OC012: The clinical significance of “sludge” during transvaginal examination of the cervix in patients with preterm labor

Abstract

Objective: To determine the inter-observer variability of cervical length (CL) measured by conventional transvaginal 2D ultrasound (2DUS) and compare it to 3D ultrasound (3DUS). Methods: Prospective, IRB-approved study of transvaginal CL by 2DUS and 3DUS in gravid subjects at high risk for preterm birth at Thomas Jefferson University from 2003–2004, using standardized examination techniques. Two independent investigators measured CL, funnel length and width (when present), before and after fundal pressure (FP) with 2DUS, and 3DUS. Investigators were blinded with respect to each other’s measurements. Each investigator measured the CL by 2DUS and then obtained and stored a 3DUS cervical volume. The CL determined from the 3DUS was assessed on a different day in order to limit bias from each investigator’s memory of the 2DUS CL results. Inter-observer variability (ITOV) and intra-observer variability (INOV) were evaluated via the intraclass correlation coefficient (ICC). Results: A total of 200 CL measurements were performed during 22 exams. Three exams (30 CL measurements) were excluded because of dynamic changes, leaving 19 exams and 170 measurements for analysis. ITOV was uniformly low (5.0% to 6.9%). INOV was also low (3.8% to 3.9%). With 2DUS before FP, ITOV was 6.9% with ICC. 931 (95%CI. 82-.98), and INOV was 3.8% with ICC. 962 (95%CI. 93-.98). With 2DUS after FP, ITOV was 5.0% with ICC. 950 (95%CI. 88-.98), and INOV was 3.9% with ICC. 961 (95%CI. 95-.98). For 3DUS, ITOV was 5.5% with CC. 945 (95%CI. 86-.98). Conclusions: This study demonstrates that CL measurements performed by blinded, independent investigators, using a standardized 2DUS and 3DUS measurement technique, are highly reproducible, with uniformly low inter-observer variabilities (5.0% to 6.9%). There was no indication of any inter-observer variability difference between 2DUS and 3DUS, although our study was not adequately powered to detect a small difference if it exists.