Interventions to prevent preterm delivery in women with short cervix before fetoscopic laser surgery for twin-twin transfusion syndrome

Preoperative short cervical length (CL) remains a major risk factor for preterm birth after laser surgery for twin-twin transfusion syndrome (TTTS), but the optimal intervention to prolong pregnancy remains elusive. The objective of this study was to compare secondary methods for the prevention of preterm birth in twin pregnancies with TTTS undergoing fetoscopic laser photocoagulation (FLP), in the setting of a short cervix at the time of FLP, in five North American Fetal Treatment Network (NAFTNet) centers.

This was a secondary analysis of data collected prospectively at five NAFTNet centers, conducted from January 2013 to March 2020. Inclusion criteria were a monochorionic diamniotic twin pregnancy complicated by TTTS, undergoing FLP, with preoperative CL < 30 mm. Management options for a short cervix included expectant management, vaginal progesterone, pessary (Arabin, incontinence or Bioteque cup), cervical cerclage or a combination of two or more treatments. Patients were not included if the intervention was initiated solely on the basis of having a twin gestation rather than at the diagnosis of a short cervix. Demographics, ultrasound characteristics, operative data and outcomes were compared. The primary outcome was FLP-to-delivery interval. Propensity-score matching was performed, with each treatment group matched (1:1) to the expectant-management group for CL, in order to estimate the effect of each treatment on the FLP-to-delivery interval.

A total of 255 women with a twin pregnancy complicated by TTTS and a short cervix undergoing FLP were included in the study. Of these, 151 (59%) were managed expectantly, 32 (13%) had vaginal progesterone only, 21 (8%) had pessary only, 21 (8%) had cervical cerclage only and 30 (12%) had a combination of treatments. A greater proportion of patients in the combined-treatment group had had a prior preterm birth compared with those in the expectant-management group (33% vs 9%; P = 0.01). Mean preoperative CL was shorter in the pessary, cervical-cerclage and combined-treatment groups (14-16 mm) than in the expectant-management and vaginal-progesterone groups (22 mm for both) (P < 0.001). There was no significant difference in FLP-to-delivery interval between the groups, nor in gestational age at delivery or the rate of live birth or neonatal survival. Vaginal progesterone was associated with a decrease in the risk of delivery before 28 weeks' gestation compared with cervical cerclage and combined treatment (P = 0.03). Using propensity-score matching for CL, cervical cerclage was associated with a reduction in FLP-to-delivery interval of 13 days, as compared with expectant management.

A large proportion of pregnancies with TTTS and a short maternal cervix undergoing FLP were managed expectantly for a short cervix, establishing a high (62%) risk of delivery before 32 weeks in this condition. No treatment that significantly improved outcome was identified; however, there were significant differences in potential confounders and there were also likely to be unmeasured confounders. Cervical cerclage should not be offered as a secondary prevention for preterm birth in twin pregnancies with TTTS and a short cervix undergoing FLP. A large randomized controlled trial is urgently needed to determine the effects of treatments for the prevention of preterm birth in these pregnancies.