TRT Statistics 2026: Use, Safety, Benefits, and Source Data

Most TRT statistics get quoted wrong because different studies count different things. Here is the strongest current number we found: an Epic Research study of 109 million men found that 0.93% of male patients who had a healthcare encounter in 2025 had an active testosterone prescription. That is 93 per 10,000, or about 1 in 108, and it is 45.3% above the 2021 rate of 0.64%.

The number to quote: In Epic Cosmos, 0.93% of male patients with a healthcare encounter had an active testosterone prescription in 2025—93 per 10,000, up 45.3% from 2021.

One limit, right up front. This is not “0.93% of all American men are on TRT.” It counts male patients who had an encounter at a health system in Epic Cosmos and had testosterone listed as active in the chart. A chart entry does not prove the prescription was filled or used.

The next number is harder to brush past. The American Urological Association guideline says “up to a third” of men placed on testosterone may not meet the criteria for testosterone deficiency. In a 2026 review of 200 men at Michigan Medicine, only 12% had the full guideline-based workup before their first prescription.

By: TRT Provider Guide · Published: September 2, 2026 · Last verified: September 2, 2026
About this page: TRT Provider Guide is an independent research and reference resource. This page has no advertising, affiliate links, or provider recommendations.

Data files: Full TRT statistics dataset · Prescribing timeline · Common-claims source ledger

---

TRT statistics at a glance

Each row answers a different question. The denominator is the part that keeps the number honest.

Sources: Epic Research; Selinger and Thallapureddy, PLOS ONE; Sinha and Papaleontiou, ENDO 2026; Dubin et al., JAMA Internal Medicine; Mulligan et al., International Journal of Clinical Practice; Araujo et al., JCEM; and the TRAVERSE trial reports. Compiled and recalculated by TRT Provider Guide. Last verified September 2, 2026.

---

What the TRT statistics show—and what they do not

Answer capsule: Testosterone prescribing fell for much of the 2010s, then rose to the highest full-year rate in the Epic series by 2025. At the same time, several studies found weak testing before treatment. The data does not give a complete U.S. headcount, and it cannot tell us whether each prescription was right for that patient.

The 2025 Epic rate of 0.93% is 12.0% above the same series’ 2013 rate of 0.83%. Use did not just return to its old level. It passed it.

But the timing needs to stay honest. Prescribing had already started to fall after 2013. The FDA added stronger label language in March 2015. Commercial-insurance data then showed a further fall through 2016. Those facts line up in time, but the studies do not prove that the FDA action caused the whole decline.

Three plain takeaways:

• Use is rising, and it peaks in middle age. Men age 50 to 64 had the highest rate in 2025.
• “How common is low testosterone?” has no one answer. Change the population or definition and the result moves from 2.1% to 38.7%.
• Benefits and risks are outcome-specific. The same trial found gains in sexual activity and anemia, no significant gain in erectile function, no excess in its main cardiac outcome, and more fractures.

And one thing nobody can prove from the prescribing data: why use rose. Epic Research says its study describes the trend but does not explain it.

---

How many men use TRT in the United States?

Answer capsule: No source in this review gives a complete national headcount. The best current rate is 0.93% of male Epic Cosmos patients with a healthcare encounter in 2025. A separate 24-state study counted 1,656,641 unique people filling testosterone in 2022. Those figures measure different things and cannot be added.

TRT means testosterone replacement therapy. Hypogonadism is a medical syndrome: symptoms or signs occur with consistently low testosterone. A full workup looks for the cause, which may involve the testes, pituitary, hypothalamus, illness, obesity, or medicines.

A national count is hard for simple reasons. Insurance data misses cash-pay care. State prescription databases miss testosterone given inside a clinic. Health records do not prove a prescription was filled. No one dataset closes all of those gaps.

Three ways current sources count testosterone use

Sources: Epic Research, 2026; Selinger and Thallapureddy, PLOS ONE 2024; Sellke et al., International Journal of Impotence Research 2024. Last verified September 2, 2026.

A calculation we did not publish. It is easy to multiply 0.93% by the U.S. adult male population and create a national estimate. We did the math and threw it away. The 0.93% denominator is male patients with a healthcare encounter in Epic Cosmos, not all adult men. Stretching it across the whole country would create a clean-looking number that the source cannot support.

A prescription drug monitoring program, or PDMP, is a state database for controlled-substance prescriptions after they are dispensed. That makes it closer to a real fill count than a chart entry, but the 2024 study still had only 24 states in its final analytic sample and did not contain diagnosis data.

---

Are TRT statistics going up or down?

