Author: drcardona

  • GLP-1s and Muscle: How Much Muscle Is Really Lost, and How to Protect It Beyond Protein

    GLP-1s and Muscle: How Much Muscle Is Really Lost, and How to Protect It Beyond Protein

    Weight Loss & GLP-1s

    The headlines say these drugs “eat muscle.” The trial data say something more precise: the proportion of lean tissue lost is ordinary for weight loss, but the kilograms are not trivial, and what you do about them matters more than protein alone.

    25 to 39%of weight lost on incretin drugs is lean mass in pooled trial data, about 31% overall
    ~½of that lean mass is actually skeletal muscle, so true muscle is roughly 10 to 25% of weight lost
    5 to 6 kgof lean mass lost in the landmark trials, alongside 14 to 22 kg of total weight loss
    +1.1 kgof lean mass saved by resistance training on tirzepatide in the T-REx trial, with strength gains instead of losses
    The short answer

    The ratio is normal. The kilograms are not trivial.

    Two adults doing dumbbell curls together in a gym.
    Two adults doing dumbbell curls together in a gym.

    Lean-mass loss on GLP-1 medicines is real. Across the body-composition studies, roughly a quarter to two-fifths of the weight lost is lean mass rather than fat. But that proportion is not a drug-specific wasting effect: in the largest pooled analysis to date, 20 randomized trials, 15,782 participants, lean mass was 25 to 39% of weight lost on incretin therapies and 26.2% with intensive lifestyle intervention, a difference that was not statistically significant (p = 0.42). The authors’ conclusion was blunt: lean loss is “a physiological consequence of weight reduction, not a disproportionate drug effect.”[1]

    No evidence shows the drugs attack muscle directly. The mechanism is the energy deficit itself, plus appetite suppression and less mechanical loading as body weight falls. On MRI, muscle-volume change on tirzepatide matched the population-expected change for the weight lost.[2]

    So why does this deserve a whole article? Because of scale. A lifestyle program that produces 6 kg of weight loss at a 26% lean share costs about 1.5 kg of lean tissue. A GLP-1 producing 20 kg of loss at the same share costs more than 5 kg, an amount comparable to a decade or more of age-related muscle loss, arriving in a little over a year. For a robust 45-year-old, that is usually manageable. For a 72-year-old who starts with little muscle in reserve, it may not be.

    The honest headline

    Not “GLP-1s eat muscle” and not “muscle loss is a myth.” It is this: the proportion lost is the same as with any effective weight loss; the absolute amount is large because the total weight loss is large, and the people with the least muscle to spare are the ones who need a plan, not just a prescription.

    Section 1 · The numbers

    First, a distinction the headlines skip: lean mass is not muscle

    Almost every alarming number in this debate comes from DXA scans, and DXA does not measure muscle. It divides the body into fat mass, bone mineral, and everything else. That “everything else” is called lean mass or lean soft tissue. It includes skeletal muscle but also water, organs, connective tissue, and the water stored with glycogen.[3]

    The 2025 joint advisory from the American College of Lifestyle Medicine, the American Society for Nutrition, the Obesity Medicine Association, and The Obesity Society estimates that muscle is only about half of lean mass. So when a trial reports that 38% of weight lost was lean, the muscle share is roughly 20%. The headline overstates muscle loss about twofold. Modeled without structured strength training, the advisory puts muscle at about 10 to 15% of weight lost in women and 20 to 25% in men.[4]

    What a DXA calls “lean mass”

    Roughly half of lean soft tissue is skeletal muscle. The rest is water, organs, connective tissue and glycogen-associated water, some of which shrinks appropriately as the body gets smaller.

    What that means for a “38% lean” headline

    STEP-1’s ~38% lean share translates to roughly 20% of total weight lost as actual muscle (about 19% muscle, 19% other lean tissue), still meaningful, but half the headline figure.

    Two more measurement caveats matter. Some early lean loss is glycogen and water, which is partly reversible and is not contractile tissue. And DXA cannot tell whether lean tissue was lost from the limbs, where it affects strength and function, or from the trunk and organs. MRI measures of thigh and appendicular muscle are better, and trials are only now using them.

    STEP-1: semaglutide

    In the DXA substudy of STEP-1 (n = 140), semaglutide 2.4 mg over 68 weeks reduced fat mass by 19.3% and lean body mass by 9.7%, while the lean proportion of the body actually rose by 3.0 percentage points. People became leaner overall even as they lost lean tissue in absolute terms. The joint advisory summarizes the substudy as a 13.6 kg loss made up of 8.3 kg of fat (62%) and 5.3 kg of lean (38%).[4]

    A discrepancy worth knowing about

    A second published summary of the same substudy (Look et al., 2025) reports 10.4 kg of fat and 6.9 kg of lean, 60% fat / 40% lean. The percentages agree (~38 to 40% lean); the kilograms differ, likely because of different samples or estimands. Some reviews quote 45% lean for STEP-1, but that figure traces to a review tabulation rather than the primary paper and could not be reconciled to a primary source. This article uses about 38 to 40% and cites both summaries rather than picking one silently.

    SURMOUNT-1: tirzepatide

    The SURMOUNT-1 DXA substudy (n = 160 with baseline and week-72 scans; 73% women, mean age 46) is the cleanest dataset we have. On tirzepatide, weight fell 21.3%, fat mass fell 33.9% (−15.9 kg), and lean mass fell 10.9% (−5.6 kg). The split of weight lost was 74% fat / 26% lean, essentially identical to placebo’s 75% / 25%, despite roughly four times the total loss. Visceral fat fell 40.1%. The proportions were consistent across age groups (including ≥65 years, at 76% fat), sex, dose, and weight-loss tertiles.[3]

    One caution: STEP-1 and SURMOUNT-1 are different trials, with different populations, scanners and durations. There is no randomized head-to-head DXA trial, so do not read these numbers as proof that one drug is gentler on muscle than the other. The pooled analysis below is the fairer comparison, and even it is indirect.

    Lean mass as a share of weight lost, by approach

    Pooled estimates with 95% confidence intervals, Eisa & Barood, 2026 (20 RCTs, 15,782 participants)

    0% 10% 20% 30% 40% Lifestyle + resistance training 17.5% Tirzepatide 25.4% Lifestyle alone 26.2% Liraglutide 26.8% Semaglutide 35.2%

    Whiskers show 95% confidence intervals. Heterogeneity across trials was substantial (I² = 68%), and a separate meta-analysis of nine placebo-controlled DXA trials (Beavers et al., 2025) put the lean share at about 30% overall.[5] Individual trials range widely, from ~15% or less to 40 to 60% in some studies, depending on population, drug and comorbidity.

    Estimate your own split

    Pick a total weight loss and an approach. The lean share comes from the pooled analysis above; estimated muscle assumes muscle is about half of lean mass, per the joint advisory. This is a population-level model, not a prediction for an individual. Your own split depends on training, protein, rate of loss and starting composition.

    14.9 kgfat lost
    5.1 kglean mass lost (muscle, water, organs, connective tissue)
    ~2.5 kgestimated skeletal muscle lost, roughly half the lean figure

    For context, the advisory’s sex-specific modeling, without structured strength training, puts muscle at about 10 to 15% of weight lost in women and 20 to 25% in men. Resistance training is what moves you toward the bottom row of the chart above.

    Section 2 · The comparison

    GLP-1s are not worse than the alternatives. They are bigger.

    Every method that produces large weight loss takes lean tissue with it. Diet-only weight loss has followed a rough “quarter rule” for decades: about 20 to 30% of the loss is fat-free mass, and the degree of caloric restriction and the speed of loss push that share up, as a systematic review of 55 cohorts showed back in 2007.[6] Bariatric surgery sits in the same band. GLP-1s do too.

    MethodLean share of lossEvidence
    Diet / intensive lifestyle (no structured exercise)Pooled lifestyle estimate 26.2% (Eisa & Barood 2026); the long-standing “quarter rule”
    Diet + resistance trainingLowest share of any approach in the pooled analysis; often under 15% in individual training trials
    GLP-1 / dual agonistsSemaglutide ~35%, tirzepatide ~25% pooled; ~30% in the Beavers 2025 DXA meta-analysis
    Bariatric surgeryRoux-en-Y gastric bypass ~27 to 31% in most cohorts; higher in the early post-operative period
    Very-low-calorie / very rapid lossGreater caloric restriction is positively associated with the fat-free-mass share of loss (Chaston 2007)

    So why does lean loss feel worse on GLP-1s? Three reasons, none of them a unique drug toxicity. First, total loss matches what surgery produces, 15 to 22% of body weight versus 5 to 10% for most lifestyle programs, so the same percentage means 5 to 6 kg of lean tissue instead of 1 to 2 kg. Second, the loss is fast, especially during dose titration. Third, appetite suppression can drive protein intake very low at exactly the time the body needs it most.

    There is one more pattern worth knowing: the leaner you get, the worse the ratio tends to become. In SURMOUNT-1, the group that lost the least weight had the lowest fat share (70%), and as fat stores shrink the body defends them and takes proportionally more from lean tissue. That is an argument for lifting harder late in a weight loss, not just at the start.

    Section 3 · Does it matter?

    Strength is mostly preserved, until you test it properly

    The reassuring version of the story has real evidence behind it. A 2026 narrative review of 21 studies in older adults found GLP-1 therapy associated with muscle-mass loss in most studies but generally without significant changes in strength, with muscle quality mostly improved. In SURMOUNT-1, patient-reported physical function improved on tirzepatide despite the lean loss, carrying 20 fewer kilograms makes daily function easier even with somewhat less muscle. And on MRI in SURPASS-3, tirzepatide reduced the fat infiltrating muscle by 0.36 percentage points: less muscle, but cleaner muscle.[2][7]

    The counterpoint is the sharpest data in this article. In the T-REx trial (below), adults aged 50 to 70 on tirzepatide without training lost 6 to 9% of their strength over 40 weeks. Bench press fell 8.4%. So “strength is preserved” is true on average in younger trial populations, and false without training in the one trial that measured strength directly in this age group. Both statements are in the literature; which one applies to you depends on your age, your reserve, and whether you train.

    Who is actually at risk

    Aging alone removes roughly 3 to 8% of muscle mass per decade after 30, faster after 60, and strength falls faster than mass. Starting reserve is therefore the key variable. The joint advisory identifies as especially vulnerable: older adults, peri- and postmenopausal women, sedentary people, and anyone not doing strength training.[4] In SURMOUNT-1’s ≥65 subgroup the fat share was 76%, no disproportionate lean loss, but that subgroup was small and not frail, and the trial authors still flag older adults at risk of sarcopenic obesity as needing individualized nutrition and activity plans.

    Scale matters too. Someone who loses 30 kg at a 25% lean share loses 7.5 kg of lean tissue, roughly the muscle lost over a decade or more of aging, and after 65, anabolic resistance makes it harder to rebuild. There is also a plausible ratchet: stop the drug, regain weight that is predominantly fat, and repeat, each cycle leaving less muscle than the last. That mechanism is a hypothesis, not something yet proven in GLP-1 cohorts, and the body composition of weight regain on modern GLP-1s is understudied.

    What about mortality?

    Sarcopenia itself is strongly linked to death: a meta-analysis of 56 studies and 42,108 participants found sarcopenia associated with roughly double the mortality risk (HR 2.00, 95% CI 1.71 to 2.34).[8] But that is prevalent sarcopenia, low mass plus low strength and function, often alongside illness, in older populations. Applying that hazard ratio to a person with obesity who lost 5 kg of lean tissue while gaining function and metabolic health is not valid, and no honest reading of the data does so.

    The counterweight is proven benefit. In SELECT, semaglutide 2.4 mg in 17,604 adults with cardiovascular disease and overweight or obesity cut major cardiovascular events from 8.0% to 6.5% (HR 0.80) over about 40 months. No GLP-1 outcomes trial has shown a mortality increase attributable to lean loss. The honest statement is this: the cardiovascular benefit is proven in high-risk adults; a mortality harm from the accompanying lean loss is hypothesized, not demonstrated, and the balance plausibly flips in frail, low-reserve older adults, where the benefit is less certain and the muscle cost more dangerous.

    What is not established, stated plainly

    No GLP-1 trial has had falls, fractures or frailty as a primary outcome, and no study establishes a “safe” threshold for lean loss. A large preprint meta-analysis (not peer reviewed) found no increase in serious falls or fractures, but it captured only serious adverse events in trial populations whose mean ages topped out around 74, not the frailest old. Whether GLP-1 lean loss causes falls or frailty in the people most at risk remains an open question, not a settled reassurance.

    Section 4 · How to protect muscle

    Beyond protein: training is the intervention, protein is the support

    Most advice on this topic starts and ends with protein. The evidence orders things differently. Resistance training is the only strategy with a direct randomized trial on a modern GLP-1, and protein, in the joint advisory’s own words, is “likely inadequate” on its own without structured resistance training. Here is the protection plan in evidence order.

    1. Resistance training: the T-REx trial

    T-REx (Green et al., Nature Medicine, November 2026) randomized 120 sedentary adults aged 50 to 70, all on tirzepatide titrated to 15 mg, to thrice-weekly supervised resistance training or usual activity for 40 weeks, with 91% adherence in the training group. Total weight loss was identical. Almost everything else was not.[9]

    Total weight−22.6 kg
    Fat mass−17.2 kg
    Lean soft tissue−4.67 kg
    Bench press−4.1 kg
    Lat pulldown−4.4 kg
    Knee extension−4.4 kg

    Without training, tirzepatide alone cost 4.67 kg of lean soft tissue and 6 to 9% of strength across the three lifts. Bars are centred on zero: left is loss, right is gain (strength bars show percent change, capped for display).

    Read that carefully, because it cuts both ways. Training saved about 1.1 kg of lean tissue, roughly a quarter less lean loss, and swung strength by about 20 percentage points, from losing 6 to 9% to gaining 10 to 18%. But training did not fully prevent lean loss: even the lifters lost 3.6 kg. The trial ran at a single center (Perth), in people aged 50 to 70 without diabetes, over 40 weeks, with no falls or function endpoints and no follow-up after stopping. It is the best evidence we have, and it says: train, and expect attenuation, not immunity.

