We welcome the opportunity to respond to Dr. Baum’s comments regarding the use of light liquid paraffin (LLP) as the study placebo and placebo-adjusted values for our recently published results of icosapent ethyl in the ANCHOR trial. The disclosures of the investigators of the ANCHOR trial accompanied the original report. Although not cited in his letter, we note that according to his on-line biography, Dr. Baum is the founder and director of clinical development for a commercial company that markets a variety of dietary supplements, including those containing docosahexaenoic and eicosapentaenoic acid.

Regarding Dr. Baum’s concern, LLP is a very highly refined mineral white oil and is a mixture of liquid hydrocarbons obtained from petroleum, which do not contain any functional groups (e.g., no carboxyl groups in contrast with fatty acids in vegetable oils). Hence, LLP is chemically inert, and absorption is minimal. The United States Food and Drug Administration (FDA) provides guidance on the use of various types of placebo controls in clinical studies. Such guidance was integral to the selection of LLP as the placebo for the Multi-Center, Placebo-Controlled, Randomized, Double-Blind, 12-Week Study With an Open-Label Extension (MARINE) and ANCHOR studies before commencement of the phase 3 clinical program. Paraffin oil is a commonly used placebo in omega-3 fatty acid clinical studies, with a documented lack of effect on lipids such as triglycerides, low-density lipoprotein (LDL) cholesterol, and fatty acids. In the omega-3 fatty acid study that included patients receiving lipid-lowering agents (which Dr. Baum cited as supporting evidence for adverse effects of paraffin oil on lipid levels), there were no significant changes from baseline in LDL cholesterol, any other lipids, markers of diabetes, or markers of inflammation in the 3 g/day paraffin oil placebo group after treatment for 2 months. Furthermore, in a study of healthy subjects comparing various oil-based placebo controls, paraffin oil at 4 g/day for 10 days had no effect on fatty acids in any lipid fraction.

Mineral oil has been used for nearly a century as a laxative at much higher doses (15 to 45 ml/day) than the LLP doses used in the MARINE and ANCHOR studies (about 2 ml twice daily with food). Paraffin oil is listed as “generally recognized as safe” in the United States and is accepted in the United Kingdom for use in food applications and in nonparenteral medicinal products. Although isolated earlier publications, such as the one cited by Dr. Baum, suggest that mineral oil may interfere with the absorption of fat-soluble substances, more recent reports indicate that this is not expected at the doses and duration used in the MARINE and ANCHOR studies. A case report of a 17-year-old female patient with chronic constipation who took 400 ml mineral oil daily for 5 months revealed normal ranges of fat-soluble vitamin serum levels and no impact on vitamins A, D, E, and K levels. Sharif et al and Gal-Ezer and Shaoul also evaluated the evidence and concluded that mineral oil–induced fat-soluble vitamin deficiency is unfounded.

In his comment, Dr. Baum notes increases in LDL cholesterol, non-high-density lipoprotein cholesterol, and high-sensitivity C-reactive protein in the placebo arm of the ANCHOR trial and alterations in lipid levels in the placebo group in the MARINE trial, and he suggests that these increases and alterations may have been due to inhibition of statin absorption by the placebo. In response, mineral oil is not included in the list of agents that cause potential drug interactions with statins. Furthermore, in the placebo arm of the MARINE trial, LDL cholesterol decreased by 3% in the intent-to-treat population, decreased by 8% in stable statin-treated patients, and had no change (0%) in patients not taking statins (p = NS; data on file). If the LLP in the placebo group had reduced statin exposure, then LDL cholesterol should have increased after 12 weeks of treatment, not decreased.

Regarding the differences in LDL cholesterol or other lipid levels after treatment with Lovaza (GlaxoSmithKline, London, United Kingdom) versus icosapent ethyl, we disagree that they are due to differences in the choice of placebos (corn oil vs mineral oil, respectively). The effects of these 2 placebos on changes from baseline in LDL cholesterol (−3.0% for mineral oil, −4.8% for corn oil) and triglycerides (+9.7% for mineral oil, +6.7% for corn oil) were very similar in the clinical studies of patients with triglyceride levels ≥500 mg/dl, which had similar designs (as noted in the Lovaza package insert). It is important to note that the ANCHOR and Combination of Prescription Omega-3 With Simvastatin (COMBOS) studies had significant differences in design, rendering cross-study comparisons difficult; patients in COMBOS had an 8-week run-in period on a fixed-dose statin, while ANCHOR enrolled patients receiving long-term background statin therapy before randomization.

Regarding data presentation, having a placebo control is a standard element of randomized clinical trial design. Given the many potential confounders within the context of entering a clinical trial, it is not at all unusual that results are often expressed as a comparison to placebo. The ANCHOR and MARINE reports published in this journal show the efficacy data as change from baseline and as placebo-adjusted data in tabular form. Thus, the change from baseline and the change from placebo were disclosed and available to the readers of these reports. Presenting the data as compared to placebo (placebo adjusted) acknowledges the changes compared to placebo, which is not only commonplace in clinical trial reporting but was the a priori statistical end point of the trials. Finally, this statistical approach was consistent with the special protocol assessment agreements with the FDA for MARINE and ANCHOR and is consistent with the MARINE icosapent ethyl analyses cited in the product label, again, as agreed upon with the FDA.

In summary, although we agree that omega-3 consumption appears to have health benefits, we respectfully disagree that the FDA-sanctioned use of an LLP placebo group compromised the validity of the ANCHOR or any other clinical studies, and we know of no data suggesting that LLP placebo increases the risk for cardiovascular events.