarean delivery after prelabor rupture of membranes at term
Predictors of Cesarean Delivery After Prelabor Rupture of Membranes at Term DAVID PELEG, MD, MARY E. HANNAH, MDCM, ELLEN D. HODNETT, RN, PhD, GARY A. FOSTER, PhD, ANDREW R. WILLAN, PhD, AND DAN FARINE, MD, FOR THE TERM PRELABOR RUPTURE OF MEMBRANES STUDY GROUP Objective: To identify the significant predictors of cesarean delivery after prelabor rupture of membranes (PROM) at term. Methods: In a multicenter study involving 72 institutions in six countries, 5041 women were randomized to induction of labor with oxytocin or prostaglandins or to expectant management. We did univariate and multivariate logistic regression analyses to determine the statistically significant independent predictors of cesarean delivery (P < .05). Results: The following variables were found to be significantly associated with cesarean delivery: delivery in Israel, versus Canada (odds ratio [OR] 0.34); delivery in Australia, versus Canada (OR 1.93); nulliparity (OR 2.81); labor lasting more than 12 hours, versus less than 6 hours (OR 2.78); labor lasting 6 –12 hours, versus less than 6 hours (OR 1.66); previous cesarean delivery (OR 2.75); epidural anesthesia (OR 2.66); clinical chorioamnionitis (OR 2.42); internal fetal heart rate monitoring (OR 2.19); birth weight of at least 4000 g (OR 2.07); use of oxytocin (OR 1.97); maternal age of at least 35 years (OR 1.44); latent period of at least 12 hours (OR 1.41); and meconium staining (OR 1.41). Conclusion: Strong predictors of cesarean delivery after PROM at term were country of birth, nulliparity, long labor, previous cesarean delivery, and epidural anesthesia. (Obstet Gynecol 1999;93:1031–5. © 1999 by The American College of Obstetricians and Gynecologists.)
From the Department of Obstetrics and Gynecology, Poriya Hospital, Tiberias, Israel; and the Department of Obstetrics and Gynaecology, Sunnybrook and Women’s College Health Sciences Centre, the Maternal, Infant and Reproductive Health Research Unit, Centre for Research in Women’s Health, the Faculty of Nursing, and the Department of Obstetrics and Gynaecology, Mount Sinai Hospital, University of Toronto, Toronto, Canada. Supported by Medical Research Council of Canada grant no. MA11392. The members of the TermPROM Study Group are listed in Hannah ME, Ohlsson A, Farine D, Hewson SA, Hodnett ED, Myhr TL, et al. Induction of labor compared with expectant management for prelabor rupture of the membranes at term. TermPROM Study Group. N Engl J Med 1996;334:1005–10.
VOL. 93, NO. 6, JUNE 1999
Approximately 8% of women with pregnancies at term have prelabor rupture of membranes (PROM). Management options include immediate induction of labor and observation in the hope that spontaneous labor will begin. Before 1996, some clinicians considered expectant management to be preferable for women because studies suggested the risk of cesarean delivery might be lower.1,2 The Term Prelabor Rupture of the Membranes (TermPROM) Study Group compared immediate induction with intravenous (IV) oxytocin or vaginal prostaglandin (PG) E2 gel, to expectant management up to 4 days followed by labor induction if labor did not begin spontaneously.3 There was no difference in the rate of cesarean delivery between the induction and expectant management groups. We did this secondary analysis of data from the TermPROM Study to determine which factors independently predict cesarean delivery in women with PROM at term. Variables under consideration included parity, maternal age, duration of the latent period, spontaneous versus induced labor, duration of active labor, cervical dilation and effacement at study entry, development of clinical chorioamnionitis, previous cesarean delivery, use of oxytocin in labor, epidural anesthesia, internal fetal heart rate (FHR) monitoring, meconium staining, birth weight, gestational age at study entry, and country of birth. We also wished to determine whether there was significant interaction between mode of labor onset (induced versus spontaneous) and the duration of the latent period with regard to cesarean delivery.
Materials and Methods The TermPROM Study, completed and reported on in 1996,3 involved 5041 women with singleton, term pregnancies and PROM from 72 institutions, in six countries.
