Testosterone replacement is an effective treatment for a real condition. It is also given to a great many men who do not have that condition, in a market that has strong financial reasons not to look too closely.
Both things are true, and this page tries to hold both.
What TRT is, and what it isn’t
Replacement restores testosterone to the normal range in a man whose own production is inadequate. That is what the drug is approved for and what the evidence supports.
Optimization pushes a man with normal or borderline testosterone toward the upper end of the range, or above it, on the theory that more is better. There is no trial evidence that this improves outcomes, and the phrase exists mostly to make it saleable.
Enhancement uses supraphysiologic doses for muscle or performance. That is a different drug regimen with a different risk profile, covered at Performance Enhancement.
The line between the first and the second is where most of the argument in this field actually sits.
Who benefits
The Endocrine Society’s July 2026 statement is direct: testosterone therapy has clear benefits for men with appropriately diagnosed hypogonadism caused by testicular, pituitary or hypothalamic disease.
That’s the population. Men with a documented cause, consistently low correctly-measured testosterone, and symptoms consistent with deficiency — especially sexual symptoms, which are the ones that actually track testosterone.
Benefits with reasonable evidence behind them: sexual desire and sexual activity, correction of anemia, bone mineral density, lean mass and fat mass, and modest improvement in mood in men with mild-to-moderate symptoms.
Benefits frequently claimed and not supported: erectile function in men with vascular disease, glycemic control, fracture prevention, cardiovascular protection, and longevity. Each of those is covered at Testosterone, the Prostate and the Heart.
Who shouldn’t receive testosterone
The Endocrine Society’s 2018 guideline lists conditions where testosterone should not be started, or started only with specialist input:
- Breast cancer or prostate cancer
- A palpable prostate nodule or induration, or a PSA above 4 ng/mL — or above 3 ng/mL in men at higher risk — without urological evaluation
- Elevated hematocrit at baseline
- Untreated severe obstructive sleep apnea
- Severe lower urinary tract symptoms
- Uncontrolled or severe heart failure
- Myocardial infarction or stroke within the previous six months
- Thrombophilia
- A man planning conception in the near term — see TRT and Fertility
We were unable to verify the exact numeric hematocrit and urinary-symptom-score thresholds in the guideline text and are not going to invent them. The hematocrit thresholds that are well established are for stopping or reducing during treatment, and appear at Testosterone, the Prostate and the Heart.
The fertility item deserves emphasis because it is the one most often skipped, and unlike the others it is invisible until it matters.
How quickly it works
Different effects arrive on different schedules. A widely used synthesis published in the European Journal of Endocrinology in 2011 by Saad and colleagues mapped the timelines:
| Effect | Onset | Maximum |
|---|---|---|
| Sexual desire | ~3 weeks | plateaus around 6 weeks |
| Erections and ejaculation | slower | may take up to 6 months |
| Quality of life | 3–4 weeks | later |
| Depressive mood | 3–6 weeks | 18–30 weeks |
| Lipids | 4 weeks | 6–12 months |
| Fat mass, lean mass, strength | 12–16 weeks | stabilizes 6–12 months, marginal gains thereafter |
| Glycemic control | — | 3–12 months |
| Erythropoiesis / hematocrit | 3 months | peaks 9–12 months |
| Bone density | detectable at 6 months | continues beyond 3 years |
| PSA and prostate volume | — | small rise, plateaus around 12 months |
Two things to take from this. Libido responds first and fastest — within weeks. And body composition takes months, which means a man who feels nothing at eight weeks has not necessarily failed treatment.
This is a synthesis of earlier trials rather than a single study, and it gives timing rather than effect size. Treat it as a map of what to expect when, not a promise of how much.
The formulations
No formulation is best. The right one depends on how stable you need levels to be, how often you’re willing to dose, cost, fertility plans, and whether there are children in the house.
Injectable — cypionate and enanthate
The most commonly used regimen in the United States. Given intramuscularly or subcutaneously, typically weekly or every two weeks.
Levels peak roughly two to five days after injection and fall to a trough before the next dose. Longer intervals produce a bigger swing, and some men feel that swing as a cycle of good days and flat days. More frequent, smaller doses flatten it.
Cheap — testosterone cypionate cash prices at major US pharmacies run roughly $20 to $32 per vial — and effective. The peaks drive the highest rates of erythrocytosis of any route.
Xyosted — subcutaneous enanthate auto-injector
Weekly, self-administered. Boxed warning for blood pressure increases. Depression and suicidal ideation appear in the warnings and precautions section — not in the boxed warning. We flag this because the two are frequently conflated, including in earlier drafts of our own research.
Aveed — intramuscular testosterone undecanoate
Long-acting, roughly every ten weeks after loading doses. Convenient in principle. Boxed warning for pulmonary oil microembolism and anaphylaxis, requiring a REMS programme with prescriber and facility certification and a 30-minute post-injection observation period. That requirement is the practical barrier.
Transdermal gels — AndroGel, Testim, Fortesta, Vogelxo
Daily application to shoulders, upper arms or abdomen. Smooth levels, no needles, easily stopped — and the formulation used in TRAVERSE, so the cardiovascular data apply most directly here.
Boxed warning for secondary exposure. Virilization has been reported in children exposed to testosterone gel through skin contact. Wash hands, cover the site once dry, and take this seriously if there are children or a pregnant partner in the household. Transfer to an adult partner is possible by the same route — see testosterone in women.
Androderm — transdermal patch
Daily, stable levels, no transference risk. No boxed warning. Skin reactions are the limiting factor and they are common: on older dosing, pruritus in 37%, blistering in 12%, and site reactions overall in 48%.
Natesto — nasal gel
Three times daily, which is the obvious drawback. The interesting property is that its short-acting pharmacokinetics appear to leave LH and FSH within the normal range and preserve semen parameters better than other exogenous testosterone routes — a genuinely distinctive feature for a man who wants replacement without shutting down his own axis. Published and replicated in direction; we could not extract the specific figures and are not quoting any.
Testopel — pellets
Implanted under the skin every three to six months. Nothing to remember. Least predictable release kinetics of the approved options, and the pellets can extrude or the site can become infected. Published rates exist; we could not verify them and will not guess.
Oral undecanoate — Jatenzo, Tlando, Kyzatrex
Twice daily, with food. These are absorbed through the intestinal lymphatics, bypassing the liver, and are not 17-alpha-alkylated — which means the hepatotoxicity that made older oral androgens dangerous does not apply. That distinction is real and worth understanding.
They carry a blood pressure warning and are not recommended in men with uncontrolled hypertension. The food requirement is not optional: for Kyzatrex, absorption rose 37% with a 16%-fat meal, 87% with 33% fat, and 94% with 45% fat, relative to fasting. Take it without food and you may absorb very little.
Striant — buccal
Twice daily against the gum. It retains an active FDA label but appears to have largely left the market. We could not confirm current availability.
What counts as a response
The Endocrine Society directs treatment toward the mid-normal range. We could not verify a specific target in ng/dL from the guideline text and won’t publish one — the commonly quoted figures are not traceable to the source.
More useful: a response means symptoms improved. Testosterone in range with no symptomatic change is not success. If a man is at target and feels no different, the answer is not more testosterone. It is to revisit whether testosterone was the problem — see When Testosterone Isn’t the Answer.
A useful benchmark: in TRAVERSE, 61% of men in both the testosterone and placebo arms discontinued. In real-world claims data covering 15,435 men on testosterone gel, only 34.7% were still on treatment at six months and 15.4% at twelve. Roughly five in six men stopped within a year.
That is not what you would expect from a drug that reliably makes men feel better. Some of it is cost, inconvenience and gel messiness. Some of it is that a substantial share of men who start testosterone did not have a testosterone problem.
Adverse effects
Erythrocytosis. The commonest reason treatment is modified. Injectables most, gels least. See Testosterone, the Prostate and the Heart.
Acne. Reported in 0.6% to 9.1% of men across studies, varying by route: oral lowest at 0.6–2.9%, transdermal gel around 5.3%, injectables highest at 6.2–8.6%. Some men discontinue because of it.
Testicular volume loss. Suppressed LH means reduced testicular activity and reduced volume. In a series of 121 men followed for two years, mean testicular volume fell from 16.5 mL to 13.7 mL — about 17% — with the reduction becoming statistically significant from six months, and injectable formulations a risk factor. This was a conference abstract rather than a full paper, so treat the figures as preliminary. hCG co-administration substantially mitigates it.
Gynecomastia. Caused by aromatization of testosterone to estradiol. The mechanism is not in doubt. We could not find trial-level incidence data and are not quoting a rate. Routine aromatase inhibition to prevent it is not supported — see Testosterone, Estradiol and DHT.
Hair loss. Here the honest answer is that nobody knows. A 2025 scoping review found the evidence insufficient either to confirm or refute that testosterone accelerates male-pattern baldness. Mechanistically it is plausible in genetically predisposed men. Anyone who tells you confidently in either direction is going beyond the data.
Blood pressure. A class-wide FDA warning since February 2025.
Fertility suppression. Near-universal and the most consequential. See TRT and Fertility.
Stopping
Two questions get asked constantly and the evidence base is thinner than you’d hope.
Can you stop? Yes. The axis has been suppressed and takes months to restart, so there is typically a period of symptomatic deficiency — potentially worse than before treatment, because you are now below your own baseline while the pituitary recovers. That is a transitional state, not a new permanent one.
Will my testosterone end up lower than before I started? This is the fear that circulates most widely, and we could find no credible peer-reviewed evidence supporting it as a general phenomenon. The suppression is expected and reversible in most men without pre-existing axis pathology. What we cannot tell you is how long recovery takes in men treated therapeutically for years, or what fraction fully recover — the good recovery data come from contraceptive trials in healthy young men on defined-duration treatment, which is a different population. See TRT and Fertility for those figures and their limits.
Is it lifelong? If the cause is permanent — Klinefelter syndrome, testicular loss, a structural pituitary problem — then yes, in the same sense that thyroid replacement is lifelong. If the cause is reversible and gets reversed, it may not be. The honest answer depends entirely on why your testosterone was low, which is why Why Is Testosterone Low? matters more than it appears to.
What remains uncertain
Whether “optimization” above mid-normal does anything. No trial evidence. The market runs ahead of the data.
Long-term outcomes beyond a few years. The largest trial ran under two years of mean treatment.
Whether injectable formulations carry the same cardiovascular profile as gel. TRAVERSE tested gel. Most American men use injections.
Whether testosterone accelerates male-pattern baldness. Genuinely unresolved.
Recovery after prolonged therapeutic use. Not characterized.
Why five in six men stop within a year. Cost and inconvenience explain some of it. Not all.
Optimal target range. Guidelines say mid-normal. What that means numerically, and whether it matters, is not well established.
Questions patients ask
TRT is basically anabolic steroids.
Same molecule, very different dose and purpose.
What the evidence showsReplacement targets the normal physiological range in a man who is deficient. Anabolic steroid use typically involves doses many times higher, often with multiple compounds, in men with normal testosterone. The risk profiles are not comparable.
What remains uncertainWhere exactly replacement ends and enhancement begins is genuinely contested, especially in "optimization" practice.
Bottom lineReplacement is not steroid use. Supraphysiologic dosing is, whoever prescribes it.
Strong
TRT will make me big.
It improves body composition in deficient men. It doesn't build a bodybuilder.
What the evidence showsLean mass and strength changes appear from 12–16 weeks and stabilize by 6–12 months. The dramatic physique changes associated with steroids come from supraphysiologic dosing, not replacement.
What remains uncertainIndividual variation in response is wide.
Bottom lineExpect a moderate change in body composition, not a transformation.
Moderate
TRT causes weight loss.
It shifts fat and muscle around. It is not a weight-loss drug.
What the evidence showsFat mass and lean mass changes take 3–4 months to appear. Scale weight often moves little because lean mass increases as fat falls.
What remains uncertainWhether body composition changes translate to clinical benefit.
Bottom lineIf weight loss is the goal, this is the wrong drug. See Obesity, GLP-1 Medications and Testosterone.
Moderate
TRT gives you more energy.
Sometimes — and fatigue was not independently associated with testosterone in the strongest population study available.
