After reading a recent study about GLP-1 treatment in the International Journal of Obesity, David B. Allison immediately became skeptical about the paper’s analysis. The article, published in May 2024, found people who combined a GLP-1 therapy with another weight loss drug added to GLP-1 therapy led to greater weight loss in patients.
“I could not really comprehend exactly what analysis they did,” Allison, chief of nutrition and director of the Children’s Nutrition Research Center at Baylor College of Medicine in Houston, told Retraction Watch. “And more so, I could not comprehend how the analysis they did would give results that would be informative of the conclusions they drew. So I was scratching my head a little bit.”
The IJO paper was a retrospective cohort study of adults with obesity who had been prescribed a GLP-1 therapy, specifically Saxenda and Ozempic. The study compared patients who received a GLP-1 alone with those receiving the GLP-1 therapy and then bupropion/naltrexone added to their regimen. The Food and Drug Administration approved bupropion/naltrexone in 2014 for chronic weight management in obese adults.
The study found the combination regimen resulted in a small but significant extra weight loss for patients, resulting in roughly 4% additional total body weight loss over six and 12 months.
Allison, who has worked for a decade on correcting statistical errors in the nutrition literature, asked some of his mentees to see if they “could make sense” of the paper’s analysis, he said. But the group also had trouble understanding it, he said. With the authors’ cooperation, Allison’s team reviewed the raw data associated with the findings, ultimately concluding the results were not fully reproducible, said Allison, who serves as a statistical editor for the journal.
After looking into Allison’s findings, IJO retracted the article in December.
Springer confirmed it launched an investigation into the article after “a member of the journal’s editorial board raised concerns about the validity of its conclusions,” Maria Hodges, publishing director for biomedicine at Springer, said by email.
“This involved an assessment of the article’s original data by one of the journal’s statistical editors and members of their research team, in addition to consultation with the article’s authors, independent experts, the journal’s Editors–in–Chief, and the Springer Nature Research Integrity Group,” Hodges told us. “Following this, the Editors-in-Chief concluded that the main finding of this article is not substantiated and that retraction was the appropriate editorial action to take.”
All authors disagreed with the retraction, the retraction notice states.
Nadia Khan, a clinician scientist at the University of British Columbia in Vancouver and the principal investigator on the study, told us she and her colleagues “felt they couldn’t accept a retraction unless they had more information on what the specific issues were, and whether the issues constituted enough of a departure to warrant a retraction.” Khan and two other authors also practice at Revolution Medical Clinic, a Vancouver clinic that provides obesity and metabolic disease treatment.
Khan said the authors were “forthcoming with immediately supplying all raw data and all statistical codes to the journal,” when asked. She declined to answer further questions, but said authors are preparing a reply letter in defense of their study to submit to the journal.
The IJO paper garnered some attention when it came out, including at least six news summaries and about 40 posts on X. The study has been cited eight times, according to Clarivate’s Web of Science.
In addition to unverifiable results, Allison said his team found “further things that seemed questionable.” He declined to go into detail about what those things were, but said his team has summarized the specifics in a report that will be published in the journal.
Allison added the IJO paper is related to the idea of “treatment response heterogeneity,” a growing field of study. Heterogeneity of treatment effect is the nonrandom variability in treatment effects for patients within a population.
“I do think the field needs to up its game in terms of getting people to understand what are reasonable and legitimate methods for studying treatment response heterogeneity,” he said.
Allison said he was pleased with how IJO handled the concerns about the article, especially considering the “tremendous variability” among journals in their willingness to fix reported errors, correct or retract articles, and take prompt action, he told us. Another paper his group critiqued in Evidence-Based Complementary and Alternative Medicine was retracted more than three years after his team first raised concerns about the paper. In other cases, editors have taken swifter action.
“I’m very pleased with the IJO journal editors that they’re taking matters seriously, and I think it would be good if more journal editors followed their example,” he said.
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on X or Bluesky, like us on Facebook, follow us on LinkedIn, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].