Induction of labor has been shown to be associated with an increased risk for cesarean birth in a variety of circumstances.1–4 Typically, the increase is considered justified when the benefits to mother or infant from delivery outweigh the maternal risk from the surgery.
Maternal age over 35 years is associated with an increased risk of having a stillborn fetus.5–12 Although a number of these fetal deaths are attributable to conditions such as hypertension and diabetes, which are more prevalent among older gravidas, a significant proportion remain unexplained.11,12 Attempts to reduce stillbirth fetus risk typically rely on antepartum fetal testing; however, any potential benefit of testing must be weighed against the risk of interventions that occur in response to positive tests. In the case of antepartum tests for fetal well-being, nonreassuring fetal tests will likely lead to delivery. As part of an ongoing project to determine the usefulness and optimal timing of antepartum testing to prevent stillborn fetuses in older women, we needed to know the risk of cesarean delivery, which would be the major risk encountered with positive tests, under the conditions that might be encountered in the proposed testing. We therefore performed a retrospective cohort study to estimate the impact of labor induction, gestational age, parity, and maternal age on the risk of cesarean delivery in a large, diverse population of pregnant women who were eligible for a trial of labor and whose care was provided in multiple practice models.
MATERIALS AND METHODS
Our sample was collected from Beth Israel Deaconess Medical Center and Brigham and Women's Hospital in Boston, the two largest obstetric services in the city. The delivery services at each of these two institutions are comprised of a variety of practice models, including resident practices, maternal–fetal medicine services, midwifery services caring for community health center patients and health maintenance organization patients, and conventional private practices staffed by one to six physicians. We identified all singleton deliveries that occurred at or after 36 completed weeks' gestation from the comprehensive birth record databases maintained by each hospital. The study population included all deliveries from January 1, 1998, through December 31, 1998, at Brigham and Women's Hospital and from January 1, 1998, through December 31, 1999, at Beth Israel Deaconess Medical Center. The human research committees of both institutions approved the study.
Data in the Brigham and Women's Hospital database were abstracted from the medical record by trained abstractors and entered into a research quality database. Data in the Beth Israel Deaconess Medical Center database were entered by the obstetrician at the time of delivery. Definitions of variables of interest in this study were comparable between institutions.
To obtain a sample of pregnancies eligible for vaginal delivery, we excluded deliveries that met one or more of the following criteria: history of cesarean delivery or myomectomy, breech or other abnormal presentation, placenta previa, active maternal herpes infection, prolapsed cord, fetal anomaly, and no trial of labor (due to maternal medical contraindication or refusal or reason unspecified). Exclusions are summarized in Table 1. Of 17,684 total deliveries, 14,409 (82%) were eligible for inclusion in our analyses. Forty-seven percent of the deliveries in the sample occurred at Brigham and Women's Hospital and 53% occurred at Beth Israel Deaconess Medical Center. Characteristics of the sample are shown in Table 2.
Gestational age was defined as completed weeks calculated from the best estimate of the date of confinement. Maternal age was completed years at delivery. Labor was classified as induced if the delivery record indicated an induction or if pitocin or prostaglandins were administered before a cervical dilatation of 4 cm. Pregnancy-induced hypertension included both proteinuric and nonproteinuric hypertension as well as the preeclampsia syndromes.
Before combining the data from the two institutions, we confirmed that the associations of induction of labor, gestational age, and maternal age group with cesarean delivery were similar in the two subpopulations.
We obtained person-level sociodemographic data from Brigham and Women's Hospital. This information was not available for the women delivering at Beth Israel Deaconess Medical Center, but we describe the sociodemographic characteristics of the entire birth cohort at the hospital during the time of the study. We then analyzed the effects of marital status, Medicaid, and race on cesarean delivery among the cases just from Brigham and Women's Hospital. Nulliparous and multiparous deliveries were modeled separately.
Wald statistics13 were calculated for each coefficient and 95% confidence intervals (CIs) were constructed for the odds ratios (ORs) in each of the unadjusted analyses.