0029-7844/99/$20.00 PII S0029-7844(98)00570-5
1031
Rupture of membranes was determined clinically and confirmed by positive nitrazine or fern test. Women were excluded if they were in active labor, there had been a previous failed attempt to induce labor, presentation was noncephalic, or there was a contraindication to either induction of labor (eg, placenta previa) or expectant management (eg, maternal fever, meconium staining, or other evidence of maternal or fetal compromise). Women were assigned by a centrally controlled telephone randomization program to one of four treatments: immediate induction of labor with IV oxytocin, immediate induction of labor with vaginal PGE2, expectant management followed by induction of labor with IV oxytocin if necessary, and expectant management followed by induction of labor with vaginal PGE2 gel if necessary. If labor was induced with an infusion of oxytocin, the rate was titrated to contractions according to local hospital policy. If labor was induced with vaginal PGE2 gel, 1–2 mg was inserted into the posterior vaginal fornix, and this was repeated 6 hours later if labor had not started. Oxytocin was infused 4 hours later if labor still had not started. The details of the study design were published previously.3 The medical personnel made other decisions concerning fetal and maternal care and mode of delivery. Approximately 95% of women in all groups had vaginal or introital cultures performed for group B streptococcus at study entry. Vaginal examinations were avoided, as part of the protocol, to minimize risk of ascending infection. If a digital or speculum vaginal examination was done before the study, information on dilation and effacement was recorded. Information on cervical dilation and effacement otherwise was not recorded before induction of labor or at the onset of spontaneous labor. For women in either of the induction groups, the cervical status at study entry most likely represented the status of the cervix before induction of labor, because labor was induced shortly after randomization. For the others, the state of the cervix might have become more favorable as time passed. Labor was considered to have begun when the cervix became dilated to at least 3 cm or effaced to at least 80% and when uterine contractions were occurring at least every 5 minutes. Clinical chorioamnionitis was considered present when one or more of the following criteria were met: maternal temperature persistently above 37.5C on two or more occasions, at least 1 hour apart, or a single temperature above 38C before delivery; maternal white blood cell count greater than 20,000 cells/mm3; or foul-smelling amniotic fluid. Of the 5041 enrolled women, we excluded 13 who had prelabor cesarean deliveries. Among the 5028 remaining women, 4532 (90.1%) delivered vaginally and 496 (9.9%) by cesarean. Variables possibly affecting the
1032 Peleg et al
Predictors of Cesarean Delivery
need for cesarean delivery among normally laboring women were identified, and the association between those variables and mode of delivery was tested using univariate analyses (x2 test). The variables associated with cesarean delivery at P # .3 on univariate analysis and those variables for which there was previous evidence for an association with cesarean delivery were considered for the multivariate regression model. Because vaginal examinations were avoided in the TermPROM Study, there was a substantial amount of missing data on cervical dilation and effacement; thus those variables were not considered for the model. Instead, we combined cervical dilation and effacement, with mode of labor onset, as one variable (induced labor with an unfavorable cervix or induced labor with a favorable cervix versus spontaneous labor), and that variable, regarding which information was available for 4113 women, also was considered for the model. An unfavorable cervix was defined as a cervix dilated less than 3 cm and effaced less than 80%. The interaction term (induced versus spontaneous labor by duration of the latent period) was considered for the model. Multivariate analysis was done by stepwise logistic regression to determine the statistically significant (P , .05) independent predictors of cesarean delivery. Logistic regression analysis is a statistical method for evaluating the effects of multiple independent variables on a dichotomous dependent variable such as mode of delivery. The purpose of presenting the results of the univariate and multivariate analyses is to demonstrate the extent to which the effect of each independent variable was affected by the others.
Results The demographic information at study entry and perinatal outcomes, by country of birth, for the study population are detailed in Table 1. On univariate analysis, the variables that were significantly (P , .001) associated with an increase in cesarean delivery were birth weight of at least 4000 g; nulliparity; meconium staining; induced labor; use of oxytocin in labor; internal FHR monitoring; epidural anesthesia; clinical chorioamnionitis; duration of latent period of at least 12 hours; duration of active labor of 6 –12 hours, versus less than 6 hours, and of at least 12 hours, versus less than 6 hours; induced labor with an unfavorable cervix at study entry, versus spontaneous labor; and delivery in Canada, versus Israel. Variables significantly (P $ .001 and # .3) associated with an increase in cesarean delivery were cervical dilation of less than 3 cm at study entry (P 5 .015), previous cesarean delivery (P 5 .132), maternal age of at least 35 years (P 5 .135), and delivery in Canada, versus Sweden (P 5 .10), Denmark (P 5
Obstetrics & Gynecology
Table 1. Demographic Information at Entry and Perinatal Outcome United Kingdom (n 5 1341)
Sweden (n 5 105)
11.0 60.7 12.9 84.0
10.5 53.3 4.8 98.1
99.0 (n 5 984) 99.1 (n 5 891) 82.6
99.0 (n 5 103) 100.0 (n 5 103) 90.5
6.9 2.9 35.2 2.6 9.7
21.0 3.8 35.2 3.8 5.7
Maternal age $35 y* Nulliparity* Gestational age $41 wk* Vaginal digital or speculum examination* Cervical dilation ,3 cm*† Cervical effacement ,80%*† ,12 h of membrane rupture* GBS positive* Previous cesarean delivery* Epidural anesthesia Clinical chorioamnionitis Cesarean delivery
Denmark (n 5 28)
Australia (n 5 375)
Israel (n 5 331)
Canada (n 5 2848)
Overall (n 5 5028)
10.7 35.7 14.3 96.4
5.3 59.2 9.3 97.3
12.7 34.1 14.5 96.4
12.4 62.4 6.6 72.9
11.5 59.5 9.0 79.9
100.0 (n 5 27) 100.0 (n 5 27) 92.9
98.5 (n 5 323) 99.3 (n 5 292) 80.3
98.1 (n 5 308) 99.0 (n 5 304) 85.5
98.3 (n 5 1858) 98.0 (n 5 1717) 77.9
98.5 (n 5 3603) 98.6 (n 5 3334) 80.2
9.6 1.9 35.7 2.9 11.5
6.3 4.5 54.4 3.0 3.0
12.2 3.8 62.3 9.6 10.7
10.3 3.5 51.7 6.7 9.9
3.6 0 7.1 3.6 3.6
GBS 5 group B streptococcus culture result from introital or vaginal specimen obtained at entry. * Values given as % unless otherwise indicated. Demographic information at study entry. † Cervical dilation and effacement were determined by vaginal digital examination, if done, otherwise by vaginal speculum examination.
