Researchers who used AI to analyze hundreds of thousands of social media posts say they have identified unreported GLP-1 side effects, including menstrual changes, fatigue and temperature-related symptoms, suggesting these patterns warrant further investigation.
In a new study, published in Nature Health, researchers said the patient-reported side effects to the popular weight loss drugs may not be fully captured in clinical trials or regulatory documents. Still they caution that this study cannot establish cause and effect.
The peer-reviewed research used artificial intelligence to analyze more than 400,000 posts on Reddit from nearly 70,000 users discussing drugs such as semaglutide, which includes brand names like Ozempic and Wegovy, and tirzepatide branded as Mounjaro or Zepbound.
The researchers analyzed this data set of posts spanning from May 2019 through June 2025.
Researchers from the University of Pennsylvania school of engineering and applied science say the findings don’t prove the medications caused these symptoms, but they point to patterns worth further investigation.
“Clinical trials generally identify the most dangerous side effects of drugs,” Lyle Ungar, professor in computer and information science and a co-author on the study, said in a press release. “But they can fail to find what symptoms patients are most concerned about; even though social media is not necessarily representative, a large collection of posts may reflect additional concerns.”
What the study found
Around 44 per cent of Reddit users analyzed reported at least one side effect, most commonly gastrointestinal issues, like nausea, which are already well documented.
The study also found:
- More than 16 per cent reported fatigue and vomiting.
- More than 15 per cent reported constipation.
- Nearly 13 per cent reported diarrhea.
But the AI analysis also surfaced less commonly discussed concerns, including symptoms related to reproductive health.
Nearly four per cent of users reporting side effects described menstrual irregularities, including heavy bleeding or missed cycles.
Users also reported temperature-related complaints like chills, feeling unusually cold, hot flashes and fever-like symptoms.
Fatigue was ranked among the most common complaints, despite appearing less frequently in clinical trial data.
Psychiatric symptoms were also reported by nearly 13 per cent of users, with four per cent noting anxiety and three per cent mentioning depression.
Neil Sehgal, a doctoral student and the study’s first author, told CTVNews.ca in a Zoom interview Monday that researchers can’t say for certain whether these drugs are causing menstrual irregularities. “But we just found these signals in the data and so we’re calling for a future investigation.”
For menstrual irregularities, Sehgal said “you would expect that it would be higher in an all-women sample.”
Sehgal emphasized that clinical trials are the gold standard but they also have limitations.
“They’re pretty short so you might not pick up on symptoms that emerge after a few years. They’re also expensive and they can’t always enroll tens of thousands of people,” Sehgal said.
Novo Nordisk Canada, a pharmaceutical company that produces insulin pens and GLP-1 medications like Ozempic and Wegovy, said in a emailed statement to CTVNews.ca that they are aware of the study.
“While we value real-world patient perspectives and take them seriously as part of our ongoing pharmacovigilance activities, social media analyses have inherent methodological limitations,” it said.
“Novo Nordisk continuously monitors the safety of all our medicines through robust pharmacovigilance systems, in close collaboration with regulatory authorities worldwide.”
Eli Lilly Canada, which produces tirzepatide, said in an email to CTVNews.ca “patient safety is Lilly’s top priority. We take any report involving patient safety seriously, and actively monitor, evaluate, and report safety information for all our medicines, including tirzepatide.”
How it worked
Researchers pulled Reddit posts where people talked about obesity and weight loss.
They filtered that down to a list of posts where people mentioned common GLP-1 drugs or used different slang — with some users calling Ozempic “Oz” and semaglutide “sema.”
AI then determined if these individuals were using the drug themselves or just talking about it. Once they had a list of people who were confirmed to be disclosing their own medication use, the researchers passed that information through another AI tool to see if those individuals mentioned side effects.
Researchers confirmed the data at each step, Sehgal said.
Social media and its limitations
The research builds on a growing field sometimes called “computational social listening,” where scientists analyze online discussions to identify health trends, according to the study’s senior author Sharath Chandra Guntuku. Using large language models — similar to those behind tools like GPT — the team was able to scan massive amounts of instructed text and map user descriptions to standardized medical terminology.
Researchers caution that Reddit users are not representative of the broader population. The platform skews younger, more male and heavily U.S.-based.
“We’re getting both extremes on Reddit,” Sehgal explained. “If GLP-1 was ‘OK’, would you leave a review about your Ozempic? Maybe not, but if it was a really amazing experience or if you had crippling side effects, you would try to share it with other people.”
Experts say self-reported data online can be influenced by perception, bias and incomplete information and should be interpreted carefully.
Sehgal said Reddit has more users, extensive posts and is one of the few platforms that gives public access to use data for research.
What happens next
GLP-1 drugs have surged in popularity in recent years for both diabetes treatment and weight loss. The speed at which their popularity has grown, researchers say, makes it even more important to monitor real-world patient experiences.
Researchers say social media may offer early clues about emerging side effects for fast-trending treatments.
The researchers hope their findings prompt further clinical study into the symptoms patients are already discussing online — particularly those not yet well understood.
They also plan to expand their analysis to beyond Reddit and English-language posts to see whether similar patterns appear across different populations.
“I think one of the bigger opportunities might just be to apply the same framework to other medications that are really exploding in popularity faster than formal safety monitoring infrastructure can keep up with,” Sehgal said.
For now, the researchers emphasize that patients should not stop or change medications based on social media reports alone and consult their health-care provider, but that those experiences may still hold valuable insights.
The authors report no outside funding for the study. However co-author Jena Shaw Tronieri, and senior research investigator at Penn’s Center for Weight and Eating Disorders, reported receiving an investigator-initiated grant on behalf of the University of Pennsylvania from Novo Nordisk, a manufacturer of GLP-1 medications, as well as consulting fees from Currax Pharmaceuticals. Other authors reported no conflicts of interest.
