Should you start testosterone? Ten questions first.
A structured way to work through the decision, and ten questions worth taking into the appointment with you.
Ten steps. Work through them in order. If you cannot answer one, that is the next thing to establish — not a reason to move on.
Research presented at ENDO 2026, reviewing 200 men started on testosterone in primary care, found only 12% had a guideline-concordant diagnostic workup. If you complete this page honestly, you will know more about your own case than most men holding a prescription.
Step 1 — Do your symptoms actually suggest testosterone deficiency?
Not all symptoms carry equal weight, and this is not widely understood.
The European Male Ageing Study screened 32 candidate symptoms in 3,369 men and found only three independently associated with testosterone, all sexual: poor morning erections, low sexual desire, and erectile dysfunction. Physical and psychological symptoms — fatigue, low energy, low mood, poor concentration — tracked age and accumulating comorbidity rather than testosterone level.
Ask yourself: are your symptoms predominantly sexual, or predominantly fatigue and mood?
If it’s the latter, testosterone is a less likely explanation than sleep, mood, medication, weight or an undiagnosed medical condition. That does not mean nothing is wrong. It means the answer is probably elsewhere.
Move on if: you have sexual symptoms, or you have a specific reason to suspect an organic cause.
Step 2 — Was your testosterone measured correctly?
Four conditions, all required:
- Early morning — between roughly 7 and 10am. Afternoon values run 20 to 25% lower in younger men.
- Fasting — eating drops testosterone by an average of 100 to 123 ng/dL within an hour, and in one study 56% of men transiently fell below 300 ng/dL after a meal.
- Not during illness, and not after a run of bad sleep.
- On a standardized assay, ideally CDC-certified. Direct analog immunoassays for free testosterone should not be used at all — every major guideline says so.
If your test was an afternoon draw, or you’d eaten, or you were unwell — you don’t have a result yet. See How Testosterone Should Actually Be Tested.
Step 3 — Was the low result confirmed?
Around 30% of men with an initially low testosterone have a normal value on repeat measurement. That figure is from the Endocrine Society’s own guideline, and it is why every major society requires two separate samples.
Three men in ten.
If you were diagnosed on one blood test, you were not adequately evaluated. This is the single most commonly skipped step, and it is not a technicality — it is the difference between a diagnosis and a coin flip weighted 70/30.
Step 4 — Is free testosterone relevant in your case?
For most men, total testosterone is adequate. It stops being adequate when SHBG is abnormal.
SHBG is likely low — depressing your total while free testosterone may be fine — if you have obesity, insulin resistance, type 2 diabetes or hypothyroidism.
SHBG is likely high — propping up your total while free testosterone may be genuinely low — if you are older, thin, hyperthyroid, on anticonvulsants, or have liver disease.
If either applies, SHBG and free testosterone belong in your workup — calculated from a reliable equation or measured by equilibrium dialysis. Not by direct immunoassay.
Note that societies disagree about when free testosterone is warranted. See Why Experts Don’t Always Agree.
Step 5 — Why is your testosterone low?
This is the step that separates a diagnosis from a number, and it costs one blood test.
LH and FSH tell you where the problem is. High LH and FSH with low testosterone points to the testes. Low or inappropriately normal LH and FSH points upstream — to the pituitary, the hypothalamus, or something suppressing them.
That distinction determines whether you need pituitary imaging, whether the cause is reversible, whether fertility can be preserved, and whether testosterone is even the right drug.
If nobody measured your LH and FSH, nobody knows why your testosterone is low. See Why Is Testosterone Low?.
Step 6 — Do you want children? Ever?
Answer this before the first dose, not after.
Testosterone therapy suppresses sperm production in most men. In the World Health Organization’s contraceptive trials, 65% of men on weekly testosterone became azoospermic, at a mean of around 120 days. Recovery usually happens — 90% by twelve months in contraceptive-trial participants — but it takes months to over a year, and it is slower after longer treatment.
Your blood testosterone will look excellent throughout. It tells you nothing about your fertility, because intratesticular testosterone runs 80 to 100 times higher than blood levels and exogenous testosterone drops it by around 94%.
If children are possible — now or in ten years — options exist: bank sperm first, use hCG alongside, or stimulate your own axis with a SERM instead of replacing testosterone. All of them require the conversation to happen first. See TRT and Fertility.
Step 7 — Are there contraindications or risks specific to you?
