May 27th, 2011

AIM-HIGH Halted: A Death Knell for the HDL Hypothesis? Six Experts Weigh In

Earlier this week, the National Heart, Lung and Blood Institute stopped the randomized clinical trial known as AIM-HIGH (Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides: Impact on Global Health). It had been designed to test the addition of high-dose, extended-release niacin to statins in people who were at risk for cardiovascular events, had well-controlled LDL levels, but also had low HDL and elevated triglyceride levels (See CardioExchange’s news item describing details about the trial and the decision to halt it.)

We at CardioExchange put this question to one of the principal investigators of AIM-HIGH and to five other experts from different domains within the cardiovascular-medicine and scientific communities:

Is the halting of AIM-HIGH the death knell for the “HDL hypothesis”: the idea that clinical benefit can be achieved by raising HDL-cholesterol levels?

Here’s how our interviewees responded. We urge you, as members of the CardioExchange community, to engage them — and one another — in dialogue about this important research development.

No, emphatically, this trial does not sound the death knell for the HDL hypothesis. VA-HIT clearly validated that one can reduce death, MI, and stroke with gemfibrozil 12 to 13 years ago, absent any LDL change. But the baseline LDL in that earlier study was 111 mg/dL; in AIM-HIGH it was 71 mg/dL — a huge difference of 40 mg/dL. We must continue to pursue research in higher-risk ACS and acute-MI patients whose HDL levels are low — and perhaps include more de novo or statin-naive patients, as we often see in real-world practice.  —William E. Boden, MD, co–principal investigator of the AIM-HIGH trial; Medical Director of Cardiovascular Services, Kaleida Health; Chief of Cardiology, Buffalo General Hospital and Millard Fillmore Gates Circle Hospital; Professor of Medicine and Preventive Medicine, University at Buffalo

Another one bites the dust! As I have maintained before, HDL is a very complex molecule that is involved in many metabolic pathways. Simply modifying the molecule qualitatively or quantitatively is not ipso facto going to translate into cardioprotective effects. It is hard to argue against a randomized, controlled outcomes-based trial as being an important tool to bust myths and entrenched beliefs! —Sanjay Kaul, MD, Director, Cardiology Fellowship Training Program, Cedars-Sinai Medical Center, Los Angeles; Professor of Medicine, UCLA School of Medicine

The discontinuation of this trial is disappointing. Of course, judging from a press release alone is extremely difficult for someone not associated with the study. Did the complexity of the relatively small study population of 1718 Niaspan-treated participants — heterogeneous risk factors (metabolic syndrome, high blood pressure, diabetes) and varying drug treatments (e.g., with and without ezetimibe) —  influence the conclusion that niacin was not effective in preventing heart attacks and strokes? If so, how? (It’s somewhat surprising to see the ILLUMINATE trial mentioned in the NIH press release without noting that off-target effects are likely to have played a major role in that study’s results.)  —Monty Krieger, PhD, Whitehead Professor of Molecular Genetics, Department of Biology, Massachusetts Institute of Technology (basic-science researcher who discovered the first HDL receptor)

The hypothesis is not dead, but it is certainly on life support. It is clear that HDL “raising” is too simplistic and that HDL manipulation is a more appropriate potential target. Nevertheless, the enticing combination of niacin added to statin therapy to improve cardiovascular risk has been dealt a serious blow. It can be resurrected only by an enticing subgroup analysis from AIM-HIGH or by a clear benefit demonstrated in HPS2-THRIVE.  —Amit Khera, MD, MSc, Assistant Professor, UT Southwestern Medical Center

The HDL hypothesis may still be viable from the standpoint of future research, but this study is another disappointment to all of us who hope to further improve outcomes for patients at risk for cardiovascular disease. From the perspective of contemporary clinical care, we must step back and recognize that the evidence for lipid-modifying therapy, in terms of meaningful patient outcomes, supports only the use of statins, with little current role for agents that raise HDL. Other lipid-modifying therapies, including niacin (which failed in AIM-HIGH) and fibrates simply should not be used regularly, if at all. My perspective would change if a study or studies showed that lipid-modifying therapies that raise HDL have important clinical benefits beyond improving a patient’s lab-test result.Frederick A. Masoudi, MD, MSPH, Associate Professor of Medicine, University of Colorado Denver; Denver Health Medical Center

This study does not overturn the strong evidence from animal models that some forms of HDL raising are likely beneficial. Nor does it prove that other approaches such as CETP inhibition, which involve both LDL lowering and HDL elevation, will fail. AIM-HIGH showed that the HDL-raising effect of extended-release niacin does not benefit patients whose LDL is tightly controlled at 40 to 80 mg/dL. This could indicate that the anti-atherogenic effects of HDL are not important in the absence of an ongoing atherogenic stimulus. Or it may imply that the particular mechanism of HDL elevation by niacin, which is unknown, is not beneficial. The data highlight, once again, the importance of testing specific hypotheses in carefully designed clinical trials. —Alan R. Tall, MD, Tilden Weger Bieler Professor of Medicine, Columbia University School of Medicine

12 Responses to “AIM-HIGH Halted: A Death Knell for the HDL Hypothesis? Six Experts Weigh In”

  1. The stroke result reflecting a statistically significant 2.3-fold increase in risk (unadjusted p=0.01 by my calculation) is somewhat concerning. Even though the study protocol lists ischemic stroke as a component of the primary composite endpoint, I am curious to find out how the stroke endpoint was adjudicated. Was it based on confirmed ischemic type on brain imaging? I can’t think of a convincing explanation for increase in ischemic strokes, but am open to the possibility of increased hemorrhagic strokes especially if on-treatment LDL levels were very low. It can be hard to differentiate the two unless confirmed by brain imaging studies. A related question is what were the on-treatment LDL levels in the treatment groups?

