Based on the fact that nearly 80% of all cardiovascular events occur in developing countries and that these countries have limited resources for cardiovascular disease (CVD) prevention strategies, Gaziano et al. assessed a non-laboratory-based risk algorithm to predict CVD risk in the NHANES I follow-up study cohort. The investigators replaced total cholesterol with body mass index (BMI) in the Framingham risk score and investigated whether this non-laboratory algorithm showed similar association with CVD risk over a 24 year follow-up of 6,186 men and women initially free from CVD. Based on the c-statistic, they showed that both algorithms performed equally well in terms of CVD risk prediction. In men, the c-statistic for the laboratory-based model was 0.784 and 0.783 for the non-laboratory-based model. In women, the respective c-statistics were 0.829 and 0.831. These findings suggest that obtaining information about CVD risk factors through non-invasive measures could simplify risk prediction in situations where laboratory measurements of CVD risk factors are inconvenient or unavailable. This paper was accompanied by an editorial by Mendis and Mohan who highlighted the limitations of this paper, noting that since NHANES is a US-based cohort, these results may not be applicable elsewhere in the world, especially in developing countries. They also suggested that waist circumference assessment may improve risk prediction on top of BMI and that BMI may not be the best tool to estimate obesity-associated CVD risk. Lastly, Mendis and Mohan suggested that, when it comes to treatment with either statins or anti-diabetic medications, laboratory-based measurements should be performed, even in low-income countries.
Documentation centre