Safety · 9 min read · Published September 8, 2026

Peptide Side Effects by Class: What Is Documented, What Is Anecdotal

A class-by-class table separating adverse events recorded in controlled trials and FDA labels (GLP-1 drugs, MK-677, elamipretide, bremelanotide) from the anecdotal reports that fill forums for BPC-157, TB-500, copper peptides and nootropic peptides.

What side effects do peptides have?

It depends entirely on the class. Approved peptides such as semaglutide, tirzepatide and bremelanotide have adverse-event rates from trials of 1,000 to 2,500 people: nausea in 25 to 44 percent, dysesthesia in up to 20.9 percent for retatrutide. Unapproved peptides such as BPC-157, TB-500 and GHK-Cu have no controlled human adverse-event data at all as of September 2026.

Regulatory status first, because it decides what is known. Semaglutide, tirzepatide, bremelanotide and elamipretide are FDA-approved drugs with labels that list adverse reactions by percentage. Retatrutide and MK-677 are investigational, with adverse events recorded in registered trials. BPC-157, TB-500, GHK-Cu (as an injectable), Semax, Selank, DSIP and Epitalon are not approved for any use in the United States; as covered in our FDA status tracker, several went through an advisory committee vote in July 2026, but none can be legally compounded today and none has a safety database. This article sorts every class into two columns: what a trial or label recorded, and what is only reported on forums.

How to read the table

A documented adverse event means a rate measured against placebo in a randomized trial or listed on an FDA label. An anecdotal report means a symptom that appears in user forums, clinic marketing or case reports without a denominator. Anecdote can invent problems that do not exist and miss ones that only appear when 2,000 people are followed for a year. Evidence tiers are labeled as human RCT, human observational, animal, in vitro or none.

GLP-1 class: what STEP 1, SURMOUNT-1 and TRIUMPH-1 recorded

In STEP 1 (2021, n=1,961, 68 weeks) nausea and diarrhea were the most common semaglutide side effects; 4.5 percent stopped for gastrointestinal events versus 0.8 percent on placebo. In SURMOUNT-1 (2022, n=2,539, 72 weeks) adverse events stopped tirzepatide in 4.3 to 7.1 percent versus 2.6 percent. TRIUMPH-1 (2026, n=2,339, 80 weeks) added 12.5 percent dysesthesia on 12 mg retatrutide.

The GLP-1 drugs are the only peptide class where the side-effect conversation can be had with real numbers. Gastrointestinal effects are dose-related, cluster during dose escalation and usually fade, which is why the trials report them as 'mild to moderate' even at rates above 40 percent. In TRIUMPH-1, the 4, 9 and 12 mg retatrutide arms recorded nausea in 28.6, 38.4 and 42.4 percent (placebo 14.8), vomiting in 10.6, 22.8 and 25.3 percent (placebo 4.8), and discontinuation for adverse events in 4.1, 6.9 and 11.3 percent (placebo 4.9).

Retatrutide also produced a side effect the other two drugs did not: dysesthesia, a skin tingling or heightened sensitivity. It was 5.1, 12.3 and 12.5 percent by dose in TRIUMPH-1 against 0.9 percent on placebo, and higher in the smaller knee-osteoarthritis trial TRIUMPH-4 (n=445, 68 weeks): 8.8 percent on 9 mg and 20.9 percent on 12 mg versus 0.7 percent. Lilly described the events as mostly mild and rarely a reason to stop. That is the 'about one in five' figure that circulates, and it belongs to the top dose of one trial, not to the drug as a whole. Our retatrutide versus tirzepatide comparison has the efficacy side of those trials.

Growth hormone secretagogues: the MK-677 glucose and edema data

The only long randomized trial of a GH secretagogue for healthy aging gave 65 adults aged 60 to 81 MK-677 25 mg daily or placebo for 2 years (Annals of Internal Medicine, 2008). Fasting glucose rose an average of about 5 mg/dL and insulin sensitivity fell; the most frequent complaints were increased appetite, transient mild lower-leg edema and muscle pain.

Those findings are what you would predict from raising growth hormone and IGF-1 back to young-adult levels: growth hormone opposes insulin, and fluid retention is a known effect of the axis. The trial called the drug 'generally well tolerated'; that phrase is quoted often, the glucose line less. For the injectable secretagogues (ipamorelin, CJC-1295, sermorelin, the GHRPs), there are pharmacokinetic and short hormone-response studies but nothing with a year of follow-up in healthy people, so the water retention and carpal-tunnel-like numbness described on forums cannot be given a rate. Mechanistically plausible, undocumented.

Healing peptides: BPC-157 and TB-500 have no human adverse-event data

As of September 2026 there is no controlled human safety data for BPC-157 or TB-500. A 2025 systematic review in HSS Journal found the human record consists of a single uncontrolled 12-patient series. The FDA's July 2026 briefing document reported no adequate clinical data and market products whose certificates of analysis did not match the label.

The animal literature for BPC-157 is large and reports no toxicity signal, and that is where the 'no known side effects' claim comes from. It is not the same as safety in people. The HSS Journal review put it plainly: adverse effects are possible from unregulated manufacturing, contamination and unknown clinical safety. The recurring theoretical concern is that a peptide chosen for its ability to grow new blood vessels and tissue could do the same for a tumor; there is no human evidence either way, and Harvard Health flags it as a reason for caution in anyone with or at risk for cancer. What users report (injection-site soreness, headache, nausea, fatigue) reads like a list of injection side effects in general.

Mitochondrial peptides: the Forzinity label

Elamipretide is the one mitochondrial peptide with a label, because it was approved in September 2025 for Barth syndrome (see our Forzinity explainer). The prescribing information is built on a 12-patient crossover trial, TAZPOWER, so the percentages are coarse: every patient had a local injection reaction on drug versus 67 percent on placebo, erythema 100 versus 25 percent, induration and pruritus 67 versus 17 percent, and there is a hypersensitivity warning covering rash, facial swelling and cough. Daily subcutaneous injection of a peptide produces skin reactions in most people; that is the documented reality behind the vaguer 'injection site issues' anecdotes for MOTS-c, which has no comparable data at all.

Copper peptides, nootropic peptides and melanocortins

Bremelanotide (Vyleesi) is the melanocortin reference: in the pooled Phase 3 label data (n=627 versus 620 placebo) nausea occurred in 40.0 percent, flushing 20.3 percent, injection site reactions 13.2 percent and headache 11.3 percent, and 18 percent discontinued. Injectable GHK-Cu, Semax and Selank have no structured adverse-event reporting; what exists is anecdote.

The Vyleesi label also records a transient blood pressure rise of about 6 mmHg systolic and 3 mmHg diastolic peaking 2 to 4 hours after a dose, and focal hyperpigmentation in 1 percent, more often in darker skin. Nausea was severe enough that 8 percent quit because of it. Melanotan II, the unapproved cousin, shares the receptor and the nausea, and adds a case-report literature of new and darkening moles that has never been quantified because nobody has run a trial.

For GHK-Cu the documented use is topical and cosmetic; skin irritation is reported but without a rate, and injectable use has no data. The nausea, hypotension and fatigue described after injection are anecdotal. Semax and Selank have Russian clinical studies that describe good tolerance without a denominator, and no Western pharmacovigilance record. Nasal irritation, headache and irritability are the common forum reports; treat them as plausible, not measured.

Why gray-market products add a second layer of risk

That distinction is the point of both the Harvard Health and American Medical Association pieces on injectable peptides published in 2026: the physicians quoted are not saying every peptide is dangerous, they are saying that for unapproved products the strength, purity, sterility and storage are unverified, that allergic reactions including anaphylaxis and contamination are the realistic acute risks, and that the long-term effects are unknown because the studies were never done. Our guide to reading a certificate of analysis explains what a COA can and cannot tell you.

The bottom line

A side-effect claim with a trial name, a sample size and a placebo rate is documented. One that comes with 'users report' is not. For GLP-1 drugs, MK-677, elamipretide and bremelanotide the documentation exists and is worth reading in full. For BPC-157, TB-500, copper peptides and nootropic peptides the accurate statement in September 2026 is that nobody has counted.