Testosterone and Blood Sugar: What the Metabolic Trials Show

In men with confirmed low testosterone, testosterone and blood sugar are linked but the effect is conditional. The T4DM trial (Wittert 2021, Lancet Diabetes and Endocrinology) randomized 1,007 men aged 50 to 74 already in a lifestyle program, and at two years 12 percent on testosterone versus 21 percent on placebo met the diabetes glucose threshold, relative risk 0.59.

Key takeaways

• In T4DM (Wittert 2021, Lancet Diabetes and Endocrinology), a 2-hour glucose of 11.1 mmol/L or higher on an oral glucose tolerance test occurred in 87 of 413 men (21 percent) on placebo versus 55 of 443 (12 percent) on testosterone at two years, relative risk 0.59 (95% CI 0.43 to 0.80, p = 0.0007).
• Every T4DM participant was enrolled in a lifestyle program, and participants were selected to be without pathological hypogonadism, so the result describes testosterone added to lifestyle change in that specific population.
• Zhang’s 2018 meta-analysis in World Journal of Urology (8 trials, 596 hypogonadal men with type 2 diabetes) reported an HbA1c reduction of 0.45 percent (95% CI -0.73 to -0.16), while Li’s 2020 meta-analysis (18 trials, 1,415 patients) reported 0.67 (95% CI -1.35 to -0.19).
• Corona’s 2015 meta-analysis in the European Journal of Endocrinology, covering 59 randomized trials with 3,029 treated and 2,049 control participants, found reduced fat mass, increased lean mass and reduced fasting glucose and insulin resistance, but no change in body weight, waist circumference or BMI.
• Skinner’s 2018 meta-analysis in the Journal of Cachexia, Sarcopenia and Muscle found a 5.7 percent increase in fat-free mass with intramuscular testosterone versus 1.7 percent with transdermal.
• The TRAVERSE diabetes substudy (Bhasin 2024, JAMA Internal Medicine) found no difference in progression to diabetes, and its authors state that testosterone alone should not be used to prevent or treat diabetes.

What did the T4DM trial find about testosterone and blood sugar?

T4DM is the strongest single result in this area. Wittert’s 2021 trial in Lancet Diabetes and Endocrinology was a randomized, double-blind, placebo-controlled, 2-year, phase 3b study in 1,007 men aged 50 to 74. At two years, a 2-hour glucose of 11.1 mmol/L or higher on an oral glucose tolerance test was present in 87 of 413 men (21 percent) on placebo and 55 of 443 (12 percent) on testosterone, a relative risk of 0.59 (95% CI 0.43 to 0.80, p = 0.0007).

That is a 41 percent relative reduction in the proportion of men meeting the diabetes glucose threshold. It was measured on top of a lifestyle program, not instead of one.

Two details that change how the result should be read

Every participant in T4DM was enrolled in a lifestyle program for the duration of the trial. The comparison is therefore testosterone plus lifestyle change against lifestyle change alone, not testosterone against doing nothing.

The second detail is the population. Wittert’s participants were selected to be without pathological hypogonadism. The trial was not testing treatment of a diagnosed deficiency, and its result should not be read as though it were.

How much did T4DM change 2-hour glucose?

By three quarters of a millimole per litre more than placebo. The mean change in 2-hour glucose was -0.95 mmol/L on placebo and -1.70 mmol/L on testosterone, a difference of -0.75 mmol/L (95% CI -1.10 to -0.40, p less than 0.0001).

Both groups improved, which is what a lifestyle program is expected to do. The testosterone group improved further.

What do meta-analyses show for HbA1c and insulin resistance?

They show a reduction, with two different estimates of its size, and the honest thing is to show both. Zhang’s 2018 meta-analysis in World Journal of Urology pooled 8 randomized trials and 596 hypogonadal men with type 2 diabetes and reported HbA1c -0.45 percent (95% CI -0.73 to -0.16), HOMA-IR -0.79 (95% CI -1.23 to -0.34), fasting glucose -0.98 (95% CI -1.13 to -0.54) and fasting insulin -2.47 (95% CI -3.99 to -0.95).

