TL;DR: A 2026 medRxiv preprint used a validated LLM to pull self-reported symptoms out of Reddit discussion of unapproved retatrutide. Among 13,589 users reporting current use, 7,823 had at least one mapped symptom — and the resulting profile roughly inverts the trial one. Instead of GI events dominating, the most frequent reports were appetite increase, fatigue, increased energy, nausea, food craving, insomnia, and increased heart rate. Appetite increase on an incretin is the counterintuitive part. This data cannot give you incidence rates. What it can do is surface signals a controlled trial is structurally unable to see.
Clinical trials are excellent at one thing: measuring what happens to a screened population given a verified drug on a fixed schedule. They are correspondingly bad at capturing what happens when thousands of people obtain a grey-market peptide, titrate it themselves, stack it with other compounds, and then write about the experience in public. Those two datasets should not match. The interesting question is how they diverge — and the divergence here is stranger than anyone expected.
What the preprint actually did
The study — Self-Reported Side Effects Among Reddit Users Taking Unapproved Retatrutide, posted to medRxiv in June 2026 — is a cross-sectional analysis of Reddit posts and comments from retatrutide-specific and broader peptide/weight-management communities, running through December 2025. A validated large language model classified whether an author was reporting personal, current retatrutide use, then extracted author-attributed symptoms and mapped them to MedDRA preferred terms — the same standardized vocabulary used to code adverse events in trials, which is what makes the comparison possible at all.
The inversion
Here is the finding that makes the paper worth reading. Retatrutide’s phase 2 program — and everything the incretin class has taught us — says the dominant adverse events should be gastrointestinal: nausea, vomiting, diarrhea, constipation, all dose-related, all clustering around escalation. That is the profile our dosing and titration walkthrough describes and the one the GLP-1 side effects guide is built around.
The Reddit corpus does not look like that. The most frequently reported experiences were, in the authors’ framing: appetite increase, fatigue, increased energy, nausea, food craving, insomnia, and increased heart rate. Nausea is still on the list — but it is sharing the podium with its own opposite. Appetite increase and food craving, reported frequently, on a triple incretin agonist whose entire commercial premise is appetite suppression.
Why you cannot read incidence rates out of this
Every limitation below is real, and none of them makes the data worthless. They just constrain what question it can answer.
- Self-selected, self-reported, unblinded. Nobody in this dataset was randomized to anything. There is no placebo arm, so there is no way to know what fraction of “fatigue” is drug, deficit, or life.
- Posting is not sampling. People post when something is novel, alarming, or worth bragging about. Uneventful weeks generate no thread. Negative and surprising experiences are structurally over-represented; a quiet, textbook-tolerable run is under-represented by design.
- No dose verification, no purity verification. Grey-market material means the actual milligrams — and in some cases the actual molecule — are unknown. A “12 mg” report and a trial’s 12 mg arm are not the same exposure.
- Stacking. A large share of this population runs retatrutide alongside other compounds. Attribution to retatrutide alone is unreliable, and the authors say so.
- Reddit skews. English-speaking, US-heavy, and demographically nothing like a phase 3 enrollment roster.
Put together: this dataset cannot tell you how often anything happens. It can tell you what people are talking about, in a population that trials cannot recruit and would not enroll. Those are different currencies. Confusing them is how a legitimate signal turns into a bad headline.
The appetite signal: three hypotheses, clearly labelled as such
Nobody has demonstrated a mechanism for the appetite reports. But there are at least three plausible contributors, and it is worth separating them rather than reaching for the most dramatic one.
1. Glucagon is doing something different. Retatrutide is a triple agonist — GLP-1, GIP, and glucagon — and glucagon is the receptor that semaglutide and tirzepatide do not touch, as our retatrutide vs. tirzepatide comparison lays out. Glucagon agonism raises energy expenditure and drives hepatic fat mobilization; its metabolic footprint is genuinely distinct from the satiety-forward GLP-1/GIP arms. An agent that increases energy expenditure while appetite suppression is thin or fading is, at least in principle, an agent that could produce hunger. The increased energy reports sit naturally alongside this. It remains a hypothesis.
2. Self-titration overshoots and undershoots. Trials hold a rung for four weeks and step up on a schedule. Self-directed protocols on grey-market material do neither reliably. Doses that are too low for meaningful satiety, gaps between injections, abrupt jumps, and material of unverified concentration all produce a population where “how much drug is actually on board” varies enormously week to week. Rebound appetite after an effective period thinning out is a mundane explanation that fits a lot of the reports.
3. It is not only retatrutide. Stacked stimulants, other peptides, aggressive deficits, and the sleep disruption that shows up in the same symptom list all feed appetite and energy reporting. In an uncontrolled dataset, these are not confounders you can subtract out.
The increased heart rate reports are the least surprising item on the list: modest chronotropic effects are a known, documented feature of the incretin class across approved agents, and finding them echoed in self-report is consistent rather than novel.
How this sits against TRIUMPH
The contrast is cleanest when you put the two side by side. In TRIUMPH-1, the phase 3 obesity trial (topline figures covered in our TRIUMPH-1 results overview), the adverse-event story was the familiar one: GI events led, they were dose-related and mostly mild-to-moderate, and they were the main driver of a discontinuation curve that climbed with dose.
Eli Lilly, TRIUMPH-1 phase 3 topline results, 2026. Percentage of participants discontinuing for any adverse event, by assigned dose. Trial-grade figures — not comparable to self-report data.
