GLP-1 Diarrhea: Why It Doesn't Follow the Dose, and What Has Actually Been Tested

Diarrhea is one of the three most common side effects of these drugs: on Wegovy it was reported by 30% of people in the trials, against 16% on placebo.^[1] It does not behave like the other two. When the dose went up, vomiting and nausea went up with it; when two drugs were compared, vomiting differed. In the two largest recent comparisons, diarrhea did neither. And for all the advice in circulation about what to do about it, no trial has tested a single treatment. Here is what is measured, and what is not.

How common it is, by drug

Every label reports it, from its own placebo-controlled trials.^[1][2][4][3] The rows below come from different trials in different people — the Ozempic figures are from people with type 2 diabetes on lower doses — so compare each drug with its own placebo column, not the drugs with one another.

Few people stop over it. On Wegovy, diarrhea led to permanent discontinuation in 0.7% of people against 0.1% on placebo.^[1] On Zepbound, stomach side effects of every kind together caused 1.9% to 4.3% to stop, depending on the dose, against 0.5% on placebo.^[2]

It does not follow the dose the way nausea does

The clearest test came when Novo Nordisk tripled the semaglutide dose. In STEP UP, the results posted to the trial registry show diarrhea in 27.3% of people on 7.2 mg and 27.9% on 2.4 mg — no difference — against 12.9% on placebo. Over the same trial, vomiting rose from 16.4% to 24.8%.^[5] Stomach side effects as a whole were more common at the higher dose, 70.8% against 61.2%; diarrhea was not what moved.^[6] The Wegovy label reflects it: its list of reactions more common at 7.2 mg than at 2.4 mg includes nausea, vomiting and constipation, and does not include diarrhea.^[1]

SURMOUNT‑5, the trial that put tirzepatide head to head with semaglutide in people with obesity, found the same thing between drugs. Its registry results list exactly 88 cases of diarrhea in each arm — 88 of 374 on tirzepatide, 88 of 376 on semaglutide — while vomiting was lower on tirzepatide, 56 against 80.^[7] And in a small trial that randomized people with type 2 diabetes to three different tirzepatide escalation schedules, diarrhea landed at 31% to 36% on every schedule while nausea ranged from 24% to 39%, though that study was not powered to separate them.^[8] Ozempic’s own table shows the pattern at low doses: diarrhea 8.5% at 0.5 mg and 8.8% at 1 mg, while nausea went from 15.8% to 20.3%.^[3]

That has a practical consequence. The standard lever for side effects is the dose: the labels tell prescribers to follow the escalation schedule to reduce stomach side effects, and Wegovy’s says to consider delaying a step by four weeks if a dose is not tolerated.^[1][2] A small 2025 pilot trial in people with type 2 diabetes found that a slower, flexible semaglutide ramp cut withdrawals for stomach side effects from 19% to 2% and roughly halved the number of days with nausea.^[16] Its abstract does not report diarrhea. Whether holding or lowering the dose helps diarrhea specifically is something nobody has measured, and the dose data above give reason not to assume it.

How long it lasts

Individual bouts are short. Pooling the STEP 1 to 3 trials of semaglutide 2.4 mg, the median diarrhea episode lasted three days, about the same as on placebo. But short bouts can keep recurring: the share of people reporting diarrhea stayed higher than on placebo for the whole 68 weeks, peaking around week 20 and falling after that.^[9] For tirzepatide no published source gives a duration in days, so we do not give one; its label says most nausea, vomiting and diarrhea happened during dose escalation and eased over time.^[2] How the full set of side effects fades is covered in how long GLP-1 side effects last.

Why a drug that slows the stomach causes diarrhea: nobody knows

These drugs slow the emptying of the stomach,^[1] which makes diarrhea the side effect that seems least to fit. It is the harder one to explain, and the researchers who pooled the STEP data say so directly: changes in lower-gut transit “may be responsible for constipation and/or diarrhoea, although evidence is lacking.”^[9] The human studies point in different directions by segment. Liraglutide sped up transit through the large bowel by 31.7% in one placebo-controlled trial, in people with type 1 diabetes and nerve damage;^[10] exenatide slowed movement through the small intestine in another.^[11]

There is a genuine paradox on top of that. In people with bile acid diarrhea — a specific, chronic cause of diarrhea — a randomized trial found liraglutide reduced it better than the standard drug: 77% on liraglutide cut their stool frequency by a quarter or more, against 50% on colesevelam.^[12] The same class of drug that causes diarrhea in some people treated one form of it in others. Until the mechanism is understood, any explanation you read of why GLP‑1 diarrhea happens is a hypothesis.

What has been tested to treat it: nothing

We searched PubMed for any study of GLP‑1 drugs, diarrhea and loperamide or any other antidiarrheal. It returned five records and not one was a trial of treating the side effect. The most-cited expert consensus on managing these side effects suggests that loperamide or probiotics “could be considered” if diarrhea persists — but it grades none of its advice, and the source it cites for that line is a 2010 expert consensus on liraglutide, not a study.^[13]

The SURMOUNT trials show how it is handled in practice without showing whether it works. Investigators first counseled people with troublesome stomach symptoms on diet — smaller meals, splitting three meals into four or more, stopping when full — and could then prescribe an antidiarrheal at their own discretion. Between 5.2% and 9.3% of people on tirzepatide used one, against 1.2% to 2.2% on placebo. Nobody measured whether it helped.^[14] A 2026 systematic review of diet strategies alongside these drugs concluded that stomach symptoms occurred despite dietary guidance and that evidence on the best nutritional approach remains limited.^[15] That review is taken apart in the diet advice nobody has tested. One registered trial has stool texture and frequency as its main outcomes; it has posted no results.

What that leaves you with is the label’s own advice, which is modest and sound: drink fluids to avoid dehydration, and tell the prescriber if diarrhea “does not go away”.^[2] Over-the-counter antidiarrheals are widely used, and the absence of trials is not evidence they fail — only that nobody has checked. Ask a pharmacist or the prescriber before adding one, particularly if you take other medicines.

When diarrhea is not the ordinary side effect

The reason the labels care about fluids is the kidneys. Each warns of sudden kidney injury, sometimes serious enough to need dialysis, and says most reported cases were in people dehydrated by stomach side effects: nausea, vomiting or diarrhea.^[1] That is covered in GLP-1 drugs and your kidneys, and it matters more if you also take metformin, for reasons set out in metformin and a GLP-1 together.

• Severe stomach pain that will not go away, with or without vomiting, sometimes felt through to the back. The Wegovy Medication Guide says to stop the drug and call right away: it is the warning sign for pancreatitis.^[1]
• Clay-colored stools, pain in the upper stomach, fever, or jaundice, where the skin or the whites of the eyes turn yellow. These are the gallbladder warning signs the Zepbound Medication Guide lists — the one change in stool the labels name as a red flag.^[2] Both risks are covered in pancreatitis and gallbladder risk, and the everyday cramps these drugs cause are in GLP-1 stomach pain: what is expected and what is a warning sign.
• Not keeping fluids down, much less urine than usual, or dizziness on standing. That is the dehydration pattern, and it is a call to the prescriber, not a week to endure.

If you take the pill

Tirzepatide’s label tells people who rely on the contraceptive pill to use a method that is not a pill, or to add condoms or another barrier method, for the first four weeks and again for four weeks after every step up in dose. The reason it gives is not diarrhea but delayed stomach emptying, which is largest after the first dose.^[2] Foundayo gives the same reason for 30 days of backup.^[4] So the advice applies whether or not you have diarrhea, and none of these labels ties reduced absorption of oral medicines to diarrhea itself.