Extremely short cervix in the second trimester: bed rest or modified Shirodkar cerclage?

Abstract

The objective of this study was to compare modified Shirodkar cerclage to bed rest for treatment of the midtrimester extremely short cervix. This study used a concurrent retrospective cohort design at two institutions over the same period, 2000-2010. Patients were included at both institutions when midtrimester endovaginal ultrasound cervical length was ≤ 15 mm and had modified Shirodkar cerclage (cerclage group) at New York Hospital Queens and bed rest (control group) at Weill Cornell Medical Center. Cerclage was placed as high on the cervix as possible. Indomethacin and antibiotics were used perioperatively. The cerclage group included 112 patients and the control group included 55 patients. Median postoperative cervical length in the cerclage group was 3.3 cm (interquartile range 3.0-3.6). Cerclage patients were less likely to deliver preterm at 37, 35, 32, and 28 weeks (P=0.0066, 0.0004, 0.0023, and 0.03 respectively) and had longer latency (median 120 vs. 94 days P<0.0001). Kaplan-Meier survival curve showed a significant benefit in favor of cerclage (P=0.0043). Our data suggest that modified Shirodkar cerclage as high as possible on the cervix with perioperative indomethacin and antibiotics is superior to bed rest for treatment of the midtrimester extremely short cervix (≤15 mm). We propose a randomized trial of this specific technique.