Premature ejaculation drugs work in trials, but the trials are mostly low quality

Drugs that delay ejaculation do lengthen the time to climax in trials, but the evidence behind them is large, messy and mostly low quality, and only one medicine anywhere is licensed specifically for the problem. That medicine, dapoxetine, is not approved in the United States, where every drug used for premature ejaculation is prescribed off-label [s2][s3].

Premature ejaculation is one of the most common male sexual complaints, and it has a formal definition. The International Society for Sexual Medicine describes lifelong premature ejaculation as ejaculation that always or nearly always occurs before or within about one minute of vaginal penetration, present from a man's first sexual experiences [s1]. Acquired premature ejaculation, which develops later, is defined by a clinically significant and bothersome reduction in latency, with a self-estimated or stopwatch cut-off of about three minutes [s1]. Both forms share two further requirements: an inability to delay ejaculation, and distress or avoidance because of it [s1]. The one-minute and three-minute thresholds matter, because much older writing on the topic used no objective measure at all and treated any dissatisfaction as a disorder.

What the drugs are

The pharmacological toolkit is small and largely borrowed. Selective serotonin reuptake inhibitors delay ejaculation as a side effect, and this is the mechanism behind most drug treatment. Dapoxetine is a short-acting SSRI developed and approved for on-demand use in premature ejaculation in a number of countries, but it is the only SSRI approved for the indication, and it is not marketed in the US [s2][s3]. The American Urological Association and the Sexual Medicine Society of North America, in their joint 2022 guideline, describe the routine options as off-label longer-acting SSRIs such as paroxetine, topical anaesthetic creams and sprays applied to the penis, phosphodiesterase type 5 inhibitors such as sildenafil, and the opioid tramadol [s2]. Behavioural techniques and, for some men, treating a coexisting erectile problem are also part of standard management [s2].

That a treatment is off-label does not make it ineffective. It means the drug was licensed for something else and is being used on the strength of trial evidence rather than a regulator's sign-off for this specific use. The question is how good that trial evidence actually is.

The evidence, weighed honestly

The most useful recent appraisal is a 2025 umbrella review, a study of studies, that pulled together 44 systematic reviews and meta-analyses covering 65 randomised controlled trials of drug treatment for premature ejaculation [s3]. Its headline is reassuring and its fine print is not. Across the trials, the treatments significantly lengthened the intravaginal ejaculatory latency time compared with placebo, and after the authors removed overlapping trials and re-ran the numbers, paroxetine produced the largest average increase [s3]. So the drugs do something measurable.

But the same review is blunt about quality. Of the 44 reviews it examined, only two rated as moderate-to-high quality on a standard appraisal tool, and of the 65 underlying trials, only six were judged to be at low risk of bias [s3]. The median follow-up across those trials was 7.9 months, which is short for a chronic condition that men may want to manage for years [s3]. In other words, a treatment can clear the bar of "better than placebo" while the studies proving it remain small, brief and methodologically shaky, an inconsistency the review attributes to varied outcome measures and uneven trial quality across the field [s3].

There is a further gap the trials rarely close: what "works" means to the man taking the drug. Stopwatch latency is easy to measure and lengthens with treatment, but control over ejaculation and the distress that brought a man to the clinic are what the definition actually turns on [s1], and those subjective outcomes are measured less consistently. A drug that adds a minute on a stopwatch has not necessarily fixed the problem the patient described.

What it means for a reader

The practical picture is that ejaculation-delaying drugs are a legitimate, evidence-supported option, but one built on a weaker foundation than the volume of published trials suggests. The safest reading of the literature is that these medicines modestly lengthen latency, that side effects are common enough to matter, and that the choice among them is not settled by high-quality head-to-head data. That is also why guidelines still frame drug treatment as one part of a broader approach rather than a cure [s2].

Premature ejaculation and its partner condition are often tangled together; the AUA guideline notes that acquired premature ejaculation frequently coexists with erectile dysfunction, and that treating the erectile problem can help [s2]. And because low sexual desire, erectile difficulty and ejaculatory complaints are routinely conflated in men's health marketing, it is worth being clear that these are distinct problems with distinct evidence, separate again from the question of whether low testosterone is the cause.

The next thing worth watching is not a new drug but better trials: longer follow-up, consistent patient-centred outcomes, and head-to-head comparisons rather than yet another placebo study. Until those arrive, the honest summary is that the pills help, the proof is thinner than it looks, and the marketing is well ahead of the data.