A summary of published research, not medical advice. See the Disclaimer.
Studies: every indexed trial and review on this topic, scored
Semaglutide is a GLP-1 receptor agonist, a drug that mimics a gut hormone involved in appetite and insulin release. It is given as a weekly injection or a daily tablet. It has large, long placebo-controlled trials of hard outcomes, which few drugs in this wiki have.
Best-supported effect: Weight loss in adults with obesity. Helps a lot (strong evidence).
Most-cited weak claim: Slowing Alzheimer's disease. Makes little or no difference (strong evidence).
How we grade evidence · Last reviewed: October 2026
What you need to know
In adults with obesity or overweight and without diabetes, weekly semaglutide injections led to large weight loss, with strong evidence: average weight change was -14.9% against -2.4% on placebo in one large trial. It also reduced heart attacks, strokes and cardiovascular deaths in adults with heart disease and overweight, and in adults with type 2 diabetes, with strong evidence. People switched to placebo regained weight, and an oral 14 mg tablet made little or no difference to early Alzheimer's disease. Nausea was reported in 44% against 16% on placebo, and severe gastrointestinal reactions, gallstones and stopping because of side effects were all more common than on placebo. The label lists a personal or family history of medullary thyroid cancer, or MEN 2, as a contraindication. Weight trials used 2.4 mg once weekly, reached by stepping the dose up over several weeks, for 68 weeks.
| Effect | Verdict | What was measured | Studied in |
|---|---|---|---|
| Weight loss | Helps a lot (strong evidence) Large manufacturer-funded trial, participants did not know what they got, weight measured directly |
Measured | Adults with obesity or overweight, without diabetes |
| Fewer heart attacks, strokes and cardiovascular deaths | Helps (strong evidence) Two large trials measuring heart and blood vessel events, with and without diabetes |
Hard outcome | Adults with heart disease and overweight, without diabetes; adults with type 2 diabetes |
| Slower kidney decline | Probably helps (moderate evidence) One trial, stopped early after a check of results partway through |
Hard outcome | Adults with type 2 diabetes and chronic kidney disease |
| Heart failure symptoms | Probably helps (moderate evidence) One modest trial relying on symptom scores people gave themselves |
Self-reported | Adults with obesity and heart failure with preserved ejection fraction |
Weight loss: In STEP 1, 1,961 adults without diabetes received weekly 2.4 mg or placebo with lifestyle support for 68 weeks. Average weight change was -14.9% versus -2.4% [1]. Helps a lot (strong evidence): a large blinded trial that measured weight directly. The manufacturer funded the trial. In STEP 4, people switched to placebo after 20 weeks regained 6.9% of body weight over the next 48 weeks, while those who continued lost a further 7.9% [2].
Cardiovascular events: SELECT randomized 17,604 adults with heart disease and a BMI of 27 or more, without diabetes. Over about 40 months, the combined rate of cardiovascular death, heart attack or stroke was 6.5% versus 8.0% (hazard ratio 0.80) [3]. In type 2 diabetes, SUSTAIN-6 (3,297 people) found a similar reduction (hazard ratio 0.74) [4]. Helps (strong evidence): two large trials of the outcome itself.
Kidney disease: FLOW randomized 3,533 people with type 2 diabetes and chronic kidney disease to 1.0 mg weekly or placebo. Major kidney events were 24% lower, and cardiovascular deaths were also lower [5]. Probably helps (moderate evidence): downgraded because it was one trial, stopped early after an interim analysis.
Heart failure: In STEP-HFpEF, 529 people with obesity and heart failure with preserved ejection fraction improved 7.8 points more on a symptom score and walked 20 m further in 6 minutes than the placebo group after 52 weeks [6]. Probably helps (moderate evidence): downgraded because it was one trial of modest size, and symptom scores are subjective.
| Effect | Verdict | What was measured | Studied in |
|---|---|---|---|
| Slowing Alzheimer's disease | Makes little or no difference (strong evidence) Two oral tablet trials, decline same as placebo |
Hard outcome | Adults with early Alzheimer's, oral 14 mg |
| Lasting weight loss after stopping | Probably makes little or no difference (moderate evidence) One trial, people switched to placebo regained weight |
Measured | Adults with obesity |
Alzheimer's: The evoke and evoke+ trials randomized 3,808 people with early Alzheimer's disease to oral semaglutide or placebo for 104 weeks. Decline on the main clinical scale was the same in both groups, and the trials were discontinued [7]. Makes little or no difference (strong evidence).
