Treating Premature Ejaculation: What It Is and What Actually Works

The International Society for Sexual Medicine (ISSM) definition has three parts, and all three have to be present.

Ejaculation that always or nearly always happens within about one minute of penetration, if you have had it your whole life. Or a clear and bothersome reduction to about three minutes or less, if it developed later.

An inability to delay ejaculation on all or nearly all occasions.

Negative personal consequences: distress, frustration, or avoiding sex because of it.

Two other things worth knowing. Lifelong PE means it has been there since you became sexually active. Acquired PE means something changed, and something changing is worth investigating, because acquired PE is more often linked to another cause such as thyroid problems, prostate inflammation, or ED. And the research base is built almost entirely on vaginal intercourse. ISSM says directly that there is insufficient information to define this for oral sex, anal sex, or same-gender sexual activity.

Lifelong PE appears to be neurobiological. ISSM describes it as suggesting an underlying neurobiological functional disturbance, and variations in genes governing serotonin and dopamine signaling have been associated with shorter latency. This is why drugs acting on serotonin work.

Acquired PE is most commonly caused by performance anxiety, relationship problems, or ED, and occasionally by prostatitis, an overactive thyroid, or drug withdrawal. The thyroid link is striking: in men with hyperthyroidism the prevalence of PE was 50%, falling to 15% once thyroid levels were normalized.

You will see 20 to 30% quoted everywhere. It comes from surveys that asked men a single yes or no question about whether they ejaculate too soon.

When researchers used stopwatches instead, the picture changed considerably.

A word on that method, because it sounds implausible. Nobody was in the room. Couples were given a timer, took it home, and recorded every episode of intercourse over four weeks. The largest study did this with 500 couples across the Netherlands, the UK, Spain, Turkey and the US.

The American Urological Association (AUA) states that while up to 30% of men self-report PE, few have a latency under two minutes, making the actual prevalence under 5%. Most researchers in the field estimate 5 to 8%.

So, what are the other 15% of men experiencing? Usually a normal ejaculation time and an expectation set by pornography.

Here is the number that resets that expectation. In a stopwatch study of 500 couples across five countries, the median time from penetration to ejaculation was 5.4 minutes. It ranged from 33 seconds to 44 minutes. It fell with age, from 6.5 minutes in men aged 18 to 30 down to 4.3 minutes in men over 51. Condom use made no significant difference. Neither did circumcision.

Five and a half minutes is the middle of the distribution. Half of all men are below it.

The gap between belief and reality is measurable: in one survey 44.1% of men said the ideal time to ejaculation is more than 15 minutes, and 45.8% reported ejaculating sooner than their own ideal. The same survey found no association between frequent pornography use and ejaculation disorders.

In the PEPA survey, only 9% of men with PE had ever consulted a physician about it. Of those who did, 91.5% reported little or no improvement as a result.

That last figure needs unpacking. It does not mean the treatments failed. It means most of these men were not offered one. PE is frequently dismissed at the primary care level as a relationship issue rather than a treatable condition.

Certain antidepressants delay ejaculation as a side effect. In depression that might be a problem. In PE it is the point.

These are prescribed off-label for PE in the US, and the effect sizes are substantial. From the ISSM guideline:

Three important findings:

Daily dosing beats on-demand dosing decisively. Paroxetine taken as needed produces a 1.4-fold increase, roughly what placebo produces. Taken daily it produces 8-fold. This is the general pattern: taken on demand, most antidepressants barely beat placebo, because they need days of accumulation to work.

Paroxetine is the strongest, and it is also the one most associated with sexual side effects including reduced libido and erectile difficulty. Sertraline and fluoxetine trade some potency for a gentler side effect profile.

It takes time. Ejaculation delay usually begins within 5 to 10 days, but the full effect can take 2 to 3 weeks. Men who try it for a weekend and conclude it failed have not tested it.

Clomipramine deserves a mention, because it is the one drug here that works well on demand. Taken daily it produces a 6-fold increase, and unlike the SSRIs it holds up taken before sex, at 4-fold. That makes it the only antidepressant available in the US with a real on-demand option. The trade-off is tolerability: it is a tricyclic, carrying dry mouth, constipation, drowsiness, blurred vision, dizziness on standing and weight gain. It is more dangerous in overdose than an SSRI and interacts with more drugs. Dapoxetine is a short-acting SSRI developed specifically for on-demand use, producing a 2.5 to 3-fold increase. It is approved in Europe but not in the United States.

