A new analysis of the REDUCE-IT study has reignited concerns that the benefit shown by the high-dose fish oil product in the study, icosapent ethyl (Vascepa, Amarin), may have been related to harms caused by the placebo mineral oil.
Results show that allocation to icosapent ethyl had minimal effects on a series of biomarkers associated with atherosclerotic disease, whereas levels of these biomarkers increased among those allocated to mineral oil.
At 12 months, the median percent increases from baseline in the mineral oil group were 1.5% for homocysteine, 2.2% for lipoprotein(a), 10.9% for oxidized low-density-lipoprotein (LDL) cholesterol, 16.2% for interleukin (IL)-6, 18.5% for lipoprotein-associated phospholipase A2, 21.9% for high-sensitivity C-reactive protein (hsCRP), and 28.9% for IL-1β. The changes were similar at 24 months. However, in the icosapent ethyl group, there were minimal changes in these biomarkers at 12 and 24 months.
The study was published online in Circulation on June 28.
The authors, led by Paul Ridker, MD, Brigham & Women's Hospital, Boston, do not voice much opinion on what the results mean, concluding that "the effect of these findings on the interpretation of the REDUCE-IT trial results remains unclear and will require further investigation."
They also say that a second icosapent ethyl trial using a nonmineral oil comparator "would help resolve ongoing controversy."
However, the authors are a mixed group; Ridker and some of his coauthors were not part of the original REDUCE-IT trial, whereas other coauthors were members of the REDUCE-IT steering committee and one was an employee of Amarin.
Lead investigator of the REDUCE-IT trial, Deepak Bhatt, MD, also from Brigham & Women's Hospital, who is the senior author of the current study, played down the new findings, saying they did not offer much new incremental information on mechanistic insight.
"These are small degrees of biomarker increases on an absolute level we are seeing with the placebo that would not be expected to produce harm," Bhatt told theheart.org | Medscape Cardiology. He also said the Circulation peer-review process had removed some of the discussion, which could lead to some "misinterpretation" of the authors' views.
Ridker was unavailable for further comment.
The publication of this study has inevitably poured fuel on the fire regarding the controversy that has long dogged the REDUCE-IT trial, with questions about the large reduction in event rates seen with icosapent ethyl without an obvious mechanistic explanation.
One of the loudest critics of the study, Steve Nissen, MD, Cleveland Clinic, described the new findings as "the closest thing I've seen to a smoking gun in medicine for a long, long time."
"The result of this new analysis shows that mineral oil increases virtually every inflammatory and lipid marker that they measured," he commented to theheart.org | Medscape Cardiology. "There are a lot of theories, but the bottom line is that something really bad happened in the mineral-oil group which makes icosapent ethyl look efficacious. In my view, this needs to be reviewed by the FDA for consideration of removing the label claim for cardiovascular benefit."
Other experts in the field not directly involved in the study voiced concern about these new findings, adding to calls for another trial.
In a Twitter thread on the issue, Harlan Krumholz, MD, describes the Circulation publication as "an exceptionally important article," adding that it is "time to rethink this drug."
"My point is...once you know you have non-neutral comparator and the effect on risk biomarkers is far from trivial…then you have introduced substantial uncertainty about the trial result, as conveyed by the authors…and no one can say what would happen with a neutral comparator," Krumholz writes.
In an accompanying editorial in Circulation, Robert Harrington, MD, professor of medicine at Stanford University, California, concludes that "the hard reality is that we are left with uncertainties and the questions raised by use of the mineral oil as placebo can only be answered by another randomized controlled trial."
"My main points are that the chosen placebo was not inert (an essential characteristic for an active control–placebo comparison), that the biomarker data show elevations in multiple markers involved with inflammatory pathways associated with atherosclerosis, and that these data create enough uncertainty in the trial interpretation that the best recourse to answer the criticisms is to do another trial with a truly inert placebo," Harrington told theheart.org | Medscape Cardiology.
He added that Bhatt's point that the changes in biomarkers are too small to really matter clinically may be right. "But for me, the uncertainties mean that I have less confidence in the true magnitude of the treatment effect than I would have if there were no changes in the inflammatory markers."
In Circulation, the authors say it is unclear why multiple biomarkers increased over time among REDUCE-IT participants allocated to mineral oil. They note that no substantive changes in these biomarkers were observed in the placebo groups over periods of 3 to 5 years in other trials, including JUPITER, CIRT, CANTOS, SPIRE, and the STRENGTH trials, which evaluated a different high-dose omega-3 oil product but used corn oil as the placebo.
"The core design of REDUCE-IT does not make it possible to resolve convincingly whether any adverse effects associated with mineral oil use as a comparator may have affected clinical outcomes," they write.
They point out that regulatory agencies evaluating REDUCE-IT estimated that approximately 3% of the net clinical benefit observed with icosapent ethyl might have been a consequence of adverse biomarker effects on LDL cholesterol and hsCRP attributable to mineral oil. But in the context of an overall 25% relative risk reduction in first events and a 30% reduction in total ischemic events observed, a potential bias of this magnitude, even if doubled in size, would be unlikely to fully attenuate the overall benefit of icosapent ethyl observed.
They add that they are not aware of a method to assess what the potential magnitude might be of a combination of the multiple effects
Bhatt: New Data Do Not Change the Debate
"We did a large, well-powered randomized trial, and this paper shouldn't change anything in how that trial should be interpreted," Bhatt told theheart.org | Medscape Cardiology.
He claims the new biomarkers evaluated in the study are correlated with LDL and CRP, data which have already been reported and analyzed, so have limited relevance.
"It's not really independent biomarker information; this is what we would expect to see when we see small increases in LDL and CRP. So, I don't think this new information fundamentally changes the debate," he said.
Bhatt also pointed out that the study highlights relative increases rather than absolute increases in the biomarkers, making it seem more alarming than is actually the case.
"The paper makes it seem like that there are large increases in these other biomarkers, but the values reported are relative increases and the absolute increases were actually rather small. In many cases, the changes reported are less than the lower limit of quantification of the assay used," he noted.
He added: "Even if one is unable to get around the placebo issue in the REDUCE-IT trial, there will always be the JELIS trial — a randomized trial with no placebo showing a 19% relative risk reduction. While the biomarker data may be interesting, what really matters in the end is clinical events. And significant reductions in two independent trials should be enough."
Bhatt says the REDUCE-IT steering committee does not believe another trial is needed. "Maybe a different population would be good — such as primary prevention, patients without elevated triglycerides — but just repeating REDUCE- IT with a different placebo would be a waste of resources," he commented.
But Nissen refuted Bhatt's claims.
"These biomarkers are not in the same pathways as LDL and CRP, and these are not small increases. In the CANTOS trial, a monoclonal antibody against interleukin-1β beta showed a significant benefit. The increase in interleukin-1β now reported in REDUCE-IT is exactly the opposite of CANTOS," he pointed out.
"The FDA did not know about these additional biomarkers when it reviewed the data on LDL and CRP. Now we have new information. It needs to be looked at again," Nissen added.
Funding for the study was provided by Amarin Pharma. Bhatt was the lead investigator of the REDUCE-IT trial. Nissen was the lead investigator the STRENGTH trial. Further disclosures of the authors can be found in Circulation.
Lead image: Medscape Illustration/Dreamstime
Medscape Medical News © 2022
Cite this: New Biomarker Data Add to Concerns Over REDUCE-IT Trial - Medscape - Jul 01, 2022.