Does Ozempic Cause Erectile Dysfunction? What the Evidence Shows

It is not established that Ozempic, Wegovy, Mounjaro or Zepbound cause erectile dysfunction. No randomized trial of semaglutide or tirzepatide has measured erections, and none of the current US labels lists erectile dysfunction as a side effect. The observational studies disagree. One database study of young men without diabetes found more new ED diagnoses after semaglutide; a larger, better-controlled study in men with diabetes found a modest rise that lost statistical significance once the authors corrected for bias; several smaller studies found erections improved. The one randomized measurement inside a GLP‑1 outcomes trial, with dulaglutide, found slightly less new ED than placebo.

What the labels say: nothing

The words erectile, impotence, sexual and libido appear nowhere in the current US prescribing information for Ozempic,^[1] Wegovy,^[2] Mounjaro,^[3] Zepbound,^[4] Rybelsus^[5] or Saxenda,^[6] read in full and not only the side-effect tables. Those tables are where incidence from randomized trials is reported. ED is in none of them.

That absence has limits. We found no published semaglutide or tirzepatide trial that reported erections with a validated questionnaire, and a label can only list what the trials collected. A silent label means no signal surfaced in the trials as they were run. It does not mean anyone tested the question.

The study behind the headlines

The “Ozempic causes ED” story traces to a 2024 analysis of TriNetX, a commercial database of US health records. Researchers took men aged 18 to 50 with a BMI over 30 and no diabetes, excluded anyone with a past ED diagnosis, low testosterone or a prior prescription for sildenafil or tadalafil, and matched 3,094 men prescribed semaglutide against 3,094 who never were.^[7]

In the semaglutide group, 1.47% went on to a new ED diagnosis or a first PDE5 inhibitor prescription, against 0.32% of the comparison group: a risk ratio of 4.5. A new diagnosis of low testosterone was also more common, 1.53% against 0.80%.^[7] The 4.5 figure is what gets quoted. The absolute gap is about one man in a hundred.

Three features of the design matter more than the ratio. The outcome counted a sildenafil or tadalafil prescription as if it were a diagnosis. The comparison men had never been prescribed semaglutide, which also means they had not just opened a new prescribing relationship, often with a telehealth service that can write a PDE5 prescription in the same visit. And ED is recorded only when it is raised. A man seeing a prescriber every month for dose changes has more chances to raise it than a man who is not. None of this proves the result is wrong. It means the study cannot separate a drug effect from a contact effect.

The better-designed study shrank the signal

A 2026 study from the University of Pennsylvania attacked that contact problem directly. Instead of comparing GLP‑1 users with non-users, it compared men with type 2 diabetes starting a GLP‑1 with men starting a DPP‑4 inhibitor such as sitagliptin, another diabetes drug prescribed to similar patients. The design, called target trial emulation, sets up the comparison the way a randomized trial would and then fills it from health records: 4,910 GLP‑1 starters and 5,524 DPP‑4 starters, followed from 2019 to 2024.^[8]

New ED was recorded at 35.2 per 1,000 person-years on a GLP‑1 and 28.0 on a DPP‑4 inhibitor, a hazard ratio of 1.26. Then the authors ran a negative-control check: they measured the same comparison on outcomes the drug should not affect, to estimate how much hidden bias the data carried, and corrected for it. After that calibration, the increase was smaller and no longer statistically significant.^[8] Their own conclusion was that the finding may reflect residual or selection bias and does not establish causation.

The evidence pointing the other way

The only randomized data on erections from a large GLP‑1 trial come from REWIND, a cardiovascular outcomes trial of dulaglutide (Trulicity) in people with type 2 diabetes. Men were offered the International Index of Erectile Function (IIEF), the standard questionnaire, at baseline and at follow-up visits. Of 5,312 men, 3,725 were analyzed, with an average age of 65.5, and more than half already had moderate or severe ED at the start.^[9]

New moderate or severe ED occurred at 21.3 per 100 person-years on dulaglutide and 22.0 on placebo, a hazard ratio of 0.92 that was statistically significant, and erectile function scores fell slightly less on the drug.^[9] It was an exploratory analysis the trial was not designed around, funded by the drug’s maker, in older men with diabetes, on a drug weaker for weight than semaglutide or tirzepatide. It is also the closest thing to a randomized answer that exists.

