Scientists from the U.S. used artificial intelligence to analyze hundreds of thousands of patient messages online and identified possible side effects of medications based on semaglutide and tirzepatide that are rarely mentioned in official instructions.
The results of the study published in the journal Nature Health.
Researchers from the University of Pennsylvania analyzed over 400,000 posts left by 70,000 users on the Reddit platform over the past five years. Using large language models, they examined the personal experiences of individuals who took popular diabetes and weight loss medications, including semaglutide and tirzepatide.
The authors of the study note that clinical trials of drugs typically focus on identifying serious and potentially life-threatening complications. However, in everyday life, patients often encounter less severe but unpleasant symptoms that are not always reported to doctors.
At the same time, people on internet forums discuss any changes in their well-being in detail, which allows for additional insights into the real experiences of using the medications.
According to the analysis results, about 44% of users mentioned at least one side effect. Most complaints were related to already known gastrointestinal disorders. However, artificial intelligence also identified a number of symptoms that are currently underrepresented in official databases and instructions.
One of the most unexpected findings was reports of menstrual cycle disturbances. Nearly 4% of users experiencing side effects complained of intermenstrual bleeding, cycle irregularities, or unusually heavy periods.
Additionally, many discussion participants reported a constant feeling of cold, chills, or, conversely, hot flashes. Another common complaint was chronic fatigue. According to the study, this symptom ranked second in frequency of mention, although it appears significantly less often in official drug manufacturer reports.
The authors emphasize that the results of the analysis of social media posts do not prove a direct link between the medications and the identified symptoms. Nevertheless, such data may help scientists and doctors pay attention to possible side effects that require further investigation.
Researchers believe that using artificial intelligence to analyze large volumes of user reviews can become an important tool for more quickly identifying potential risks of medications after they hit the market.
