Twenty‑One Times a Month

The number travels well because it is specific. Twenty-one times a month. Five times a week, give or take. A tidy male-health hack that sounds like it came from a lab rather than a questionnaire. The research behind it is better than a meme and narrower than the version that circulates.

The figure comes from the Health Professionals Follow-up Study, a Harvard cohort of mostly White US clinicians and other health workers. In 1992, men were asked how many ejaculations they had, on average, each month in their twenties, in their forties, and in the previous year. Jennifer Rider, Kathryn Wilson, Lorelei Mucci, Edward Giovannucci and colleagues then watched who was later diagnosed with prostate cancer. The 2016 update in European Urology is the paper people mean when they say “21 a month.”

It is a strong observational finding. It is not a randomised trial, not a demonstrated cause, and not a measured gain in years of life.

Where the number actually comes from

Prostate cancer has almost no established levers a healthy man can pull. Age, ancestry, family history, and inherited variants dominate the risk list. That vacuum is why sexual behaviour has been studied for decades, and why a single clean number was always going to travel.

The biological idea is older than the Harvard paper. In 1983, John Isaacs argued that secretions sitting in the prostate might give carcinogens more time to do damage: the prostate stagnation hypothesis. Flush the gland more often, the thinking goes, and you reduce that dwell time. Later papers added other possible mechanisms: fewer prostatic crystalloids, a delayed metabolic switch in prostate cells, and a quieter sympathetic drive on the gland. None of those pathways has been proven in humans. They are the stories that make an association biologically plausible.

The first large study to measure ejaculation itself, rather than partners or intercourse as a proxy, was Australian. In 2003, Graham Giles and colleagues at Cancer Council Victoria published a case-control study in BJU International: 1,079 men with prostate cancer diagnosed before 70, and 1,259 age-matched controls in Melbourne, Sydney, and Perth. Men who averaged five or more ejaculations a week in their twenties had an odds ratio of 0.66 (95% CI 0.49–0.87) compared with men who did that less often. Number of sexual partners did not matter. Peak ejaculations in 24 hours did not matter. The exposure that showed a signal was total ejaculatory frequency, especially early in adult life.

Five a week is about 21 a month. The viral number and the Australian threshold are the same bin wearing different units.

A year later, Michael Leitzmann and the Harvard group published the first prospective version in JAMA. Among 29,342 men followed for eight years, 1,449 were diagnosed. Compared with 4–7 ejaculations a month, men reporting 21 or more had a multivariate relative risk of 0.89 in their twenties, 0.68 in their forties, 0.49 in the year before the questionnaire, and 0.67 averaged across adult life. That “50 percent lower” line is the one that entered popular coverage. The authors themselves flagged the problem: the strongest result sat in the year just before the form, which is exactly where undiagnosed disease, or the worry that follows it, can change sexual behaviour.

The 2016 paper is that same cohort with another decade of follow-up, more than double the cases, and a harder look at screening, tumour type, and competing death.

What the 2016 update actually found

The 1992 form asked a blunt question: “On average, how many ejaculations did you have per month during these ages?” The bins were none, 1–3, 4–7, 8–12, 13–20, and more than 20. The researchers did not ask whether the ejaculation came from intercourse, masturbation, or nocturnal emission. They chose 4–7 a month as the reference because too few men sat in the 0–3 group.

Frequency fell with age, as you would expect. Fifty-seven percent of men reported 13 or more a month in their twenties. That share dropped to 32 percent in their forties. Forty percent stayed in the same bin across those two decades; 47 percent dropped by one category.

Men in the highest bin at ages 40–49 were a little heavier, a little more active, a little more likely to be divorced, and they drank and smoked a bit more. They also reported more sexually transmitted infections. PSA testing, number of PSA tests, and biopsy after a raised PSA were similar across bins, which matters: a lower diagnosis rate is less interesting if the high-frequency men were simply screened less.

After adjustment, men reporting 21 or more ejaculations a month had a lower risk of a later prostate-cancer diagnosis than men reporting 4–7:

• Ages 20–29: hazard ratio 0.81 (95% CI 0.72–0.92)
• Ages 40–49: hazard ratio 0.78 (95% CI 0.69–0.89)
• Year before the 1992 form: hazard ratio 0.76 (95% CI 0.61–0.94)

The trend tests were tight (p < 0.0001 at both younger ages). Excluding men with erectile dysfunction, excluding cases in the first four years, and restricting the analysis to men who had already had a PSA test did not wipe the association out. That is the core of why this paper is still cited.

