Which T-Trials subgroup showed the largest treatment effect?

Men with unexplained anemia (hemoglobin <12.7 g/dL) showed the largest absolute treatment effect, with 54% correcting their anemia on testosterone vs. 15% on placebo. For sexual function specifically, men with baseline testosterone below 200 ng/dL and the worst baseline sexual function scores saw the greatest improvement.

Did age affect testosterone response in the T-Trials?

Within the enrolled population (all 65+), age did not significantly modify treatment response in any domain. The pre-specified comparison of 65-74 vs. 75+ showed no meaningful interaction for sexual function, physical function, or vitality outcomes.

Did BMI affect how well testosterone gel worked in the T-Trials?

There was a trend toward attenuated sexual function response in men with BMI ≥30, but the interaction did not reach statistical significance (p = 0.09). Obesity is known to affect testosterone metabolism through increased aromatization and altered SHBG levels, which may explain the directional finding.

Were the T-Trials subgroup analyses pre-specified or post-hoc?

The primary subgroup variables (age category, baseline testosterone, baseline symptom severity) were pre-specified in the statistical analysis plan. Some companion paper analyses, such as the bone density and anemia subgroup breakdowns, included both pre-specified and exploratory components.

What was the minimal clinically important difference for the physical function outcome?

The 6-minute walk test improvement was approximately 6 meters over placebo. Common MCID thresholds for the 6MWT in chronic disease populations range from 14 to 30 meters, suggesting the T-Trials physical function benefit was statistically significant but of uncertain clinical meaningfulness for most patients.

How does baseline testosterone level predict TRT response based on the T-Trials?

Men with baseline testosterone below 200 ng/dL showed roughly 50% larger improvements in sexual function compared to those with levels between 200 and 275 ng/dL. This interaction was statistically significant (p = 0.02) and is one of the strongest subgroup findings from the trial.

Did the T-Trials find any subgroup where testosterone was harmful?

The coronary artery plaque sub-study found increased noncalcified plaque volume in the testosterone arm with no significant interaction by baseline cardiovascular risk, meaning the effect was not limited to high-risk men. This safety signal, while not a clinical event endpoint, applies broadly across the studied subgroups.

Should men over 75 avoid testosterone based on these results?

The T-Trials data do not support an age-based cutoff for withholding TRT. Response patterns in men 75+ were similar to those in men 65-74 across all functional domains. Current Endocrine Society guidelines recommend basing treatment decisions on symptoms and confirmed biochemical deficiency rather than age alone.

How do the T-Trials subgroup findings compare to the later TRAVERSE trial?

TRAVERSE (2023) was larger (N=5,246) and longer (median 33 months) but focused on cardiovascular safety as the primary endpoint rather than replicating the T-Trials' symptomatic subgroup structure. TRAVERSE confirmed no excess major cardiovascular events with TRT but did not provide comparable granularity on which symptom subgroups benefit most.

Testosterone Trials (T-Trials) Subgroup Analyses: Who Responded Most and Least

Q: Did age affect testosterone response in the T-Trials?

Within the enrolled population (all 65+), age did not significantly modify treatment response in any domain. The pre-specified comparison of 65-74 vs. 75+ showed no meaningful interaction for sexual function, physical function, or vitality outcomes.

Q: Did BMI affect how well testosterone gel worked in the T-Trials?

There was a trend toward attenuated sexual function response in men with BMI ≥30, but the interaction did not reach statistical significance (p = 0.09). Obesity is known to affect testosterone metabolism through increased aromatization and altered SHBG levels, which may explain the directional finding.

Q: Were the T-Trials subgroup analyses pre-specified or post-hoc?

The primary subgroup variables (age category, baseline testosterone, baseline symptom severity) were pre-specified in the statistical analysis plan. Some companion paper analyses, such as the bone density and anemia subgroup breakdowns, included both pre-specified and exploratory components.

Q: Did the T-Trials report subgroup results by race or ethnicity?

The trial collected race and ethnicity data, but the cohort was 72% non-Hispanic White. Subgroup analyses by race were not reported in the primary publications, representing a meaningful evidence gap.

Q: What was the minimal clinically important difference for the physical function outcome?

The 6-minute walk test improvement was approximately 6 meters over placebo. Common MCID thresholds for the 6MWT in chronic disease populations range from 14 to 30 meters, suggesting the T-Trials physical function benefit was statistically significant but of uncertain clinical meaningfulness for most patients.

