Years Later, MIST Trial Investigator Suspended in Long-Running Controversy Over PFO-Closure Study Misconduct
Andrew Dowson, MD, the headache specialist who was co-PI of the controversial Migraine Intervention with STARflex Technology (MIST) trial has been suspended from practicing medicine in the United Kingdom, having lost all but one of his appeals over misconduct and dishonesty charges in a British High Court. The lone charge overturned by the judge pertained to MIST data he presented at the TCT 2007 meeting.
Dowson’s suspension, according to the UK’s General Medical Council (GMC) website, went into effect December 2, 2015.
Earlier this year, the GMC’s Fitness to Practice Panel ruled that Dowson was guilty of 7 counts of misconduct—including misconduct with dishonesty—and recommended a 4-month practice suspension. Dowson appealed. Last month Mr. Justice Edis, reviewing the Fitness to Practice Panel rulings, upheld 6 of the 7 charges.
MIST, sponsored by NMT Medical, failed to show a benefit of PFO closure for migraine. Final results were published in Circulation in 2008. The trial findings themselves, however, were swiftly eclipsed by allegations brought forward by the other co-PI, cardiologist Peter Wilmshurst, MD, who spoke out at the TCT 2007 meeting about misconduct by both the trial sponsor and Dowson.
At the same meeting, Dowson himself presented data from MIST I and MIST III, misrepresenting himself as a neurologist and reporting a successful PFO closure rate in MIST of 94%. Wilmshurst, for his part, spoke in a separate session claiming that he had run afoul of the study sponsor following his own echo review showing much lower rates of device success. He had subsequently been refused access to the full dataset, denied an explanation on the discordant data, and ultimately removed as co-PI on the study.
Wilmshurst himself refused to sign off on the Circulation paper and is not named as an author. Of note, the paper includes results that differ from those originally presented and makes no mention of Wilmshurst’s echo review, nor of a review by the trial monitor.
Some of Wilmshurst’s concerns, when reported in the media, led to a costly long-running libel suit that eventually collapsed when NMT Medical went out of business.
The charges upheld by Edis center primarily on Dowson’s failure to disclose a prior breach of protocol in an earlier trial, for which he was removed as PI and sanctioned by the GMC. What is clear from Edis’s judgment is that NMT Medical was fully aware of Dowson’s troubled research history, but Dowson himself neglected to disclose it to fellow investigators and steering committee members, or to the research ethics committee overseeing the MIST trial.
Edis also upheld 4 panel conclusions citing misconduct but no dishonesty. One of these pertained to Dowson’s failure to disclose to the MIST research ethics committee “extensive work” for which he charged NMT Medical 200 British pounds per hour. Edis also noted that Dowson failed to disclose a bonus provision in his agreement with NMT as well as the fact that he later bought and sold shares in the company, although the panel found these charged “unproved.”
The ‘TCT Charge’
The only charge that was not upheld by the High Court was the one Edis identified as the “TCT Charge.” This covers the allegation that Dowson intentionally misled attendees at TCT 2007 by quoting a closure success rate of 94%—the same rate subsequently published in Circulation—and failing to mention Wilmshurst’s review showing much higher rates of residual shunt.
Edis concluded, however, that the reporting of the MIST trial results was a “collective action” and that Dowson was merely presenting the numbers that would subsequently be in the paper, written by multiple authors. “It appears to me that a determination that a single individual in this chain of events was guilty of misconduct when he did no more than accurately foreshadow the paper in his remarks to the TCT conference is flawed,” he noted.
Speaking with TCTMD, Wilmshurst characterized this decision as “a legal technicality, rather than an exoneration” of Dowson.
Wilmshurst is now semi-retired, although he practices 2 days per week within the National Health Service and continues to research PFO closure. He also works part-time “investigating research misconduct.”
While stopping short of saying that the judge’s decision offers a resolution to what has been a long battle, Wilmshurst acknowledges that he “won’t get any more judicial action” in his complaints against trial investigators. “I still feel that I need to get some further correction of the paper and the information in the paper,” he said, “[although] I don’t know [everything that’s incorrect] because I never got all of the data. No investigator got all of the data.”
Circulation did issue a correction to the original MIST publication, but it did not, in Wilmshurst’s mind, go far enough. He said he believes Circulation editors made a key mistake by asking Dowson, the person Wilmshurst sees as responsible for allowing errors to be published in the first place, to also write the correction.
MIST was one of just 2 randomized, sham-controlled trials testing the theory of PFO closure for migraine yet, as a result of its flaws, provides few insights. Wilmshurst, for one, remains convinced that there is more to explore, although the field’s appetite for doing so may be weak.
“There’s no doubt that there is a strong link between PFOs and other types of right-to-left shunts and migraine with aura. Even in MIST, ... the very first thing we found was that there was a very high prevalence of right-to-left shunts in patients with severe migraine, and they weren’t all PFOs. There was a very large prevalence of pulmonary arteriovenous malformations, which is very interesting.”
Calls and emails to Dowson from TCTMD went unanswered.
- Headache or No Headache, PFO-Related Auras Have Same Cause
- MIST Results Fail to Show Migraine-PFO Connection, Trial Design Issues Cited
- MIST Trial: No Clear Answers