Abstract
BACKGROUND: Few cohort studies examine the association of cardiovascular risk scores with cardiovascular mortality in India. This study assessed the association of baseline Framingham Risk Scores (FRS) with 10-year incidence of fatal CVD events in rural Tamil Nadu, India. METHODS: Using a retrospective cohort study design, we analysed the association of baseline FRS categories assessed in 2011-12 through a STEPS risk factor survey, with CVD deaths over 10 years. Causes of death for the survey participants aged 30-64 years at baseline (2011-12), were obtained through established vital event surveillance, while baseline FRS CVD scores were calculated using original and published recalibration equations. RESULTS: 3418 participants (1480 males, 1938 females), free of CVD at baseline, were followed up for mortality for 10.22 years (median). The CVD mortality rate was 3.01 per 1000 person-years among males and 1.36 in females. Those with baseline original lipid-based FRS ≥20 % had higher CVD mortality risk (Hazard Ratio males: 11.18, 95 % CI: 4.67-26.79; females: 17.51, 95 % CI: 6.07-50.55) compared to those with scores <10 %, with similar results using recalibrated scores. Discrimination statistics (Harrell's C) were 0.755 and 0.751 for original and recalibrated lipid-based scores in males, compared to 0.734 and 0.842 in females. CONCLUSIONS: FRS had good predictive validity for cardiovascular mortality in a rural Indian population, confirming its clinical usefulness.