A weekly injection produces a weekly opportunity for the skin to object, and the advice that follows is always the same: rotate the site. That advice is reasonable and it is also borrowed. What the trials of these drugs actually measured is a reaction rate that differs by a factor of eighteen between two molecules in the same class, and a within-drug difference of eleven-fold that has nothing to do with where the needle went. The place to start is the same one that governs storage and handling: what was in the vial, and what the label recorded.
What the two labeled programs recorded
In the pooled semaglutide weight-reduction trials in adults, injection-site reactions — the term covers pruritus, erythema, inflammation, induration and irritation — were reported by 1.4% of treated patients and 1% of those on placebo.[1] A 0.4-point excess over a 68-week trial is close to nothing, and the symptom does not appear at all in the label’s main adverse-reaction table, which requires an event to reach 2% and to exceed placebo.
The tirzepatide program recorded something different. Against 2% on placebo, injection-site reactions were reported by 6% at 5 mg, 8% at 10 mg and 8% at 15 mg — a composite that includes injection-site bruising, erythema, pruritus, pain, rash and reaction.[2] That is a four-to-six point excess, enough to clear the table threshold, and it makes this one of the few adverse effects on which the two molecules genuinely diverge rather than differ in degree. The rest of that comparison is in the molecule article.
The head-to-head where the ranking reverses
SUSTAIN 3 randomized 813 patients with type 2 diabetes to once-weekly semaglutide 1.0 mg or exenatide extended-release 2.0 mg for 56 weeks. Injection-site reactions were reported by 22.0% of the exenatide group and 1.2% of the semaglutide group.[3]
In the same trial, gastrointestinal adverse events ran the other way: 41.8% on semaglutide against 33.3% on exenatide.[3] One randomized comparison, two tolerability questions, opposite winners. That is the strongest available argument against a single “better tolerated” ranking for this class, and it also explains why extended-release depot formulations have a reputation for nodules that the weekly analogs do not deserve.
The reaction tracks antibodies, not technique
The tirzepatide label reports the finding that reframes the whole symptom. Among treated patients in the pooled weight-reduction trials, injection-site reactions occurred in 11.3% of those who developed anti-tirzepatide antibodies and 1% of those who did not.[2] Anti-tirzepatide antibodies were detected in 64.5% (1,591 of 2,467) of treated patients in those studies.
An eleven-fold difference by immune response, inside one trial population injecting under one protocol, is a much larger signal than anything technique could plausibly produce. It means the person who reacts is mostly not the person who injected badly.
The obvious next step — comparing that 64.5% against semaglutide’s antibody rate — is the one the label forbids. Anti-semaglutide antibodies were detected in 3% (50 of 1,709) of patients on 2.4 mg across two 68-week studies, and in 11.2% (34 of 304) of patients on the same 2.4 mg dose across two later studies using a new assay, alongside 15.4% (202 of 1,311) on 7.2 mg.[1] The same dose of the same drug reads 3% on one instrument and 11.2% on another, and the label states plainly that assay differences preclude meaningful comparison across studies. A number that moves by a factor of nearly four when the assay changes cannot carry a cross-molecule conclusion.
What the postmarketing instrument adds, and what it cannot say
A 2026 review extracted adverse-event reports for tirzepatide, semaglutide, liraglutide, exenatide, dulaglutide and lixisenatide and combined injection-site events with skin events. Of 442,567 reports, 137,412 (31.0%) fell into that category, making it the third most frequently reported group. The proportion differed sharply by agent: exenatide 53.1%, dulaglutide 33.5%, tirzepatide 32.6%, liraglutide 17.1%, semaglutide 12.2% and lixisenatide 6.9%. Most reactions were mild to moderate and were managed without discontinuation.[4]
A separate analysis of 123,145 tirzepatide reports found injection-site reactions among the strongest signals in the file, alongside medication errors.[5] Neither result is an incidence. A spontaneous reporting database has no denominator, is shaped by how recently a drug was launched and how loudly it was marketed, and counts a report rather than a patient. What the two do establish is that the rank order by molecule holds across the randomized and the postmarketing instruments, which is more than most safety signals manage.
What was in the vial the trials tested
The single-dose presentations are short lists. Each milliliter of the semaglutide single-dose pen or prefilled syringe contains disodium phosphate dihydrate, sodium chloride and water for injection.[1] Each tirzepatide single-dose pen or vial contains sodium chloride, sodium phosphate dibasic heptahydrate and water for injection.[2] No preservative, because a container used once does not need one.
