Retatrutide Side Effects: Trial Data

The single most useful number on this page is the last one. Across the TRIUMPH-1 dose ladder, discontinuation for adverse events climbed from 4.1% at 4mg to 6.9% at 9mg to 11.3% at 12mg, against 4.9% on placebo. At the bottom of the ladder, retatrutide was better tolerated than placebo on that measure. At the top, it drove people out of the trial at more than twice the placebo rate.

Everything else in this filing is context for that climb: which events cause it, how they scale with dose, and where the published data stops.

Three habits make the difference between reading this data and misreading it.

Compare within a trial, not across trials. The four readouts enrolled four different populations: general obesity, type 2 diabetes, severe obesity with cardiovascular disease, and obesity with knee osteoarthritis. Baseline symptom burden and concomitant medication differ between them, and so does how investigators solicit and code symptoms. The 9mg to 12mg comparison inside one trial is clean. The TRIUMPH-1 to TRIUMPH-4 comparison of the same event is not.

Read the placebo column every time. Every trial in this programme reported double-digit gastrointestinal rates on placebo, because these are common background symptoms in this population. Nausea at 14.8% on placebo is the correct baseline for reading nausea at 42.4% on 12mg. The difference is the drug effect. The raw number is not.

Incidence is not severity. These are counts of participants who reported an event at least once, not measures of how bad or how persistent it was. A 42.4% nausea rate does not mean 42.4% of people were meaningfully unwell, which is exactly why the discontinuation rates matter more.

Rates as reported in the Eli Lilly readout release for each trial. Percentages are the proportion of participants in each arm reporting the event. A dash means the arm did not exist in that trial or the event was not reported separately there.

Trial sizes: TRIUMPH-1 n=2,339 over 80 weeks; TRIUMPH-2 n=1,152 over 80 weeks; TRIUMPH-3 n=1,949 over 80 weeks; TRIUMPH-4 n=445 over 68 weeks. Population detail and efficacy results are in the Phase 3 results filing.

Nausea, diarrhoea, constipation and vomiting are not four unrelated findings. They are one mechanism showing up in four ways. Incretin receptor agonism slows gastric emptying and alters gut motility and satiety signalling, which is a large part of how the compound produces weight reduction in the first place. The gastrointestinal events are the same effect experienced as symptoms.

That is why they scale with dose so cleanly. In TRIUMPH-1, nausea moved 28.6% to 38.4% to 42.4% across the ladder and vomiting more than doubled from 10.6% to 25.3%. In TRIUMPH-2, nausea moved 13.7% to 20.8% to 28.0%. The signal is present at every step, in every trial.

It is also why titration matters as much as dose. These effects are strongest when receptor exposure changes, and they attenuate as exposure stabilises, which is why every trial in the programme escalated in steps rather than starting at target dose. The Phase 2 investigators had already observed that starting at 2mg rather than 4mg reduced the burden. In practical terms the gastrointestinal profile is a function of three things, not one: the dose reached, the speed of the climb toward it, and how long the participant has been stable at each step.

One quiet detail in the table is worth flagging. Diarrhoea in TRIUMPH-1 was slightly lower at 12mg, at 32.0%, than at 9mg, at 34.1%, and the same inversion appears in TRIUMPH-3 and TRIUMPH-4. Small non-monotonic wobbles like that are normal in adverse event reporting and should not be read as 12mg being gentler on the gut. The overall direction across the cluster is unambiguous.

Dysesthesia means an altered or unpleasant skin sensation: tingling, burning, or ordinary touch feeling wrong. It is the one event in the table that no retatrutide article written before December 2025 mentions, because it does not appear as a distinct adverse event in the 2023 Phase 2 publication. It appears in all four Phase 3 readouts.

The headline figures come from TRIUMPH-4, which reported 8.8% at 9mg and 20.9% at 12mg against 0.7% on placebo, and where Lilly described the events as generally mild and rarely leading to discontinuation. TRIUMPH-1, four times the size, reported a lower 5.1%, 12.3% and 12.5% against 0.9%. TRIUMPH-2 and TRIUMPH-3 came in lower still.

The mechanism has not been established. Candidate explanations exist and none has been demonstrated, which is the whole point of the dedicated dysesthesia filing, where the cross-trial figures, the sponsor characterisation and the hypotheses are set out properly.

A modest resting heart rate increase is a recognised effect across the incretin class rather than a retatrutide peculiarity. A published network meta-analysis of twelve randomised trials in people with overweight or obesity put the class effect at a mean increase of 3.47 beats per minute against placebo, and placed retatrutide at 3.46 beats per minute with a 95% confidence interval of 1.74 to 5.18. Tirzepatide came in at 2.05, semaglutide at 3.35 and liraglutide at 2.37 in the same analysis.