In multivariate logistic regression analyses, we modeled the probability of cesarean delivery as a function of spontaneous or induced labor, adjusted for the variables found to be significant in the univariate analyses (gestational age, maternal age group, birth weight, hypertension, diabetes, hospital, and year of delivery; data not shown but available on request). We included Medicaid and race in additional analyses using only the cases from Brigham and Women's Hospital. Nulliparous and multiparous deliveries were modeled separately. Wald statistics were calculated for each coefficient and 95% CIs were constructed for the ORs. Terms for the interaction between labor and maternal age group were included initially but only retained in the final analysis if their coefficients were significant at P ≤ .05 or if they had an impact on the coefficients of other predictors in the model. In the case of variables that had more than two possible values (eg, gestational age) we made no assumptions about possible trends and modeled all predictors using dummy variables.
We also examined the distribution of indications for cesarean delivery among nulliparas. Indications for cesarean delivery were classified as failed induction, failure to progress, nonreassuring fetal status, failure to progress and nonreassuring fetal status, or other indication. A failed induction was defined as a cesarean delivery performed before the cervix was dilated to 4 cm in the absence of nonreassuring fetal status for patients delivering at Brigham and Women's Hospital. A failed induction at Beth Israel Deaconess Medical Center was so noted by the delivering obstetrician. Failure to progress was assigned as an indication if arrest of dilatation occurred after 4 cm or arrest of descent was recorded in the patient's record.
Cochran–Mantel–Haenszel statistics were calculated to test the association between labor induction and each indication for cesarean delivery, stratified by gestational age. All statistical analyses were performed using SAS 8.02 (SAS Institute Inc., Cary, NC).
Overall, the primary cesarean delivery rate in the cohort was 10.8%. The difference in cesarean delivery rates between the two hospitals was a function of different rates of assisted vaginal delivery. The proportions of women in both institutions who delivered spontaneously were virtually identical (80.3% versus 80.0%).
Among the 7372 nulliparous women in the sample, 2227 (30%) had induced labor. Induced nulliparas had a 24.7% cesarean delivery rate, compared with a 13.7% cesarean delivery rate among spontaneously laboring nulliparas.
There were 7027 multiparous women in the sample, of whom 1920 (27%) had induced labor. The rate of cesarean delivery was 4.5% among induced multiparas, compared with 2.4% among those with spontaneous labor. In both nulliparous and multiparous women, the unadjusted effect of labor induction on the probability of cesarean delivery was significant (P < .001).
The effect of gestational age on the probability of cesarean delivery is illustrated in Figure 1. Among multiparas, gestational age has little impact on the risk of cesarean delivery for either induced or spontaneously laboring women—the curves for both of these groups are relatively flat for all gestational ages. In contrast, gestational age has a significant association with the rate of cesarean delivery among nulliparous women. For spontaneously laboring nulliparas, the probability of cesarean delivery is lowest between 36 and 40 weeks and then rises dramatically after 40 weeks. For induced nulliparas, the cesarean delivery rate is lowest between 36 and 38 weeks and then rises, with the largest increment after 40 weeks.
In the multivariate analysis (Table 3), which controlled for maternal age group, gestational age at delivery, presence of maternal diabetes and hypertension, birth weight, hospital, and year, labor induction was significantly associated with a higher probability of cesarean delivery among nulliparous women, with an adjusted OR of 1.70 (95% CI 1.48, 1.95). Black and Hispanic women and women whose race was defined as American Indian, other, or unknown had a greater probability of cesarean delivery than white women (respective ORs 2.37 [95% CI 1.68, 3.43], 2.09 [95% CI 1.39, 3.14], and 1.57 [95% CI 1.08, 2.27]). Asian race did not impact the cesarean delivery rate.
The effect of labor induction was slightly attenuated though still significant among multiparous women, with an adjusted OR of 1.49 (95% CI 1.10, 2.00). The inclusion of sociodemographic variables did not alter the signs on the coefficients for the other predictors in the model, although the magnitude of some ORs changed slightly. Terms for the interaction between labor induction and maternal age group were not statistically significant and had no impact on other covariates, and were therefore excluded from the final multivariate logistic regression models.