    Earlier work points the same way. In the S-LiTE trial, after an initial diet-induced loss, the combination of supervised exercise plus liraglutide maintained a 9.5 kg loss at one year, better than either alone, and cut body-fat percentage by 3.9 points, about twice either single treatment. A year after all treatment stopped, the former combination group was still 5.1 kg lighter and 2.3 points leaner than the former liraglutide-only group, and the exercise groups were the only ones still holding a ≥10% loss on average.[10] Carry the caveats with those numbers: only 109 of 195 participants returned for the follow-up, the exercise was supervised and vigorous, the drug was liraglutide (older and weaker), and the initial loss came from diet. The finding is about having built an exercise habit, not a post-stopping prescription proven in its own trial. And in adults over 65 losing weight by diet (Villareal 2017), resistance training limited lean loss to 2% versus 5% with aerobic exercise alone, while also halving hip bone-density loss.

    The dose: the joint advisory recommends strength training at least 3 times per week plus at least 150 minutes per week of moderate aerobic activity; general guidance sets muscle-strengthening on at least 2 days per week as a floor. Treat 2×/week as the minimum and 3×/week, the T-REx dose, as the target.

    2. Protein: necessary, proposed, and not proven on these drugs

    The joint advisory proposes 1.2 to 1.6 g of protein per kilogram of body weight per day during active weight loss, well above the 0.8 g/kg RDA, which is a minimum to prevent deficiency, not a target for someone losing 20 kg. A practical alternative the advisory gives is an absolute 80 to 120 g/day. It also sets a floor: do not fall below 0.4 to 0.5 g/kg/day, a range associated with muscle atrophy and functional impairment, and avoid prolonged intakes at or above 2 g/kg/day.[4]

    Protein target calculator

    114 to 152 gproposed advisory range at 1.2 to 1.6 g/kg/day
    76 gRDA minimum at 0.8 g/kg/day, a deficiency floor, not a weight-loss target
    80 to 120 gthe advisory’s practical absolute range, for comparison

    Two honest caveats. First, the advisory itself says it is unclear which weight to use, and that actual body weight can significantly overestimate requirements in obesity. Adjusted or ideal weight, or fat-free mass, are listed alternatives, which is why the basis selector exists. Second, the 1.2 to 1.6 g/kg figure is proposed, borrowed from weight-loss, aging and bariatric research. It has not been established in prospective trials in GLP-1 users. Kidney disease and other conditions require individualized targets; this calculator is a starting point for a conversation, not a prescription.

    The practical problem is rarely knowledge. It is intake. Appetite suppression and nausea can make even 80 g/day hard during titration. That is one reason side-effect management, meal timing, and protein-first eating are part of muscle protection, even though none of them has a GLP-1-specific muscle trial behind it.

    3. Everything else, graded honestly

    1. 1

      Resistance training, 2 to 3 times per weekDirect RCT evidence

      T-REx on tirzepatide, plus S-LiTE and Villareal by extension. The only intervention shown to change both lean loss and strength on a modern GLP-1. It attenuates loss; it does not abolish it.

    2. 2

      Protein at 1.2 to 1.6 g/kg/dayProposed advisory target

      Recommended by the four-society joint advisory, borrowed from adjacent literature. Supports training; the advisory is explicit that protein without structured resistance training is likely inadequate on its own.

    3. 3

      Keep the rate near 0.5 to 1% of body weight per weekExtrapolated

      In athletes, losing at 0.7%/week preserved or gained lean mass while 1.4%/week cost lean mass and strength, and restriction degree predicts the fat-free share of loss. No trial has randomized fast versus slow loss on a GLP-1 and measured lean mass. Treat this as a guardrail, not a proven rule.

    4. 4

      Creatine, if you are liftingIndirect evidence only

      No completed trial has tested creatine in GLP-1 users. In people doing resistance training generally, creatine adds about +1.1 kg of lean mass over training alone across 35 trials; without exercise, it does nothing for lean mass. Note that creatine also adds 0.5 to 2 kg of water weight, which can mask fat loss on the scale. The first trial of creatine in GLP-1 users was still recruiting in 2026.

    5. 5

      Slower dose titrationSensible, unproven for muscle

      Slower escalation is standard for gastrointestinal tolerability and makes it easier to actually eat enough protein. No trial shows it preserves lean mass. Do not let anyone sell it to you as a proven muscle strategy.

    6. 6

      HMBWeak and mixed

      The leucine metabolite HMB (~3 g/day) shows small effects in older and clinical muscle-wasting populations, with mixed results, and, like creatine, no data at all in GLP-1 users. It is optional, not a headline recommendation. The same applies to essential amino acids, leucine, omega-3s and vitamin D for muscle specifically, though vitamin D, B12, iron, thiamine and calcium are worth monitoring on GLP-1s for deficiency.

    Measure the right things. If you track anything, track strength and function, grip strength, a chair-stand test, your working weights, alongside a DXA if you have access to one, rather than scale weight alone. Consumer bioimpedance scales are hydration sensitive, which makes them especially noisy on these drugs.

    Section 5 · In development

    The muscle-preserving drugs: striking scans, no strength data yet

    A new class of drugs aims to change the ratio itself. Myostatin and the activins are natural brakes on muscle growth, acting through activin type II receptors. Block the brake, and muscle can be maintained, or even grown, while fat is lost. Three agents now have Phase 2 data alongside GLP-1s. All of them share the same limits: none is approved for obesity or muscle preservation, every result is DXA based, and none has yet shown that people get stronger or function better.

    Bimagrumab Phase 2 · Not approved for obesity

    Activin type II receptor blocker (Eli Lilly) · IV infusion every 4 weeks

    In the BELIEVE trial (n = 507, 48 weeks), semaglutide alone lost 3.9 kg of lean mass and 9.6 kg of fat. The high-dose bimagrumab combination lost only 1.3 kg of lean while losing 15.6 kg of fat. 92.3% of the combination’s weight loss was fat, versus 71.1% on semaglutide alone. By week 72 the combination had lost 22.1% of body weight with lean mass down just 2.9%; bimagrumab alone actually gained lean mass (+2.5%) while losing fat. An earlier trial in type 2 diabetes showed the same signature: fat mass −20.5% with lean mass up 3.6%.

    Lean loss, combo vs sema alone
    −1.3 kg vs −3.9 kg (wk 48)
    Fat share of loss
    92.3% vs 71.1%
    Watch-outs
    Muscle spasms, diarrhea, acne; an LDL-cholesterol rise (+0.47 mmol/L) reported in a pooled analysis of earlier trials

    Status as of late 2026: no Phase 3 obesity trial listed, and a planned bimagrumab + tirzepatide trial in type 2 diabetes was withdrawn before enrolment. One figure to hold loosely: secondary summaries of BELIEVE’s week-72 semaglutide lean change differ slightly (−7.4% vs −7.9%); this article uses −7.4%.

    Apitegromab Phase 2 for this use · Approved only for spinal muscular atrophy

    Selective inhibitor of myostatin activation (Scholar Rock) · IV every 4 weeks, added to tirzepatide

    In EMBRAZE (n = 102, 24 weeks), adding apitegromab to tirzepatide cut lean loss from 3.5 kg to 1.6 kg, 1.9 kg less lean loss (95% CI 1.2 to 2.7), a 54.9% retention, bringing the lean share of weight lost down from 30.2% to 14.6%, with similar total weight loss in both arms.

    Lean loss
    −1.6 kg vs −3.5 kg
    Lean share of loss
    14.6% vs 30.2%
    Watch-outs
    No improvement in strength, physical function or cardiometabolic markers at 24 weeks; 8 weeks after stopping, the lean advantage had shrunk from 1.9 kg to 0.9 kg

    Design caveats: this was a proof-of-concept trial powered at 75% for a 2 kg difference, analyzed at a two-sided α of 0.20 without multiplicity adjustment. Apitegromab was FDA approved in September 2026 as Isembyld, for spinal muscular atrophy only. That approval is not for weight loss, and the drug is not prescribable for muscle preservation.

    Trevogrumab Phase 2 · Investigational

    Anti-myostatin (anti-GDF8) antibody (Regeneron) · Added to semaglutide in the COURAGE trial

    In COURAGE (n ≈ 975 to 1,005, results presented autumn 2026), semaglutide plus placebo lost 7.3% of lean mass over 52 weeks, versus 5.8% with trevogrumab 25 mg and 4.2% with 75 mg, a 42.5% preservation at the higher dose (difference 3.1 percentage points, 95% CI 1.3 to 4.9). An MRI substudy suggested about 70% of thigh muscle volume was preserved. In a higher-dose portion, the lean share of loss fell from 33% on semaglutide alone to 16.8% with trevogrumab 200 mg.

    Lean-mass change
    −4.2% (75 mg) vs −7.3% (sema alone)
    Thigh muscle volume (MRI)
    ~70% preserved
    Watch-outs
    See the safety note below. It concerns the combination arm, not trevogrumab alone

    Evidence level: at the time of research, the COURAGE numbers rested on the sponsor’s press release and secondary summaries of the journal publication, not on a full peer-reviewed text verified for this article. Hold them one notch looser than BELIEVE and EMBRAZE. Safety: the trial’s triple-therapy arm (trevogrumab plus garetosmab, a different antibody, plus semaglutide) recorded 14 serious adverse events among 149 participants, 48 discontinuations and 2 deaths, with no causal association identified. Those events belong to the triple combination, not to trevogrumab alone, but they should not be omitted, and they are part of why this class is not ready for routine use.

    What about testosterone, SARMs or growth hormone, which circulate online for exactly this purpose? There is no randomized evidence for any of them in GLP-1 users, and each carries significant safety and regulatory issues. They are not an evidence-based strategy, whatever the forums say.

    The balanced read on the new drugs

    These agents prove the fat-to-lean ratio is modifiable. That matters. But they prove it on DXA scans, over 24 to 72 weeks, in Phase 2, without showing preserved strength or function, and with open questions about cost, infusions, lipids, long-term safety, and whether preserved lean mass persists after the drug stops. Promising is the right word. Available is not.

    Section 6 · Claim checker

    Six claims you will hear, checked

    Tap any claim for the verdict and the evidence behind it.

    Overstated ~2×

    The ~38% figure from STEP-1 is lean mass, muscle plus water, organs and connective tissue. Muscle is about half of lean mass, so the muscle share is roughly 20% of weight lost. The confusion between lean mass and muscle mass is the single biggest distortion in this debate.

    Not supported

    Pooled data put the lean share at 25 to 39% on incretin drugs versus 26.2% for intensive lifestyle intervention, not significantly different (p = 0.42). In SURMOUNT-1 the split on tirzepatide (74% fat / 26% lean) matched placebo almost exactly. What differs is the total amount lost, not the proportion.

    Depends who you are

    An estimated 2 to 3 kg of true muscle within a 15 to 20 kg loss is small for a robust 45-year-old. It is not small for a 75-year-old with sarcopenia losing 25 kg. That can equal more than a decade of age-related loss, and muscle is harder to rebuild after 65. Starting reserve and total loss decide whether this claim is true for you.

    Mostly, with a sharp caveat

    Function scores improved in SURMOUNT-1, most studies in older adults show no significant strength change, and muscle quality (fat infiltration) improves on MRI. But in T-REx, tirzepatide alone caused a 6 to 9% strength loss over 40 weeks in people aged 50 to 70. “Preserved” holds on average in younger populations, and fails without training in the one direct strength trial.

    Backwards

    The joint advisory states plainly that higher protein without structured resistance training is likely inadequate to preserve muscle. And the protein target itself (1.2 to 1.6 g/kg/day) is proposed, not proven in GLP-1 users. Training is the intervention with the randomized trial; protein is its support act. Even together, T-REx lifters still lost 3.6 kg of lean tissue.

    Evidence points the other way

    Weight regain after stopping is predominantly fat, and lean regain is incomplete, especially without training. In the S-LiTE follow-up, only the groups that had exercised were still holding a ≥10% loss a year after treatment ended. One honest limit: the body composition of regain on modern GLP-1s specifically is understudied, so the size of this ratchet effect is not yet pinned down.

    Not yet

    Bimagrumab, apitegromab and trevogrumab all shift body composition impressively in Phase 2, up to 92% of weight lost as fat with bimagrumab plus semaglutide. But none is approved for this use, none has shown a strength or function benefit, and apitegromab’s lean advantage halved within 8 weeks of stopping. DXA composition is a surrogate; function is the outcome that matters.

    Section 7 · Open questions

    What we still do not know

    An honest article ends with its gaps. As of this research (15 November 2026), the literature has no answer to any of the following. Anyone who claims otherwise is ahead of the evidence:

    • A “safe” threshold for lean loss. No validated cut-off exists below which lean loss is harmless.
    • Falls, fractures and frailty as outcomes. No GLP-1 trial has measured them as primary endpoints; direct evidence in adults over 75, or with established frailty or sarcopenia, is essentially absent.
    • What regain is made of. The body composition of weight regained after stopping semaglutide or tirzepatide specifically has not been well characterized.
    • Whether slower titration preserves muscle. Plausible, sensible for tolerability, and untested for lean mass.
    • Creatine or HMB on GLP-1s. Trials are recruiting (the first creatine trial, with supervised resistance training, is due around 2027); there are no results to cite.
    • The trials that would settle protection. LEAN-PREP (232 adults; semaglutide/tirzepatide with resistance training, protein, both, or control; MRI quadriceps as the primary outcome) is not due to complete until 2029, and RESIST (tirzepatide plus resistance training) until 2027.
    The bottom line

    GLP-1s do not uniquely destroy muscle. They produce large, fast weight loss. Large, fast weight loss always takes lean tissue with it, at roughly the same proportion as every other effective method. About half of that lean loss is real muscle, amounting to roughly 5 to 6 kg of lean tissue in the landmark trials. For most people that is a manageable cost next to proven cardiovascular and metabolic benefit. For older adults and anyone starting with little muscle, it is a cost that deserves a plan: lift two to three times a week, eat protein toward 1.2 to 1.6 g/kg/day, keep the rate of loss sane, and measure strength, not just the scale. Protein alone was never going to be the whole answer. It is not even the first item on the list.

