Retatrutide, a next-generation GLP-1 weight-loss drug, is not yet approved for use. Nevertheless, online forums reveal that people are purchasing unapproved versions of the drug, which are not intended for human consumption, and sharing their experiences.

Now, researchers at the University of Pennsylvania have used AI to take a closer look at these online conversations, analyzing nearly 150,000 posts and comments from roughly 40,000 Reddit users, to provide the first systematic look at the kind of side effects these users claim to be experiencing.

“Our goal is to identify hypotheses in online conversations that we could test. AI helps us find recurring patterns in people’s self-reports, including experiences that may be missed by conventional drug-safety monitoring,” says Sharath Chandra Guntuku, Research Associate Professor in Computer and Information Science (CIS), Director of Penn’s Computational Social Listening Lab and senior author of a new paper in JAMA Network Open.

Among Reddit users who shared classifiable potential symptoms, the most frequently reported symptom was increased appetite, which differs from retatrutide’s Phase 2 clinical trials, where gastrointestinal complaints were identified as the most common side effects.

Because of the Reddit sample’s limitations, the researchers underscore that their findings should be treated with caution. These differences do not establish that retatrutide causes the reported symptoms or that the experiences of the Reddit users reflect those of the clinical trial participants, but rather point to the kinds of patterns that AI-assisted analysis can uncover.

“We can’t say that retatrutide is causing these symptoms, or even that Reddit users are actually taking genuine retatrutide, since the drug is not legally available, and gray market formulations vary,” says Lyle Ungar, Professor in CIS and a co-author of the study.

In other words, rather than being taken as evidence of previously unrecognized side effects, the results should serve as a starting point for other public health researchers. “This is the first glimpse into the experiences of a population that is otherwise outside conventional drug safety monitoring because they are taking unapproved versions of a drug that has yet to finish clinical trials,” says Neil Sehgal, the study’s first author and a doctoral student in CIS.

The new study builds on earlier research by the same Penn team, which used large language models (LLMs) to analyze more than 400,000 Reddit posts about semaglutide and tirzepatide, identifying potential side effects — including menstrual irregularities — that may not have been fully captured in clinical trials or regulatory documents. “We wanted to apply the same methodology to gray-market retatrutide,” says Sehgal.

Unlike semaglutide, which activates one hormone receptor involved in regulating appetite and blood sugar, or tirzepatide, which activates two, retatrutide activates three, giving the researchers reason to investigate whether users might experience a different pattern of symptoms.

The researchers used AI to sift through nearly 150,000 Reddit posts and comments, identifying approximately 14,000 users who described personally taking an unapproved version of retatrutide. The AI tools then extracted symptoms those users attributed to the drug and matched their descriptions to standardized medical terminology, winnowing the sample to 7,823 users reporting at least one classifiable symptom beyond expected effects like weight loss or appetite suppression.

“A researcher might be able to examine a few thousand Reddit posts by hand, but looking at nearly 150,000 would be an enormous undertaking,” says Sehgal. “AI makes it possible to analyze these conversations at scale, identifying patterns in what people are sharing that would otherwise be difficult to recognize.”

The team found that, while nausea and other gastrointestinal symptoms were the most common side effects reported in retatrutide’s Phase 2 trials (the drug is currently in Phase 3 trials, the final step before FDA approval), the Reddit users claiming to take unapproved versions of retatrutide more frequently described a different set of symptoms, including increased appetite, an unexpected finding given retatrutide’s demonstrated weight-loss effects in Phase 2 clinical trials.

A similar contrast emerged when the researchers compared their findings with their earlier study of semaglutide and tirzepatide. Reddit posts about gray-market retatrutide featured proportionally more reports of appetite-related symptoms, sleep disturbances, increased heart rate and muscle-related changes, and fewer reports of nausea and other gastrointestinal symptoms.

“Retatrutide works differently from existing GLP-1 drugs, so it’s possible that its effects on the body are different, too,” says Jena Shaw Tronieri, Senior Research Investigator at Penn’s Center for Weight and Eating Disorders and a co-author of the study.

“But the people discussing it on Reddit also may be different,” she notes. “Users willing to obtain an unapproved medication on the gray market likely differ from other clinical populations. Some may be taking multiple experimental peptides, and we can’t verify exactly what they’re taking. We need more research to understand what, if anything, these differences in self-reported symptoms mean.”

The researchers hope that AI-assisted social media analysis can help identify potential safety concerns earlier, giving clinicians and public health researchers a starting point for further investigation.

Sehgal points to the team’s previous GLP-1 study as an example of how that process can work.
After the Penn researchers identified reports of menstrual irregularities among semaglutide and tirzepatide users, another research group was inspired to investigate the same symptoms using the FDA’s adverse-event database.

Their subsequent study, published in Obstetrics & Gynecology, also identified disproportionate reports of certain menstrual changes, providing additional evidence that the symptoms merit investigation.

“That’s what we hope will happen with this study,” says Sehgal. “Other researchers can take what we find on Reddit and look for the same patterns in other datasets, using entirely different methods. Whether those signals hold up or not, the goal is to build a better understanding of what people are actually experiencing.”

This study was conducted at the University of Pennsylvania School of Engineering and Applied Science and Perelman School of Medicine. Additional co-authors include Benjamin Rader of Boston Children’s Hospital. This project was funded in part by the University of Pennsylvania’s Penn Medical Communication Research Institute. Tronieri reports receiving an investigator-initiated grant, on behalf of the University of Pennsylvania, from Novo Nordisk and receiving consulting fees from Currax Pharmaceuticals, LLC. The other authors report no conflicts of interest.