At the American Urological Association 2026 Annual Meeting (May 15-18, Washington D.C.), a retrospective cohort of more than 1,600 men reframed what GLP-1 drugs do to testosterone. Earlier research said the testosterone gain on semaglutide and tirzepatide was a side effect of weight loss. The AUA 2026 data say it is not — at least not entirely.
Median total testosterone rose from 320 ng/dL to 419 ng/dL after starting a GLP-1 receptor agonist. Median free testosterone rose from 9 ng/dL to 10.4 ng/dL. Critically, the increase was independent of age and baseline BMI, and the change in testosterone did not align directly with the change in BMI. "The GLP-1s are somehow improving testosterone levels outside of weight loss," said Tobias Köhler of Mayo Clinic, summarizing the finding.
For men weighing GLP-1 against TRT, this changes the math.
Key Takeaways
AUA 2026 retrospective cohort of 1,600+ men: GLP-1s raise total and free testosterone after starting therapy
Median total testosterone +99 ng/dL (320 -> 419 ng/dL); median free testosterone +1.4 ng/dL
Increases were independent of age and baseline BMI and did not track tightly with BMI change
Suggests a direct hormonal effect beyond weight loss alone
Holds across semaglutide, dulaglutide, liraglutide, and tirzepatide — class effect, not molecule-specific
Strengthens the GLP-1-instead-of-TRT case for obese or diabetic men with mild-to-moderate low T
The cohort was a retrospective electronic-health-record analysis of more than 1,600 men prescribed a GLP-1 receptor agonist — semaglutide, dulaglutide, liraglutide, or tirzepatide — for type 2 diabetes or obesity. Researchers pulled paired pre- and post-treatment testosterone labs and looked at the change.
The headline numbers:
Median total testosterone: 320 ng/dL at baseline -> 419 ng/dL post-treatment (median gain: 99 ng/dL)
Median free testosterone: 9 ng/dL at baseline -> 10.4 ng/dL post-treatment (median gain: 1.4 ng/dL)
Statistical significance: Both gains were significant at p < 0.001
Independent of age and baseline BMI: The effect did not require a particular age window or starting body weight
Not aligned with BMI change: The testosterone gain did not track tightly with how much weight each man lost
That last point is the news. Prior work — including a 2026 systematic review of 639 men in The Journal of Sexual Medicine — attributed the testosterone benefit of GLP-1 to weight loss alone. The proposed mechanism: lose fat, reduce aromatase, reduce estradiol, release the brake on LH and FSH, restore testicular function. That model predicts the testosterone gain should scale with weight loss. The AUA 2026 cohort says it does not — or at least not entirely. Some other mechanism is doing work in parallel.
Why the Independent-of-Weight-Loss Finding Matters
The weight-loss-mediated model implied that lean men with normal BMI would not benefit from GLP-1 for low T. The AUA 2026 data leave that open. If part of the effect is direct — at the hypothalamus, pituitary, or Leydig cells — it could extend to men who are not significantly overweight, or to men who plateau on weight loss but still see testosterone gains.
The research team did not propose a specific mechanism. Candidates being investigated in animal and human pharmacology work include:
Direct GLP-1 receptor expression on hypothalamic neurons that regulate GnRH release
Reduced systemic inflammation beyond what weight loss alone produces — GLP-1s have anti-inflammatory effects independent of weight
Improved insulin sensitivity at the hypothalamus, restoring kisspeptin signaling
Direct testicular effects — GLP-1 receptors have been identified on Leydig cells in some preclinical studies
None of these mechanisms is fully established. What is established is that the testosterone change does not cleanly track weight loss, which means the simple model is wrong or incomplete.
How This Updates the GLP-1-vs-TRT Decision
For the past two years, the standard framing was: TRT for primary hypogonadism, GLP-1 for obesity-related (functional) low T because weight loss restores the axis. The AUA 2026 data soften the second clause.
Three scenarios shift:
Scenario 1: Obese man, moderate low T, modest weight loss expected. Before AUA 2026, the case for GLP-1 over TRT depended on hitting meaningful weight loss (10 percent or more) to drive a meaningful testosterone gain. Now, even men who lose less weight may see useful testosterone increases — making GLP-1 a reasonable first move before committing to lifelong TRT.
Scenario 2: Overweight diabetic man already on a GLP-1, considering adding TRT. If you started semaglutide or tirzepatide for diabetes and your testosterone has been rising independently, you may not need TRT layered on top. Recheck labs at 6 and 12 months before adding exogenous testosterone.
Scenario 3: Lean man with low T and metabolic features. The independent-of-BMI signal opens a thin case for considering GLP-1 in men with normal BMI who have metabolic syndrome or insulin resistance. This is off-label and outside current guideline language, but the AUA 2026 cohort included a wide BMI range, and the effect did not require obesity.
The median gain in the AUA 2026 cohort was 99 ng/dL of total testosterone. That number deserves context.
