No firm conclusions about HDL cholesterol can be drawn from JUPITER sub-analysis

In the Lancet study, Paul Ridker and colleagues, from Brigham and Women’s Hospital (Boston, MA, USA), undertook a retrospective post-hoc analysis of the JUPITER trial. The results show that if a normal, healthy individual has level of low density lipoprotein (LDL), known as “bad cholesterol”, substantially lowered with a potent statin, then the level of HDL “good cholesterol” in that person no longer bears any relation to the remaining cardiovascular risk.
The original JUPITER trial was designed to answer the critical question of whether rosuvastatin prevents cardiovascular disease among healthy people with normal LDL cholesterol levels, but increased levels of high-sensitivity C-reactive protein, a marker of chronic low level inflammation, considered a new risk factor for cardiovascular events.
The current Lancet study showed that when 17,802 subjects were divided into quartiles of HDL cholesterol concentrations, HDL cholesterol concentrations were inversely related to vascular risk at the end of study for individuals randomised to placebo, with the top quartile having a 46% reduced risk compared to the bottom quartile (p=0.0039). In contrast, however, among those subjects given active treatment with rosuvastatin, vascular risk was calculated to be similar for subjects in both the top and bottom HDL quartiles (p=0.82)
“Although measurement of HDL–cholesterol concentration is useful as part of initial cardiovascular risk assessment, HDL-cholesterol concentrations are not predictive of residual vascular risk among patients treated with potent statin therapy who attain very low concentrations of LDL cholesterol,” the authors of the study conclude.
ESC spokesperson Professor Dan Atar, from Oslo University Hospital, Norway, believes there are dangers in interpreting the study as showing that raising HDL levels produces no beneficial cardiovascular effects. “It’s a matter of statistics. If you’re looking at populations with a very low incidence of cardiovascular events, and then with an intervention of any kind you reduce the risk of events even further, it’s logical that you’ll washout the influence of any other effect. These patients already have achieved such low levels of LDL that no other marker will prevail as a predictor of the few remaining events.”
He added that he had concerns that readers of the paper might not appreciate that more data was needed before the scientific community could make a qualified decision about whether raising HDL levels was beneficial or not. “With subgroup analyses, such as the one presented here, you just can’t make such judgements,” he said.
In fact, previous studies, such as the Helsinki Heart Study and the VA-HIT Study , have been successful in raising HDL and reducing cardiovascular events, using gemfibrozil, however this agent also concomitantly lowers LDL. Additionally, the drug nicacin has been shown to be effective at elevating HDL and reducing cardiovascular morbidity in the Coronary Drug Project Study (5), but this strategy could not easily be implemented into clinical practice due to unpleasant side effects, notably flushing. It is hoped that laropiprant, a novel flushing pathway inhibitor, will overcome this limitation.
A clearer indication of the benefit of raising HDL, Atar added, will come from the ongoing phase III Heart Protection Study 2-Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2-THRIVE), which has enrolled 25,000 patients to investigate whether the combination of niacin/laropiprant can further reduce the risk for myocardial infarction, stroke and the need for revascularisation in patients already treated to lower LDL.