.221), and the United Kingdom (P 5 .30). The variables not significantly (P . .3) associated with cesarean delivery on univariate analysis were delivery in Australia, versus Canada (P 5 .67); induced labor with a favorable cervix, versus spontaneous labor (P 5 .96); cervical effacement of less than 80% at study entry (P 5 .394); and gestational age of at least 41 weeks (P 5 .305). All variables, other than cervical dilation and effacement, were considered for the multivariate regression model. On multivariate analysis, the strongest predictor of cesarean delivery among women with PROM at term was country of birth (Table 2). Women in Israel were significantly less likely to deliver by cesarean than were women in Canada (odds ratio [OR] 0.34; 95% confiTable 2. Independent Predictors of Cesarean Delivery
Country of birth Israel vs Canada Australia vs Canada Parity 0 vs $1 Active labor .12 h vs ,6 h 6 –12 h vs ,6 h Previous cesarean delivery Epidural anesthesia Clinical chorioamnionitis Internal fetal heart rate monitoring Birth weight $4000 vs ,4000 g Use of oxytocin in labor Maternal age $35 vs , 35 years Latent period $12 vs , 12 hours Meconium staining
OR
95% CI
P
0.34 1.93 2.81
0.15, 0.81 1.29, 2.88 1.95, 4.05
.01 .001 ,.001
2.78 1.66 2.75 2.66 2.42 2.19 2.07 1.97 1.44 1.41 1.41
2.01, 3.85 1.24, 2.22 1.46, 5.16 1.94, 3.65 1.80, 3.25 1.74, 2.75 1.50, 2.86 1.47, 2.65 1.03, 2.02 1.08, 1.85 1.03, 1.93
,.001 .001 .002 ,.001 ,.001 ,.001 ,.001 ,.001 .034 .013 .031
OR 5 odds ratio; CI 5 confidence interval.
VOL. 93, NO. 6, JUNE 1999
dence interval [CI] 0.15, 0.81), and women in Australia were significantly more likely to deliver by cesarean than were women in Canada (OR 1.93; 95% CI 1.29, 2.88). Other strong (OR greater than 2.5) predictors of cesarean delivery were nulliparity (OR 2.81; 95% CI 1.95, 4.05); duration of active labor exceeding 12 hours, versus less than 6 hours (OR 2.78; 95% CI 2.01, 3.85), previous cesarean delivery (OR 2.75; 95% CI 1.46, 5.16) and epidural anesthesia (OR 2.66; 95% CI 1.94, 3.65) (Table 2). Other less strong but statistically significant predictors were clinical chorioamnionitis; internal FHR monitoring; birth weight of at least 4000 g; use of oxytocin in labor; duration of active labor of 6 –12 hours, versus less than 6 hours; latent period of at least 12 hours, versus less than 12 hours; and meconium staining (Table 2). Gestational age at study entry, induced versus spontaneous labor, and induced labor with an unfavorable cervix or a favorable cervix versus spontaneous labor were not associated with cesarean delivery on multivariate analysis. There was no significant interaction between duration of latent period and induced versus spontaneous labor.
Discussion Much of the debate surrounding PROM at term has centered around whether induction of labor increases the risk of cesarean delivery. The TermPROM Study and meta-analyses of randomized controlled trials have not found a statistically significant difference in the rate of cesarean delivery between induction of labor with oxytocin or PGs and expectant management.3–7 Despite
Peleg et al
Predictors of Cesarean Delivery
1033
those results, we thought it appropriate to explore further the association between induced or spontaneous labor and cesarean delivery. After controlling for other fetal and maternal factors for which we had information, we did not find induced labor to be associated with a higher risk of cesarean delivery, regardless of duration of the latent period and status of the cervix at study entry. For women with PROM at term but otherwise without complications, the independent fetal and maternal risk factors for cesarean delivery were nulliparity, prolonged labor, previous cesarean delivery, epidural anesthesia, clinical chorioamnionitis, use of internal FHR monitoring, larger infants, use of oxytocin in labor, older maternal age, longer duration of latent period, and meconium staining. Except for duration of latent period, those risk factors are similar to those associated with cesarean delivery in women at term without PROM.8 –10 The association between duration of latent period and cesarean delivery might be due to subclinical chorioamnionitis, because the effect was independent of the development of clinical chorioamnionitis and we previously found longer latent periods to be associated with increased risk of clinical chorioamnionitis.11 We found that country of birth was a strong predictor of cesarean delivery, after controlling for fetal and maternal factors. Variations in rates of cesarean delivery were noted previously among countries, institutions, and physicians and nurses.8,12–15 In 1991–1992, the rates of cesarean delivery in community hospitals, without facilities for neonatal intensive care, in Ontario, Canada, varied from 6.6 to 30.3%.13 Those large variations in rates of cesarean delivery for different settings likely are due to differences in environmental factors including variations in quality and degree of supportive care provided to women in labor, supervision of labor and birth by midwives versus physicians, timing of admission to labor and delivery units, use of fetal scalp sampling or interpretation of FHR monitor tracings to diagnose unreassuring FHR, routine versus selective use of epidural anesthesia, availability of practitioners with expertise in operative vaginal delivery, and accepted upper limits for length of first and second stages of labor, all of which can have an effect on the likelihood of cesarean delivery.16 –22 It is not known specifically why Israel had a significantly lower rate of cesarean delivery than Canada, but this merits further investigation. This study was a retrospective analysis of data, so there might be other fetal and maternal factors associated with cesarean delivery after PROM concerning which we had no or inadequate information. We had no information on environmental factors other than coun-
