A comparative study of transvaginal ultrasound scan and digital vaginal examination in prediction of preterm delivery at 32 to 36 weeks in Kitui District Hospital.

Abstract

Introduction: Preterm labour occurs when regular uterine contractions associated with cervical shortening and dilatation occurs before 37 weeks of gestation. Preterm deliveries are those that occur at less than 37 weeks of gestation (1). Every year approximately 15million preterm births occur and more than 1million neonates die from prematurity(1). Although there are no accurate world wide data; estimates of preterm births range from 5% in developed countries to 18% in developing countries (1). As labour approaches the cervix shortens, softens, rotates anteriorly and dilates(2). Both digital and ultrasound examinations have been used to show that cervical shortening is a risk factor for preterm delivery(2). Vaginal examination as a screening tool for preterm delivery gives conflicting results with false positive range of 20-60%(3); which may be due to lack of precision, inter examiner differences and lack of reproducibility of the result (3). Ultrasound examination provides an objective means of assessing the cervix. No studies have been conducted locally to determine the best method for prediction of preterm delivery and there are no specific guidelines on prediction of preterm delivery in symptomatic women. Broad objective: To compare transvaginal ultrasound scan and digital vaginal examination in prediction of preterm delivery between 32-36 weeks of gestation. Specific objectives: Among the women presenting with preterm labour at 32 to 36 weeks of gestation to determine; the prevalence of cervical parameters (shortening and dilatation) by transvaginal ultrasound scan and digital vaginal examination; the gestation at delivery; and compare the sensitivity, specificity, positive predictive value and negative predictive value of transvaginal ultrasound scan versus digital vaginal examination in determining the risk of preterm delivery. Methodology: This was a prospective comparative study conducted in Kitui County, at the Kitui District Hospital. Study participants were women between 32-36 weeks of gestation presenting with preterm labour. The main symptomatology during recruitment was lower abdominal pain; with drainage of liquor and per vaginal bleeding being the main aspects in exclusion criteria. Participants were evaluated to determine cervical changes, using DVE and TVS assessment. The cervical changes evaluated were cervical length, cervical dilatation and presence or absence of funnelling. Participants were managed depending on the clinical findings, though there were no specific hospital protocols for management of preterm labour. Participants were followed up until delivery or 37 completed weeks of gestation. The primary outcome was delivery before 37 weeks of gestation. Results: Between October 2014 to May 2015, a total of 110 women were enrolled in the study, 5 were lost to follow up and 3 had incomplete questionnaires. The remaining 102 proceeded till the end of the study. 54.9% of the participants had preterm delivery. TVS had low sensitivity and high specificity in detecting cervical shortening and almost comparable sensitivity and specificity in detecting dilatation. Vaginal examination had low sensitivity and high specificity for detecting both cervical shortening and dilatation. The PPV of a short cervix to predict preterm delivery by TVS was 90.5% as compared to that of DVE 77.8%. The PPV of a dilated cervix to predict preterm delivery by TVS was 63.3% as compared to that of DVE 77.4%. Prevalence of cervical shortening on VE was 27/102 26.5% ± 4.3 (95% CI 17.8-35.2). Prevalence of cervical shortening on TVS was 21/102 20.6% ± 4 (95% CI 12.6-28.6). Both methods (VE&TVS) had comparable sensitivity in prediction of PTB. From the study cervical length was a better predictor of preterm labour but when both length and dilatation were used the chances of predicting PTB was much higher. Conclusion: A transvaginal ultrasound scan is a better predictor of preterm birth but VE is also fairly accurate and quite useful since it is more accessible especially in our setup due to limited resources