March 9th, 2013

HPS2-THRIVE: A ‘Disappointing But Clear’ Result

For more of our ACC.13 coverage of late-breaking clinical trials, interviews with the authors of the most important research, and blogs from our fellows on the most interesting presentations at the meeting, check out our Coverage Headquarters.

The results of HPS2-THRIVE were “disappointing but clear,”  said Jane Armitage, who presented the results this morning at the ACC meeting in San Francisco.

HPS2-THRIVE randomized 25,673 high-risk patients who could tolerate niacin to either placebo or extended-release niacin plus laropiprant (Tredaptive, Merck), an anti-flushing agent, in addition to background therapy. The primary endpoint was the time to first major vascular event, defined as the composite of non-fatal MI or coronary death, any stroke, or any arterial revascularization.

Major vascular events occurred in 13.2% of the niacin arm and 13.7% of the placebo arm (p=0.29), despite causing average reductions in LDL of 10 mg.dL and triglycerides of 33 mg/dL, in addition to a 6 mg/dL increase in HDL. Armitrage reported that based on data from previous trials and observational studies, “it was anticipated such lipid differences might translate into a 10-15% reduction in vascular events.”

Significantly more patients in the niacin arm stopped their study treatment during the course of the trial (25.4% versus 16.6%). Serious adverse events also occurred more often in the niacin group, including new onset diabetes (9.1% vs. 7.3%), diabetic complications (11.1% vs. 7.5%), gastrointestinal problems  (4.8% vs. 3.8%), and skin problems such as itching and rashes (0.7% vs. 0.4%).

The investigators were also surprised to find higher rates of bleeding (2.5% vs. 1.9%) and infections (8.0% vs. 6.6%) in the niacin group. They offered no explanation for these findings.

“We are disappointed that these results did not show benefits for our patients,” said Jane Armitage, lead author of the HPS2-THRIVE study, in a press release. “Still, finding out a drug is not helping people is just as important as finding that it has benefits — the net result is that people are healthier. Niacin has been used for many years in the belief that it would help patients and prevent heart attacks and stroke, but we now know that its adverse side effects outweigh the benefits when used with current treatments.”

Armitage said that the trial results are unlikely to be due to the side effects of laropriprant, since the main finding of a lack of benefit is consistent with the AIM-HIGH study and many of the side effects are known to be associated with the use of niacin.

During the panel discussion Donna Arnett noted that the trial was performed in a patient population that had already achieved low LDL levels with statins. The results of HPS2-THRIVE establish that “if you have a low LDL on a statin there is no benefit” with niacin. “So it really is calling into question whether raising HDL in the context of low LDL is important,” she said.

Another panelist, Christie Ballantyne, agreed with Arnett, and went further, arguing that the study did not shed light on the patient population that is most likely to take niacin. He said the majority of niacin patients in the U.S. have LDL levels over 70. He said it was still possible that niacin could be found beneficial in this group.

8 Responses to “HPS2-THRIVE: A ‘Disappointing But Clear’ Result”

  1. Pablo Corral, MD says:

    Who is surprised with the results of HPS2 THRIVE?
    Who prescribe niacin in this group of patients with LDL-C 63 mg / dl? We should understand and learn the mechanism of action of niacin, and check that their most important effects are on LDL-c (evidenced by the decrease in LDL particles [LDL-C]) and not on the particles of HDL (HDL -p), concluding that it is more important the effect of niacin on LDL-p and not-p HDL.
    What should we do with the studies: CDP, HATS, FATS, CLAS, and Oxford Niaspan ARBITER??? Forget, pretend it did not exist???
    It is very difficult to draw clear conclusions from these studies (including the AIM HIGH), studies with clear design errors that lead to the danger of misinterpreting current indications of this group of drugs.
    Finally no one disputes that this Trial, under no circumstances tests the theory of HDL.

  2. Jean-Pierre Usdin, MD says:

    At the moment in France we have an important controversial about Statin therapy in spite of evidence based medicine results.
    Arguing without real scientific data that Statin are useless or harmful, an university retired ‘famous’ professor makes many patients uncomfortable in continuing their essential, well tolerated anti-cholesterol treatment.
    I bet this distinguished retired colleague, will not argue about Niacin!
    The results of all the trials including this ‘natural medicine’ are speaking by themselves!
    Is it not time to stop these studies concerning Niacin alone or combined with a supposed anti-sideeffect product?