What the evidence showsQuality of life improves from 3–4 weeks in the published timeline. But the European Male Ageing Study screened 32 symptoms in 3,369 men and found only three sexual symptoms independently associated with testosterone; fatigue tracked age and comorbidity.
What remains uncertainHow much of the energy improvement men report is a direct hormonal effect.
Bottom linePossible. Not the effect with the strongest evidence, despite being the most advertised.
Limited
How quickly does TRT work?
Libido in about three weeks. Body composition in three to four months. Bone density in years.
What the evidence showsThe published timeline gives sexual desire onset at around 3 weeks plateauing near 6; mood 3–6 weeks with maximum at 18–30 weeks; fat and lean mass 12–16 weeks; bone density detectable at 6 months and continuing past 3 years.
What remains uncertainThese are timings from a synthesis of earlier trials, not effect sizes.
Bottom lineJudge sexual symptoms at six weeks. Judge body composition at six months.
Moderate
TRT will cure my erectile dysfunction.
Usually not, particularly if you're older or diabetic.
What the evidence showsIn TRAVERSE, testosterone improved sexual desire and activity with **no significant improvement in erectile function**. Meta-analysis suggests the erectile benefit concentrates in men with more severe deficiency — below about 8 nmol/L — and is near-absent in mild cases.
What remains uncertainWhich individual men have a hormonally responsive component.
Bottom lineExpect desire to improve. Don't count on erections. See [Testosterone, Libido and Sexual Function](/sexual-health/).
Strong
TRT is lifelong.
It depends entirely on why your testosterone was low.
What the evidence showsPermanent causes — Klinefelter syndrome, testicular loss, structural pituitary disease — mean permanent replacement. Reversible causes may not, if the cause is actually addressed.
What remains uncertainHow reliably men recover adequate function after prolonged replacement.
Bottom lineIf nobody established why your testosterone was low, nobody can tell you whether this is lifelong.
Moderate
You can't stop TRT.
You can. There is usually a rough transitional period.
What the evidence showsThe axis is suppressed and takes months to restart, so symptoms may be worse than baseline temporarily. Recovery data in men treated therapeutically for years are poor.
What remains uncertainDuration of recovery and what fraction fully recover.
Bottom lineStopping is possible and is not comfortable. Plan it with a clinician rather than just quitting.
Limited
My testosterone will be permanently lower than before if I stop.
We found no credible evidence for this.
What the evidence showsSuppression during treatment is expected. Recovery follows in most men without pre-existing axis pathology. The claim that baseline ends up permanently lower appears to be a forum belief rather than a documented finding.
What remains uncertainRecovery after many years of therapy is genuinely not well characterized, so "we found no evidence" is not the same as "this never happens."
Bottom lineThe transitional dip is real. The permanent downgrade is not evidenced.
Limited
Testosterone injections damage the testicles.
They shrink from disuse. That is not damage in the structural sense.
What the evidence showsIn 121 men followed two years, mean testicular volume fell from 16.5 mL to 13.7 mL — around 17% — significant from six months, with injectables a risk factor. This was a conference abstract, so treat as preliminary.
What remains uncertainHow completely volume recovers after stopping. We could not verify recovery data.
Bottom lineExpect shrinkage. hCG substantially reduces it.
Limited
TRT causes acne.
In a minority, and it depends on the formulation.
What the evidence showsReported incidence ranges from 0.6% to 9.1%: oral lowest at 0.6–2.9%, gel around 5.3%, injectables 6.2–8.6%. Some men discontinue over it.
What remains uncertainWhether dose or peak level drives it.
Bottom lineUncommon overall, more likely with injections, often manageable by switching route.
Moderate
TRT causes hair loss.
Possibly in predisposed men. The evidence genuinely doesn't settle it.
What the evidence showsA 2025 scoping review concluded the available evidence is insufficient either to confirm or refute that testosterone accelerates male-pattern baldness. One trial noted scalp thinning; another reported abnormal hair growth in up to 5.3%.
What remains uncertainAlmost all of it.
Bottom linePlausible if you're genetically predisposed. Anyone certain in either direction is overstating.
Limited
TRT causes gynecomastia.
It can, through conversion to estradiol.
What the evidence showsThe mechanism — aromatization of testosterone to estradiol — is well established. We could not find reliable trial-level incidence figures.
