GLP-1 Adverse Event Reports: Volume, Reporting Regimes, and Clinical Terms

Five approved GLP-1 receptor agonists appear in 368,915 adverse event reports received from January 2020 through June 2026. Tirzepatide accounts for 171,728 molecule matches, while exenatide accounts for 14,211. Reporters marked 116,347 cohort reports serious and identified hospitalization on 46,541 reports. The crude serious share varies substantially by molecule, but reporting geography explains much of that variation. United States reports have a 20.3 percent serious share, while major non-US sources range from 97.4 to 99.7 percent. Within United States reports, molecule-level serious shares remain separated but narrow to an 11.0–48.1 percent range. Clinical interpretation therefore depends on reported terms, administration failures, indication, and concomitant drug burden rather than report volume alone.

1. Cohort and counting rules

The cohort includes reports naming semaglutide, tirzepatide, dulaglutide, liraglutide, or exenatide in structured drug fields. FDA received dates define the window from 1 January 2020 through 30 June 2026. A report is a submitted form, not a confirmed patient or an adjudicated causal event. One report can name several molecules, several drugs, and several MedDRA reaction terms.

2. Report volume by molecule

Tirzepatide appears in 171,728 reports, followed by semaglutide in 109,952 and dulaglutide in 65,236. Liraglutide appears in 17,374 reports, while exenatide appears in 14,211. Molecule totals exceed the unique cohort because 8,287 reports name at least two cohort molecules. Volume measures filing activity and does not establish comparative risk or treatment exposure.

3. Report volume by calendar year

Unique cohort volume increased from 24,437 reports in 2020 to 101,878 in 2025. Tirzepatide increased from 6,980 molecule matches in 2022 to 64,477 during 2025. Semaglutide increased from 4,752 matches in 2020 to 30,803 during 2025. Dulaglutide, liraglutide, and exenatide declined or remained stable across completed years. The 2026 half-year remains outside the figure because partial periods distort longitudinal comparison.

4. Patient characteristics recorded in reports

Recorded sex and age describe submitted reports rather than treated populations. Female patients account for 61.6 percent of tirzepatide matches and 46.8 percent of exenatide matches. Age is populated for 50.0–65.9 percent of molecule matches, above the prespecified 40 percent threshold. Mean recorded age ranges from 54.9 years for tirzepatide to 64.9 years for exenatide.

5. Reporter qualification

Consumers filed 295,807 unique cohort reports and marked 23.2 percent serious. Health professionals filed 67,375 reports and marked 63.7 percent serious. Lawyers filed 4,013 reports, including 3,990 reports carrying a serious flag. Reporter composition therefore changes the observed serious share independently of molecule identity.

6. Serious share by molecule

Liraglutide has the highest crude serious share at 67.5 percent, followed by semaglutide at 52.8 percent. Dulaglutide records 30.4 percent, exenatide records 19.9 percent, and tirzepatide records 18.2 percent. This comparison reflects reporting systems, reporter qualifications, indications, and case composition alongside clinical events. Act III therefore tests whether the five molecule denominators are comparable.

7. Composition of the serious flag

The serious designation contains six non-exclusive sub-flags, so component counts exceed 116,347 serious reports. Other serious appears on 84,173 reports, while hospitalization appears on 46,541. Life-threatening events appear on 5,363 reports, and death appears on 4,988. Disabling outcomes appear on 4,243 reports, while congenital anomalies appear on 154.

8. Reaction outcomes as reported

Reaction outcomes provide a second field that does not reproduce the serious flag. Unknown outcomes account for most reaction rows, while not-recovered and recovered rows form the next largest groups. A report can contain several reactions with different outcomes, so reaction rows exceed report counts. Fatal reaction rows and death-flagged reports also use different units and must not be equated.

9. Serious share by reporting country

United States reports comprise 302,660 unique cohort reports and carry a 20.3 percent serious share. United Kingdom reports carry a 99.7 percent share, while Japan records 97.4 percent. Canada, European Union, France, Brazil, Germany, Australia, and China each exceed 98 percent. The country pattern indicates different submission regimes rather than uniform case capture across markets.

10. Serious share within United States reports

Restricting the comparison to United States reports changes every molecule estimate. Liraglutide remains highest at 48.1 percent, followed by semaglutide at 41.2 percent. Dulaglutide records 23.8 percent, exenatide records 17.7 percent, and tirzepatide records 11.0 percent. Geography does not remove all separation, but it eliminates the non-US near-complete serious-report selection.

11. Concomitant drug burden

Liraglutide reports name 13.61 drugs on average, compared with 4.02–5.40 for the other molecules. A total of 5,343 liraglutide reports name at least ten drugs. Semaglutide has the next largest ten-drug count at 18,550, but across a much larger report denominator. Concomitant drug burden indicates different patient and reporting complexity across molecule files.

12. Administration and device terms against clinical terms

A fixed list of 22 MedDRA terms identifies dosing errors, dose omissions, and device-handling events. These terms appear in 97,140 cohort reports, representing 26.3 percent of the cohort. Exenatide has the highest administration share at 71.9 percent, while liraglutide has the lowest at 9.8 percent. Administration reports carry a 9.1 percent serious share, compared with 39.6 percent elsewhere.

13. Reported reaction terms by molecule

The five ranked term profiles separate molecule files more clearly than crude serious shares. Tirzepatide leads with 32,304 incorrect-dose reports, followed by nausea and injection-site pain. Semaglutide leads with 16,525 nausea reports, followed by vomiting and off-label use. Exenatide is dominated by device leakage and device-use terms, while liraglutide contains mainly clinical symptoms.

14. Named clinical signals

Named terms identify reporting patterns but do not establish causal associations. Pancreatitis appears most often among the selected signals, with 2,515 semaglutide and 2,001 tirzepatide reports. Intestinal obstruction appears on 2,402 semaglutide and 1,125 tirzepatide reports. Suicidal ideation, aspiration, gallbladder terms, and thyroid neoplasm occur less often.

15. Approved indications and other recorded uses

Recorded indications combine approved uses, off-label uses, missing values, and free-text coding differences. Type 2 diabetes dominates dulaglutide and exenatide records, while weight control is prominent for tirzepatide and semaglutide. Tirzepatide records also name sleep apnoea and polycystic ovarian syndrome. These counts describe suspect-drug rows rather than unique patients or prescribing prevalence.

16. Supported interpretation

The record supports four conclusions within its reporting limits. Tirzepatide dominates recent filing volume, but volume cannot estimate comparative treatment risk. Crude serious shares combine molecule identity with reporting country, reporter qualification, and concomitant drug burden. Administration and device terms materially change report composition, especially for exenatide and tirzepatide. Clinical comparison therefore requires term-level profiles and controlled denominators rather than one aggregate severity percentage.

17. Limitations

One patient can generate several reports, and one report can name several medicines and reactions. Reporter-set serious flags do not represent independent clinical adjudication across countries or organizations. Drug exposure, prescriptions, treatment duration, dose, and untreated comparison populations are absent from this dataset. Term searches depend on recorded MedDRA wording and can miss related concepts or include broader categories. Coverage ends on 30 June 2026, and later periods are absent rather than zero.