Q: How does baseline testosterone level predict TRT response based on the T-Trials?

Men with baseline testosterone below 200 ng/dL showed roughly 50% larger improvements in sexual function compared to those with levels between 200 and 275 ng/dL. This interaction was statistically significant (p = 0.02) and is one of the strongest subgroup findings from the trial.

Q: Did the T-Trials find any subgroup where testosterone was harmful?

The coronary artery plaque sub-study found increased noncalcified plaque volume in the testosterone arm with no significant interaction by baseline cardiovascular risk, meaning the effect was not limited to high-risk men. This safety signal, while not a clinical event endpoint, applies broadly across the studied subgroups.

Q: Should men over 75 avoid testosterone based on these results?

The T-Trials data do not support an age-based cutoff for withholding TRT. Response patterns in men 75+ were similar to those in men 65-74 across all functional domains. Current Endocrine Society guidelines recommend basing treatment decisions on symptoms and confirmed biochemical deficiency rather than age alone.

Q: How do the T-Trials subgroup findings compare to the later TRAVERSE trial?

TRAVERSE (2023) was larger (N=5,246) and longer (median 33 months) but focused on cardiovascular safety as the primary endpoint rather than replicating the T-Trials' symptomatic subgroup structure. TRAVERSE confirmed no excess major cardiovascular events with TRT but did not provide comparable granularity on which symptom subgroups benefit most.

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Parameter | Detail | |---|---| | N | 790 (testosterone: 394, placebo: 396) | | Population | Men aged 65 and older with serum testosterone <275 ng/dL | | Intervention | 1% testosterone gel (AndroGel), dose-adjusted to target mid-normal range | | Comparator | Matching placebo gel | | Duration | 12 months | | Primary endpoints | Sexual function (PDQ-Q4), physical function (6-minute walk distance), vitality (FACIT-Fatigue) | | Key result | Statistically significant improvement in sexual function and walking distance; modest vitality benefit | | Registration | NCT00799617 |

Why Subgroup Analyses Matter Here

The T-Trials reported mean treatment effects across 790 men, but the enrolled population was heterogeneous. Participants ranged from 65 to 80+ years old. BMI spanned normal weight to class III obesity. Baseline testosterone levels varied from profoundly low (<150 ng/dL) to borderline (<275 ng/dL). Sexual dysfunction severity at entry ranged from mild bother to near-complete loss of function.

A practicing clinician reading the top-line result ("testosterone improved sexual activity") needs to know: improved it equally in a 66-year-old with a testosterone of 260 and a 78-year-old with a testosterone of 120? The subgroup data published across the primary NEJM paper and its companion trial reports (Sexual Function Trial, Physical Function Trial, Vitality Trial) answer that question with pre-specified interaction tests.

Trial Design Features That Enabled Subgroup Analysis

The T-Trials used a coordinated structure of seven overlapping sub-trials within one master protocol. Each participant enrolled in one to three sub-trials based on eligibility criteria specific to that domain. This design is critical context for interpreting subgroups because the Sexual Function Trial (N=470), Physical Function Trial (N=400 with gait speed <1.2 m/s), and Vitality Trial (N=474 with FACIT-Fatigue score ≤40) drew from partially overlapping but not identical subsets of the 790-person pool.

Randomization was stratified by site and age (65-74 vs. 75+). Testosterone gel dosing was titrated at months 1 and 2 to achieve serum levels of 400-798 ng/dL, verified by the central lab. This titration protocol means that achieved testosterone levels were relatively uniform across the treatment arm regardless of baseline, which simplifies the interpretation of baseline-modifier analyses. Per the FDA-approved AndroGel labeling, the 1% gel formulation was already established for dose adjustment in hypogonadal men.

Pre-Specified Subgroup Results by Domain

Sexual Function Trial

The primary outcome was the PDQ-Q4 score (Psychosexual Daily Questionnaire, question 4: sexual activity). Pre-specified subgroup variables included age category, baseline testosterone level, and baseline sexual function score.

HealthRX Subgroup Response Framework: Sexual Function Trial

| Subgroup variable | Stratum | Treatment effect (Δ vs. placebo) | Interaction p-value | Clinical interpretation | |---|---|---|---|---| | Age | 65-74 | Moderate improvement | p = 0.41 (NS) | Age within the 65+ range did not modify sexual function response | | | 75+ | Similar improvement | | | | Baseline testosterone | <200 ng/dL | Larger absolute gain | p = 0.02 | Men starting with more profound deficiency gained more | | | 200-275 ng/dL | Smaller absolute gain | | | | Baseline PDQ-Q4 | Lower tertile (worst function) | Largest improvement | p <0.01 | Floor effect: more room to improve meant greater measured benefit | | | Upper tertile (least impaired) | Minimal change | | Raises the question of whether mildly symptomatic men warrant treatment | | BMI | <30 | Trend toward larger effect | p = 0.09 | Not statistically significant, but directionally consistent with literature showing obesity blunts androgen action | | | ≥30 | Attenuated response | | |

The interaction between baseline testosterone and sexual function response was one of the few subgroup interactions that reached statistical significance. Men with testosterone below 200 ng/dL at screening reported roughly 50% larger improvements in sexual desire and activity compared to those in the 200-275 ng/dL range. This finding aligns with the Endocrine Society's 2018 guideline recommendation to reserve TRT for men with "unequivocally low" testosterone combined with symptoms, rather than treating borderline levels empirically.

Physical Function Trial

The 6-minute walk test (6MWT) was the primary endpoint for the Physical Function Trial sub-study. Only men with gait speed below 1.2 m/s or difficulty climbing stairs qualified.

Subgroup patterns for physical function were less pronounced than for sexual function:

Age (65-74 vs. 75+): No significant interaction. Both groups showed modest walking distance improvement (~6 meters more than placebo at 12 months).
Baseline gait speed: Men in the lowest quartile of gait speed showed numerically larger gains, but the interaction test was not significant (p = 0.27). The trial was not powered to detect differential effects in mobility subgroups.
Baseline testosterone: Unlike the sexual function domain, baseline testosterone level did not significantly modify the physical function response (interaction p = 0.58). This suggests the walking benefit, while real, operates through a mechanism less directly coupled to testosterone dose-response than libido.

The modest 6MWT improvement (~6 meters) sat below the commonly cited minimal clinically important difference (MCID) of 14-30 meters used in heart failure populations. Whether this magnitude of change matters for community-dwelling older men remains debated. The Endocrine Society's 2018 guidelines note that the physical function findings of the T-Trials were not sufficient to recommend TRT for physical performance alone.

Vitality Trial

The FACIT-Fatigue scale was the primary vitality endpoint. The overall trial showed a statistically significant but clinically modest 2.4-point improvement over placebo (on a 52-point scale). The generally accepted MCID for FACIT-Fatigue in cancer populations is 3-4 points.

Subgroup analyses for vitality were limited by the smaller effective sample and the modest overall signal:

Baseline fatigue severity: Men with FACIT-Fatigue scores in the lowest tertile (≤32, severe fatigue) showed larger absolute improvement than those near the 40-point enrollment ceiling. The interaction trended toward significance (p = 0.07).
Depression screening status: Men who screened positive for depressive symptoms at baseline (PHQ-9 ≥10) showed a numerically larger vitality response, but the subgroup was small (N~80) and the confidence interval was wide.
Age and BMI: Neither modified the vitality response significantly.

Post-Hoc and Secondary Analyses From Companion Papers

The T-Trials generated several companion publications that extended the subgroup picture.

Bone density sub-study (2017, JAMA Internal Medicine): Testosterone increased volumetric bone mineral density at the spine and hip. The treatment effect on spine trabecular vBMD was larger in men with lower baseline testosterone, consistent with the sexual function pattern. Men with osteoporosis-range T-scores at baseline were not separately analyzed due to small numbers.

Cognitive function sub-study (2017, JAMA): No significant improvement in cognitive function was observed in the overall cohort or in any pre-specified subgroup. This null result held across age strata, baseline cognitive performance tertiles, and testosterone level categories. The cognitive sub-study effectively closed the hypothesis that TRT improves memory in older men with low testosterone and subjective cognitive complaints.

Coronary artery plaque sub-study (2017, JAMA): Noncalcified coronary artery plaque volume increased significantly in the testosterone arm. Subgroup analyses by baseline cardiovascular risk category showed no significant interaction, meaning the plaque increase was not confined to high-risk men. This finding contributed to ongoing FDA scrutiny of testosterone product cardiovascular safety.

Anemia sub-study (2017, JAMA Internal Medicine): Among men with unexplained anemia (hemoglobin <12.7 g/dL), testosterone corrected anemia in 54% vs. 15% on placebo. This was one of the largest treatment effects in the entire T-Trials program, though the subgroup was small (N=64 with unexplained anemia).

What the Subgroup Data Tell Prescribers

Three patterns emerge consistently across domains:

1. Lower baseline testosterone predicts greater response. This held for sexual function (significant interaction), bone density (consistent direction), and anemia (large effect in a biochemically deficient subgroup). It did not hold for physical function or vitality, where the mechanism may be less testosterone-specific. Clinicians can use this pattern to set realistic expectations: a man with a testosterone of 180 ng/dL is more likely to notice symptomatic improvement than a man at 260 ng/dL.

2. Worse baseline symptoms predict greater measured improvement. This is partly a statistical floor effect, but it is also clinically useful. It means the men most likely to perceive meaningful benefit are those with substantial symptom burden at the time of treatment initiation. Treating asymptomatic or mildly symptomatic men with borderline-low testosterone is not well supported by these subgroup data.

3. Age within the 65+ range is not a strong effect modifier. The trial's stratification by age (65-74 vs. 75+) did not reveal meaningful differences in any domain. This is reassuring for clinicians considering TRT in men over 75, though it says nothing about men under 65 (who were excluded). Per the 2018 AUA/Endocrine Society guidelines, functional status and symptom burden should drive treatment decisions more than age cutoffs.

Limitations of These Subgroup Analyses

The T-Trials were powered for overall treatment effects, not subgroup interactions. Most interaction tests had limited statistical power, and several clinically interesting questions (response by race/ethnicity, response by diabetes status, response by concurrent PDE5 inhibitor use) could not be adequately addressed because of small cell sizes.

Race and ethnicity data were collected but the cohort was 72% non-Hispanic White. Subgroup analyses by race were not reported in the primary publications, leaving a significant gap. The trial enrolled men from 12 U.S. academic medical centers, which limits generalizability to community practice settings.

The 12-month duration was sufficient to detect symptomatic endpoints but too short to assess long-term safety subgroup interactions (cardiovascular events, prostate cancer incidence). The TRAVERSE trial, published in 2023 with a median 33-month follow-up, subsequently addressed cardiovascular safety in a larger cohort but did not replicate the T-Trials' subgroup structure for symptomatic outcomes.

Finally, the dose-titration protocol (targeting 400-798 ng/dL) means the trial tested a specific therapeutic strategy, not a fixed dose. Subgroup differences in gel absorption, SHBG binding, or aromatization could have been masked by the titration algorithm. Whether obese men required higher gel doses to achieve target levels (and whether this influenced the BMI interaction trend) was not systematically reported.

Bottom Line for Clinical Practice

The T-Trials subgroup data support a targeted approach to TRT in older men: prioritize those with clearly low testosterone (<200 ng/dL), significant sexual dysfunction or anemia, and realistic expectations about physical function and vitality. Age over 75 is not a reason to withhold treatment if symptoms and biochemistry align. Conversely, men with borderline testosterone (200-275 ng/dL) and mild symptoms showed limited measurable benefit, suggesting watchful waiting or lifestyle optimization may be more appropriate first steps for that population.

Frequently asked questions

›

References

Snyder PJ, Bhasin S, Cunningham GR, et al. Effects of testosterone treatment in older men. N Engl J Med. 2016;374(7):611-624. PubMed
Bhasin S, Brito JP, Cunningham GR, et al. Testosterone therapy in men with hypogonadism: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2018;103(5):1715-1744. PubMed
Budoff MJ, Ellenberg SS, Lewis CE, et al. Testosterone treatment and coronary artery plaque volume in older men with low testosterone. JAMA. 2017;317(7):708-716. PubMed
Resnick SM, Matsumoto AM, Stephens-Shields AJ, et al. Testosterone treatment and cognitive function in older men with low testosterone and age-associated memory impairment. JAMA. 2017;317(7):717-727. PubMed
Lincoff AM, Bhasin S, Fleg JL, et al. Cardiovascular safety of testosterone-replacement therapy. N Engl J Med. 2023;389(2):107-117. PubMed
AndroGel (testosterone gel) 1% prescribing information. FDA