1034 Peleg et al
Predictors of Cesarean Delivery
try of birth. Future prospective research should involve evaluation of changes in environment, while controlling for fetal and maternal factors. The ongoing Nursing SCIL Trial, funded by the National Institutes of Health, coordinated in Toronto by EDH, is an example of such an evaluation. The Nursing SCIL Trial is a multicenter, randomized controlled trial evaluating the effect of continuous, supportive care in labor by nurses in institutions in the United States and Canada. The present study confirmed that the fetal and maternal risk factors for cesarean delivery for women with PROM at term are similar to those for women with other uncomplicated term pregnancies. After country of birth was controlled for, the greatest risks were for nulliparous women, those with long labors, those who had previous cesarean deliveries, and those receiving epidural anesthesia. The risk, however, was not increased if labor was induced. An approach that decreases the duration of the active phase of labor and the use of epidural anesthesia, particularly among nulliparous women, might decrease the rate of cesarean delivery in women with PROM at term.
References 1. Wagner MV, Chin VP, Peters CJ, Drexler B, Newman LA. A comparison of early and delayed induction of labor with spontaneous rupture of the membranes at term. Obstet Gynecol 1989;74: 93–7. 2. Grant JM, Serle E, Mahmood T, Sarmandal P, Conway DI. Management of prelabour rupture of the membranes in term primigravidae: Report of a randomized prospective trial. Br J Obstet Gynaecol 1992;99:557– 62. 3. Hannah ME, Ohlsson A, Farine D, Hewson SA, Hodnett ED, Myhr TL, et al. Induction of labor compared with expectant management for prelabor rupture of the membranes at term. TermPROM Study Group. N Engl J Med 1996;334:1005–10. 4. Mozurkewich EL, Wolf FM. Prelabor rupture of membranes at term: A meta-analysis of three management schemes. Obstet Gynecol 1997;89:1035– 43. 5. Tan B, Hannah ME. Oxytocin for prelabour rupture of the membranes at or near term. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. 6. Tan B, Hannah ME. Prostaglandins for prelabour rupture of the membranes at or near term. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. 7. Tan B, Hannah ME. Prostaglandins vs. oxytocin for prelabour rupture of the membranes at or near term. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. 8. Goyert GL, Bottoms SF, Treadwell MC, Nehra PC. The physician factor in cesarean birth rates. N Engl J Med 1989;320:706 –9. 9. Cnattingius R, Cnattingius S, Notzon FC. Obstacles to reducing cesarean rates in a low-cesarean setting: The effect of maternal age, height, and weight. Obstet Gynecol 1998;92:501– 6. 10. Lieberman E, Lang JM, Cohen A, D’Agostino R, Datta S, Frigoletto FD. Association of epidural analgesia with cesarean delivery in nulliparas. Obstet Gynecol 1996;88:993–1000. 11. Seaward PG, Hannah ME, Myhr TL, Farine D, Ohlsson A, Wang EE, et al. International Multicenter Term Prelabor Rupture of Membranes Study: Evaluation of predictors of clinical chorioam-
Obstetrics & Gynecology
12. 13.
14. 15.
16.
17. 18.
19.
20.
nionitis and postpartum fever in patients with prelabor rupture of membranes at term. Am J Obstet Gynecol 1997;177:1024 –9. Notzon FC. International differences in the use of obstetric interventions. JAMA 1990;263:3286 –91. Hospital-specific and regional-level utilization information on cesarean sections, appendectomy and breast cancer surgery. In: Naylor CD, Anderson GM, Goel V, eds. Patterns of health care in Ontario. Ottawa: Canadian Medical Association, 1994;137– 42. Learman LA. Regional differences in operative obstetrics: A look to the south. Obstet Gynecol 1998;92:514 –9. Radin TG, Harmon JS, Hanson DA. Nurses’ care during labor: Its effect on the cesarean birth rate of healthy, nulliparous women. Birth 1993;20:14 –21. Hodnett ED. Support from caregivers during childbirth. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. Harvery S, Jarrell J, Brant R, Stainton C, Rach D. A randomized, controlled trial of nurse-midwifery care. Birth 1996;23:128 –35. McNiven P, Williams IJ, Hodnett K, Kaufman K, Hannah M. The effect of early labour assessment on cesarean section rates for low risk nulliparous women. Birth 1998;25:5–10. Thacker SB, Stroup DF. Continuous electronic fetal heart monitoring during labor. (Cochrane Review) In: the Cochrane Library; Issue 3. Oxford: Update Software, 1998. Howell CJ. Epidural vs non-epidural analgesia in labour. (Cochrane review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998.