Li’s 2020 meta-analysis in the International Journal of Endocrinology pooled 18 randomized trials and 1,415 patients (767 treated, 648 control) with type 2 diabetes or metabolic syndrome and reported larger figures: HbA1c -0.67 (95% CI -1.35 to -0.19, p = 0.006) and HOMA-IR -1.94 (95% CI -2.65 to -1.23, p less than 0.0001).

Zhang’s authors note that effects were inconsistent across the individual studies that went into the pooled figure. Two analyses of overlapping literature producing estimates that differ by about half an HbA1c point is a reason to treat the exact number as uncertain, even though the direction is consistent.

Does testosterone change fat mass and lean mass?

Yes, in randomized trials. Corona’s 2015 meta-analysis in the European Journal of Endocrinology pooled 59 randomized controlled trials with 3,029 treated participants and 2,049 controls, and found a significant reduction in fat mass, an increase in lean mass, and reductions in fasting glucose and insulin resistance. The glucose effect was larger in younger men and in men with metabolic disease.

That same analysis found no significant change in body weight, waist circumference or BMI, and no improvement in HDL cholesterol or blood pressure. Fat mass falling while lean mass rises and total weight stays the same is internally consistent, and it is the reason a scale is a poor way to judge this treatment.

Does it change body weight and waist size?

The observational data say yes, the randomized data say no, and the label on the source matters. Corona’s 2016 meta-analysis in the Journal of Endocrinological Investigation pooled 32 observational and registry studies covering 4,513 patients with a mean age of 51.7 years, and reported a weight change at 24 months of -3.50 kg (95% CI -5.21 to -1.80) and a waist circumference change of -6.23 cm (95% CI -7.94 to -4.76), with reduced fat mass, increased lean mass and reduced fasting glucose and insulin resistance.

These are observational figures, not trial results. The same authors note that placebo-controlled trial data on weight are more conflicting and that the finding needs a purpose-built randomized trial. Li’s 2020 randomized-trial analysis did report body weight -3.91 kg (95% CI -4.14 to -3.69) and waist circumference -2.80 cm (95% CI -4.38 to -1.21, p = 0.0005) in men with type 2 diabetes or metabolic syndrome, which sits against Corona 2015’s null result on the same outcomes in a broader trial population.

Does the route of administration matter for lean mass and strength?

It appears to. Skinner’s 2018 meta-analysis in the Journal of Cachexia, Sarcopenia and Muscle covered 31 randomized trials for fat-free mass (1,213 treated, 1,168 placebo) and 17 trials for strength (2,572 treated, 2,523 placebo). It found increases in fat-free mass (effect size 1.20, 95% CI 0.91 to 1.49), total body strength (0.90, 95% CI 0.67 to 1.14), lower-extremity strength (0.77, 95% CI 0.45 to 1.08) and upper-extremity strength (1.13, 95% CI 0.78 to 1.47), all at p less than 0.001.

Broken down by route, intramuscular testosterone gave a 5.7 percent increase in fat-free mass and 10 to 13 percent increases in total body strength, while transdermal gave a 1.7 percent increase in fat-free mass. The authors also report that transdermal produced no change in lower-extremity strength versus placebo (effect size 0.26, 95% CI -0.19 to 0.70, p = 0.26). Choice of route is a clinical decision, not a self-selected one, and it belongs in a conversation with your doctor or pharmacist.

Who does this evidence apply to?

It applies to men with confirmed low testosterone, and it does not apply to men with normal testosterone levels. Health Canada’s product monograph for oral testosterone undecanoate capsules, authorized 12 December 2025, requires deficiency to be “clearly demonstrated by clinical features and confirmed by two separate validated biochemical assays (morning testosterone),” and states the product should not be used “to treat non-specific symptoms suggestive of hypogonadism if testosterone deficiency has not been demonstrated.” What testing your own case needs is a decision for your doctor or nurse practitioner.