  3. Carlos Cuneo, MD says:

    I’m not surprised with HPS2 results, because I knew the results of AIM-HIGH.
    But, I also know that many, many physicians are still using niacin in a high risk coronary patient, statin treated, with LDL 63 mg/dl and HDL < 35 or 30 mg/dl). Should they be surprised ?

  4. Indu Poornima, MD says:

    In the setting of negative results in HPS2 and AIM-HIGH, I am interested in experts’ approach to high Lp(a) levels in a patient with vascular disease who does not have LDL abnormalities

  5. Indu, I am not a Lp(a) expert, but I think the most elegant solution to your dilemma about Lp(a) levels might be to stop measuring your patients’ Lp(a) levels.

    Lp(a) is a risk factor for cardiovascular disease, but I am not aware of evidence that treating patients to lower their Lp(a) levels makes a difference to outcomes that matter. For the patients you describe (with established cardiovascular disease) risk is high and so I’d like to offer a statin to those patients who can tolerate it. The negative HPS2 and AIM-HIGH lead me to suspect there would be no benefit from adding niacin for these patients on a statin.

    In summary, why would I test Lp(a), if I already know my patient is at high risk, and the test result is not going to change my management?

  6. Pablo Corral, MD says:

    Regarding the commentary of Indu, if a patient has vascular disease with normal LDL levels, is a clear indication to measure Lp(a), or seek nontraditional factors. If the levels of Lp(a) are high I would start with a statin, and then most likely add Niacin and aspirin. The Lp(a) is a recognized risk factor and we must not forget that it is a modified LDL with prothrombotics properties.

  7. Absolutely clear cut results. Laropiprant should not be used to treat heart disease, it as all other anti-prostaglandin medications seems increase heart disease.

    Aim High did not show a benefit due to the fact that niacin was the placebo and the study was inappropriately terminated early.

    I refer you to the Meta-analysis in December’s JACC which demonstrated a dramatic benefit from the use of niacin in coronary prevention.

  8. The meta-analysis you refer to (presumably the one by Lavigne and Karas, J Am Coll Cardiol 2013;61(4):440-6 has, in my view, many problems.

    First, it included many papers where niacin was only one of several components of the intervention being studied. It’s not surprising that niacin, in combination with proven agents like statins or bile acid sequestrants, seemed to have a cardiovascular effect. However, trials of such design do not tell us whether niacin has any additional benefit beyond what we would expect from statins or bile acid sequestrants alone. The inclusion of such papers in this meta-analysis is likely to bias the apparent effect of niacin in the direction of benefit.

    Second, the forest plots in the Lavigne and Karas meta-analysis demonstrate much smaller effect sizes in the larger trials, and much larger and generally much more beneficial effects in the smaller trials. In my view, such a pattern raises serious concerns about publication bias.

    Third, the authors did not report on total mortality.

    Studies of niacin as a lone intervention, unsullied by the bias of other intervention components, are few in number. The big one was the Coronary Drug Project, which showed a 27% reduction in non-fatal MI but no significant effect on all-cause mortality when the trial ended. However, a 15-yr follow-up of the Coronary Drug Project did find a mortality benefit from niacin — this was an 11% relative risk reduction after 15 yrs in secondary prevention ( To put this into context, based on the 4S trial of simvastatin, we’d expect a 30% relative reduction in all-cause mortality after 5.4 yrs with simvastatin ( Statins, then, seem much more impressive. For patients who tolerate a statin, and who are at significant cardiovascular risk, I imagine many of us would agree that statins are first choice.

    So, in our current era of widespread statin prescribing, the important question was: in a patient already on a statin, does adding niacin add any further protection? The AIM-HIGH study found that it didn’t ( I don’t know why you say that “niacin was the placebo” in this study; in fact the study was of niacin versus placebo in statin-treated patients. The study findings could be negatively biased by the non-niacin arm patients getting a bit more statin and ezetimibe. However, AIM-HIGH lends no support, in my view, to the strategy of adding niacin to statins. From what we know so far, the same seems true of HPS2-THRIVE.