Abstract
Induction of labor is one of the most common procedures in modern obstetrics, with an incidence of approximately 20% of all deliveries. Not all of these inductions result in vaginal delivery; some lead to cesarean sections, either for emergency reasons or for failed induction. That’s why, It seems necessary to outline strategies for the improvement of the success rate of induced deliveries. Traditionally, the identification of women in whom labor induction is more likely to be successful is based on the Bishop score. However, several studies have shown it to be subjective, with high variation and a poor predictor of the outcome of labor induction. Transvaginal sonography for cervical measurement can be a more objective criterion in assessing the success of labor induction. Many studies have been done recently to compare cervical measurement and Bishop Score in labor induction.
This paper reviewed the literature that evaluated sonographic cervical length measurement to predict induction of labor outcome.
Introduction
Induction of labor is one of the most common procedures in modern obstetrics, with an incidence of approximately 20% of all deliveries [1–3]. As not all of these inductions result in vaginal delivery [4]; it is essential to determine the most appropriate method to predict successful induction of labor and vaginal delivery. The traditional method of predicting whether an induced labor will result in successful vaginal delivery is the Bishop score. However, assessment is subjective and it is a poor predictor of the outcome of labor induction [5–8] especially when the external os is closed [7, 9]. Many studies compared the performance of ultrasonographic cervical assessment with the Bishop score to predict the outcome of labor induction [10–15]. In this review we analyzed current evidence comparing Bishop Score to measured cervical length in prediction of labor induction outcome and vaginal delivery.
Cervical length measure technique [15, 16]
Transvaginal ultrasonographic assessment of the cervical length is performed with an empty bladder to avoid a deceptive elongation of the cervix. The transducer is placed on the cervix and identifies the internal os, the cervical canal, and the external os (Fig. 1). Next, pressure on the transducer is relaxed to the point at which the image blurs. Gentle pressure is then performed to recreate a clear image. The distance between the internal and the external os is then measured [16] (Fig. 2).
Acceptability and tolerability
In a prospective study, P C Tan et al. [13] performed before labor induction an ultra-sonographic measurement of cervical length and determined the Bishop score for 249 women. A 10-point visual analog scale (VAS) for procedure related pain was obtained. Transvaginal sonography was significantly less painful than digital examination for Bishop Score assessment (mean difference in VAS score 3.46; P < 0.001). The authors concluded that transvaginal sonography for cervical length measurement is better tolerated than digital examination for Bishop score assessment [13].
Cervical length and spontaneous labor onset
The results of the studies that focused on the relationship between cervical length measured by ultrasound and spontaneous onset of labor are discordant [17–22]. This can probably be explained by the small study groups and the differences in the methodology between these studies [20]. Rozenberg et al. [17] examined 128 singleton pregnancies at 39–41 Weeks, and reported a correlation between cervical length and labor onset within 7 days but not on the risk of caesarian section [20]. Vimercati et al. [18] measured cervical length in 120 patients with singleton pregnancies. He reported that a shortened cervix is a useful predictor of the likelihood of spontaneous onset of labor in a 7-day period. Strobel et al. [19] studied 97 pregnancies at 291–296 days’ gestation. He concluded that cervical length is a significant predictor of spontaneous onset of labor and delivery within 1–2 days but not within 4 days in either nulliparous or parous women.
The two largest series published on the subject are those of Ramanathan et al. [22] and Rao et al. [20]. The first study involved 1571singleton low risk pregnant women. Cervical length at 37 weeks of gestation and parity were both significantly associated to spontaneous labor onset and delivery within 31 days [22]. The chances of delivery within 31 days and of vaginal birth diminished when cervical length increased.
Rao et al. [20] in a prospective study including 1864 pregnancies at a term of 41 weeks of gestation or more, tried to predict the chances of spontaneous labor onset within the subsequent 10 days and the delivery mode by combining cervical length and maternal characteristics. The authors found that cervical length is independently associated to both spontaneous onset of labor and delivery mode.
Comparison between Bishop Score and sonographic measurement of cervical length in labor induction success prediction
The results of studies that relate pre-induction ultrasonographic evaluation of the cervix are controversial. The disparity among these studies is probably attributable, in part, to a variety of primary outcome variables, including vaginal delivery [15, 23], vaginal delivery within 24 h of initiation of induction [21, 23], and reaching the active phase of labor [23, 24] as the definition of induction success. Moreover, these contradictory results can be explained by theirnumerous bias; like the heterogeneity of the studied groups regarding the indication for medical induction [16, 25] and gestational ages [23, 26]; knowing that previous studies described the limitation of including women with different gestational ages and the use of various induction agents [16]. Another reason for the discrepant results among these studies may be related, in part, to the study population of participants (i.e. nulliparity or multiparity) [11, 27] because the response to oxytocic drugs and preinduction cervical dilatation among parous women are different from those among nulliparous women [27]. In fact, a previous history of vaginal delivery, have a positive beneficial effect on cervical compliance and parous women have innate lower cervical resistances due to their previous labors [27].
For all these reasons, meta-analysis published on the subject are also difficult to interpret.
