Clustering of cardiovascular disease risk factors

In a population-based cardiovascular disease study we analyzed the associations among five risk factors—cholesterol, triglycerides, blood pressure, obesity, and cigarette smoking. Two methods were used: standard correlation analysis, and a percentile analysis method limited to associations at higher levels of these risk factors. The study population, 4,839 men and women aged 30–89, showed significantly positive standard and age-adjusted correlation coefficients for all comparisons between any two risk factors except for those comparisons involving smoking. In the percentile analysis, subjects with moderate or greater (?70th percentile) or high (?90th percentile) levels on one of these risk factors showed clustering of elevations in other risk factors in that observed/expected ratios were generally greater than unity, again excepting smoking comparisons. Clustering was strongest in subjects at the highest levels of these risk factors, a phenomenon which has not, to our knowledge, previously been reported. Because of the particularly high risk of subjects with several risk factors, this finding should caution clinicians who identify a patient with an elevation of one of these risk factors to evaluate the patient carefully for elevations of other risk factors.