Pharma’s Cutting Edge

This is the press release from Merck describing the MEDAL study results. You’ll recall that MEDAL is the prospective, comparative trial of Arcoxia (a COX-2 inhibitor) and diclofenac (an NSAID) with the primary outcome of cardiovascular events. Before the results were release, the study design had been criticized on a couple of key grounds. In particular, Curt Furberg, an outspoken, influential critic of Merck’s handling of Vioxx, writing in the August issue of the American Heart Journal, was critical of the use of diclofenac as the comparator. Now that the unpublished results of MEDAL apparently demonstrate no difference in cardiovascular outcomes with the two drugs, Furberg’s criticism will become the real challenge for Merck. In an interview published in today’s WSJ, Furberg was quoted as saying that: “Most people didn’t expect a difference” …. The take-home message is stay away from diclofenac. The Cox-2 drugs have harmful effects and this study documented that diclofenac is no safer.” Probably not the reaction Merck was hoping for.

But surely Merck knew that diclofenac would be criticized as a comparator when it designed the MEDAL study. So, why didn’t Merck choose Naproxen as the comparator? I suspect we’ll know the answer to this question in short order. In the meantime, what should Merck do to proactively address this crucial criticism? There’s a few things they can do, none of which will be definitive. First, they need to scour the published literature on diclofenac (I’m assuming they already have) and provide support for an independent case-control study or two of diclofenac using data from prescription-event monitoring databases. Both of the efforts have the goal of establishing the relative safety of diclofenac in the “real-world” setting. Merck should also explore preclinical experimentation to determine how/if diclofenac differentiates itself from Vioxx with regards to proposed mechanisms of increased thrombotic risk. Finally, it will be useful for Merck to compare the rates of cardiovascular events in MEDAL with those from a body of recent cardiovascular outcomes studies, using primary data when possible. In particular, it will be important to match MEDAL subjects taking diclofenac with subjects from contemporaneous studies on key factors that would be likely to influence absolute event rates (e.g. age, geographical region, smoking status, concomitant meds, etc.). If MEDAL event rates appear to be the same as those from other studies in which subjects didn’t use diclofenac, it will be harder to argue that diclofenac itself is dangerous.