AIM-HIGH: Results Raise Controversy Over Early Stopping

November 15, 2011

November 15, 2011 (Orlando, Florida) — Final results of the AIM-HIGH trial appear to suggest that the signal of increased ischemic stroke with niacin, which was one of the reasons the study was stopped early, could have been the play of chance, with the final p value for ischemic stroke coming in at a nonsignificant 0.11.

This has provoked a backlash from researchers in the field against the National Institutes of Health (NIH) sponsors of the study who made the decision to terminate the trial.

Dr Steven Nissen (Cleveland Clinic, OH), who was not involved in AIM-HIGH but is a key player in lipid research, was particularly vocal on this issue. He commented to heartwire : "I do not agree with the decision to stop this trial. It was completely inappropriate. The NIH sponsors saw a weak signal of stroke and panicked, and when all the data have come in, this doesn't appear to be an issue. Now we have lost the opportunity to properly answer the very important question of whether niacin adds any benefit in this population with low LDL levels."

AIM-HIGH, Over and Out

The AIM-HIGH trial randomized 3414 patients with established heart disease, low HDL levels, and raised triglycerides to extended-release niacin (1500–2000 mg per day) or placebo. All patients also received simvastatin plus ezetimibe if needed to maintain LDL levels below 80 mg/dL (2.07 mmol/L).

The trial was stopped earlier this year after a mean follow-up of three years. The statement from the National Heart Lung and Blood Institute at the time said the trial had been stopped because niacin was showing no additional benefits over placebo and there was also a small, unexplained increase in ischemic stroke in the niacin group.

Results presented today at the American Heart Association (AHA) 2011 Scientific Sessions and published online simultaneously in the New England Journal of Medicine show that at two years, niacin therapy had increased HDL levels from a median of 35 to 42 mg/dL, lowered triglyceride levels from 164 to 122 mg/dL, and lowered LDL levels from 74 to 62 mg/dL.

The primary end point--the first event of a composite of CHD death, nonfatal MI, ischemic stroke, hospitalization for ACS or symptom-driven coronary or cerebral revascularization--was similar in the two groups, occurring in 282 patients (16.4%) in the niacin group vs 274 patients (16.2%) on placebo. There was also no difference in two secondary composite end points.

End points Niacin (%) Placebo (%) HR (95% CI) p
Primary end point 16.4 16.2 1.02 (0.87–1.21) 0.80
CHD death/ nonfatal MI/ ischemic stroke/ high-risk ACS 9.3 10.0 1.08 (0.87–1.34) 0.49
CHD death/ nonfatal MI/ ischemic stroke 8.1 9.1 1.13 (0.90–1.42) 0.30

Focus on the Low LDL Level

Speculation on the reason for the lack of benefit with niacin is focusing on the low LDL level achieved in the trial.

Lead investigator Dr William Boden (University at Buffalo School of Medicine, NY) commented to heartwire : "if you are starting off with an LDL of 70, it may not be possible to show any incremental benefit." He also pointed out that patients in the trial were very well-treated in terms of other medical therapy, and many patients had been on statins for several years. But he said these patients were not representative of the real world, where only about 15% to 20% of patients actually get their LDL levels down to 70. "I would urge people not to overreact to these results. They should not be extrapolated to patients with LDLs of 100 or more, which is much more the norm. We must not throw the baby out with the bathwater."

But Boden and Nissen both stressed that reducing LDL to 70 with high-dose statins is still the first priority. Nissen commented: "I am less likely to use niacin after this trial, but I would still use it in patients who can't get to LDL goals with statins."

In an editorial accompanying the paper [2], Dr Robert Giugliano (Brigham and Women's Hospital, Boston) says: "Given the lack of efficacy shown in this trial, the frequent occurrence of flushing with niacin therapy that some patients find intolerable, and the unresolved issue of ischemic stroke, one can hardly justify the continued expenditure of nearly $800 million per year in the US for branded extended-release niacin." He adds, however, that "before holding a final retirement party for niacin, it would be appear to more prudent to assign it to part-time work such as in statin-intolerant patients."

Stroke: "A Causal Association or the Play of Chance"?

On the stroke issue, the AIM-HIGH investigators report in the paper that ischemic stroke occurred as the first event in 27 niacin patients (1.6%) vs 15 placebo patients (0.9%). However, eight of the strokes in the niacin group occurred between two months and four years after discontinuation of niacin treatment. "When all the patients with ischemic stroke were considered, rather than just those in whom the stroke was the first study event, a nonsignificant higher trend persisted in the niacin group," they say. They add: "The overall rate of ischemic stroke was low, and it is not clear whether the increase seen with niacin reflects a causal association or the play of chance."

Boden added to heartwire : "When the NIH stopped the trial, there was an increase in ischemic stroke of borderline significance. I think they saw neutral outcome data, and because of a potential ethical safety issue they decided to stop the trial. But there was no signal of an increase in fatal stroke or of all-cause or cardiac mortality. And now that we have all the data, the effect is far from significant, and when we look just at strokes that occurred when the patients were actually on treatment, we have 21 on niacin vs 18 on placebo. The earlier numbers were clearly an aberration--just the play of chance."

Boden said he was "disappointed that we couldn't complete the trial as designed," especially in view of the previous VA-HIT trial, which was conducted before statin use became so widespread and showed a reduced event rate with gemfibrozil (which also raises HDL), but this didn't start to appear until 18 months into the study. "We had planned a 4.6-year follow-up, but we only got three years, so we are unable to know for sure the effect of niacin in this patient group."

Blumenthal More Cautious on Stroke Finding

Commenting for heartwire , Dr Roger Blumenthal (Johns Hopkins University, Baltimore, MD) agreed that, in retrospect, AIM-HIGH "probably shouldn't have been stopped, as it may not have been long enough to show a benefit." He noted that in the CARE trial of pravastatin vs placebo, the event curves also didn't start to diverge until two years.

On the stroke data, Blumenthal said "a p value of 0.11 is still a strong trend, and I think it is still potentially a worry. I am not completely convinced it was the play of chance." He pointed out that in one of the early trials of niacin--the Coronary Drug Project--there was an excess of atrial fibrillation, which is one of the main risk factors for stroke. But atrial fibrillation is not reported in the AIM-HIGH paper.

Result Questions the Value of CIMT Studies

Assuming that the neutral result was a true finding, Blumenthal said this again calls in to question the value of surrogate-end-point trials, particularly those focusing on carotid intima-media thickness (CIMT). "One of the reasons this trial was conducted was because of the ARBITER-6 study, which showed a small improvement in CIMT with niacin vs ezetimibe in just 200 patients already taking simvastatin. The authors of that trial said the results merited strong consideration of niacin for patients with low HDL. But here we are again with a larger trial that hasn't confirmed the results of a small surrogate-end-point study."

More data on the use of niacin will come from other studies now ongoing, in particular HPS2-THRIVE, which with 24 000 patients is much bigger and should give a definitive answer. Unlike AIM-HIGH, this trial is enrolling "all comers"--patients do not have to have low HDL and will not have such good LDL levels, so it should be more relevant to a real-world population.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....