  2. Disappointing is an understatement… Am among those who believe “something may have been overlooked”.

    HDL biology is complex, and HDL quality improvement remains viable.

    The importance of primordial prevention, even for those who reach aggressive LDL targets with statins is underscored. In the group with low HDL, aerobic exercise and both total calorie and carbohydrate calorie control are still at the top of the list. There is, literally, no free (big) lunch.
    Richard Kones

  3. Anil Virmani, MD, DRM says:

    The results are indeed very disappointing. They further strenghthen the belief that by relying too much on surrogate markers, we have always been going up the wrong path !To wait for outcome studies to guide our treatment strategies is not always practical and, what should we do ?

    Competing interests pertaining specifically to this post, comment, or both:

  4. Two very separate questions are being muddled together here. First, what is the effect of niacin, when added to a statin, on atherothrobmotic cardiovascular disease events? Second, what is the effect of raising HDL levels, in whatever setting by whatever mechanism, on CVD events?
    Addressing the first question, many prior studies of niacin have shown CVD benefit; however, none did so in addition to background statin therapy. It seems unlikely (but might eventually prove true) that niacin, which seems to work in monotherapy and with a bile sequestrant, cannot reduce CVD when added to a statin. It seems more likely that extra ezetimibe, higher statin doses and possibly greater dropout rates in the placebo group would explain the unexpected loss of benefit. Conceivably, niacin works LESS well in patients with low baseline HDL–a counter-intuitive concept which was never tested before AIM HIGH. Addressing the 2nd question, AIM-HIGH was not a definitive test of HDL-raising because of the above confounders and because niacin has other lipid effects. Surely, there is much yet to learn about low HDL levels and their treatment. In neither case can we give intelligent answers without further data and analysis of AIM-HIGH.
    I strongly recommend that we avoid the urge to jump to conclusions which may make great sound bites but do not serve our patients well. More complete AIM-HIGH data and analyses will be available later this year. Further, in a couple of years we will have results from HPS2-THRIVE, which tests the same general question as AIM-HIGH but which is much larger and avoids many limitations of AIM-HIGH. Patients with residual low HDL-C despite aggressive statin therapy are at high CVD risk, and I suggest that we not be eager to dismiss evidence-based treatment options prematurely.

    Competing interests pertaining specifically to this post, comment, or both:
    Speaker and/or consultant for Abbott, Amarin, Daiichi-Sankyo, Essentialis, GSK, Lupin, Merck, Regeneron and Sanofi-Aventis.

  5. Robin Motz, M.D., Ph.D. says:

    I always wait for the second, confirming study. Remember the study that “proved” that coffee but not tea caused cancer of the pancreas?

  6. A lesson I learned a long time ago was not to over react to a single study. In this case, a single, relatively small, uniquely atypical population study failed to reach statistical significance for an outcome. There are too many reasons that this could be the case. Compliance? Other risk factors? Blood pressure? Diabetes?

    It is unfortunate that this study was stopped without a compelling reason. I caution clinicians to judge slowly and await the sub-study analysis.

    Perhaps the very high statin doses did something to counter the benefit of the niacin.

    I know that in my practice, I use a lot of niacin with low dose statins and fish oil derived omega-3 fatty acids and I see very very few heart attacks or ischemic strokes.

    Competing interests pertaining specifically to this post, comment, or both:

  7. What do you think now about the eagerly awaited CETPi (anacetrapib and dalcetrapib) studies? In vivo, they increased HDL up to 150%, will that correlates in clinical outcomes?

  8. Robin Motz, M.D., Ph.D. says:

    We always need more that one study whether or not it confirms our favorite hypothesis. Just consider the recent study of Zetia, which lowered the cholesterol level (as it promised the fda it would, compared ton an allizatir whidh just srang at and driwned its vuictims,

  9. Can’t say anything about anacetrapib yet… to me, AIM-HIGH is not about the HDL hypothesis – it is about the importance of testing each drug for its net effect.It is still plausible that a drug that raises HDL could lower risk… but we cannot assume the benefit based on the modification of the risk factor. And that goes for other risk factors too.

  10. Dr. Motz, were you taking 2 grams of Niaspan(tm) when you wrote your second sentence?

  11. Joseph Wildman, MD says:

    While it may be premature to abandon anacreapib and the whole idea of raising HDL, surely a few conclusions CAN be drawn:
    1. Patients similar to those in the study, on statins with low HDL, should NOT be given niacin.
    2. The Abbott advertising campaign for Niaspan needs to be re-evaluated.
    3. The terms of the FDA approval of Niaspan need to be re-evaluated.
    4. Simplistic concepts of “correcting” suboptimal lipid profiles need to be vewed with great skepticism.

    Competing interests pertaining specifically to this post, comment, or both:
    No conflicts

  12. Brett Forge, MB BS FRACP says:

    There were no details of the treatments or the lipid results. How much did the HDL increase? This trial may just be showing that ezetrol is as good as niacin, or that the control group finished up getting more statins in which case niacin is no better than increasing the dose of statin. Or it appears there was an unexpected excess of strokes (many were not even on niacin at the time) which may be an artefact and that there was in fact a reduction in CVS events. Seems very bad to stop the trial when no harm has been done. The safety monitoring panel is presumably there to monitor safety, but there was no increase in events so why are they intervening?

    Competing interests pertaining specifically to this post, comment, or both: