From Pages to Practice

By Suellen Li, MD

Published March 13, 2024


Testosterone plays a critical role in men’s skeletal health, increasing bone mineral density throughout adulthood and maintaining it as they age. Men who undergo androgen deprivation therapy to reduce testosterone levels have decreased bone mineral density and increased risk of fracture. Conversely, older men with low testosterone who receive testosterone replacement therapy have increased bone mineral density and bone strength. These data led many to hypothesize that testosterone therapy could reduce fracture risk, but high-quality evidence does not exist to support this theory.

The TRAVERSE trial, published in 2023, offered an opportunity to study the effect of testosterone replacement on fracture risk. This randomized, double-blind, placebo-controlled trial examined the cardiovascular safety of daily transdermal testosterone gel in more than 5,000 men (age range, 45–80 years) with symptomatic hypogonadism. Participants were at high risk for cardiovascular disease and had low testosterone levels.

Although the results indicated that testosterone replacement therapy was noninferior to placebo with respect to cardiovascular safety, the big surprise came when researchers analyzed the clinical fracture data. The Fracture trial was a subtrial of the TRAVERSE trial and powered to detect a 30% lower risk of fracture in the testosterone group. However, after a median follow-up of 3 years, more participants in the testosterone group than in the placebo group experienced a clinical fracture (3.5% vs. 2.5%).

Given that the researchers had expected the opposite outcome, they struggled to find a good explanation for these findings. They cited one study in which testosterone improved the structural and mechanical parameters of trabecular bone but negatively affected measures of bone strength in cortical bone. However, fracture incidence in the TRAVERSE trial increased almost immediately after treatment initiation, which is arguably too fast for testosterone to have significantly affected bone strength.

Editorialists proposed an alternative explanation, noting that most excess fractures in the testosterone group were ankle or rib fractures, which are typically traumatic. We also know that men on testosterone therapy perform better on the 6-minute walk test and report higher physical function. Could the improved physical function in the testosterone group have led to those participants engaging more in high-risk activities that could cause fractures?

Ultimately, any proposed explanations are speculative, as the TRAVERSE trial did not track physical activity or risk behavior. Thus, this study raised more questions about testosterone and fracture risk in men than it answered. Future trials are needed to elucidate the potential mechanisms behind testosterone’s effect on bone health. For now, men who are considering testosterone therapy should be counseled about this potential adverse effect of treatment.

Read the following NEJM Journal Watch summary for more details of this study.

Does Testosterone Therapy Prevent Fractures in Older Men with Mild Hypogonadism?

Allan S. Brett, MD, reviewing Snyder PJ et al. N Engl J Med 2024 Jan 18 Grossmann M and Anawalt BD. N Engl J Med 2024 Jan 18

Testosterone therapy is thought to increase bone density in older men with relatively low testosterone levels, but whether it prevents fractures is unclear. To address this question, researchers conducted a “subtrial” within TRAVERSE, the previously published randomized trial designed primarily to examine the cardiovascular (CV) safety of testosterone supplementation (NEJM JW Gen Med Jul 15 2023 and N Engl J Med 2023; 389:107).

In TRAVERSE, 5200 middle-aged and older men (mean, 63 years) with testosterone levels <300 ng/dL (median, 227 ng/dL) received testosterone gel or placebo. During median follow-up of about 3 years, the overall incidence of fractures (excluding sternum, fingers, toes, skull, and face) was significantly higher in the testosterone group than in the placebo group (3.5% vs. 2.5%). Fracture incidences were too low to draw meaningful conclusions by fracture site: For example, major osteoporotic fractures (i.e., hip, spine, humerus, or wrist) occurred in only 36 testosterone recipients and 30 placebo recipients.

Comment: These results surprised the researchers, who expected a possible reduction, not an increase, in fracture incidence with testosterone therapy; neither the study authors nor editorialists have a good explanation for the findings. Editorialists speculate that perhaps testosterone therapy led to behavioral changes that led to excess fracture risk, but this explanation seems questionable: Patients were blinded to treatment, and treatment only raised average serum testosterone levels to ≈350 to 400 ng/dL. In any case, for patients like these study participants, we now need to mention a small excess risk for fracture as a possible adverse effect of testosterone supplementation. But the results don't apply to men with hypogonadism from clearly documented hypothalamic-pituitary or testicular disease. We recently summarized another substudy from TRAVERSE that addressed potential effects of testosterone therapy on the prostate (NEJM JW Gen Med Feb 1 2024 and JAMA Netw Open 2023; 6:2348692).

Browse more From Pages to Practice »

Suellen Li, MD, is a 2023–2024 NEJM Editorial Fellow. She is a hospitalist at Massachusetts General Hospital, where she also completed her internal medicine residency.