BLOG: Dissecting DREAM, part 2: Masked, prospective chaos

ByDarrell E. White, MD

DREAM has been touted in numerous press releases as a randomized, controlled, masked prospective trial to investigate whether fish oil is effective in treating the symptoms of moderate or severe dry eye. It is my contention that it was not controlled, nor was it randomized. No less than the American Academy of Ophthalmology Preferred Practice Pattern agrees with me on that. Did you know there was a DED PPP from the AAO? Neither did I. On page P310, it describes DREAM as a “large scale, masked, prospective trial.” Conspicuously absent would be “randomized” and “controlled.” Even more damning is the absence of a “strength” rating given by the committee; DREAM is apparently too weak to merit any rating at all.

Pro level shade from AAO.

We can trust that it was masked (neither investigators nor subjects were aware of whom was receiving active ingredient). For me, pretty much nothing else really fits, even the prospective part. In order for a study to be truly prospective (treatment or placebo begin and end at discrete points), one must ensure that subjects are “clear” of the agent under study. Most DED docs feel that both the buildup and decline of 0-3 in RBCs is slow. Some subjects in both groups were taking up to 1,250 mg of fish oil both before and during the study. When did treatment start for these subjects?

This also contributes to the problematic nature of randomization in DREAM. Prior to initiating the study, the subject group had a significantly higher EPA level than did the placebo group. One should assume that a higher level of EPA leads to a dampening of symptoms, thereby making it harder to show an effect through the addition of more fish oil. This was “addressed” through some sort of statistical manipulation of the OSDI data in this group. Funny thing, that: In order to adjust for the effect of higher EPA on OSDI, you have to assume that EPA makes people feel better.

Still, it is the total lack of control of DED treatment regimens (plural emphasized) that nullifies and renders irrelevant conclusions that single out any individual component of those treatments, including fish oil. First, there are simply too many treatment variables to isolate a single treatment. It’s like trying to redo the Framingham Heart Study with all 10 variables but only 400 subjects. Second, no effort was made to isolate fish oil from other highly potent treatments. Nor was there any screening of the appropriateness of those treatments. For example, which patients treated with CsA experienced a decrease in symptoms? How long had they been treated? Were they among the patients whose treatment was altered during the trial?

Ah, the treatments that were changed. In a controlled trial, the one thing that is under control is the treatment. Seventy-four percent of patients in the fish oil treatment group and 75% of the patients in the placebo group had at least one change in their treatment over the course of the 12 months of the study. Note that I deliberately use “patient” to describe them and not “subjects”; the individuals observed in this study were not being treated with strict adherence to a study control but were rather “controlled” by their treating doctors. Both the ongoing signs and symptoms of DED were assessed and treated in an ad hoc manner based on physician preference.

In the context of a study, that’s not control, that’s chaos.

Because both groups got better, there is only one logical conclusion that can be drawn from the DREAM study: Aggressive multimodal treatment of DED with a regimen that includes fish oil leads to a lower symptom burden for DED patients. In order to evaluate the precise effect (or lack thereof) of fish oil, additional study of a dramatically larger “real-world” cohort, or a tightly controlled study that isolates the effect of DHA/EPA, would be necessary.

Disclosure: White reports he is a consultant to Allergan, Shire, Sun, Kala, Ocular Science, Rendia, TearLab, Eyevance and Omeros; is a speaker for Shire, Allergan, Omeros and Sun; and has an ownership interest in Ocular Science and Eyevance.