Those numbers are trial-grade: verified drug, verified dose, a denominator you can trust. The Reddit corpus has none of that, which is precisely why the two should never be stacked on the same axis. The preprint’s contribution is not a competing incidence estimate — it is a different list of what shows up, generated from a population the trial excluded.
One thing the two datasets do agree on, in their different registers: retatrutide’s neurological/skin-sensation signal is real and unusual. Dysesthesia was flagged at 9 mg and 12 mg in both TRIUMPH-1 and TRIUMPH-4, and it is also the thing community discussion is most preoccupied with — the “death itch” and tingling threads that dominate forum chatter far out of proportion to how often the trials logged it. We have covered that in depth in retatrutide dysesthesia: the tingling, explained, so we will not re-litigate the mechanism here. The point for this article is narrower: intense community attention to a symptom is evidence of salience, not of frequency. A memorable, weird, hard-to-ignore sensation generates threads. Constipation does not.
The de-escalation pattern
A recurring practical observation in this corpus, and one that would be invisible to a trial: a meaningful number of researchers move back down to tirzepatide after a retatrutide run. Trials do not have a “switch to the previous drug” arm. Grey-market use effectively does, and it is one of the few genuinely useful things this kind of data can surface — a revealed preference rather than a rating scale.
The stated reasons vary and are, again, self-reported: tolerability, the sensation-related effects, cost, supply, or simply that the appetite suppression felt more predictable on a two-receptor agonist. Our semaglutide vs. tirzepatide vs. retatrutide comparison covers where each sits on the efficacy/tolerability curve, and the retatrutide primer explains what the third receptor buys you in exchange.
What this data is actually for
The honest summary: a preprint mining self-report from an unblinded, self-selected, unverified-exposure population cannot tell you how likely you are to experience anything. It is not evidence that retatrutide’s trial profile is wrong. Trials remain the instrument for incidence.
What it is good for is generating hypotheses that trials would never think to test — because a trial with fixed titration, screened participants and pharmaceutical-grade drug has no mechanism for noticing that people running the compound their own way report hunger, energy and insomnia more than they report vomiting. That observation is now on the record, mapped to a standard vocabulary, at a scale big enough to be worth someone’s time. Whether the appetite signal survives contact with a controlled study is exactly the right open question, and exactly the sort of thing a preprint like this exists to raise.
Frequently asked questions
Does retatrutide increase appetite?
In the phase 2 and phase 3 trials, no — appetite suppression is a core, measured effect. In the 2026 Reddit preprint, appetite increase was among the most frequently reported experiences from self-directed users. Those are not contradictory findings so much as findings from incompatible datasets: the trial measures a verified drug at a verified dose in a screened population; the preprint captures what an unblinded, self-selecting, stacking, grey-market population chose to write about. The appetite reports are a signal worth investigating, not an established drug effect.
Is the Reddit retatrutide study peer-reviewed?
No. It is a preprint posted to medRxiv in June 2026 and has not been through peer review. The authors describe the findings as hypothesis-generating and relevant to pharmacovigilance, which is the right frame — it flags a pattern for further study rather than establishing one.
Why don’t the Reddit symptoms match the TRIUMPH trial results?
Several structural reasons at once. Reddit posters self-select and skew toward novel or bothersome experiences; there is no placebo control to subtract background noise; dose and even compound identity are unverified on grey-market material; many users stack other compounds; and rapid self-titration produces exposure patterns no trial protocol would allow. Any one of those would shift the reported profile. Together, they make a divergence from trial data the expected outcome, not an anomaly.
Can this data tell me how common a retatrutide side effect is?
No, and the paper does not claim it can. There is no reliable denominator — you cannot know how many people used the compound without posting, or posted without describing symptoms. The dataset supports statements about what is frequently discussed, not about what is frequently experienced. For incidence, the TRIUMPH trial data remains the only trustworthy source.
Why is glucagon the receptor everyone points at?
Because it is the one retatrutide has and its closest comparators do not. Semaglutide targets GLP-1; tirzepatide adds GIP; retatrutide adds glucagon on top of both. Glucagon agonism raises energy expenditure and mobilizes hepatic fat — a metabolic profile meaningfully different from pure satiety signaling. That makes it the natural suspect whenever retatrutide reports something the other two do not, from dysesthesia to the energy and appetite signals here. Natural suspect is not the same as demonstrated cause.
References
- Self-Reported Side Effects Among Reddit Users Taking Unapproved Retatrutide. medRxiv preprint, 2026. doi:10.64898/2026.05.28.26352819 (not peer-reviewed)
- Jastreboff AM, Kaplan LM, Frías JP, et al. Triple–Hormone-Receptor Agonist Retatrutide for Obesity — A Phase 2 Trial. New England Journal of Medicine. 2023;389(6):514-526.
- Eli Lilly and Company. Lilly’s triple agonist retatrutide delivered powerful weight loss in the pivotal phase 3 TRIUMPH-1 obesity trial. Company press release, 2026.
- Coskun T, Urva S, Roell WC, et al. LY3437943, a novel GLP-1, GIP, and glucagon receptor triagonist. Cell Metabolism. 2022;34(9):1234-1247.
- Rosenstock J, Frías J, Jastreboff AM, et al. Retatrutide, a GIP, GLP-1 and glucagon receptor agonist, for people with type 2 diabetes: a phase 2 trial. The Lancet. 2023;402(10401):529-544.
Research use only. This article summarizes published and preprint research for educational purposes. It is not medical advice and does not recommend any dose, protocol, or human use.