VOL. 93, NO. 6, JUNE 1999
21. Hannah WJ, workshop participants. The Canadian consensus on breech management at term. J SOGC 1994;16:1839 –58. 22. Maresh M, Choong KH, Beard RW. Delayed pushing with lumbar epidural analgesia in labour. Br J Obstet Gynaecol 1983;90:623–7.
Address reprint requests to:
David Peleg, MD c/o Mary E. Hannah, MDCM Maternal, Infant and Reproductive Health Research Unit Room 751 University of Toronto 790 Bay Street Toronto, ON M5G 1N8 Canada E-mail: [email protected]
Received August 31, 1998. Received in revised form November 25, 1998. Accepted December 10, 1998. Copyright © 1999 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Peleg et al
Predictors of Cesarean Delivery
1035
From the Department of Obstetrics and Gynecology, Poriya Hospital, Tiberias, Israel; and the Department of Obstetrics and Gynaecology, Sunnybrook and Women’s College Health Sciences Centre, the Maternal, Infant and Reproductive Health Research Unit, Centre for Research in Women’s Health, the Faculty of Nursing, and the Department of Obstetrics and Gynaecology, Mount Sinai Hospital, University of Toronto, Toronto, Canada. Supported by Medical Research Council of Canada grant no. MA11392. The members of the TermPROM Study Group are listed in Hannah ME, Ohlsson A, Farine D, Hewson SA, Hodnett ED, Myhr TL, et al. Induction of labor compared with expectant management for prelabor rupture of the membranes at term. TermPROM Study Group. N Engl J Med 1996;334:1005–10.
VOL. 93, NO. 6, JUNE 1999
Approximately 8% of women with pregnancies at term have prelabor rupture of membranes (PROM). Management options include immediate induction of labor and observation in the hope that spontaneous labor will begin. Before 1996, some clinicians considered expectant management to be preferable for women because studies suggested the risk of cesarean delivery might be lower.1,2 The Term Prelabor Rupture of the Membranes (TermPROM) Study Group compared immediate induction with intravenous (IV) oxytocin or vaginal prostaglandin (PG) E2 gel, to expectant management up to 4 days followed by labor induction if labor did not begin spontaneously.3 There was no difference in the rate of cesarean delivery between the induction and expectant management groups. We did this secondary analysis of data from the TermPROM Study to determine which factors independently predict cesarean delivery in women with PROM at term. Variables under consideration included parity, maternal age, duration of the latent period, spontaneous versus induced labor, duration of active labor, cervical dilation and effacement at study entry, development of clinical chorioamnionitis, previous cesarean delivery, use of oxytocin in labor, epidural anesthesia, internal fetal heart rate (FHR) monitoring, meconium staining, birth weight, gestational age at study entry, and country of birth. We also wished to determine whether there was significant interaction between mode of labor onset (induced versus spontaneous) and the duration of the latent period with regard to cesarean delivery.
Materials and Methods The TermPROM Study, completed and reported on in 1996,3 involved 5041 women with singleton, term pregnancies and PROM from 72 institutions, in six countries.
0029-7844/99/$20.00 PII S0029-7844(98)00570-5
1031
Rupture of membranes was determined clinically and confirmed by positive nitrazine or fern test. Women were excluded if they were in active labor, there had been a previous failed attempt to induce labor, presentation was noncephalic, or there was a contraindication to either induction of labor (eg, placenta previa) or expectant management (eg, maternal fever, meconium staining, or other evidence of maternal or fetal compromise). Women were assigned by a centrally controlled telephone randomization program to one of four treatments: immediate induction of labor with IV oxytocin, immediate induction of labor with vaginal PGE2, expectant management followed by induction of labor with IV oxytocin if necessary, and expectant management followed by induction of labor with vaginal PGE2 gel if necessary. If labor was induced with an infusion of oxytocin, the rate was titrated to contractions according to local hospital policy. If labor was induced with vaginal PGE2 gel, 1–2 mg was inserted into the posterior vaginal fornix, and this was repeated 6 hours later if labor had not started. Oxytocin was infused 4 hours later if labor still had not started. The details of the study design were published previously.3 The medical personnel made other decisions concerning fetal and maternal care and mode of delivery. Approximately 95% of women in all groups had vaginal or introital cultures performed for group B streptococcus at study entry. Vaginal examinations were avoided, as part of the protocol, to minimize risk of ascending infection. If a digital or speculum vaginal examination was done before the study, information on dilation and effacement was recorded. Information on cervical dilation and effacement otherwise was not recorded before induction of labor or at the onset of spontaneous labor. For women in either of the