GLP-1 Side Effects by Week: When They Start, Peak and Fade

The most reliable statement about when GLP-1 side effects happen is not a forum thread or an average of anecdotes. It is the sentence each manufacturer agreed with the FDA and printed in the prescribing information, and the four weekly injectables say nearly the same thing.

The Ozempic label (Novo Nordisk, semaglutide, revised May 2026 on DailyMed) reports that gastrointestinal adverse reactions occurred in 15.3% of placebo patients and in roughly a third of Ozempic-treated patients in the pooled placebo-controlled trials, then states: "The majority of reports of nausea, vomiting, and/or diarrhea occurred during dose escalation." The Wegovy label (Novo Nordisk, semaglutide, revised June 2026) reports gastrointestinal adverse reactions in 73% of Wegovy injection-treated adults versus 47% on placebo, with nausea at 44% versus 16%, diarrhea at 30% versus 16% and vomiting at 24% versus 6%, and then says: "These reactions were most frequently reported during dosage escalation."

The two tirzepatide labels go one clause further. The Zepbound label (Eli Lilly, revised August 2026) reports gastrointestinal adverse reactions in 56% of Zepbound-treated patients versus 30% on placebo and states: "The majority of nausea, vomiting, and/or diarrhea events occurred during dose escalation and decreased over time." The Mounjaro label (Eli Lilly, revised August 2026) gives 20.4% for placebo and between 37.1% and 43.6% for Mounjaro depending on dose, with the same sentence: "The majority of reports of nausea, vomiting, and/or diarrhea occurred during dose escalation and decreased over time."

The European labels agree. The EMA summary of product characteristics for Wegovy says the most frequently reported adverse reactions "were mainly seen in the dose-escalation period," and its patient leaflet puts it in plain language: these effects "are mainly seen during dose escalation and usually go away over time." So the timing answer is settled at the regulator level: the loud weeks are escalation weeks. What the labels do not do is promise a particular week for any one person, and neither does this page. For the list of effects themselves, including the rare serious ones, see the overview of GLP-1 side effects.

"Week 3" means very little on its own, because every one of these labels describes a starting dose that is held for a fixed period and then stepped up, with each subsequent step also held for a period, before a maintenance dose is reached. The Ozempic label holds its starting dose for four weeks and describes that dose as being for treatment initiation; the Wegovy label lays out an escalation table in four-week steps; the Zepbound label holds its starting dose for four weeks, says that dose is not approved as a maintenance dose, and increases only after at least four weeks on the current dose. The general shape of a GLP-1 dosing schedule is the same across all four: a staircase, with landings.

That structure is why the labels talk about escalation rather than about calendar weeks. Each step up is a fresh exposure to a higher level of the drug, and the labels' own explanation for the staircase is tolerability: the Ozempic and Zepbound labels both instruct prescribers to follow the escalation "to reduce the risk of gastrointestinal adverse reactions." The Wegovy label goes further and tells the prescriber what to do when a step is not tolerated: "If patients do not tolerate a dose during dosage escalation, consider delaying dosage escalation for 4 weeks." The EMA text for Wegovy says the same in its own words, that delaying escalation or returning to the previous dose "could be considered" in case of significant gastrointestinal symptoms.

Read those sentences carefully, because they are addressed to the prescriber. They are not an invitation to a reader to change a schedule, and this page does not describe one. Their value to a reader is different: they mean the person setting the schedule already expects the first weeks after a step to be the rough ones, and already has a documented option when they are rougher than expected. A dated log of exactly which week a symptom landed in is what turns that option into a decision made on evidence instead of on memory. The first 12 weeks guide covers the whole experience of those weeks; this page stays on the side effects.

The labels say "during escalation." The published trial analyses put numbers on it, and two are worth naming because they are the sources behind most of what gets repeated elsewhere.

For semaglutide, Wharton and colleagues pooled the STEP 1, 2 and 3 trials (2,117 participants on once-weekly semaglutide for weight management, 1,262 on placebo, 68 weeks) in Diabetes, Obesity and Metabolism in 2022. Their finding on timing is direct: "GI AEs were reported more frequently within the dose-escalation period, with the cumulative incidence of participants experiencing their first GI AE plateauing after week 20," and that the prevalence of nausea, diarrhoea and vomiting in the semaglutide arm "peaked at about week 20 and decreased thereafter." The STEP escalation ran 16 weeks, so the peak sits just after the final step, which is exactly what the label's language predicts. On severity, 98.1% of gastrointestinal events were mild or moderate, and 4.3% of semaglutide participants stopped treatment because of one, against 0.7% on placebo. On duration, the median nausea episode lasted 8 days, diarrhea 3 days, vomiting 2 days, and constipation 47 days. The EMA summary of product characteristics for Wegovy prints the same four medians from the STEP 1 to 4 trials, which is a useful cross-check that these are the numbers regulators accepted.

For tirzepatide, SURMOUNT-1 (Jastreboff and colleagues, New England Journal of Medicine, 2022) ran 72 weeks with a 20-week dose-escalation period, and its summary states that gastrointestinal events were the most common adverse events, "most were mild to moderate in severity, occurring primarily during dose escalation." Rubino and colleagues then pooled SURMOUNT-1 through 4 in Diabetes, Obesity and Metabolism in 2025 and wrote that in clinical trials tirzepatide's GI AEs "were typically mild-to-moderate in severity, transient and occurred during dose escalation," adding that the first use of a medication to manage a gastrointestinal symptom was most frequent during the first 24 weeks, which included escalation. Neither tirzepatide paper publishes a week-by-week prevalence curve or median episode durations, so this page does not invent them; the semaglutide figures are semaglutide figures.

