This article is more than 5 years old. Information may no longer be current.
Genetic variants that increase HDL levels do not decrease risk for MI
Click Here to Manage Email Alerts
Results of a study by Voight and colleagues in collaboration with several international genetic consortia, including CARDIoGRAM, have made an interesting observation: Genetic variants associated solely with increased levels of plasma HDL cholesterol are not associated with the expected decrease in risk for MI.
Genetic variants, cholesterol and associated risk
HDL and LDL cholesterol plasma concentrations are the most commonly measured and most important variables utilized in the prevention of the No. 1 killer — CVD. While it has been definitively proven that decreasing LDL cholesterol levels lowers the risk for CAD and MI, it has been claimed for decades that increased levels of HDL cholesterol also lowers the risk for CAD and MI. That claim has recently been questioned, partly because the results of most studies are confounded by other variables; for example, the drugs utilized to increase HDL cholesterol, such as statins, also decrease LDL cholesterol levels.
Robert Roberts
Genetic influences account for 60% to 70% of the variation in levels of HDL and LDL cholesterol. Many genetic variants regulating HDL and LDL cholesterol were previously identified in a large GWAS study with a sample size of more than 100,000. In that study, we confirmed 95 genetic variants of genome-wide significance to be associated with plasma lipids, of which 22 were associated with LDL cholesterol and 31 with HDL cholesterol. While many variants affected more than one lipid, some were specifically associated with HDL cholesterol.
SNPs identified in recent studies
The Voight and colleagues study was planned to determine whether genetic variants due to single nucleotide polymorphisms (SNPs) associated with an isolated increase in HDL cholesterol, without other confounding factors, is associated with decreased risk for MI. It is of note that most of our randomized, placebo-controlled clinical trials expose the individual to the intervention for 3 to 5 years. A major advantage of assessing the effect of a genetic variant utilizing Mendelian randomization is the individual will have been exposed to the genetic intervention from birth.
Under the leadership of Sekar Kathiresan, an analysis was performed using large sample sizes available from various international consortia, including the CARDIoGRAM consortium, which pursues genes related to CAD. The current study included a Mendelian randomization analysis of three types of genetic variants: (1) a SNP in the endothelial lipase gene (LIPG) associated solely with an increase of HDL cholesterol; (2) 14 common SNPs associated solely with an increase in HDL cholesterol; and (3) 13 SNPs associated with increased LDL cholesterol levels, as a positive control.
A SNP identified in LIPG substitutes asparagine for serine at nucleotide 396 (LIPG Asn396Ser) and occurs in 2.6% of the population. Carriers of LIPG Asn396Ser are associated with an increase solely of HDL cholesterol (0.14 mmol/L per copy) compared with noncarriers (P=8 x 10-3). The other plasma lipid levels, including LDL cholesterol and triglycerides, in the carriers were similar to those of the noncarriers. Based on epidemiological studies, one would expect LIPG Asn396Ser to be associated with a decrease in risk for MI of 13%. The sample size was designed with 90% power to detect a 13% reduction in risk for MI. Carriers of LIPG Asn396Ser exhibited an average increase of 0.29 standard deviation (SD) units in HDL cholesterol compared with noncarriers (P=8x10-13). Other variables such as LDL cholesterol and BP in the carriers were similar to noncarriers.
A prospective analysis was performed in 50,763 participants, of which 4,228 developed a first MI. Carriers of LIPG Asn396Ser were not associated with decreased risk for MI, despite elevated levels of HDL cholesterol (OR=1.1; P=.37).
In a case-control study involving an additional 16,685 cases of MI and 48,872 controls, LIPG Asn396Ser was also not associated with reduced risk for MI (OR=0.94; P=.41).
Finally, in a meta-analysis combining the prospective and case-control studies (20,913 cases and 95,407 controls), LIPG Asn396Ser showed no association with risk for MI (OR=0.99; P=.85).
In our previous GWAS pursuing genes related to plasma lipids involving more than 100,000 individuals, we observed 13 common SNPs of genome-wide significance (P<5x10-8) to be associated with LDL cholesterol, but with no association with triglycerides or HDL cholesterol. As a positive control, these SNPs predisposing to increased LDL cholesterol were summarized in a score and analyzed in a sample size of 53,146 cases and controls from CARDIoGRAM. Each increase of 1 SD unit in LDL cholesterol was associated with increased risk for MI, as expected from epidemiological studies (OR=2.13; P=2x10-11). A similar analysis of 14 genetic variants of genome-wide significance associated with increased HDL cholesterol and no association with LDL cholesterol or triglycerides with Mendelian randomization showed no reduction in MI (OR=0.93; P=.63). Genetic variants in the gene for cholesterol ester transfer protein (CETP) were associated with increased plasma levels of HDL cholesterol and a decreased risk for MI, but these genetic variants also decreased plasma levels of LDL cholesterol.
Consistent results were observed in this study indicating that isolated increased plasma HDL cholesterol without a decrease in LDL cholesterol was not associated with any change in risk for MI. It is also important to note that in the Mendelian randomization study, the cases were exposed to higher HDL cholesterol levels from birth for several decades. This was true for the potent isolated LIPG genetic variant, as well as several other SNPs associated with isolated increases in HDL cholesterol. These findings are in keeping with a previous study in women showing that hormone replacement therapy, which increased HDL cholesterol levels, offered no protection from CAD or MI. In the AIM-HIGH trial, adding long-acting niacin increased HDL cholesterol, but did not lower the risk for CV events.
Conclusion
There are many components to HDL and different interventions, such as exercise and alcohol, which increase plasma levels. Elucidation of the various fragments of HDL cholesterol and the interventions whereby its plasma level is increased may be necessary to determine whether there is indeed a beneficial effect from increased HDL cholesterol. In the meantime, one can no longer assume that HDL cholesterol is a surrogate for the risk of CAD or MI.
Disclosure: Roberts reports no relevant financial disclosures.