How to Increase Testosterone Naturally: What the Evidence Shows

In men carrying extra weight, losing it has the clearest measured effect: across 44 studies of 1,774 men, diet-led weight loss raised total testosterone by about 2.5 nmol/L. Across 18 sleep studies, a night or more without sleep lowered it, while a few short nights did not; hard exercise raised it for under 30 minutes, about 12 weeks of training left resting levels unchanged in 11 trials, and regular drinking went with lower levels in 21 studies.^{1, 2, 3, 4, 5} Low-fat and very high-protein, low-carbohydrate diets lowered it; honey, ginger, garlic and receipts have no trial in men; a vasectomy made no hormone difference; cycling studies measured erection problems, not testosterone.^{6, 7, 8, 9} The NHS says the fall with age is "steady at about 1% a year from around the age of 30 to 40, and this is unlikely to cause any problems in itself"; symptoms go to a GP.¹⁰ More in the Wellgard men's health guide.

Evidence at a glance

• Rests on: pooled studies of weight loss (44 studies, 1,774 men), sleep loss (18 studies, 252 men), training (11 trials, 421 men) and alcohol (21 studies, 10,199 men).^{1, 2, 4, 5}
• Largest study following men over time: 2,395 European men aged 40 to 79, about four years.¹¹
• Strongest measured lifestyle outcome: total testosterone about 2.5 nmol/L higher after diet-led weight loss (19 studies, 735 men; median follow-up 26 weeks across all 44).¹
• Not tested: honey, ginger, onions or garlic in a randomised trial in men, or sleeping naked in any study (PubMed, 13 October 2026).

Sources checked on 13 October 2026.

Does testosterone fall with age, and what lowers it?

Slowly, and the fall rarely explains symptoms. The NHS "male menopause" page (last reviewed 13 October 2022; review overdue) calls the term unhelpful and says "in many cases the symptoms are nothing to do with hormones".¹⁰ The causes it lists include lack of sleep, a poor diet, lack of exercise, drinking too much alcohol, smoking, and stress, depression or anxiety (how to lower cortisol covers stress). Late-onset hypogonadism is "not a normal part of ageing" and is more likely in men who are obese or have type 2 diabetes.¹⁰ Medical causes are in low testosterone: symptoms and the blood test.

Does losing weight raise testosterone?

Yes, in men carrying extra weight. Across 44 studies of 1,774 men followed for a median of 26 weeks, total testosterone rose by about 2.5 nmol/L after diet-led weight loss and 7.2 nmol/L after weight-loss surgery, more in heavier men and those starting lower.¹ In 2,395 European men followed for about four years, losing weight went with a proportional rise in testosterone and gaining it with a fall; free testosterone moved only with a weight change of 15% or more.¹¹ An association, not proof. Once a deficiency is confirmed, UK specialist guidance does not support lifestyle change alone as the first treatment, so symptoms still go to a GP.¹²

Does sleep affect testosterone?

Losing a whole night does. Across 18 studies of 252 healthy men, 24 hours or more without sleep lowered testosterone; short-term partial sleep loss made no significant difference.² In a publicly funded laboratory study of 10 young men, a week of 5-hour nights cut daytime testosterone by 10% to 15%.¹³ Two small publicly funded randomised studies found no change after five nights of 4 hours in bed (14 men), and levels slightly lower but rising over six weeks of 90 minutes less (13 men).¹⁴ Sleeping naked is untested; in 656 men at a fertility centre, boxers went with higher sperm counts, not different testosterone.¹⁵ More in the Wellgard sleep guide.

Does exercise raise testosterone? Weights, legs and running

Briefly. Across 48 studies of 569 men, moderate and hard exercise raised testosterone within 30 minutes but not after; mild activity did not. The analysis graded effort, not muscle groups.³ In 11 randomised trials of 421 inactive men with normal testosterone, about 12 weeks of training left resting levels unchanged.⁴ A review describes lower resting testosterone in some men after years of heavy endurance training.¹⁶ The NHS says exercise and relaxation can help with stress.¹⁰

Does alcohol lower testosterone?

Regular drinking goes with lower levels. Across 21 studies of 10,199 men, alcohol went with lower total and free testosterone in healthy men who drank regularly, not after a single session.⁵ The NHS advises no more than 14 units a week.¹⁷

Which foods and drinks affect testosterone?

No single food has a human trial showing a rise; the one placebo-controlled pomegranate juice trial in men, nine weightlifters, with no external funding, found testosterone lower just after training.¹⁸ In 6 small studies of 206 men, low-fat diets lowered testosterone against higher-fat diets.⁷ In 27 studies of 309 men, low-carbohydrate diets moderate in protein had no consistent effect; those with 35% or more of energy from protein lowered resting testosterone in three short studies of 26 men (moderate certainty, in the review's grading).⁶ See how much protein you need; soya has its own page. Zinc contributes to the maintenance of normal testosterone levels in the blood, and its trials were in men low in zinc (foods high in zinc; zinc for men). Vitamin D contributes to the maintenance of normal bones, the maintenance of normal muscle function and the normal function of the immune system; its testosterone trials are in do testosterone supplements work? and the Wellgard vitamin D guide.

