Plasma HDL Cholesterol and Risk of Myocardial Infarction: A Mendelian Randomisation Study

This Mendelian randomization study used genetic variants associated with higher HDL-C — including those affecting CETP (cholesteryl ester transfer protein), the principal determinant of HDL-C levels — as instrumental variables in 116,320 individuals to test whether genetically elevated HDL-C reduces myocardial infarction risk.

Genetically elevated HDL-C was not associated with lower MI risk. The CETP variant rs3764261, which raises HDL-C substantially, showed no effect on MI risk — and a CETP-specific allele score had no MI benefit despite raising HDL-C by 6.4 mg/dL. This contrasted with the strong protective observational association between HDL-C and MI risk, which had driven decades of HDL-raising drug development.

The Mendelian randomization result means the observational HDL-C protective association is not causal — HDL-C concentration is a marker of other metabolic states that reduce CV risk, but raising HDL-C itself does not reduce MI risk. This genetic finding has been confirmed by multiple failed clinical trials: torcetrapib (a CETP inhibitor) raised HDL-C by 72% and increased cardiovascular events due to off-target effects; dalcetrapib raised HDL-C and showed no cardiovascular benefit; evacetrapib raised HDL-C by 130% and showed no benefit.

The clinical implication is fundamental: HDL-C concentration should not be a therapeutic target. The HDL particle’s function — reverse cholesterol transport capacity — matters more than the cholesterol concentration it carries. This study also provided early validation of the Mendelian randomization approach for testing lipid causal hypotheses in very large genetic datasets.

We rate the evidence strong. A large Mendelian randomization study in 116,320 participants demonstrating that genetically elevated HDL-C does not reduce MI risk — resolving the decade-long debate about whether HDL-C is a causal cardiovascular risk factor or merely a marker.