Largest tirzepatide FAERS study finds medication errors top the signal list

Researchers analyzing the FDA Adverse Event Reporting System (FAERS) say the largest look yet at tirzepatide — the molecule in Mounjaro and Zepbound — found that medication errors and injection-site reactions produced stronger reporting signals than the gastrointestinal side effects the drug is best known for ^[1].

The study pulled FAERS reports that listed tirzepatide as the primary suspected drug from the second quarter of 2022 through the fourth quarter of 2025, covering 123,145 reports and 251,435 individual adverse events ^[1]. Women accounted for 62.30% of reports, and the mean age was 53.7 years ^[1]. Tirzepatide was approved by the FDA for type 2 diabetes in 2022 and for chronic weight management in 2023 ^[1].

What the analysis found

The authors used four standard signal-detection methods — ROR, PRR, IC and EBGM — and coded events using MedDRA terminology ^[1]. Major signals included medication errors, injection-site reactions and gastrointestinal disorders, with the authors concluding that medication errors and injection-site reactions were the strongest ^[1].

Timing mattered. Among reports with valid onset data, the median time to onset was 13 days, and 67.1% of events occurred within 30 days of starting treatment ^[1]. The authors wrote that this pattern "suggest[s] the importance of closer monitoring during the first month of therapy" ^[1].

The study also flagged several "novel unlabelled signals" — reported events that are not listed in the current product labeling — including eructation (burping), hunger, food craving and starvation ketoacidosis ^[1]. The authors said these signals "warrant further investigation" ^[1]. The abstract does not report how many cases of each were involved, or how severe they were.

How tirzepatide compared with semaglutide

The team also ran a head-to-head comparison against semaglutide, the molecule in Ozempic, Wegovy and Rybelsus. Compared with semaglutide, tirzepatide showed lower reporting odds of vomiting, constipation and decreased appetite ^[1]. The abstract does not give the specific odds ratios, nor does it list any events for which tirzepatide had higher reporting odds than semaglutide.

It is important to understand what this kind of study can and cannot show. Disproportionality analysis compares how often an event is reported for one drug versus everything else in the database; it measures reporting patterns, not the actual rate of side effects in patients, and it does not establish that the drug caused the event ^[1]. FAERS reports are submitted voluntarily by patients, clinicians and manufacturers, so the number of reports depends partly on how widely a drug is used and how much attention it gets in the news.

Why it matters for patients

The medication-error finding is the most unusual part of this analysis. Both tirzepatide and semaglutide are injected with pens or, in compounded form, with vials and syringes, and "medication error" in FAERS covers a broad range of problems — wrong dose, wrong device handling, wrong timing. The study does not break down what types of errors were reported, so the specific causes are not yet known from this source. But the fact that errors topped the signal list, alongside injection-site reactions, suggests that how the drug is administered is a meaningful part of its real-world safety picture, not just what the molecule does in the body ^[1].

The timing data gives patients a rough sense of when problems tend to surface in reports: about two-thirds within the first month, with a median of under two weeks ^[1]. That does not mean later side effects don't happen — reports without complete onset data were excluded from that calculation ^[1].

The comparison with semaglutide may be of interest to people weighing the two drugs, but lower reporting odds for vomiting, constipation and decreased appetite are not the same as a head-to-head clinical trial result. Reporting behavior differs between drugs and over time.

What happens next

The sources do not describe any FDA action, label change or regulatory review tied to this analysis. The authors' stated next step is further investigation of the unlabelled signals, including starvation ketoacidosis ^[1]. The paper carries a 2026 copyright from John Wiley & Sons, and the authors declared no conflicts of interest ^[1]. Whether any of the newly flagged signals leads to labeling changes for Mounjaro or Zepbound is not yet known.