Weight after stopping: In STEP 4, people switched from semaglutide to placebo regained weight [2]. Probably makes little or no difference (moderate evidence): one trial.
| Claim | Verdict | What was measured | Studied in |
|---|---|---|---|
| Less drinking in alcohol use disorder | May help a little (weak evidence) One small 9-week trial, total drinking days unchanged |
Measured | 48 adults, 9-week trial |
Alcohol use disorder: In a trial of 48 adults, low-dose semaglutide reduced alcohol consumed in a laboratory drinking session, drinks per drinking day and craving. Total drinking days did not change [8]. May help a little (weak evidence): one small, short trial.
Shown in humans
Shown only in animals or cells
Theory only
| Side effect | Verdict | How often vs placebo | Source | Studied in |
|---|---|---|---|---|
| Nausea | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 44% vs 16% on placebo |
44% vs 16% | [1][15] | Adults with overweight or obesity |
| Diarrhea | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 30% vs 16% on placebo |
30% vs 16% | [1][15] | Adults with overweight or obesity |
| Vomiting | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 24% vs 6% on placebo |
24% vs 6% | [15] | |
| Constipation | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 24% vs 11% on placebo |
24% vs 11% | [15] | |
| Abdominal pain | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 20% vs 10% on placebo |
20% vs 10% | [15] | |
| Indigestion | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 9% vs 3% on placebo |
9% vs 3% | [15] | |
| Headache | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 14% vs 10% on placebo |
14% vs 10% | [15] | |
| Fatigue | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 11% vs 5% on placebo |
11% vs 5% | [15] | |
| Dizziness | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 8% vs 4% on placebo |
8% vs 4% | [15] | |
| Hair loss | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 3% vs 1% on placebo |
3% vs 1% | [15] | |
| Abnormal skin sensations (tingling, burning) | Probably happens (moderate evidence) People reported this themselves, so graded no higher than moderate, 2% vs 1% at 2.4 mg |
2% vs 1% at 2.4 mg; 22% vs 0% at 7.2 mg | [15] | |
| Severe gastrointestinal reactions | Happens (strong evidence) Events counted across combined trials, 4.1% vs 0.9% on placebo |
4.1% vs 0.9% | [15] | |
| Stopping because of side effects | Happens (strong evidence) Dropouts counted in combined trials, plus the STEP 1 and SELECT trials |
6.8% vs 3.2% (pooled weight trials); 4.5% vs 0.8% for gastrointestinal reasons in STEP 1; 16.6% vs 8.2% in SELECT | [1][3][15] | Adults with overweight or obesity |
| Gallstones and gallbladder inflammation | Happens (strong evidence) Events counted across combined trials, gallstones 1.6% vs 0.7% |
Gallstones 1.6% vs 0.7%; cholecystitis 0.6% vs 0.2% | [15] | |
| Low blood glucose in type 2 diabetes | Happens (strong evidence) Glucose measured in diabetes trials, 6% vs 2% without insulin |
6% vs 2% without insulin; higher with insulin or sulfonylureas | [15] | |
| Faster resting heart rate | Happens (strong evidence) Measured in trials, average rise of 1 to 4 beats per minute |
Measured in trials, mean rise of 1 to 4 beats per minute; peak rise of 20 or more in 26% vs 16% | [15] | |
| Diabetic retinopathy complications | Probably happens (moderate evidence) SUSTAIN-6 trial in type 2 diabetes, 3% vs 1.8% on placebo |
3% vs 1.8% (SUSTAIN-6, type 2 diabetes) | [4][15] | Patients with type 2 diabetes; People with existing diabetic retinopathy |
| Optic nerve stroke (NAION) | Unknown (very weak evidence) Two studies of medical records disagree, one at a single centre, one larger across several countries |
Hazard ratio 4.28 (type 2 diabetes) and 7.64 (obesity) in one single-centre cohort; no significant association in a larger multinational records study | [9][10] | |
| Acute pancreatitis | Probably no harm (moderate evidence) Pooled analysis of 21 trials and 34,721 people, results could not rule out no effect |
Odds ratio 0.7 (95% CI 0.5 to 1.2) across 21 trials and 34,721 people; postmarketing reports include fatal cases | [15][17] | 34,721 people |
| Hip and pelvis fractures in people 75 and older | May happen (weak evidence) Drug label notes more in the SELECT trial, no rates reported |
Not reported; more in the semaglutide group in SELECT | [15] | |
| Acute kidney injury from dehydration | Unknown (very weak evidence) Only reports of single patients after the drug went on sale, some needing dialysis |
