Introduction

Induction of labor (IoL) by ripening the cervix is a common obstetrical intervention, performed in approximately 20 % of pregnancies [1]. It is routinely advised for gravid women with an unfavorable cervix in whom maternal or fetal health risks outweigh the anticipated risk of ongoing gestation [2]. As such, specific indications for IoL may include, among others, premature rupture of membranes (PROM), diabetes, hypertensive disorders, maternal chronic diseases, fetal growth restriction, postdate pregnancy, oligohydramnios, and cholestasis of pregnancy. At the same time, it is important to consider that IoL itself may be associated with some complications, namely, uterine tachysystole with or without corresponding fetal heart rate changes [3–6], inadequate labor progression [7, 8], and cesarean delivery [9–16]. Factors increasing the risk of failed IoL are unfavorable cervix, low parity, older maternal age, early gestational age, high neonatal birth weight, maternal short stature, and increased body mass index [17–19].

Most studies of IoL to date have included highly varied populations in terms of gestational age and method of cervical ripening and indications. Whether the specific indications for IoL have different effects on the likelihood of its success or failure has hardly been investigated [20, 21]. The studies were also limited by the bias of defining cesarean delivery as a primary outcome for IoL, as cesarean delivery may also be performed for reasons unrelated to the induction procedure, long after active labor has been achieved.

Therefore, the aim of the present study was to evaluate the impact of indications for labor induction on the rate of induction failure and the risk of cesarean delivery.

Materials and methods

The database of a tertiary university-affiliated medical center was retrospectively reviewed for all women with a viable singleton pregnancy who underwent cervical ripening between January 2013 and December 2014. The study was approved by our local institutional review board (Approval No. 0351-13).

Study population

Inclusion criteria were vertex presentation, Bishop Score ≤7 and gestational age 34–41 completed weeks. Exclusion criteria were multiple gestation, non-vertex presentation, prior cesarean delivery, stillbirth, and chromosomal or major structural fetal anomalies. Women who had elective IoL for nonmedical reasons and women for whom data were insufficient to properly ascertain the indication for IoL were excluded as well.

Data collection

The following data were collected for all eligible patients from the departmental health records: demographic and obstetric characteristics, physical characteristics, pregnancy-related complications, indication for labor induction, findings on cervical examination at admission, gestational age at delivery, mode of delivery, indications for cesarean delivery, and neonatal birth weight.

Induction of labor protocol

According to our departmental protocol, a prostaglandin E2 (PGE2) slow-release vaginal insert placed in the posterior fornix (Propess 10 mg, Ferring Pharmaceuticals AB, Kiel, Germany) was used for labor induction in all the cases. Physician assessment and fetal heart rate monitoring were performed immediately before and 1 h after insertion and again at 12 and 24 h or in the event of painful uterine contractions, vaginal bleeding, or suspected PROM. The insert was removed on findings of cervical ripening (Bishop Score >7), uterine hyperstimulation (>5 contractions in 10 min), non-reassuring fetal heart rate, or if none of these has occurred at 24 h after placement. Women in whom cervical ripening was achieved were transferred to the delivery room, and if necessary, labor was augmented with oxytocin and/or amniotomy under continuous electronic fetal heart rate monitoring. If cervical ripening was not achieved, a subsequent trial was undertaken using either oxytocin or an extra-amniotic balloon, depending on the Bishop score, pattern of uterine contractions and physician discretion.

Definitions

Reasons for inducing labor were categorized into four groups on the basis of previous medical literature [22]: (1) maternal indication: diabetes mellitus, renal disease, pulmonary disease, cardiovascular disease, or other chronic illnesses (2) any hypertensive disorder of pregnancy; (3) PROM; and (4) fetal indication: suspected fetal growth restriction (sonographic estimated fetal weight <10th percentile according to local reference values [23]), oligohydramnios (amniotic fluid index ≤5 cm), suspected macrosomia (sonographic estimated fetal weight ≥4000 g), cholestasis of pregnancy, or postdate pregnancy (gestational age ≥41 + 0 weeks). IoL was considered failed under the following conditions: (1) Bishop score <7 at 24 h after PGE2 administration (lack of cervical ripening); (2) Cesarean delivery due to prolonged latent phase. This was defined as labor dystocia 24 h after PGE2 administration and dilatation of <4 cm after at least 12 h of oxytocin infusion with adequate contractions in the presence of ruptured membranes; or (3) removal of the PGE2 insert before 24 h due to hyperstimulation and non-reassuring fetal heart rate, leading to an emergency cesarean delivery.