For comparison:
Men starting from 280 ng/dL move to roughly 380 ng/dL — still well within the diagnostic gray zone but symptomatic for many
Men starting from 320 ng/dL move to roughly 420 ng/dL — into the lower-normal range
Men starting from 400 ng/dL move to roughly 500 ng/dL — mid-normal
A 100 ng/dL shift can be the difference between hypogonadal symptoms and no symptoms for men in the borderline range. For severely hypogonadal men (under 200 ng/dL), a 100 ng/dL gain on GLP-1 likely will not get them to where they need to be — TRT remains the better tool.
This is why the AUA 2026 data fit best in the borderline-to-moderate hypogonadism population. If your total testosterone is 250 to 400 ng/dL, you have metabolic features, and you want to preserve fertility, GLP-1 is now a more credible first step than it was last year. See the normal vs optimal testosterone levels and total vs free testosterone guides for how to read these numbers in context.
Free Testosterone Moved Less
The free testosterone gain — 1.4 ng/dL — is smaller in absolute terms than the total testosterone gain, but the relative jump (from 9 to 10.4 ng/dL, roughly 15 percent) is consistent with the total testosterone change after accounting for SHBG. This matters because free testosterone is the active fraction. A total testosterone gain that does not translate to free testosterone gain is a sign that SHBG is binding the increase. The AUA 2026 cohort showed both moved together, which is what you want to see.
The free testosterone calculator (Vermeulen) is the standard tool for estimating free T from total, SHBG, and albumin — useful for tracking your own response on a GLP-1.
Fertility Preservation Is Still the Strongest GLP-1 Advantage
Even before the AUA 2026 cohort, the most defensible case for GLP-1 over TRT was fertility. Conventional TRT shuts down LH and FSH at the pituitary, which crashes spermatogenesis. Recovery takes 6 to 24 months after stopping, and is not always complete.
GLP-1s do the opposite. The published data — admittedly small — show preserved or increased LH and FSH on GLP-1, and either stable or improved semen parameters. A 24-week head-to-head trial of semaglutide versus testosterone undecanoate in 25 men with type 2 diabetes, obesity, and functional hypogonadism showed:
Semaglutide: stable sperm concentration, improved sperm morphology, modest testosterone gain, significant weight loss
Testosterone undecanoate: larger testosterone gain, decreased sperm count and total sperm number, no weight change
For men under 40 who want kids in the next few years, the choice is clear. The AUA 2026 data make GLP-1 more attractive even for men who do not care about fertility, but the fertility-preservation argument is independent and still strong.
1. Retrospective design. Men were not randomized to GLP-1 vs control. Confounding by diet, exercise, and other lifestyle changes during GLP-1 treatment is real and not fully controlled.
2. No long-term follow-up. The testosterone gains were measured within 6 to 12 months of starting GLP-1. Whether the gains persist at 3 or 5 years is unknown.
3. Mechanism is unproven. "Independent of BMI change" is a statistical statement, not a mechanistic one. The biological pathway driving the weight-loss-independent effect has not been confirmed in humans.
4. No symptom data. The cohort tracked lab values. Whether the testosterone increase translated into improved libido, energy, mood, or erectile function was not the primary endpoint.
5. No comparison arm. There was no parallel TRT or placebo group. The 99 ng/dL gain is in absolute terms, not relative to what comparable men would have seen without GLP-1.
These limitations do not invalidate the finding. They constrain how strongly the conclusion can be stated. The right summary is: GLP-1 raises testosterone, the effect is real, the magnitude is meaningful for borderline-hypogonadal men, and at least some of the effect is not explained by weight loss alone.
How Clinics Are Likely to Respond
The AUA 2026 cohort lands in the middle of a broader pivot in men's health: telehealth platforms that treated GLP-1 and TRT as parallel services are starting to integrate them. Hims & Hers expanded both verticals in 2025 and 2026. Peter MD and similar clinics increasingly offer metabolic care alongside hormone optimization.
For a man with metabolic features and borderline testosterone, the right clinic should:
Test metabolic markers — HbA1c, fasting insulin, lipids — before reaching for TRT
Offer GLP-1 as a first-line option if BMI is elevated or HbA1c is borderline
Recheck testosterone at 6 and 12 months on GLP-1 before considering TRT
Layer TRT only if symptoms persist after GLP-1 has done its job
Track LH, FSH, and SHBG, not just total testosterone
Total T 250 to 400 ng/dL and BMI over 30: Ask your clinician about a GLP-1 trial before starting TRT. The AUA 2026 data support this as a credible first move.
Total T 250 to 400 ng/dL and BMI 25 to 30 with metabolic features: The case for GLP-1 is thinner but reasonable, especially if you want to preserve fertility.
Total T under 200 ng/dL or you have primary hypogonadism (high LH, low T): GLP-1 will not solve this. TRT is the right tool.
Already on a GLP-1: Recheck testosterone at 6 and 12 months. If it has risen into the symptom-free range, you may not need TRT layered on.
The decision is not all-or-nothing. Many men benefit from GLP-1 first, weight loss and testosterone improvement together, and TRT only if symptoms persist.
Where This Sits in the Bigger Picture
The AUA 2026 cohort joins a growing pile of 2025-2026 evidence that the boundary between metabolic and hormonal medicine is blurring:
The AUA 2026 GLP-1 cohort adds a new tile: for the subset of men with metabolic-driven low T, you may not need TRT at all — or you may need less of it.