Testosterone should not be started, or started only with specialist input, if you have:
- Breast cancer or prostate cancer
- A palpable prostate nodule, or a PSA above 4 ng/mL without urological evaluation
- An elevated hematocrit at baseline
- Untreated severe obstructive sleep apnea
- Severe lower urinary tract symptoms
- Uncontrolled or severe heart failure
- A myocardial infarction or stroke within the past six months
- Thrombophilia
- Plans to conceive in the near term
Notably, in the ENDO 2026 cohort, 55% of men started on testosterone had obstructive sleep apnea and 4% had a prior prostate cancer diagnosis.
Step 8 — Have reversible causes been addressed?
The Endocrine Society’s 2018 guideline states that many men with secondary hypogonadism have potentially reversible causes “that may be managed without need for testosterone treatment.”
Has anyone looked at your weight — where dieting raises testosterone by around 83 ng/dL on average and bariatric surgery by around 251? Your opioids, where prevalence of suppression runs above 50% in most studies? Your glucocorticoids? Your prolactin, where treating an elevation restores testosterone in 50 to 65% of cases without any replacement? Your thyroid? Your sleep? Your alcohol? Whether you are eating enough for your training load?
One honest caveat, because this site would rather be accurate than tidy: treating sleep apnea does not appear to raise testosterone. A meta-analysis of 232 men found CPAP had no significant effect. Treat it anyway — it is dangerous and it causes the same symptoms — but not on the promise that your number will move.
See When Testosterone Isn’t the Answer.
Step 9 — Would testosterone plausibly improve your symptoms?
Match the treatment to the complaint.
Good evidence testosterone helps: sexual desire, sexual activity, correction of anemia, bone mineral density, lean and fat mass.
Poor evidence: erectile function in men with vascular disease — TRAVERSE found no significant improvement; glycemic control — TRAVERSE found no benefit; fracture prevention — fractures were higher on testosterone (HR 1.43, 95% CI 1.04–1.97); cardiovascular protection; longevity.
If your main complaint is erectile dysfunction and you are 60 with diabetes and hypertension, testosterone is unlikely to fix it. Expect desire to improve and plan for the erections separately.
Step 10 — How will you know whether it’s working?
Agree this before starting, in writing if you can.
Which symptoms are you treating? Name them. By when? Sexual desire responds in about three weeks and plateaus around six. Mood in three to six weeks with maximum at eighteen to thirty. Body composition at three to four months. Bone density beyond a year. What gets monitored? Testosterone, hematocrit, blood pressure, PSA where appropriate, symptoms — with timing depending on formulation. See Monitoring Testosterone Therapy. What’s the stopping rule? If your testosterone reaches target and your symptoms haven’t moved, that is information. It means testosterone was not the cause. The answer then is to look elsewhere — not to raise the dose.
For scale: in TRAVERSE, 61% of men discontinued. In real-world data on 15,435 men using testosterone gel, only 15.4% were still on treatment at twelve months. Roughly five in six men stop within a year. Deciding in advance what success looks like is how you avoid becoming a statistic in either direction.
The 10 questions to ask before you start
Take this to your appointment.
- Do I actually meet the diagnostic criteria?
- Was my testosterone measured in the morning, fasting, on a standardized assay?
- Was the low result confirmed on a second sample?
- What were my LH and FSH, and what do they tell us about the cause?
- Is there a reversible cause we haven’t addressed?
- What does this mean for my fertility, and what are my options?
- What alternatives are there, including doing nothing for now?
- What are my individual risks, given my history?
- Which formulation makes sense for me, and why that one?
- What are we measuring to decide whether this is working, and when do we stop if it isn’t?
If a clinician cannot answer questions 3, 4 and 6, they have not evaluated you. They have measured you.
TRT should be the result of a diagnosis — not the diagnosis itself.
A testosterone level is a piece of information. It is not a diagnosis.
A prescription is not a treatment plan.
And testosterone replacement is not the same thing as testosterone optimization, anti-aging medicine, or performance enhancement.
Don’t treat the number. Treat the man.
Questions patients ask
My level is low, so I should start treatment.
A low level is a reason to investigate, not a reason to prescribe.
What the evidence showsAround 30% of initially low results are normal on repeat, and the Endocrine Society's July 2026 statement requires at least two early-morning fasting measurements in a man who has symptoms consistent with deficiency. LH and FSH then locate the problem, and a substantial share of causes are reversible.