    References

    Sources

    1. Eisa N, Barood O. Lean mass changes with incretin therapy versus lifestyle intervention: a systematic review and meta-analysis of randomised controlled trials. Diabetes, Obesity and Metabolism. 2026;28:4818 to 4827. dom-pubs.onlinelibrary.wiley.com/doi/10.1111/dom.70666
    2. Sattar N, et al. Tirzepatide and muscle composition changes in people with type 2 diabetes (SURPASS-3 MRI): a post-hoc analysis. Lancet Diabetes & Endocrinology. 2025;13:482 to 493. pubmed.ncbi.nlm.nih.gov/40318682
    3. Look M, et al. Body composition changes during weight reduction with tirzepatide in the SURMOUNT-1 study. Diabetes, Obesity and Metabolism. 2025;27(5):2720 to 2729. pmc.ncbi.nlm.nih.gov/articles/PMC11965027
    4. Mozaffarian D, et al. Nutritional priorities to support GLP-1 therapy for obesity: a joint advisory from the American College of Lifestyle Medicine, the American Society for Nutrition, the Obesity Medicine Association, and The Obesity Society. American Journal of Clinical Nutrition. 2025;122:344 to 367. pmc.ncbi.nlm.nih.gov/articles/PMC12304835
    5. Beavers KM, et al. GLP-1 receptor agonist, based therapies and DXA-acquired musculoskeletal health outcomes: a focused meta-analysis of placebo-controlled trials. Obesity. 2025;33:225 to 237. pubmed.ncbi.nlm.nih.gov/39710882
    6. Chaston TB, Dixon JB, O’Brien PE. Changes in fat-free mass during significant weight loss: a systematic review. International Journal of Obesity. 2007;31:743 to 750. pubmed.ncbi.nlm.nih.gov/17075583
    7. GLP-1 receptor agonist therapy and skeletal muscle: a narrative review synthesizing adult evidence with implications for older adults. Journal of Frailty & Aging. 2026. pmc.ncbi.nlm.nih.gov/articles/PMC13499509
    8. Xu J, Maier AB, et al. Sarcopenia is associated with mortality in adults: a systematic review and meta-analysis. Gerontology. 2022;68:361 to 376. pubmed.ncbi.nlm.nih.gov/34315158
    9. Green DJ, et al. Resistance exercise to attenuate lean mass loss during tirzepatide therapy for obesity: a randomized controlled trial (T-REx). Nature Medicine. Published online 9 November 2026. pubmed.ncbi.nlm.nih.gov/43003902
    10. Jensen SBK, et al. Healthy weight loss maintenance with exercise, GLP-1 receptor agonist, or both combined followed by one year without treatment (S-LiTE follow-up). eClinicalMedicine. 2024;69:102475. pubmed.ncbi.nlm.nih.gov/38544798
    11. Heymsfield SB, et al. Bimagrumab plus semaglutide, alone or in combination, for the treatment of obesity (BELIEVE): a randomised phase 2 trial. Nature Medicine. 2026;32:869 to 882. nature.com/articles/s41591-026-04204-0
    12. Heymsfield SB, et al. Effect of bimagrumab vs placebo on body fat mass among adults with type 2 diabetes and obesity: a phase 2 randomized clinical trial. JAMA Network Open. 2021;4:e2033457. pubmed.ncbi.nlm.nih.gov/33439265
    13. Pratley RE, et al. Apitegromab for lean mass preservation during tirzepatide-induced weight loss (EMBRAZE): a randomized, double-blind, placebo-controlled phase 2 trial. Nature Medicine. 2026. nature.com/articles/s41591-026-04440-4
    14. Regeneron Pharmaceuticals. Phase 2 COURAGE trial: trevogrumab + semaglutide (results as announced, 1 October 2026; numbers from sponsor release and secondary summaries, see the evidence-level note in Section 5). drugs.com COURAGE announcement
    15. Also cited in text: Wilding JPH, et al. Once-weekly semaglutide in adults with overweight or obesity (STEP-1). NEJM. 2021;384:989 to 1002 · Jastreboff AM, et al. Tirzepatide once weekly for the treatment of obesity (SURMOUNT-1). NEJM. 2022;387:205 to 216 · Lundgren JR, et al. Healthy weight loss maintenance with exercise, liraglutide, or both combined. NEJM. 2021;384:1719 to 1730 · Villareal DT, et al. Aerobic or resistance exercise, or both, in dieting obese older adults. NEJM. 2017;376:1943 to 1955 · Lincoff AM, et al. Semaglutide and cardiovascular outcomes in obesity without diabetes (SELECT). NEJM. 2023;389:2221 to 2232 · Neeland IJ, Linge J, Birkenfeld AL. Changes in lean body mass with GLP-1, based therapies and mitigation strategies. Diabetes, Obesity and Metabolism. 2024;26(Suppl 4):16 to 27 · Desai I, et al. Creatine supplementation and resistance training: meta-analysis. Journal of Strength and Conditioning Research. 2024 · Garthe I, et al. Effect of two different weight-loss rates on body composition and performance in elite athletes. International Journal of Sport Nutrition and Exercise Metabolism. 2011;21:97 to 104.

    This article is educational and does not replace medical advice. GLP-1 and GIP/GLP-1 medicines are prescription drugs with real risks and contraindications; whether any of them is appropriate for you belongs in an evaluation with your own physician, not in an article.

  • Testosterone, the Black Sheep of Medicine

    Testosterone, the Black Sheep of Medicine

    Why primary care hesitates, and why the story is less about one hormone than about sport, scandal, a decade-long heart scare, and clinics that made saying yes look like the problem.

    Medical glass vials and a sterile syringe on a white background.
    Medical glass vials and a sterile syringe on a white background.

    The short version

    • Testosterone is the only common replacement hormone that is also a Schedule III controlled substance and a banned doping drug. Thyroid hormone, insulin and estrogen are not controlled substances. Estrogen is expressly excluded from the anabolic-steroid definition in US law.
    • A 2013 to 2014 cardiovascular scare drove prescribing down 48% in established users and 62% in new users between 2013 and 2016, on the strength of observational studies and a meta-analysis, not a randomized safety trial.
    • TRAVERSE (2023, n=5,246) found testosterone gel noninferior to placebo for major cardiovascular events (7.0% vs 7.3%; HR 0.96). In February 2025 the FDA removed the boxed warning on cardiovascular risk, and added a blood-pressure warning at the same time.
    • The guidelines never banned testosterone. They require symptoms plus two low early-morning levels, a search for the cause, and monitoring. Much of the reluctance is about everything around the prescription, not the molecule alone.

    The drug with three identities

    Ask a primary care doctor to replace thyroid hormone and nobody blinks. Levothyroxine is deficiency medicine: measure, replace, recheck. Insulin is the same. Estrogen, for all its own controversies, is prescribed as hormone therapy. Testosterone should sit in that same row, a hormone, replaced when it is genuinely low, monitored like any other.

    It does not sit in that row. Testosterone is the only common replacement hormone that carries three identities at once: medicine, crime, and cheating. It is a legitimate treatment for hypogonadism. It is also a Schedule III controlled substance under the US Controlled Substances Act, regulated by the DEA in the same tier as ketamine and codeine combination products, a category defined by accepted medical use alongside potential for abuse. And it is the archetypal doping drug, the substance behind stripped medals and fallen champions.

    That triple identity is not rhetoric. It is written into the statute itself. Federal law defines an anabolic steroid as any drug or hormonal substance “chemically and pharmacologically related to testosterone (other than estrogens, progestins, and corticosteroids) that promotes muscle growth” (21 U.S.C. §802). The definition is built around testosterone, and estrogen is carved out of it by name. Thyroid hormone and insulin were never in it at all.

    So a clinician who prescribes testosterone is, in regulatory terms, prescribing a drug of abuse, from a DEA-registered practice, for a documented legitimate medical purpose, on a prescription that expires in six months. Possession without a prescription is a federal crime. No other hormone on the shelf asks the prescriber to work inside that frame. Clinicians absorbed the framing along with the pharmacology: this is the hormone you defend, not the hormone you simply replace.

    Every other replacement hormone is treated as deficiency medicine. Testosterone alone arrives at the visit with a criminal record and a doping scandal attached.

    One honest limit before going further: no study found in the research for this article directly measured how much of today’s reluctance the legal history caused. The link is inferential. The law made the drug symbolically different, and practice culture followed. What is documented is the law, the scare that followed, and the measurable collapse in prescribing when that scare peaked. Those come next.

    How sport made a hormone a controlled substance

    Synthetic testosterone was isolated and synthesized in 1935 and used medically for decades, for hypogonadism, delayed puberty, some anaemias and some breast cancers, long before anyone scheduled it. The drug predates its criminalization by 55 years. What changed was not the medicine. What changed was sport.

    1935

    A medicine first

    Testosterone is isolated and synthesized. For the next five decades it is a prescription medicine for defined deficiencies, not a federally controlled substance.

    1988

    Seoul, and the scandal era

    Ben Johnson is stripped of Olympic gold at the 1988 Seoul Games after testing positive for a steroid. Congressional hearings follow in 1988 to 1990. The stated objective of the legislation that follows is to remove steroids from competitive sport and from cosmetic and bodybuilding use, not to regulate endocrinology.

    1990

    Anabolic Steroids Control Act of 1990

    Enacted 29 November 1990 as part of the Crime Control Act (Pub. L. 101 to 647), effective 27 February 1991. Anabolic steroids, originally 27 compounds, defined around testosterone, are placed in Schedule III. Non-medical use and distribution are criminalized.

    2004

    The net widens

    After the BALCO scandal, the Anabolic Steroid Control Act of 2004 (signed 22 October 2004, effective 20 January 2005) broadens the definition to capture designer steroids and prohormone precursors such as androstenedione sold as supplements. Sources disagree on the expanded list’s exact count, 49 versus 59 named substances, so no precise count is claimed here.

    2014

    Designer Anabolic Steroid Control Act

    About two dozen further designer steroids are added, with new offences for false labeling. The framework built for sport now fully surrounds a hormone doctors had prescribed since the 1930s.

    There is a contested footnote worth flagging rather than repeating as fact. Multiple secondary histories report that the AMA, FDA, DEA, NIDA and NIH testified against scheduling in 1988 to 1990, arguing anabolic steroids did not meet the Controlled Substances Act’s dependence criterion, and that Congress proceeded anyway to “send a message.” That account appears in secondary and partisan sources; it was not verified against a primary hearing record in the research for this article, so treat it as a widely repeated claim, not an established one.

    What needs no caveat is the direction of travel. A hormone became a controlled substance because of what athletes and bodybuilders did with it. Every prescriber since has inherited that decision, including the ones seeing a 58-year-old with two low morning levels and a textbook symptom picture, who has never been near a podium or a stage.

    The cardiovascular scare, 2013 to 2015

    If the law gave testosterone its reputation, the years 2013 to 2015 gave primary care its reason. Three publications, none of them a randomized safety trial, raised the possibility that testosterone therapy increased heart attacks, strokes and death. The FDA responded, labels changed, and prescribing fell off a cliff.

    On 31 January 2014, the FDA issued a Drug Safety Communication announcing it was evaluating the risk of stroke, heart attack and death with approved testosterone products. That September, an FDA advisory committee voted 20 to 1 to restrict the indicated population and require a cardiovascular safety trial. On 3 March 2015, the FDA required label changes: testosterone was approved only for men with low testosterone caused by certain medical conditions; benefit and safety for low testosterone due to aging were not established; and labels had to carry information about a possible increased risk of heart attack and stroke. The FDA’s own wording was careful, “possible” with some studies showing risk and others not. In practice, careful wording rarely survives contact with a busy clinic. What survived was the warning.

    Observational · VA cohort

    Vigen et al., JAMA 2013

    Design: Retrospective national VA cohort of 8,709 men with total testosterone <300 ng/dL who had coronary angiography (2005 to 2011); 1,223 started testosterone.

    Finding: Composite of death, MI and ischaemic stroke at 3 years: 25.7% (testosterone) vs 19.9% (no testosterone) on Kaplan, Meier estimates; adjusted HR 1.29 (95% CI 1.04 to 1.58).

    Limit: Retrospective and not randomized. Who gets treated, and why, cannot be separated from the result. Critics noted the raw event count was lower in the treated group (123/1,223, 10.1% vs 1,587/7,486, 21.2%) and that the headline figures were model-based estimates; JAMA issued corrections. The claim that the study is therefore “reversed” is itself contested.

    Observational · Claims data

    Finkle et al., PLoS One 2014

    Design: Claims-database cohort of 55,593 men with a first testosterone prescription, comparing non-fatal MI in the 90 days after the prescription with the prior year.

    Finding: Post/pre rate ratio 1.36 (95% CI 1.03 to 1.81). In men ≥65, RR 2.19; in men <65, excess risk appeared only with prior heart disease (RR 2.90).

    Limit: The paper itself lists them: no data on indication, labs, race or lifestyle; no data on how much drug was actually taken; non-fatal MI only. In absolute terms the increase was small, about 1 additional MI per 1,000 person-years (3.48 to 4.75) in secondary analyses.

    Meta-analysis · 27 RCTs

    Xu et al., BMC Medicine 2013

    Design: 27 placebo-controlled trials, 2,994 mainly older men, 180 cardiovascular-related events in total.

    Finding: Pooled OR 1.54 (95% CI 1.09 to 2.18) for cardiovascular events, but the effect split sharply by funding: non-industry trials OR 2.06 vs industry-funded trials OR 0.89 (p for interaction 0.03).

    Limit: No included trial was designed to measure cardiovascular events; events were harvested from adverse-event reports with inconsistent definitions. A contemporaneous meta-analysis by Corona et al. (75 studies) found no cardiovascular risk (OR 1.07).

    A small 2010 randomized trial in frail older men (Basaria et al., the TOM trial) had also been stopped early for excess cardiovascular events, and it formed part of the background signal, though it was not one of the three studies the FDA action rested on.

    What the warning did to prescribing

    Baillargeon and colleagues tracked nearly 10 million commercially insured men in JAMA (2018). Testosterone use had climbed from 0.52% in 2002 to 3.20% in 2013, then fell to 1.67% by 2016. The relative decrease from 2013 to 2016 was 48% in established users and 62% in new users. The authors’ own words: “The steepest decrease coincided with 2 published reports of testosterone-associated adverse cardiovascular events and an FDA communication.” A VA study over the same period found a 40% decrease.

