Developing a Preterm Birth Prediction Model in Pregnant People With HIV Using a Population-Based Cohort

To identify independent risk modifiers of and develop a prediction model for preterm births for pregnant people living with HIV (PLWH).

We analyzed the British Columbia Perinatal HIV Surveillance Database for singleton births from January 1997 to December 2022. Risk modifiers for preterm birth (<37 weeks gestational age) were identified through univariate analyses. We then completed a multivariate logistic regression using clinically relevant risk factors and risk modifiers with P < 0.1 in the univariate analysis. Using stepwise regression with the Akaike information criterion, the prediction model was developed.

Of 578 singleton pregnancies in pregnant PLWH, 111 (19.2%) had preterm births. After adjusting for history of preterm births, CD4 count at delivery, and antiretroviral regimens, independent risk factors include substance use in pregnancy (odds ratio [OR]: 1.9; 95% CI: 1.03 to 3.50; and P = 0.041), hepatitis C in pregnancy (OR: 1.98; 95% CI: 1.02 to 3.83; and P = 0.043), unsuppressed viral load at delivery (OR: 2.03; 95% CI: 1.09 to 3.79; and P = 0.026), and sexually transmitted infections or bacterial vaginosis (STIBV) in pregnancy (OR: 2.06; 95% CI: 1.04 to 4.10; and P = 0.039). A prediction model with substance use, STIBV, hepatitis C in pregnancy, unsuppressed viral load, and low CD4 count was developed Area Under the Curve (AUC) = 0.73). For PLWH without any risk factors, the predicted probability of preterm birth is 9%.

Independent risk factors associated with preterm births in pregnant PLWH include substance use in pregnancy, unsuppressed viral load, hepatitis C co-infection, and STIBV in pregnancy. Our simplified preterm birth prediction tool provides a personalized approach to caring for pregnant PLWH.