T4DM is the exception worth naming, because its participants were selected to be without pathological hypogonadism. That is why its glucose result is best described as a finding about metabolic risk in that trial’s population, rather than as a benefit of treating a confirmed deficiency. How the diagnosis is actually made in Canada is covered in how low testosterone is diagnosed, and the other measured effects are collected in what testosterone therapy does, by the numbers.

What the evidence does not show

The largest cardiovascular outcomes trial did not show diabetes prevention. The TRAVERSE diabetes substudy (Bhasin 2024, JAMA Internal Medicine) covered 1,175 men with prediabetes and 3,880 men with diabetes, and found that progression to diabetes did not differ between the testosterone and placebo groups. The authors state that testosterone alone should not be used to prevent or treat diabetes. T4DM showed prevention in men selected for impaired glucose tolerance and enrolled in a lifestyle program; TRAVERSE, in a different population without that program, did not. The metabolic benefit is real but conditional, and both results have to be read together. The wider trial is covered in the TRAVERSE explainer.

T4DM also produced a safety signal that has to be reported alongside its glucose result. A hematocrit above 54 percent occurred in 106 of 491 men (22 percent) on testosterone versus 6 of 484 (1 percent) on placebo, and the authors describe increases in hematocrit as potentially treatment limiting. That is the single most common reason treatment is stopped or adjusted, and it is why hematocrit appears on every monitoring schedule. What gets checked and when is set out in monitoring on testosterone therapy.

The randomized body composition data are narrower than the marketing version of this topic suggests. Corona 2015, across 59 randomized trials, found no significant change in body weight, no change in waist circumference, no change in BMI, no improvement in HDL cholesterol and no improvement in blood pressure. The weight and waist figures that circulate most widely, -3.50 kg and -6.23 cm at 24 months, come from Corona’s 2016 analysis of 32 observational studies, and the same authors say a purpose-built randomized trial is needed.

Two further limitations belong here. Zhang’s 2018 authors report that the glycaemic effects were inconsistent across the individual studies. Li’s 2020 analysis found that erectile function measured by the IIEF-5 did not improve, even though Aging Males’ Symptoms scores did; that finding sits with the rest of the sexual function evidence in testosterone and erectile function.

Frequently asked questions

Does testosterone therapy lower blood sugar?

In men with confirmed low testosterone it can, but the size depends on the population. Zhang’s 2018 meta-analysis in World Journal of Urology, covering 8 trials and 596 hypogonadal men with type 2 diabetes, reported an HbA1c reduction of 0.45 percent (95% CI -0.73 to -0.16). Li’s 2020 meta-analysis of 18 trials reported 0.67 (95% CI -1.35 to -0.19).

What did the T4DM trial actually show?

T4DM (Wittert 2021, Lancet Diabetes and Endocrinology) randomized 1,007 men aged 50 to 74, all enrolled in a lifestyle program. At two years, a 2-hour glucose of 11.1 mmol/L or higher occurred in 87 of 413 men (21 percent) on placebo and 55 of 443 (12 percent) on testosterone, relative risk 0.59 (95% CI 0.43 to 0.80, p = 0.0007).

Should testosterone be used to prevent or treat type 2 diabetes?

No. The TRAVERSE diabetes substudy (Bhasin 2024, JAMA Internal Medicine), covering 1,175 men with prediabetes and 3,880 with diabetes, found that progression to diabetes did not differ between the testosterone and placebo groups. The authors state that testosterone alone should not be used to prevent or treat diabetes. Speak with your doctor or pharmacist about diabetes management.

Does testosterone cause weight loss?

The randomized evidence does not show it. Corona’s 2015 meta-analysis in the European Journal of Endocrinology, covering 59 randomized trials, found reduced fat mass and increased lean mass but no significant change in body weight, waist circumference or BMI. The weight and waist reductions of -3.50 kg and -6.23 cm at 24 months come from Corona’s 2016 analysis of 32 observational studies, not from trials.

Why does hematocrit matter on testosterone therapy?

Because it rises. In T4DM (Wittert 2021), a hematocrit above 54 percent occurred in 106 of 491 men (22 percent) on testosterone versus 6 of 484 (1 percent) on placebo, and the authors describe the increase as potentially treatment limiting. The European Association of Urology guideline lists a hematocrit of 54 percent or above as a contraindication.