Percentages hide the scale. At ages 40–49, the absolute rate was 8.94 diagnoses per 1,000 person-years in the 4–7 group and 6.74 in the 21-or-more group. The difference is 2.20 cases per 1,000 person-years. Over a decade that is roughly two fewer diagnoses per 100 men, not a rewriting of male life expectancy.

Low-risk tumours, not a clear win against lethal disease

This is the part that usually falls off the slide.

Of the 3,839 diagnoses, 1,585 were localised low-risk disease, 1,493 were intermediate-risk, 604 were high-risk, and 157 already had regional or distant spread. The inverse association was driven by low-risk tumours. At ages 40–49, men with 13 or more ejaculations a month had a 28 percent lower risk of low-risk disease (HR 0.72) than the 4–7 group. High-risk disease (HR 0.89) and metastases at diagnosis (HR 0.96) were not significantly lower. Lethal disease, meaning death from prostate cancer or later distant spread, did not show a clean inverse trend for frequency in the forties.

There was a suggestive higher risk of advanced and lethal disease among men in the highest bin in the year before the questionnaire. That signal faded when the team dropped diagnoses in the first four years of follow-up. The honest reading is reverse causation, or chance, not a claim that frequent ejaculation causes aggressive cancer.

That pattern changes what the finding is for. Prostate cancer is the most commonly diagnosed cancer in US men, with an estimated 313,780 new cases and 35,770 deaths in 2025. About one in eight men is diagnosed in a lifetime. Five-year relative survival is 98 percent, and 15-year survival is 97 percent, because most tumours are found while they are still local or regional. An estimated 3.5 million US men were living with a prostate-cancer history in 2022. Many of those men will die of something else.

The American Cancer Society’s 2025 review is blunt about the implication: longevity after diagnosis makes treatment harm, not only death, the relevant outcome. Surgery and radiation can cost continence and erections. Active surveillance is now the default for many low-risk tumours for that reason. If more frequent ejaculation mainly reduces the chance of being diagnosed with a tumour that was never going to kill you, the longevity math is smaller than the headline, and the healthspan math may still be real: fewer biopsies, fewer treatments, fewer side effects.

Rider’s own conclusion said as much. More frequent ejaculation, “in the absence of risky sexual behaviours,” might cut the cost and the sexual harm of diagnosing and treating low-risk tumours, “even though it appears to be less strongly associated with aggressive disease.”

What later reviews added

The Harvard paper is still the best single prospective dataset. Later work has mostly asked whether the signal survives when you stop privileging one US cohort.

Zhongyu Jian and colleagues, writing in The Journal of Sexual Medicine in 2018, pooled 22 studies (21 of them case-control). They did not find a clean linear dose-response for ejaculation. They did find a lower risk at a moderate frequency of two to four times a week (OR 0.91, 95% CI 0.87–0.96). That band is 8–16 times a month, which overlaps the Harvard 13–20 bin more than it anoints 21 as a magic threshold. The same review found a higher risk with more female partners and a slightly lower risk with later first intercourse: a reminder that “sexual activity” is not one exposure.

A December 2025 update in BMC Cancer, led by Abouzar Raeisvandi and Jalal Poorolajal, pushed the pool to 29 studies and 315,193 men. Higher ejaculation frequency was the only sexual metric that stayed significant: pooled OR 0.83 (95% CI 0.77–0.90), with a significant inverse linear trend. Frequency of intercourse, number of female partners, age at first intercourse, and masturbation frequency as a separate item were all compatible with no effect. The authors rated the ejaculation evidence as moderate-confidence and the rest as weak. Most of the included studies were still case-control. Self-report, recall, and residual confounding do not disappear because you average them.

That split is the useful one. Ejaculation frequency is the exposure with a repeated signal. More partners is not a prostate-cancer strategy, and it raises infection risk. The Australian study already argued against infection as the main story in that population, because partner count was null while ejaculation count was not.

Twenty-one is a bin, not an optimum

The form had to put a top category somewhere. “More than 20” became ≥21 in the papers, then “21 times a month” in the retelling. The data do not show that 21 is better than 18, or that 25 is better than 21. They show that the highest bin, compared with a middling bin, had a lower diagnosis rate, and that the trend across bins mostly ran in that direction.

Place yourself on the same scale the study used.

