Background: Previous spontaneous preterm birth (sPTB) is a strong risk indicator for recurrent preterm birth (PTB). Cervical cerclage is an accepted intervention to prevent recurrent PTB in high risk patients. Cervical pessary might be a less invasive alternative. The objective of this study is to determine whether a cervical pessary is non-inferior to cerclage in the prevention of recurrent PTB.

Methods: We performed an international, open-label, non-inferiority, randomised controlled trial in 21 hospitals between March 2014 and December 2022. We included singleton pregnancies with an indication for cerclage based on either multiple previous sPTBs <34 weeks or with a previous sPTB <34 weeks and an asymptomatic mid-trimester short cervix (≤25 mm). Randomisation was 1:1, stratified by centre and indication, to cervical pessary or vaginal cerclage. Primary outcome was PTB <32 weeks. Secondary outcomes included (s)PTB rates, obstetric, and maternal outcomes and a composite of adverse perinatal outcomes including perinatal mortality and severe neonatal morbidity. Analysis was by intention-to-treat. Treatment effect was expressed as relative risk (RR), absolute risk difference (aRD) and 95% confidence intervals (CI). Sample size was calculated at 400 participants with a non-inferiority margin for pessary of 10%, meaning that non-inferiority is proven if the upper limit of the CI of the risk difference is <10%. Trial registration at ICTRP: NL-OMON26958.

Findings: We randomised 261 participants to pessary (n = 133) or cerclage (n = 128). After the third interim analysis (n = 228 participants), recruitment was halted due to safety concerns and the apparent challenge in establishing non-inferiority of pessary treatment. PTB <32 weeks occurred in 44/130 cases after pessary vs 30/125 cases after cerclage (33.8% vs 24.0% aRR 1.4, 95% CI 0.95-2.1, p = 0.09, aRD 9.8% 95% CI -1.2 to 20.9). The composite of adverse perinatal outcomes occurred in 42 cases after pessary compared to 29 cases in cerclage (32.2% vs 23.2%; RR 1.4 95% CI 0.93-2.1 p = 0.1) and consisted mainly of perinatal death (22.3% vs 14.4% RR 1.5 95% CI 0.9-2.6 p = 0.1).

Interpretation: Non-inferiority of cervical pessary compared to cerclage in preventing recurrent PTB <32 weeks was not proven. Cerclage is the recommended treatment.

Funding: ZonMw (#837002406), a Dutch Organisation for Health Research and Development.