    Testosterone use among commercially insured US men ≥30, 2002 to 2016

    All users: 0.52% (2002) → 3.20% (2013) → 1.67% (2016). Source: Baillargeon et al., JAMA 2018;320:200 to 202, Clinformatics Data Mart, 9,962,538 men. Commercially insured (employed) men, not generalisable to retired or Medicare populations. Testosterone use had also tripled from 2001 to 2011, mostly in men without a clear indication, so part of the fall was likely correction of overprescribing, not just undertreatment.

    That last point matters, and honest writing about this topic has to hold both halves of it. Some of the prescribing that disappeared after 2013 should never have happened, the pre-2013 boom was driven substantially by men without a clear indication, in the era of “low T” advertising. And some of the men who lost access, or never got assessed, had the real disease. A blunt warning cannot tell those two groups apart. Clinicians, understandably, resolved the ambiguity by saying no more often, to everyone.

    What TRAVERSE actually showed

    The trial the FDA committee asked for in 2014 arrived in 2023. TRAVERSE (Lincoff et al., New England Journal of Medicine 2023) was the first adequately powered randomized cardiovascular safety trial of testosterone: 5,246 men aged 45 to 80, at 316 US sites, each with symptoms of hypogonadism, two fasting testosterone levels below 300 ng/dL, and either established cardiovascular disease or high cardiovascular risk. Men were randomized to daily 1.62% testosterone gel, dose adjusted to keep levels at 350 to 750 ng/dL, or matching placebo. Mean treatment duration was 21.7 months; mean follow-up was 33.0 months.

    TRAVERSE: the primary result

    5,246
    Men randomized, 45 to 80 years, confirmed low testosterone + symptoms + high CV risk
    7.0% vs 7.3%
    Major adverse cardiovascular events (CV death, nonfatal MI, nonfatal stroke): testosterone vs placebo
    HR 0.96
    95% CI 0.78 to 1.17 · P<0.001 for noninferiority, upper limit well below the 1.50 margin
    33.0 mo
    Mean follow-up (mean treatment duration 21.7 months), gel only, physiological target range

    In plain terms: in this population, at physiological doses, testosterone did not increase major cardiovascular events versus placebo. Noninferiority is not superiority. TRAVERSE did not show testosterone prevents events.

    What it showed

    • No increase in the primary MACE composite over ~3 years of follow-up in men with confirmed hypogonadism and high cardiovascular risk.
    • Similar results in sensitivity analyses censoring at treatment discontinuation, and for the secondary composite including coronary revascularization.
    • Enough reassurance that on 28 February 2025 the FDA removed the boxed-warning language on increased cardiovascular risk from all testosterone labels and added the TRAVERSE results to them.

    What it did not show

    • Secondary signals: higher incidence on testosterone of atrial fibrillation, acute kidney injury and pulmonary embolism (stated in the abstract without numbers; precise percentages circulating in secondary summaries were not verified against the NEJM full text for this article).
    • Long-term safety beyond ~3 years; safety of injections (gel only); safety at supraphysiological, bodybuilding doses; safety in men without confirmed hypogonadism.
    • Efficacy for energy, vitality or cognition. It was a safety trial, not an efficacy trial for those outcomes. Prostate-cancer numbers were too few to prove equivalence.

    The limitations are real, too

    A widely cited critique in JACC Advances (2023), an opinion piece, and worth reading as one, makes points any honest summary should carry: about 61% of participants discontinued their gel or placebo, so much of the follow-up happened off treatment, diluting any true difference in an intention-to-treat analysis; roughly 18% were lost to follow-up; achieved testosterone levels hovered near the bottom of the 350 to 750 ng/dL target (men started at a mean of 227 ng/dL and rose only ~148 ng/dL at 12 months); the trial enrolled 5,246 of a planned ~6,000 men; it was funded by AbbVie and other manufacturers; and men with recent MI or stroke or uncontrolled heart failure, the highest-risk subgroup, were excluded.

    The accurate framing, then, is not “case closed.” It is: the best randomized evidence available excludes a large increase in major cardiovascular events over about three years, at physiological doses, in confirmed hypogonadism, and leaves the secondary signals open.

    And note what the FDA did not do in February 2025. It retained the limitation of use for age-related hypogonadism. It added a blood-pressure warning, after ambulatory monitoring studies confirmed a class-wide rise in blood pressure. Testosterone remains Schedule III. The label change is not a recommendation to prescribe, nor a declaration of safety for everyone. One specific warning came off; another went on. (Reports in 2026 secondary sources that HHS/FDA has requested further changes, removing the age-related limitation and narrowing the prostate contraindication, describe a request, not a completed change, and were not verified on an FDA primary page for this article.)

    Why primary care still hesitates

    If the heart question has been substantially answered, why does the hesitation persist? Because the cardiovascular scare was only ever one layer of it. Several real frictions stack up at once. Some are documented. Some are not. It helps to keep the two apart.

    ⚠ An honesty note on the evidence for this section

    No quantitative US survey of primary-care physicians’ attitudes toward testosterone prescribing was found in the research for this article. Reasons such as “minimal hormone training in residency” “fear of litigation or board action” and “prior-authorization burden” appear repeatedly in clinic and trade commentary, but no curriculum study, malpractice dataset or payer-policy source was found to verify them. They are presented below as commentary-level claims, plausible, widely repeated, and not survey-established fact.

    The one peer-reviewed attitudinal study found is British, qualitative and small: interviews with 20 men (aged 45 to 62), 30 GPs and 10 endocrinologists, published in the Journal of Men’s Health in 2024 and commissioned by Besins Healthcare UK. Its findings are recognisable on either side of the Atlantic. Many GPs were unaware of the full range of testosterone-deficiency symptoms and did not think to check testosterone for non-specific complaints like fatigue and low energy, which were attributed to aging or work stress. Many lacked confidence in managing the condition and wanted standardized, guideline-based care. Stigma, embarrassment and “machismo” inhibited men from raising symptoms at all, and limited consultation time blocked the longer conversation. Endocrinologists, for their part, believed many referred men could be managed appropriately in primary care. Small, qualitative and industry-funded, but it points to knowledge gaps, low confidence, stigma and time pressure, not to a single cause.

    The documented frictions

    Documented

    A diagnosis that takes work

    Guideline diagnosis requires symptoms plus two early-morning fasting measurements, then LH/FSH, PSA and haematocrit, then a search for the cause, depression, hypothyroidism, sleep apnoea, obesity, opioids and pituitary disease can all sit behind the same complaints. A single low number is not a diagnosis, and ruling the mimics out is real work in a short visit.

    Documented

    A monitoring burden that never ends

    Both major guidelines require baseline and ongoing monitoring of testosterone, haematocrit and prostate risk (details in the next section). That is a standing obligation, repeat labs, dose decisions, stop rules, attached to every prescription, indefinitely.

    Documented

    Fertility cuts the other way

    Exogenous testosterone suppresses LH and FSH and with them sperm production. Both guidelines advise against starting it in men planning fertility in the near term; the AUA states it should not be prescribed to men currently trying to conceive. In younger men, hesitation is often the guideline working as intended.

    Documented

    Controlled-substance gravity

    Schedule III status brings DEA registration, legitimate-medical-purpose documentation and diversion concern to every script. The regulatory facts are established; how heavily they weigh on individual clinicians’ decisions is inferred, not surveyed.

    Commentary-level

    Training, liability and paperwork

    Clinic commentary routinely cites minimal hormone training, fear of liability, prior-authorization fights and 15-minute visit economics. Each is plausible; none was verified against a primary source in this research. Treat them as the profession’s folklore until someone surveys it properly.

    Documented, indirectly

    Guilt by association

    A GP who prescribes testosterone risks being read as practicing “low-T clinic” medicine, because, as the next sections show, a visible part of that market prescribes without diagnosis, monitoring or thresholds. No survey quantifies the reputational effect, but the overprescribing it reacts to is documented.

    Stack those together and “no” becomes the path of least resistance: a safety scare that has been superseded but is still seared in, a genuinely demanding diagnostic and monitoring protocol, a controlled-substance framework, and a cultural association with doping and anti-aging marketing. Not because the doctor has read TRAVERSE and rejected it. Often because the doctor trained during the warning years, never built the monitoring workflow, and has a waiting room full of problems that come with fewer questions attached.

    What the guidelines actually say

    Here is the part that surprises patients most: the guidelines are not against testosterone. The Endocrine Society (Bhasin et al., 2018), the American Urological Association (Mulhall et al., 2018) and, most narrowly, the American College of Physicians (2020) all describe a legitimate treatment for a defined condition. What they are against is treating a marketing category. “Low T” as advertised, fatigue, low energy, weight gain and fading vitality in aging men with borderline or unmeasured levels, is not the diagnosis any of them recognizes.

    Endocrine Society: Bhasin et al., J Clin Endocrinol Metab 2018

    • Diagnose only in men with symptoms/signs consistent with testosterone deficiency and unequivocally, consistently low serum testosterone.
    • First test: fasting morning total testosterone with an accurate assay, then confirm by repeating it. Add free testosterone (equilibrium dialysis or accurate formula) when total T is near the lower limit or SHBG is altered. Then evaluate the cause.
    • Treat symptomatic deficiency after discussing benefits, risks and monitoring, shared decision-making, aiming for mid-normal concentrations.
    • Against routine screening of the general population, and against routinely prescribing to all men ≥65.
    • Do not start with: plans for near-term fertility; breast or prostate cancer; palpable prostate nodule; PSA >4 ng/mL (>3 if at increased risk) without urological evaluation; elevated haematocrit; untreated severe sleep apnoea; severe lower urinary tract symptoms; uncontrolled heart failure; MI or stroke within 6 months; thrombophilia.
    • Monitor: symptoms, adverse effects and compliance; serum testosterone and haematocrit; prostate-cancer risk during the first year.

    Full text detail beyond the published abstract (exact monitoring intervals, haematocrit thresholds) is widely summarized secondhand; figures not in the abstract were not treated as verified for this article.

    The diagnosis, step by step

    What a guideline-concordant work-up actually looks like, the standard a prescription should be measured against:

    1. Symptoms firstConsistent symptoms or signs of testosterone deficiency, not a number looking for a patient, and not fatigue alone without the rest of the picture.

    2. Two early-morning fasting levelsTotal testosterone measured on two separate occasions, both early morning and fasting, with a reliable assay, below the cut-off both times (AUA: <300 ng/dL as a reasonable threshold).

    3. Find the causeLH (and prolactin if LH is low), plus a search for reversible drivers, obesity, diabetes, opioids, sleep apnoea, pituitary disease, before replacing anything.

    4. Screen the contraindicationsFertility plans, prostate risk (PSA, examination), haematocrit, sleep apnoea, heart failure, recent MI or stroke, thrombophilia.

    5. Treat to mid-normal, and monitorShared decision-making, a mid-normal target (not an enhancement target), then scheduled testosterone, haematocrit and prostate monitoring, with a willingness to stop if symptoms do not improve.

    Where the caution is justified

    A piece arguing that testosterone is medicine’s black sheep would be dishonest if it pretended the wariness has no basis. Some of it is the correct clinical response to a drug with real effects beyond the one being treated.

    Erythrocytosis: the most consistent risk

    Testosterone thickens the blood. A systematic review commissioned for the Endocrine Society guideline found testosterone increased the risk of haematocrit above 54% (RR 8.14, 95% CI 1.87 to 35.40); in the T4DM trial of injected testosterone undecanoate, haematocrit >54% occurred in 22% on testosterone vs 1% on placebo. This is why haematocrit monitoring is not bureaucratic decoration. (Figures via secondary summary of Ponce et al., JCEM 2018.)

    Fertility suppression

    Exogenous testosterone suppresses LH and FSH, lowers intratesticular testosterone and reduces, sometimes abolishes, sperm production. Recovery takes months to years and is not guaranteed. Prescribing it to a man who wants children, without that conversation, is a harm, full stop.

    Prostate: monitoring, not panic

    Neither guideline finds definitive evidence that testosterone causes prostate cancer, and TRAVERSE found no difference, with too few cancers to prove equivalence. But PSA rises slightly more on treatment, and active prostate cancer remains a contraindication. Monitoring is mandatory precisely because the question is managed, not settled.

    Sleep apnoea & blood pressure

    Untreated severe obstructive sleep apnoea is a guideline contraindication to starting, because testosterone can worsen it. And the FDA’s February 2025 action added a blood-pressure warning on ambulatory-monitoring evidence even as it removed the cardiovascular boxed warning. “The warning came off” is only half of that headline.

    And the low-T clinic problem is real

    The strongest evidence that some caution is warranted comes from the sellers, not the sceptics. In a secret-shopper study published in JAMA Internal Medicine (Dubin et al., 2022), researchers presented to seven direct-to-consumer testosterone platforms as a 34-year-old man with normal testosterone who wanted children. Six of the seven offered him testosterone anyway. Only one of seven asked about recent cardiovascular events or desire for fertility; six of seven had no testosterone threshold at all; half quoted a treatment target above 1,000 ng/dL, roughly double the mid-normal aim the guidelines describe.

    The pattern repeats at scale. A 2026 global analysis of 253 websites offering testosterone prescriptions found 9.9% prescribed for men with normal testosterone, 20.6% claimed testosterone reduces cardiovascular risk, and 9.9% marketed anti-aging effects, with US clinics the worst performers. And a single-center chart review presented at ENDO 2026, conference data, not yet peer reviewed, so weigh it accordingly, found that of 200 men prescribed testosterone between 2020 and 2025, only 12% had received a guideline-concordant evaluation: two low morning levels, LH/FSH, and no contraindications. A quarter of the prescriptions followed a patient request.

    That is the environment a primary care doctor is looking at when a patient arrives asking for testosterone by name. The refusal is sometimes aimed at the man in the room. Often it is aimed at the market behind him.

    The line worth holding

    Strip the history away and the evidence supports a clear, narrow position. Treating symptomatic, confirmed hypogonadism, two low morning levels plus symptoms, cause investigated, no contraindications, to mid-normal levels, with monitoring, carries the reassurance of a 5,246-man randomized trial on major cardiovascular events over about three years. The evidence does not support testosterone for energy, vitality or cognition alone; at supraphysiological targets; without baseline labs; in men with normal levels who want children; or as a substitute for asking why the level is low in the first place.

