In the ANCHOR trial publication, Ballantyne et al, evaluating AMR101 (icosapent ethyl) for the treatment of mixed dyslipidemia, concluded that omega-3 fatty acids containing only eicosapentaenoic acid (EPA) lower low-density lipoprotein (LDL) cholesterol. The investigators’ interpretation of the data regarding the effects of pure EPA on LDL cholesterol is misleading for 2 important reasons. First, the placebo used in the AMR101 trials was light liquid paraffin, more commonly known as mineral oil, a biologically active substance. Second, the use of “placebo-adjusted results” skews the interpretation of the data so significantly that the conclusions are invalid.

Mineral oil has been known for decades to block the absorption of medications and therefore has been seldom used as a placebo for double-blind clinical trials. On the basis of the adverse drug events reported in the ANCHOR trial, the placebo arm (mineral oil) produced a much greater side-effect profile than the drug (all gastrointestinal disorders 17.2% with mineral oil vs 11.6% with AMR101). More important than the adverse drug events (which are simply further evidence of the active nature of this placebo), mineral oil adversely affected the lipid profiles of the patients in the trial. In the mineral oil control group, LDL cholesterol, non-high-density lipoprotein chole-sterol, and high-sensitivity C-reactive protein increased by 9%, 10%, and 17%, respectively, over 12 weeks of treatment in a relatively large sample size of patients receiving statin therapy (n = 233). The adverse effects of mineral oil on lipid levels were also demonstrated in AMR101’s other pivotal trial for patients with severe hypertriglyceridemia, the Multi-Center, Placebo-Controlled, Randomized, Double-Blind, 12-Week Study With an Open-Label Extension (MARINE), as well as another clinical trial with omega-3 fatty acids. A potential explanation for mineral oil’s deleterious effects on the lipid parameters in these studies was the inhibition of the absorption of the statins used concomitantly during the trials.

The adverse effects of mineral oil on lipid levels also call into question the conclusions based solely on “placebo-adjusted results.” The primary end point for a clinical trial is the difference in the efficacy of therapy from baseline to the end of treatment compared to “placebo.” Discussing only placebo-adjusted results fails to take into account a potential adverse effect of an “active control.” Consider LDL cholesterol (although this approach holds true for all other end points as well). LDL cholesterol increased by 1.5% in the AMR101 4 g group yet “decreased” by 6.2% compared to placebo, which increased LDL cholesterol by 8.8%. Thus, placebo adjustment rendered a beneficial reduction of LDL cholesterol in this AMR101 group (p = 0.0067), although in absolute terms, AMR101 elevated LDL cholesterol. The investigators stated that LDL cholesterol was improved by pure EPA compared to prescription omega-3 fatty acids containing both EPA and docosahexaenoic acid (Lovaza; GlaxoSmithKline, London, United Kingdom). This conclusion is invalid because AMR101 and Lovaza increased LDL cholesterol to a similar level in patients with mixed dyslipidemia taking statins. The difference between the effects in these trials is likely the “placebo” that was used (mineral oil vs corn oil). A comparison of the positive benefits of EPA and docosahexaenoic acid for cardiovascular benefits is outside the scope of this comment, but there is ample evidence that the 2 fatty acids are essential for optimal health. The conclusion stated by the investigators of the ANCHOR publication may potentially mislead clinicians to believe that EPA-only therapy provides superior lipid efficacy compared to omega-3 therapy that contains EPA and docosahexaenoic acid.

In sum, using “placebo-adjusted results” makes the ANCHOR trial’s findings unjustified. The ongoing Reduction of Cardiovascular Events With EPA–Intervention Trial (REDUCE-IT) is evaluating the effects of AMR101 compared to a “placebo” for cardiovascular outcomes in high-risk patients with mixed dyslipidemia taking statins for 5 years. If mineral oil is the “placebo” in this trial, as it was in the ANCHOR trial, there may be a difference in outcomes in the trial that could be due to harmful effects of the mineral oil rather than a beneficial effect of AMR101. This issue may compromise the validity of the trial and, more important, may put the enrolled patients taking statins at risk for increased cardiovascular events.