Industry & Medicine: It’s Complicated 3

Some people think industry exerts a uniformly negative force on medicine, or at least that’s the only aspect they focus on when they write or talk about the issue. Others focus exclusively on the beneficial effects of industry, and exhibit amnesia in their failure to recall the numerous instances in recent years in which the influence of industry has caused tremendous harm. Both perspectives possess a certain measure of truth, but the insistence on a one-sided, black-and-white view ultimately causes more harm than good.

One recent event serves as a case study of both of these extreme views. I want to outline where I think each side went wrong, and very briefly propose a perspective that allows for a more nuanced perspective on industry.

The Black Knight: Arnold Relman

Not long ago the former NEJM editor Arnold Relman said that NEJM should not have published a recent paper, the ARISTOTLE study, which compared apixaban to warfarin for the prevention of stroke in AF, and which was widely hailed as not just a positive trial but a well-performed trial with a very clear message.

Relman stated that ARISTOTLE was fatally flawed due to conflict of interest issues because many of its authors had deep ties to industry (including 3 who work for Bristol Myers Squibb), and that the current editors of NEJM were insufficiently critical of the trial. He pointed to 2 major omissions in the discussion section of the NEJM paper as evidence of this problem.

The first omission was the lack of discussion in the paper of the 7,000 patient European subgroup (out of 18,000 total patients), in which no difference between apixaban and warfarin was observed. However, the trial was not powered to achieve a significant difference in the geographic subgroups, the p value for interaction was 0.44, and although no benefit was observed in this subgroup, neither was there any hint of harm. This does not strike me as a major flaw of the trial.

The second omission, according to Relman, was the authors’ failure to discuss the fact that 35% of the time the patients taking warfarin were not taking a therapeutic dose of the drug. Relman said this could explain why warfarin-treated patients had a worse outcome than the apixaban-treated patients. On this point Relman is simply wrong. It’s widely agreed within the anticoagulant community that the 65% time within the therapeutic range (TTR) in the trial is about as good as it gets in the real world. (For more on this topic, see the recent FDA panel about the ROCKET AF trial, which spent a great deal of time talking about exactly this issue.)

Overall, I think Relman chose a poor example of the dangers of conflict of interest, overstated his position, and by doing so actually harmed his position. Relman said “the study was not well peer-reviewed” and that neither Marcia Angell (another former editor of NEJM) nor himself “would have accepted this paper for publication.” But really this is just silly: no editor of NEJM would ever outright reject a paper of the importance and significance of ARISTOTLE. He would not have lasted 5 minutes in his job if that had been his practice at the time.

The White Knight: Chris Cannon

Here’s a comment sent to CardioExchange (registration required) from Chris Cannon, in response to readers who supported Relman:

Earth to everyone – ALL trials of new drugs are funded by the company that makes the drug. Thus EVERY bit of our current medical therapy/evidence based medicine is based on trials funded by the manufacturer of the drug (which usually also includes scientists who discovered and developed it). Thus all the chest beating of how these trials are bad is ridiculous and naive. This and all major trials are conducted to high standards, and if successful, they are then reviewed by FDA, and often approved – thus they meet VERY high levels of scrutiny. People need to get off their high horses and stop trying to make headlines. And former editors might look back at the papers they did publish and see that they too had many drug trial sponsored by drug companies (that had the same high level of excellent conduct).

Cannon’s perspective is a valuable correction to Relman’s, but it replaces the simplistic “black” of Relman’s viewpoint with the  equally simplistic “white” of an industry defender. We only need to look at the multitude of industry scandals over the last decade– in cardiovascular medicine, Vioxx, Avandia, and Vytorin come immediately to mind– to recognize  that not every drug has been subject to the “VERY high levels of scrutiny” as stated by Cannon and that not all trials have been conducted with the highest of standards.

It’s a mistake to think these problems have now been fixed after the scandals of the past. I’ll cite just one very recent and ongoing example: dronedarone (Multaq), which is a case study of all that can go wrong when conflicts of interest run amok. I’ve written extensively about dronedarone in the past so I won’t repeat myself here, but it’s clear that without the all-pervasive effect of industry influence, the drug would not have been approved with a broad indication and adopted by European and US guidelines with such haste. Now, of course, we are in the midst of a rapid and embarrassing reversal of this ill-considered process, and we may never fully understand the extent of the damage that may have been caused by this drug.

It’s Not So Simple

How to avoid the twin pitfalls of a blanket condemnation or a wholesale endorsement of industry? What would a more balanced perspective look like? The ROCKET trial strikes me as a reasonable place to start. Like ARISTOTLE, ROCKET compared a new anticoagulant, rivaroxaban, to warfarin for stroke prevention in AF patients. When ROCKET was first presented last November and later published in NEJM it also received glowing reviews. The ROCKET investigators, many of whom were colleagues of the ARISTOTLE investigators, also had deep ties to industry, and perhaps these relationships helped lead the investigators to develop an overly sunny view of the trial and the drug.

But then  a funny thing happened when ROCKET got to the FDA. The FDA reviewers raised all sorts of serious questions about the trial, and although the FDA advisory committee ultimately voted to recommend approval of the drug, we now have a much less positive view of the trial, and of rivaroxaban, than we had before, and its commercial prospects have been considerably dampened. (It will be interesting to see what the FDA reviewers find when they get their hands on ARISTOTLE.)

But even after acknowledging that they are not perfect drugs, and that the overly sunny view proposed by the drugs’ makers and their academic champions is not realistic, we should not forget that apixaban and rivaroxaban represent a significant advance, one of the great achievements of big pharma in recent years. Economic issues aside, they’re surely preferable in most respects to the drug they’re designed to replace, the universally despised warfarin. In a decade or so, when these drugs start to go off patent, the entire world will reap the benefits, just as it is now enjoying the benefits of generic statins, ACE inhibitors, and many other drug classes.

In summary, I don’t think we should either demonize or praise industry. Instead, we need to retain our ability to think critically, and remember that the world is neither black nor white.


  1. When I compare the meds available now with what was available when I started practice in 1980, I am very thankful for advances brought forth by industry. It would be naive to think that these advances could take place without a profit motive. The important thing is to keep focused on the data and results–usually truth wins out in the long run.

  2. Last weekend I attended the National Physicians Alliance annual conference in Washington, D.C. Dr. Relman was one of the speakers and I went up to him during a break and discussed some of these issues with him. My understanding from talking to him is that he would not have published the article in the form it was published, but would have considered the article for publication if the authors had revised the discussion section to discuss these two issues discussed in Larry’s post.

  3. One of the most interesting parts of your post is the fact that the FDA advisory committee that reviewed the ROCKET trial did their job. Unfortunately, these committees can be hit or miss in terms of their own conflict of interest problems and their ability to look carefully at the data. (One has only to review the transcript of the committee that reviewed the VNS device (which Jeanne Lenzer and I wrote about for the BMJ to see that some committees do a better job than others.

Leave a Reply