Preinduction cervical assessment has traditionally been accomplished through the Bishop score.1 This evaluation is essentially subjective, and both inter‐ and intra‐examiner variability exists, which could affect reliability. Transvaginal ultrasound allows visualization of the cervix beyond a closed external os. It has been demonstrated that 50% of the cervix is not palpable on digital examination because of the supravaginal segment of the cervix, allowing ultrasound to more accurately measure the cervical characteristics commonly used.2 Controlled studies have confirmed decreased fetal morbidity associated with a policy of planned induction at 41 weeks' gestation for post‐term pregnancies, making this the most common indication for induction.3–5 Induction of labor, particularly in nulliparous women, however, has been associated with an increased risk for operative delivery.6
A diagnostic tool that could more accurately identify those patients likely to have a successful induction would be a valuable instrument in potentially decreasing the risk of operative delivery. This study was undertaken to determine if transvaginal ultrasound, given its ability to objectively measure cervical dimensions, could predict successful induction better than clinical examination in post‐term women.
MATERIALS AND METHODS
A total of 122 women were prospectively recruited between November 1997 and January 2000 at the Grace General Hospital, St. John's, Newfoundland. The study protocol was approved by the Human Investigation Committee of the Faculty of Medicine, Memorial University of Newfoundland, and the hospital. Informed written consent was obtained from all study participants before examination. All women 41 weeks or greater (by rigorous menstrual history or an ultrasound done before 20 weeks), with singleton vertex pregnancies and intact membranes, were eligible.7 Exclusion criteria were previous cesarean delivery or any contraindication to vaginal birth.
Cervical examination by transvaginal ultrasound and digital clinical assessment was performed by separate examiners, blinded to each other's results, within 1 hour before starting induction of labor. The ultrasound was performed using a 5–9‐MHz ATL transducer (Advanced Technology Laboratories, Bothwell, WA). The woman's bladder was emptied before ultrasound, and care was taken not to compress the cervix with the probe. Ultrasound measurements were made in the sagittal plane and included cervical length, cervical dilatation, and the presence of funneling. Cervical length was measured along the length of the endocervical canal with simultaneous visualization of the internal and external os. The shortest of three measurements obtained was taken as the cervical length. Dilatation of the cervix on ultrasound was measured as the maximum width of echolucency across the endocervical canal. Funneling was defined as a V or U shaped indentation of the internal os, and the measurement was taken from the apex of the funnel to the plane of the internal os. Digital clinical examination assessed the five components of the Bishop score.1 The ultrasound variables were measured three times, and then once during manual suprapubic pressure on the maternal abdomen. In the last 40 patients, a visual analogue pain scale was used to assess patient discomfort with each method (0–10 cm, 0 = no pain, 10 = severe pain).8
Induction of labor was initiated within 1 hour of the cervical assessments. The method of induction chosen was left to the discretion of the attending obstetrician, based on the results of the digital examination. Methods used at our center include amniotomy, intravenous oxytocin, vaginal dinoprostone gel, controlled release vaginal dinoprostone suppository, and oral or vaginal misoprostol tablet. Successful induction was defined as vaginal delivery.
The primary outcome of interest was the rate of vaginal delivery. The sample size requirement was based on the confidence interval around the estimated positive predictive value,9 assuming alpha = 0.05 and margin of error = 0.10, finding 96 vaginal deliveries required. Given a 20% cesarean delivery rate at our center, we anticipated 120 patients would need to be enrolled to provide 96 vaginal births. Secondary outcomes were the frequency of active labor in 12 hours, vaginal delivery in 12 and 24 hours, latent phase duration, and induction to delivery interval. Student t test, χ2, and Fisher's exact test were used for univariate analysis to compare the ultrasound and digital examination measurements with the primary and secondary outcomes. Receiver operating characteristic (ROC) curves were constructed to determine appropriate cutoffs for continuous variables. Simple linear and stepwise multiple regression models were generated to identify variables that were significantly associated with the outcomes of interest. Paired Student t test was used to compare discomfort (by visual analogue scale) with transvaginal ultrasound and digital examination. Statistical analysis was performed with Statistix 4.1 (Analytical Software, Tallahassee, FL).
