Risk Factors for Cesarean Delivery following Labor Induction in Multiparous Women

Objective. To identify potential risk factors for cesarean delivery following labor induction in multiparous women at term. Methods. We conducted a retrospective case-control study. Cases were parous women in whom the induction of labor had resulted in a cesarean delivery. For each case, we used the data of two successful inductions as controls. Successful induction was defined as a vaginal delivery after the induction of labor. The study was limited to term singleton pregnancies with a child in cephalic position. Results. Between 1995 and 2010, labor was induced in 2548 parous women, of whom 80 had a cesarean delivery (3%). These 80 cases were compared to the data of 160 parous women with a successful induction of labor. In the multivariate analysis history of preterm delivery (odds ratio (OR) 5.3 (95% CI 1.1 to 25)), maternal height (OR 0.87 (95% CI 0.80 to 0.95)) and dilatation at the start of induction (OR 0.43 (95% CI 0.19 to 0.98)) were associated with failed induction. Conclusion. In multiparous women, the risk of cesarean delivery following labor induction increases with previous preterm delivery, short maternal height, and limited dilatation at the start of induction.


Introduction
Induction of labor is a common and essential element of the contemporary obstetric practice and now accounts for approximately 20% of all deliveries [1][2][3]. Induction of labor is thought to be associated with an increase in the risk of cesarean delivery both for nulliparous and multiparous women [4]. is has been demonstrated both for inductions on medical grounds and for elective inductions [5,6]. More recent randomized comparisons have demonstrated that the effect of the induction of labor on the risk of cesarean delivery is limited. In postterm women as well as in women with prolonged rupture of membranes at term and in women with hypertensive disease, induction of labor is more effective than expectant management [7][8][9].
Data in parous women undergoing labor induction have revealed con�icting results. Some observational studies suggest that the rate of cesarean delivery in multiparous women with an elective induction is similar to that in those women with a spontaneous onset of labor [6,10,11]. Other studies report an increased risk for those who were electively induced [12][13][14][15]. One recent study even reported a lower cesarean delivery rate in multiparous women in whom labor was induced preventively, in order to ensure that pregnant women entered labor at an optimal time for the mother-baby pair [16].
Not much is known about factors related to a cesarean delivery aer induction of labor in multiparous women.
In women where cesarean delivery is required, the procedure not only carries the operative risks in the index pregnancy, but also increases risks for future pregnancies [17]. Consequently, it would be useful to understand which factors are related to a cesarean delivery aer induction of labor in multiparous women. In view of this question, we performed a case-control study to examine which factors are associated with a failed induction of labor in multiparous women.