Indeed, Verhoeven CJ et al. meta-analysis [28], published in 2013, included 31 studies with heterogeneous populations made of nulliparous and multiparous women at different gestational ages. In addition, the endpoints of these studies were different as well as the protocols used for cervical ripening.
As for the meta-analysis of Ezebial IU et al. [29] published in 2015, it had included two randomized studies [24, 30]. The first study [30] had involved a heterogeneous group of women in virtue of parity. However, after stratification, the authors found that the use of sonographic cervical length for assessing the cervix prior to induction of labor can reduce the need for prostaglandin administration in nullipara at term. The second study [24] included only nulliparous women and also concluded to the superiority of the sonographic measurement of cervical length in this group of women. Each taken alone, these two randomized studies are relevant regarding methodologies and results. However, their inclusion in a meta-analysis without stratification on parity introduces bias in the methodology. In addition, the cut-offs of Bishop score and cervical length used to define unfavorable cervix differ between the two studies introducing another methodology bias and making interpretation of results difficult.
In this review we analyzed separately studies that included only nulliparous women or stratified on parity, studies that included only parous women, and those with heterogeneous study population made of nulliparous and parous women. We focused only on studies that included women at gestational ages >37 weeks of gestation. Indeed, cervical ripening is a dynamic process, occurring late in the third trimester before the onset of labor [16]. Therefore, the interpretation of sonographic cervical assessment might be different according to gestational age [16].
Comparison in nulliparous women
In our literature review, we found 5 articles [15, 32] that concerned only nulliparous women and 3 articles that stratified on parity [4, 34]. Only one of these studies didn’t conclude to the superiority of ultrasound cervical length measurement [4]. In this study, Adelaide Cubal et al. [4] found that Bishop Score and cervical length measured by transvaginal ultrasound are good predictors of a successful induction of labor only in nulliparous women but this study was not able to find one test to be better than the other [4]. The other studies concluded to the superiority of ultrasound cervical length measurement compared to the Bishop score in preinduction cervical assessment. The cut-off value of cervical length in these studies ranged from 20.0 to 32.5 mm. Meijer-hoogeveenet al. [34] found that only cervical length measured in the upright position was a significant predictor of the need for a Cesarean section. The authors found that an increase of 1 mm in cervical length was associated with a 14% increase in the odds of a cesarean section for failure to progress. Park et al. [24] found that in comparison with the Bishop score, the use of sonographic cervical length for assessing the cervix prior to induction of labor can reduce the need for prostaglandin administration by approximately 50% without adversely affecting the outcome of induction in nulliparae at term. Of note, in this study, an unfavorable cervix was defined by a Bishop score≤4 or by a cervical length of 28 mm or more [24]. This can constitute a bias given that the results are dependent on these cut-offs used [24].
Comparison in parous women
Park et al. [25] in a prospective observational study enrolling 110 parous women at term examined the predictive value of previous obstetric history, Bishop Score, and sonographic measurement of cervical length for predicting failed induction of labor in parous women at term. The authors concluded that the previous obstetric history (i.e. only previous mid-trimester loss or preterm delivery) and the Bishop score independently predicted the failure of labor induction in parous women; however sonographic measurement of the cervical length appeared to have a poor predictive value for the risk of failed induction. Cubal et al. [4] found after stratification on parity that in parous women neither Bishop score nor ultrasound measurement of cervical length are associated to induction outcome.
Comparison in heterogeneous populations made of nullipara and parous women at term
Data of these studies are divergent. Roman et al. [9] found in their study about 106 pregnant women with a Bishop score≤5 undergoing labor induction that compared to the Bishop score, cervical length by ultrasound is not a better predictor for the outcome of labor induction in an unfavorable cervix [9]. Nevertheless, the Bishop score was of poor predictive value for failed induction of labor [9]. However, 5 other studies found in our literature review including women at term with heterogeneous parity concluded to the superiority of cervical length measurement for the prediction of labor induction outcome [11–13, 35]. The cut-off value of cervical length ranged in these studies from 19 to 28 mm.
Comparison in twin pregnancies
Park et al. [36] evaluated in a prospective study enrolling 72 twin pregnancies the predictive value of body mass index (BMI), Bishop score, and sonographic measurement of cervical length for predicting successful labor at >36.0 weeks’ gestation who were scheduled for induction of labor. They found that BMI independently predicted the success of labor induction but the sonographic measurement of the cervical length and Bishop Score had poor predictivevalues.
Other parameters studied for predicting labor induction outcome
Besides cervical length measurement, there are some trials on other sonographic markers for the outcome of labor induction like anterior or posterior cervical angle as well as fetal head station, fetal head–perineum distance and cervical wedging [4, 35].
Some biochemical markers have also been studied. These include fetal fibronectin [9] and insulin-like growth factor binding protein [33]. The results of these studies are not consistent with sometime conflicting results, thus further studies might be needed.
Conclusion
It is difficult to compare different published studies concerning the prediction of the results of labor induction, because of major differences in series characteristics [9]. Nevertheless, it seems actually clear that ultrasound cervical length measurement is useful and very probably better predictor than Bishop Score for the outcome of labor induction in nulliparous women at term. However, in parous women, this parameter appears to have a poor predictive value in pre-induction cervical assessment.
Declaration of interest
The authors report no conflicts of interest.