    The strongest version of the doctor’s case

    “I am not against testosterone. I am against this testosterone, prescribed without a diagnosis, without monitoring, at enhancement doses, by a clinic that profits from saying yes.”

    The strongest version of the patient’s case

    “I do meet the guideline criteria, and I am being refused on the basis of a 2013 scare that a 5,246-man randomized trial and a 2025 FDA label change have superseded.”

    Both of those people are right about the case they are describing. The failure happens when they are describing different cases and talking past each other, when a blanket “we don’t do testosterone here” meets a man who has done everything the guidelines ask, or when a legitimate prescription is expected to behave like the enhancement protocol a clinic advertised on a podcast.

    Testosterone earned its black-sheep status honestly, in stadiums and in strip-mall clinics, long before it reached the exam room. But a hormone is not its worst users. The drug synthesized in 1935 for men whose bodies could not make enough of it is still that drug. The task for primary care is not to rehabilitate its reputation or to keep punishing it, it is the duller, harder work the guidelines already describe: measure twice, find the cause, treat the deficiency, monitor the man. Do that, and testosterone stops being the black sheep. It becomes what thyroid hormone got to be all along, just another hormone, replaced properly.

    Open questions this article will not pretend are closed

    • No quantitative US survey of primary-care attitudes toward testosterone was found. The training, liability and prior-authorization explanations remain commentary level until one exists.
    • TRAVERSE’s secondary signals (atrial fibrillation, acute kidney injury, pulmonary embolism) are stated without numbers in the trial abstract; precise percentages in circulation come from secondary summaries.
    • Whether the reported 2026 request to remove the age-related limitation of use becomes a completed FDA label change is unresolved, as of the sources checked, it is a request, not a change.
    • Long-term safety beyond ~3 years, injectable formulations, and men at the very highest cardiovascular risk remain outside what TRAVERSE tested.
    Educational only. This article is general information about the evidence and history of testosterone therapy. It is not medical advice and does not establish a doctor-patient relationship. Testosterone is a Schedule III controlled substance available only by prescription after an individual evaluation. If you think you may have testosterone deficiency, the guideline-concordant next step is a conversation with your own clinician and two early-morning fasting measurements, not a single test, and not a direct-to-consumer platform that skips them.

    Follow @dr.francisco_md for the Saturday hormone-health series: protein after 40, peptides, estrogen in men, visceral fat and alcohol are next in the queue.

    References

    1. FDA. FDA issues class-wide labeling changes for testosterone products (28 Feb 2025), TRAVERSE results added; cardiovascular boxed warning removed; age-related limitation retained; blood-pressure warning added. fda.gov
    2. FDA Drug Safety Communication (3 Mar 2015): label changes, approved uses clarified; possible increased risk of heart attack and stroke. drugs.com
    3. Lincoff AM, et al. Cardiovascular safety of testosterone-replacement therapy (TRAVERSE). N Engl J Med. 2023;389(2):107 to 117. doi:10.1056/NEJMoa2215025. PubMed 37326322
    4. Bhasin S, et al. Testosterone therapy in men with hypogonadism: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2018;103(5):1715 to 1744. PubMed 29562364
    5. Mulhall JP, et al. Evaluation and management of testosterone deficiency: AUA guideline. J Urol. 2018;200(2):423 to 432. PubMed 29601923
    6. Qaseem A, et al. Testosterone treatment in adult men with age-related low testosterone: ACP clinical guideline. Ann Intern Med. 2020. acponline.org
    7. Vigen R, et al. Association of testosterone therapy with mortality, myocardial infarction, and stroke in men with low testosterone levels. JAMA. 2013;310(17):1829 to 1836. PubMed 24193080
    8. Finkle WD, et al. Increased risk of non-fatal myocardial infarction following testosterone therapy prescription in men. PLoS One. 2014;9(1):e85805. PMC3905977
    9. Xu L, et al. Testosterone therapy and cardiovascular events among men: systematic review and meta-analysis. BMC Med. 2013;11:108. PMC3648456
    10. Baillargeon J, et al. Testosterone prescribing in the United States, 2002 to 2016. JAMA. 2018;320:200 to 202. jamanetwork.com
    11. Dubin JM, et al. Guideline-discordant care among direct-to-consumer testosterone therapy platforms. JAMA Intern Med. 2022;182(12):1321 to 1323. medicalxpress.com summary
    12. Testosterone Replacement: Is it the TRAVERSE Trial or a Travesty? JACC Advances. 2023;2(10):100742. PMC11198716
    13. Barriers to diagnosis and accessing effective treatment and support for testosterone deficiency. J Men’s Health. 2024. doi:10.22514/jomh.2024.001. pharmacymagazine.co.uk summary
    14. Anabolic Steroids Control Act of 1990, Schedule III placement; statutory definition at 21 U.S.C. §802. en.wikipedia.org; DEA Schedule III status summary: legalclarity.org
    15. Global analysis of websites offering testosterone prescriptions. J Clin Endocrinol Metab. 2026;111(6):1651. academic.oup.com
    16. Testosterone prescribing often ignores guidelines, single-centre review presented at ENDO 2026 (conference report, not peer-reviewed). medscape.com

    Research verified 15 November 2026. Where a figure rests on a secondary summary rather than a primary text, the article says so at the point of use.

  • Muscle Is a Metabolic Organ, Not Just for Looks

    Muscle Is a Metabolic Organ, Not Just for Looks

    Muscle and Longevity From the book Muscle Is Medicine

    Your muscles clear most of the glucose insulin sends into the bloodstream, bank the amino acids you fall back on in illness, and secrete hundreds of signalling molecules that talk to fat, liver, bone, and brain. The case for treating muscle as an organ is strong. So are the limits of the slogan. Here is both.

    ~80%of glucose disposal happens in skeletal muscle, under insulin-stimulated conditions. That condition always comes with the headline number (DeFronzo 2009)
    −11%lower insulin resistance (HOMA-IR) for every 10% more muscle relative to body weight, in 13,644 US adults. An association, not proof of cause (Srikanthan 2011)
    1.16hazard ratio for death per 5 kg lower grip strength, across 139,691 adults in 17 countries. Grip outpredicted systolic blood pressure (PURE, Leong 2015)

    Ask most people what muscle is for and they will say movement, or appearance. Both answers undersell the tissue by a wide margin. Skeletal muscle is roughly 40% of your body mass, it is where most of a meal’s glucose ends up, it is the protein reserve your body raids when you are sick, and it is a secretory organ in its own right. Medicine has been slow to treat it that way: we measure blood pressure, cholesterol, and blood sugar obsessively, and muscle almost never. This piece covers what muscle does metabolically, what happens as it disappears with age, what rebuilds it, and where the “muscle is medicine” slogan runs ahead of the evidence.

    An organ, not an accessory

    An older woman doing a dumbbell row in a gym.
    An older woman doing a dumbbell row in a gym.

    Calling muscle an endocrine organ is no longer a metaphor. Reviews in the field now describe skeletal muscle as essential for movement, breathing, and heat production, and as a tissue that secretes signalling molecules and handles most insulin-driven glucose disposal.[1] Bente Pedersen’s group at the University of Copenhagen formalized the idea: muscle fibres produce and release peptides, myokines, that act on the muscle itself, on neighbouring tissue, and through the bloodstream on distant organs. By 2012 they could write that “the muscle secretome consists of several hundred secreted peptides” and that muscles “communicate with other organs, such as adipose tissue, liver, pancreas, bones and brain.”[2]

    Muscle’s jobs beyond movement:

    • Glucose sink. Muscle is the main site of insulin-stimulated glucose uptake and the predominant site of glycogen storage. It is the body’s largest single destination for blood sugar after a meal.
    • Amino-acid reservoir. Muscle protein is the bank the body draws on during fasting, illness, trauma, or sepsis, releasing amino acids, especially alanine and glutamine, for the liver to turn into glucose and for immune cells and wounds to build with. In a classic 1976 study, 60 hours of fasting raised muscle’s release of glycogenic amino acids by about 69% on average.[3] Low muscle means low reserve: the same illness or week of bed rest costs a person with little muscle proportionally more function.
    • Heat and energy. Muscle is essential for thermogenesis and is a major consumer of glucose and contributor to whole-body energy metabolism.[1]
    • Endocrine signalling. The myokines, covered in their own section below.

    The glucose sink, and the condition attached to the famous 80%

    The number you will see quoted everywhere is that muscle handles about 80% of glucose disposal. It is a real number from real physiology, with a condition attached that usually gets dropped. It comes from the euglycemic insulin clamp, the laboratory technique in which insulin is infused at a high, steady rate. Under those insulin-stimulated conditions, Ralph DeFronzo’s group established that roughly 80% of glucose uptake occurs in skeletal muscle, and about three-quarters of it is stored rather than burned. It is converted to glycogen, and impaired glycogen synthesis is one of the earliest defects in type 2 diabetes.[4]

    Now the contrast the slogan leaves out. In the fasting, postabsorptive state, which covers most of your day, the majority of glucose uptake, about 70 to 75%, happens in tissues that do not need insulin at all: the brain, red blood cells, and the gut and liver region. Only about a quarter goes to insulin-sensitive tissues like muscle.[4] Toggle between the two states to see how different the picture is:

    One tissue, two very different states

    Where glucose uptake happens

    Source: DeFronzo RA & Tripathy D. Diabetes Care. 2009;32(Suppl 2):S157, S163. Shares are of total glucose uptake in each state, not of everything you eat.

    The mechanism explains a lot of the rest. Insulin docks at its receptor on the muscle cell, a signalling chain (IRS-1 → PI3-kinase → Akt → AS160) moves GLUT4 transporters to the cell surface, and glucose flows in to be burned or stored as glycogen. Insulin resistance in muscle is a failure of that chain: less transport, less storage.[4] And there is a second door: muscle contraction itself stimulates glucose uptake through an insulin-independent pathway. That is why a walk after a meal lowers blood glucose even in people whose insulin signalling is impaired. The contracting muscle pulls glucose in whether insulin’s key works or not.[1]

    Where type 2 diabetes begins

    DeFronzo’s central claim, and it has held up well, is that skeletal muscle insulin resistance is “the initiating or primary defect” in type 2 diabetes, detectable decades before blood sugar rises. The cleanest evidence: lean, normal-glucose-tolerant young adults whose both parents have type 2 diabetes already show moderate-to-severe muscle insulin resistance, with glucose uptake reduced 33 to 43% at clamp insulin levels, explained entirely by reduced glycogen synthesis, while their livers suppress glucose normally. In lean people who already have diabetes, insulin-stimulated leg-muscle glucose uptake is reduced by about half.[4]

    Keep the nuance in mind. It limits everything that follows: “primary” in this literature means earliest detectable, not “the only cause.” Liver, fat tissue, gut, and pancreatic defects all contribute, and DeFronzo is explicit that beta-cell failure is the sine qua non: without it, diabetes never becomes overt, however insulin resistant the muscle. Muscle is where the disease starts. It is not the whole disease.

    More muscle, better insulin sensitivity, mostly association

    The population evidence linking muscle to metabolic health is large and consistent. In NHANES III (13,644 US adults), Preethi Srikanthan and Arun Karlamangla measured skeletal muscle mass relative to body weight and found that, after adjusting for age, ethnicity, sex, and both general and central obesity, each 10% increase in relative muscle mass was associated with an 11% relative reduction in insulin resistance (HOMA-IR) and a 12% relative reduction in the prevalence of prediabetes or overt diabetes.[5] A 2025 systematic review of 25 prospective cohort studies, 278,475 participants in the muscle mass analyses and 400,181 in the strength analyses, found that muscle mass relative to body weight, and grip strength relative to BMI, were consistently and inversely associated with developing type 2 diabetes.[6]

    The worst phenotype is the combination: sarcopenic obesity, excess fat with low muscle, which a large UK Biobank analysis reported as carrying the highest diabetes risk of any body-composition group, above obesity alone and above low muscle alone. That figure currently rests on a news summary of the paper rather than the paper itself, so treat the exact multiple as provisional.[7] A 16-year Korean cohort presented at ENDO 2024 points the same way. Diabetes risk rose as weight-adjusted muscle mass fell. But that is a conference abstract, not yet a full peer-reviewed paper.[8]

    Two honest limits on this section. First, the flagship NHANES analysis is cross sectional: it photographs one moment and cannot show that gaining muscle causes the improvement; its muscle mass was also estimated by bioelectrical impedance, an imprecise method. Second, and this matters for how you act on it, the intervention evidence does not isolate kilograms of muscle. Resistance training improves insulin sensitivity acutely and chronically through contraction-mediated glucose uptake, more GLUT4, and mitochondrial and vascular adaptations, including in people who gain little visible mass. Attribute the metabolic benefit to training and muscle quality, not to the scale of your lean mass alone.

    Muscle talks: the myokines

    If the glucose story is muscle as a sink, the myokine story is muscle as a broadcaster. A 2026 review in the Archives of Pharmacal Research sorts the field by evidence strength: Tier 1 myokines have demonstrable human interventional or causal evidence; Tier 2 have defined targets but largely correlative human data; Tier 3 is exploratory.[9] Only four signals make Tier 1: IL-6, the myostatin pathway, FGF21, and apelin. Tap through the five names you will meet most often:

    Explore the signals

    Five myokines, honestly graded

    Tiering: “Exercise-induced myokines in metabolic regulation.” Arch Pharm Res. 2026. Primary sources as cited per molecule.

    What this evidence supports: contracting muscle genuinely is a secretory organ, exercise changes the circulating levels of hundreds of proteins, and IL-6 anchors the concept in humans. What it does not support: a blood test that prescribes your training, a single molecule that explains exercise, or an “exercise pill.” No myokine mimetic is approved for metabolic disease or sarcopenia. The myokine field is young, and much of it is still animal work. The irisin story above is the cautionary tale the field itself tells.

