Testosterone and Muscle Growth: What Human Trials Show

The 1996 trial: testosterone and training together

Bhasin and colleagues randomized 43 healthy men to testosterone or placebo, with or without supervised weight training. The experimental exposure was 600 mg of testosterone enanthate weekly for 10 weeks; training occurred three times weekly.^[1]

The combined group also increased bench-press strength by 22 ± 2 kg and squat capacity by 38 ± 4 kg. MRI documented greater muscle size. These are changes within that group, not placebo-adjusted effects. Standard errors describe uncertainty in the estimated mean, not the range of individual responses.^[1]

Testosterone also increased muscle size and strength in the non-training comparison. That finding demonstrates an anabolic effect under the experiment’s conditions. It does not show that training is unnecessary or that 6.1 kg was entirely new contractile muscle. Ten weeks also cannot establish long-term safety.^[1]

Read the 1996 study summary

The 2001 dose-response study: a gradient, not a calculator

In 61 healthy men aged 18–35, researchers suppressed endogenous testosterone with a GnRH agonist, then assigned one of five weekly testosterone-enanthate doses for 20 weeks. Food energy and protein intake were standardized. The three groups below had the fat-free-mass gains reported numerically in the abstract.^[2]

View the chart data

Bhasin et al. (2001), 20 weeks

Weekly exposure	Mean nadir testosterone	Fat-free mass change
125 mg	542 ng/dL	+3.4 kg
300 mg	1,345 ng/dL	+5.2 kg
600 mg	2,370 ng/dL	+7.9 kg

These nadir blood levels were group means, not treatment targets. The findings cannot be read as “add this dose to your natural production.” Nor can the line between studied groups reliably predict an untested dose, a different formulation or a longer duration.^[2]

Read the dose-response study summary

What do replacement-style trials show?

The 2013 Hildreth trial randomized 167 older men with baseline testosterone of 200–350 ng/dL to gel or placebo and resistance training or no training for 12 months. Of these, 143 completed the study. Testosterone improved body composition, but among those doing resistance training it did not add a significant strength or functional advantage over placebo. In non-exercisers, upper-body strength improved.^[5]

This is a useful counterweight to assuming that a better body-composition result must produce better performance. It also shows why the large high-dose gains in healthy young men cannot be used as an expected TRT result.

Longer follow-up adds another perspective. The three-year TEAAM muscle report found modest improvements in selected strength, power and lean-mass outcomes in older men. The investigators left the clinical importance and effect on disability unresolved. A statistically detectable difference is not automatically a noticeable improvement in everyday life.^[6]

In the older-men graded-dose experiment, anabolic responsiveness was retained, but hemoglobin increases and adverse effects were more concerning. The comparison used earlier young-men data rather than randomly assigning people to age groups. Age can change the risk profile even when muscle remains responsive.^[4]

Newer evidence: mobility in 2024, muscle metabolism in 2026

A 2024 trial studied 148 men aged at least 70 with mobility problems and low to low-normal testosterone. Its four groups tested testosterone undecanoate, a training-and-nutrition program, their combination, or control over 20 weeks. The combined intervention included protein, calcium and vitamin D.^[8]

The comparison favors the combined program, but it should not be used to promise a similar benefit from testosterone alone. The four-arm design and the contributions of training and nutrition matter.^[8]

A 2026 LITROS report examined muscle metabolites in 44 of 83 randomized older men with obesity, hypogonadism and frailty. During 26 weeks of lifestyle therapy, adding testosterone changed glycolysis-related metabolites. This helps investigate a mechanism; it does not establish a new hypertrophy dose, quantify extra muscle for young lifters, or prove that the metabolite changes caused clinical improvement.^[9]

Is the gain real muscle, or just water?

The evidence supports genuine hypertrophy, while also showing why body-composition numbers need care. A related analysis obtained muscle biopsies from 39 men in the dose-response research. Larger muscle volumes were accompanied by increases in muscle-fiber area and myonuclear number. It is related evidence from the same research population, not another independent trial confirming the entire result.^[3]

That makes “it was all water” too dismissive. But calling every kilogram of fat-free mass a kilogram of new muscle is also unjustified. A strong reading considers whole-body composition alongside imaging, biopsy and strength results. None is a complete substitute for the others.

For your own interpretation of a study, match the wording to the measurement: “fat-free mass increased” when that is what was measured, “muscle volume increased” when imaging supports it, and “strength improved” when a performance test actually changed.

Read the muscle-fiber analysis

Lean mass, muscle size and contractile tissue are different measurements

Fat-free mass is everything outside the fat compartment, including muscle, water, organs and bone mineral. DXA lean soft tissue excludes bone mineral but still does not isolate skeletal muscle. Publications sometimes use “lean body mass” loosely, so the methods section matters more than the headline. A scan can accurately detect a change in its measured compartment while still being unable to tell you how much new contractile protein was built.

Glycogen, creatine and water can move the result

Muscle glycogen is stored carbohydrate, and changes in fuel storage accompany changes in tissue water. Creatine loading can also alter water content. Those are real changes in the body, not automatically new contractile tissue. They can affect body-composition estimates over a much shorter period than substantial muscle-protein accumulation would be expected.

Bone and colleagues tested glycogen and creatine manipulation in 18 trained male cyclists. Glycogen loading with and without creatine loading increased DXA lean-mass estimates by approximately 3% and 2%, respectively. The authors emphasized that metabolites and water altered the estimates during a period when little muscle-protein change was likely. This illustrates measurement sensitivity; it does not establish that any fixed fraction of a testosterone trial’s gains was water.^[18]

There is no defensible universal subtraction such as “remove 30% from every lean-mass gain to find real muscle.” To separate the components, a study needs additional measurements. Starting creatine, changing carbohydrate intake or moving from a depleted state to a fed, rehydrated state can complicate comparisons. The correct response is better standardization and more specific outcomes, not dismissing every gain as an artifact.

Even cross-sectional area needs context

Anatomical imaging is more specific to muscle than a whole-body lean-mass estimate, but a larger cross-section is not a direct measurement of newly synthesized myofibrillar protein. A pump, exercise-related edema and changes in tissue hydration can enlarge the measured area. Ultrasound is also sensitive to probe placement and technique; one site may not represent growth along an entire muscle.

In a small study of ten untrained men, Damas and colleagues found that early increases in ultrasound muscle cross-sectional area occurred alongside indicators of swelling and muscle damage. This cautions against calling every early size increase pure hypertrophy. It does not mean that later growth is unreal or that the exact swelling pattern applies to trained people using testosterone.^[19]

A biopsy adds information about individual fibers and their nuclei, as in the testosterone research described above. But it samples a small region. Tissue orientation, sampling and fiber selection influence interpretation. The most convincing case combines sustained anatomical growth with compatible fiber-level findings and functional outcomes, while acknowledging that strength also depends on skill and neural adaptation.^[1][3]

What a stronger before-and-after comparison looks like

As a practical interpretation checklist, look for the same instrument and analysis method, a consistent feeding and hydration state, comparable carbohydrate and creatine exposure, and a standardized interval after the last workout. Check whether the size of the change exceeds the method’s measurement error. These controls improve interpretability; they cannot turn DXA into a direct assay of muscle protein.

For a research claim, also ask whether assessors were blinded, whether the same anatomical region was measured, whether all randomized participants were included and whether the reported outcome was specified before the results were known. “Lean mass increased,” “quadriceps volume increased” and “muscle-fiber area increased” are all useful findings. The mistake is treating them as identical.

Fat loss, recomposition and the percentage trap

Body-fat percentage is a ratio. It can fall because fat mass decreased, because non-fat mass increased, or both. As a simplified arithmetic example, someone with 20 kg of fat at 100 kg body weight has 20% body fat. If fat stays at 20 kg while total weight rises to 105 kg, the percentage falls to about 19%. That is a lower percentage without a kilogram of fat loss. The example describes arithmetic, not an expected treatment response.

For recomposition claims, therefore, read the absolute changes in fat mass and lean mass separately. Ask whether visceral fat was measured directly or whether the study only reported total fat. A change in waist circumference can be informative but cannot establish which fat depot changed. Likewise, a rise in scale weight cannot distinguish muscle, water and fat.

The estradiol experiment discussed in the next chapter is especially relevant because muscle and adiposity need not respond through identical hormonal pathways. More lean tissue and less fat are separate outcomes, and improvement in one does not guarantee improvement in the other. A paper reporting better body composition also needs a performance test before claiming better function.^[5][7]

Why sustained high exposure is a different experiment

Supraphysiologic means above normal physiological exposure. It is not a universal weekly milligram cutoff, and it is not a switch that makes normal-range testosterone irrelevant. In the 2001 experiment, higher assigned doses produced higher mean nadir concentrations and progressively larger fat-free-mass gains. Yet the 125 mg group, with a mean nadir of 542 ng/dL, also gained fat-free mass. That observation alone rules out a simple claim that muscle effects begin only above the reference range.^[2]

At the same time, a nadir is a low-point sample, not the average exposure across the injection interval. Endogenous production was experimentally suppressed, and administration produced a particular exposure pattern. The 542 ng/dL group is not interchangeable with a natural lifter who happens to measure 542 once. Nor can the 1,345 or 2,370 ng/dL group means be converted into personal targets. The whole protocol, not just the laboratory number, defines the experiment.

