Framingham ten-year general cardiovascular disease risk: Agreement between BMI-based and cholesterol-based estimates in a South Asian convenience sample

Abstract

The goal of this analysis was to determine the agreement between body mass index-based and cholesterol-based ten-year Framingham general cardiovascular disease risk scores among a convenience sample of 773 South Asian Canadian adults attending community-based screening clinics. Scores were calculated using age, systolic blood pressure, antihypertensive use, current smoking, diabetes, and total cholesterol and high density lipoprotein (for cholesterol-based risk) or height and weight (for body mass index-based risk). Mean risk score differences (body mass index-based risk minus cholesterol-based risk) were estimated using paired t-tests. Bland-Altman plots were used to assess agreement between scores. Finally, agreement across risk categories (low [<10%], moderate [10% to <20%], high [> = 20%]) was examined using the kappa statistic. Average agreement between the two risk scores was quite good overall (mean differences of 0.6%for men and 0.5%for women), but increased to about 3% among participants 60-74 years of age. However, Bland-Altman plots revealed that the differences between the two scores and the variability of the differences increased with increasing average 10-year risk. In terms of clinical importance, the limits of agreement were reasonable for women < 60 years (95%confidence interval: -3.2% to 3.1%), but of concern for women 60-74 years (95%confidence interval: -6.0% to 12.3%), men < 60 years (95%confidence interval: -7.1% to 7.3%) and men 6-074 years (95%confidence interval: -13.8% to 18.8%). Agreement across categories was moderate for most sex and age groups examined (kappa values: 0.51 for women < 60 years, 0.50 for women 60-74 years, 0.65 for men < 60 years), except for men 60-74 years, where agreement was only fair (kappa = 0.26). In light of these disagreements, evaluation of a participant's change in cardiovascular disease risk over time will necessitate use of the same risk score (i.e., either body mass index-based or cholesterol-based) at all screening sessions.