In 2008, The Lancet published a phase 2 study of tesofensine as an adjunct to a hypocaloric diet in the treatment of obesity.1 The study was also published in Ugeskrift for Læger.2 In 2011, The Danish Health and Medicines Authority decided to inspect completed trials at random. This study was among those selected and was inspected in November of that year. Concerns were expressed about two of the five sites where the trial was undertaken, and were brought to the attention of The Lancet in 2012 by the Article’s corresponding author.
Three major points were raised by the inspection. First, Danish regulations require that informed consent is undertaken in its entirety by a medical practitioner, but in one site this duty had been delegated to non-medical personnel. Second, the integrity of blinding was questioned in an earlier inspection in 2007, and rebutted by the investigators. Third, the recording and assessment of adverse events by the contract research organisation was incomplete, because their staff did not consider as noteworthy the recurrence of events that trial participants had experienced on occasions prior to the study. In particular, headache, migraine, stress, and depression were not reliably recorded in people who had experienced these conditions before enrolling in the trial. Therefore, the report concluded that “the overall side-effect profile that is…published in The Lancet is not in accordance with the actual course of the trial”.3
The aforementioned observations raise concerns about this study of which readers should be aware while the editors await further clarification on details from the Danish Health and Medicines Authority.
The study, led by Arne Astrup of the University of Copenhagen, has been cited 50 times, according to Thomson Scientific’s Web of Knowledge. Here’s the original editor’s note that ran on top of it:
Editors’ note: Obesity is becoming more prevalent and poses major health problems for individuals and societies. Weight loss is difficult, but when achieved, is associated with improved health outcomes. In a phase II trial lasting 6 months, obese adults in Denmark who took tesofensine lost 4.5-10.6% of their weight. Tesofensine inhibits uptake of noradrenaline, dopamine, and serotonin. It was well tolerated, although gastrointestinal events, altered mood, and increased heart rate were observed.
NeuroSearch, the Danish company that’s developing tesofensine, had hoped it would be effective against Parkinson’s and Alzheimer’s, but disappointing results ended those efforts. However, researchers noticed weight loss among those who participated in the studies, and decided to investigate the drug’s potential use against obesity. In addition to the Lancet study, a later phase II study showed some efficacy.
Meanwhile, the company seems to be on the verge of closing. On March 18, they announced that they would dismiss most of their employees, and four days later, they said they’d sold all of their lab equipment and furniture.
Update, 7 a.m. Eastern, 4/10/13: Astrup tells us:
Though I am first author of the original paper in The Lancet, and corresponding author, the issues raised in the report of the Danish Medicine Agency (DMA) inspection do not concern my own department, or my role, but two of the other Danish centres, the sponsor and the Contract Research Organisation (Cyncron). Cyncron has since discontinued trading.
As principal investigator I was not involved in the daily running of the project at other centres, and was not made aware of the points raised by DMA until I was notified by the sponsor (Neurosearch A/S) in February 2012, and received a copy of the report. Cyncron was responsible for monitoring and controlling procedures at all centres.
As I mentioned earlier, the leaders of the two centres inspected by DMA and the sponsor do not concur with all the DMA’s the centre-specific issues.
Though the DMA found it was possible to break the code by holding the envelopes up to a strong light, there is absolutely no reason to believe that a member of staff would have done this, and the envelopes were kept in secure storage.
In fact, the investigators believe that only the possible underreporting of some of the adverse events is relevant. However, DMA clearly state in their conclusion that the data on primary and secondary end-points, and serious adverse events, were reported in accordance with the source data.
From my perspective this case reminds us that the proper execution of a trial does not only depend on the investigators (and particularly not only the PI), but also the company and the CRO doing the monitoring and audit. The two latter can fail without the PI have any chance to be aware of the flaws. That is how I see this, and this is my personal lesson to learn from this case.
I drafted an erratum and informed The Lancet, and we have had a very constructive dialogue, so I think that part of the story is positive.
And NeuroSearch pointed us to this press release from last June in which they disclosed the DMA’s findings.