The Phase 2 trial reported a dose-dependent heart rate increase that peaked around week 24 and declined thereafter, which is the pattern generally described for this class. Lilly did not report heart rate in any of the four Phase 3 topline releases, so the Phase 3 picture is a genuine gap rather than a reassuring silence. TRIUMPH-3, in a population with established cardiovascular disease, is the readout where that omission is most conspicuous. The detail sits in the heart rate filing.

An event list tells you what people noticed. A discontinuation rate tells you what they could not live with. It counts participants who left a trial they had volunteered for, in a monitored setting, because of adverse events. It is the closest thing the dataset has to a tolerability verdict.

The TRIUMPH-1 row is the one to anchor on, because it is the pivotal trial and the largest. Read it against the efficacy ladder from the same trial: 19.0% at 4mg, 25.9% at 9mg, 28.3% at 12mg. The last dose step buys roughly two and a half percentage points of additional weight reduction, and takes the discontinuation rate from 6.9% to 11.3%. The step before it buys nearly seven percentage points and costs less than three points of discontinuation.

That asymmetry is the most important shape in the retatrutide dataset. Efficacy flattens as you climb. Tolerability does not.

Two other things in the table deserve honesty. TRIUMPH-2 is non-monotonic: 11.6% at 9mg against 7.7% at 12mg. In a 1,152-participant trial split across four arms, that is well within the range where random variation and arm size explain the inversion, and it should not be read as 12mg being better tolerated than 9mg in diabetes. And TRIUMPH-4 ran markedly higher than every other trial across the board, at 12.2% and 18.2%. It was also the smallest trial, and its own higher-BMI subgroup came in far lower at 8.8% and 12.1%, which shows how much a small denominator moves a percentage.

Topline sponsor releases are the primary source for everything above, and they are also the limit of it. Being explicit about the gaps is more useful than filling them.

No heart rate figures in Phase 3. None of the four releases reported heart rate, blood pressure or ECG findings.

No severity or duration breakdown. Events are reported as incidence, with a one-line characterisation of severity for dysesthesia only. Grade distribution, time to onset, duration and resolution rates are not in the topline record.

No serious adverse event detail. The releases give discontinuation rates but not a breakdown of serious events, deaths or events of special interest such as gallbladder, pancreatic or thyroid findings, which are standard concerns for this class.

No long-term data. The longest published exposure is the 104-week extension in a 532-participant subgroup. Cardiovascular and renal safety over years is what TRIUMPH-OUTCOMES, with roughly 10,000 participants, is built to answer, and it has not reported.

Peer-reviewed publication of the full TRIUMPH safety datasets, and any FDA review documents that follow the planned Q1 2027 filing, are where those gaps close.

Retatrutide is not approved anywhere and is not registered with FDA Philippines, so it exists in this country as research material rather than as a medicine. That makes the safety record something to read rather than something to act on, and it is the reason this page reports rates in trial populations and stops there.

The country picture sits in the retatrutide Philippines filing, the regulatory background in the FDA Philippines guide, and the wider class comparison in the GLP-1 side effects overview. Handling in a tropical climate is covered in the storage filing.

1. [01]TrialLilly · TRIUMPH-1 readout, 21 May 2026

   Gastrointestinal and dysesthesia rates by arm, and discontinuation at 4.1%, 6.9% and 11.3% against 4.9% placebo.

   investor.lilly.com · TRIUMPH-1 release
2. [02]TrialLilly · TRIUMPH-2 and TRIUMPH-3 readout, 23 July 2026

   Adverse event and discontinuation rates in the type 2 diabetes and cardiovascular disease populations.

   investor.lilly.com · TRIUMPH-2 and TRIUMPH-3 release
3. [03]TrialLilly · TRIUMPH-4 readout, 11 December 2025

   Adverse event rates including dysesthesia at 8.8% and 20.9%, and discontinuation at 12.2% and 18.2% against 4.0% placebo.

   investor.lilly.com · TRIUMPH-4 release
4. [04]Meta-analysisEur J Med Res 2026 · GLP-1 receptor agonists and heart rate

   Systematic review and network meta-analysis, twelve randomised trials, 15,313 participants. Class mean difference 3.47 bpm; retatrutide 3.46 bpm, 95% CI 1.74 to 5.18.

   link.springer.com · 10.1186/s40001-026-03933-9
5. [05]TrialNEJM 2023 · Retatrutide Phase 2 obesity trial

   Jastreboff AM et al. N Engl J Med 2023;389:514-526. The Phase 2 safety record, which does not report dysesthesia as a distinct event.

   https://www.nejm.org/doi/full/10.1056/NEJMoa2301972