A Nonlinear Association of Serum Cholesterol Level with Incident Coronary Heart Disease in Elderly Men

P62 Increased serum cholesterol has been found to be a direct risk factor for coronary heart disease (CHD) in middle aged individuals in epidemiologic follow-up studies. However, this relationship has not been consistently reproduced in elderly populations. From 1991-93, 2424 elderly (71-93 years old) Japanese American men from the Honolulu Heart Program cohort, who did not have prevalent CHD and who were not on cholesterol lowering drugs, were examined. The subsequent 6-year follow-up provided an opportunity to closely examine the relationship between their baseline cholesterol and their incidence of CHD as determined by surveillance and a physician panel review. The study revealed a significant non-linear association of the baseline cholesterol with the incidence of CHD (p=0.033). Among the 1524 elderly men with cholesterol values below 200mg/dl the age adjusted CHD rate decreased as the mean cholesterol level increased. In those with cholesterol levels of greater than or equal to 200mg/dl (900) the age adjusted CHD rates increased with increasing level of cholesterol. The lowest CHD rates were seen in men with cholesterol levels between 200 and 219 mg/dl. Multivariate adjustment for other know risk factors (BMI, smoking, hypertension, diabetes, and HDL), decreased the strength of this non-linear relationship but it remained significant (p=0.049). A measure of frailty (10% weight loss since an exam 20 years before) further reduced the strength of the cholesterol relationship slightly but did so by reducing the magnitude of the cholesterol/CHD relationship in both those higher cholesterol (positive relationship) and those with lower cholesterol (negative relationship). Thus it is unlikely that frailty can fully explain the reverse cholesterol/CHD relationship seen in those with lower cholesterol levels. In this study the majority of men already had cholesterol levels below the standard cutpoint of 200mg/dl and had the opposite pattern of risk as elderly men with cholesterols above that cutpoint. This would indicate that further data is needed before recommending lowering of cholesterol to levels below 200mg/dl in older individuals.