Comparison of modifiable determinants of lipids and lipoprotein levels among African-Americans, Hispanics, and Non-Hispanic Caucasians > or =65 years of age living in New York City.

Information on determinants of plasma lipids and lipoproteins and how these factors would differ among race/ethnic groups in elderly populations is scarce. We studied cross-sectionally the distribution and predictors of lipids and lipoproteins in 1,118 free-living elderly subjects (> or =65 years of age), in a multiethnic urban community (22% non-Hispanic Caucasian, 34% African-American, and 44% Hispanic). Mean levels of total cholesterol, total/high-density lipoprotein (HDL) cholesterol ratio, and triglycerides decreased with increasing age (p <0.001). Low-density lipoprotein cholesterol and total cholesterol were lower among men, whereas women had higher levels of HDL cholesterol (p <0.001). Hispanics had lower low-density lipoprotein cholesterol, total cholesterol, and HDL cholesterol levels, whereas African-Americans had a lower total/HDL cholesterol ratio and triglyceride levels along with higher HDL cholesterol levels (p <0.001). Diabetes was more prevalent among Hispanics and African-Americans (p = 0.002), and body mass index was higher in African-Americans (p = 0.009). Hispanics were less likely to drink alcohol (p <0.0001), but more likely to drink larger amounts of coffee (p <0.0001). A greater proportion of African-Americans were active smokers (p <0.001). In multivariate regression models, body mass index was a significant independent predictor of total cholesterol (beta = 0.74, p <0.001). Waist circumference predicted lower HDL cholesterol levels (0.57 mg/dl lower per inch, p <0.001) and a higher total/HDL cholesterol ratio (beta = 0.05, p <0.001). Alcohol intake (>2 g/day) predicted higher HDL cholesterol (beta = 6.20, p <0.001). Diabetic status predicted lower HDL cholesterol (beta = -2.47, p <0.05), higher total/HDL cholesterol ratio (beta = 0.35, p <0.001), and triglycerides (beta = 37.10, p <0.001). Physical activity, coffee intake, and a calorie-adjusted atherogenicity index did not show independent predictive value. These findings indicate that obesity, alcohol intake, and diabetes mellitus are potentially modifiable independent determinants of lipids and lipoprotein levels in an elderly multiethnic population.