What remains uncertainHow often it occurs, and who is susceptible.
Bottom lineA recognized effect. Not a reason for routine aromatase inhibition.
Moderate on mechanism, Limited on frequency
Everyone on TRT needs an aromatase inhibitor.
No, and routine use has a real downside.
What the evidence showsNo controlled evidence supports routine co-prescription. Estradiol is the dominant regulator of bone resorption in men, and aromatase inhibitor trials show adverse bone-turnover signals.
What remains uncertainWhich minority of men with genuine estrogen-excess symptoms benefit.
Bottom lineOne of the clearest gaps between clinic practice and evidence. See Testosterone, Estradiol and DHT.
Limited to Unsupported for routine use
Which formulation is best?
There isn't one, and a clinic that offers only one is telling you about its business model.
What the evidence showsEach has a distinct profile. Injectables are cheap with the biggest peaks and most erythrocytosis. Gels give smooth levels and carry a boxed warning for transferring to children. Patches avoid transference but cause skin reactions in nearly half. Nasal dosing appears to preserve fertility but requires three doses daily. Pellets need nothing remembered and have the least predictable kinetics. Orals need food and raise blood pressure.
What remains uncertainWhether the cardiovascular reassurance from TRAVERSE, which studied gel, extends to injections.
Bottom lineMatch the formulation to the man.
Strong
Weekly injections are better than every two weeks.
More frequent dosing gives steadier levels, which many men prefer.
What the evidence showsLevels peak two to five days after injection and trough before the next. Longer intervals mean bigger swings. Some men experience the swing symptomatically.
What remains uncertainWhether steadier levels improve hard outcomes or only comfort.
Bottom lineIf you feel good early in the cycle and flat at the end, more frequent smaller doses is the standard fix.
Moderate
Testosterone gel is safe to use around my family.
Only with proper precautions, and this carries a boxed warning for a reason.
What the evidence showsVirilization has been reported in children secondarily exposed to testosterone gel through skin contact. Apply to shoulders, upper arms or abdomen; wash hands; cover the site once dry. A partner can be exposed by the same route; see [testosterone in women](/testosterone-in-women/).
What remains uncertainNothing about the risk. It is documented and it is why the warning exists.
Bottom lineWith young children at home, consider a patch, injection or nasal formulation instead.
Strong
Oral testosterone destroys your liver.
The old ones did. The current approved ones are a different drug design.
What the evidence showsOlder oral androgens were 17-alpha-alkylated to survive first-pass hepatic metabolism, which caused direct liver toxicity. Jatenzo, Tlando and Kyzatrex are testosterone undecanoate absorbed through the intestinal lymphatics, bypassing the liver, and are not 17-alpha-alkylated. They do carry a blood pressure warning.
What remains uncertainLong-term data are shorter than for injectables and gels.
Bottom lineThe hepatotoxicity concern does not apply to the modern orals. The blood pressure concern does.
Strong
Is a telehealth TRT prescription legitimate?
It can be. Ask what was actually done before the prescription.
What the evidence showsResearch presented at ENDO 2026 found only 12% of 200 men starting testosterone had a guideline-concordant workup — two low morning samples, LH and FSH, contraindication screening. The route of prescribing matters less than whether those steps happened.
What remains uncertainHow prescribing quality actually compares between telehealth and in-person care. We could not find reliable comparative data.
Bottom lineTwo morning tests, LH and FSH, a cause, and a fertility conversation. If those happened, the platform is secondary. If they didn't, so is the convenience.
Moderate
If TRT isn't working, I need a higher dose.
Usually the opposite conclusion is warranted.
What the evidence showsGuidelines direct re-evaluation for other causes of symptoms rather than dose escalation. If testosterone is at target and symptoms have not moved, testosterone was probably not the cause.
What remains uncertainWhether a minority of men need higher levels for symptomatic benefit. No trial supports it.
Bottom lineIn range and no better means look elsewhere, not push higher.
Moderate
Where to go next
Formulations in detail: Which Testosterone Formulation? What to check and when: Monitoring Testosterone Therapy Before you start: Should You Start Testosterone? If children are a possibility: TRT and Fertility