Does the route of administration matter?

For muscle it appears to. Skinner’s 2018 meta-analysis in the Journal of Cachexia, Sarcopenia and Muscle found intramuscular testosterone gave a 5.7 percent increase in fat-free mass and 10 to 13 percent increases in total body strength, while transdermal gave a 1.7 percent increase in fat-free mass and no change in lower-extremity strength versus placebo.

Does this evidence apply to men whose testosterone is normal?

No. The randomized evidence on glycaemic and body composition outcomes comes from men with confirmed low testosterone. Health Canada’s Information Update of 15 July 2014 states that testosterone products should not be used in men for non-specific symptoms if laboratory tests have not confirmed a low testosterone level and other causes have not been excluded.

References

1. Wittert G, et al. 2021. Testosterone treatment to prevent or revert type 2 diabetes in men enrolled in a lifestyle programme (T4DM): a randomised, double-blind, placebo-controlled, 2-year, phase 3b trial. Lancet Diabetes and Endocrinology. PMID 33338415. DOI: https://doi.org/10.1016/S2213-8587(20)30367-3
2. Zhang J, et al. 2018. Effects of testosterone supplement treatment in hypogonadal adult males with T2DM: a meta-analysis and systematic review. World Journal of Urology. PMID 29511802. DOI: https://doi.org/10.1007/s00345-018-2256-0
3. Li SY, et al. 2020. Metabolic Effects of Testosterone Replacement Therapy in Patients with Type 2 Diabetes Mellitus or Metabolic Syndrome: A Meta-Analysis. International Journal of Endocrinology. PMID 33061966. DOI: https://doi.org/10.1155/2020/4732021
4. Corona G, et al. 2015. THERAPY OF ENDOCRINE DISEASE: Testosterone supplementation and body composition: results from a meta-analysis study. European Journal of Endocrinology. PMID 26537862. DOI: https://doi.org/10.1530/EJE-15-0262
5. Corona G, et al. 2016. Testosterone supplementation and body composition: results from a meta-analysis of observational studies. Journal of Endocrinological Investigation. PMID 27241317. DOI: https://doi.org/10.1007/s40618-016-0480-2
6. Skinner JW, et al. 2018. Muscular responses to testosterone replacement vary by administration route: a systematic review and meta-analysis. Journal of Cachexia, Sarcopenia and Muscle. PMID 29542875. DOI: https://doi.org/10.1002/jcsm.12291
7. Bhasin S, et al. 2024. Effect of Testosterone Replacement Therapy on Progression to Diabetes. JAMA Internal Medicine. PMID 38315466. DOI: https://doi.org/10.1001/jamainternmed.2023.7862
8. Lincoff AM, et al. 2023. Cardiovascular Safety of Testosterone-Replacement Therapy. New England Journal of Medicine. PMID 37326322. DOI: https://doi.org/10.1056/NEJMoa2215025
9. Bhasin S, et al. 2018. Testosterone Therapy in Men With Hypogonadism: An Endocrine Society Clinical Practice Guideline. Journal of Clinical Endocrinology and Metabolism. PMID 29562364. DOI: https://doi.org/10.1210/jc.2018-00229
10. European Association of Urology. Guidelines on Sexual and Reproductive Health, male hypogonadism chapter. 2026 edition, limited update March 2026. https://uroweb.org/guidelines/sexual-and-reproductive-health/chapter/male-hypogonadism
11. Health Canada. Product monograph, oral testosterone undecanoate capsules. Authorized 12 December 2025. https://pdf.hres.ca/dpd_pm/00082762.PDF
12. Health Canada. Information Update, Possible cardiovascular problems associated with testosterone products. 15 July 2014. https://recalls-rappels.canada.ca/en/alert-recall/information-update-possible-cardiovascular-problems-associated-testosterone-products

This page is educational information, not medical advice. Testosterone is a prescription medication and a controlled substance in Canada. Talk to your doctor or pharmacist about your own situation.