Abstract / Summary
Background and Objectives: More research is required to determine the associations of high-density lipoprotein (HDL) cholesterol concentration and particle count with mortality. The objectives of this analysis were to investigate whether there were non-linear associations of HDL concentration with all-cause mortality, whether cardiovascular disease (CVD) deaths, cancer deaths or other deaths drove these associations, and whether small, medium, or large HDL particle count was associated with mortality. Methods and Results: The sample included 47,123 men and 97,589 women from the Mexico City Prospective Study. Models were adjusted for confounders and competing risks. There were 11,264 deaths in men and 15,486 in women. There was a J-shaped association of HDL concentration with all-cause mortality in men driven by other deaths. For example, compared with the >40th-60th percentile, the hazard ratio (95% confidence interval) for other mortality was 1.27 (1.11, 1.44) in the >95th-99th percentile and 1.82 (1.45, 2.28) in the >99th-100th percentile. There was a U-shaped association in women driven by CVD deaths at lower concentrations and other deaths at higher concentrations: the hazard ratio for CVD mortality was 1.23 (1.05, 1.43) in the two lowest percentiles groups and the hazard ratio for other mortality was 1.33 (1.18, 1.50) in the two highest percentile groups. Small HDL particle count was associated with reduced risk of all-cause, CVD, and other mortality in men and women. For example, compared with the bottom quartile, the hazard ratio for all-cause mortality was 0.79 (0.73, 0.86) in the top quartile in men and 0.77 (0.72, 0.82) in the top quartile in women. Conclusions: There is a J-shaped association of HDL concentration with mortality in men and a U-shaped association in women. More research is required to confirm the novel finding that small HDL particle count reduces the risk of all-cause, CVD, and other mortality in men and women.