Ejaculation Frequency and Prostate Cancer: What the Harvard Study of 31,925 Men Found, and What Abstinence Really Does to Testosterone

In a Harvard cohort of 31,925 men, those reporting 21 or more ejaculations a month had a lower rate of prostate cancer than men reporting 4 to 7. The link was driven by low-risk disease, and the day-7 testosterone peak from abstinence comes from a study that was later retracted.

Ejaculation frequency and prostate cancer have been studied together for more than two decades, and the headline finding has held up: men who ejaculate more often are diagnosed with prostate cancer somewhat less often. The largest study, from the Health Professionals Follow-up Study at Harvard, followed 31,925 men from 1992 to 2010. Men who reported 21 or more ejaculations a month in their 40s had a hazard ratio of 0.78 for prostate cancer compared with men who reported 4 to 7 (Rider and colleagues, European Urology, 2016).

That number gets repeated online as "ejaculate 21 times a month to prevent prostate cancer." The research does not say that. The data are observational, the link was driven by low-risk tumors, and the men who ejaculated most also differed from the others in diet, alcohol, weight and relationship history. A second popular claim, that skipping ejaculation for a week spikes testosterone, rests largely on a 2003 study that has since been retracted. This guide walks through what each study measured and what a man can reasonably take from it, including one practical detail that does matter: ejaculation can raise a PSA result for a day or two.

Frequent ejaculation is associated with a lower risk of prostate cancer, but no trial has shown that it causes that lower risk. In the Health Professionals Follow-up Study, men answered a 1992 questionnaire about their average number of ejaculations per month at ages 20 to 29, at ages 40 to 49 and in the year before the survey. Researchers then tracked them to 2010. Over 480,831 person-years, 3,839 of the men were diagnosed with prostate cancer.

Compared with men reporting 4 to 7 ejaculations a month, men reporting 21 or more had a hazard ratio of 0.81 (95% CI 0.72 to 0.92) based on frequency in their 20s and 0.78 (95% CI 0.69 to 0.89) based on frequency in their 40s. A hazard ratio below 1 means the high-frequency group was diagnosed at a lower rate during follow-up, after adjusting for other factors. The same cohort had produced a smaller analysis in JAMA in 2004 (Leitzmann and colleagues) with 1,449 cancers. Both are plotted below.

The 2016 update is the stronger estimate. It added about a decade of follow-up and more than twice as many cancer cases as the 2004 paper, which narrowed the confidence intervals. The 2004 figure for ages 20 to 29, 0.89, crossed 1 (95% CI 0.73 to 1.10) and was not statistically significant on its own. By 2016 the estimate for that decade was 0.81 and significant.

The Harvard studies compared men reporting 21 or more ejaculations a month with men reporting 4 to 7. The JAMA authors wrote that most frequency categories were unrelated to risk and that the difference showed up at the high end. That comparison is where the popular "21 a month" figure comes from, but the study did not test 21 as a target or threshold. It was simply the top category on the questionnaire.

A 2018 dose-response meta-analysis in the Journal of Sexual Medicine (Jian and colleagues) pooled 21 case-control studies and 1 cohort study with 55,490 participants. It found no linear association between ejaculation frequency and prostate cancer. Moderate frequency, 2 to 4 times per week, was linked to a lower risk (odds ratio 0.91, 95% CI 0.87 to 0.96). A 2025 meta-analysis in BMC Cancer (Raeisvandi and colleagues) of 29 studies and 315,193 participants put the pooled odds ratio for higher ejaculation frequency at 0.83 (95% CI 0.77 to 0.90).

The association between ejaculation frequency and prostate cancer has been consistent across studies, but confounding has not been ruled out. In the Harvard cohort, men with higher ejaculation frequency in their 40s also had higher body mass index, more physical activity, more divorce, more history of sexually transmitted infections, and higher intake of calories and alcohol. The researchers adjusted for these in their models, yet no observational study can adjust for everything that separates one group of men from another.

Screening was the obvious worry. If men with more active sex lives also saw doctors less, they might simply have had fewer cancers found. The 2016 paper checked this directly. PSA test use by 2008, the number of PSA tests and the frequency of prostate biopsy were similar across ejaculation categories, and the association held when the analysis was restricted to men who had been PSA screened. The authors also judged it unlikely that deaths from other causes explained the pattern.

The meta-analyses carry their own caveat. Jian and colleagues noted that most of the studies they pooled were case-control designs, in which men already diagnosed with cancer recall their sexual history from decades earlier. That recall can be distorted, which is why the 2004 JAMA authors stressed the value of a prospective cohort in the first place.

That 2025 meta-analysis is useful because it separated sexual behaviors. Intercourse frequency (OR 1.02), number of female partners (OR 1.01) and masturbation frequency (OR 1.12) were not significantly associated with prostate cancer. Only the overall ejaculation frequency measure was. The authors concluded that most sexual behaviors are not significantly linked to prostate cancer risk, while higher ejaculation frequency appears to have a protective association.

Nobody has established how ejaculation frequency could influence prostate cancer, and the biology is still at the hypothesis stage. The Harvard group took one step toward an answer in a 2018 European Urology paper (Sinnott and colleagues). They examined prostate tissue from men in the cohort who later developed cancer: 157 tumor samples and 85 samples of adjacent normal tissue.

In the tumor tissue, gene expression showed no association with ejaculation frequency after correcting for multiple testing. In the adjacent normal tissue, 409 genes and six biological pathways differed across ejaculation frequency categories. The authors described these as potential biological links. That is a reasonable lead for future research, but a difference in gene activity is not proof that ejaculation prevents tumors.