induction groups, the cervical status at study entry most likely represented the status of the cervix before induction of labor, because labor was induced shortly after randomization. For the others, the state of the cervix might have become more favorable as time passed. Labor was considered to have begun when the cervix became dilated to at least 3 cm or effaced to at least 80% and when uterine contractions were occurring at least every 5 minutes. Clinical chorioamnionitis was considered present when one or more of the following criteria were met: maternal temperature persistently above 37.5C on two or more occasions, at least 1 hour apart, or a single temperature above 38C before delivery; maternal white blood cell count greater than 20,000 cells/mm3; or foul-smelling amniotic fluid. Of the 5041 enrolled women, we excluded 13 who had prelabor cesarean deliveries. Among the 5028 remaining women, 4532 (90.1%) delivered vaginally and 496 (9.9%) by cesarean. Variables possibly affecting the
1032 Peleg et al
Predictors of Cesarean Delivery
need for cesarean delivery among normally laboring women were identified, and the association between those variables and mode of delivery was tested using univariate analyses (x2 test). The variables associated with cesarean delivery at P # .3 on univariate analysis and those variables for which there was previous evidence for an association with cesarean delivery were considered for the multivariate regression model. Because vaginal examinations were avoided in the TermPROM Study, there was a substantial amount of missing data on cervical dilation and effacement; thus those variables were not considered for the model. Instead, we combined cervical dilation and effacement, with mode of labor onset, as one variable (induced labor with an unfavorable cervix or induced labor with a favorable cervix versus spontaneous labor), and that variable, regarding which information was available for 4113 women, also was considered for the model. An unfavorable cervix was defined as a cervix dilated less than 3 cm and effaced less than 80%. The interaction term (induced versus spontaneous labor by duration of the latent period) was considered for the model. Multivariate analysis was done by stepwise logistic regression to determine the statistically significant (P , .05) independent predictors of cesarean delivery. Logistic regression analysis is a statistical method for evaluating the effects of multiple independent variables on a dichotomous dependent variable such as mode of delivery. The purpose of presenting the results of the univariate and multivariate analyses is to demonstrate the extent to which the effect of each independent variable was affected by the others.
Results The demographic information at study entry and perinatal outcomes, by country of birth, for the study population are detailed in Table 1. On univariate analysis, the variables that were significantly (P , .001) associated with an increase in cesarean delivery were birth weight of at least 4000 g; nulliparity; meconium staining; induced labor; use of oxytocin in labor; internal FHR monitoring; epidural anesthesia; clinical chorioamnionitis; duration of latent period of at least 12 hours; duration of active labor of 6 –12 hours, versus less than 6 hours, and of at least 12 hours, versus less than 6 hours; induced labor with an unfavorable cervix at study entry, versus spontaneous labor; and delivery in Canada, versus Israel. Variables significantly (P $ .001 and # .3) associated with an increase in cesarean delivery were cervical dilation of less than 3 cm at study entry (P 5 .015), previous cesarean delivery (P 5 .132), maternal age of at least 35 years (P 5 .135), and delivery in Canada, versus Sweden (P 5 .10), Denmark (P 5
Obstetrics & Gynecology
Table 1. Demographic Information at Entry and Perinatal Outcome United Kingdom (n 5 1341)
Sweden (n 5 105)
11.0 60.7 12.9 84.0
10.5 53.3 4.8 98.1
99.0 (n 5 984) 99.1 (n 5 891) 82.6
99.0 (n 5 103) 100.0 (n 5 103) 90.5
6.9 2.9 35.2 2.6 9.7
21.0 3.8 35.2 3.8 5.7
Maternal age $35 y* Nulliparity* Gestational age $41 wk* Vaginal digital or speculum examination* Cervical dilation ,3 cm*† Cervical effacement ,80%*† ,12 h of membrane rupture* GBS positive* Previous cesarean delivery* Epidural anesthesia Clinical chorioamnionitis Cesarean delivery
Denmark (n 5 28)
Australia (n 5 375)
Israel (n 5 331)
Canada (n 5 2848)
Overall (n 5 5028)
10.7 35.7 14.3 96.4
5.3 59.2 9.3 97.3
12.7 34.1 14.5 96.4
12.4 62.4 6.6 72.9
11.5 59.5 9.0 79.9
100.0 (n 5 27) 100.0 (n 5 27) 92.9
98.5 (n 5 323) 99.3 (n 5 292) 80.3
98.1 (n 5 308) 99.0 (n 5 304) 85.5
98.3 (n 5 1858) 98.0 (n 5 1717) 77.9
98.5 (n 5 3603) 98.6 (n 5 3334) 80.2
9.6 1.9 35.7 2.9 11.5
6.3 4.5 54.4 3.0 3.0
12.2 3.8 62.3 9.6 10.7
10.3 3.5 51.7 6.7 9.9
3.6 0 7.1 3.6 3.6
GBS 5 group B streptococcus culture result from introital or vaginal specimen obtained at entry. * Values given as % unless otherwise indicated. Demographic information at study entry. † Cervical dilation and effacement were determined by vaginal digital examination, if done, otherwise by vaginal speculum examination.