Outcome measures

The primary outcome measure was the IoL failure rate and the secondary outcome measure—cesarean delivery rate, both stratified by the preliminary indication for IoL. As pre-labor PROM seems to be the most natural indication for IoL with the lowest IoL failure rates, we evaluated the other indications with reference to it. In addition, as parity is considered a major determinant for successful IoL, outcomes were analyzed separately for nulliparous and multiparous women.

Statistical analysis

Data analysis was performed with the SPSS package, v21.0 (Chicago, IL). Student’s t test and Mann–Whitney U test were used to compare continuous variables with and without normal distribution, respectively, between groups. Chi-square and Fisher’s exact tests were used for categorical variables, as appropriate. Logistic regression analysis was used to adjust IoL failure rates to known confounders, such as maternal age, gestational age at delivery and neonatal birth weight. Differences were considered significant at p < 0.05.

Results

Of the 15,564 women who gave birth at our center during the study period, 1066 (6.8 %) met the inclusion criteria. IoL failed in 213 (20 %), due to lack of cervical ripening (n = 162, 76 %), hyperstimulation and non-reassuring fetal heart rate (n = 40, 18.8 %), and labor dystocia (n = 11, 5.2 %) (Fig. 1).

Fig. 1
figure 1

Study population. IoL induction of labor, PGE2 prostaglandin E2, PROM premature rupture of membranes

Full size image

About half the cohort was nulliparous (n = 538, 50.5 %) and half multiparous (n = 526, 49.5 %). The nulliparous group accounted for a considerably higher percentage of the induction failures (148/538, 27.5 %) than the multiparous group (65/528, 12.3 %). In both the groups, the most common reason for performing IoL was fetal indication, followed by maternal indication. The relative proportion of IoL procedures for maternal indication was higher in the multiparous group (27.7 vs. 17.7 %) (Table 1).

Table 1 Indications for labor induction stratified by parity
Full size table

The demographic and obstetrical characteristics of the study cohort, stratified by IoL failure or success, are presented in Table 2. Women who failed IoL were significantly more likely to be nulliparous (69.5 vs. 45.7 %, p < 0.0001), older (median age 31 vs. 30 years, p = 0.047), and at an earlier median gestational age at delivery (39.4 vs. 40.0 weeks, p < 0.0001). There were no differences between the groups in pre-gestational body mass index or neonatal birth weight.

Table 2 Characteristics of the study cohort stratified by failure/success of induction of labor
Full size table

Failed IoL and cesarean delivery rate by indication: nulliparous women

Analysis of the nulliparous group revealed a significant association of the indication for induction with induction failure. The risk of failure was higher when IoL was performed because of maternal indication (41.1 %) than hypertensive disorders (33.3 %), fetal indication (23.7 %), or PROM (20.7 %) (Fig. 2). On logistic regression analysis controlling for maternal age, gestational age at delivery, and neonatal birth weight, only maternal indication remained significantly associated with the failure of IoL (aOR 2.52, 95 % CI 1.28–4.95, p = 0.007) and with cesarean delivery (aOR 2.36, 95 % CI 0.40–2.29, p = 0.019) (Table 3).

Fig. 2
figure 2

Rates of failure of induction of labor and cesarean delivery stratified by indication in nulliparous and multiparous women. IoL induction of labor, PROM premature rupture of membranes

Full size image
Table 3 IoL failure in nulliparous women, stratified by indication
Full size table

Failure of IoL and cesarean delivery rate by indication: multiparous women

Analysis of the multiparous group revealed a significant association of the indication for induction with induction failure. The risk of failure was higher when IoL was performed because of hypertensive disorders (26.7 %) than maternal indication (15.8 %), fetal indication (8.5 %), or PROM (5.9 %) (Fig. 2). On logistic regression analysis controlling for maternal age, gestational age at delivery, and neonatal birth weight, hypertensive disorders (aOR 7.26, 95 % CI 1.89–27.87, p = 0.004), and maternal indication (aOR 4.22, 95 %CI 1.14–15.58, p = 0.031) remained significantly associated with the failure of IoL but not with cesarean delivery (Table 4).