Maternal age over 35 years was associated with an increased risk for cesarean delivery among nulliparas, and maternal age over 40 years was associated with an increased risk in multiparas. Consistent with the data shown in Figure 1, the relative risk for cesarean birth varied by gestational age. In nulliparas, gestational age below 39 weeks was associated with a decreased risk relative to term (40 weeks), whereas gestational age at or after 41 weeks was associated with an increased risk. In multiparas, cesarean delivery risk was increased only at 41 weeks.
The indications for cesarean delivery in each labor group and gestational age in the nulliparas are presented in Table 4. Induction increased the frequency of cesarean delivery for nonreassuring fetal status (P < .01) and nonreassuring fetal status plus failure to progress (P < .01) independent of gestational age. Failed induction was the indication for cesarean delivery in 3.3% of all cesarean deliveries and was independent of gestational age. The frequency with which cesarean delivery was performed because of failure to progress was much higher in the induced patients and increased with gestational age (P < .001).
Our study demonstrates that labor induction, parity, maternal age, and birth weight extremes played significant roles in determining risk for cesarean delivery in our cohort of ethnically diverse and relatively low-risk obstetric patients. Women with hypertension and diabetes also had an increased risk for cesarean delivery but did not contribute substantially to the overall cesarean delivery rate because of their low prevalence. Induction was associated with a 70% increase in the risk of cesarean delivery among nulliparas and a 50% increase among multiparas who were eligible for a trial of labor. Although the relative risks are similar in nulliparous and multiparous women, the overall cesarean delivery rate in multiparas is so much lower that the absolute risk increased only 2.1%. The magnitude of the increase in risk for cesarean delivery with induction in nulliparas agrees very well with that recently reportedly by Seyb et al.2 In their study population, elective induction of labor was associated with a 90% increase in cesarean delivery rate and medically indicated induction was associated with a 70% increase.
Our data are also remarkably similar to those of Seyb et al2 with regard to the influence of gestational age on cesarean delivery risk in nulliparas. Both studies find that, overall, rates of cesarean delivery in nulliparas are lowest between weeks 36 and 40 and rise significantly after 40 weeks. In our more detailed analyses of nulliparas, the cesarean delivery rates are lowest between 36 and 40 weeks for the spontaneously laboring patients, whereas the rate begins to rise at 39 weeks in the induced patients. Multiparas in our study do not demonstrate any substantive effects of gestational age on cesarean delivery rates.
Because enforced practice guidelines proscribe elective induction before the 39th week of gestation in both Brigham and Women's Hospital and Beth Israel Deaconess Medical Center, the inductions performed between 36 and 38 weeks in our study were likely to be medically indicated. These women would be expected to have the highest proportion of unfavorable cervices. The fact that the cesarean delivery rate was no higher in these pregnancies than in those of the term patients suggests that current induction techniques, which routinely involve the use of cervical ripening agents, are successful.14–16 This observation is further supported by the finding that the proportion of failed inductions in nulliparas remained steady at 2% of all deliveries independent of gestational age. Fetal size, as indicated by the increasing proportion of women undergoing cesarean delivery for failure to progress, may play a significant role in the increased cesarean delivery rate after 38 weeks' gestation.
The effect of advancing maternal age in increasing the cesarean delivery rate is not surprising, as an increase has been reported in the majority of studies assessing this outcome. A number of factors have been hypothesized to contribute to this increase, including disproportionate numbers of large for gestational age and small for gestational age infants, uterine dysfunction, and a lower threshold among patients and providers to perform a cesarean delivery in older mothers.17–20
The strengths of our study include the size and diversity of the population studied and its multiinstitutional basis. Although the cesarean delivery rates differed between the two hospitals, they varied in inverse proportion to the operative delivery rates and likely reflect differences in the approach to the second stage of labor. Such differences in practice style are likely to be encountered in institutions with a number of provider models. The fact that the initial unadjusted analyses of the variables of interest for each institution were similar speaks to the robustness of the observations.
