The science
Built on evidence. Allergic to hype.
Every claim in TestoTrack — from why trends beat single tests to which lifestyle levers actually move your levels — is grounded in peer-reviewed research. Here's the foundation, with the citations attached and the limits stated.
Last reviewed August 2026
The evidence
What does the research actually say?
E1
Is testosterone really declining, or is that a marketing line?
It is real, and it is measured. The Massachusetts Male Aging Study tracked the same population across three waves and found serum testosterone falling roughly 1.2% per year between 1987 and 2004 — a drop that persisted after controlling for age, meaning it is not simply men getting older. What it does not do is tell you about you: a population trend is not a personal diagnosis.
E2
Why is one blood test not enough?
Because testosterone moves. It follows a daily rhythm, and repeat draws from the same man days apart vary enough that a single value can land on either side of a clinical threshold by chance. This is why the Endocrine Society does not accept one number: diagnosis requires a repeated fasting morning measurement, in a man who also has symptoms. One result is a data point. A trend is information.
E3
Can lifestyle changes actually move the number?
For some men, measurably. In a controlled study, one week of sleep restricted to five hours a night lowered daytime testosterone in healthy young men by 10–15%. And a meta-analysis of weight-loss studies found that losing weight raises testosterone in men whose levels were suppressed by obesity. These are inputs you can change and outputs you can measure — which is the entire premise of tracking.
E4
How common is low testosterone, really?
More common than the diagnosis rate suggests. In the HIM study, 38.7% of men aged 45 and over presenting to US primary-care offices met the biochemical criteria for hypogonadism — and the odds were meaningfully higher in men with obesity, diabetes, hypertension, or hyperlipidaemia. Prevalence figures vary with the threshold and assay used, which is another argument for watching your own line rather than a population average.
E5
Why does TestoTrack keep talking about 90 days?
Because that is roughly how long the body takes to answer you. The original 1963 measurement put one cycle of the seminiferous epithelium at 16 days and the whole of spermatogenesis at about 64 days; the figure usually quoted since is 74 days, plus a further stretch for transit and maturation before sperm appear in an ejaculate — which is where the familiar ~90-day window comes from. Worth saying plainly: a 2008 review holds that the accepted duration may be wrong by around six days and that the underlying 1960s data is neither robust nor precise. So treat 90 days as a sensible measurement interval, not a biological constant. The same logic applies to testosterone: give a change enough time to show up before you retest.
Reference ranges differ between laboratories and assays, and study populations are not you. Everything above describes what has been observed across groups of men — useful for framing a question, never sufficient to answer it about one person.
Where we won't sell you something
Does testosterone therapy affect fertility?
Yes — and in the direction most men do not expect. Exogenous testosterone suppresses the hypothalamic-pituitary-gonadal axis, which lowers intratesticular testosterone and can drive sperm counts toward zero; it has been studied as a male contraceptive for exactly this reason. Suppression is usually reversible, with most men recovering within several months of stopping and a median around 110 days, but recovery can take a year or more, and is less predictable after long-term use. If you are trying to conceive, this is a conversation to have with a physician before you start, not after.
We have no financial interest in whether you start testosterone therapy. TestoTrack does not sell hormones, supplements, or clinic referrals, and the coach will not recommend a treatment. That is deliberate: an app that profits from a decision should not be the thing advising you on it.
How the AI coach stays honest
How does the coach avoid making things up?
General-purpose chatbots improvise. TestoTrack's coach uses retrieval-augmented generation: before answering, it searches a curated library of endocrinology research and clinical guidelines, then combines what it finds with your actual data. If the evidence isn't there, the coach says so — and suggests talking to your doctor.
- R1
Your question + your data
Age, latest results, and trend context are assembled into the query.
- R2
Evidence retrieval
FAISS vector search pulls the most relevant passages from vetted research.
- R3
Grounded answer
Claude composes a plain-language answer constrained to the retrieved evidence.
- R4
Safety layer
Medical-boundary checks ensure the coach informs — it never diagnoses or prescribes.
Medical boundary
TestoTrack is a tracking and education tool, not a medical device, and the AI coach is not a physician. It will never diagnose a condition or recommend medication. For clinical decisions — including any form of hormone therapy — always work with a qualified healthcare provider.