BackgroundThere is substantial evidence that a low serum level of HDL cholesterol (HDLC) is a risk factor for coronary deaths. However, data on older people are scarce, and previous studies have not examined this association in relation to alcohol intake.Methods and ResultsCoronary mortality, all-cause mortality, and mortality due to alcohol and violence were related to HDLC levels among 7052 male smokers 50 to 69 years old in south and west Finland enrolled from 1984 to 1988 in the ATBC (AT, alpha-tocopherol; BC, beta-carotene) Study placebo group. During the average follow-up period of 4.7 years, 620 men died; 222 of these deaths were from coronary heart disease and 82 from causes related alcohol and violence. HDLC levels were inversely associated with coronary mortality, irrespective of age, whereas high total cholesterol was positively associated with coronary mortality among the younger men, 50 to 59 years of age, but not among the older men, 60 to 69 years old. Correction for temporal variation in HDLC measurement indicated a 43% stronger inverse association between HDLC and coronary mortality compared with that based only on a single value. The inverse association of HDLC and coronary mortality was less marked at higher levels of alcohol intake. All-cause and alcohol- and violence-related mortality were positively associated with HDLC among the younger men. All-cause mortality showed a U-shaped dose response among men ≥ 60 years old.ConclusionsPrevious studies may have underestimated the beneficial effect of high HDLC because of regression-dilution bias and the confounding effect of heavy alcohol intake. This study supports the view that, particularly among older men, lipoprotein fractions may be more appropriate for screening than total cholesterol.