Data were available for 120 of the 122 subjects enrolled. Seventy‐eight women (64%) were nulliparous, and 44 (36%) were multiparous. Ninety‐eight women (80%) had vaginal deliveries. Demographic characteristics of the study population (mean ± standard deviation) were as follows: age 27.9 ± 5.2 years, height 164.1 ± 12.8 cm, weight 89.6 ± 18.5 kg, gestation 288.7 ± 2.0 days. One woman was indigenous, and the others (99.2%) were white. With regard to induction method, 50 (40%) women received oral or vaginal misoprostol, 34 (28.3%) vaginal dinoprostone gel, 18 (15%) intravenous oxytocin, 18 (15%) amniotomy, and two (1.7%) controlled release vaginal dinoprostone suppository.
The ROC curves failed to identify an appropriate cutoff for continuous variables relating to sonographic cervical measurements. These variables were, therefore, analyzed as continuous variables in the regression models. Dependent variables for logistic regression were vaginal delivery or not, vaginal delivery in 12 and 24 hours, and active labor in 12 hours. Linear regression analysis was performed using latent phase duration and induction to vaginal delivery interval as dependent variables. The Bishop score and each of its components, sonographic cervical length, dilatation, presence of funneling, maternal age, weight, height, gestation, parity, and method of induction were independent variables.
Independent predictors of vaginal delivery (Table 1) included Bishop score, cervical position, and maternal age. Although no appropriate cutoff for these continuous variables could be identified by ROC curves, examples of sensitivity, specificity, positive and negative predictive values for these predictors are shown in Table 2. Variables significantly associated with secondary outcomes are shown in Tables 3 and 4. Maternal weight, cervical dilatation, and cervical effacement predicted active labor in 12 hours. Cervical dilatation, induction method, and maternal weight independently predicted vaginal delivery in 12 hours. Vaginal delivery in 24 hours was predicted independently by cervical effacement and maternal parity. Factors independently associated with latent phase labor duration were maternal weight, cervical position, and cervical dilatation. Predictors of induction to vaginal delivery interval were parity, cervical effacement, and maternal weight.
No cervical ultrasound method measurement showed a significant association with any outcome of interest. Patient discomfort by visual analogue scale was significantly less with sonographic assessment of the cervix compared with digital examination (mean difference 3.95, P < .001).
Transvaginal ultrasound has gained increasing applications in obstetrics. Its use has primarily focused on detecting cervical change in women at risk for preterm delivery. In this regard, transvaginal ultrasound characteristics, particularly cervical length, have been found to be predictive of impending delivery.10,11 Evaluation of its clinical usefulness in predicting successful labor induction would appear to be a logical progression.
Early investigations involving transvaginal ultrasound and labor induction indicated that certain sonographic characteristics might be associated with successful induction. In a pilot study, Paterson‐Brown et al12 performed transvaginal ultrasound on a cohort of 50 women undergoing labor induction. In addition to the clinical measurement of Bishop score exceeding 5, posterior cervical angle greater than 70 degrees was determined to be significantly associated with vaginal delivery. Similarly, in a prospective study of 53 patients, Boozarjomehri et al13 found that the presence of cervical funneling was independently predictive of latent phase labor and total labor duration. Both these studies contained a small number of patients and thus may have lacked the power to detect other potential factors contributing to successful induction.
Subsequently, two larger studies have been published, which compared Bishop score and transvaginal ultrasound in preinduction cervical assessment. In a study of 109 women, Watson et al14 used regression modeling to determine factors associated with successful induction. They determined that only cervical dilatation, as assessed by clinical examination, was a predictor of induction success. Likewise, Gonen et al15 prospectively evaluated 86 study subjects and found that only Bishop score and parity were independent predictors of vaginal delivery in induced labor. These studies evaluated women undergoing induction for a wide variety of indications.
Our findings are consistent with these latter two investigators in that only measurements obtained by clinical digital examination were significantly associated with a successful outcome of induction. Unlike the other studies, however, we specifically focused on the post‐term population. Cervical ripening is a dynamic process, with changes occurring in the third trimester weeks before the onset of labor.16 Preinduction evaluation in women undergoing medical induction at earlier gestations may therefore be subject to different sonographic norms than the post‐term group.
An interesting finding in our study was the observation that maternal weight was inversely associated with successful labor induction. A progressive increase in maternal weight was independently associated with a longer latent phase and induction to delivery interval, as well as a decreased likelihood of reaching active labor by 12 hours. This cannot be explained by its association with birth weight because when birth weight was placed in the regression model, maternal weight remained significant. Although an increased risk of operative delivery with maternal obesity has been reported in the literature, there is little information on the progress of labor in this cohort.17 The majority of investigations looking at the influence of maternal weight and obstetric outcomes primarily focus on pregnant women at extremes of body weight, not specifically at women of varying body weights. Additionally, studies of predictors of successful induction suggest that cervical parameters are the strongest determinants of induction success, with maternal body habitus playing a less significant role.