    Aging: strength leaves three times faster than mass

    The word is sarcopenia, from the Greek sarx (flesh) and penia (loss). Irwin Rosenberg proposed it in 1989. The current European consensus (EWGSOP2, 2019) defines it as a muscle disease and made low strength, not low mass, the primary characteristic: low muscle quantity or quality confirms the diagnosis, and poor physical performance marks it severe.[10] The cutoffs the consensus uses:

    The diagnostic lines · EWGSOP2, 2019

    When low becomes a diagnosis

    Grip strength<27 kg men · <16 kg womenLow strength is the entry point: “probable sarcopenia.” Chair stand: >15 seconds for five rises is the alternative test.
    Muscle quantity<7.0 kg/m² men · <5.5 kg/m² womenAppendicular muscle mass divided by height squared, usually by DXA. Low quantity confirms the diagnosis.
    Walking speed≤0.8 m/sSlow gait on top of low strength and low mass means severe sarcopenia.

    Source: Cruz-Jentoft AJ et al. Age Aging. 2019;48(1):16 to 31. Sarcopenia has been a coded disease (ICD-10-CM M62.84) since 2016.

    How common is it? A 2022 meta-analysis of 151 studies and 692,056 people found prevalence of 10 to 27% depending on the definition used. Across the wider literature, reported prevalence runs from under 1% to over 85% purely because definitions differ. Anyone who gives you a single prevalence figure without naming the definition is selling something.[11]

    The key aging finding is a ratio. In the Health, Aging and Body Composition Study, 1,880 older adults followed for three years, leg strength fell at roughly 3% per year while leg lean mass fell at roughly 1% per year. Strength drains about three times faster than size, and the authors’ conclusion deserves emphasis: “maintaining or gaining muscle mass does not prevent aging-associated declines in muscle strength.” What leaks away is muscle quality: neural drive, fiber type, fat infiltrating the muscle, not just quantity.[12]

    Health ABC, annualized change

    The gap that defines aging muscle

    −3%/yrLeg strength. Range across the cohort: 2.6% per year (white women) to 4.1% per year (black men).
    −1%/yrLeg lean mass. Size shrinks slowly; force shrinks fast. That gap is muscle quality, and it is why strength, not size, is the clinical target.

    Source: Goodpaster BH et al. J Gerontol A Biol Sci Med Sci. 2006;61:1059 to 1064 (Health ABC).

    That is why strength, measured cheaply, predicts so much. In the same Health ABC cohort, Anne Newman and colleagues found that quadriceps strength predicted mortality, while muscle size by CT or DXA “was not strongly related to mortality” in the same people. Grip strength gave similar risk estimates to the quadriceps measurement.[13] The largest test of that idea is the PURE study: 139,691 adults aged 35 to 70 in 17 countries, followed a median of four years. Each 5 kg lower grip strength was associated with a 16% higher risk of death (HR 1.16, 95% CI 1.13 to 1.20), with similar figures for cardiovascular and non-cardiovascular death, and grip strength was “a stronger predictor of all-cause and cardiovascular mortality than systolic blood pressure.”[14] The outcomes data line up behind it: a 2026 meta-analysis of 39 cohort studies in community-dwelling older adults found sarcopenia associated with roughly 79% higher odds of death (OR 1.79) and 90% higher odds of functional decline (OR 1.90).[15]

    Two cautions on those striking numbers. First, PURE itself contains a useful negative: grip strength showed no significant association with developing diabetes, hospital admission for pneumonia or COPD, injury from falls, or fractures in that analysis. Grip does not predict everything. Second, all of this is observational. Weaker people die sooner, but illness also makes people weak, and the PURE authors say so themselves: research is still needed “to test whether improvement in strength reduces mortality.” Prediction is not yet proof that building strength buys back the risk.

    What actually builds and preserves it

    The guideline floor is the World Health Organization’s: on top of 150 to 300 minutes a week of moderate aerobic activity, adults should do muscle-strengthening activities involving all major muscle groups on two or more days a week; older adults add balance and functional training on three or more days.[16] Cohort data suggest the bar for benefit is lower than most people fear: a 2022 meta-analysis of 16 prospective studies found muscle-strengthening activity associated with 10 to 17% lower risk of death, cardiovascular disease, cancer, and diabetes. It followed a J-shaped dose response, with maximum benefit around 30 to 60 minutes a week, and the effect of much higher volumes is unclear.[17] Observational and mostly self reported, but a useful corrective to the idea that the dose must be heroic.

    And it works late. A 2026 meta-analysis of 25 randomized trials in 1,302 older adults with sarcopenia found resistance training improved strength (effect size 0.71), grip (0.55), and walking ability (0.41), while lean mass moved far less (0.22). A second meta-analysis in frail and pre-frail adults over 65 found the same pattern: lower-limb strength effect 0.93, muscle mass 0.29.[18][19] The classic proof of principle is older still: in 1994, Fiatarone and colleagues put nursing-home residents in their nineties on high-intensity progressive resistance training and improved their strength and function. The trial is cited for showing it is possible and effective in the very old, which is the point that matters.[20] The practical prescription the evidence supports: two to three full-body sessions a week, multi-joint movements for the major muscle groups, progressive load, roughly 2 to 4 sets of 6 to 15 reps at 60 to 80% of one-rep max, leaving a few reps in reserve.

    Protein: the RDA is a floor, not a target

    The 0.8 g/kg/day RDA is the minimum to avoid deficiency, not an optimum. Older muscle is “anabolically resistant” and needs more protein per meal to trigger building. The PROT-AGE study group’s evidence-based targets, the sports-nutrition position stand, and the largest protein meta-analysis give a consistent ladder:

    Grams per kilogram of body weight, per day

    Protein targets

    WhoTargetSource
    General adult, the RDA minimum0.8 g/kgInstitute of Medicine
    Healthy older adult, over 651.0 to 1.2 g/kgPROT-AGE, 2013
    Older adult who exercises≥1.2 g/kgPROT-AGE, 2013
    Older adult with acute or chronic illness1.2 to 1.5 g/kgPROT-AGE, 2013
    Exercising adult building or keeping muscle1.4 to 2.0 g/kgISSN position stand, 2017
    Plateau for training gains~1.6 g/kgMorton et al., 2018

    Bauer J et al. J Am Med Dir Assoc. 2013;14(8):542 to 559 (PROT-AGE); Morton RW et al. Br J Sports Med. 2018;52(6):376 to 384. Exception: severe kidney disease (eGFR <30, not on dialysis). Protein is restricted and individualized with a clinician.

    Do the arithmetic

    Your daily protein target

    Enter your body weight and pick the row that describes you. This uses only the published targets above. It is a starting point for a conversation, not a prescription.

    Two cautions before you chase a number. In Morton and colleagues’ meta-analysis of 49 trials, protein supplementation on top of training added, on average, 0.30 kg of fat-free mass. Real, but small; intakes above roughly 1.6 g/kg/day did not add further gains, and the effect shrank with age. Protein helps; training is the main driver, and 1.6 is a population plateau, not a personal minimum.[21] Spread matters too: PROT-AGE suggests roughly 25 to 30 g of protein per meal for older adults rather than the usual pattern of almost none at breakfast and a mountain at dinner. One safety point. These higher targets appear safe with normal kidney function in the expert reviews. People with kidney disease should set their target with their clinician, not with an article.

    Testosterone: a treatment for a diagnosis, not a tonic

    Because I prescribe hormone therapy, I get asked whether testosterone is the shortcut to keeping muscle. It is not, and the Endocrine Society’s 2026 statement is blunt about the boundary: replacement is for men with symptoms plus consistently low, accurately measured testosterone, at least two early-morning fasting tests, with a common threshold near 300 ng/dL, and not as population screening or an anti-aging prescription. In the trials, its benefits for muscle strength and physical function are described as potential and modest, alongside a consistent, modest benefit for libido.[22] The safety ledger from TRAVERSE, over 5,200 men treated for one to four years, found no meaningful increase in heart attack or stroke, but an approximately 50% relative increase in pulmonary embolism and an increased incidence of fractures, with long-term safety, including prostate cancer, still unestablished. The European menopause society’s position adds that testosterone should not be used routinely in older men to improve exercise capacity or cognition.[23] For overweight men whose low testosterone has no other cause, weight loss is typically first line. Testosterone treats diagnosed hypogonadism. Training and protein treat aging muscle.

    Where the slogan overreaches

    “Muscle is medicine” is a framing I use in my own book, and like any framing, it can be stretched past what the evidence holds. Here are the versions you will hear, with the verdicts the literature supports:

    The bottom line

    • Muscle is the main insulin-stimulated glucose sink, about 80% of disposal under clamp conditions, and a genuine secretory organ, with IL-6 anchoring a secretome of hundreds of peptides.
    • Relative muscle mass, and especially strength, are consistently associated with better insulin sensitivity, less type 2 diabetes, less disability, and lower mortality. Fat mass is the key confounder, and causation is not yet proven by a mortality trial.
    • Strength falls about three times faster than mass with age, and strength and function predict outcomes better than size. The goal is strong, functional muscle, not just a bigger DXA number.
    • The prescription that works: progressive resistance training at least two days a week, plus protein at 1.0 to 1.2 g/kg/day for healthy older adults (up to about 1.6 g/kg/day for those training hard). It works at 85 and beyond.
    • Testosterone is for diagnosed hypogonadism only. Myokine drugs for healthy aging do not exist. Muscle is a lever, powerful, modifiable, and undermeasured, alongside aerobic exercise, weight, sleep, and actual medical treatment.

    What we still don’t know

    • Whether building strength extends life. No trial has randomized adults to strength training with mortality as the endpoint; every mortality link above is observational, with confounding and reverse causation impossible to exclude.
    • The exact sarcopenic-obesity risk. The UK Biobank finding (highest diabetes risk of any body-composition group) was read for this article via a news summary; the primary Diabetes Care paper’s adjusted hazard ratios and definitions have not been verified here.
    • The Korean 16-year cohort linking falling muscle mass to incident diabetes exists so far as a 2024 conference abstract; the full peer-reviewed paper has not been verified.
    • Whether irisin does anything in humans. Its detection in human blood by mass spectrometry is reported in secondary sources; no human trial of irisin exists, and its fat-browning effect at real-world concentrations is unproven.
    • Per-meal protein dosing. The precise per-meal figures sometimes quoted for older adults (about 0.40 g/kg per meal, 2.5 to 3 g leucine) come from secondary summaries in this research, not verified primary papers, which is why this article uses PROT-AGE’s 25 to 30 g per meal instead.
    • Vitamin D, growth hormone, and sleep for muscle. All plausibly matter; no verified effect size or guideline for “optimizing” them for muscle was established in this research, so this article makes no claim for them.

    Sources

    1. Feraco A, et al. Exploring the role of skeletal muscle in insulin resistance: lessons from cultured cells to animal models. Int J Mol Sci. 2021;22:9327. pubmed.ncbi.nlm.nih.gov/34502235
    2. Pedersen BK, Febbraio MA. Muscles, exercise and obesity: skeletal muscle as a secretory organ. Nat Rev Endocrinol. 2012;8:457 to 465. ncbi.nlm.nih.gov/pubmed/22473333
    3. Pozefsky T, et al. Effects of brief starvation on muscle amino acid metabolism in nonobese man. J Clin Invest. 1976;57:444 to 449. doi:10.1172/JCI108295.
    4. DeFronzo RA, Tripathy D. Skeletal muscle insulin resistance is the primary defect in type 2 diabetes. Diabetes Care. 2009;32(Suppl 2):S157, S163. pmc.ncbi.nlm.nih.gov/articles/PMC2811436
    5. Srikanthan P, Karlamangla AS. Relative muscle mass is inversely associated with insulin resistance and prediabetes: findings from NHANES III. J Clin Endocrinol Metab. 2011;96(9):2898 to 2903. pubmed.ncbi.nlm.nih.gov/21778224
    6. Putranata H, Hengky A, Hartoko B. Unraveling the role of muscle mass and strength in predicting type 2 diabetes risk: a systematic review. Acta Diabetol. 2025;62(2):157 to 176. pubmed.ncbi.nlm.nih.gov/39760787
    7. Guan Z, et al. Sarcopenic obesity and risk of incident type 2 diabetes: a prospective cohort study and landmark analysis from the UK Biobank. Diabetes Care. 2026 (doi reported: 10.2337/dc26-0797). As summarised in: medicalxpress.com, news summary; primary paper not opened for this article.
    8. Mun H, et al. Skeletal muscle mass, insulin sensitivity, and beta-cell function in the development of type 2 diabetes: a 16-year prospective cohort study. J Endocr Soc. 2024;8(Suppl 1), ENDO 2024 conference abstract.
    9. Exercise-induced myokines in metabolic regulation: mechanisms, mimetics, and translational potential. Arch Pharm Res. 2026. doi:10.1007/s12272-026-01624-x. link.springer.com
    10. Cruz-Jentoft AJ, et al. Sarcopenia: revised European consensus on definition and diagnosis (EWGSOP2). Age Ageing. 2019;48(1):16 to 31. pmc.ncbi.nlm.nih.gov/articles/PMC6322506
    11. Petermann-Rocha F, et al. Global prevalence of sarcopenia and severe sarcopenia: a systematic review and meta-analysis. J Cachexia Sarcopenia Muscle. 2022;13:86 to 99. pubmed.ncbi.nlm.nih.gov/34816624
    12. Goodpaster BH, et al. The loss of skeletal muscle strength, mass, and quality in older adults: the Health, Aging and Body Composition Study. J Gerontol A Biol Sci Med Sci. 2006;61:1059 to 1064.
    13. Newman AB, et al. Strength, but not muscle mass, is associated with mortality in the Health, Aging and Body Composition Study cohort. J Gerontol A Biol Sci Med Sci. 2006;61(1):72 to 77. pubmed.ncbi.nlm.nih.gov/16456196
    14. Leong DP, et al. Prognostic value of grip strength: findings from the Prospective Urban Rural Epidemiology (PURE) study. Lancet. 2015;386(9990):266 to 273. pubmed.ncbi.nlm.nih.gov/25982160
    15. Zhao et al. Long-term impact of sarcopenia on functional decline and mortality in community-dwelling older adults: a systematic review and meta-analysis. Front Nutr. 2026;12:1652386. pmc.ncbi.nlm.nih.gov/articles/PMC12823505
    16. World Health Organization. WHO Guidelines on Physical Activity and Sedentary Behaviour. Geneva, 2020. who.int/publications/i/item/9789240015128
    17. Momma H, Kawakami R, Honda T, Sawada SS. Muscle-strengthening activities are associated with lower risk and mortality in major non-communicable diseases: a systematic review and meta-analysis of cohort studies. Br J Sports Med. 2022;56:755 to 763. doi:10.1136/bjsports-2021-105061. researchgate.net
    18. Tan Z, et al. Optimizing prescription of resistance training in older adults with sarcopenia: a systematic review and meta-analysis. Eur Rev Aging Phys Act. 2026;23:8. doi:10.1186/s11556-025-00399-2. link.springer.com
    19. Benefits of resistance training in frailty and sarcopenia: a systematic review and meta-analysis. J Clin Med. 2021;10:1630. mdpi.com/2077-0383/10/8/1630
    20. Fiatarone MA, et al. Exercise training and nutritional supplementation for physical frailty in very elderly people. N Engl J Med. 1994;330:1769 to 1775.
    21. Morton RW, et al. A systematic review, meta-analysis and meta-regression of the effect of protein supplementation on resistance training-induced gains in muscle mass and strength in healthy adults. Br J Sports Med. 2018;52(6):376 to 384. doi:10.1136/bjsports-2017-097608. ovid.com
    22. Endocrine Society. Statement on Testosterone Replacement Therapy. 16 July 2026, updated 9 September 2026. endocrine.org
    23. European Menopause and Andropause Society. Testosterone replacement therapy in older men: position statement. Maturitas. 2023;178:107854. emas-online.org (PDF)
    24. Additional sources cited in the myokine explorer: Pedersen BK, Febbraio MA. Muscle as an endocrine organ: focus on muscle-derived interleukin-6. Physiol Rev. 2008;88:1379 to 1406 (PMC5538267); McPherron AC, Lawler AM, Lee SJ. Regulation of skeletal muscle mass in mice by a new TGF-β superfamily member. Nature. 1997;387:83 to 90; Boström P, et al. A PGC1-α-dependent myokine that drives brown-fat-like development of white fat and thermogenesis. Nature. 2012;481:463 to 468; Timmons JA, et al. Is irisin a human exercise gene? Nature. 2012;488:E9, E10 (nature.com/articles/nature11364); Kanzleiter T, et al. The myokine decorin is regulated by contraction and involved in muscle hypertrophy. Biochem Biophys Res Commun. 2014;450(2):1089 to 1094 (pubmed.ncbi.nlm.nih.gov/24996176); Erickson KI, et al. Exercise training increases size of hippocampus and improves memory. Proc Natl Acad Sci USA. 2011;108:3017 to 3022 (hippocampal finding as reported in secondary summaries).