The multi-dose presentations are not short lists, and this is where the tidy story about compounded vials collapses. Each dose from the tirzepatide multi-dose vial or KwikPen contains benzyl alcohol 5.4 mg, glycerin 4.8 mg and phenol 1.08 mg alongside the salts.[2] The semaglutide multi-dose pen contains phenol 5.5 mg and propylene glycol 14 mg.[1] Benzyl alcohol was named Allergen of the Year for 2026 by the American Contact Dermatitis Society, on the grounds that it causes allergic contact dermatitis with prolonged or repeated exposure and is absent from many standard patch-test series, so its role is underrecognized.[6]
The dividing line, in other words, is multi-dose against single-dose, not compounded against brand. A preserved vial is a preserved vial whoever filled it, which is the same logic that governs what is in a compounded vial.
The blends this market sells were never in a trial
A 2026 survey identified 33 compounded semaglutide and tirzepatide products advertised as differing from the approved ones. Forty-eight percent combined the active drug with some mixture of cyanocobalamin, glycine, niacinamide, docusate or ondansetron. Only 18% published a beyond-use date and preferred storage conditions, and 28% published storage conditions at all. The authors concluded there is little justification for adding nutrients or docusate sodium to a subcutaneous product, and that while there is a rationale for adding ondansetron, evidence for giving it subcutaneously is lacking.[7]
Every incidence figure earlier on this page comes from a formulation whose complete excipient list is printed on a label, and neither list contains cyanocobalamin, glycine or niacinamide. The rates are therefore rates for a product that a B12-and-glycine blend is not. Compounded drugs are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before they are dispensed, which means the added ingredient has no trial behind it either alone or in combination.
The pharmacovigilance picture is consistent with that gap. Of 81,078 GLP-1 reports, 707 involved compounded products, and those carried higher reporting odds of contamination (19.00), preparation errors (48.92), compounding or manufacturing issues (8.51) and hospitalization (2.35).[8] Contamination and preparation are the two failure modes that present at the skin. The FDA has separately described a report of redness, site swelling, pain and a red lump at the injection site from a product labeled as compounded tirzepatide that came from a pharmacy which had not compounded it, recommends discarding a multi-dose vial 28 days after first use even when the compounder’s instructions allow longer, and instructs a new sterile needle and syringe for every dose.[9] How to check the pharmacy behind a vial is covered in the vetting article.
The rotation evidence, and where it comes from
Both labels instruct rotating the injection site with each dose, and both list the abdomen, thigh and upper arm.[1][2] The tirzepatide instructions for use give a reason: rotating within the chosen area reduces the risk of lipodystrophy — described on the label as pits or thickened skin — and of localized cutaneous amyloidosis, described as lumps.[2] What neither label cites is a trial, because there is no randomized rotation study in this drug class. The evidence exists, and it is about insulin.
A 2026 systematic meta-analysis of 51 studies of lipohypertrophy in insulin-treated diabetes found the strongest contributor to be incorrect site rotation, at a prevalence odds ratio of 8.85 (95% CI, 5.10 to 15.33), followed by needle reuse at 3.20(1.99 to 5.13), more than five years of therapy at 2.62 and more than two daily injections at 2.27. Sex, age and device type were not significant risk factors.[10]
Two things follow. The first is that both leading risk factors are behaviors rather than fixed characteristics. The second is that lipohypertrophy is a fatty thickening produced by repeated daily injection of insulin into the same tissue, and it is not the same lesion as a weekly erythematous reaction driven by antibody status. Borrowing the rotation finding across that gap is a reasonable precaution, not a result.
The one prospective test of acting on that advice is worth reading for its subtitle. A Belgian study followed 171 insulin-injecting patients given 4 mm needles and structured injection-technique education. At baseline 95.9% were reusing needles and 37.0% were rotating incorrectly; afterward needle reuse fell to 21.2% and incorrect rotation to 4.1%. Severe hypoglycemic episodes fell from 15.8% to 4.1% and unexplained hypoglycemia from 46.6% to 16.4%, both at P < 0.001 — and HbA1c did not move at all.[11] The instrument a clinic would have chosen to prove the intervention worked showed nothing; the outcome that mattered to patients moved a great deal.
What none of this decides
Injection-site reactions on semaglutide are, in trial terms, barely distinguishable from placebo. On tirzepatide they are real, they are mostly mild, and they concentrate in the majority of patients who develop anti-drug antibodies. Rotation and a fresh needle each dose cost nothing and have an evidence base from an adjacent field. Disposal of those needles is its own subject, in the sharps article.
A reaction that spreads beyond the site, produces hives, swelling of the face or lips, or difficulty breathing is not an injection-site reaction — both labels carry serious hypersensitivity warnings, including anaphylaxis and angioedema reported postmarketing, and that presentation is an emergency. Warmth, spreading redness, pus or fever days after an injection points at infection rather than irritation, and belongs with a prescriber the same day.