The weekly column is there because that is how people actually think about the habit. It is also how Giles asked the question. If you want a translation that does not pretend to a precision the bins do not have: a few times a week, sustained through adult life, is the pattern the data keep circling. Not a daily quota. Not a score to hit in a tracking app.

Does this increase longevity?

Prostate cancer is common enough, and deadly enough in its aggressive form, that a true prevention effect would matter for male healthspan. The leap from “fewer low-risk diagnoses” to “more years of life” is the part the papers do not make.

Rider’s team modelled competing risks because they worried that high ejaculation frequency was just a marker of vigour, or that low frequency marked men who would die of something else before a prostate cancer could be found. Men in both the lowest and highest bins had a slightly higher risk of dying from other causes. The drop in prostate-cancer diagnoses in the high-frequency group was larger than that extra other-cause mortality, so competing death does not fully explain the result. That still leaves you with a lower diagnosis rate, not a measured gain in lifespan.

No trial has assigned men to a frequency and waited for deaths. No study in this literature reports extra years lived. The Million Veteran Program can put a number on stacked lifestyle factors because it counted deaths. This literature counted diagnoses.

A fair longevity reading is therefore modest:

• Prostate cancer remains the second-leading cause of cancer death in US men. Preventing lethal disease would add years. This dataset does not show that.
• Avoiding an unnecessary low-risk diagnosis can spare treatment that shortens the usable years you already have. That is healthspan, and it is the claim the 2016 authors were willing to make.
• Sexual function, relationship quality, and lower distress are longevity-adjacent for other reasons, including the social-connection factor that keeps showing up in male longevity lists. Those benefits do not require a quota of 21.

What this study cannot tell you

One more practical limit. The questionnaire did not distinguish safe from unsafe sex. If “get to 21” is read as “add partners,” you have left the evidence and picked up infection risk the authors told you to avoid.

What to take from it

Three things hold up.

The research is real, and it is specific. The 21-a-month figure is not wellness folklore. It is the top bin of a 1992 Harvard questionnaire, updated in 2016, prefigured by Giles in 2003, and still visible in a 2025 meta-analysis. If you have been wondering whether the number was invented on the internet, it was not.

Ejaculation is the exposure, not a partner tally and not a performance contest. The studies that survive contact with the evidence count ejaculations from any source. They do not reward riskier sex.

Treat it as a male-health footnote, not a longevity stack. The boring factors that move death rates, the ones in the eight-habit veteran analysis, still dwarf this. If a few times a week is already part of your life, the literature gives you no reason to stop. If it is not, this is a weak reason to force a quota and a good reason not to turn a prostate into a productivity target.

Prostate symptoms, screening decisions, and treatment choices belong with a clinician who has your history. A questionnaire from 1992 cannot do that job.

Sources

• Rider JR, Wilson KM, Sinnott JA, Kelly RS, Mucci LA, Giovannucci EL. Ejaculation Frequency and Risk of Prostate Cancer: Updated Results with an Additional Decade of Follow-up. European Urology, 2016;70(6):974–982 (PMID 27033442)
• Leitzmann MF, Platz EA, Stampfer MJ, Willett WC, Giovannucci E. Ejaculation Frequency and Subsequent Risk of Prostate Cancer. JAMA, 2004;291(13):1578–1586
• Giles GG, Severi G, English DR, et al. Sexual factors and prostate cancer. BJU International, 2003;92(3):211–216
• Isaacs JT. Prostatic structure and function in relation to the etiology of prostatic cancer. The Prostate, 1983;4:351–366
• Jian Z, Ye D, Chen Y, Li H, Wang K. Sexual Activity and Risk of Prostate Cancer: A Dose-Response Meta-Analysis. Journal of Sexual Medicine, 2018;15:1300–1309
• Raeisvandi A, Omidi S, Javaheri M, Moradi H, Poorolajal J. Updated dose-response meta-analysis of sexual activity and prostate cancer risk. BMC Cancer, 2025;26:47
• Kokori E, Olatunji G, Isarinade DT, et al. Ejaculation Frequency and Prostate Cancer Risk: A Narrative Review of Current Evidence. Clinical Genitourinary Cancer, 2024;22(3):102043
• Kratzer TB, Mazzitelli N, Goel N, et al. Prostate cancer statistics, 2025. CA: A Cancer Journal for Clinicians, 2025
• Boston University School of Public Health summary of the 2016 Rider paper