Do masturbation, a vasectomy, cycling, cold showers, receipts or smoking change testosterone?

None lowered resting testosterone where it was measured; a cold soak after weights blunted the short post-training rise, and the best-known abstinence figure is retracted.

When is low testosterone a GP question rather than a lifestyle one?

When symptoms last. The NHS advises a GP for symptoms such as loss of sex drive, erection problems, depression or tiredness; the GP may order a blood test and, if it suggests a deficiency, refer you to a specialist, who may offer testosterone replacement, as an injection or a gel, once the diagnosis is confirmed.¹⁰ The test is in low testosterone: symptoms and the blood test; supplements are in do testosterone supplements work? Under-18s with puberty questions: see a GP.

What we don't know

• Whether months of short sleep lower testosterone: the studies ran five nights to six weeks.^{13, 14}
• Whether a lifestyle-driven rise changes how men feel: the pooled weight and training studies measured hormones.^{1, 4}
• Whether training much longer than 12 weeks moves resting testosterone in men with normal levels.⁴

Frequently asked questions

How fast can testosterone rise with lifestyle changes?

Over months: the weight-loss studies ran a median of 26 weeks, and the rise after a workout was gone after 30 minutes.^{1, 3}

Does not ejaculating for 7 days increase testosterone?

Not on sound evidence: the day-7 figure comes from a 28-man study its journal retracted in 2021.²⁰ In 10 men studied in 2001, testosterone was higher after three weeks without ejaculating; the study measured hormones, not health.¹⁹

Which drink is best for testosterone?

No drink has a trial showing a lasting rise. In 13 men, a low dose of alcohol raised testosterone briefly (funding not checked), while regular drinking went with lower levels; in nine weightlifters, pomegranate juice left it lower than placebo just after training.^{24, 5, 18}

Can low testosterone go back to normal?

In men carrying extra weight, testosterone rose as weight fell.^{1, 11} Other causes need a GP and sometimes a specialist; see low testosterone: symptoms and the blood test.¹⁰

Sources

1. Ken-Dror et al., 2024. Pooled analysis, 44 studies, 1,774 men. PubMed 37345263.
2. Su et al., 2021. Pooled analysis, 18 studies, 252 men. PubMed 34801825.
3. D'Andrea et al., 2020. Pooled analysis, 48 studies, 569 men. PubMed 32297287.
4. Potter et al., 2021. Pooled analysis, 11 randomised trials, 421 men. PubMed 35134000.
5. Santi et al., 2024. Pooled analysis, 21 studies, 10,199 men. PubMed 37705506.
6. Whittaker and Harris, 2022. Pooled analysis, 27 studies, 309 men; no funding; corrected 2022. PubMed 35254136.
7. Whittaker and Wu, 2021. Pooled analysis, 6 studies, 206 men; corrected 2026. PubMed 33741447.
8. Zhao et al., 2018. Observational, 2,425 men. PubMed 30356207.
9. Baran et al., 2014. Review. PubMed 27784566.
10. NHS. The "male menopause". Page last reviewed 13 October 2022 (review overdue). nhs.uk.
11. Camacho et al., 2013. Observational, 2,395 men, about 4 years. PubMed 23425925.
12. Hackett et al., 2023. Guideline, 25 statements; funded by the issuing society; the lead author declares advisory roles, and a co-author travel grants, with testosterone-treatment companies. PubMed 36876744.
13. Leproult and Van Cauter, 2011. Laboratory study, 10 men, 1 week; publicly funded; figures from the full text. PubMed 21632481.
14. Smith et al., 2019. Two randomised crossover studies, 14 and 13 men, 5 nights and 6 weeks; publicly funded. PubMed 31416797.
15. Mínguez-Alarcón et al., 2018. Observational, 656 men. PubMed 30102388.
16. Hackney et al., 2005. Review. PubMed 16268050.
17. NHS. Erectile dysfunction (impotence). Page last reviewed 28 July 2023 (review overdue). nhs.uk.
18. Ammar et al., 2020. Randomised crossover trial, 9 men; no external funding. PubMed 32143716.
19. Exton et al., 2001. Repeated-measures study, 10 men, 3 weeks; funding not checked. PubMed 11760788.
20. Jiang et al., 2003. 28 men. Retracted in 2021. PubMed 12659241.
21. Earp et al., 2019. Randomised crossover trial, 11 men; funding not checked. PubMed 31222379.
22. Commission Regulation (EU) 2016/2235, Annex, as retained in UK law. legislation.gov.uk.
23. Zhao et al., 2016. Pooled analysis, 22 observational studies, 13,317 men. PubMed 26763163.
24. Sarkola and Eriksson, 2003. Randomised crossover trial, 13 men; funding not checked. PubMed 12711931.

Spotted something out of date? Email support@wellgard.co.uk with the page address and the sentence.