One potential limitation of our study was that we did not classify patients by method of induction. Method of induction was found to be independently associated with the likelihood of vaginal delivery in 12 hours. Grouping patients by induction method would serve to reduce any confounding effect this characteristic may have had on the regression model. However, we did control for this potential confounder in the regression models. Alternatively, assessing the utility of transvaginal ultrasound on patients undergoing labor induction with a variety of methods serves to increase the generalizability of our findings and may be viewed as a possible strength of our study.
It is of interest that sonographic measurements of the cervix, despite being a more objective assessment, particularly of cervical length, have limited utility in preinduction cervical assessment. Although transvaginal ultrasound has demonstrated its usefulness in the preterm population, digital examination of the cervix appears to more effective in prediction of successful labor induction.10,11 A question that remains, related to induction in post‐term pregnancy, is whether ultrasonographic cervical assessment is useful in predicting success in that group of women where the cervix is closed clinically, and effacement cannot be determined.
1. Bishop EH. Pelvic scoring for elective induction. Obstet Gynecol 1964;24:266–8.
2. Jackson GM, Ludmir J, Bader TJ. The accuracy of digital examination and ultrasound in the evaluation of cervical length. Obstet Gynecol 1992;79:214–8.
3. Hannah ME, Hannah WJ, Hellmann J, Hewson S, Milner R, Willan A. Induction of labor as compared with serial antenatal monitoring in post-term pregnancy: A randomized controlled trial. N Engl J Med 1992;326:1587–92.
4. Society of Obstetricians and Gynecologists of Canada. Post-term pregnancy. SOGC committee opinion no. 15. Ottawa, Canada: Society of Obstetricians and Gynecologists of Canada, 1997.
5. Crowley P. Interventions for preventing or improving the outcome of delivery at or beyond term (Cochrane review). In: The Cochrane Library, Issue 3. Oxford: Update Software, 2000.
6. Seyb ST, Berka RJ, Socol ML, Dooley SL. Risk of cesarean delivery with elective induction of labor at term in nulliparous women. Obstet Gynecol 1999;94:600–7.
7. Manning FA. General principles and applications of ultrasonography. In: Creasy RK, Resnik R, eds. Maternal fetal medicine. 4th ed. Philadelphia: WB Saunders, 1999: 169–206.
8. Ho K, Spence J, Murphy MF. Review of pain measurement tools. Ann Emerg Med 1996;27:427–32.
9. Arkin CF, Wachtel MS. How many patients are necessary to assess test performance? JAMA 1990;263:275–8.
10. Iams JD, Goldenburg RL, Meis PJ, Mercer BM, Moawad A, Das A, et al. The length of the cervix and the risk of spontaneous preterm delivery. N Engl J Med 1996;334:567–9.
11. Crane JMG, Van den Hof M, Armson BA, Liston R. Transvaginal ultrasound in the prediction of preterm delivery: Singleton and twin gestations. Obstet Gynecol 1997;90:357–63.
12. Paterson-Brown S, Fisk NM, Edmonds DK, Rodeck CH. Preinduction cervical assessment by Bishop's score and transvaginal ultrasound. Eur J Obstet Gynecol Reprod Biol 1991;40:17–23.
13. Boozarjomehri F, Timor-Trisch I, Chao C, Fox HE. Transvaginal ultrasonographic evaluation of the cervix before labor: Presence of cervical wedging is associated with shorter duration of induced labor. Am J Obstet Gynecol 1994;171:1081–7.
14. Watson WJ, Stevens D, Welter S, Day D. Factors predicting successful labor induction. Obstet Gynecol 1996;88:990–2.
15. Gonen R, Degani S, Ron A. Prediction of succesful induction of labor: Comparison of transvaginal ultrasonography and the Bishop score. Eur J Ultrasound 1998;7:183–7.
16. Kushnir O, Vigil DA, Izquierdo L, Schiff M, Curet LB. Vaginal ultrasonographic assessment of cervical length changes during normal pregnancy. Am J Obstet Gynecol 1990;162:991–3.
© 2001 The American College of Obstetricians and Gynecologists
17. Johnson SR, Kolberg BH, Varner MW, Railsback LD. Maternal obesity and pregnancy. Surg Gynecol Obstet 1987;164:431–7.