High Density Lipoprotein pathway as a therapeutic target for coronary heart disease: individual participant meta-analysis in 28,597 individuals with 4197 coronary events

Importance: Cholesterol content in high-density lipoprotein particles (HDL-C) is associated inversely with coronary heart disease (CHD), but findings from Mendelian randomization studies and randomized trials of HDL-C raising drugs have questioned whether this link is causal. However, these analyses do not exclude a causal role for specific HDL sub-fractions of different density, mobility, size and composition. Objective: To determine whether sub-components of the HDL pathway exhibit differing relationships with CHD risk. Design: In seven longitudinal studies, we used factor analysis to reduce 21 measures of HDL particle size and lipid content to a smaller number of factors representing different components of the HDL pathway. We constructed factor scores and modelled their associations on CHD risk in adjusted Cox regression analyses. We pooled results using random-effects meta-analysis. Setting: Seven population-, individual-, occupational- or community-based longitudinal studies in the UK and Finland. Participants: 28,597 participants (49% female, mean age 59.6 years) contributed to the analysis. Exposures: Sub-components of the HDL pathway, characterized by 21 measures of HDL size and lipid content based on nuclear magnetic resonance spectroscopy. Main Outcomes: Incident fatal or non-fatal CHD. Results: We identified 4 HDL components with highly replicable across studies; 3 were indices of particle size/composition (extra-large (XL), large (L) and medium/small (MS)), and the other an index of triglycerides (TG) carried in HDL of all sizes. After up to 17 years of follow-up, 4179 incident CHD cases occurred. After adjusting for age, sex, ethnicity, smoking, systolic blood pressure, body mass index, diabetes and LDL-C, higher levels of the XL and MS factors were linked to a reduced risk of CHD (hazard ratio per 1 standard deviation (SD) increase 0.88 [95% CI 0.85, 0.92] and 0.91 [0.87, 0.94]). In contrast, a SD increase in the level of the TG factor was associated with increased risk of CHD (1.10 [1.07, 1.14]). Conclusions and Relevance: We found qualitative differences between sub-components of the HDL pathway and the risk of developing CHD. Discovery of the biological determinants of these components, possibly through genetic analysis, will facilitate selection of drug targets and inform trial design.