Miller, Emily S. MD, MPH; Grobman, William A. MD, MBA
Cesarean rates have been well documented to be rising in both developed and developing countries. There are a myriad of reasons for this increase: changes in the medical–legal climate, patient expectations, and medical practice, as well as increased rates of risk factors for cesarean delivery, such as maternal obesity, multiple gestations, and advanced maternal age.1 At present, the cesarean delivery rate in the United States is approximately 30%.2
In general, women who already have had a vaginal delivery have a relatively low chance of cesarean delivery.3 Yet, it is not well understood whether and to what degree the interval between a previous vaginal delivery and a subsequent pregnancy is associated with the risk of cesarean delivery. Anecdotally, on our labor and delivery ward, it has appeared that women with longer interdelivery intervals have a greater chance of cesarean delivery. One observational study, based on data from a Michigan-wide database, demonstrated a small but statistically significant increase in rates of “labor dystocia” with increasing interpregnancy intervals.4 However, several variables, such as maternal obesity and birth weight relative to previous deliveries, that may have confounded the association between interdelivery interval and dystocia, were not included in the analysis. Also, this analysis focused on “labor dystocia” as defined by International Classification of Diseases, 9th Revision coding, which included many different types of diagnoses (eg, delayed delivery) and was not concerned with cesarean delivery specifically. In a small case–control study from Nigeria, the chance of cesarean delivery did not differ between women with interbirth intervals of less than 6 years and those with intervals of at least 6 years.5 The sample size in this study, however, was such that type II error could be responsible for the lack of observed association. Because of these conflicting results as well as the paucity of data available, the relationship between interbirth interval and cesarean delivery remains poorly understood.
In this study, we have explored the association between interdelivery interval and frequency of cesarean delivery. We hypothesized that increasing intervals between deliveries, independent of other factors, would be associated with an increased probability of cesarean delivery.
MATERIALS AND METHODS
This was a retrospective cohort study of all multiparous women who delivered at Northwestern Memorial Hospital from June 1, 2006, through December 31, 2007. These women were identified through use of a perinatal database that documents parity in addition to multiple other patient characteristics. Women were included in the present analysis if, in addition to having had a previous vaginal delivery, they had a singleton in the vertex presentation at greater than 34 weeks of gestation. Women who presented for and underwent a scheduled cesarean delivery, including those who arrived before their scheduled date because of early labor or rupture of membranes but who did not desire a trial of labor, were excluded. Women with contraindications to a trial of labor (ie, placenta previa, previous cavity-entering myomectomy) or with anomalous fetuses also were excluded. Of note, women with a previous cesarean and a previous vaginal delivery were included in the analysis.
The charts of those women meeting inclusion criteria were abstracted for demographic information including length of time between the last vaginal delivery and the present delivery. To calculate this interval, the date of the most recent previous vaginal delivery was abstracted from the medical record. If only the month was recorded (and not the exact date), the delivery date was recorded as the 15th of that month. Similarly, if the year was available but not the month, the delivery date was taken to be June 15 of that year. If none of this information was available, the woman was excluded from analysis. Other demographic data, as well as the route of delivery during the admission within the study period, were recorded as well.
Women were grouped according to 5-year increments of interbirth interval (ie, 0–4 years, 5–9 years, 10–14 years, and at least 15 years), and the primary analysis used this categorical exposure variable.
To account for changes in birth weight between pregnancies, a categorical variable (Δ birth weight) was created to represent a clinically significant difference in birth weights. It was defined as being present when the birth weight in the current pregnancy was at least 1 pound (454 g) greater than that of a woman's largest previous vaginal delivery. Body mass index (BMI, calculated as weight (kg)/[height (m)]2) at delivery was categorized as nonobese (less than 30), class 1 obesity (30.0–34.9), or class 2–3 obesity (35 or more).
Other potential confounders examined included previous cesarean delivery, number of previous vaginal deliveries, induction of labor (compared with spontaneous labor), gestational or pre-existing diabetes mellitus, preeclampsia, cervical examination on admission (including dilation, effacement, and station), oligohydramnios, and the use of assisted reproductive technology for conception.
