Development and validation of a nomogram for individualized prediction of spontaneous extremely preterm birth at < 28 weeks in twin pregnancy: a retrospective cohort study.

PURPOSE: To develop a nomogram to predict spontaneous preterm birth at < 28 weeks in pregnant women with twin pregnancies.

METHODS

We retrospectively studied the medical records of twin-pregnancy women with asymptomatic cervical dilation or cervical shortening between December 2015 to February 2022 in two hospitals. Data from one center was used to develop the model and data from the other was used to evaluate the model.

RESULTS

A total of 270 twin pregnancies were enrolled in the study. We incorporated 4 items (cervical length, cervical dilation, C-reactive protein and the use of cerclage) to build the 28-week nomogram with satisfactory discrimination and calibration when applied to the validation sets. The C index for the 28-week nomogram in the development and external cohort was 0.88 (95% CI, 0.84-0.93) and 0.89 (95% CI, 0.80-0.98), respectively. The nomogram reached a sensitivity of 70.70%, specificity of 97.10%, positive predicted value of 95.61% and negative predicted value of 78.77%. Moreover, the decision curve analysis indicated that the nomogram showed positive clinical benefit.

CONCLUSION

We developed and validated a nomogram with good performance in predicting individual risk of spontaneous preterm birth at < 28 in twin pregnancy.