A second TriNetX study, in men aged 18 to 70 with type 2 diabetes, compared tirzepatide with three other drugs using the same composite of an ED diagnosis or a PDE5 prescription. Tirzepatide came out lower in each comparison: a risk ratio of 0.70 against sitagliptin, 0.67 against injectable semaglutide and 0.55 against dulaglutide.^[10] That puts semaglutide behind tirzepatide on this outcome, which fits the Able result, but it is the same database and the same blind spots.

Smaller clinic studies, mostly from Italy and mostly in men who already had ED, diabetes or low testosterone, report better questionnaire scores on a GLP‑1:

Table sources: Able,^[7] Tang,^[8] Cowart,^[10] REWIND,^[9] Lisco,^[11] Defeudis,^[12] La Vignera 2023,^[13] La Vignera 2025,^[14] Giagulli^[15] and Seminara.^[16] Most of the positive studies enrolled men who already had ED or low testosterone, so they had the most room to improve, and few had a comparison group that got no new treatment.

A genetic study approached the question from a different angle. It used natural variation in the GLP‑1 receptor gene as a stand-in for lifelong drug exposure and found that the variants mimicking the drug went with roughly half the odds of ED.^[17] That kind of analysis, called Mendelian randomization, avoids prescribing bias but measures a lifetime of mild receptor activity, not a year of a weekly injection.

Adverse event reports: signals, not rates

FAERS, the FDA’s adverse event database, holds voluntary reports from patients, clinicians and manufacturers. A report means someone took a drug and later had a problem; it does not mean the drug caused it, and there is no denominator, so FAERS cannot give a rate. Disproportionality studies ask only whether a problem is reported more often with one drug than with everything else in the database.

• A 2026 FAERS analysis found 49 ED reports with semaglutide and a reporting odds ratio of 1.53. Sitagliptin and dapagliflozin, two diabetes drugs with no proposed link to erections, produced the same-sized signal (1.52 and 1.51). Dulaglutide, linagliptin and metformin did not. The authors called the findings preliminary and said they do not establish causality.^[18] A signal shared by three unrelated diabetes drugs points as much toward diabetes as toward any one of them.
• A 2025 FAERS analysis of all six GLP‑1 drugs from 2003 to early 2024 found 182 reports of male sexual problems (ED, orgasm problems or low libido), with a reporting odds ratio of 0.41: these were reported less often with GLP‑1s than with other drugs in the database.^[19] Again, a reporting pattern, not a rate.
• The raw counts. As of the FDA’s July 2026 data update, 105 FAERS reports listing semaglutide included the term erectile dysfunction, out of 73,001 reports listing semaglutide for any reason. For tirzepatide the figures were 61 of 160,663, and for liraglutide 58 of 50,704.^[20] These are reports, not cases shown to be caused by the drug.

What the reviews conclude

Pooled analyses lean toward benefit, built mostly from the small studies in the table. A 2026 meta-analysis of seven studies covering 237 men found erectile function scores 2.73 points higher on a GLP‑1, and called for larger randomized trials.^[21] A 2026 meta-analysis of eight studies and 375 men found positive erectile outcomes, though testosterone therapy did better on orgasm and satisfaction.^[22] A 2025 review of diabetes drugs found GLP‑1s ahead of metformin for ED, from three studies.^[23]

The most rigorous synthesis was narrower. A 2026 network meta-analysis restricted to 23 randomized trials (1,899 men with low testosterone driven by obesity) reported a significant IIEF improvement only for testosterone therapy combined with structured lifestyle change, and rated confidence in many comparisons as low or very low.^[24] A 2026 counseling review in Andrology drew the practical line:^[25]

These drugs should be selected for metabolic indications, not as treatments for ED or infertility.

— Broul et al., Andrology, 2026

No review we found concluded that GLP‑1 drugs cause ED. None concluded they should be prescribed to treat it.