    Educational content only; not medical advice and not a substitute for an individual clinical assessment. Protein and training targets are population guidelines from the cited sources. Adjust them with your clinician if you have kidney disease, joint limitations, or other conditions. Figures are drawn from the cited studies as researched on 15 November 2026; items flagged in the text as conference abstracts, news summaries, or secondary reports had not been verified against their primary papers at that date.

  • “Sports TRT” Is a Steroid Cycle With Better Marketing

    “Sports TRT” Is a Steroid Cycle With Better Marketing

    The pitch is that a prescription and a clinic visit turn a performance dose of testosterone into medicine. They do not. The category is defined by the dose and the indication, not by who signed for it.

    The pitch arrives with a medical vocabulary. Not a cycle, a protocol. Not steroids, optimization. Not a dealer, a men’s clinic. The dose is higher than replacement, the goal is performance and physique, and the thing that is supposed to make it different from what a bodybuilder does in a garage is that a licensed provider wrote for it.

    I spent four years competing in Men’s Physique and earned my IFBB Pro card in 2017. I also prescribe testosterone every week as an internist. From both chairs, the honest description of “sports TRT” is this: it is a steroid cycle, run at a steadier dose, with a prescription attached. The prescription changes the legal exposure and the quality of the vial. It does not change the pharmacology.

    The verdict in one paragraph

    Replacement therapy and performance dosing are separated by the serum level you are aiming for, not by the setting. Guidelines define testosterone replacement as treating a diagnosed deficiency and targeting the middle of the normal range, roughly 450 to 600 ng/dL, using the lowest dose that gets there. Doses chosen to exceed that range are, by the Endocrine Society’s own definition in its scientific statement on performance-enhancing drugs, supraphysiologic androgen use. The health consequences that follow, erythrocytosis, suppressed fertility, cardiac remodeling, and a hypothalamic-pituitary axis that may not recover, are dose-dependent and do not care what the regimen is called.

    The claimVerdictWhy
    “It’s TRT because a doctor prescribed it”FalseGuidelines define the therapy by diagnosis and target range, not by prescriber. A supraphysiologic dose is supraphysiologic on a prescription pad
    “Pharmaceutical-grade product makes it safe”FalseSterility and accurate concentration are real advantages over underground product. They do not change the dose-dependent effects on hematocrit, fertility or the heart
    “Monitoring makes the risk manageable”Partly trueMonitoring catches erythrocytosis and allows dose reduction. It does not prevent testicular suppression, and no monitoring protocol has been shown to prevent the cardiac findings seen in long-term users
    “TRAVERSE proved testosterone is cardiac-safe”OverstatedTRAVERSE tested replacement dosing, titrated to 350 to 750 ng/dL, in men with confirmed deficiency. It says nothing about performance doses in men with normal baseline levels
    “Steady weekly dosing is safer than a classic cycle”Plausible, unprovenAvoiding the crash of an on-off cycle is a reasonable hypothesis. No trial has compared continuous supraphysiologic dosing against cycled dosing on hard outcomes
    “It’s a steroid cycle”True, when the dose exceeds replacementSupraphysiologic androgen administration to a man who is not deficient is the textbook definition, whatever the label on the box

    What testosterone replacement is, as medicine defines it

    Three elements define the therapy, and all three have to be present.

    A diagnosis. The American Urological Association’s guideline uses a total testosterone below 300 ng/dL as the cutoff supporting a diagnosis, measured on two separate early-morning draws, and makes the clinical diagnosis only when those low levels sit alongside symptoms or signs. [Mulhall et al., 2018] Thresholds differ a little by society. The Endocrine Society uses 264 ng/dL, the AACE has used 200. [AUANews, 2023] No society defines deficiency by how a man feels at a normal level.

    A target. The AUA directs clinicians to aim for the middle third of the normal range, commonly cited as about 450 to 600 ng/dL, using the minimum dose that achieves it. The Endocrine Society’s phrasing is the mid-normal range. TRAVERSE, the FDA-mandated safety trial, titrated to 350 to 750 ng/dL. [Lincoff et al., 2023] Every one of those targets sits inside the physiologic range of a healthy man.

    A monitoring schedule. Symptoms, testosterone, and hematocrit at baseline, at three to six months, then annually, with therapy held if hematocrit exceeds 54%, plus PSA assessment in men over 40. [Bhasin et al., 2018]

    Notice what is absent from all of this: any provision for a man with a normal baseline who wants to be stronger. That is not an oversight. It is the boundary of the indication.

    What “sports TRT” actually is

    The term has no formal definition, which is part of its usefulness to the people selling it. In practice it describes one of two things.

    The first is a man with a genuinely normal baseline, often 400 to 600 ng/dL, who is prescribed testosterone anyway, typically 150 to 250 mg per week, with the stated goal of reaching 900 to 1,200 ng/dL or higher. The second is a man with a borderline level who is started at a dose far above what would be needed to correct it, and then titrated upward by symptoms that were never specific to testosterone in the first place.

    Both are the same thing pharmacologically. You are adding androgen to a man who does not lack it, to push him above his own physiology. The Endocrine Society’s 2014 scientific statement on performance-enhancing drugs describes exactly this population and is blunt about who it is: “the vast majority of PED users are not athletes but rather nonathlete weightlifters, and the adverse health effects of PED use are greatly underappreciated.” [Pope et al., 2014]

    The one real difference. I want to be fair about what the clinic actually provides. A prescription means a sterile, accurately dosed, pharmacy-sourced product instead of an underground lab vial of unknown concentration. It means bloodwork. It means someone notices when your hematocrit hits 56%. Those are not nothing, and harm reduction is a legitimate framework in medicine. But harm reduction is an argument for supervizing a risk, not a claim that the risk is gone, and it is not what is being advertised.

    Where replacement ends: the dose-response data

    We do not have to speculate about where the line sits. The question was answered directly by Bhasin’s dose-response work, which remains the cleanest human data we have on graded testosterone exposure.

    Healthy young men received a GnRH agonist to shut down their own production, then weekly testosterone enanthate at 25, 50, 125, 300, or 600 mg for 20 weeks. The resulting mean nadir testosterone concentrations were 253, 306, 542, 1,345, and 2,370 ng/dL respectively. [Bhasin et al., 2001]

    Read that series carefully, because it maps the entire argument.

    Weekly doseMean nadir testosteroneChange in fat-free mass over 20 weeksWhere this sits
    25 mg253 ng/dL−0.1 kgBelow the normal range
    50 mg306 ng/dL+1.1 kgLow-normal
    125 mg542 ng/dL+3.4 kgMid-normal: the replacement target
    300 mg1,345 ng/dL+5.2 kgRoughly double the top of normal
    600 mg2,370 ng/dL+7.9 kgFrank supraphysiologic exposure

    Fat-free mass changes from Bhasin et al., 2001; nadir concentrations measured one week after the preceding injection, so peak levels were higher still.

    Two things follow. The first is that about 125 mg per week lands a healthy man squarely in the middle of the normal range. That is what replacement looks like in a controlled setting. The companion trial in men aged 60 to 75 found the same thing and said so explicitly, concluding that the best trade-off between benefit and adverse effects was achieved at the 125 mg dose. [Bhasin et al., 2005]

    The second is that the muscle keeps coming above that line. Fat-free mass rose from +3.4 kg at 125 mg to +7.9 kg at 600 mg. This is the uncomfortable part, and pretending otherwise destroys your credibility with anyone who has spent time in a gym: more androgen does build more muscle. It just does so by leaving the range the human body is built to run in, and everything else on the dose-response curve rises with it.

    What the higher doses actually do

    Red cells thicken the blood, predictably and by dose

    Erythrocytosis is the most common dose-limiting effect of testosterone therapy. In the same dose-response cohorts, hemoglobin and hematocrit rose in a linear, dose-dependent fashion in both young and older men, and the effect was more pronounced with age. [Coviello et al., 2008] In the older cohort, several men had to be withdrawn for hematocrit above 54%; none of the young men crossed that threshold even at 600 mg weekly. [Bhasin et al., 2005] This is why 54% is the stop-and-reassess number in the guidelines, and it is the single most reliable consequence of pushing the dose up.

    Fertility: suppression is the expected result, not a rare complication

    The clearest data on this come from an unexpected source. The World Health Organization ran weekly 200 mg testosterone enanthate as a male contraceptive in a ten-center, seven-country trial of 271 healthy fertile men. By six months, 157 of them, a cumulative rate of 65%, had reached azoospermia on three consecutive semen samples, and the regimen was then used as the couples’ only form of contraception for twelve months. [WHO Task Force, Lancet 1990] The follow-up study across 15 centers found suppression to azoospermia or severe oligozoospermia in 97.8% of 399 men. [WHO Task Force, Fertil Steril 1996]

    When a dose is effective enough at shutting down sperm production to be studied as birth control, calling that effect a side effect understates it. Median time from azoospermia back to a sperm concentration of at least 20 million/mL after stopping was 3.7 months in the 1990 trial, which is reassuring, but those were healthy men on a defined protocol for a defined period, not men three years into continuous use.

    The same trial recorded gynecomastia in 24 of 271 men. [WHO Task Force, Fertil Steril 1996] Aromatization tracks total dose. More testosterone means more substrate.

    The heart: the finding that changed my own thinking

    Baggish and colleagues at Massachusetts General compared 86 long-term anabolic steroid users against 54 non-using weightlifters, all experienced lifters aged 34 to 54, using echocardiography and coronary CT angiography. Men who were on-drug at evaluation had meaningfully reduced left ventricular systolic function compared with non-users, with an ejection fraction of 49±10% versus 58±10%, along with reduced diastolic function. Users also carried greater coronary plaque volume than non-users, and, critically, lifetime cumulative dose was strongly associated with plaque burden. [Baggish et al., Circulation 2017]

    That last clause is the one that matters for this discussion. A dose-response relationship with coronary atherosclerosis is not a finding you can supervise your way out of. It is an argument about total exposure over years, which is precisely what continuous “sports TRT” delivers, arguably more faithfully than a cycled approach does.

    Imaging findings raise the question of whether events follow. A 2025 Danish registry study answered it by tracking cardiovascular diagnoses rather than scans, following men sanctioned for AAS use against age- and sex-matched controls over a mean of 11 years. The adjusted hazard ratios were not subtle: acute myocardial infarction 3.00 (95% CI 1.67 to 5.39), coronary intervention or bypass grafting 2.95, venous thromboembolism 2.42, and arrhythmia 2.26. [Windfeld-Mathiasen et al., Circulation 2025]

    These men had a mean age around 27 at entry. A threefold rate of myocardial infarction in a cohort that young is not a subtle population signal. The same caveats apply as to any registry study, and one specific limitation matters here: the investigators could not track whether individuals continued using after their sanction, so the exposure is “ever sanctioned” not a measured dose over time.

    Recovery is not guaranteed

    A Danish case-control study compared 37 current users, 33 former users who had been off for a median of about 2.5 years, and 30 healthy controls. Former users still had significantly lower total testosterone than controls, 14.4 nmol/L (about 415 ng/dL) versus 18.8 nmol/L (about 542 ng/dL), and 27.2% of former users were below the lower reference limit years after stopping. [Rasmussen et al., PLoS One 2016]

    Roughly one in four men who stopped had not gotten their own production back. This is the risk that is almost never priced into the decision at the front end, and it is the one that most often walks into my office: a man in his late thirties who started “optimization” at 32, stopped, and now needs genuine replacement for an iatrogenic deficiency he did not have before.