We did not stratify our data by elective versus medical indications for induction or by cervical dilatation at admission because our goal was to produce overall baseline rates by gestational age for use in decision making. Seyb et al2 reported similar risks for cesarean delivery in elective and medically indicated induction, suggesting that stratification by type of induction would not substantially change our point estimates. We believe that the method we have used for estimating cesarean delivery risk is the most appropriate for determining the downside impact of introducing antepartum testing to prevent unexplained stillborn fetuses in older gravidas.
1. Macer JA, Macer CI, Chan LS. Elective induction versus spontaneous labor: A retrospective study of complications and outcome. Am J Obstet Gynecol 1992;166:1690–7.
2. Seyb ST, Berka RJ, Socol ML, Dooley SL. Risk of cesarean delivery with elective induction of labor at term in nulliparas. Obstet Gynecol 1999;94:600–7.
3. Maslow AS, Sweeny AL. Elective induction of labor as a risk factor for cesarean delivery among low-risk women near term. Obstet Gynecol 2000;95:917–22.
4. Crowley P. Interventions for preventing or improving outcome of delivery at or beyond term (Cochrane review). Cochrane Database Syst Rev 2000;2:CD000170.
5. Naeye RL. Maternal age, obstetric complications, and the outcome of pregnancy. Obstet Gynecol 1983;61:210–6.
6. Forman MR, Meirik O, Berendes HW. Delayed childbearing in Sweden. JAMA 1984;252:2135–9.
7. Kiely JL, Paneth N, Susser M. An assessment of the effects of maternal age and parity in different components of perinatal mortality. Am J Epidemiol 1986;123:444–54.
8. Cnattingius S, Forman MR. Berendes HW, Isotalo L. Delayed childbearing and risk of adverse perinatal outcome: A population-based study. JAMA 1992;268:886–90.
9. Cnattingius S, Berendes HW, Forman MR. Do delayed childbearers face increased risk of adverse pregnancy outcomes after the first birth? Obstet Gynecol 1993;81:512–6.
10. Fretts RC, Schmittdiel J, McLean FH, Usher RH, Goldman MB. Increased maternal age and the risk of fetal death. N Engl J Med 1995;333:953–7.
11. Huang DY, Usher RH, Kramer MS, Yang H, Morin L, Fretts R. Determinants of unexplained antepartum fetal deaths. Obstet Gynecol 2000;95:215–21.
12. Froen JF, Armestad M, Frey K, Vege A, Saugstad OD, Stray-Pedersen B. Risk factors for sudden intrauterine unexplained death: Epidemiological characteristics of singleton cases in Oslo, Norway, 1986–1995. Am J Obstet Gynecol 2001;184:694–702.
13. Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons, 1989.
14. Hofmeyr GJ, Gulmezoglu AM. Vaginal misoprostol for cervical ripening and induction of labour (Cochrane review). Cochrane Database Syst Rev 2003;1:CD000941.
15. Kelly AJ, Kavanagh J, Thomas J. Vaginal prostaglandin (PGE2 and PGF2α) for induction of labour at term (Cochrane review). Cochrane Database Syst Rev 2001;2:CD003101.
16. Wing DA, Rahall A, Jones MM, Goodwin TM, Paul RH. Misoprostol: An effective agent for cervical ripening and labor induction. Am J Obstet Gynecol 1995;172:1811–6.
17. Martel M, Wacholder S, Lippman A, Brohan J. Maternal age and primary cesarean section rates: A multivariate analysis. Am J Obstet Gynecol 1987;156:305–8.
18. Leich CR, Walker JJ. The rise in caesarean section rate: The same indications but a lower threshold. Br J Obstet Gynaecol 1998;105:621–6.
19. Rosenthal AN, Paterson-Brown S. Is there an incremental rise in the risk of obstetric intervention with increasing maternal age? Br J Obstet Gynaecol 1998;105:1064–9.
20. Bell JS, Campbell DM, Graham WJ, Penney GC, Ryan M, Hall MH. Do obstetric complications explain high caesarean section rates in women over 30? A retrospective analysis. BMJ 2001;322:894–5.