.221), and the United Kingdom (P 5 .30). The variables not significantly (P . .3) associated with cesarean delivery on univariate analysis were delivery in Australia, versus Canada (P 5 .67); induced labor with a favorable cervix, versus spontaneous labor (P 5 .96); cervical effacement of less than 80% at study entry (P 5 .394); and gestational age of at least 41 weeks (P 5 .305). All variables, other than cervical dilation and effacement, were considered for the multivariate regression model. On multivariate analysis, the strongest predictor of cesarean delivery among women with PROM at term was country of birth (Table 2). Women in Israel were significantly less likely to deliver by cesarean than were women in Canada (odds ratio [OR] 0.34; 95% confiTable 2. Independent Predictors of Cesarean Delivery
Country of birth Israel vs Canada Australia vs Canada Parity 0 vs $1 Active labor .12 h vs ,6 h 6 –12 h vs ,6 h Previous cesarean delivery Epidural anesthesia Clinical chorioamnionitis Internal fetal heart rate monitoring Birth weight $4000 vs ,4000 g Use of oxytocin in labor Maternal age $35 vs , 35 years Latent period $12 vs , 12 hours Meconium staining
OR
95% CI
P
0.34 1.93 2.81
0.15, 0.81 1.29, 2.88 1.95, 4.05
.01 .001 ,.001
2.78 1.66 2.75 2.66 2.42 2.19 2.07 1.97 1.44 1.41 1.41
2.01, 3.85 1.24, 2.22 1.46, 5.16 1.94, 3.65 1.80, 3.25 1.74, 2.75 1.50, 2.86 1.47, 2.65 1.03, 2.02 1.08, 1.85 1.03, 1.93
,.001 .001 .002 ,.001 ,.001 ,.001 ,.001 ,.001 .034 .013 .031
OR 5 odds ratio; CI 5 confidence interval.
VOL. 93, NO. 6, JUNE 1999
dence interval [CI] 0.15, 0.81), and women in Australia were significantly more likely to deliver by cesarean than were women in Canada (OR 1.93; 95% CI 1.29, 2.88). Other strong (OR greater than 2.5) predictors of cesarean delivery were nulliparity (OR 2.81; 95% CI 1.95, 4.05); duration of active labor exceeding 12 hours, versus less than 6 hours (OR 2.78; 95% CI 2.01, 3.85), previous cesarean delivery (OR 2.75; 95% CI 1.46, 5.16) and epidural anesthesia (OR 2.66; 95% CI 1.94, 3.65) (Table 2). Other less strong but statistically significant predictors were clinical chorioamnionitis; internal FHR monitoring; birth weight of at least 4000 g; use of oxytocin in labor; duration of active labor of 6 –12 hours, versus less than 6 hours; latent period of at least 12 hours, versus less than 12 hours; and meconium staining (Table 2). Gestational age at study entry, induced versus spontaneous labor, and induced labor with an unfavorable cervix or a favorable cervix versus spontaneous labor were not associated with cesarean delivery on multivariate analysis. There was no significant interaction between duration of latent period and induced versus spontaneous labor.
Discussion Much of the debate surrounding PROM at term has centered around whether induction of labor increases the risk of cesarean delivery. The TermPROM Study and meta-analyses of randomized controlled trials have not found a statistically significant difference in the rate of cesarean delivery between induction of labor with oxytocin or PGs and expectant management.3–7 Despite
Peleg et al
Predictors of Cesarean Delivery
1033
those results, we thought it appropriate to explore further the association between induced or spontaneous labor and cesarean delivery. After controlling for other fetal and maternal factors for which we had information, we did not find induced labor to be associated with a higher risk of cesarean delivery, regardless of duration of the latent period and status of the cervix at study entry. For women with PROM at term but otherwise without complications, the independent fetal and maternal risk factors for cesarean delivery were nulliparity, prolonged labor, previous cesarean delivery, epidural anesthesia, clinical chorioamnionitis, use of internal FHR monitoring, larger infants, use of oxytocin in labor, older maternal age, longer duration of latent period, and meconium staining. Except for duration of latent period, those risk factors are similar to those associated with cesarean delivery in women at term without PROM.8 –10 The association between duration of latent period and cesarean delivery might be due to subclinical chorioamnionitis, because the effect was independent of the development of clinical chorioamnionitis and we previously found longer latent periods to be associated with increased risk of clinical chorioamnionitis.11 We found that country of birth was a strong predictor of cesarean delivery, after controlling for fetal and maternal factors. Variations in rates of cesarean delivery were noted previously among countries, institutions, and physicians and nurses.8,12–15 In 1991–1992, the rates of cesarean delivery in community hospitals, without facilities for neonatal intensive care, in Ontario, Canada, varied from 6.6 to 30.3%.13 Those large variations in rates of cesarean delivery for different settings likely are due to differences in environmental factors including variations in quality and degree of supportive care provided to women in labor, supervision of labor and birth by midwives versus physicians, timing of admission to labor and delivery units, use of fetal scalp sampling or interpretation of FHR monitor tracings to diagnose unreassuring FHR, routine versus selective use of epidural anesthesia, availability of practitioners with expertise in operative vaginal delivery, and accepted upper limits for length of first and second stages of labor, all of which can have an effect on the likelihood of cesarean delivery.16 –22 It is not known specifically why Israel had a significantly lower rate of cesarean delivery than Canada, but this merits further investigation. This study was a retrospective analysis of data, so there might be other fetal and maternal factors associated with cesarean delivery after PROM concerning which we had no or inadequate information. We had no information on environmental factors other than coun-
1034 Peleg et al
Predictors of Cesarean Delivery
try of birth. Future prospective research should involve evaluation of changes in environment, while controlling for fetal and maternal factors. The ongoing Nursing SCIL Trial, funded by the National Institutes of Health, coordinated in Toronto by EDH, is an example of such an evaluation. The Nursing SCIL Trial is a multicenter, randomized controlled trial evaluating the effect of continuous, supportive care in labor by nurses in institutions in the United States and Canada. The present study confirmed that the fetal and maternal risk factors for cesarean delivery for women with PROM at term are similar to those for women with other uncomplicated term pregnancies. After country of birth was controlled for, the greatest risks were for nulliparous women, those with long labors, those who had previous cesarean deliveries, and those receiving epidural anesthesia. The risk, however, was not increased if labor was induced. An approach that decreases the duration of the active phase of labor and the use of epidural anesthesia, particularly among nulliparous women, might decrease the rate of cesarean delivery in women with PROM at term.
