Retraction Watch

Tracking retractions as a window into the scientific process

Retractile dysfunction? Author says journal yanked paper linking Viagra, Cialis to vision problem after legal threats

without comments

The British Journal of Ophthalmology has retracted a 2006 paper which purported to show a link between drugs for erectile dysfunction and a rare form of sudden vision loss called non-arteritic anterior ischaemic optic neuropathy, more commonly known as “Viagra blindness.”

That wouldn’t be terribly interesting, except for this: One of the authors of the paper, a researcher at the University of Alabama named Gerald McGwin Jr., told us that the journal retracted the article because it had become a tool in a lawsuit involving Pfizer, which makes Viagra, and, presumably, men who’d developed blindness after taking the drug:

The article just became too much of a pain in the rear end. It became one of those things where we couldn’t provide all the relevant documentation [to the university, which had to provide records for attorneys]

Ultimately, however, McGwin said that the BJO pulled the plug on the paper.

It was really the journal’s decision to take it out of the literature.

The retraction notice is mute on the reason for the retraction of the blindness paper (and, so far, our requests for comment seem to have fallen on deaf ears). Here’s all it says:

G McGwin Jr, M S Vaphiades, T A Hall, et al. Non-arteritic anterior ischaemic optic neuropathy and the treatment of erectile dysfunction. Br J Ophthalmol 2006;90:154e7. This article has been retracted.

Although the retraction notice just appeared on PubMed, we’ve been aware of it for several months and have been trying to track down more information about it. What we hadn’t seen was this letter in BJO from two Pfizer executives in response to the Alabama research:

Dear Editor,

RE: Non-arteritic anterior ischaemic optic neuropathy and the treatment of erectile dysfunction.

McGwin et al.[1] suggest that treatment of erectile dysfunction with sildenafil (Viagra®) or tadalafil (Cialis®), two phosphodiesterase  inhibitors (PDE5-I), may increase the odds of non-arteritic anterior  ischaemic optic neuropathy (NAION) in men with a history of myocardial  infarction (MI) or hypertension (HTN). We believe, however, that this  study is fraught with significant limitations (as the authors acknowledge) that preclude drawing any conclusions about this relationship.

Case-control studies are susceptible to several biases that must be  carefully considered and controlled for in the study design,  implementation, and analysis.[2] For example, accurate and complete  ascertainment of exposure is critical in a retrospective case-control  study because both disease and exposure occur prior to study initiation.   The authors note that the interviewers were not blinded to the case or  control status of the patient, making it possible that the interviewer  may, even unconsciously, probe cases differently from controls for  exposure to a PDE5-I (interviewer bias).  Another obstacle is that  patients were not consistently interviewed at the time of NAION event (or  index date for controls) about their PDE5-I usage.  Hence, NAION patients  may have been more likely to remember drug usage (recall bias).    Furthermore, exposure misclassification may have occurred as timing, dose, and duration of drug use relative to event onset were not captured  (exposure bias). This information is crucial for drugs used as needed  such as PDE5-I and particularly for short half-life drugs like sildenafil.

Perhaps, the most troublesome weakness of the study was the limited  sample size and differential participation rates of cases and controls,  likely resulting in selection bias that distorts the conclusion.  The  authors note that almost one-fifth of the cases and one-third of the  controls refused to participate.  The baseline cardiovascular  characteristics, while not significantly different (with the exception of  MI) between cases and controls, were consistently more prevalent in the  NAION group.  This finding is not surprising given that these  cardiovascular conditions, especially in combination, are also risk  factors for NAION.[3] Thus, the MI and HTN subgroup analyses presented in Table 3 should be interpreted with skepticism.

Exacerbating the inherent problems are subgroup analyses that had no  a priori hypothesis. The dangers of unplanned subgroup analyses in  research are well documented.[4] Compounding matters is the sparse number  of patients, reflected in the exceptionally wide confidence intervals  (Table 3). The robustness of such extremely small cell numbers must also be questioned, as the observed statistical significance for patients with  MI can be eliminated if only one or two patients are switched to an  alternative category. The authors also did not provide individual  patient totals by exposure group with and without MI, rendering it  impossible to replicate their results.  Furthermore, there appear to be  errors in the numbers/percentages and crude odds ratios presented in Table 2.

In summary, the methodological limitations call into serious question the authors’ conclusions.  For men with a history of MI or HTN, therefore, this study does not provide any valid evidence that the use of Viagra or  Cialis may increase the risk of NAION.

Rachel E. Sobel, MPH,
Pfizer Inc,
150 E 42nd St., MS#150-3-72,
New York, NY 10017.

[e-mail address deleted]

Joseph C. Cappelleri, PhD, MPH,
Pfizer Inc,
Global Research and Development,
Groton, CT.

References

1 McGwin G, Vaphiades MS, Hall TA, Owsley C.  Non-arteritic anterior  ischaemic optic neuropathy and the treatment of erectile dysfunction.  Br  J Ophthalmol 2006; 90:154-157.

2 Rothman KJ, Greenland S. Modern Epidemiology. Philadelphia, PA:  Lippincott-Raven Publishers, 1998.

3 Hatzichristou, D.  Phosphodiesterase 5 Inhibitors and Nonarteritic  Anterior Ischemic Optic Neuropathy (NAION): Coincidence or Causality? J  Sex Med 2005; 2:751–758.

4 Assman SF, Pocock SJ, Enos LE, Kasten LE. Subgroup analysis and  other (mis)uses of baseline data in clinical trials. Lancet 2000; 35:1064- 1069.

(We’re not sure why, but that letter has two dates on it: One says it was accepted on March 9, 2006, while another says it was published online March 18, 2011.)

Now, responses of this sort are hardly unusual. What’s troubling, of course, is if the journal buckled under pressure from either side of a lawsuit. And at the risk of sounding off prematurely,  the notion that a journal would retract a paper not because it was scientifically unsound but because it was causing administrative headaches is bizarre and offensive to the publishing process.

If the paper is valid, retracting it deprives the literature — and therefore other researchers and clinicians — of valuable data. Indeed, the editors thought enough of the paper to commission an editorial about it when it was first accepted, and the paper has been cited a few dozen times. And while Pfizer’s letter poses some legitimate questions, it doesn’t actually claim the paper is invalid, and doesn’t call for retraction as a remedy.

We look forward to being able to update this post with more information.

Written by amarcus41

May 3rd, 2011 at 12:05 pm

Comments