Köhler T et al. Retrospective cohort analysis of GLP-1 receptor agonist therapy and serum testosterone in more than 1,600 men. AUA 2026 Annual Meeting, Washington D.C., May 15-18, 2026.
GLP-1 RAs May Independently Boost Testosterone Levels. Medscape coverage of AUA 2026 abstract, May 2026.
Sansone A et al. GLP-1 receptor agonists and testosterone in men with obesity and type 2 diabetes: a systematic review. The Journal of Sexual Medicine, 2026.
Portillo Canales S et al. Anti-obesity medications can normalize testosterone levels in men with obesity or type 2 diabetes. ENDO 2025, San Francisco.
Frequently Asked Questions
What did the AUA 2026 GLP-1 testosterone study actually find?
A retrospective cohort of more than 1,600 men treated with GLP-1 receptor agonists showed significant increases in both total and free testosterone after starting therapy. Median total testosterone rose from 320 ng/dL to 419 ng/dL — a 99 ng/dL gain. Median free testosterone rose from 9 ng/dL to 10.4 ng/dL. The increases were independent of patient age and baseline BMI, and the changes did not align directly with change in BMI. Investigators concluded that GLP-1 drugs may have a hormonal effect beyond weight loss alone. The data were presented at the American Urological Association 2026 Annual Meeting (May 15-18, Washington D.C.).
How is this different from prior GLP-1 testosterone research?
Earlier evidence — including a 2026 systematic review of 639 men — attributed the testosterone gain on GLP-1 to weight loss. Lose adipose tissue, reduce aromatization, restore HPG-axis signaling, testosterone rises. The AUA 2026 cohort breaks that link. The testosterone increase was not tightly coupled to BMI change, which means another mechanism is also at work — possibly direct GLP-1 receptor effects on the hypothalamus, pituitary, or Leydig cells. The mechanism is still being worked out, but the clinical signal is real. See our prior coverage of the [GLP-1 testosterone obese men 2026 study](/treatment/glp1-testosterone-obese-men-2026-study) for the weight-loss-mediated mechanism.
Should I take a GLP-1 instead of TRT?
It depends on why your testosterone is low. If you have obesity, type 2 diabetes, or functional hypogonadism (testes work, brain signaling is suppressed), a GLP-1 can be a legitimate alternative to TRT and preserves fertility. If you have primary hypogonadism (testicular failure), GLP-1 will not fix the underlying problem. The AUA 2026 data add weight to the GLP-1 option even when the weight loss is modest. Talk to a clinic that handles both metabolic and hormone care — compare options at [/clinics](/clinics?from=glp1-testosterone-independent-weight-loss-aua-2026).
What is the magnitude of testosterone gain on GLP-1?
In the AUA 2026 cohort, median total testosterone rose by 99 ng/dL (320 to 419 ng/dL) and median free testosterone rose by 1.4 ng/dL (9 to 10.4 ng/dL). That is enough to move many men out of the symptomatic-low range and into the lower-normal range, but it does not match what most men get on weekly testosterone cypionate injections. For severely hypogonadal men, TRT will produce a larger and faster response. For mild-to-moderate hypogonadism — especially in the obesity-related pattern — GLP-1 may be sufficient.
Does GLP-1 preserve fertility better than TRT?
Yes. Conventional TRT suppresses LH and FSH at the pituitary, which shuts down spermatogenesis. Recovery after stopping TRT takes 6 to 24 months and is not guaranteed. GLP-1 receptor agonists do the opposite: they either preserve or increase LH and FSH while raising testosterone, which keeps the testes producing sperm. A small 24-week head-to-head trial showed semaglutide improved sperm morphology while testosterone undecanoate worsened sperm count. See the [TRT fertility complete guide](/treatment/trt-fertility-complete-guide) and [HCG for fertility on TRT](/treatment/hcg-for-fertility-on-trt) for fertility-preserving protocols if you take the TRT path.
Which GLP-1 drug raises testosterone the most?
The AUA 2026 cohort did not isolate one drug as superior. The effect held across semaglutide, dulaglutide, liraglutide, and tirzepatide, with most of the testosterone gain seen in the first 6 to 12 months of treatment. The signal is consistent with a class effect rather than a molecule-specific one. That said, tirzepatide produces the largest weight loss in head-to-head obesity trials, which historically has correlated with the largest testosterone gain. If the AUA 2026 finding holds — that part of the effect is weight-loss-independent — drug choice may matter less than getting on a GLP-1 in the first place.
Will a GLP-1 prescription cover the testosterone benefit?
No. GLP-1 receptor agonists are FDA-approved for type 2 diabetes and obesity. The testosterone benefit is off-label observational data. Insurance will pay for the GLP-1 if you meet the diabetes or obesity criteria, but you cannot get one prescribed for low testosterone specifically. The men in the AUA 2026 cohort were treated for metabolic indications, and the testosterone change was an incidental finding. If you have both metabolic and hormone concerns, a clinic that handles both lets you stack the benefits — see [/clinics](/clinics?from=glp1-testosterone-independent-weight-loss-aua-2026).
