Nausea is the question people ask before they ask about the price, and the useful answer is a shape rather than a number. It tends to arrive with a dose change, build over a few days, and settle while the dose holds. The pivotal trials recorded that shape clearly enough to plan a calendar around, which is what this page is for.
The first sixteen weeks are the ones to plan for
Semaglutide labeling begins at 0.25 mg once weekly and steps up every four weeks until it reaches 2.4 mg, so the maintenance dose arrives at week sixteen. Tirzepatide begins at 2.5 mg and rises in 2.5 mg increments after at least four weeks on each rung. Every one of those steps is a fresh adjustment rather than a continuation of the last one.
That is the part worth internalizing. A rough first week at the starting dose does not predict a rough week three months later, and a quiet first month is not evidence that the fourth will be quiet. A 2026 review of tolerability across this drug class describes the gastrointestinal effects as emerging during dose escalation and resolving over time.[1] The pattern repeats at each rung; it does not run once and finish.
Which molecule you are climbing matters less than most marketing suggests, and the two schedules are laid out side by side in the molecule comparison. What the effects themselves were, and how often, is a separate question with its own page.
What STEP 4 measured almost by accident
STEP 4 was designed to test whether weight loss holds when the drug is withdrawn, and to do that it ran every participant through a twenty-week lead-in first: sixteen weeks of escalation plus four weeks at the maintenance dose. That design produced an unusually clean reading of the escalation phase on its own.
Of 902 people who started the lead-in, 803 — 89.0% — reached 2.4 mg and went on to be randomized.[2] Roughly one in nine did not get there. That figure is the closest thing the literature offers to a direct answer on how many people the climb itself stops.
The phase after the climb reads differently. Across the following forty-eight weeks at a steady dose, gastrointestinal events were still reported by 49.1% of those who continued semaglutide against 26.1% on placebo — so the effects did not disappear. Yet discontinuation for adverse events was 2.4% against 2.2%, which is the same number twice.[2]
Read those two together and the finding is specific: at a held dose the effects persist at a level people largely absorb, while the decisions to quit are concentrated earlier. Part of that is selection: the lead-in had already removed the 11% who could not climb, which the trial design states openly.
The longest exposure on record
SELECT followed 17,604 people with a mean drug exposure of 34.2 months, which is far longer than any obesity trial. Adverse events led to permanent discontinuation in 16.6% of the semaglutide group against 8.2% on placebo.[3]
The gap of roughly eight percentage points over nearly three years is the number to carry, not the raw 16.6%. Placebo groups stop too, for life reasons that have nothing to do with the drug, and a rate quoted without its comparator overstates the case badly.
SELECT also differs from the weight-loss trials in a way that cuts against easy comparison. Its participants were 45 or older with existing cardiovascular disease, and they stayed on drug far longer than anyone in STEP 1 or STEP 4. A higher discontinuation rate over a longer exposure in an older population is not the same finding as a higher rate of nausea.
What none of these trials publish is the shape most readers actually want: a day-by-day curve of how bad a given week is. Trials record whether an event occurred, how severe it was graded, and whether it led to stopping. They do not record how a Tuesday felt. Anyone promising you a precise week-three forecast is working from something other than this evidence.
What actually shortens a bad stretch
The lever with the most evidence behind it is the schedule itself: holding at the current dose rather than climbing on time. That is a prescriber’s call, and it is the single most useful thing to raise at a check-in, because the alternative most people reach for — quietly skipping a week — is a different intervention with a different result.
Holding a dose has a billing consequence that nobody mentions at intake. A seller that re-prices at every milligram step charges on the assumption that you climb on schedule, and a paused month can land awkwardly against that. Sellers that hold one price regardless of dose are collected on the flat-pricing board, and the wider price picture sits on the cost board.
When it is worth a call rather than a wait
Most of the above describes ordinary tolerability. A few things sit outside it: abdominal pain that is severe and persistent, vomiting that prevents keeping fluids down, or symptoms that start abruptly after months of stability. Those are reasons to contact a prescriber promptly rather than to wait out another week.
None of this is medical advice, and no number here predicts an individual. Two further caveats belong on the record. Every figure above comes from trials of branded product at labeled doses under supervision, while most sellers listed in the provider reviews supply compounded versions instead. And tolerability is what decides whether treatment continues at all, which has its own consequences — stopping is not a neutral event.