Orlistat, a weight-loss drug taken by millions of patients during the last two decades, has been buoyed by problematic clinical studies that “systematically understated” the drug’s potential harms, according to a new analysis. Danish researchers who reviewed data summaries and published journal articles found that seven drug trials funded by Roche in the 1990s downplayed the frequency of apparent adverse effects, such as diarrhea and incontinence.
The new research didn’t include the data reports that were submitted to the FDA when orlistat was approved in the United States––first in 1999 as the prescription drug Xenical (Roche) and later as the over-the-counter product Alli (GlaxoSmithKline).
According to STAT News, sales of orlistat have slumped in recent years partly because of its reputation for unpleasant gastrointestinal adverse effects. A 2007 study of approximately 17,000 patients found that only 6% of participants were still taking Xenical after one year; after two years, it was down to 2%.
Under the Freedom of Information Act, the researchers obtained clinical study reports (CSRs) from seven randomized, placebo-controlled trials of orlistat submitted to the European Medicines Agency (EMA) by Roche in the 1990s. Two researchers independently extracted data on adverse events (AEs) from protocols and CSRs. Corresponding published papers were identified on PubMed, and AE data were extracted from that source as well. All three sources were compared.
The researchers found that none of the protocols or CSRs contained instructions for investigators on how to question participants about AEs. In the CSRs, gastrointestinal AEs were coded only if the participants reported that they were “bothersome,” a condition that was not specified in the protocols for two of the trials. Serious AEs were assessed for their relationship to the drug by the sponsor, and all adverse events were coded by the sponsor using a glossary that could be updated by the sponsor. According to the investigators, only between 3% and 33% of the total number of investigator-reported adverse events from these trials were reported in publications because of post hoc filters, although six of the seven papers stated that “all adverse events were recorded.”
For one trial, the researchers identified an additional 1,318 AEs that were not listed or mentioned in the CSRs but could be identified through manually counting individual AEs reported in an appendix. They found that most of the patients had multiple episodes of the same AEs that were counted only once, although this was not described in the CSRs. The investigators also discovered that participants treated with orlistat experienced twice as many days with AEs compared with participants treated with placebo (22.7 day versus 14.9 days, respectively; P < 0.0001). Moreover, compared with the placebo group, AEs in the orlistat group were more severe. “None of this was stated in the CSR or in the published paper,” the investigators noted.
All of the studies still stand in the scientific literature.
“The reporting of trials of orlistat in the 1990s understated harms in the summarized results submitted to the EMA for drug approval and in the published papers,” the researchers concluded. “Based on the characteristics of harms observed and reported in these trials, we suggest that reports of harms include duration of adverse effects. We also suggest that systematic reviews of drugs might be improved by including protocols and CSRs in addition to published articles.”
A spokeswoman for Roche told STAT News that reporting on the trials “was done according to the standards of the time” and that the company “continuously update[s] our methods” to comply with changing standards.