The AUA lists daily antidepressants as a first-line treatment, at its strongest recommendation grade.

This combination is the most interesting finding in the field, and it is where the evidence has moved most in the last decade.

A 2017 systematic review and meta-analysis pooled six randomized trials covering 521 men, comparing an ED medication plus an antidepressant against the antidepressant alone. The combination increased time to ejaculation by an additional 1.52 minutes. A further trial specifically testing tadalafil plus sertraline against sertraline alone found 1.02 minutes of additional benefit at six weeks.

The European Association of Urology now states that combining these two drug classes outperforms antidepressant monotherapy, at its highest level of evidence, and recommends using ED medication alone or in combination for PE even in men without ED.

Worth noting that the guidelines disagree here. ISSM's 2014 position was that ED medication should not be used for PE in men with normal erections. The EAU's 2026 position is the opposite. The evidence accumulated in between is the reason.

Why would combining them help? Several plausible contributors. Many men with PE have some degree of erectile anxiety, and a firmer, more reliable erection reduces the urgency to finish. ED medication shortens the recovery period between erections. And these drugs act on a different pathway than antidepressants do, so the effects add rather than overlap.

For men who have both PE and ED, which is a common combination, treating the ED is a guideline recommendation in its own right.

There is a second consideration if you are weighing a daily tablet. Remember that in these cases, you are exploiting a side effect of a drug that is used for significant mental health disorders. Antidepressants taken long term carry their own trade-offs, including reduced libido, which is a strange thing to accept in a treatment for a sexual problem. Daily tadalafil sits differently: it has been studied for years of continuous use, and the research points toward broader cardiovascular benefit. Our article on the case for daily tadalafil covers that evidence.

Tramadol is a painkiller that also delays ejaculation, and it occupies a useful niche: it is the only on-demand option with real evidence behind it. Every antidepressant except dapoxetine, which is unavailable in the US, needs daily dosing to work.

In the largest randomized trial, involving 604 men, tramadol increased median time to ejaculation by 1.2 minutes at 62 mg and 1.5 minutes at 89 mg, against 0.6 minutes on placebo. A 2015 meta-analysis found a pooled difference of 1.24 minutes favoring tramadol, and a 2021 meta-analysis found a mean difference of 139.6 seconds. For a drug taken only when you want it, roughly doubling to trebling your latency is a meaningful result.

One study directly compared it against daily paroxetine over 12 weeks. At 6 weeks tramadol produced a 7-fold increase and paroxetine 11-fold. By 12 weeks tramadol had declined to 5-fold while paroxetine had risen to 22-fold.

Tramadol is a “weak” opioid. It is a Schedule IV controlled substance, the same category as most sleep and anxiety medications, rather than Schedule II where oxycodone and morphine sit. The AUA notes its potential for addiction or abuse appears low in most populations, citing abuse rates of 0.7% against 1.2% for hydrocodone. It is not risk-free: it lowers the seizure threshold even at normal doses and carries a risk of serotonin syndrome, which may rule out combining it with an antidepressant. The AUA lists it as second-line.

Topical lidocaine and prilocaine reduce sensation in the head of the penis, which raises the threshold for ejaculation.

The best data comes from two phase 3 trials of a metered spray. Pooled, time to ejaculation rose from 0.58 minutes to 3.17 minutes on the active spray, against 0.56 to 0.94 minutes on placebo. That is roughly a 6-fold increase against 1.7-fold.

In the US, the specific product from those trials is not approved. What is available over the counter is lidocaine spray, sold under brand names including Promescent, and benzocaine wipes. Lidocaine and prilocaine cream are prescription products approved for other purposes and used off-label here.

There are practical and obvious drawbacks. Numbness can reduce your own pleasure, and the agent transfers to a partner and numbs them too, which is why a condom or wiping off before intercourse is usually advised. It also tastes unpleasant, which makes oral sex a problem unless thoroughly removed. Applied too heavily, or for too long, it can contribute to loss of erection.

The AUA lists topical anesthetics as first-line, alongside daily antidepressants.

These are the classic behavioral treatments.

The stop-start method, described by Semans in 1956, involves stimulation until you approach the point of no return, then stopping entirely until arousal subsides, then resuming. Repeated over weeks, the aim is to learn to recognize and control that threshold.