References 1. Wagner MV, Chin VP, Peters CJ, Drexler B, Newman LA. A comparison of early and delayed induction of labor with spontaneous rupture of the membranes at term. Obstet Gynecol 1989;74: 93–7. 2. Grant JM, Serle E, Mahmood T, Sarmandal P, Conway DI. Management of prelabour rupture of the membranes in term primigravidae: Report of a randomized prospective trial. Br J Obstet Gynaecol 1992;99:557– 62. 3. Hannah ME, Ohlsson A, Farine D, Hewson SA, Hodnett ED, Myhr TL, et al. Induction of labor compared with expectant management for prelabor rupture of the membranes at term. TermPROM Study Group. N Engl J Med 1996;334:1005–10. 4. Mozurkewich EL, Wolf FM. Prelabor rupture of membranes at term: A meta-analysis of three management schemes. Obstet Gynecol 1997;89:1035– 43. 5. Tan B, Hannah ME. Oxytocin for prelabour rupture of the membranes at or near term. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. 6. Tan B, Hannah ME. Prostaglandins for prelabour rupture of the membranes at or near term. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. 7. Tan B, Hannah ME. Prostaglandins vs. oxytocin for prelabour rupture of the membranes at or near term. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. 8. Goyert GL, Bottoms SF, Treadwell MC, Nehra PC. The physician factor in cesarean birth rates. N Engl J Med 1989;320:706 –9. 9. Cnattingius R, Cnattingius S, Notzon FC. Obstacles to reducing cesarean rates in a low-cesarean setting: The effect of maternal age, height, and weight. Obstet Gynecol 1998;92:501– 6. 10. Lieberman E, Lang JM, Cohen A, D’Agostino R, Datta S, Frigoletto FD. Association of epidural analgesia with cesarean delivery in nulliparas. Obstet Gynecol 1996;88:993–1000. 11. Seaward PG, Hannah ME, Myhr TL, Farine D, Ohlsson A, Wang EE, et al. International Multicenter Term Prelabor Rupture of Membranes Study: Evaluation of predictors of clinical chorioam-
Obstetrics & Gynecology
12. 13.
14. 15.
16.
17. 18.
19.
20.
nionitis and postpartum fever in patients with prelabor rupture of membranes at term. Am J Obstet Gynecol 1997;177:1024 –9. Notzon FC. International differences in the use of obstetric interventions. JAMA 1990;263:3286 –91. Hospital-specific and regional-level utilization information on cesarean sections, appendectomy and breast cancer surgery. In: Naylor CD, Anderson GM, Goel V, eds. Patterns of health care in Ontario. Ottawa: Canadian Medical Association, 1994;137– 42. Learman LA. Regional differences in operative obstetrics: A look to the south. Obstet Gynecol 1998;92:514 –9. Radin TG, Harmon JS, Hanson DA. Nurses’ care during labor: Its effect on the cesarean birth rate of healthy, nulliparous women. Birth 1993;20:14 –21. Hodnett ED. Support from caregivers during childbirth. (Cochrane Review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998. Harvery S, Jarrell J, Brant R, Stainton C, Rach D. A randomized, controlled trial of nurse-midwifery care. Birth 1996;23:128 –35. McNiven P, Williams IJ, Hodnett K, Kaufman K, Hannah M. The effect of early labour assessment on cesarean section rates for low risk nulliparous women. Birth 1998;25:5–10. Thacker SB, Stroup DF. Continuous electronic fetal heart monitoring during labor. (Cochrane Review) In: the Cochrane Library; Issue 3. Oxford: Update Software, 1998. Howell CJ. Epidural vs non-epidural analgesia in labour. (Cochrane review) In: The Cochrane Library; Issue 3. Oxford: Update Software, 1998.
VOL. 93, NO. 6, JUNE 1999
21. Hannah WJ, workshop participants. The Canadian consensus on breech management at term. J SOGC 1994;16:1839 –58. 22. Maresh M, Choong KH, Beard RW. Delayed pushing with lumbar epidural analgesia in labour. Br J Obstet Gynaecol 1983;90:623–7.
Address reprint requests to:
David Peleg, MD c/o Mary E. Hannah, MDCM Maternal, Infant and Reproductive Health Research Unit Room 751 University of Toronto 790 Bay Street Toronto, ON M5G 1N8 Canada E-mail: [email protected]
Received August 31, 1998. Received in revised form November 25, 1998. Accepted December 10, 1998. Copyright © 1999 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Peleg et al
Predictors of Cesarean Delivery
1035