Clinical decision support · Multisystem Medicine
Guideline algorithms answer a population question: does this man fall below the threshold? The clinic asks a different one: is this man androgen-insufficient relative to his own physiology, and is anything about it reversible? This pathway keeps the population threshold as the entry gate, then adds the three things standard figures leave out — reversible drivers, the individual's own reference, and a pre-committed endpoint.
Does he belong to the hypogonadal population? Answered by threshold, repeated on two mornings, with an accurate assay. Necessary, and the only part with outcome data behind it.
Is he androgen-insufficient for himself — and why? Answered by his own historical values, his axis, his end-organ evidence, and what happens when a driver is corrected.
Case detection, not population screening. Symptoms alone never make the diagnosis; a level alone never makes it either.
Most “low testosterone” in practice is a pre-analytical problem, not an endocrine one.
If he is taking, or recently stopped, testosterone, AAS, SARMs, prohormones, or hCG obtained anywhere — this is drug exposure, not a diagnosis. Suppressed LH with a low or variable T is the expected pharmacology. Recovery of the axis takes months and is incomplete in a meaningful minority. Do not assign a diagnosis of hypogonadism until ≥3 months (preferably 6–12) after full withdrawal, and document the exposure history in the chart. Direct-to-consumer platforms are a common and underdisclosed source.
Free T adds information only when SHBG is, or is likely to be, abnormal. When it is, it changes the diagnosis in both directions.
Free testosterone is the single most error-prone number in this workup because three units circulate simultaneously: pg/mL, ng/dL, and nmol/L. 1 ng/dL = 10 pg/mL. A “free T of 6” is profoundly low if it is pg/mL and entirely normal if it is ng/dL — and both appear on real reports. Published algorithms quoting “<52” mean pg/mL. Before acting on a free T, confirm the unit and the method on the report itself, not on the number.
Biochemical androgen deficiency established. Proceed to localize the axis.
High-SHBG state masking true deficiency — classically the older, lean, or hyperthyroid man. Treat as hypogonadal and localize.
Pseudohypogonadism. Low SHBG from adiposity, insulin resistance, or nephrotic syndrome. The bioavailable fraction is intact. Testosterone is not the intervention — go to the metabolic drivers in Stage 4.
Not androgen-deficient by population criteria. Standard figures stop here. Two things still deserve attention: the drivers of his actual symptoms (Stage 4), and — only if a documented prior baseline was far higher and symptoms are specific — the individualization sequence (Stage 5).
LH and FSH, drawn with the confirmatory testosterone. Prolactin and free T4 whenever gonadotropins are not elevated.
An LH in the upper half of the reference range alongside a low-normal testosterone is the hypothalamic-pituitary unit reporting that it is working harder than usual to hold that level. It is not a diagnosis, and no guideline scores it — but it is real information about this man's set point, and it belongs in the assembled picture at Stage 5. The mirror image is equally informative: a mid-range LH with a testosterone of 250 in an obese man says the axis is not trying, which points to the drivers rather than the testis.
The safeguard of the whole pathway. Every one of these can produce a testosterone in the 200s with an unelevated LH, and every one of them responds to something other than testosterone. Work these before the diagnosis is written down, because once “hypogonadism” is in the problem list it rarely comes out.
After ≥3–6 months of driver-directed therapy delivered to target — not merely prescribed — repeat morning total T, free T where indicated, and LH. Functional hypogonadism that corrects is not hypogonadism; it was a symptom of something else. Only what persists after this gate earns the diagnosis. Document the drivers addressed, the degree of correction achieved, and the post-correction values.
The most common man in a concierge practice has a partially reversible driver and a genuinely low set point — 40 lb of visceral fat, moderate OSA, and a testis that was never going to make 700. Correcting the driver moves him from 240 to 340 and he still feels the way he felt. That result is not a failure of the reversibility stage; it is the reversibility stage doing its job, because now the residual deficit is measured rather than assumed, and the therapeutic target is anchored to a corrected baseline rather than a confounded one.
Four independent channels of evidence. None is diagnostic alone; the value is in whether they agree.
A documented prior baseline is a legitimate target once a diagnosis has been established. It is not, by itself, a basis for making the diagnosis. Relative decline within the reference range remains unvalidated as a diagnostic criterion, and treating it as one is precisely how this framework would be misused. Everything in Stage 5 exists to keep that distinction visible.
The androgen receptor CAG repeat length genuinely explains variance in individual set point — shorter repeats confer greater receptor sensitivity, and men with longer repeats sit at higher free testosterone for the same phenotype. It is mechanistically satisfying and, on current evidence, adds little predictive value over serum testosterone and does not change management. In-vitro androgen bioassays and tissue-level androgen action remain research tools. Do not order these, and be skeptical of platforms that do.
Set each channel, then name the pattern out loud in the note. “Discordant, with these four findings” is a defensible clinical position. “Low T” is not.
Set each evidence channel to what you actually have. The pattern updates below.