What remains uncertainWhich men with borderline results benefit from treatment at all.
Bottom lineConfirm it, then find out why. The prescription is the last step, not the first.
Strong
I should wait until my testosterone is very low before treating.
No — but the expected benefit does shrink as the number rises.
What the evidence showsThe erectile-function benefit concentrates below 8 nmol/L, and the Testosterone Trials found their clearest effects in men who were genuinely deficient. TRAVERSE, in a broader population, found no erectile benefit at all.
What remains uncertainWhether 8 nmol/L is a biological threshold or a convenient analytic cut point.
Bottom lineSeverity predicts benefit. It does not set a bar you must fall below.
Moderate
If I start, I am on it for life.
Usually in practice, not necessarily in principle.
What the evidence showsExogenous testosterone suppresses LH and FSH, and the axis takes months to recover once it stops. But in real-world claims data covering 15,435 men on gel, only 34.7% were still taking it at six months and 15.4% at twelve — five in six stopped within a year.
What remains uncertainHow much of that discontinuation reflects lack of benefit rather than cost or inconvenience.
Bottom lineStopping is common and possible. Plan for a trial with an endpoint, not an open-ended commitment.
Moderate
I can decide about fertility later.
This is the one decision that is genuinely hard to reverse.
What the evidence showsTestosterone suppresses spermatogenesis, in most men to severe oligospermia or azoospermia. Recovery is usual but not guaranteed, and can take a year or more. Evaluating the original cause afterwards means stopping and waiting for the axis to recover.
What remains uncertainWhich men fail to recover, and why.
Bottom lineSettle this before the first dose. See TRT and Fertility.
Strong
A symptom questionnaire told me I probably have low testosterone.
Those instruments are built to be sensitive, which makes them poor at ruling anything in.
What the evidence showsOne validated screening questionnaire flagged roughly half the men who did not have low testosterone. Fatigue, low mood and poor concentration are the complaints men bring in and, on European Male Ageing Study data, are not the ones that track testosterone — the three that do are sexual.
What remains uncertainNothing about their specificity; it is well documented.
Bottom lineA quiz is a marketing funnel, not a diagnostic step.
Strong
My doctor said my level is normal, so there's nothing to discuss.
Ask which threshold, and ask what else was measured.
What the evidence showsPublished thresholds span roughly 231 to 346 ng/dL depending on the society. A man at 320 is normal under the AUA and the Endocrine Society and potentially hypogonadal under the EAU. Free testosterone also matters when SHBG is abnormal, and a total result alone can be misleading in obesity, diabetes, thyroid disease or on certain drugs.
What remains uncertainWhether any threshold identifies the men who benefit; none has been validated against treatment response.
Bottom line"Normal" is a decision someone made, not a fact about you. See Why Experts Disagree.
Strong
Trying testosterone for a few months costs me nothing.
It costs you the ability to answer the original question.
What the evidence showsOnce treatment starts, LH and FSH are suppressed, so the workup that would have identified the cause no longer works. Fertility is suppressed. Hematocrit rises. And a trial without a pre-agreed endpoint tends to continue by default.
What remains uncertainHow often a properly conducted therapeutic trial changes the eventual decision.
Bottom lineA trial is reasonable — after the workup, with a defined endpoint and a defined stopping rule.
Moderate
If treatment doesn't help, we can just raise the dose.
If a physiologic dose has not helped, the diagnosis is the thing to revisit.
What the evidence showsThe Endocrine Society directs treatment to the mid-normal range; the AUA names 450 to 600 ng/dL, described as the middle tertile of the reference range and labelled expert opinion. There is no trial evidence that pushing above the reference range improves symptoms, and supraphysiologic exposure raises hematocrit and suppresses fertility further.
What remains uncertainWhether any subgroup benefits from a higher target; it has not been tested.
Bottom lineNon-response is information. It usually means testosterone was not the problem.
Moderate
Where to go next
Testing done properly: How Testosterone Should Actually Be Tested Establishing the cause: Why Is Testosterone Low? Fertility: TRT and Fertility Risks: Testosterone, the Prostate and the Heart Alternatives: When Testosterone Isn’t the Answer
Ready for testosterone care built on a diagnosis?
The first step is a full endocrine evaluation, not a prescription. We see patients across San Diego County and welcome referrals from other physicians.