Answer capsule: Up, after a long fall. In Epic Cosmos, the rate moved from 0.83% in 2013 to 0.64% in 2021, then rose to 0.93% in 2025. That is a 45.3% relative increase from the low point and a 0.29 percentage-point absolute increase.

The current Epic health-record series

Source: Epic Research, July 29, 2026. Per-10,000 rates and relative changes calculated by TRT Provider Guide. Last verified September 2, 2026.

The headline math is:

(0.93 − 0.64) ÷ 0.64 × 100 = 45.3%

That is the relative increase. The absolute increase is 0.29 percentage points. Both are true. Print both when space allows.

Epic also says the number of contributing organizations grew over time. Early years came from a thinner health-system footprint. The series is useful, but it is not a fixed national panel.

Arithmetic audit 1: the PLOS fill-count increase

The PLOS ONE PDMP study prints a 27% increase from 1,216,982 people in 2018 to 1,656,641 in 2022.

Using the standard start-value formula gives:

(1,656,641 − 1,216,982) ÷ 1,216,982 × 100 = 36.1%

Using the ending value as the denominator gives:

439,659 ÷ 1,656,641 × 100 = 26.5%

That rounds to the paper’s 27%. The paper does not state the formula used for that sentence. The safest way to report it is the raw change: 1,216,982 to 1,656,641, up 439,659 people. When a percentage is needed, label 27% as the paper’s printed value and 36.1% as the standard recalculation.

The same paper prints a rise in population share from 0.76% to 0.98% as about 30%. That one matches the start-value formula:

(0.98 − 0.76) ÷ 0.76 × 100 = 28.9%

The two sentences in the same paper therefore appear to use different denominators.

Arithmetic audit 2: the Medicare claim increase

The Medicare study prints 542,040 testosterone claims in 2016 and 751,976 in 2019. It also prints a 54.6% increase.

Those raw counts give:

(751,976 − 542,040) ÷ 542,040 × 100 = 38.7%

A 54.6% increase would require a different starting value. We found no second denominator in the paper that reconciles the claim totals with 54.6%. This page and the data files therefore use 38.7% calculated from the published counts and preserve 54.6% only as the percentage printed by the paper.

The long view: separate sources, separate populations

Sources: Baillargeon et al., JAMA Internal Medicine 2013; Baillargeon et al., JAMA 2018; Sellke et al., International Journal of Impotence Research; Selinger and Thallapureddy, PLOS ONE; Epic Research. These series must not be joined into one continuous national line.

Who writes the prescriptions?

In the 2026 Michigan Medicine review, primary-care physicians wrote 45% of the first prescriptions, urologists 35.5%, endocrinologists 18%, and other specialists 1.5%.

Medicare data points the same way. Family medicine and internal medicine produced the most claims among physician specialties. From 2016 to 2019, the number of nurse practitioners prescribing testosterone rose 118%, and the number of physician assistants rose 78%. Advanced-practice providers’ share of claims moved from 7.5% to 11.6%.

That does not make TRT “a primary-care drug” in every setting. It does show that specialty clinics are not the whole story.

---

Which age group uses TRT the most?

Answer capsule: Men age 50 to 64. In 2025, 1.57% of male Epic patients in that age group had an active testosterone prescription. That is 157 per 10,000, about 1 in 64, and 11.2 times the rate for men age 18 to 29.

Source: Epic Research, 2026, among male patients with a healthcare encounter. Per-10,000, ratio, and “1 in” conversions by TRT Provider Guide.

The highest use and the fastest growth are not the same thing. In the 24-state PDMP study, the population-adjusted share filling testosterone from 2018 to 2022 rose 120% for age 24 and younger, 86% for age 25–34, 45% for age 35–44, 35% for age 45–54, 17% for age 55–64, and 12% for age 65 and older.

A small base can grow fast and remain small. Younger groups grew faster. Men age 50 to 64 still had the highest active-prescription rate in the Epic series.

The youngest PDMP group also cannot be used as a trend for young men alone. The study included men and women, and women became the majority of testosterone patients in the group age 24 and younger over the study period. The authors linked that shift to growth in gender-affirming care.

---

How common is low testosterone?

Answer capsule: The result depends on who was tested and how “low testosterone” was defined. One primary-care study classified 38.7% of men age 45+ as hypogonadal. A community study found 5.6% of men age 30–79 had low total and free testosterone plus symptoms. A European study found 2.1% under its strict symptom-and-lab definition.

Sources: Mulligan et al., International Journal of Clinical Practice 2006; Araujo et al., JCEM 2007; Wu et al., New England Journal of Medicine 2010. Untreated HIM rate calculated by TRT Provider Guide from the study’s printed counts.

The 38.7% HIM statistic needs two labels

The HIM study is real, but the popular wording “40% of men over 45 have low T” drops two facts.

First, the men were already in primary-care offices. They were not a random U.S. sample. The same paper found low testosterone was more common in men with obesity and diabetes.

Second, the study’s 836-man numerator included 80 men already receiving testosterone, whether or not their measured value at that visit was below 300 ng/dL. Remove those treated men from both sides and the measured low-testosterone rate among untreated participants is:

(836 − 80) ÷ (2,162 − 80) × 100 = 36.3%

The blood was drawn between 8 a.m. and noon. The paper does not report a fasting requirement. Current guidelines call for repeat morning testing before diagnosis.

The 2.1% EMAS statistic needs the right denominator

EMAS recruited 3,369 men. The main diagnostic analysis used 2,966 men with complete hormone and symptom data. Sixty-three met the strict definition:

63 ÷ 2,966 × 100 = 2.1%

Attaching 2.1% to all 3,369 recruited men quietly changes the denominator. The difference is small here, but the rule matters.

What current guidelines require

The American Urological Association guideline says clinicians should use total testosterone below 300 ng/dL as a reasonable cutoff, confirm it with two separate early-morning tests, and diagnose only when low levels are paired with symptoms or signs.

The Endocrine Society guideline also requires symptoms or signs plus consistently low, accurately measured testosterone, confirmed with a repeat morning fasting test.

The often-quoted 264 ng/dL number is not a universal diagnosis line. It is the 2.5th percentile in a harmonized reference range from healthy, nonobese men age 19–39. That range was 264 to 916 ng/dL and was built for assays tied to a CDC reference method. A reference limit helps interpret a lab. It does not replace symptoms, repeat testing, assay quality, or clinical judgment.

---

Why do TRT statistics disagree so much?

Answer capsule: Most disagreements are denominator problems. A low lab value, a diagnosed syndrome, an active prescription, a filled prescription, and a trial outcome are five different facts. Label the fact and the disagreement usually clears up.

We call this denominator drift.

Sources: HIM, BACH, Epic Research, Selinger and Thallapureddy, and Lincoff et al. Framework and table by TRT Provider Guide.

Before using a number, ask five questions: What was counted? Who was counted? What year? What definition was used? What can the number not prove?

Every row in the downloadable dataset carries those fields, a source URL, a verification date, and a caveat.

---

How many men get the recommended tests before TRT?

Answer capsule: In a 2026 review of 200 men at Michigan Medicine, only 12% had the full guideline-based workup before their first prescription. In a separate secret-shopper study, 6 of 7 direct-to-consumer platforms offered testosterone even though the shopper’s testosterone results were normal.

Michigan Medicine chart review

The ENDO 2026 study reviewed a random sample of 200 males assigned at birth, with a mean age of 52.5, who received a first outpatient testosterone prescription from 2020 through 2025.

To count as the full workup, the chart had to show:

• two low testosterone results drawn from 5 a.m. to 10 a.m.;
• luteinizing hormone and/or follicle-stimulating hormone testing, which helps find the cause; and
• no listed contraindication to testosterone therapy.

Source: Sinha and Papaleontiou, ENDO 2026, presented June 13, 2026. This is a one-system retrospective chart review reported at a conference, not a national estimate.

The same review found obesity in 63%, obstructive sleep apnea in 55%, high blood pressure in 52%, depression in 40%, diabetes in 28%, and arthritis in 28%. Four percent had prior prostate cancer, and 1.5% had a PSA above 4 ng/mL before the prescription.

That is useful evidence, but it has a clear weight: one health system, 200 charts, and conference-level reporting. We treat the direction as important and the national size as unknown.

Direct-to-consumer secret shopper

A 2022 JAMA Internal Medicine study used one standardized secret shopper with normal testosterone results and a stated wish to preserve future fertility.

Source: Dubin et al., JAMA Internal Medicine 2022. This was one shopper and seven platforms, so it is a direct snapshot, not a prevalence estimate for the whole industry.

The two studies do not prove how often poor workups happen nationwide. They do show that the problem exists in both a large academic system and direct-to-consumer care.

There is also an undertreatment problem. In the older BACH community survey, only 12.2% of men who met the study’s full definition of symptomatic androgen deficiency were receiving treatment. Overtreatment and undertreatment can happen at the same time, to different men.

---

Does TRT work? What the large trials measured

Answer capsule: There is no one TRT success rate. Testosterone improved sexual activity and desire, corrected anemia in more men than placebo, and helped some older men on selected measures. It did not significantly improve erectile function in the largest sexual-function trial, and the main vitality outcome in the Testosterone Trials was not significant.