The strongest distinction is the size and persistence of the intervention. The classic high-dose trials deliberately maintained a substantial pharmacological exposure for weeks. A supplement claim might describe a small average change, an isolated endpoint or an effect in people with a nutritional or hormonal limitation. Those designs do not establish equivalent anabolic conditions. Larger observed gains in high-exposure experiments also do not imply that escalating exposure is a favorable health tradeoff.^[1][2][4]

Total testosterone, free testosterone and timing

Total testosterone includes hormone bound to proteins as well as the small unbound fraction. Changes in sex hormone-binding globulin can complicate interpretation of total testosterone. The clinical guideline recommends appropriate free-testosterone assessment in selected situations, such as altered SHBG or borderline total results. Neither total nor free testosterone, however, is a validated calculator for kilograms of muscle gained.^[12]

For endogenous production, morning sampling and confirmation of a low result matter. For administered testosterone, timing relative to treatment also matters. Comparing a post-dose peak with another person’s trough can make two exposure patterns appear more different, or more similar, than they really are. A well-reported experiment identifies the assay, sampling schedule and achieved concentrations instead of treating every blood result as interchangeable.

The post-workout hormone spike is another separate question

West and colleagues compared arm training performed with differing exercise-induced systemic hormone responses. The higher acute hormonal response did not enhance elbow-flexor hypertrophy or strength. Brief elevations after exercise therefore cannot be assumed to reproduce the effects of sustained pharmacological exposure.^[17]

This also explains why “this workout raises testosterone” is an incomplete reason to choose it. A program should demonstrate useful training outcomes. Evidence against a benefit from short-lived hormonal spikes does not negate the high-dose trials, and those trials do not validate workouts or supplements merely because they change a hormone measurement.

Testosterone is not the whole body-composition story

In a 16-week suppression-and-replacement experiment, Finkelstein and colleagues studied testosterone with or without blocking conversion to estradiol. Androgen deficiency primarily explained losses in lean mass, muscle size and strength, while estrogen deficiency primarily explained increases in body fat. Both hormones contributed to sexual-function changes.^[7]

This helps explain why the same intervention can affect muscle and fat through different pathways. It does not establish a universal estradiol target or justify suppressing estrogen to improve physique. The experiment deliberately altered hormone production and conversion, so its design needs to stay attached to its conclusions.^[7]

Why a testosterone booster is not automatically a muscle builder

A label promising a testosterone increase of 150 or 300 ng/dL sounds persuasive because the number is concrete. But there are three separate claims to test: does the product reliably raise testosterone, does it improve muscle growth beyond the same training with placebo, and is the hormone change responsible for that benefit? Evidence for the first claim does not establish the other two. The example increases here are hypothetical marketing-style scenarios, not an established average effect of over-the-counter boosters.

A change needs a starting point, a comparator and a duration

An increase from 250 to 550 ng/dL is not the same clinical situation as an increase from 650 to 950 ng/dL. Symptoms, repeated measurements and the cause of a low result matter. Neither arithmetic example predicts a particular muscle gain. Correcting genuine androgen deficiency can improve body composition within a physiological range; it is therefore inaccurate to say that testosterone only affects muscle once it exceeds the laboratory range.^{[5][6][7][12]}

Next ask whether the reported increase is relative to placebo. If a supplement group rises by 150 ng/dL but the placebo group rises by 100, the observed difference in changes is 50, not 150. That is an illustration, not a result from a trial. Confidence intervals, baseline balance and repeated testing help show whether a difference is credible. A single before-and-after blood draw cannot separate a treatment effect from ordinary fluctuation or regression toward the mean.

Blood testosterone is not a direct readout of muscle responsiveness

In a 2018 analysis of 49 resistance-trained young men, circulating hormone measurements did not consistently explain changes in muscle-fiber area or lean mass after 12 weeks of training. Muscle androgen-receptor content was associated with hypertrophy in analyses of the highest and lowest responders. This supports looking beyond a serum number, but it is an association, not proof that increasing receptor content causes growth or that a supplement can do so.^[13]

The result also cannot demonstrate that every sustained 150–300 ng/dL increase is biologically irrelevant. Natural differences among participants and experimentally changing an individual’s exposure are different questions. The defensible conclusion is narrower: a modest rise within the usual range is not enough evidence to promise visible additional hypertrophy, particularly in an already healthy, trained person.

