Evaluation of endometrial thickness by transvaginal ultrasound and baseline risk factors as a predictor for endometrial abnormalities in postmenopausal women

Abstract

Introduction/Purpose: To evaluate the endometrial thickness (ET) as a predictor of endometrial abnormalities in postmenopausal women and whether consideration of baseline risk factors increases diagnostic accuracy. Methods: This is a retrospective observational study of postmenopausal women presenting with bleeding or thickened endometrium (≥4 mm) on ultrasound, between 2003 and 2012. Risk factors for endometrial abnormality were analysed using logistic regression. Of 301 women, 220 were symptomatic and 81 were asymptomatic. The median ET was 6 mm (IQR 4–9) for symptomatic women and 9 mm (IQR 6–12) for asymptomatic women. Results: Abnormal pathology was found in 35 symptomatic (15.9%) and 6 asymptomatic women (7.4%). For each 1 mm increase in ET, the odds of an abnormal diagnosis increased by 16.3% (95% CI 9.6–23.5) for symptomatic and 19.9% (95% CI 3.1–39.3) for asymptomatic women. The Youden's index method identified an ET threshold of ≥7.1mm for symptomatic and ≥14.5mm for asymptomatic women. In symptomatic women the sensitivity was 88.6% (95% CI 72.3–96.3) and specificity 69.2% (95% CI 61.9–75.6), while in asymptomatic women the sensitivity was 50.0% (95% CI 13.9–86.1) and specificity was 89.3% (95% CI 79.5–95.0). The addition of age in the symptomatic women model reduced the sensitivity (82.9% (95% CI 65.7–92.8)) but increased the specificity (72.4% (95% CI 65.3–78.6)). Conclusion: ET is a significant predictor of abnormality. In the absence of risk factors, our study suggests that invasive procedures may be withheld until the ET is ≥7.1 mm with bleeding and ≥14.5 mm in asymptomatic women with no bleeding.