Risk Score Development to Predict Postpartum Cardiovascular Disease Incorporating Pregnancy and Nonpregnancy Factors:...

Area of Science:

• Cardiology
• Public Health
• Reproductive Health

Background:

• Pregnancy complications are linked to future cardiovascular disease (CVD).
• No existing models identify postpartum individuals at high risk for CVD.
• Severe maternal morbidity necessitates better risk stratification post-delivery.

Purpose of the Study:

• To develop a predictive risk score for postpartum CVD.
• To utilize pregnancy and non-pregnancy risk factors for CVD prediction.
• To generate likelihood ratios (LRs) for postpartum CVD hospitalization.

Main Methods:

• Utilized Canadian national administrative hospitalization data (2008-2021).
• Employed Cox proportional hazards models to predict CVD hospitalization.
• Generated risk scores and LRs based on model coefficients; assessed performance using C-statistics.

Main Results:

• A cohort of 1,992,972 individuals was analyzed with a median follow-up of 3.7 years.
• Likelihood ratios for CVD within 1 year varied significantly by risk score (<0: 0.55, 0 to <3: 1.42, ≥3: 7.81).
• The multivariable Cox model achieved a C-statistic of 0.669.

Conclusions:

• The developed risk score offers an initial framework for estimating future cardiac risk postpartum.
• Limitations include the absence of sociodemographic, clinical, and behavioral data, impacting predictive performance.
• This score may guide the intensity and necessity of postpartum CVD surveillance.