One negative supplement trial, one positive trial

Melville and colleagues tested D-aspartic acid during 12 weeks of supervised resistance training. Nineteen men completed the study. The studied 6 g/day regimen did not improve testosterone or training-related hypertrophy and strength compared with placebo. This challenges the product’s proposed benefit in that setting. It does not test what happens when testosterone actually rises by 300 ng/dL, because the intervention did not produce that hormonal result.^[14]

A small 2015 ashwagandha trial provides a useful counterexample. Fifty-seven men with little resistance-training experience were randomized for eight weeks, and 50 completed the study. The extract group had larger improvements in selected strength and muscle-size measures, alongside a testosterone increase of about 96 ng/dL versus 18 ng/dL with placebo. Thigh-size differences were not significant. These findings should be included rather than dismissed, but they do not show that the testosterone change caused the training benefits. The intervention was a botanical extract, not isolated testosterone, and some size outcomes used anthropometric estimates rather than direct muscle imaging.^[15]

These trials do not justify treating all “boosters” as one class. Different ingredients, extracts, participants and outcome methods can produce different results. A favorable trial for one preparation cannot validate every product sharing its plant name, and a testosterone-only endpoint cannot stand in for a muscle-growth endpoint. The relevant question for a lifter is whether the exact intervention adds meaningful, reproducible tissue growth to otherwise comparable training.

Restoring a deficiency is a different proposition

The classic zinc study included a small group of marginally zinc-deficient older men whose testosterone rose with supplementation. That is evidence about nutritional status and hormones, not proof that zinc builds extra muscle in zinc-replete lifters. It did not establish a hypertrophy benefit or compare supplementation with placebo in that older group.^[16]

The same reasoning applies when evaluating a “testosterone support” claim: identify what was deficient, who was studied and what actually improved. A valid correction of a deficiency can matter without becoming a universal anabolic strategy. Conversely, failure to demonstrate extra muscle growth does not mean that correcting a clinically relevant deficiency has no value.

Muscle benefits and safety need separate evidence

TRAVERSE enrolled 5,246 men with symptomatic hypogonadism and cardiovascular disease or elevated risk. Testosterone gel was titrated to 350–750 ng/dL. Major cardiovascular events occurred in 7.0% with testosterone and 7.3% with placebo, meeting the trial’s noninferiority criterion. The hazard ratio was 0.96, with a 95% confidence interval of 0.78–1.17.^[10]

This is reassuring within the studied replacement setting. It does not establish safety for supraphysiologic exposure or combinations of anabolic drugs. Atrial fibrillation, acute kidney injury and pulmonary embolism occurred more often in the testosterone group. The trial was funded by AbbVie and others.^[10]

In February 2025, FDA announced removal of boxed-warning language about increased cardiovascular outcomes while requiring blood-pressure warnings. Those changes address different evidence and should be read together. A change to one warning is not a declaration that testosterone is risk-free.^[11]

The 2018 Endocrine Society guideline frames treatment around symptoms, consistently low testosterone, evaluation of the cause and monitoring. Fertility plans and contraindications matter. A muscle-building goal alone does not answer that clinical assessment.^[12]

Diagnosis and monitoring guideline · TRAVERSE reading note

Questions to ask before applying a result

Can a study predict gains on 150 or 200 mg per week?

No exact prediction follows from these trials. Those doses are not interchangeable with the studied groups, and an individual’s baseline hormones, achieved exposure, training and measurement method can differ. An interpolated number would look more precise than the evidence allows.

How fast should changes happen?

The research here measured outcomes over 10 weeks, 20 weeks, 12 months and three years in different populations. These are assessment windows, not interchangeable growth rates. Dividing a trial’s total gain by its weeks and extending that rate indefinitely would be misleading.

Does a higher dose mean a better result?

Some anabolic outcomes increased with exposure in controlled experiments. “Better” also includes adverse effects, fertility, function and long-term health. These studies do not identify a universally optimal tradeoff.^[2][4][12]

How much of the gain remains after stopping?

The results highlighted here largely concern active treatment. They do not provide a reliable percentage of retained muscle after withdrawal. That question requires follow-up that measures retained tissue and performance after treatment ends.

Which study should I compare myself with?

Start with the population and purpose, not the largest number. A replacement trial in older men with low testosterone, a frailty trial with nutrition support and a high-dose experiment in healthy men answer different questions. Even the closest match provides a group estimate rather than a personal forecast.