Sexual activity, desire, and erections

The TRAVERSE Sexual Function Study included 1,161 men with low libido: 587 assigned to testosterone gel and 574 to placebo. Mean age was 63.7.

Source: Pencina et al., Journal of Clinical Endocrinology & Metabolism 2024.

That last row is easy to lose. Testosterone helped these men want sex more and have it more often. It did not show a significant benefit for erections in this trial.

Anemia

In TRAVERSE, 815 of 5,204 men had anemia at the start. The corrected paper reports these rates of anemia remission:

Source: Pencina et al., JAMA Network Open 2023, with the journal correction published January 23, 2024. Six-month counts also match ClinicalTrials.gov record NCT03518034.

The correction matters. Some early coverage paired 143 of 349 with 122 of 360 at six months. But 122 of 360 is 33.9%, the 12-month placebo value. The corrected six-month placebo count is 103 of 375, or 27.5%.

The Testosterone Trials in older men

The earlier Testosterone Trials enrolled 790 men age 65+ with two morning values averaging below 275 ng/dL and treated them for one year.

Sexual activity and desire improved. The main vitality outcome did not improve significantly. The main walking outcome did not improve significantly inside the Physical Function Trial, although the pooled analysis across all three main trials did show a difference.

Researchers screened 51,085 men to enroll 790:

51,085 ÷ 790 = 64.7

That is about 65 men screened for every one enrolled. It is not a prevalence estimate. It shows how selective this trial population was.

What a usable benefit claim needs

A benefit number needs four parts: the outcome, the population, the time point, and the comparison group. “TRT works for 80% of men” has none of them. Without those four parts, the percentage is decoration.

---

Does TRT raise heart attack or stroke risk?

Answer capsule: In TRAVERSE, testosterone gel was noninferior to placebo for the prespecified composite of cardiovascular death, nonfatal heart attack, and nonfatal stroke: 7.0% versus 7.3%. That result applies to selected men age 45–80 with symptoms, two fasting testosterone values below 300 ng/dL, and cardiovascular disease or high risk. It does not prove that every form or dose is “heart-safe.”

TRAVERSE randomized 5,246 men at 316 U.S. sites; 5,204 were included in the main analysis. The treatment was 1.62% testosterone gel, adjusted to a target of 350 to 750 ng/dL. Mean treatment time was 21.7 months, and mean follow-up was 33.0 months.

Noninferior means the trial met a preset test showing testosterone was not unacceptably worse than placebo for the main outcome. It does not mean testosterone protects the heart.

Source: Lincoff et al., New England Journal of Medicine 2023. Main-outcome hazard ratio 0.96, 95% CI 0.78 to 1.17. Absolute differences calculated by TRT Provider Guide.

Atrial fibrillation, acute kidney injury, and pulmonary embolism occurred more often in the testosterone group. These were not the prespecified primary endpoint, and the trial was not built to settle every secondary safety signal on its own. They still belong beside the main result. Leaving them out turns a narrow result into a blanket claim.

About six in ten participants stopped the assigned gel: 61.4% in the testosterone group and 61.7% in the placebo group. That high and nearly equal discontinuation rate is another limit on how far the result should be stretched.

---

What other TRT risks showed up in the trials?

Answer capsule: The clearest unexpected signal was fracture: 3.50% on testosterone versus 2.46% on placebo. Prostate-cancer events were too rare to settle long-term risk. Testosterone can also raise hematocrit and blood pressure, which is why lab and blood-pressure monitoring remain part of care.

Fractures

The TRAVERSE fracture study followed the 5,204 men for a median 3.19 years. Clinical fractures occurred in 3.50% of men assigned to testosterone and 2.46% assigned to placebo.

• Testosterone: 91 fractures
• Placebo: 64 fractures
• Hazard ratio: 1.43
• 95% confidence interval: 1.04 to 1.97
• Absolute difference: 1.04 percentage points

Testosterone can raise bone density. It did not lower the real-world fracture outcome in this trial. That is why a better lab marker cannot stand in for the event patients care about.

Prostate outcomes

The TRAVERSE prostate substudy followed a screened population over 14,304 person-years.

Source: Bhasin et al., JAMA Network Open 2023.

The differences were not statistically significant, but the event counts were tiny. The authors said the data should not be read as proof that the risks are equal.

The trial also excluded men at higher baseline risk, including men with prior prostate cancer, a prostate nodule, severe urinary symptoms, or PSA above its study limits. Prostate cancer can take years to appear. Short-term low event counts do not close the long-term question.

Hematocrit and blood pressure

Testosterone can cause erythrocytosis, meaning too many red blood cells. The Endocrine Society guideline uses hematocrit above 54% as the point to stop therapy, check for low oxygen or sleep apnea, and restart at a lower dose after the level falls. The guideline also calls for hematocrit checks before treatment, after 3 to 6 months, and then yearly.

There is no one honest “erythrocytosis rate” for all TRT. The result changes with the form, dose, blood level reached, patient age, study threshold, and follow-up time. A single range made from unlike thresholds and treatment forms would look precise but would not describe one shared outcome, so it is not used here.

In February 2025, the FDA required class-wide label changes after ambulatory blood-pressure studies found increases across testosterone products. The agency removed the cardiovascular boxed-warning language and added or strengthened blood-pressure warnings at the same time.

---

What happens to fertility, and how many men stay on TRT?

Answer capsule: Exogenous testosterone suppresses sperm production. Recovery is common after stopping, but it can take months and the best recovery data comes from healthy men in contraception studies, not long-term TRT patients. Persistence is also low in two very different datasets: 17% met a topical-therapy adherence threshold at 12 months, and about 61% stopped assigned gel in TRAVERSE.

Sources: Liu et al., Lancet 2006; Grabner et al., Journal of Sexual Medicine 2018; Lincoff et al., New England Journal of Medicine 2023.

The AUA/ASRM male infertility guideline says testosterone monotherapy should not be prescribed to men interested in current or future fertility. The Endocrine Society also recommends against starting testosterone when fertility is planned in the near term.

The recovery table should not be turned into a promise. Those men were healthy and took hormones for a defined study period. Age, starting sperm count, length of treatment, and the drug used affected recovery. A man who used TRT for years is not the same population.

The adherence number also needs its real label. The 17% figure means 17% met that paper’s adherence threshold at 12 months. It does not prove that 83% permanently quit treatment, felt no benefit, or stopped every form of TRT.

---

Are men’s testosterone levels falling?

Answer capsule: A published analysis of U.S. NHANES samples found mean total testosterone in males age 15–39 fell from 605.39 ng/dL in 1999–2000 to 451.22 ng/dL in 2015–2016, about 25.5%. But the survey cycles used three assay methods, so the size of the biological decline cannot be separated cleanly from measurement change.

Source: Lokeshwar et al., European Urology Focus 2021, analysis of 4,045 males age 15–39 in NHANES.

The raw change is:

(451.22 − 605.39) ÷ 605.39 × 100 = −25.5%

The authors reported that the trend remained after adjustment for several health and lifestyle factors. The assay issue still stays beside the result. NHANES used different methods across the cycles, and the harmonized-reference-range study showed that assay differences can explain a meaningful part of variation between testosterone datasets.

An older longitudinal study, the Massachusetts Male Aging Study, also found a birth-cohort effect: men measured in later calendar years had lower average testosterone than same-age men measured earlier. That points in the same direction, but it does not give one fixed annual decline that applies to every man.

We found no single “testosterone falls X% each year” number strong enough to publish as a universal rate. The estimate changes with age, cohort, adjustment, and whether total, free, or bioavailable testosterone is measured.

---

What changed in U.S. testosterone policy in 2025 and 2026?

Answer capsule: The FDA removed cardiovascular boxed-warning language in February 2025 and added a class-wide blood-pressure warning. In June 2026, HHS said FDA was requesting more label updates, including removal of the age-related limitation and changes to prostate and enlarged-prostate language. “Requested” is not the same as every product label already being changed.

Sources: U.S. Food and Drug Administration, February 28, 2025; U.S. Department of Health and Human Services, June 18, 2026; Endocrine Society, July 16, 2026. Last verified September 2, 2026.

The clean current summary is:

• The 2025 FDA action was a class-wide labeling change.
• The June 2026 announcement described requested further changes.
• The Endocrine Society still says diagnosis requires symptoms plus consistently low, accurately measured testosterone, and it does not support population screening of asymptomatic men.

This page does not claim that every product label already carries every requested 2026 revision. A product-specific claim should be checked against that product’s current FDA label.

---

How big is the TRT market?

Answer capsule: There is no open, reproducible 2026 market total in the primary sources we reviewed. The strongest public historical benchmark is a 41-country study that found global testosterone sales rose 12-fold, from $150 million in 2000 to $1.8 billion in 2011.

The Medical Journal of Australia study used IMS Health wholesale sales for every testosterone product in 41 countries.

Source: Handelsman, Medical Journal of Australia 2013. Dollar figures are wholesale sales, not patients, prescriptions, profit, or a current market valuation. The 90.2% figure is our calculation from 37 of 41.

The paper treated sales as a stand-in for prescribing and converted products into defined monthly doses. It also stated the assumptions: a steady relation between purchase, prescribing, and use, and a similar adult-male share across countries.

We do not print a current multibillion-dollar market estimate from a paid forecast or a news story. Those estimates use different product scopes and are not reproducible from public underlying data. The honest current answer is that a verified open total is not available here.

---

20 common TRT claims checked against primary sources

This ledger replaces a vague “most quoted” list. We did not run a quote-frequency study, so we do not claim these are the 20 most repeated claims on the internet. They are 20 common claims that can be checked against the original source.

Primary sources: Epic Research; PLOS ONE; Sellke et al.; Baillargeon et al.; HIM; BACH; EMAS; AUA; ENDO 2026; Dubin et al.; TRAVERSE reports; FDA; Liu et al.; Grabner et al.; and Lokeshwar et al. Full URLs, denominators, dates, and notes are in the downloadable ledger and full dataset.

---

How we built this page

Answer capsule: Every publishable number came from an original study, trial record, regulator, guideline body, or original data producer. We recorded the denominator, date, source URL, confidence, and caveat. We recalculated any derived figure from printed source values and did not pool unlike populations.

What we let in

A source qualified when it was one of these:

• an original peer-reviewed study;
• a clinical-trial report or registry record;
• an issuing regulator or government department;
• a specialty guideline body; or
• the producer of the underlying dataset.

A number also had to carry enough detail to label what was counted and who was counted. We excluded clinic testimonials, vendor outcome claims, paid market forecasts, and secondary numbers whose underlying report could not be checked.

The arithmetic

We used three basic formulas:

• Relative change: (new − old) ÷ old × 100
• Rate per 10,000: percentage ÷ 100 × 10,000
• Absolute difference: treatment percentage − comparison percentage

We did not merge active prescriptions, filled prescriptions, low lab values, diagnoses, trial outcomes, claims, or sales into one rate. We did not turn an encounter-based rate into a national headcount.

The checks done by hand

Four findings on this page come from recomputing or reconciling the source itself:

1. PLOS arithmetic: 1,216,982 to 1,656,641 is 36.1% by the standard start-value formula, while the paper prints 27%.
2. Medicare arithmetic: 542,040 to 751,976 is 38.7%, while the paper prints 54.6%.
3. HIM treated-patient adjustment: removing the 80 men already on testosterone changes the untreated measured rate from the paper’s 38.7% classification to 36.3%.
4. Anemia correction: the corrected six-month placebo result is 103 of 375, or 27.5%; 122 of 360 belongs to month 12.

The first three are original calculations from published source values. The fourth is a source-version check against the journal correction and trial registry.

Confidence labels used in the data file

• High: peer-reviewed primary report, official agency action, registry record, or original data producer with a clear denominator.
• Moderate: original conference report or a narrow study with a small sample that still reports its method clearly.
• Excluded: underlying source unavailable, denominator unclear, source conflict unresolved, or claim outside what the study measured.

Authorship and review

The visible byline is TRT Provider Guide. No medical reviewer is claimed. This is educational reference material, not a diagnosis or treatment plan.

---

What this data cannot tell you

Answer capsule: No dataset here can say exactly how many U.S. men use TRT or how many prescriptions are appropriate. The safety trial tested gel in a selected older population. The best U.S. symptom-plus-lab prevalence estimate is based on data collected more than 20 years ago.

• No national user count exists in these public sources. Each system misses cash pay, in-office treatment, nonparticipating states, or people outside its health systems.
• The prescribing datasets do not pair every prescription with a full clinical diagnosis. They cannot measure appropriateness.
• Epic is not a fixed panel. Its contributing organizations grew over time, and early years have thinner representation.
• The PDMP study is a 24-state analytic sample. It includes men and women, has no diagnosis, and misses clinic-administered testosterone.
• The strongest U.S. symptomatic-prevalence estimate is old. BACH used 2002–2005 data.
• TRAVERSE tested gel in men age 45–80 with low testosterone and cardiovascular disease or risk. It does not settle injections, higher targets, younger men, or long-term use.
• About six in ten TRAVERSE participants stopped assigned treatment. That limits how far results can be stretched.
• The 12% workup figure is one health system and conference-level evidence. It is not a national percentage.
• Testosterone assays are not interchangeable. That limits comparisons across years, labs, and cohorts.
• The 2026 Epic figure is partial. It covers data through June 22, not a full year.
• The 2026 HHS label announcement described requested changes. Product-specific wording must be checked on the current product label.

---

How to cite this page

These are neutral publication details for accurate attribution.

Web page
TRT Provider Guide. “TRT Statistics 2026: Use, Safety, Benefits, and Source Data.” TRT Provider Guide. Published and last verified September 2, 2026. https://trtproviderguide.com/research/trt-statistics/

Dataset
TRT Provider Guide. TRT Statistics Dataset, 2026 edition. Published and last verified September 2, 2026. https://trtproviderguide.com/research/trt-statistics/#data

For one statistic, the primary source named in that row is the source of the underlying fact. This page is the compilation. Its original work is the denominator labeling, source reconciliation, arithmetic audits, derived conversions, and common-claims ledger.

---

Frequently asked questions

What percentage of men are on TRT?

No study covers all U.S. men. The best current benchmark is that 0.93% of male Epic Cosmos patients with a healthcare encounter had an active testosterone prescription in 2025. That is 93 per 10,000. It is not a population rate for every U.S. man.

How many men use TRT in the United States?

No nationwide public source in this review gives a complete headcount. A 24-state study counted 1,656,641 unique people filling testosterone in 2022, but it included men and women and missed clinic-administered treatment.

Is TRT use increasing?

Yes in the current Epic series. The rate fell from 0.83% in 2013 to 0.64% in 2021, then rose to 0.93% in 2025. That is a 45.3% relative rise from 2021.

What age group uses TRT the most?

Men age 50 to 64 had the highest 2025 rate in Epic Cosmos: 1.57%, or about 1 in 64 male patients with an encounter.

What percentage of men have low testosterone?

It depends on the definition. BACH found 5.6% of community men age 30–79 had low total and free testosterone plus symptoms. HIM classified 38.7% of men age 45+ visiting primary care, but that definition included men already receiving testosterone.

Is 264 ng/dL the cutoff for low testosterone?

Not by itself. It is the lower end of a harmonized reference range in healthy, nonobese men age 19–39 using standardized assays. Diagnosis still requires symptoms or signs and consistently low results on repeat morning testing.

How many men get the full workup before TRT?

A 2026 Michigan Medicine review found 12% in a random sample of 200 first-time prescriptions. It was one health system and conference-level evidence, not a national estimate.

What is the TRT success rate?

There is no one rate. In separate trial analyses, testosterone added 0.49 sexual acts per day at six months and corrected anemia in 41.0% versus 27.5% on placebo. It did not significantly improve erectile function in the largest sexual-function trial.

Is TRT safe for the heart?

TRAVERSE found its main cardiac outcome in 7.0% on testosterone gel and 7.3% on placebo, meeting noninferiority. Atrial fibrillation, acute kidney injury, and pulmonary embolism occurred more often on testosterone, so a blanket “heart-safe” claim goes too far.

Does TRT raise fracture risk?

In TRAVERSE, clinical fractures occurred in 3.50% on testosterone and 2.46% on placebo over a median 3.19 years. The hazard ratio was 1.43.

Does TRT cause prostate cancer?

TRAVERSE had too few prostate-cancer events and too little follow-up to settle long-term risk. Short-term differences were not statistically significant, but the study excluded higher-risk men.

Does TRT cause infertility?

It suppresses sperm production. Recovery was common after stopping in pooled studies of healthy men, but timing varied and those data do not guarantee recovery after long-term TRT. Guidelines advise against testosterone when fertility is desired.

How many men stay on TRT?

One claims study found 17% met its adherence threshold for topical therapy at 12 months. In TRAVERSE, 61.4% on testosterone and 61.7% on placebo stopped the assigned gel. Those measures are not the same as permanent discontinuation of all TRT.

Are men’s testosterone levels declining?

A NHANES analysis found mean total testosterone in males age 15–39 fell 25.5% from 1999–2000 to 2015–2016. Because the survey cycles used different assay methods, the exact biological decline remains uncertain.

What changed in FDA testosterone rules in 2025 and 2026?

In February 2025, FDA removed cardiovascular boxed-warning language and added blood-pressure warning changes while retaining the age-related limitation. In June 2026, HHS said FDA was requesting removal of that limitation and other label updates. A request is not proof that every product label already changed.

How current are these TRT statistics?

Every publishable source and calculation was rechecked on September 2, 2026. The 2026 Epic prescribing figure covers data through June 22 and is labeled partial.

---

Primary sources

1. Bartelt K, Cox C. After Nearly a Decade-Long Decrease, Testosterone Prescribing Rates Rising Again — Epic Research, July 29, 2026.
2. Selinger S, Thallapureddy A. Cross-sectional analysis of national testosterone prescribing through prescription drug monitoring programs, 2018–2022 — PLOS ONE 2024;19:e0309160.
3. Baillargeon J, Kuo YF, Westra JR, Urban RJ, Goodwin JS. Testosterone Prescribing in the United States, 2002–2016 — JAMA 2018;320:200–202.
4. Baillargeon J, Urban RJ, Ottenbacher KJ, Pierson KS, Goodwin JS. Trends in androgen prescribing in the United States, 2001 to 2011 — JAMA Internal Medicine 2013;173:1465–1466.
5. Sellke N, et al. Trends in testosterone prescription during the release of society guidelines — International Journal of Impotence Research 2024;36:380–384.
6. Mulligan T, Frick MF, Zuraw QC, Stemhagen A, McWhirter C. Prevalence of hypogonadism in males aged at least 45 years: the HIM study — International Journal of Clinical Practice 2006;60:762–769.
7. Araujo AB, et al. Prevalence of symptomatic androgen deficiency in men — Journal of Clinical Endocrinology & Metabolism 2007;92:4241–4247.
8. Hall SA, et al. Treatment of symptomatic androgen deficiency: results from the Boston Area Community Health Survey — Archives of Internal Medicine 2008;168:1070–1076.
9. Wu FCW, et al. Identification of late-onset hypogonadism in middle-aged and elderly men — New England Journal of Medicine 2010;363:123–135.
10. Travison TG, et al. Harmonized Reference Ranges for Circulating Testosterone Levels in Men of Four Cohort Studies in the United States and Europe — Journal of Clinical Endocrinology & Metabolism 2017;102:1161–1173.
11. Bhasin S, et al. Testosterone Therapy in Men With Hypogonadism: An Endocrine Society Clinical Practice Guideline — Journal of Clinical Endocrinology & Metabolism 2018;103:1715–1744.
12. Evaluation and Management of Testosterone Deficiency: AUA Guideline — American Urological Association.
13. Testosterone therapy in men may be overprescribed, inconsistent with clinical guidelines — Endocrine Society, ENDO 2026, June 13, 2026.
14. Dubin JM, et al. Guideline-Discordant Care Among Direct-to-Consumer Testosterone Therapy Platforms — JAMA Internal Medicine 2022;182:1321–1323.
15. Lincoff AM, et al. Cardiovascular Safety of Testosterone-Replacement Therapy — New England Journal of Medicine 2023;389:107–117.
16. Snyder PJ, et al. Testosterone Treatment and Fractures in Men with Hypogonadism — New England Journal of Medicine 2024;390:203–211.
17. Bhasin S, et al. Prostate Safety Events During Testosterone Replacement Therapy in Men With Hypogonadism — JAMA Network Open 2023;6:e2348692.
18. Pencina KM, et al. Efficacy of Testosterone Replacement Therapy in Correcting Anemia in Men With Hypogonadism — JAMA Network Open 2023;6:e2340030; correction January 23, 2024.
19. TRAVERSE registry record, NCT03518034 — ClinicalTrials.gov.
20. Pencina KM, et al. Effect of Testosterone Replacement Therapy on Sexual Function and Hypogonadal Symptoms — Journal of Clinical Endocrinology & Metabolism 2024;109:569–580.
21. Snyder PJ, et al. Effects of Testosterone Treatment in Older Men — New England Journal of Medicine 2016;374:611–624.
22. Liu PY, et al. Rate, extent, and modifiers of spermatogenic recovery after hormonal male contraception — Lancet 2006;367:1412–1420.
23. Grabner M, et al. Topical Testosterone Therapy Adherence and Outcomes Among Men With Primary or Secondary Hypogonadism — Journal of Sexual Medicine 2018;15:148–158.
24. Diagnosis and Treatment of Infertility in Men: AUA/ASRM Guideline — American Urological Association and American Society for Reproductive Medicine.
25. Lokeshwar SD, et al. Decline in Serum Testosterone Levels Among Adolescent and Young Adult Men in the USA — European Urology Focus 2021;7:886–889.
26. Travison TG, et al. A population-level decline in serum testosterone levels in American men — Journal of Clinical Endocrinology & Metabolism 2007;92:196–202.
27. FDA issues class-wide labeling changes for testosterone products — U.S. Food and Drug Administration, February 28, 2025.
28. HHS Announces Requested Updates to Testosterone Therapy Product Labels — U.S. Department of Health and Human Services, June 18, 2026.
29. Statement on Testosterone Replacement Therapy — Endocrine Society, July 16, 2026.
30. Handelsman DJ. Global trends in testosterone prescribing, 2000–2011 — Medical Journal of Australia 2013;199:548–551.

---

TRT Provider Guide is an independent research and reference resource. This page has no advertising, affiliate links, or provider recommendations. It is educational material, not medical advice.