REDUCE-IT vs STRENGTH: Still Missing Pieces in the Omega-3 Puzzle

In a paper being presented during ESC Congress 2021, authors from the University of Copenhagen use a registry to mimic 2 well-known trials, finding explanations for some, but not all, of the gap in risk reduction between them.

A paper presented Sunday during ESC Congress 2021 addresses an ongoing question in cardiology research: How have 2 recent high-profile clinical trials involving omega-3 fatty acids produced such different results?

The effort by a team based at the University Copenhagen raises questions of its own. The authors’ analysis accounts for 12% of the difference in risk reduction for atherosclerotic cardiovascular disease (ASCVD) between the 2 trials, but cannot explain the remaining 13%. So far, 7% of the difference is attributed to the placebo in the REDUCE-IT study.

One study author, Børge G. Nordestgaard, MD, MSc, of the University of Copenhagen, told The American Journal of Managed Care® in an interview that the missing 13% difference remained a “mystery.”

The analysis, published simultaneously in the European Heart Journal, used patient data from the Copenhagen General Population Study (CGPS) to mimic designs for the trials:

  • REDUCE-IT, which in November 2018 reported a 25% risk reduction (HR, 0.75) for ASCVD for patients taking icosapent ethyl, a purified formulation of EPA. REDUCE-IT used a mineral oil placebo.
  • STRENGTH, which in November 2020 reported no risk reduction (HR, 0.99) for ASCVD with carboxylic acid, a combination of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). STRENGTH used a corn oil placebo.

During the presentation of STRENGTH, investigators from Cleveland Clinic raised the issue of the mineral oil placebo in REDUCE-IT, saying there were signs it had influenced the results of the earlier trial. Questions about the mineral oil placebo in REDUCE-IT had been raised previously, but never by other scientists presenting their own results.

At the May meeting of the American College of Cardiology, Cleveland Clinic’s Steven Nissen, MD, called for a head-to-head trial. “Additional research is needed, with trials designed to compare corn oil with mineral oil, and to compare purified EPA with other formulations of omega-3 fatty acids,” Nissen said.

Meanwhile, REDUCE-IT’s lead investigator, Deepak L. Bhatt, MD, MPH, of Harvard University, has noted that the reduction in ischemic events has been borne out in 2 trials and is attributable to the EPA-only formulation, not the mineral oil placebo. Other commentators have said that a risk reduction of 25% is too large to be attributable to the placebo.

Since presenting the first results, Bhatt has offered regular subgroup analyses at major conferences, including one during ESC Congress 2021 on patients who had suffered a prior heart attack. Among these patients, icosapent ethyl was associated with a 26% reduction in first major adverse cardiac events (MACE) and a 35% drop in in total MACE.

Bhatt and others have noted that mineral oil was approved by the FDA, but Nordestgaard said that occurred before the study results were known.

“I think the mineral [oil] certainly does explain part of the 25% reduced risk,” he said during the AJMC® interview. “All this thing about the FDA approved it—well, they approved it before they had seen the results of REDUCE-IT. This is the data that we have now.”

The authors write that their findings have important clinical implications. Icosapent ethyl, approved by FDA in 2013 to treat high triglycerides, received a 2019 indication as an add-on therapy for ASCVD. Amarin, which sells the product as Vascepa, reported a second quarter income of $154.5 million, a 14% increase over the prior year period. The company has plans to enter several markets in Europe, including Germany, starting in September.

Analysis. The research team selected participants from the CGPS to mimic the trial designs; they selected 5684 who met inclusion criteria for REDUCE-IT, including 852 with ASCVD, and 6862 for STRENGTH, including 697 with ASCVD. Patients were followed for the median durations of REDUCE-IT (4.9 years) and STRENGTH (3.5 years).

Researchers then performed statistical analyses, in which they evaluated the changes that took place in the EPA arm of REDUCE-IT for triglycerides, low-density lipoprotein (LDL) cholesterol, and C-reactive protein (CRP) during the study, and then observed the hazard ratios for ASCVD for where there was a comparable drop in triglycerides, LDL cholesterol, and CRP, both individually and combined. A parallel analysis was done for the CGPS group set up to mimic the EPA-DHA arm of the STRENGTH study. The hazard ratios (HR, all 95% CI) were as follows:

For the CGPS mimicking REDUCE-IT, the HRs for ASCVD were:

  • For –20% change in triglycerides: 0.97
  • For –1% change in LDL cholesterol: 1.00
  • For –14% change in C-reactive protein (CRP): 0.97
  • Combined: 0.96

For the CGPS mimicking STRENGTH, the HRs for ASCVD were:

  • For –19% change in triglycerides: 0.97
  • For +1% change in LDL cholesterol: 1.00
  • For –20% change in CRP: 0.96
  • Combined: 0.94

Similar calculations were done for the placebos. The HR for the mineral oil was 1.07, while the HR for the corn oil was 0.99.

Combining these results, the active ingredient vs placebo ratio was an HR of 0.88 in the CGPS mimicking REDUCE-IT compared with 0.75 in the actual study. For STRENGTH, the CPGS mimicking STRENGTH was an HR of 0.96 compared with 0.99 in the actual study.

Explanations. The authors offered an expansive discussion on their analysis as well as reasons for the chasm between REDUCE-IT and STRENGTH. “We estimate that changes in risk of ASCVD can be explained by the observed effects on lipid traits and C-reactive protein by active oils and comparator oils. Importantly, our approach does not exclude other possible effects of active and comparator oils on ASCVD unrelated to lipid traits and C-reactive protein.”

They note the analysis could not account for 13% of the risk reduction, which they said could be due to “other beneficial effects of EPA, or deleterious effects of mineral oil.” The wrote that EPA may have positive effects on ASCVD risk factors, such as “blood pressure, platelet activation, oxidative stress, inflammation, endothelial function, plaque phenotype, and lipid levels and metabolism not accounted for,” which could prevent ASCVD or the rupture of plaques that trigger cardiac events.

Indeed, the authors discuss results from EVAPORATE, presented in phases in November 2019 and August 2020, that show icosapent ethyl appears to cause a retreat of coronary plaques among patients with coronary artery disease, offering a possible explanation for the reduction in cardiac events.

But the Copenhagen authors conclude with arguments that the mineral placebo is suspect, possibly interfering with statin absorption and elevating risk in the placebo arm. “In the REDUCE-IT mineral oil arm, LDL cholesterol, non–HDL cholesterol, apolipoprotein B, and C-reactive protein were all elevated at the end of the study period. In contrast, such changes were minimal in the STRENGTH corn oil arm, even with a trend towards potential protection against ASCVD due to lower C-reactive protein,” the authors write.

“Surprisingly, changes in the 3 causal risk factors led to similar estimated reduced risk of ASCVD in the active oil groups of EPA in REDUCE-IT and of EPA/DHA in STRENGTH. Taken together, increased lipid traits and C-reactive protein may account for increased risk of ASCVD in the mineral oil arm of REDUCE-IT, explaining part of the contrasting results of REDUCE-IT vs STRENGTH.”

Maggie L. Shaw contributed to this report.

Reference

Doi T, Langsted A, Nordestgaard BG. A possible explanation for the contrasting results of REDUCE-IT vs STRENGTH: cohort study mimicking trial designs. Eur Heart J. Published online August 29, 2021. doi:10.1093/eurheartj/ehab555