GLP-1 Side Effects Week by Week: When Nausea Peaks and When It Resolves
Gastrointestinal effects cluster around dose increases rather than spreading evenly across treatment. In STEP-1, nausea affected roughly 44% of participants against about 18% on placebo, and about 7% discontinued for an adverse event against about 3%. The titration schedule exists to manage this: doses step up at intervals, and most reports concentrate in the days after each step.
What the pivotal trials reported
| Effect | Semaglutide 2.4 mg (STEP-1) | Tirzepatide (SURMOUNT-1) | Placebo (STEP-1) |
|---|---|---|---|
| Nausea | ~44% | ~29% | ~18% |
| Diarrhoea | ~30% | ~23% | ~16% |
| Vomiting | ~24% | ~13% | ~6% |
| Constipation | ~24% | ~17% | ~11% |
| Discontinued for an adverse event | ~7% | ~4–7% | ~3% |
Separate trials with different populations and durations. Cross-trial comparison of tolerability is indicative, not a head-to-head result.
Why the timing follows the dose, not the calendar
The approved schedule steps the dose up at intervals rather than starting at a maintenance dose. Each step is a new exposure, and reports cluster in the days after it. That is why a person can be four months in, feel settled, step up, and have a difficult week — nothing has gone wrong, and it is the pattern the schedule anticipates.
Percentages are indicative of the reported pattern across the STEP programme rather than tabulated values from a single trial. Direction, not precision.
What the label says about slowing down
Approved labelling permits staying longer at a tolerated dose before stepping up. That is a prescriber's decision, not a self-directed one, and it is the documented route when a step is poorly tolerated. The relevant point for a buyer is commercial as much as clinical: if you may need to spend longer at a lower dose, a prepaid multi-month plan commits you through exactly that window.
| Pricing structure | Effect of staying longer at a lower dose |
|---|---|
| Flat at every covered dose | No cost change — the rate does not follow the dose |
| Dose-tiered | Lower cost while at a lower dose, rising as you step up |
| Prepaid multi-month | Already paid — check what is refundable before committing |
| Membership plus medication | Membership continues regardless of dose |
About 7% of STEP-1 participants discontinued for an adverse event, most during titration.
What is not established
- That slower titration improves final weight loss. It manages tolerability; it has not been shown to change the endpoint.
- Which individuals will struggle. Genetic associations with nausea have been reported at population level, but no validated test predicts an individual.
- That one molecule is better tolerated. The figures above come from different trials. Only a head-to-head could settle it, and tolerability was not its primary endpoint.
The effects that are not gastrointestinal
Nausea dominates the conversation because it is the most common. The labelling covers several others that appear less often and matter more.
| Risk | What the labelling says | What to do |
|---|---|---|
| Pancreatitis | A recognised risk | Severe persistent abdominal pain — seek care promptly |
| Gallbladder disease | Reported, associated with rapid weight loss | Right upper abdominal pain, fever, jaundice |
| Thyroid C-cell tumours | Boxed warning, based on rodent studies | Contraindicated with MTC or MEN 2 history |
| Diabetic retinopathy | Complications reported in trials | Relevant if you have existing retinopathy |
| Hypoglycaemia | Mainly with insulin or sulfonylureas | Requires review of other medicines |
| Aspiration under anaesthesia | Delayed gastric emptying | Tell your anaesthetist you take a GLP-1 |
This is a summary of labelled risks, not a complete safety profile. Read the medication guide.
Muscle mass, and what the imaging sub-studies actually found
A substantial share of weight lost on any effective weight-loss intervention is lean mass, and body composition sub-studies of the GLP-1 trials have reported the same pattern. What those studies cannot establish is the functional consequence — strength, mobility, fall risk over years — because they were imaging sub-studies, not trials powered for that endpoint.
Protein intake and resistance training are the standard recommendations, and they are sensible. Neither has been tested in a randomised trial against a control in this population, so anyone presenting a protocol as evidence-based is describing a reasonable extrapolation rather than a result.
What reduces the odds of a bad week
- Do not skip the titration steps. The schedule exists because effects cluster at increases.
- Smaller meals, stopped earlier. Slowed gastric emptying means the usual portion arrives on a fuller stomach.
- Hydration. Vomiting and diarrhoea are the routes to the dehydration that turns a bad week into a hospital visit.
- Raise a poorly tolerated step with your prescriber rather than stopping abruptly. Labelling permits staying longer at a tolerated dose.
What is not known
Long-term safety beyond the trial windows is, by definition, outside what has been observed. The pivotal trials ran 68 to 72 weeks. Post-marketing surveillance continues, and it is the mechanism by which rare effects are found — which means the absence of a signal today is not the same as its absence.
What compounding adds to the side-effect question
Every figure on this page comes from trials of approved products. A compounded preparation introduces variables the trials did not test, and they are worth naming plainly.
| Variable | Why it matters |
|---|---|
| Salt form | The FDA has said sodium and acetate forms are not the same active ingredient |
| Concentration in mg/mL | Determines the volume drawn; a different concentration means a different draw |
| Added ingredients | B12 and other additives are common and were not in any trial |
| Dose measurement | A vial and syringe require the patient to measure; a pen does not |
| Beyond-use date | Shorter than a manufactured product's shelf life |
None of these makes a compounded preparation unsafe. All of them are unknowns a trial figure cannot cover.
The measurement error nobody warns about
An approved pen delivers a fixed dose. A vial requires the patient to draw the dose themselves, and dosing errors are a recognised risk of that format — including tenfold errors where units and millilitres are confused. If you are drawing from a vial, confirm with the prescriber exactly how many units correspond to your dose, in writing, and confirm again after any concentration change.
What is not established
- That one molecule is genuinely better tolerated. The rates on this page come from separate trials with different populations.
- Who will struggle. Population-level genetic associations with nausea have been reported; no validated test predicts an individual.
- Long-term effects beyond the trial windows. The pivotal trials ran 68 to 72 weeks. Post-marketing surveillance is how rarer effects are found, which means today's absence of a signal is not its absence.
- Anything about compounded preparations specifically. No trial has studied one.