A McGill University–led team developed a prediction model integrating pregnancy-related factors to identify cardiovascular disease risk in women aged 15 to 45.
A McGill University–led research team has developed and validated a prediction model designed to identify cardiovascular disease (CVD) risk in reproductive-aged women—a population largely overlooked by existing risk stratification tools. The model, published in JACC: Advances, incorporates pregnancy-related and female-specific risk factors that conventional tools fail to capture, potentially enabling earlier intervention in a group historically undertreated for heart disease.¹
"Millions of women who give birth each year are never considered candidates for cardiovascular risk assessment simply because of their age," said coauthor Kristian Filion, Professor in the Departments of Medicine and of Epidemiology, Biostatistics, and Occupational Health at McGill University. The gap is clinically consequential: heart disease remains the leading cause of death in women globally, yet most validated risk calculators were designed and calibrated in older cohorts.
Using routinely collected health data from more than 260 000 women in the United Kingdom aged 15 to 45 who had delivered at least one child, researchers developed and internally validated a CVD risk prediction model. Participants were followed for approximately four years postpartum, according to the study.
KEY FACTS
The model incorporated several predictors not included in standard risk tools, including hypertensive disorders of pregnancy, gestational diabetes, preterm birth, polycystic ovary syndrome (PCOS), depression, thyroid disorders, oral contraceptive use, and social deprivation. The study was published May 27, 2026, and was supported by the Canadian Institutes of Health Research.¹
The researchers reported that the model successfully identified younger women who may carry a meaningfully elevated CVD risk earlier than conventional frameworks would predict—a population often assigned low-risk status by default due to age alone.
Current cardiovascular risk calculators—including the Pooled Cohort Equations used in US guidelines—are derived predominantly from middle-aged and older adult populations and do not systematically account for obstetric history or reproductive health factors.² The postpartum period represents a missed opportunity for risk stratification, particularly given growing evidence linking adverse pregnancy outcomes (APOs) such as preeclampsia and gestational diabetes to accelerated atherosclerosis and incident CVD years later.³
Senior author Robert Platt, Professor in the Department of Epidemiology, Biostatistics, and Occupational Health, noted that while the association between pregnancy complications and future cardiac risk is established, no validated tool has existed to quantify individualized risk in this younger cohort.
The model addresses a plausible and well-documented biological and epidemiological gap. However, internal validation within a single national dataset—the UK Biobank or analogous administrative data source—limits immediate generalizability. The researchers acknowledge the next step is external validation in North American populations, including Canada and the United States, where healthcare infrastructure, demographic composition, and obstetric practice patterns differ from the UK.
Whether incorporation into postpartum clinical workflows would translate to measurable reductions in cardiovascular events remains to be demonstrated in prospective or interventional studies. The tool's value will ultimately depend on whether risk identification prompts evidence-based management changes—lifestyle counselling, lipid monitoring, or specialist referral—and whether those interventions reduce downstream morbidity.
The study was conducted using UK health data, and external validation in Canadian and US cohorts has not yet been completed. The approximately four-year follow-up window may be insufficient to capture CVD events that manifest over decades. The computational modeling methodology also does not establish causality between the identified predictors and cardiovascular outcomes. The research team's stated long-term goal is integration of a practical risk calculator into electronic health records to facilitate point-of-care identification of higher-risk patients.¹
References
