SAN FRANCISCO – Full results from the HPS2-THRIVE study of Merck & Co. Inc.’s Tredaptive have heightened scrutiny about outcomes studies for cardiovascular drugs and may lead to changes in the way these studies are designed.

The more than 25,000-patient trial failed to show a significant improvement in cardiovascular outcomes with treatment with Merck’s combination of extended-release niacin and laropiprant, an anti-flushing compound intended to mitigate a common side effect of niacin. Announcing top-line results in December, Merck had said that it was dropping U.S. development of Tredaptive and soon after said that it was discontinuing marketing in other countries where it is approved.

But HPS2-THRIVE is just the latest in a string of failed outcomes studies for therapies that aim to improve CV health by raising HDL-cholesterol along with lowering LDL-cholesterol and triglycerides, including AIM-HIGH for Abbott Laboratories Inc.’s Niaspan (extended-release niacin), the ACCORD study for Abbott’s Trilipix (fenofibric acid) and the dal-OUTCOMES trial for Roche’s CETP inhibitor dalcetrapib Also see "Dalcetrapib Failure Raises Yet More Questions On Value Of Tinkering With HDL Cholesterol" - Pink Sheet, 14 May, 2012.).

HPS2-THRIVE shows that the days of “big dumb” trials need to finally end in cardiology, argues Baylor’s Christie Ballantyne.

In the past, a variety of therapies have been added to statins for patients with low HDL-C, persistently elevated LDL-C despite statin therapy and/or high triglycerides – including niacin, fibrates and GlaxoSmithKline PLC’s omega 3 fish oil Lovaza. Use has been directed at improving surrogate markers, but generally has not been supported by outcomes studies.

Now, a range of new classes are being positioned as add-on therapies in the wake of the patent expiry of statins, including PCSK9 inhibitors, CETP inhibitors and new omega 3 fish oil products Also see "Branded Life After Lipitor: Pharma Targets Unmet Needs In Dyslipidemia" - In Vivo, 23 April, 2012..

Expectations for these drugs could be affected by the full results from HPS2-THRIVE reported at the American College of Cardiology meeting in San Francisco on March 9. In the trial, average baseline lipid levels were 63 mg/dL LDL, 44 mg/dL d/L, with triglycerides at 125 mg/dL. On average, the combination raised HDL by 6 mg/dL (13.6%), lowered LDL by about 10 mg/dL (15.8%) and lowered triglycerides by 33 mg/dL (26%).

Investigators at Oxford University had expected a 10% to 15% reduction in the event rate. However, they wound up reporting a numerical but not statistically significant 4% reduction in events for the combination and an unacceptably high rate of serious adverse events, most of which required hospitalization, compared to the comparator placebo arm (see related story, Also see "Tredaptive Safety Hazards Cast Shadow Over AbbVie’s Niaspan" - Pink Sheet, 18 March, 2013.).

Implications For Other Classes?

At an ACC press conference, HPS2-THRIVE investigators faced questions about the use of drugs without evidence from outcomes studies and what the results mean for new drug classes in development.

Lead investigator Jane Armitage acknowledged that in clinical practice, “we are using a lot of drugs out there that haven’t been scrutinized in the way that regulatory agencies would expect if a new drug was coming on the market.” There may well be “other drugs out there that should be evaluated more carefully,” she said.

In particular, more work could identify subpopulations that could see benefit. She pointed out that while fenofibrates didn’t quite meet the primary endpoint in ACCORD, there were benefits in some subgroups and no clear safety signal, so fibrates remain useful for high-risk patients.

As for CETP inhibitors, which raise HDL like niacin, sponsors have already committed to running outcomes studies prior to approval – though none have yet met with success. Dalcetrapib development was terminated due to low efficacy, as was the pioneer in the class, Pfizer Inc.’s torcetrapib. But there is still hope for Merck’s anacetrapib and Eli Lilly & Co.’s evacetrapib, which have more dramatic effects on HDL and also lower LDL significantly Also see "Merck, Lilly Vow To Press On With CETP, Despite Dalcetrapib Disaster" - Pink Sheet, 7 May, 2012..

HDL remains an important predictor for risk, said Armitage, who is on the steering committee of the REVEAL study of anacetrapib, which is being run by Oxford University.

“I certainly think it’s premature to say raising HDL is not going to work,” Oxford University’s Armitage said.

As HDL was only increased a small amount in HPS2-THRIVE, “I don’t believe it really tells us anything very useful whether or not other mechanisms might still be beneficial,” Armitage said. Niacin’s mechanism of action isn’t well understood, “so I think it may well be that other mechanisms of raising HDL may reverse some of the epidemiological risk. But we just don’t know that at the moment. I certainly think it’s premature to say raising HDL is not going to work.”

After the meeting, she clarified that the population in the REVEAL study is similar to the one studied in HPS2-THRIVE, but that since the drugs are very different, investigators can’t draw any conclusions about what might happen in REVEAL.

Panelist Calls For End To “Big, Dumb Studies”

Investors have been jittery about another outcomes study from Merck – the IMPROVE-IT study for Vytorin, the combination of the firm’s Zetia (ezetimibe) and Zocor (simvastatin). But like other analysts who are not worried about a negative outcome, ISI Group analyst Mark Schoenebaum noted that Vytorin is very different in that it is aimed at lowering LDL, whereas niacin is used primarily to raise HDL. He pointed out that the baseline LDL for the first 10,000 patients enrolled in IMPROVE-IT (out of a total of 18,000), who were already on lipid-lowering therapy, was much higher than the HPS2-THRIVE baseline: 97 mg/dL compared to 63 mg/dL ([A#14130312002]). This additional room to fall in IMPROVE-IT could theoretically translate into a higher rate of reduction in events.

In HPS2-THRIVE, enrollees with higher LDL had a trend toward a benefit, but this was not statistically significant and was still outweighed by risks.

During the presentation of results at the ACC meeting, Baylor College of Medicine’s Christie Ballantyne, who has been a member of steering committees on a number of major CV trials, including IMPROVE-IT and dal-OUTCOMES, suggested trials need to be enriched with patients most likely to benefit from treatment.

In an interview after the meeting, he explained that based on the baseline lipid profile in HPS2-THRIVE, many patients in the study should not have been getting niacin as an add-on therapy. In practice, doctors use niacin as a third-line therapy for reducing LDL, or in patients with low HDL. Given the baseline lipid levels (low LDL and HDL over 40 mg/dL), adding a drug with known side effects could not possibly have resulted in a favorable risk-benefit ratio, he said.

Ballantyne chalked up the failure to the tendency in cardiology to do large studies of all comers, that is all patients at high risk, as opposed to matching the mechanism of action with a smaller patient population most likely to benefit. All-comer trials are easier to enroll and give drug companies a shot at the broadest population possible, but they carry the risk of perhaps not capturing a drug’s actual efficacy.

He suggested companies should start smaller with patient populations enriched with those likely to benefit, and then expand to larger populations. Whereas cancer trials have moved in this more personalized direction, cardiology has been slower to adapt.

“Big, dumb studies” are a thing of the past, Ballantyne argued. “It’s kind of a brute force approach [and] you assume you have the power to test everything,” especially if you don’t know how the drug works, he said. “Maybe that works if you have no side effects,” but it is antiquated. “People talk about precision medicine or personalized therapy – this is the antithesis of that.”

However, in the case of HPS2-THRIVE, the mechanism of action was not fully understood.

In the era of statins, having a trial with all comers means you have some patients with relatively low LDL-C. New drugs should be tested in those patients most likely to use the medication, that is those who can’t get to treatment goals on existing therapies rather than all comers, Bill Sasiela, director of Regeneron Pharmaceuticals Inc.’s PCSK9 development program, agrees. This kind of trial is harder to enroll but is more likely to demonstrate a treatment effect.

However, he added, “we may be able to demonstrate in the future that even lower LDL-C levels will provide additional CV benefits to patients, but you need to design and power the trial for this appropriately.”

Data from major LDL-lowering trials shows that the most solid predictor of CV event reduction is the absolute reduction in LDL-C and not the percent reduction, Sasiela explained. A 1 mmol/L or 38.6 mg/dl drop in LDL-C results in a 20% to 22% reduction in major CV events. Most drugs that lower LDL-C lower do so at the same percentage, regardless of the baseline level.

Major CV event trials for LDL-lowering drugs must be designed with attention to expected baseline LDL-C levels as the basis of the trial assumptions and powering. Since the drug’s efficacy in terms of percent lowering is fixed, starting with a patient population with a relatively low baseline LDL-C means it will be tougher to detect a difference in event rates. It particularly gets tough if you have a drug that only provides a 15% to 20% reduction in LDL-C, as shown in the HPS2-THRIVE study, Sasiela said. The dose of niacin provided a reduction in LDL-C but with a starting LDL-C of 63 mg, that only resulted in an absolute difference of 10 mg/dL, which translated into a small reduction in events.

HPS2-THRIVE Post-Mortem: Lessons Learned For CV Drug Developers

More from Clinical Trials

More from R&D