The effect of transvaginal ultrasound, vaginal examination, or coitus on fetal fibronectin results: individual participant data from 6 cohort studies

Abstract

This study aimed to determine whether recent cervical manipulation via transvaginal ultrasound, sterile vaginal examination, or coitus affects the accuracy of fetal fibronectin results. An electronic search was performed in PubMed, Scopus, Embase, Ovid MEDLINE, ClinicalTrials.gov, Cochrane Library, and CINAHL using a combination of pertinent key words from inception to June2019. We included all observational studies that provided individual-level data on fetal fibronectin results after recent transvaginal ultrasound, sterile vaginal examination, or coitus. Studies were appraised using the Newcastle-Ottawa Quality Assessment Scale for cohort studies. Individual participant data from the included studies were pooled for each intervention. The primary outcome was agreement between pre- and postmanipulation swabs, estimated using proportion agreement and kappa statistics with 95% confidence intervals. Secondary outcomes included frequency in which the fetal fibronectin result changed after cervical manipulation and percentage of discordant pairs. Baseline fetal fibronectin swabs were not obtained in studies examining coitus; therefore, the results of these articles were examined separately. Outcome data were combined to estimate the relative risk of a positive qualitative fetal fibronectin result after coitus and differences in the concentration of quantitative fetal fibronectin. Of 807 studies identified, 6 were included. Three studies assessed the effect of transvaginal ultrasound (n=346 specimen pairs), 2 of sterile vaginal examination (n=122 specimen pairs), and 2 of coitus (n=262 specimen pairs) on fetal fibronectin results, with 1 study assessing the effect of more than 1 intervention. The proportion agreement between specimen pairs before and after transvaginal ultrasound and sterile vaginal examination was 93.4% (kappa, 0.69; 95% confidence interval, 0.57-0.81) and 88.5% (kappa, 0.69; 95% confidence interval, 0.54-0.84), respectively. For both transvaginal ultrasound and sterile vaginal examination, discordance with a positive preintervention fetal fibronectin and negative postintervention fetal fibronectin occurred more frequently than the converse. Patients reporting coitus within 24 to 48 hours were more likely to have a positive fetal fibronectin result than controls (39.7% vs 7.1%; relative risk, 5.6; 95% confidence interval, 3.0-10.6). Cervical manipulation via transvaginal ultrasound or sterile vaginal examination does not significantly affect fetal fibronectin results; therefore, its use after these exposures is clinically acceptable. Conversely, the use of fetal fibronectin in the setting of recent coitus should continue to be discouraged.