    The population-level signal

    Two Danish registry studies, built on men sanctioned for doping in fitness centers, give a view no clinic cohort can. In the first, 545 users matched to 5,450 controls, mortality was three times higher among users (HR 3.0, 95% CI 1.3 to 7.0), and acne, gynecomastia and erectile dysfunction each affected more than 10% of users. [Horwitz et al., J Intern Med 2019] The larger follow-up tracked 1,189 sanctioned men, mean age 27.4, against 59,450 controls over a mean of 11 years: 2.8% of users died versus 1% of controls, a hazard ratio of about 2.8, with hazard ratios of 3.6 for unnatural deaths and 2.2 for natural ones. [Windfeld-Mathiasen et al., JAMA 2024]

    Read these honestly. These are observational cohorts of men sanctioned for doping, not of patients in supervised clinics. Confounding is substantial: risk-taking behavior, polypharmacy and other substance use plausibly drive a large share of the unnatural-death excess, and the authors say so. These studies cannot tell you the mortality risk of supervised supraphysiologic dosing. What they establish is that the population using these doses dies earlier, and that the gap is not small.

    The part nobody puts in the advertisement

    Everything above is a physical risk, and physical risks can be weighed. This one is different, because it changes who is doing the weighing.

    Androgens produce a dependence syndrome. Reviewing seven studies, Kanayama, Brower, Wood, Hudson and Pope concluded that about 30% of illicit AAS users develop dependence, defined as continued use despite adverse effects on physical, psychosocial or occupational functioning. [Kanayama et al., Addiction 2009] An online survey of men active on bodybuilding and steroid forums found 23.4% of 479 users meeting dependence criteria. [de Zeeuw et al., Eur Addict Res 2023]

    The mechanism is worth understanding, because it is not the one people expect. Androgens are not acutely intoxicating; nobody is chasing a high. The trap is built from the withdrawal state instead. Stopping leaves a man hypogonadal for weeks to months while his own axis recovers, and that state produces fatigue, low libido and genuine depressive symptoms. As the same group put it, men experiencing prolonged dysphoric effects or frank major depression from that withdrawal hypogonadism may desire to resume AAS, thus contributing to dependence. [Kanayama et al., Drug Alcohol Depend 2010]

    Why this matters more, not less, in a clinic setting. A classic cycle has a built-in end date. A prescription renews. The commercial model that sells continuous “optimization” removes the natural off-ramp, and it does so for a drug where roughly a third of users struggle to stop, where stopping feels terrible, and where the person you would turn to for an honest assessment of whether you still need it is also the person billing you monthly for it. I am not alleging bad faith across an industry. I am pointing out an incentive structure that would concern us immediately in any other area of medicine.

    If you want the plain test of whether this applies to you, it is not how much you take. It is how you feel about stopping. A man who can say “I would come off tomorrow if the labs said to” is in a different position from one who has already decided that is not on the table.

    What TRAVERSE does and does not say

    TRAVERSE is the trial that gets cited at me most often, usually incorrectly. It enrolled 5,198 men aged 45 to 80 with symptoms and two fasting testosterone levels below 300 ng/dL, all with existing or high risk of cardiovascular disease, and randomized them to transdermal testosterone titrated to 350 to 750 ng/dL or placebo. Major adverse cardiac events occurred in 7.0% of the testosterone group versus 7.3% on placebo (HR 0.96, 95% CI 0.78 to 1.17), meeting noninferiority. [Lincoff et al., NEJM 2023]

    That is a genuinely important and reassuring result, and I use it to counsel patients who need treatment. But notice the three constraints built into it: diagnosed deficiency, a target inside the normal range, and a 33-month horizon. None of them describe sports TRT.

    And even inside those constraints, the trial was not uniformly clean. Significantly more men on testosterone had atrial fibrillation (3.5% vs 2.4%, P=.02), acute kidney injury (2.3% vs 1.5%, P=.04), pulmonary embolism (0.9% vs 0.5%), and nonfatal arrhythmias warranting intervention (5.2% vs 3.3%, P=.001). The investigators wrote that these events were not expected. [Cleveland Clinic summary] The trial’s own lead authors cautioned that the findings “should not be used as a justification for widespread prescription of these products to large numbers of men.” [Cleveland Clinic newsroom, 2023]

    Using a replacement-dose trial to vouch for performance dosing is a category error. If anything, the arrhythmia and thromboembolic signals at physiologic targets argue for more caution above them, not less.

    The definition anti-doping bodies already use

    There is one arena where this line has been drawn precisely, with money and careers on it, and it is worth knowing where it sits. Under WADA’s therapeutic use exemption framework, a TUE for testosterone is granted only for organic primary or secondary hypogonadism, a structural or genetic cause. Functional hypogonadism does not qualify. As the guidance is summarized for masters athletes: “TUE should only be approved for hypogonadism that has an organic etiology. TUE should not be approved for androgen deficiency due to functional disorder.” Andropause is explicitly not an acceptable diagnosis.

    USADA’s physician worksheet is blunter still: low levels of an endogenous hormone “absent a defined organic etiology will not justify the granting of a TUE, even if within standard of care & prescribed by a physician.”

    Read that last clause again, because it is the entire argument of this article stated by someone else: even if prescribed by a physician. Anti-doping authorities worked out decades ago that a prescription does not change what a substance does. They had to, because the alternative was a system where anyone with a sympathetic clinic could compete on exogenous androgen and call it medicine.

    Why the clinic model is the thing to look at

    It would be convenient to frame this as rogue operators versus real medicine. The more useful question is what the business model rewards, because that predicts the clinical behavior better than anyone’s intentions do.

    Consider that the FDA’s own review of testosterone prescribing, the review that ultimately led it to mandate the TRAVERSE trial, found that “many men are treated with testosterone products even though their testosterone blood levels have never been checked.” That is the baseline problem in this field, and it predates the current wave of telehealth.

    Now add a subscription structure. In a model where revenue depends on continuing therapy, four things tend to follow, and you can check any clinic against them:

    • The diagnostic threshold drifts upward. If 300 ng/dL disqualifies most prospective patients”optimal” ranges that start at 600 or 800 solve a business problem elegantly. Watch for any clinic whose stated target range has no source.
    • Symptoms do the diagnostic work. Fatigue, low mood, poor sleep and reduced libido are real, common, and caused by many things. A questionnaire that routes nearly everyone toward treatment is a sales funnel wearing a lab coat.
    • Reversible causes go unexamined. Treating the sleep apnea, the obesity, the alcohol intake or the untreated depression produces a patient who does not need a monthly prescription. That is a clinical success and a revenue loss.
    • Deprescribing never comes up. In legitimate practice”let’s see whether you still need this” is a routine sentence. Ask when you last heard it.

    If you are evaluating a clinic, five questions will tell you most of what you need to know. What was my pre-treatment level, measured twice, in the morning? What is the target range you are treating me to, and whose guideline is that from? Did you check LH, FSH, SHBG and prolactin to find out why my level is low? What is your plan if my hematocrit reaches 54%? And if I want children in five years, what changes today?

    A clinic practicing real medicine answers all five without friction. I would not stay at one that cannot.

    The strongest case, both ways

    The strongest case for the clinics

    • Men are going to use these drugs regardless. Supervision with real labs and pharmaceutical product is safer than an underground vial and a forum protocol.
    • The 300 ng/dL cutoff was derived largely from trials in men over 45 and fits a symptomatic 35-year-old poorly. A level of 320 is not automatically fine.
    • Continuous steady dosing avoids the post-cycle crash, which is when much of the psychiatric and cardiovascular turbulence of classic cycling appears.
    • Medicine has always treated conditions that sit on a spectrum rather than a cutoff, and the line at 300 is a convention, not a biological cliff.

    Where that case breaks

    • Harm reduction describes supervizing a risk. It does not license marketing the result as a health intervention, which is what is happening.
    • The cutoff debate is an argument for better diagnosis, not for targeting 1,200 ng/dL. Nobody resolves a flawed threshold by tripling the target.
    • Continuous dosing means greater cumulative lifetime exposure, which is the variable Baggish found tracked with coronary plaque.
    • Roughly a quarter of former users had not recovered their own production years later. Started at 30, that is a four-decade prescription.
    • Roughly 30% develop dependence. A model with no end date and a monthly bill is poorly designed to catch that.

    I hold both of these at once, and I think any honest clinician does. A man determined to use these compounds is better off with real labs, sterile product and a physician watching his hematocrit than with a forum protocol and a vial from a stranger. What I object to is not supervision. It is supervision marketed as optimization, which turns a risk being managed into a benefit being sold.

    What I tell patients

    I do not open with a lecture, because the man in front of me usually already knows what he is asking for. I work through four things.

    First, get the diagnosis right. Two early-morning draws, fasting, on separate days, plus LH, FSH, SHBG, prolactin and a hematocrit. Free testosterone when SHBG is abnormal. A single afternoon total testosterone from a direct-to-consumer panel is not a diagnosis.

    Second, treat the things that lower testosterone. Roughly half of men with obesity have reduced levels, [Cleveland Clinic J Med, 2024] and sleep apnea, alcohol, opioids, overtraining and untreated depression all pull the number down. Several of my patients who arrived asking for testosterone did not need it six months later.

    Third, if you are deficient, treat properly. Lowest dose that reaches mid-normal, hematocrit and PSA on schedule, fertility discussed and hCG or a SERM considered before starting if children are a possibility, not after.

    Fourth, if you want performance doses, I will tell you what they are. I am not going to pretend a 200 mg weekly protocol aimed at 1,100 ng/dL in a man with a baseline of 520 is replacement therapy. I will tell him it is a cycle, that it will likely suppress his fertility, that his hematocrit will climb, that a quarter of men in his position do not recover their own production, and that the cardiac data on long-term users is genuinely concerning. Then it is his decision, made with accurate information. That is a different conversation than the one he gets from an ad.

    Bottom line

    Testosterone replacement therapy is one of the more satisfying treatments in medicine when it is used for what it is for. A man with real deficiency, treated to a normal level and monitored properly, often gets his life back. I have watched it happen many times, and the TRAVERSE data should make physicians less skittish about providing it.

    “Sports TRT” is a different product wearing that product’s clothes. The dose is above replacement, the indication is appearance and performance, and the clinical category it belongs to is supraphysiologic androgen use. You are allowed to make that choice as an adult. What you are not entitled to is the belief that a prescription repeals the dose-response curve.

    References

    1. Mulhall JP, et al. Evaluation and Management of Testosterone Deficiency: AUA Guideline. J Urol. 2018;200(2):423-432. PubMed 29601923
    2. Bhasin S, et al. Testosterone Therapy in Men With Hypogonadism: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2018;103(5):1715-1744. PubMed 29562364
    3. Bhasin S, et al. Testosterone dose-response relationships in healthy young men. Am J Physiol Endocrinol Metab. 2001;281(6):E1172-E1181. PubMed 11701431
    4. Bhasin S, et al. Older men are as responsive as young men to the anabolic effects of graded doses of testosterone on the skeletal muscle. J Clin Endocrinol Metab. 2005;90(2):678-688. PubMed 15562020
    5. Coviello AD, et al. Effects of graded doses of testosterone on erythropoiesis in healthy young and older men. J Clin Endocrinol Metab. 2008;93(3):914-919. PubMed 18160461
    6. World Health Organization Task Force on Methods for the Regulation of Male Fertility. Contraceptive efficacy of testosterone-induced azoospermia in normal men. Lancet. 1990;336(8721):955-959. PubMed 1977002
    7. World Health Organization Task Force. Contraceptive efficacy of testosterone-induced azoospermia and oligozoospermia in normal men. Fertil Steril. 1996;65(4):821-829. PubMed 8654646
    8. World Health Organization Task Force. Effects of testosterone enanthate in normal men: experience from a multicenter contraceptive efficacy study. Fertil Steril. 1996;65(3):626-636. PubMed 8774299
    9. Baggish AL, et al. Cardiovascular Toxicity of Illicit Anabolic-Androgenic Steroid Use. Circulation. 2017;135(21):1991-2002. PubMed 28533317 · PMC5614517
    10. Rasmussen JJ, et al. Former Abusers of Anabolic Androgenic Steroids Exhibit Decreased Testosterone Levels and Hypogonadal Symptoms Years after Cessation. PLoS One. 2016;11(8):e0161208. PMC4988681
    11. Horwitz H, Andersen JT, Dalhoff KP. Health consequences of androgenic anabolic steroid use. J Intern Med. 2019;285(3):333-340. PubMed 30460728
    12. Windfeld-Mathiasen J, et al. Mortality Among Users of Anabolic Steroids. JAMA. 2024;331(14):1229-1230. JAMA
    13. Windfeld-Mathiasen J, et al. Cardiovascular Disease in Anabolic Androgenic Steroid Users. Circulation. 2025. doi:10.1161/CIRCULATIONAHA.124.071117. Record and abstract
    14. Kanayama G, Brower KJ, Wood RI, Hudson JI, Pope HG Jr. Anabolic-androgenic steroid dependence: an emerging disorder. Addiction. 2009;104(12):1966-1978. PMC2780436
    15. Kanayama G, Brower KJ, Wood RI, Hudson JI, Pope HG Jr. Treatment of anabolic-androgenic steroid dependence: emerging evidence and its implications. Drug Alcohol Depend. 2010;109(1-3):6-13. PMC2875348
    16. de Zeeuw TI, Brunt TM, van Amsterdam J, et al. Anabolic Androgenic Steroid Use Patterns and Steroid Use Disorders in a Sample of Male Gym Visitors. Eur Addict Res. 2023;29(2):99-108. Karger
    17. Pope HG Jr, et al. Adverse Health Consequences of Performance-Enhancing Drugs: An Endocrine Society Scientific Statement. Endocr Rev. 2014;35(3):341-375. PubMed 24423981
    18. Lincoff AM, et al. Cardiovascular Safety of Testosterone-Replacement Therapy (TRAVERSE). N Engl J Med. 2023;389(2):107-117. NEJM · PubMed 37326322
    19. Testosterone deficiency diagnosis and thresholds. Cleve Clin J Med. 2024;91(2):93-95. CCJM
    20. WADA TUE Physician Guidelines, Male Hypogonadism; USADA Testosterone TUE Physician Worksheet. USADA worksheet · WMA summary of WADA TUEC guidelines

    Educational content only. This article is not medical advice for any individual and does not establish a physician-patient relationship. Do not start, stop or change testosterone therapy without consulting your own physician. If you are currently on a supraphysiologic protocol, stopping abruptly carries its own risks and should be managed clinically.