Evidence that abstaining from ejaculation meaningfully raises testosterone is thin, and the best-known study behind the idea has been retracted. In 2003, Jiang and colleagues reported in the Journal of Zhejiang University Science that serum testosterone in 28 volunteers peaked on the 7th day of abstinence at 145.7% of baseline. That day-7 figure still circulates widely in abstinence and "NoFap" content. The journal issued a retraction note for the paper in December 2021, and Crossref now lists it as a retracted article.

The other study usually cited is much smaller. Exton and colleagues, in the World Journal of Urology in 2001, measured hormones in ten healthy men during masturbation-induced orgasm, once before and once after a 3-week period of sexual abstinence. Orgasm itself did not change plasma testosterone. Testosterone concentrations were higher after the abstinence period. With ten men and no long-term follow-up, that result cannot tell anyone whether abstinence produces a lasting or clinically meaningful change.

If low testosterone is the concern, the reliable route is a properly timed morning blood test, repeated to confirm, rather than changing sexual habits. Our guide to testing testosterone accurately covers timing and the mistakes that produce a falsely low result, and our article on low testosterone symptoms in men over 40 explains which symptoms actually track with low levels.

Ejaculation can raise PSA for a day or two, so many clinicians ask men to avoid it before the blood draw. Tchetgen and colleagues, in Urology in 1996, measured PSA in 64 men aged 49 to 79 before ejaculation and at intervals afterward. PSA rose in 87% of them. The average relative increase was 41% at 1 hour and stayed elevated at 6, 24 and 48 hours. The authors recommended abstaining from ejaculation for 48 hours before a PSA test.

Most men in that study returned to baseline within 24 hours (92%), but not all. A second study by Herschman, Smith and Catalona (Urology, 1997) followed 20 men with a mean age of 59. After accounting for normal biological variation, 40% still had total PSA above their baseline 24 hours after ejaculation, and percent free PSA was below baseline in 35%. A falsely high total PSA or a falsely low free PSA fraction can push a borderline result toward more testing.

For men over 40, the ejaculation and prostate cancer research is reassuring rather than prescriptive. Frequent ejaculation has not been linked to any increase in prostate cancer risk, which was the original concern that prompted the 2004 JAMA study. The data point the other way, modestly, and mostly for low-risk cancers. There is no basis for a quota, and no evidence that changing frequency later in life changes a man's personal risk.

• Treat ejaculation frequency as one of many observational associations, not a prevention strategy.
• Do not let the research replace a screening conversation with a doctor, especially with a family history of prostate cancer.
• Avoid ejaculating for 48 hours before a PSA test so the result reflects your baseline.
• Ignore claims that a week of abstinence produces a testosterone spike; the study behind them was retracted.
• If you are on TRT, see our guide to PSA and prostate monitoring on testosterone therapy for how often to test.

This article is for educational purposes only and is not medical advice. It does not diagnose, treat, cure or prevent any disease. Speak with a doctor about prostate cancer screening, and seek care for blood in the urine or semen, new urinary symptoms or bone pain.

• Rider JR et al. Ejaculation Frequency and Risk of Prostate Cancer: Updated Results with an Additional Decade of Follow-up. European Urology, 2016. https://doi.org/10.1016/j.eururo.2016.03.027
• Leitzmann MF et al. Ejaculation frequency and subsequent risk of prostate cancer. JAMA, 2004. https://doi.org/10.1001/jama.291.13.1578
• Giles GG et al. Sexual factors and prostate cancer. BJU International, 2003. https://doi.org/10.1046/j.1464-410x.2003.04319.x
• Papa NP et al. Ejaculatory frequency and the risk of aggressive prostate cancer. Urologic Oncology, 2017. https://doi.org/10.1016/j.urolonc.2017.03.007
• Jian Z et al. Sexual Activity and Risk of Prostate Cancer: A Dose-Response Meta-Analysis. Journal of Sexual Medicine, 2018. https://doi.org/10.1016/j.jsxm.2018.07.004
• Raeisvandi A et al. Updated dose-response meta-analysis of sexual activity and prostate cancer risk. BMC Cancer, 2025. https://doi.org/10.1186/s12885-025-15410-3
• Sinnott JA et al. Differential Gene Expression in Prostate Tissue According to Ejaculation Frequency. European Urology, 2018. https://doi.org/10.1016/j.eururo.2018.05.006
• Jiang M et al. A research on the relationship between ejaculation and serum testosterone level in men. 2003 (retracted). Retraction note: https://doi.org/10.1631/jzus.2003.r236
• Exton MS et al. Endocrine response to masturbation-induced orgasm in healthy men following a 3-week sexual abstinence. World Journal of Urology, 2001. https://doi.org/10.1007/s003450100222
• Tchetgen MB et al. Ejaculation increases the serum prostate-specific antigen concentration. Urology, 1996. https://doi.org/10.1016/S0090-4295(99)80486-5
• Herschman JD et al. Effect of ejaculation on serum total and free prostate-specific antigen concentrations. Urology, 1997. https://doi.org/10.1016/S0090-4295(97)00209-4

Continue Reading

Related Articles

Hormones

PSA on TRT: The Prostate Number Men Misread in Both Directions

12 min read

Read Article →

Hormones

How to Test Testosterone Accurately: Timing, Fasting and the Mistakes That Fake a Low Result

11 min read

Read Article →

Hormones

Low Testosterone Symptoms in Men Over 40: The Three Signs That Actually Predict Low T, and the Ones That Don't

10 min read

Read Article →