The squeeze technique, from Masters and Johnson in 1970, adds a firm squeeze at the base of the head of the penis at the same point.

Masters and Johnson reported a 97.8% success rate. That figure has never been replicated, and Cochrane says so directly. The original work had no control group, no randomization, no stopwatch measurement, and no validated outcome measure. Treat it as a 1970 uncontrolled claim rather than a result.

The honest modern picture is more modest. A three-year follow-up of couples treated with these methods found early gains that were largely not sustained. The AUA states that improvements have been shown to last 3 to 6 months after treatment stops, and that there are no studies of longer duration. In a survey of clinicians, 59% said fewer than 20% of their patients reported a response to behavioral therapy.

That is not a reason to skip it. It costs nothing, has no side effects, and teaches you something about your own threshold that no tablet does. It works while you keep doing it, the way physiotherapy works while you keep doing the exercises. Treat it as ongoing maintenance alongside medication rather than a course you complete.

Pre-sex masturbation is probably the most common management strategy men try, and there is no controlled trial of it.

The one study that examined it directly measured a second attempt after an initial ejaculation. Men with PE went from 2.4 minutes on the first attempt to 4.8 minutes on the second. Men without PE went from a longer baseline to 9.9 minutes.

So, it does something. But 57% of the men studied held a negative view of the approach, citing reduced desire, fatigue, and concerns about being able to get an erection for the second attempt. The authors concluded it is not universally applicable.

It is free and it is harmless, although it may increase anxiety with a failed sexual experience. It may also stop working as a strategy when your recovery time lengthens with age.

The pelvic floor is the sheet of muscle running between your pubic bone and tailbone that supports the bladder and bowel. Two of those muscles, the ischiocavernosus and bulbocavernosus, are directly involved in erection and ejaculation. Training them means learning to identify and contract them, then building strength and control, usually with a physiotherapist and sometimes with biofeedback. It gets almost no attention: only a handful of studies exist, no standard protocol has been agreed, and the 2022 AUA guideline on ejaculation disorders does not mention the pelvic floor once.

An Italian trial treated 40 men with lifelong PE and a baseline time to ejaculation of 31.7 seconds with 12 weeks of pelvic floor muscle rehabilitation. At 12 weeks the mean had risen to 146.2 seconds, a more than fourfold increase, and 33 of the 40 men, or 82.5%, had gained control of the ejaculatory reflex.

The trial was small and uncontrolled, and only 13 men reached six-month follow-up, where the mean had settled to 112.6 seconds. But the effect size is large, the intervention has no side effects, and it costs nothing.

The AUA's first-line recommendation, at its strongest grade, is daily antidepressants and topical anesthetics. It also lists two on-demand options with the caveat 'where available': clomipramine, which works on demand but is harder to tolerate, and dapoxetine, which is not approved in the US.

It recommends considering referral to a mental health professional with sexual health expertise.

It lists tramadol as second-line, for men who have failed first-line treatment.

And it makes a specific combination recommendation: men should be advised that combining behavioral and drug approaches may be more effective than either alone. The quantified benefit is roughly an additional minute over medication by itself, with some studies suggesting 1 to 3 minutes.

The pattern across every guideline is the same. No single treatment here is dramatic on its own. Stacking them is what produces a meaningful result.

Get assessed first, particularly if this is new. Acquired PE can point to a thyroid problem, prostate inflammation or ED, and those change the treatment.

Reset the expectation. The population median is 5.4 minutes, and half of all men are below it. If you last four minutes and it is causing distress, that distress is worth addressing, but your latency may not be the thing that is wrong.

Start behavioral work now, since it costs nothing and the guidelines recommend it alongside medication rather than instead of it. Pelvic floor training belongs here too.

Expect daily rather than on-demand dosing if you go the medication route, and give it two to three weeks.

If one drug at a full dose is not enough, the evidence supports adding a second mechanism rather than escalating the first. That is what the combination data shows.

Topical anesthetics produce roughly a 6-fold increase. Tramadol is the only on-demand option with real evidence, roughly doubling latency, and is a weak Schedule IV opioid with low reported abuse rates, though it cannot be combined with antidepressants. Behavioral techniques add about a minute and their benefit fades without maintenance. The strongest evidence is for combining mechanisms: adding an ED medication to an antidepressant adds 1.52 minutes over the antidepressant alone, and the EAU now recommends that combination at its highest level of evidence.