Nearly 8,000 people posted symptoms from a weight-loss drug the FDA has not approved

A University of Pennsylvania team used artificial intelligence to read nearly 150,000 Reddit posts and comments about retatrutide, an experimental weight-loss injection that has not been approved by the FDA, and found 7,823 users who described at least one symptom that could be mapped to a medical term. Increased appetite was the most common. Every figure comes from what strangers chose to write online, and the researchers stress that none of it shows the drug caused anything.

Tech news worth reading: Get Techpresso, the daily email with the most interesting tech news and insights, read by 400,000+ professionals from Google, Apple and OpenAI.

Get Techpresso →

We may earn a commission if you sign up.

Retatrutide is still in phase 3 clinical trials, the last stage before a regulatory decision, and it is not legally available to patients. People who write about taking it are, by definition, using something outside the oversight that normally collects safety data. The Penn analysis is an attempt to read that unmonitored experience for hints worth testing in proper studies.

The Penn team’s haul of posts

The work appeared as a research letter in JAMA Network Open on Oct. 14, 2026, and Penn Engineering announced it the next day. According to the paper, the team analyzed 148,640 retatrutide-related posts and comments from 38,936 users across 27 communities, covering May 2021 through December 2025. Of those users, 13,589 said they were currently taking the drug, and 7,823 of them used at least one term the team could map to a symptom. That last group is the basis for the figure of nearly 8,000.

The number is a count of people who posted, not a count of patients harmed. The researchers excluded expected effects such as weight loss and reduced appetite before counting, so the 7,823 users are those who described something beyond the effects the drug is meant to produce. The Penn release as republished by ScienceDaily on Oct. 21 gives the same totals, rounded to roughly 150,000 posts, 40,000 users and 14,000 self-described users of the drug.

The symptoms people described

Among the terms the paper lists as most frequent were increased appetite at 17.1 percent, fatigue at 15.2 percent, increased energy at 12.4 percent, nausea at 11.6 percent, food craving at 7.3 percent, insomnia at 6.1 percent and increased heart rate at 4.8 percent. Appetite appearing first is the surprising part, because the drug is meant to suppress it. Those shares are the paper’s own tallies of symptom terms, and they describe what people chose to write about, not how likely any one symptom is for a given person.

The team compared the pattern with its earlier study of more than 400,000 posts about the approved medicines semaglutide and tirzepatide, and with phase 2 trial participants, who most often reported digestive complaints such as nausea. In the retatrutide posts, appetite changes, sleep disturbances, increased heart rate and muscle-related symptoms made up proportionally more of the reports, and nausea and other digestive symptoms made up less. The paper describes that comparison as descriptive only, with no statistical testing or matching between groups.

Jena Shaw Tronieri, a senior research investigator at Penn’s Center for Weight and Eating Disorders and a co-author, noted in the Penn Engineering release that retatrutide works differently from existing GLP-1 drugs, so its effects on the body may differ too.

Why the authors call it a hypothesis

Lyle Ungar, a Penn computer science professor and co-author, put the limit plainly, saying the analysis cannot show that retatrutide is causing these symptoms, or even that the Reddit posters are taking genuine retatrutide. Senior author Sharath Chandra Guntuku, who directs Penn’s Computational Social Listening Lab, framed the goal as identifying hypotheses in online conversations that could later be tested. First author Neil K. R. Sehgal, a doctoral student, described the result as a first glimpse into a group of people who sit outside conventional drug safety monitoring.

The reasons for caution are listed in the paper itself. Products that circulate outside regulation can differ in what they contain, some users may be taking several experimental compounds at once, symptoms can be misattributed, and people who post on Reddit may differ from volunteers in a clinical trial. Any of those could explain the pattern as readily as the drug could. The authors say the data cannot establish how often a symptom occurs, whether it is caused by the drug, or how safety compares with approved medicines.

What they do suggest is a change in how clinicians and regulators listen. The paper recommends nonjudgmental screening for use of unapproved peptides, counseling about product quality, reporting of adverse events through the FDA’s MedWatch system, and prospective studies. The peer-reviewed research letter also calls the signals worth formal pharmacovigilance attention.

The study disclosed its funding and conflicts: it was funded in part by the Penn Medical Communication Research Institute, and Tronieri reported an investigator-initiated grant from Novo Nordisk, awarded on behalf of Penn, and consulting fees from Currax Pharmaceuticals. The funding note matters for context, though the paper does not tie any of the disclosed relationships to the Reddit analysis itself. Readers weighing any personal decision about weight-loss medicine should talk with a clinician, since an approved treatment exists and this one remains under study.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

Tech news worth reading: Get Techpresso →

We may earn a commission if you sign up.