Along with the increasing rate of cesarean section (CS) births has been a concomitant decrease in the rate of vaginal birth after cesarean (VBAC), mostly due to concerns about uterine rupture during a trial of labor (TOL). The risk of uterine rupture in laboring women with a previous CS must be balanced against the increasing current and future risk of peripartum hysterectomy with repeat cesarean. Improved prediction of uterine rupture could allow women at low risk to have a TOL, and those at high risk for uterine rupture could undergo a planned CS. Thinning in the lower uterine segment (LUS) measured by ultrasonography is a possible predictor of uterine rupture. The present study was performed to identify an optimal LUS thickness cutoff value and to define groups of women with a history of previous CS to whom TOL either should not be offered or could be offered safely.
PubMed and EMBASE were searched for relevant articles published during 1980 to 2011. Included studies reported on pregnant women with at least 1 prior CS and on the sonographic appearance of the LUS during pregnancy in relation to uterine defects observed during or immediately after delivery. Studies were scored on methodologic and clinical characteristics. For each study, a 2 × 2 table was constructed, cross-classifying LUS thickness measured by ultrasound and the presence of LUS defect after delivery. Sensitivity and specificity were determined for each study in the receiver operating characteristic (ROC) space along with a corresponding summary ROC curve.
From an initial 147 citations, 21 studies (all cohort design) were included and involved 2776 women. The median sample size was 71 (range, 10–642; interquartile range, 149.5). Lower uterine segment measurements were obtained at 34 to 39 weeks’ gestation. Cutoff values used to define an insufficient LUS ranged from 0.5 to 3.0 mm in studies measuring myometrial LUS thickness and from 1.5 to 5.1 mm in those using full LUS thickness. Full LUS thickness measurement using cutoffs of 2.0 to 3.0 mm had a specificity of 0.91 (95% confidence interval [CI], 0.80–0.96) at a sensitivity of 0.61 (95% CI, 0.42–0.77). Full LUS thickness measurement using cutoffs of 3.1 to 5.1 mm had a specificity of 0.63 (95% CI, 0.30–0.87) at a sensitivity of 0.96 (95% CI, 0.89–0.98). The accuracy of transvaginal or transabdominal sonography could not be compared statistically because of the limited number of studies.
The main finding is the strong negative correlation between LUS thickness and risk of uterine defect. The similar shapes of the ROC curves for myometrial and full LUS thickness indicated no significant difference in accuracy, shape, and position. An ideal screening test to predict uterine rupture could affect medical decision making by allowing accurate selection of women with a scarred uterus unlikely to have uterine rupture and thus able to have a TOL.
Departments of Obstetrics and Gynaecology (N.K., I.M.d.G., B.W.M., E.P.) and Anaesthesiology (I.C.W.) and Clinical Research Unit (B.C.O.), Academic Medical Centre, Amsterdam, the Netherlands