Case reports only (postmarketing), some needing dialysis | [15] | |
| Anaphylaxis and angioedema | Unknown (very weak evidence) Only reports of single patients on the drug label, from after it went on sale |
Case reports only (postmarketing) | [15] | |
| Bowel obstruction and ileus | Unknown (very weak evidence) Only reports of single patients on the drug label, from after it went on sale |
Case reports only (postmarketing) | [15] | |
| Inhaling stomach contents under anaesthesia | Unknown (very weak evidence) Only rare reports of single patients on GLP-1 drugs (semaglutide's class) after they went on sale |
Case reports only (rare postmarketing reports with GLP-1 drugs) | [15] | |
| Thyroid C-cell tumours | Unknown (very weak evidence) Only animal data, in mice and rats |
Animal data only (mice and rats) | [15] | Mice and rats |
| Harm to a fetus | Unknown (very weak evidence) Only animal data, drug label warns of possible harm to a fetus |
Animal data only | [15] | |
| Hospitalization and dosing errors with compounded products | Unknown (very weak evidence) Reports of single patients, plus higher odds of being reported in FDA safety reports |
Case reports only; higher reporting odds than branded products in FDA adverse event reports | [11][12] |
Under our grading rules, self-reported symptoms such as nausea or headache are graded no higher than moderate, while counted events and measured outcomes can be graded strong. Rates are from the label's three pooled 2.4 mg weight-loss trials (2,116 on semaglutide, 1,261 on placebo) unless noted [15]. In STEP 1, nausea and diarrhea were usually mild to moderate and faded over time [1]. Two of the three poison centre cases were tenfold dosing errors with vials and syringes measured in units or millilitres instead of prefilled pens [11].
This lists what studies and labels report. A pharmacist or prescriber can say whether it applies to a specific person.
This lists what studies and labels report. A pharmacist or prescriber can say whether it applies to a specific person.
| Combined with | What happened | Verdict | Source | Studied in |
|---|---|---|---|---|
| Insulin or sulfonylureas | Low blood glucose, including severe episodes, was more common in glycaemic control trials | Happens (strong evidence) Blood sugar control trials, more low blood glucose including severe episodes |
[15] | |
| Levothyroxine (with semaglutide tablets) | Thyroxine exposure (AUC) rose 33%; peak level unchanged | May happen (weak evidence) One tablet study, the drug label singles out the 33% rise |
[15] | |
| Other oral drugs with a narrow therapeutic range | The label states delayed stomach emptying could change absorption; no specific drugs measured | Unknown (very weak evidence) Drug label concern from delayed stomach emptying, no specific drugs measured |
[15] | |
| Oral contraceptive (ethinylestradiol/levonorgestrel) | Ethinylestradiol exposure unchanged; levonorgestrel exposure 20% higher; no reduction in either | May not increase it (weak evidence) One 43-person study, drug label lists no medically meaningful change |
[15][18] | |
| Lisinopril, warfarin, metformin, digoxin, furosemide, rosuvastatin | No clinically significant change in drug levels in interaction studies | May not increase it (weak evidence) Studies of taking both drugs together, no medically meaningful change in drug levels |
[15] | |
| Omeprazole (with semaglutide tablets) | No clinically significant change in semaglutide levels | May not increase it (weak evidence) No medically meaningful change in semaglutide levels with tablets |
[15] |
These are the amounts used in the studies. They are not recommendations.
Weight trials used 2.4 mg once weekly by injection, reached by stepping the dose up over several weeks, for 68 weeks [1][2]. SELECT used 2.4 mg weekly for a mean exposure of 34 months [3]. Diabetes trials used 0.5 mg or 1.0 mg weekly [4][5]. The Alzheimer's trials used a 14 mg daily tablet [7].
Semaglutide is FDA-approved as Ozempic for type 2 diabetes and Wegovy for overweight or obesity [13]. During national shortages, compounding pharmacies and other suppliers sold their own versions, some legally and some not, and fraudulent batches have been reported [13]. Reviews and case reports describe regulators' concerns about dosing errors, quality control and adverse events [11][13]. After the shortage was declared resolved, compounders could legally sell only products that differ from the approved ones, such as sublingual forms or added ingredients. Their efficacy and safety are largely unknown [14].