Key result
Model with female-specific risk factors yields modest discrimination for predicting incident CVD in postpartum women.
Why the study?
Current cardiovascular disease risk scores were developed in older populations and lack sex-specific factors, limiting their accuracy in younger women.
Can a prognostic model incorporating female-specific risk factors accurately predict incident cardiovascular disease in reproductive-aged women postpartum?
Cohort (n=262,891)
Can a prognostic model incorporating female-specific risk factors accurately predict incident cardiovascular disease in reproductive-aged women postpartum?
Effect estimate: C-statistic 0.637 (95% CI 0.623-0.650)
A newly developed prognostic model incorporating female-specific risk factors demonstrated modest discrimination but good calibration for predicting incident cardiovascular disease in postpartum women.
Captured external expert commentary on this paper, strongest first. Original sources are linked where available.
“Heart disease is the leading cause of death in women, yet existing risk tools were developed in older populations and ignore factors unique to women.”
“Millions of women who give birth each year are never considered candidates for cardiovascular risk assessment simply because of their age.”
Modest discrimination limits near-term clinical use; leaves open whether female-specific factors meaningfully improve CVD prediction in reproductive-aged women.
BACKGROUND: Currently available risk scores for cardiovascular disease (CVD) were developed in older populations and do not incorporate sex-specific factors, which limits their ability to provide accurate estimates in younger women. OBJECTIVES: The objective of the study was to develop and validate a prognostic model, including female-specific risk factors, that can identify women at high risk. METHODS: We created a cohort of 262,891 women aged 15 to 45 years with 1 randomly selected delivery from 1999 to 2017 in the Clinical Practice Research Datalink database to develop models for screening women within the first-year postpartum. The primary outcome was incident CVD. The least absolute shrinkage and selection operator method and an accelerated failure time Weibull model were used to determine the inclusion of predictors and to estimate the final model. Internal validation via bootstrapping was used to estimate the optimism-corrected measures of model performance. RESULTS: A total of 943 women (0.81 per 1,000 person-years) experienced a cardiovascular event over a median follow-up of 3.8 years (Q1-Q3 1.5-7.9). Predictors in the final model included traditional CVD risk factors, along with social deprivation, polycystic ovary syndrome, prior use of oral contraceptives, depression, thyroid disorders, hypertensive disorders of pregnancy, gestational diabetes, preterm birth, small-for-gestational-age birthweight, parity, and history of pregnancy complications. Optimism-corrected performance median (Q1-Q3) measures showed modest discrimination (C-statistic: 0.637 [0.623-0.650]) and good calibration (slope: 0.919 [0.905-0.930]). CONCLUSIONS: Although the model showed modest predictive accuracy and performance, the findings highlight the importance of sex-specific risk factors for identifying women at high risk of CVD in the postpartum period.
No takes yet. Share an insight, caveat, or question.
Grandi et al. (2026) conducted a cohort in Cardiovascular disease (n=262,891). Prediction model for cardiovascular risk was evaluated on Incident CVD (C-statistic 0.637, 95% CI 0.623-0.650). A prognostic model incorporating female-specific risk factors demonstrated modest discrimination (C-statistic 0.637; 95% CI 0.623-0.650) for predicting incident CVD in postpartum women.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: