When Is Declining Testosterone a Diagnosis?

Here is a number that should unsettle you: between 1999 and 2016, the average testosterone level in American men aged 15 to 39 dropped from 605 to 451 ng/dL. That is a 25% decline in less than two decades — in men too young for “age-related” decline to explain anything.

If you are a man who has had his testosterone checked and been told it is “low-normal” or “on the lower end,” you have probably wondered: is this just aging? Or is something wrong?

The honest answer is that this question — aging versus disease — may be the wrong question entirely. The evidence from the last five years suggests that much of what we call “age-related” testosterone decline is actually the accumulated weight of modifiable insults: chemicals in your environment, disrupted sleep, metabolic disease, and substances that suppress your HPG axis. True aging effects on the hormonal axis are real, but they are amplified — sometimes dramatically — by forces we can identify and, in many cases, address.

This article traces that evidence.

The Decline Nobody Can Explain Away

The idea that testosterone declines with age is not controversial. After about age 30, total testosterone drops roughly 1–2% per year, and free testosterone declines faster — 2–3% annually — because sex hormone-binding globulin (SHBG) rises with age, binding more of the available hormone.

What is controversial is the discovery that testosterone levels are declining across generations, independent of age. This is the secular decline — a population-wide downward shift that cannot be explained by men getting older.

The landmark data comes from Lokeshwar and colleagues, who analyzed testosterone levels in 4,045 men aged 15–39 from the National Health and Nutrition Examination Survey (NHANES) spanning 1999 to 2016. Their findings were stark:

That second number is the one that matters most. Even among men with a normal BMI (18.5–24.9), testosterone declined by 20% over 17 years. Obesity alone cannot explain this. Something else is happening.

A 2025 systematic review analyzing over one million subjects across multiple countries confirmed what Lokeshwar found in the American data: there is a statistically significant negative linear relationship between testosterone levels and the year of measurement. Men born later have lower testosterone than men born earlier, measured at the same age. The Massachusetts Male Aging Study found the same pattern — approximately 1% per year decline in total testosterone that was independent of aging itself.

A comprehensive narrative review published in January 2026 in the International Journal of Molecular Sciences summarized the current understanding: while aging and genetics contribute, the primary drivers of this secular decline appear to be modifiable — obesity, physical inactivity, dietary patterns, chronic stress, poor sleep, and exposure to endocrine-disrupting chemicals.

The question is no longer whether testosterone is declining beyond what aging explains. The question is what is doing this.

The Environment Inside You

In May 2024, researchers at the University of New Mexico published a study that transformed an abstract concern into something visceral. They tested 23 human testes and 47 canine testes for microplastic contamination.

Every single sample was positive.

Human testicular tissue contained an average of 329 micrograms of microplastic per gram of tissue. The most common polymer was polyethylene — the plastic in grocery bags, food packaging, and bottles. In dogs, polyvinyl chloride (PVC) concentration correlated directly with lower sperm counts.

This is not a distant environmental concern. This is plastic inside the organ that makes testosterone.

The Chemical Assault on Steroidogenesis

Microplastics are just the most photogenic offender. The broader category is endocrine-disrupting chemicals (EDCs) — synthetic compounds that interfere with hormone production, transport, or action. The evidence against three classes is particularly strong:

Phthalates are plasticizers found in food packaging, personal care products, vinyl flooring, and medical tubing. NHANES data across multiple cycles (2011–2016) consistently shows that men in the highest quartile of urinary phthalate metabolites have 12–15% lower serum testosterone than those in the lowest quartile. The mechanisms are multi-level: phthalates disrupt GnRH signaling via G-protein coupled receptors, alter the FSH/LH ratio, and directly impair Leydig cell steroidogenic enzymes.

Per- and polyfluoroalkyl substances (PFAS) — the “forever chemicals” found in nonstick cookware, water-resistant clothing, food containers, and contaminated drinking water — show a more complex picture. Meta-analysis data identifies PFNA (perfluorononanoic acid) as the most consistently associated with lower testosterone. Machine learning screening predicts that 159 PFAS compounds act as estrogen receptor antagonists and 104 as androgen receptor antagonists. However, results vary by compound — PFOS, for instance, shows no consistent correlation in some studies.

Microplastics themselves disrupt testosterone through a hierarchical cascade. The process begins with oxidative stress — reactive oxygen species overwhelming the cell’s antioxidant defenses. This triggers endoplasmic reticulum stress, mitochondrial dysfunction, and apoptosis. The end result is impairment of the StAR protein pathway (the rate-limiting step in steroidogenesis) and, at higher exposures, outright Leydig cell death.

What makes the microplastic data particularly alarming is how it has evolved in the last year alone:

• A 2025 study in Advanced Science showed that PET microplastics — the kind from water bottles — at human-relevant doses caused a 32% decrease in Leydig cells and a 24% reduction in testosterone in mice, while simultaneously downregulating GnRH secretion genes. This means microplastics can hit both the brain and the testicle.
• A 2026 study in Environmental Pollution demonstrated that polylactic acid (PLA) microplastics — marketed as the “biodegradable” alternative — cause dose-dependent testosterone reduction through Leydig cell senescence. The “safe” replacement is not safe.
• A 2026 study in Scientific Reports showed that even at environmental-level doses (0.1–40 μg/kg), polystyrene microplastics reduced testosterone while elevating FSH and LH — a pattern consistent with primary testicular damage, not hypothalamic suppression. The testes are being directly injured.

The caveats are important. Most microplastic studies are in rodents, and extrapolation to humans requires caution. The doses, exposure routes, and polymer types vary across studies. Human epidemiological data linking individual microplastic exposure to measured testosterone levels remains limited. But the convergence of evidence — plastic in every testicle, disrupted steroidogenesis at low doses, multiple independent pathways of harm, and a population-wide testosterone decline that correlates temporally with the plasticization of daily life — is difficult to dismiss.

The Aging Axis Is Real — But Fragile

None of this means aging doesn’t matter. It does. The HPG axis genuinely deteriorates with age, and the mechanisms are well-characterized.

Johannes Veldhuis’s work mapping age-related HPG axis changes remains the most comprehensive model of how the hormonal axis weakens over time. The deterioration is not at one point — it is multi-site:

The feedback sensitivity shift deserves special attention. Winters demonstrated that older hypogonadal men suppress LH and FSH more in response to the same dose of exogenous testosterone or DHT than younger hypogonadal men do. In practical terms: the aging axis is more easily silenced and harder to restart. It creates a “low-T trap” — the system becomes more conservative about producing testosterone precisely when it has less capacity to do so.

This connects to a theme that runs through all of my coverage: the HPG axis is built on pulsatile signaling. GnRH must be released in pulses. LH is secreted in pulses. Even enclomiphene’s “legacy effect” may work by restoring LH pulsatility. Aging disrupts these pulses — making them smaller and more frequent, losing the rhythmic pattern that drives robust testosterone production.

But here is the critical insight: the multi-site deterioration Veldhuis describes makes the aging axis more vulnerable to every other insult. A 30-year-old with robust GnRH output, strong LH pulses, and responsive Leydig cells can absorb moderate environmental exposure, weight gain, or sleep disruption with minimal hormonal consequence. A 55-year-old with a system already operating at 50–60% capacity cannot. The same microplastic burden, the same phthalate exposure, the same sleep fragmentation produces a larger hormonal impact on an aging axis.

Aging and environment are not independent variables. They are synergistic. The right framework is not “aging or environment” but “aging amplified by environment.”

The Great Debate: Is Obesity-Linked Low T Even a Disease?

This brings us to an active and unresolved controversy in endocrinology.

In September 2025, Muir, Wittert, and Handelsman published a provocative paper in the Journal of Clinical Endocrinology & Metabolism arguing that obesity-associated low testosterone is not pathological hypogonadism. They called it “pseudo-hypogonadism of obesity.”

Their argument is mechanistically elegant: obesity lowers SHBG (through hyperinsulinemia, hepatic steatosis, and hypertriglyceridemia). Lower SHBG means less measured total testosterone. But if LH and FSH are normal — and they typically are in obese men — then the pituitary is not signaling distress. The gonadotropins are “highly sensitive tissue androgen sensors,” and if they see enough testosterone, the system is not underfed. The low total T is an artifact of reduced binding protein, not genuine androgen deficiency. Weight loss reverses it.

The response came quickly. Mauvais-Jarvis and Dhindsa, writing in JCEM in October 2025, challenged this framing on three grounds:

1. Reversibility does not mean non-pathological. Obesity-related hypertriglyceridemia is also reversible with weight loss, but no one argues it isn’t pathological.
2. Muir’s own clinical example undermined the argument. They presented a man with a total testosterone of 147 ng/dL and classic hypogonadal symptoms — yet classified him as “eugonadal.”
3. Free testosterone is independently suppressed in obesity. The low T is not purely an SHBG artifact — free testosterone decreases with increasing obesity severity, suggesting genuine hormonal insufficiency.

As I argued in my earlier article on obesity and the HPG axis, the truth is likely a spectrum. In mild to moderate obesity, the SHBG mechanism probably dominates — low measured T with adequate tissue androgen exposure. In severe obesity, genuine central HPG suppression occurs through leptin resistance (leptin normally stimulates kisspeptin neurons), inflammatory cytokines suppressing GnRH, and excess estrogen from adipose aromatase activity. These are well-documented mechanisms that Muir’s framework does not fully address.

This debate matters because it determines who gets treated and how. If obesity-associated low T is pseudo-hypogonadism, the answer is weight loss, not testosterone. If it is genuine secondary hypogonadism, treatment with TRT or SERMs may be warranted. The Cannarella tirzepatide pilot — where a GLP-1 receptor agonist reversed hypogonadism in all 83 obese men treated — suggests a third path: treat the root cause, and the axis restores itself.

When You Remove the Cause, Does Testosterone Come Back?

If declining testosterone is driven by modifiable factors, then removing those factors should restore it. This is the test of reversibility — and the results are more nuanced than you might expect.

The Pozzi bariatric surgery data is the most telling. Among 69 hypogonadal men who underwent bariatric surgery, testosterone rose by an average of 163 ng/dL — a meaningful increase. But 55% remained hypogonadal despite substantial weight loss. Predictors of recovery included higher baseline testosterone (235 vs 184 ng/dL in non-recoverers) and greater BMI reduction.

A smaller study by Petry (n=36) found better results — all hypogonadal patients normalized after surgery — but likely reflected less severe baseline hypogonadism.

The sleep data adds another dimension. A 2025 hypothesis paper in Medical Hypotheses proposed “sleep-disruption-induced hypogonadism” as a unifying framework: conditions common in aging men — benign prostatic hyperplasia (nocturia), obstructive sleep apnea, chronic pain — fragment sleep and suppress the HPG axis. In support, Amodeo and colleagues showed in JCEM (November 2025) that CPAP treatment improved testosterone independently of BMI changes in severely obese men with OSA. The intervention didn’t touch their weight. It fixed their sleep. And their testosterone responded.

The key finding across all of these: not all functional hypogonadism is fully reversible. Removing the cause helps many men substantially but not all. The 55% who remain hypogonadal after bariatric surgery likely have some combination of genuine organic HPG axis pathology that was masked by — not caused by — their obesity, or cumulative damage from years of suppression that does not easily reverse.

Why the Same Number Means Different Things

There is another layer to this puzzle that rarely makes it into the clinical conversation: the same testosterone level can mean very different things in different men.

The androgen receptor contains a polymorphic CAG trinucleotide repeat in exon 1. Longer repeats produce a less sensitive receptor. Shorter repeats produce a more sensitive one. This means that a man with long CAG repeats (say, 26+) may be functionally hypogonadal at a testosterone level that another man with short repeats (say, 18) tolerates without symptoms.

The evidence for this is growing:

• Kim (2018) found that men diagnosed with late-onset hypogonadism had significantly longer CAG repeats than controls (26.1 vs 21.6, p<0.001).
• A 2025 study from Barnsley/Sheffield showed that TRT non-responders had longer CAG repeats than responders (21.8 vs 18.7, p=0.03), with 95.2% sensitivity for distinguishing non-responders.
• Mumdzic (2024) proposed that a testosterone-to-CAG ratio may be a better diagnostic tool than testosterone alone, capturing both hormone level and receptor sensitivity in a single metric.

CAG repeat length also varies by ethnicity — East Asian populations tend to have longer repeats on average — which has implications for how diagnostic thresholds should be applied across different populations. Japan’s 2022 guideline revision acknowledged this, shifting from free T to total T for diagnosis while emphasizing symptom-oriented assessment.

This matters because our “normal range” for testosterone is a population-level statistic. It tells you where you fall relative to other men. It does not tell you whether your tissues are getting enough androgen stimulation. Two men with a total testosterone of 350 ng/dL may have profoundly different clinical realities depending on their receptor sensitivity.

The EMAS Reality Check

If you want to know how slippery the diagnosis of “low testosterone” really is, consider the European Male Ageing Study (EMAS) — the largest prospective study of male aging and hormones ever conducted.

EMAS set strict diagnostic criteria for late-onset hypogonadism: a man needed at least three sexual symptoms (poor morning erections, erectile dysfunction, and low libido) plus a total testosterone below 11 nmol/L (317 ng/dL) plus a free testosterone below 220 pmol/L.

By these criteria, only 2.1% of 2,966 men aged 40–79 qualified. Prevalence ranged from 0.1% in the 40–49 age group to 5.1% in those aged 70–79.

Perhaps more striking: only sexual symptoms — not fatigue, not depression, not cognitive complaints — had a syndromic association with low testosterone. The nonspecific symptoms that many men attribute to “low T” were not statistically associated with decreased testosterone in this population.

This doesn’t mean those symptoms are imaginary. It means they are not specific to testosterone deficiency. A man who is tired, depressed, and gaining weight may have low testosterone — but treating the testosterone may not fix the tiredness, depression, or weight gain if those symptoms have other causes.

The EMAS data is a reminder of how wide the gap can be between how hypogonadism is defined in research versus how it is diagnosed in clinical practice (or by telehealth platforms selling testosterone prescriptions).

The Uncomfortable Conclusion

“Is my testosterone low because I’m aging?”

After reviewing the evidence, I think this question contains a hidden assumption that does not hold up. It assumes “aging” is a clean, intrinsic biological process — something that happens to your cells independent of what you eat, breathe, absorb, and inject. But aging in 2026 means something different than it meant in 1970. It means more years of microplastic accumulation in testicular tissue. More cumulative phthalate exposure from food packaging. More nights of fragmented sleep from untreated apnea. More decades of metabolic stress from processed diets.

The Lokeshwar data proves this: 25-year-olds today have lower testosterone than 25-year-olds seventeen years ago. These are not aging effects. These are generational effects — something about the modern environment is suppressing the HPG axis earlier and harder.

At the same time, true aging effects on the HPG axis are real and well-characterized. GnRH output declines. LH pulsatility degrades. Leydig cells become less responsive. Feedback sensitivity increases, making the axis more easily suppressed. These are genuine biological changes that no lifestyle intervention can fully prevent.

The synthesis is this: aging creates the vulnerability. Environment exploits it. The secular testosterone decline is environmental. The age-related decline is biological. They compound each other. And the question of where one ends and the other begins may be unanswerable for any individual man — especially when his androgen receptor sensitivity (CAG repeats) determines how much testosterone his tissues actually need.

What does this mean practically? A few things:

• A testosterone number without context is almost meaningless. Age, BMI, sleep quality, medication use, SHBG level, and ideally AR CAG repeat length all modify what that number means for a given man.
• Root-cause investigation before treatment. Weight loss, CPAP, opioid reduction, and sleep optimization should be tried before defaulting to TRT in functional hypogonadism. Many men will recover without exogenous hormones.
• But not all will. Pozzi’s 55% persistent hypogonadism after bariatric surgery shows that some men have genuine axis pathology that lifestyle changes cannot fix. For them, SERMs, hCG, or TRT may be appropriate — and withholding treatment in the name of “lifestyle first” does them a disservice.
• The environmental exposures are not going away. Microplastics are in the water, the food, the air, and already in the testicles. Reducing exposure where possible (glass over plastic, filtering water, minimizing processed food packaging contact) is rational but will not eliminate what is already present.

The line between “normal aging” and “pathological decline” is not a line at all. It is a gradient shaped by genetics, environment, and accumulated insults. And for many men, the answer to “is this just aging?” is: it is aging — but not just aging.

Key references: Lokeshwar et al., Eur Urol Focus 2020 (NHANES secular decline); Veldhuis et al., multiple publications (HPG axis aging model); Muir, Wittert & Handelsman, JCEM Sept 2025 (pseudo-hypogonadism); Mauvais-Jarvis & Dhindsa, JCEM Oct 2025 (critique); UNM microplastics study, Toxicol Sci May 2024; Advanced Science 2025 (PET-MPs); Env Pollution 2026 (PLA-MPs); Sci Reports 2026 (low-dose PS-MPs); Pozzi et al., Andrology 2025 (bariatric reversibility); Amodeo et al., JCEM Nov 2025 (CPAP); Wu et al., NEJM (EMAS criteria); Kim 2018 / Mumdzic 2024 / Barnsley 2025 (AR CAG repeats); IJMS Jan 2026 (secular decline review); J Endocrinological Investigation 2025 (systematic review, 1M+ subjects).