Introduction

The preterm birth (PTB) rate has had a marginal decline and most recently was noted to account for 11.4% of all deliveries in 2013 down from 11.6% in 2012. This small decline is likely, in part, due to effective strategies and interventions that help decrease PTB among women found to be at highest risk. However, many women with a spontaneous PTB (sPTB) present without any known or identifiable risk factor. This highlights the importance of continued research to investigate other etiologies and risk factors for sPTB.

Our previous work demonstrated an increased risk of sPTB in a subsequent pregnancy when a full-term cesarean delivery is performed in the second stage of labor as compared to when it is performed in the first stage of labor or compared to women with a term vaginal delivery. Prior to this work, a term cesarean delivery in the second stage of labor was not identified as a risk factor for sPTB.

In addition to the mode of delivery in the second stage of labor, a prolonged second stage of labor has been hypothesized to be a risk factor for cervical insufficiency. During the second stage of labor, it is plausible that the pressure of the fetal head against a completely dilated cervix may lead to changes in the cervical integrity and cervical strength. Therefore, a longer second stage of labor, regardless of mode of delivery, may increase the risk for structural cervical injury lending it susceptible to premature dilation and sPTB in a subsequent pregnancy. This theory, however, has not yet been investigated.

The objective of this study was to specifically evaluate the effect that the length of second stage of labor in one pregnancy has on the risk of sPTB in a subsequent pregnancy. Our hypothesis is that women with a longer second stage of labor have an increased risk of subsequent sPTB.

Materials and Methods

This was a planned secondary analysis of a large retrospective cohort study of women with 2 consecutive deliveries at the Hospital of the University of Pennsylvania from 2005 through 2010. The original study compared women with a term (≥37 weeks) induction to term spontaneous labor and evaluated rates of subsequent sPTB. Institutional review board approval was obtained prior to initiating the study.

Our hospital-based electronic database was used to identify women with >1 delivery from 2005 through 2010. The first pregnancy during this time period was considered the index pregnancy, which may or may not equate to the woman’s first pregnancy, as multiparous women were included. The second consecutive pregnancy during this time period was considered to be the subsequent pregnancy if delivery occurred at ≥16 weeks’ gestation. Patients were included in the study only once. Women were excluded from the parent study if they had a PTB in their index pregnancy of if they had a known history of PTB. Only women who reached the second stage of labor were included in the analysis of our current study.

The primary outcome was sPTB, defined as spontaneous labor and delivery or preterm premature rupture of membranes <37 weeks’ gestation. We first evaluated the overall length of the second stage of labor and its effect on sPTB rates in a subsequent pregnancy. We then evaluated various cut-points for length of the second stage of labor and the individual rates of sPTB for those cut-points (<1, <2, ≥2, ≥3, and ≥4 hours). Based on distribution of data, our sample size, and clinical utility, a prolonged second stage was defined as ≥3 hours. The rates of sPTB in a subsequent pregnancy were then compared between women with a second stage that was ≥3 hours and women with a second stage <3 hours.

Bivariate analyses were performed using χ ² and Fisher exact tests to compare categorical data. Student t tests and Mann Whitney U tests were used to compare parametric and nonparametric data, as appropriate. Linear and logistic regression were used to calculate odds ratios (OR). Multivariable logistic regression was used to adjust for confounders. To build our multivariable model, we included risk factors that had an association with the exposure or outcome at a level of P < .2. These identified risk factors were evaluated to be confounders or effect modifiers of the model. Any risk factor that was found to have an effect size of at least 10% was retained in the model as a confounder. Using backwards stepwise elimination, we created our final parsimonious model. Confounders maintained in our model were mode of delivery in the index pregnancy and parity. While race was not found to be a confounder, it was retained in the final model given the biological plausibility of being associated with both the exposure and outcome.

Data analysis was performed using software (STATA 12.0 for Windows; STATA Corp, College Station, TX). Statistical significance was set at P < .05. There was a fixed sample size for this study based on the parent study (n = 757), therefore a post hoc power analysis was performed. Based on a sPTB rate of 7.9% in the group with a second stage <3 hours (n = 709) and a ratio of 15:1, we had >90% power to detect a 2.5-fold difference in sPTB rate for women with a second stage ≥3 hours (n = 48).

Materials and Methods

This was a planned secondary analysis of a large retrospective cohort study of women with 2 consecutive deliveries at the Hospital of the University of Pennsylvania from 2005 through 2010. The original study compared women with a term (≥37 weeks) induction to term spontaneous labor and evaluated rates of subsequent sPTB. Institutional review board approval was obtained prior to initiating the study.

Our hospital-based electronic database was used to identify women with >1 delivery from 2005 through 2010. The first pregnancy during this time period was considered the index pregnancy, which may or may not equate to the woman’s first pregnancy, as multiparous women were included. The second consecutive pregnancy during this time period was considered to be the subsequent pregnancy if delivery occurred at ≥16 weeks’ gestation. Patients were included in the study only once. Women were excluded from the parent study if they had a PTB in their index pregnancy of if they had a known history of PTB. Only women who reached the second stage of labor were included in the analysis of our current study.

The primary outcome was sPTB, defined as spontaneous labor and delivery or preterm premature rupture of membranes <37 weeks’ gestation. We first evaluated the overall length of the second stage of labor and its effect on sPTB rates in a subsequent pregnancy. We then evaluated various cut-points for length of the second stage of labor and the individual rates of sPTB for those cut-points (<1, <2, ≥2, ≥3, and ≥4 hours). Based on distribution of data, our sample size, and clinical utility, a prolonged second stage was defined as ≥3 hours. The rates of sPTB in a subsequent pregnancy were then compared between women with a second stage that was ≥3 hours and women with a second stage <3 hours.

Bivariate analyses were performed using χ ² and Fisher exact tests to compare categorical data. Student t tests and Mann Whitney U tests were used to compare parametric and nonparametric data, as appropriate. Linear and logistic regression were used to calculate odds ratios (OR). Multivariable logistic regression was used to adjust for confounders. To build our multivariable model, we included risk factors that had an association with the exposure or outcome at a level of P < .2. These identified risk factors were evaluated to be confounders or effect modifiers of the model. Any risk factor that was found to have an effect size of at least 10% was retained in the model as a confounder. Using backwards stepwise elimination, we created our final parsimonious model. Confounders maintained in our model were mode of delivery in the index pregnancy and parity. While race was not found to be a confounder, it was retained in the final model given the biological plausibility of being associated with both the exposure and outcome.

Data analysis was performed using software (STATA 12.0 for Windows; STATA Corp, College Station, TX). Statistical significance was set at P < .05. There was a fixed sample size for this study based on the parent study (n = 757), therefore a post hoc power analysis was performed. Based on a sPTB rate of 7.9% in the group with a second stage <3 hours (n = 709) and a ratio of 15:1, we had >90% power to detect a 2.5-fold difference in sPTB rate for women with a second stage ≥3 hours (n = 48).