Table 4 IoL failure in multiparous women, stratified by indication
Full size table

Discussion

This retrospective study of women undergoing cervical ripening with PGE2 sought to evaluate the association between the indication for IoL and rates of IoL failure and cesarean delivery. There were several important findings: (1) the overall IoL failure rate was 20 % and the overall rate of cesarean delivery, 14.2 %; (2) women who failed IoL were significantly more likely to be nulliparous, older and at an earlier gestational age; and (3) among nulliparous women, a maternal indication for IoL was associated with a higher risk of failed IoL and cesarean delivery, while among multiparous women, a maternal indication and hypertensive disorders were associated with a higher risk of unsuccessful IoL but not cesarean delivery.

The literature reports IoL failure rates ranging from 20 to 30 % [24, 25] and cesarean delivery rates of approximately 20 % [26, 27]; both of these are compatible with our findings. As in the present study, others have identified nulliparity [10], advanced maternal age [19], and earlier gestational age [28] as risk factors for IoL failure. There are also reports of a significant association of gestational age with cesarean delivery [19–21, 28–30]. Interestingly, in the only two studies focusing on the effect of specific indications for induction on the risk of post-induction cesarean delivery, Parkes et al. [20] found a significant association of IoL failure with older gestational age (≥40 weeks; aOR = 2.028, 95 % CI 1.143–3.594, p = 0.0016), and Gerli et al. [21] found that postdate pregnancy was the only significant risk factor. Thus, gestational age may have a bimodal effect in the setting of IoL. Several studies have suggested that the optimal timing for labor induction is 36–40 weeks of gestation and that the success rate is lower before and after this period [29, 30].

In terms of indications for IoL, Parkes et al. [20] reported that among the 343 nulliparous women out of the 796 women evaluated, a fetal indication for IoL was carried with a significantly higher risk of cesarean delivery (aOR = 3.76, 95 % CI 1.153–12.265) than the other indication categories. However, in the multiparous group, no association was found for any of the indication categories. They suggested that these results represent another aspect of the well-established risk of perinatal morbidity and mortality in compromised fetuses. By contrast, in our nulliparous group, only maternal indication was significantly associated with failed IoL and with cesarean delivery among the indication categories. We speculate that this discrepancy is attributable to the different outcome measures of our study, or alternatively to the specific indications within the fetal indication group, which may have a different impact on the success of induction (growth restriction, macrosomia or oligohydramnios). We limited the definition of failed IoL to a failure to achieve cervical ripening or active labor. This was done to avoid introducing a bias, as the decision to perform intrapartum cesarean delivery may also be influenced by other factors stemming from the active labor itself, long after induction has been successfully achieved. The assumption that IoL is a better outcome measure than cesarean delivery was further strengthened by our detection of a significant effect of both maternal indication and hypertensive disorders on IoL failure in the multiparous group. As such, our results probably represent the true association of IoL indications with the process of induction itself, without confounding factors. The study of Gerli et al. [21] is also subject to the same bias. In addition, it was limited by the categorization of the cohort of 324 women by nine individual indications for IoL, which led to relatively small numbers for each.

The strengths of our study include the inclusion of a large uniform cohort of women all of whom underwent the same management protocol in terms of IoL and delivery as well as and the case-by-case review of the data by specialists. The indications were grouped by category to ensure an adequate sample. Furthermore, we used more accurate outcome measures: earlier studies evaluated rates of cesarean delivery regardless of the success or failure of induction, whereas we evaluated the success or failure of IoL it terms of the achievement of active labor, regardless of the ultimate mode of delivery. The main study limitation is its retrospective design.

In conclusion, the indications for IoL apparently have an important impact on its success or failure. Our data suggest that maternal indications for IoL pose a significant risk of failure in both nulliparous and multiparous women.