How this guide was assembled

Selected evidence guide, checked September 30, 2026. We examined foundational randomized testosterone studies and searched for newer human resistance-training and muscle reports. This is not a systematic review, exhaustive trial inventory or pooled effect estimate. Numerical trial results were checked against primary publication abstracts; full adverse-event tables and supplements were not independently reviewed for this guide. The FDA communication was read directly. This expanded edition also includes primary research on supplements, circulating hormones and body-composition measurement; selected accessible full-text methods were consulted where stated in the references. Related reports from the same study are identified to avoid counting them as independent replications.

AI-assisted editorial synthesis by PED Evidence, with editorial responsibility described on our About page. No independent clinician review is claimed. Trial funding and conflicts should be checked in each original publication; disclosure review was not comprehensive. Corrections and editorial policy.

1. Bhasin et al. (1996). Testosterone, muscle size and strength in healthy men. DOI: 10.1056/NEJM199607043350101.
   Primary abstract; 43 randomized men; 10 weeks.
2. Bhasin et al. (2001). Testosterone dose-response relationships in healthy young men. DOI: 10.1152/ajpendo.2001.281.6.E1172.
   Primary abstract; 61 randomized men; endogenous production suppressed.
3. Sinha-Hikim et al. (2002). Testosterone-induced muscle fiber hypertrophy. DOI: 10.1152/ajpendo.00502.2001.
   Primary abstract; biopsy analysis from the dose-response research, not an independent replication.
4. Bhasin et al. (2005). Responses to graded testosterone doses in older men. DOI: 10.1210/jc.2004-1184.
   Primary abstract; 60 older men randomized; comparison with earlier young-men data.
5. Hildreth et al. (2013). Testosterone and progressive resistance exercise in older men.
   Primary abstract; 167 randomized men; 143 completed 12 months.
6. Storer et al. (2017). Three years of testosterone, muscle performance and physical function. DOI: 10.1210/jc.2016-2771.
   Primary abstract; secondary muscle/function report from TEAAM.
7. Finkelstein et al. (2013). Gonadal steroids, body composition, strength and sexual function. DOI: 10.1056/NEJMoa1206168.
   Primary abstract; 16-week suppression/replacement experiment, with and without aromatase inhibition.
8. Midttun et al. (2024). Exercise, testosterone, vitamin D, calcium and protein in older men. DOI: 10.1002/jcsm.13498.
   Primary abstract; 148 men with mobility problems; four-group trial.
9. Viola et al. (2026). Testosterone plus lifestyle therapy and skeletal muscle glycolysis. DOI: 10.3389/fendo.2025.1719749.
   Primary abstract; metabolomic substudy in 44 of 83 randomized LITROS participants.
10. Lincoff et al. (2023). Cardiovascular safety of testosterone replacement, TRAVERSE. DOI: 10.1056/NEJMoa2215025.
    Primary abstract; 5,246 men; gel titrated to 350–750 ng/dL.
11. FDA (February 28, 2025). Class-wide testosterone labeling changes.
    Official regulatory communication, including blood-pressure warnings.
12. Bhasin et al. (2018). Endocrine Society hypogonadism guideline. DOI: 10.1210/jc.2018-00229.
    Clinical practice guideline; diagnosis and monitoring context, not a hypertrophy trial.
13. Morton et al. (2018). Muscle androgen receptors, circulating hormones and training-induced hypertrophy.
    Primary publication; 49 trained men, with receptor analyses in selected high and low responders; association, not randomized receptor manipulation.
14. Melville et al. (2017). D-aspartic acid during three months of resistance training.
    Primary publication and accessible methods; 19 completers, 12 weeks; no additional hypertrophy or testosterone benefit.
15. Wankhede et al. (2015). Ashwagandha, muscle strength and recovery.
    Primary publication and accessible methods; small eight-week trial; positive selected outcomes do not establish testosterone mediation.
16. Prasad et al. (1996). Zinc status and serum testosterone.
    Primary abstract; deficiency and supplementation research, not a muscle-growth trial.
17. West et al. (2010). Exercise-induced hormonal elevations and arm hypertrophy.
    Primary publication; acute exercise-related hormone elevations, not sustained exogenous exposure.
18. Bone et al. (2017). Creatine and glycogen manipulation changes DXA body-composition estimates. DOI: 10.1249/MSS.0000000000001174.
    Primary abstract; 18 trained cyclists; short-term substrate and water changes, not a testosterone trial.
19. Damas et al. (2016). Early muscle cross-sectional area increases and edema. DOI: 10.1007/s00421-015-3243-4.
    Primary abstract; ten untrained men; ultrasound and damage markers during ten weeks of training.