Methods
We performed a case-control study in the Department of Obstetrics and Gynecology of the Máxima Medical Centre in Veldhoven, in the period 1995-2010. e Máxima Medical Centre is a large teaching hospital with a regional function for tertiary perinatal care in the south of e Netherlands. We included multiparous women with a singleton pregnancy above 37 weeks in whom labor was induced. Women with a pregnancy with a known fetal anomaly or a fetus in noncephalic presentation were excluded. Cases were de�ned as multiparous women with at least one previous vaginal delivery, in whom labor was induced and had resulted in a cesarean delivery. For each case, we selected two multiparous women with successful inductions of labor as controls, one successful induction immediately before a case and one immediately aer a case. Successful induction of labor was de�ned as achieving a vaginal delivery any time aer the onset of induction of labor.
Induction of labor was performed according to a standard protocol. Patients with an unfavorable cervix (Bishop score < 6) received 2 mg Prostin gel (1 mg if there was prelabor rupture of membranes) in the posterior fornix of the vagina, repeating the dose aer 6 hours if necessary, depending on the Bishop score (Prostin E2, dinoprostonum 1 mg/3 g (RVG 13823) or 2 mg/3 g (RVG 13824), P��er B.V., Capelle aan de IJssel, e Netherlands). Amniotomy was performed in women with a favorable cervix (Bishop score ≥ 6). Oxytocin augmentation was started in cases with unsatisfactory progress of labor or when contractions were absent 60 minutes aer amniotomy. Continuous monitoring of fetal heart beat and uterine contractions was registered on the cardiotocograph in all cases. e analgesics used were mostly pethidine or epidural analgesia, the latter as a continuous infusion of 0.1% Ropivacaine and 0.5 g/mL Sufentanil. Cesarean delivery was performed as necessary for failure to progress in labor or for fetal distress, as assessed by cardiotocography and/or fetal blood sampling.
For cases and controls, we collected maternal age, height, body mass index (BMI), parity, reason for induction, gestational age, Bishop score, need for epidural, birth weight, and reason for cesarean delivery.
In our hospital, ethics committee approval is not required for a retrospective study.
Statistical analyses were performed using SPSS soware (SPSS Inc, Chicago, IL).
e association between potential risk indicators and the occurrence of cesarean delivery was addressed using logistic regression analysis. In the univariable analysis, we expressed the association between each variable and the occurrence of cesarean delivery with an odds ratio (OR) and a 95% con�dence interval (CI). Multivariable logistic regression analysis was used to estimate the independent contribution of potential risk indicators for the occurrence of cesarean delivery. In the multivariable analysis, we entered all variables with a value < .50 in the univariable analysis. Variables stayed in the model when their value was < .25.

Results
Between 1995 and 2010, labor was induced in 2548 parous women, of whom 80 had a cesarean delivery (3%). ese 80 cases were compared to the data of 160 parous women who had a successful induction of labor (controls) ( Figure 1). In 41 (51%) of the cases, the reason for cesarean delivery was failure to progress, in 34 (43%) fetal distress, and in 5 (6%) both. Table 1 shows the characteristics of the study population at the start of induction of labor. ere was no signi�cant difference in maternal age, parity, or reasons for induction of labor between cases and controls.
e characteristics of the study population during induction of labor are shown in Table 2. Prostaglandins for induction were used in 52 of the cases (65%) and in 80 (50%) of the controls (OR 1.9, 95% CI 1.1 to 3.2, value 0.029). Otherwise, labor was induced by arti�cial rupture of membranes, mostly followed by oxytocin intravenous or by oxytocin alone in case of prolonged rupture of membranes. Furthermore, the need for epidural analgesia during labor differed between the cases and controls (OR 3.1 (95% CI 1.5 to 6.1, value 0.001)). e neonates born aer cesarean delivery more oen had a low Apgar score aer one minute (15% versus 6%) and an arterial umbilical pH under 7.15 (15% versus 8%). ese children were more oen admitted to the neonatal care (20% versus 7%) (OR 3.4 (95% CI 1.5 to 7.7), value 0.004) ( Table 3).
Furthermore, dilatation (OR 0.64 (95% CI 0.50 to 0.83)) for every centimeter more dilated at the start of induction, nonengagement of the fetal head (OR 3.1 (95% CI 1.1 to 8.4)), and the need for using prostaglandins as induction method (OR 1.9 (95% CI 1.1 to 3.2)) were all associated with the risk of cesarean delivery. e amount of effacement of the cervix at the start of the induction did not differ between the cases and controls. In terms of previous obstetric history, women with only previous preterm delivery had a signi�cantly higher risk of cesarean delivery than those with at least one previous term delivery (OR 3.5 (95% CI 1.8 to 6.9)). No difference in birth weight was observed between the two groups.
In the multivariable analysis (Table 4), the risk of cesarean delivery was signi�cantly associated with low maternal height with an adjusted OR of 0.87 (95% CI 0.80 to 0.95, ), a history of preterm delivery (adjusted OR 5.3 (95% CI 1.1 to 25), ) and the amount of dilatation at the start of induction of labor (adjusted OR 0.43 (95% CI 0.19 to 0.98), ). In the cesarean delivery group, there were ten women (13%) with the presenting fetal head at the level of the pelvic inlet, so at station -3. Indications for cesarean delivery were failure to progress because of neglected transverse lie ( ), prolapsed fetal arm ( ), prolapsed umbilical cord ( ), non-engaged persistent occiput anterior position ( ), failure to progress of dilatation ( ) and fetal distress ( ). In 50% of them, the method of induction was amniotomy, mostly followed by oxytocin. In the control group, in seven women (4%) the presenting fetal head at the start of the induction of labor was at station -3. In only 29% (2/7) of these seven women amniotomy was performed, and they all delivered spontaneously.
We also analy�ed our data strati�ed for the reason for cesarean delivery: failure to progress or fetal distress. ere were no signi�cant differences between those groups. Both in the univariable and the multivariable regression analysis the point estimates of the odds ratios were approximately the same as in the analysis for the whole group.