Why the signals conflict: what else causes ED

Erections depend on blood flow, nerves, hormones and mood, and the men who get GLP‑1 prescriptions carry more of the risk factors for all four. ED affects an estimated 35% to 75% of men with diabetes.^[23] Any study comparing men on these drugs with men who are not has to untangle these:

• Obesity and diabetes themselves. Both damage the small blood vessels and nerves erections rely on. Men are prescribed a GLP‑1 because they have one or both.
• Low testosterone. Common with obesity and often improved by weight loss; the Able study also found more low-testosterone diagnoses after semaglutide.^[7] The hormone evidence is covered in our testosterone and fertility review.
• Antidepressants. The Zoloft label, for one, warns that SSRIs can cause erectile dysfunction, delayed ejaculation and lower libido in men, and tells prescribers to ask about sexual function before starting.^[29] A man who starts an SSRI and a GLP‑1 in the same year has two candidates. How the two drug types interact is covered in antidepressants and a GLP-1.
• Heart and blood-vessel disease. ED and cardiovascular disease share the same blood-vessel damage, and ED can show up before coronary disease causes symptoms.^[28] In the genetic study, part of the GLP‑1 association ran through lower rates of diabetes, obesity, high blood pressure and heart disease.^[17]
• Desire versus function. Some men report less interest in sex on these drugs, which can be mistaken for an erection problem. That question has its own evidence, covered in GLP-1 drugs and sex drive. Questions about apparent size after weight loss are in GLP-1 drugs and penis size.
• Feeling unwell. Weeks of nausea, fatigue and eating little can dampen sex for reasons unrelated to blood vessels or hormones. If that is the cause, it should ease as side effects settle. No study has tested it.

Taking Viagra or Cialis with a GLP-1

No label lists an interaction between a PDE5 inhibitor and a GLP‑1. The Viagra and Cialis labels do not mention semaglutide, tirzepatide or GLP‑1 drugs,^[26][27] and the six GLP‑1 labels do not mention sildenafil, tadalafil or PDE5 inhibitors. No study has measured the pair together. The overlap that matters is blood pressure.

Both PDE5 labels warn that combining them with blood pressure medicines or alpha-blockers can drop blood pressure, and Viagra’s label names low fluid volume as one of the variables that affects that risk.^[26] Cialis adds heavy drinking, five or more units, to the list.^[27] On the GLP‑1 side, low blood pressure is uncommon but measured:

Sources: Wegovy^[2]; Zepbound.^[4] The GLP‑1 labels also warn of kidney injury in people dehydrated by vomiting or diarrhea.^[1][3] Put together, the practical risk window is a week of stomach trouble, especially right after a dose increase, in a man who also takes blood pressure medicine. That is our reading of two sets of labels, not a measured interaction. Lightheadedness on a GLP‑1 alone is covered in GLP-1 dizziness and blood pressure pills.

GLP‑1 drugs slow stomach emptying, and their labels say this could affect how oral medicines are absorbed; Ozempic’s label adds that its pharmacology studies found no clinically relevant effect on the oral drugs tested.^[1] Whether it changes how quickly a PDE5 tablet starts working has not been studied.

In the event of an erection that persists longer than 4 hours, the patient should seek immediate medical assistance.

— Viagra (sildenafil) prescribing information

The absolute rule on both labels is unchanged by a GLP‑1: no PDE5 inhibitor for anyone using organic nitrates such as nitroglycerin, in any form, even occasionally.^[26][27] Bremelanotide (PT‑141), sometimes sold to men online for erections, is approved only for premenopausal women; see PT-141 (bremelanotide): what the evidence shows.

If erections change after starting a GLP-1

1. Do not stop the drug on your own. Stopping brings back appetite and, for men with diabetes, blood sugar; what happens when you stop a GLP-1 covers the trial data. A dose change is a prescriber decision.
2. Tell the prescriber, with dates. When it started, whether it began after a dose increase or a stretch of nausea, and whether desire, erections or both changed.
3. Ask to be checked, not just prescribed. ED sometimes shows up before heart disease does,^[28] so new ED is worth a look at blood pressure, blood sugar and cholesterol whether or not a GLP‑1 is involved, and at testosterone where symptoms fit.
4. Bring the full medicine list. Blood pressure tablets, antidepressants and anything bought online for weight or sex all matter here.
5. If a PDE5 inhibitor is offered, mention any nitrate, alpha-blocker or blood pressure medicine, and ask what to do during a bout of vomiting or diarrhea, when fluid loss is the risk both sets of labels describe.
6. Get urgent care for chest pain during sex, fainting, or an erection lasting more than four hours.