Researchers fed 400,000 Reddit posts into an artificial intelligence system and found weight-loss drug users complaining about symptoms that never made it into official drug labels.
Quick Take
- University of Pennsylvania researchers analyzed 410,198 Reddit posts from over 67,000 people who said they took semaglutide or tirzepatide.
- The artificial intelligence tool found 43.5% of these users described at least one side effect, with nausea and constipation matching known clinical trial results.
- A new pattern emerged around menstrual irregularities and temperature problems like chills and hot flashes, symptoms not well covered in current drug labels.
- The study’s lead author says the findings show patterns, not proof that the drugs cause these symptoms.
A Massive Online Data Set Reveals New Patterns
Semaglutide and tirzepatide, sold under brand names like Ozempic, Wegovy, and Mounjaro, have become some of the most talked-about drugs in America. Penn researchers scanned Reddit posts from May 2019 through June 2025, pulling out 410,198 mentions of these drugs. Within that pile, more than 67,000 people identified themselves as actual users, giving scientists a huge real-world sample to study.
The artificial intelligence model sorted through this mountain of posts looking for patterns humans would take years to spot by hand. Nearly half the self-identified users, 43.5%, described at least one side effect in their posts. That’s not a small number. It suggests a huge chunk of people taking these drugs are dealing with something worth talking about publicly.
Familiar Complaints Line Up With What Doctors Already Know
Nausea, vomiting, and constipation topped the list of complaints. That’s not a surprise. Those symptoms already appear on the official warning labels and showed up constantly in the original clinical trials that got these drugs approved. Researchers said this overlap actually strengthens confidence in the tool itself, because it managed to catch real, already-documented side effects using nothing but public social media chatter.
Finding known symptoms first matters because it proves the artificial intelligence system wasn’t just picking up noise. When a method correctly identifies things scientists already know are true, it earns credibility for the new things it claims to have found. That credibility is exactly why the next set of findings grabbed so much attention.
Menstrual Changes and Temperature Swings Emerge As New Signals
Beyond the expected stomach issues, the Penn team flagged something less discussed: reproductive symptoms. Close to 4% of users reported irregular menstrual cycles, along with heavier bleeding and disrupted timing. Users also described chills and hot flashes tied to body temperature regulation, complaints rarely mentioned in official drug materials.
Four percent might sound small, but across tens of thousands of users, that’s a real number of women describing the same unexpected experience. These are the kinds of symptoms that rarely surface in short clinical trials, which typically run for months, not years, and don’t always ask detailed questions about menstrual health or temperature regulation.
The Researcher Behind the Study Urges Caution on Causation
Neil Sehgal, the study’s lead author, made clear this research identifies patterns, not proof. He explained the artificial intelligence model detected how often certain experiences were mentioned by self-reported drug users, but the study does not claim the drugs actually caused those symptoms. That distinction matters enormously for anyone reading headlines about “hidden” side effects.
Separate pharmacovigilance research using the Food and Drug Administration’s adverse event database has also picked up menstrual reporting signals tied to semaglutide, giving this Reddit-based finding some outside company. That research similarly stopped short of proving cause and effect, noting the data shows association, not confirmed harm.
Why This Still Matters for Patients and Doctors
Drug labels get built from clinical trials that often last only months and enroll carefully screened patients. Real-world use over years, by millions of people with different bodies and health histories, can surface things trials miss entirely. This Reddit-based approach gives doctors and drug makers an early warning system worth taking seriously, even while more rigorous follow-up studies get built.
Patients considering these drugs deserve straight talk from their doctors about symptoms beyond the standard nausea warnings. If tens of thousands of real people keep describing the same unusual experience, it’s worth a conversation before starting treatment, not dismissal because it wasn’t in the fine print.
Expect follow-up studies using medical records and insurance claims data to test whether these online patterns hold up under stricter scientific review. Until then, this Reddit-mined research gives patients and doctors a genuinely useful early signal, built from the voices of people who actually took the medication and spoke up.
Sources:
medicalxpress.com, 6abc.com, peptidemethods.com