Pattern C — discordant
The channels do not agree. This is the man at 400, and this is where judgment is actually required rather than a threshold. Proceed to a designed n-of-1 trial, or to a period of continued driver optimization with a scheduled re-read.
Set the channels above to update.Persistently low levels after Stage 4, specific symptoms, corroborating end-organ evidence, an axis that fits. Treat, with the standard fertility, hematocrit, and prostate conversations first.
Say it plainly: he is not androgen-deficient. His symptoms are real and belong to something else — sleep, mood, iron, thyroid, anemia, deconditioning, alcohol, relationship or life context. Concluding sufficient is a successful visit, not a failed one, and it is the discipline that keeps this framework from becoming an optimization protocol.
Gray-zone or low-normal level, real symptoms, some surrogates supportive and others not, or a documented prior baseline far above the current value. A therapeutic trial is legitimate here — provided it is designed rather than drifted into.
The single most useful discipline in this pathway is naming the pattern in the chart with its evidence attached: “Discordant: TT 400 by LC-MS/MS ×2 after 6 months of OSA therapy and 12% weight loss, SHBG normal, cFT low-normal, LH 7.8, documented TT 800 at age 44, hemoglobin fallen 15.6 → 14.1 over six years, AMS 42, no androgen exposure.” That sentence survives a chart review, a second opinion, and your own re-reading in two years. “Low T, started TRT” does none of those things.
Six things written down before the first dose. What separates individualized medicine from indefinite hormonal drift is that the endpoint was specified while you could still be wrong about it.
Most harm in this field lives in the execution, not the decision.
| Formulation | When to draw | What you are asking |
|---|---|---|
| Cypionate / enanthate, weekly SC or IM | Mid-interval, or trough just before the next dose | Peak–trough excursion; a mid-interval mid-normal level with symptom relief is the goal |
| Cypionate / enanthate, every 2 weeks | Trough, and consider a peak at day 2–3 | Whether the swing itself is the symptom generator — often the fix is more frequent, smaller dosing |
| Transdermal gel | 2–8 h after application, steady state | Absorption adequacy; also ask about transference precautions |
| Testosterone undecanoate IM | Trough before the next injection | Adequacy at the end of the interval; observe for POME/anaphylaxis per REMS |
| Oral undecanoate | Per product labeling, after the morning dose | Levels are highly meal-dependent; check blood pressure — a labeled warning |
| Subcutaneous pellets | Trough near end of interval | Duration adequacy; note the inability to dose-adjust once implanted |
Estradiol is a male hormone. It is required for bone maintenance, libido, and normal fat distribution in men, and much of what is attributed to testosterone is estradiol-mediated. Aromatase inhibitors are not standard of care in testosterone replacement and should not be prescribed reflexively for an estradiol number in the absence of clinically significant, persistent gynecomastia. Suppressing estradiol to “optimize” a ratio trades a laboratory aesthetic for bone and sexual function. If estradiol is high, the usual answer is that the testosterone dose or the peak is too high.
An FDA advisory panel in December 2025 discussed broadening testosterone eligibility, revising prostate-related warnings, and the drug's controlled-substance status; the Endocrine Society issued a statement in July 2026 reiterating that diagnosis requires both compatible symptoms and consistently low, properly measured concentrations, and that symptoms alone are not sufficient. Expect the labeled indications and the prescribing environment to shift faster than the diagnostic evidence does. Verify current status before relying on any of it.
A separate adjudication with a stricter standard than clinical care. Everything upstream in this pathway is exactly what a TUE committee is trying to reconstruct — the organic/functional determination at Stage 4 is the determination they will make, and they will make it on your documentation alone.
Any man competing under a signatory of the World Anti-Doping Code — national and international federations, Olympic and Paralympic pathways, and masters and amateur athletes in sanctioned events, who are frequently unaware they are covered. Collegiate athletes fall under separate NCAA medical-exception rules; tactical and military populations have their own policies. Ask at the first visit. A man who discloses competition after the first injection has a materially worse problem than one who discloses before it.
For a tested athlete, this pathway's Stage 4 is not a clinical nicety — it is the whole case. A committee that cannot see an unequivocal organic lesion will deny the application no matter how carefully the biochemistry was performed, and the athlete then faces a choice between an untreated diagnosis and an anti-doping violation. Say this at the first visit, before any expectation has formed.
This is the one place in the pathway where the individual reference has no standing. A documented personal baseline, a compensating LH, a coherent symptom pattern at 400 ng/dL — all of it is good medicine and none of it is a TUE. The committee is applying a categorical rule for a reason: within-range testosterone still confers advantage, so a permissive individualized standard would be unworkable in sport. Recognizing that boundary early is what keeps a well-intentioned trial of therapy from ending an athlete's career.
Built-in safeguards — remove any of these and this becomes the thing it was written against
Reversibility before diagnosis. Functional hypogonadism that corrects was never hypogonadism.
The discipline to conclude sufficient. Not treating is a legitimate and frequent output of this pathway.
A stopping rule written and spoken before the first dose, and honored at the pre-specified date.
Competition status established at the first visit, before any expectation of therapy has formed.