Univariable comparisons between groups were performed using the Student t test and χ2 analysis for continuous and categorical variables, respectively. A χ2 test of trend (EpiInfo 3.3.2) was also performed to assess for the presence of a trend in the association between increasing interbirth interval and cesarean delivery.
The independence of the association between interbirth interval and cesarean delivery then was evaluated by multivariable logistic regression. The “interbirth interval” variables were entered into the final regression as multiple dichotomous variables with the “0–4 year” interval as the referent group. Maternal age also was included in the final regression given its previously reported association with risk of cesarean delivery and its potential confounding relationship with interbirth interval. This was included as a categorical variable (less than 25 years, 25–34 years, and 35 years or more). Other variables were assessed in the regression if they had an association with cesarean delivery at a P<.10 in univariable analysis. All variables that retained significance at a level of P<.05 were retained in the final regression, which was a binary logistic regression using the maximum likelihood estimate technique. Odds ratios and 95% confidence intervals for each of the variables in the final regression were calculated. All tests were two-tailed.
Additional regressions were performed to estimate whether the associations were robust to changes in the composition of the study population (ie, for a population with women with previous cesarean deliveries excluded) or the form (continuous compared with categorical) in which the variables were entered into the regression. Analyses were performed with Minitab 14. The Institutional Review Board of Northwestern University approved this study.
During the period of study, 2,528 multiparous women presented to labor and delivery and met the inclusion criteria. Of these women, 66 (2.6%) delivered via cesarean delivery. The clinical characteristics of the population and their relationship with cesarean delivery are presented in Table 1. As illustrated, women whose labor resulted in cesarean delivery were heavier, more likely to have had a previous cesarean, and more likely to have a neonate who was at least 1 lb larger than their largest previous neonate who was delivered vaginally. Women with interbirth intervals of less than 5 years, 5 to 9 years, 10 to 14 years, and at least 15 years had cesarean delivery rates of 1.9%, 5.0%, 7.7%, and 13.3%, respectively. The increase of these frequencies demonstrated a significant trend (P<.001).
Of these cesarean deliveries, 48.5% were done for a nonreassuring fetal heart tracing. The remaining 51.5% were performed for protracted or arrested labor (30.3% arrest of dilation, 21.2% arrest of descent). The distribution of these indications, stratified by interdelivery interval, is shown in Table 2. As can be noted, the distribution for the indication for cesarean delivery during each interdelivery interval was similar.
The results of the multivariable regression that evaluated the relationship of cesarean delivery, interdelivery interval, and other covariates are shown in Table 3. Variables that remained significantly associated with cesarean delivery included BMI, previous cesarean delivery, and Δ birth weight. Of particular note, even after controlling for these variables, interbirth interval continued to be significantly associated with risk of cesarean delivery, with each additional 5-year interval between deliveries having an increasingly higher odds ratio. Two additional regressions were performed. One regression used the same variables but was based on a population that excluded all women with a previous cesarean delivery. The associations that had previously been found to be significant (with the exception of previous cesarean delivery) remained unchanged (data not shown). Additionally, BMI, age, and Δ birth weight were entered into the equation as continuous variables. Once again, the association of interdelivery interval with cesarean delivery remained unchanged.
In this study, we have found that increasing interbirth interval is associated with a greater chance of cesarean delivery, independent of other risk factors for cesarean delivery. This trend reaches a more than 7-fold increased risk for women whose interbirth interval is 15 years or greater. This trend cannot simply be explained by the association of maternal age with interdelivery interval or by changes in other maternal characteristics (such as BMI), as controlling for these factors did not eliminate the association between interbirth interval and cesarean delivery.
The reasons for this association remain uncertain. The exact physiologic or anatomic changes or both that occur after an initial vaginal delivery, and that result in marked shortening in the lengths of both the first and second stages of labor, as well as a decrease in the subsequent chance of cesarean delivery, are not well understood.6–8 It is possible, for example, that these changes are not indelible, in which case one would expect a steady reversion of the epidemiologic characteristics of parturition toward those of a nulliparous woman.2,9–11 Indeed, this steady increase in the chance of cesarean delivery with increasing interdelivery interval is what is seen. Nevertheless, the reason for the observed association, including the potential of a biologic basis, cannot be known from this study.