Discussion
In our hospital, we found that induction of labor in multiparous women resulted in a cesarean delivery rate of 3%. is study demonstrates that maternal height, the amount of dilatation at the start of induction and a history of preterm delivery played signi�cant roles in determining the risk of cesarean delivery in induced multiparous women.
In this study, in which we only included women with at least one previous vaginal birth, we considered the need for a cesarean aer induction of labor in a multiparous woman as the outcome measure failure of induction of labor. Although a woman might have reached active labor, we were not interested in another de�nition of failed induction, such as not reaching vaginal delivery within 24 hours. Although the latter outcome is frequently used, we consider it as less relevant, since in our opinion the birth of a healthy child from a healthy mother as well as spontaneous delivery, are of more importance than a quick delivery.
Women with a previous preterm delivery had a higher risk of cesarean delivery aer induced labor than those with at least one previous term delivery. is �nding corresponds with the results of the study of Park et al. [18]. He examined ). Our results are in line with the results of Park, indicating that the course of induction in women with a history of preterm delivery differs from women with a term delivery.
In 13% (10/80) of the cases, the fetal head at the start of induction was at station -3, whereas in the control group this was only 4% (7/160). Taking the reasons for cesarean delivery in these cases into account, one may conclude that one should be careful performing amniotomy if the fetal head is not properly engaged.
Although we are aware of the additional risks inducing women with a history of cesarean delivery, we included in both cases and controls six women with a previous cesarean delivery, besides a previous vaginal delivery. A history of previous cesarean delivery was, in our study, not signi�cantly associated with the occurrence of cesarean delivery aer induction of labor.
e 3% rate of cesarean delivery in multiparous women in whom labor was induced should be compared with the cesarean delivery rate in multiparous women with spontaneous labor. Heffner reports a cesarean delivery rate of 2.4% among multiparous women with spontaneous labor [19]. e study of Jacquemyn et al. reported a cesarean delivery rate of 1.5% in women with spontaneous onset of labor, as compared to 2.8% in women with induced labor [12]. In the study of Nicholson the cesarean delivery rate was 9.9% [16].
As with other retrospective analyses, this study does have limitations of relying on data evaluated and collected in a nonstandard fashion and the inability to obtain clari�cation when information is not clearly delineated in the patient records. As a result of this, we did not have any information on the position of the fetal head at the start of the induction, nor had we enough information on the position and the consistency of the cervix and the weakness to enable us to calculate the Bishop score for every patient in our study. Furthermore, it would have been interesting to look at the time between deliveries to investigate whether this variable had predictive value. Unfortunately, these data were not present.
Although matching facilitates study precision, we did not match cases and controls to prevent statistical bias because of overmatching. Instead, we used multivariable regression to control for confounding. e absolute risk of cesarean delivery aer induction of labor in multiparous women is low. However, the multivariable analysis showed that this risk is increased in case of a history of preterm delivery, lower maternal height, and little dilatation at the start of induction of labor. is information will allow more accurate counseling and better informed consent in the decision-making process regarding induction of labor in multiparous women.
��n��c� �f �n�eres�s e authors report no declaration of interests.