Two cautions belong here. These are trial populations on trial schedules, and a prescriber may be running a slower staircase. And a median is a midpoint: half of nausea episodes lasted longer than 8 days, and constipation's 47-day median is why it deserves its own paragraph below.

Because the sources report by escalation phase rather than by calendar week, the bands below are anchored to the four-week steps every label describes. If a prescriber has set a different pace, shift the bands to match the actual dates in the dose log.

Weeks 1 to 2 are the first exposure to the starting dose, which the Ozempic and Zepbound labels describe as an initiation dose rather than a maintenance dose. Every label's wording places the bulk of nausea, vomiting and diarrhea reports inside escalation, and this is the first escalation window. The semaglutide medians suggest what an episode looks like if one arrives: nausea around a week, vomiting or diarrhea a few days.

Weeks 3 to 4 are the back half of the first four-week step. On the semaglutide medians, a nausea episode that began in week one has, at the midpoint, ended by now, so weeks 3 to 4 are the first place a log can show a direction. It is also the window in which, on the label's schedule, the prescriber is deciding whether the first step-up happens on time.

Weeks 5 to 8 typically contain the first step-up, and the labels' language means a repeat is expected: a new dose, a new short run of the same effects. The useful comparison in the log is not week 5 against week 4 but week 5 against week 1, because that is the same event at a higher step. A run that is shorter and milder than the first one is the label's "decreased over time" showing up in one person's data.

Weeks 9 to 12 usually hold the second step-up on a four-week staircase. By now there are two escalation events in the log to compare, which is enough to see a shape. Constipation, if present, is on a different clock entirely: its 47-day median in the STEP data means a case that began in week 3 may still be running here without being unusual.

Beyond week 12, the semaglutide trials place the peak of nausea prevalence at about week 20, right after the STEP staircase finished at week 16, with the prevalence falling after that; the tirzepatide trials describe a 20-week escalation with events concentrated inside it. So the honest answer to "when does it settle" is: after the last step, not after the first, and for constipation, later still. Per-drug detail lives on the Ozempic, Wegovy, Mounjaro and Zepbound side-effect pages.

"Fading" is a claim about direction, and direction needs at least two points on the same scale. Memory supplies neither; a log supplies both. Three entries per symptom are enough: the date, how strong it felt on the same simple scale every time, and a short note, all kept next to the date of each dose. The how-to for keeping a symptom journal walks through the setup; what follows is how to read it.

The first pattern is the run length after a dose. Count the days with any entry for a symptom in the seven days after each injection. A run of four days after the first dose, three after the second and one after the fourth is the label's "decreased over time" in miniature. The second pattern is the peak. If the strongest entry in each run is lower than the strongest entry in the previous run, the effect is fading even if it is still present. The third pattern is the escalation signature: a run that jumps back up in the week after a step, then shortens again over the following weeks. That shape is what the Ozempic, Wegovy, Mounjaro and Zepbound labels describe, and seeing it in one's own data is usually more reassuring than any average.

The pattern that is not fading is just as easy to read: runs that stay the same length or grow while the dose is being held, a peak that climbs step after step, or a symptom that never clears between doses at all. That log is not a failure; it is the evidence the Wegovy label's "consider delaying dosage escalation" sentence was written to act on, and it belongs in front of the prescriber before the next step, with a summary prepared for the appointment.

This is the practical layer where GLP-1 Tracker sits. It is a free iPhone app that logs each dose with its date and each symptom with its timing, how strong it felt and a note, and keeps the two timelines on one screen so the run after each dose is visible without arithmetic. It also records dose changes as dated events, so the escalation signature has a marker to line up against. It keeps the one-to-five rating you enter but does not compute a severity score, decide what a pattern means, or suggest a schedule; it makes the shape visible so the person who sets the schedule can read it.

Timing is reassuring in one direction only. "Most reports occurred during dose escalation" is a statement about nausea, vomiting and diarrhea, the common effects that the labels expect to be mild to moderate and transient. It says nothing about the serious reactions the same labels warn about, and severity does not wait for a convenient week. Persistent vomiting, severe abdominal pain, or any symptom that is getting worse rather than better is a reason to contact the prescriber or seek care now, whatever the week number. The general nausea page covers everyday comfort measures people commonly share; it does not replace that call either.

Timing also cannot say what the next step will feel like. The trial curves are averages over thousands of participants, and the labels' "decreased over time" describes the group, not a guarantee for any individual. Some people report almost nothing at any step; some find the last step harder than the first; some tolerate the injection well and find constipation the effect that outlasts everything else. A log that shows a person's own runs shortening is better evidence about that person than any figure on this page.

The pharmacist is the underused resource here. Pharmacists see the escalation schedule, the label text and the person's other medications together, and a question like "this is week six and the second step is due; here is what the log shows" is exactly the kind they are trained to answer. Bring the log, bring the schedule, and let the two people who can change anything decide whether anything should change.