Prediction of preterm birth in multiple pregnancies: development of a multivariable model including cervical length measurement at 16 to 21 weeks' gestation.

OBJECTIVE

To develop a multivariable prognostic model for the risk of preterm delivery in women with multiple pregnancy that includes cervical length measurement at 16 to 21 weeks' gestation and other variables.

METHODS

We used data from a previous randomized trial. We assessed the association between maternal and pregnancy characteristics including cervical length measurement at 16 to 21 weeks' gestation and time to delivery using multivariable Cox regression modelling. Performance of the final model was assessed for the outcomes of preterm and very preterm delivery using calibration and discrimination measures.

RESULTS

We studied 507 women, of whom 270 (53%) delivered < 37 weeks (preterm) and 66 (13%) < 32 weeks (very preterm). Women with cervical length < 30 mm delivered more often preterm (hazard ratio 1.9; 95% CI 0.7 to 4.8). Other independently contributing predictors were previous preterm delivery, monochorionicity, smoking, educational level, and triplet pregnancy. Prediction models for preterm and very preterm delivery had a c-index of 0.68 (95% CI 0.63 to 0.72) and 0.68 (95% CI 0.62 to 0.75), respectively, and showed good calibration.

CONCLUSION

In women with a multiple pregnancy, the risk of preterm delivery can be assessed with a multivariable model incorporating cervical length and other predictors.