Limitations of this study include its observational design. Not only can the reasons for the association not be known, but causality cannot be inferred. Furthermore, these data are from a population within a single institution. Thus, generalizability to other populations, both within the United States as well as internationally, cannot be certain. This difference in patient populations, as well as the difference in sample size, may help to explain the lack of association between interbirth interval and cesarean delivery that was found by Orji and colleagues in their study of Nigerian women.5 Although we attempted to control for likely confounding factors, it is possible that other factors that are unaccounted for could exist and account for the observed association.
It is also possible that the observed association is not reflective of differences within patients, but reflective of differences in approaches of physicians. An obstetrician's decision to perform a cesarean delivery is influenced by many factors, including some criteria with a subjective component.12,13 For example, it is possible that knowing the interdelivery interval predisposed obstetricians toward choosing cesarean delivery. However, we do not believe this is likely. As previously noted, there has not been clear evidence for the association between interdelivery interval and cesarean delivery, thus making it less likely that this factor systematically influenced physicians' decisions regarding mode of delivery.
Our findings corroborate those from the study that used administrative data from linked records of births and hospital discharges in Michigan to demonstrate an association between interbirth interval and increased rates of labor dystocia.4 The present study extends those findings by documenting the relationship between interdelivery interval and cesarean delivery specifically. Indeed, it appears that cesarean deliveries are equally likely to be attributed to nonreassuring fetal status as to labor dystocia, as evidenced by the data in Table 2. Moreover, by accounting for other potential confounding factors, as well as by directly abstracting and confirming data from the primary medical record, the present study lessens the possibility of confounding and ascertainment bias.
This study estimated the relationship between interbirth interval and cesarean delivery. In our cohort, longer interbirth intervals were associated with a marked increase in the risk of cesarean delivery. However, it should be noted that this risk reached only 13.3% in women with an interbirth interval of greater than 15 years; accordingly, although their risk for cesarean delivery may be higher than that for women with shorter interdelivery intervals, this risk still remains well below that of women attempting a first vaginal birth.11
1. Ecker JL, Frigoletto FD Jr. Cesarean delivery and the risk-benefit calculus. N Engl J Med 2007;356:885–8.
3. Yeast JD, Jones A, Poskin M. Induction of labor and the relationship to cesarean delivery: a review of 7001 consecutive inductions. Am J Obstet Gynecol 1999;180(3 pt 1):628–33.
4. Zhu BP, Grigorescu V, Le T, Lin M, Copeland G, Barone M, et al. Labor dystocia and its association with interpregnancy interval. Am J Obstet Gynecol 2006;195:121–8.
5. Orji EO, Shittu AS, Makinde ON, Sule SS. Effect of prolonged birth spacing on maternal and perinatal outcome. East Afr Med J 2004;81:388–91.
6. Huey JR Jr, Al-Hadjiev A, Paul RH. Uterine activity in the multiparous patient. Am J Obstet Gynecol 1976;126:682–6.
7. Arulkumaran S, Gibb DM, Lun KC, Heng SH, Ratnam SS. The effect of parity on uterine activity in labor. Br J Obstet Gynaecol 1984;91:843–8.
8. Friedman EA. Labor in multiparas; a graphicostatistical analysis. Obstet Gynecol 1956;8:691–703.
9. Zhu BP, Rolfs RT, Nangle BE, Horan JM. Effect of the interval between pregnancies on perinatal outcomes. N Engl J Med 1999;340:589–94.
10. American College of Obstetricians and Gynecologists Task Force on Cesarean Delivery. Evaluation of cesarean delivery. Washington, DC: American College of Obstetricians; 2000:1–59.
11. Coonrod DV, Drachman D, Hobson P, Manriquez M. Nulliparous term singleton vertex cesarean delivery rates: institutional and individual level predictors. Am J Obstet Gynecol 2008;198:694.e1–11.
12. Murthy K, Grobman WA, Lee TA, Holl JL. Association between rising professional liability insurance premiums and primary cesarean delivery rates. Obstet Gynecol 2007;110:1264–9.
13. Althabe F, Belizán JM, Villar J, Alexander S, Bergel E, Ramos S, et al. Mandatory second opinion to reduce rates of unnecessary caesarean sections in Latin America: a cluster randomised controlled trial. Lancet 2004;363:1934–40.