A person’s bone strength and the risk of falls determine fracture risk. Over one-third of community-dwelling adults who are at least 75 years of age will fall each year, making it a major public health concern.

As men age, decreasing serum levels of testosterone and bioavailable testosterone have been reported, which are associated with reduced muscle strength, lean mass loss, and impaired mobility, resulting in increased fall risk.

To date, available data related to testosterone therapy and incident fall risk are inconsistent.

Previous studies have used radio-immunoassay methods to measure serum sex steroid levels; however, the limited specificity and accuracy of these methods provide concerns, especially at low concentrations.

This article published in the June 2017 issue of the Journal of Bone and Mineral Researchwas written by Liesbeth Vandenput, PhD, from the University of Gothenburg, Sweden, and colleagues, and evaluates the relationship between serum sex steroid levels and incident falls in a large population of older men.

“No data are available documenting the association between serum sex steroid levels measured by mass spectrometry, considered the gold standard, and the risk of incident falls in men,” the authors wrote in the Journal of Bone and Mineral Research.

Study design and participant population

This multicenter Osteoporotic Fractures in Men (MrOS) study included older men in the United States, Sweden, and Hong Kong. The US sample included 5994 community-dwelling men of at least 65 years old who could walk without assistance, including 70.3% non-Hispanic whites, 11.5% African Americans, 8.8% Asians, and 9.4% other races and ethnicities. The Swedish sample included 3014 men aged 69 to 81 years who could walk without assistance. The Hong Kong sample included 2000 Chinese men aged 65 years or older who could walk without assistance.

Participants who had missing information related to body mass index, prevalent falls, or incident falls; who had surgical or chemical castration; or who had received androgen or antiandrogen treatment were excluded from the study. The total sample eligible for analyses included 1919 men in the US, 2495 men in Sweden, and 1469 men in Hong Kong.

A standardized questionnaire was used to obtain information related to age, race/ethnicity, prevalent diseases, and use of central nervous system medication. Factors such as alcohol consumption, dizziness, use of walking aids, falls, and mobility limitation were provided by participant self-report.

To estimate physical performance, measurements of grip strength, a timed chair stands test, a 20-cm narrow walking test, and a 6-m walking test were used. Grip strength was measured using a Jamar hand dynamometer, with the maximum value from two trials of both hands analyzed. In the timed chair stands test, research staff recorded the time to complete 5 chair stands without using arms to rise. As a measure of dynamic balance, participants walked a 6-m course within a 20-cm narrow path.

Tri-annual mailed questionnaires (Sweden and US) or telephone calls (Hong Kong) were used to collect information related to incident falls. Participants were followed until the end of the study, death, emigration, or discontinuation from the study.

Gas chromatography mass spectrometry was used to assess sex hormone levels, using a valid testosterone measurement as a reference, of 2022 men in the US, 2639 men in Sweden, and 1489 men in Hong Kong.

In the combined cohort, the average age was 74 years. A total of 50.7% of the men reported 1 or more falls during an average follow-up of 5.7 years. The incident fall rate was lowest in Hong Kong and highest in the US.

Sex steroids as risk factors for incident falls

Participants with prevalent falls at baseline had increased odds of incident falls in each sample and in the overall study population (OR = 2.68, 95% CI 2.57-2.80). In models adjusted for age, race, site, study cohort, and prevalent falls, serum levels of total and bioavailable testosterone were associated with an increased likelihood of incident falls. Serum levels of estradiol, bioavailable estradiol, and sex hormone-binding globulin were not associated with incident falls.

A significant non-linear relationship between serum total or bioavailable testosterone and fall risk was found using quadratic models, with cubic spline analyses indicating that men with total or bioavailable testosterone levels in the lowest quartile had the most prominent increased risk of falls. In participants with hypogonadism, both low total and bioavailable testosterone were associated with increased falls.

“To investigate whether the associations between serum total testosterone and bioavailable testosterone and incident falls were mediated by lean mass or physical performance, we added physical performance variables (grip strength, timed chair stand, walking speed, and narrow walk) and total body lean mass to the basic models,” noted Vandenput and colleagues in the Journal of Bone and Mineral Research.

Total and bioavailable testosterone and fall risk remained significantly inversely associated after adjustment for each physical performance variable individually. Furthermore, only a slight attenuation was noted after adjusting for all variables and lean mass simultaneously. Adjustment for fall-related comorbidities such as alcohol use, medications, and prevalent diseases did not significantly alter the associations between serum total or bioavailable testosterone and falls. Importantly, the multivariate models did not reveal significant associations of estradiol, bioavailable estradiol, or sex hormone-binding globulin with falls.


In this study, Vandenput and colleagues describe an increased fall risk in participants with low total or bioavailable serum testosterone levels. “The association between both total testosterone or bioavailable testosterone and falls was at least partly independent of physical performance and lean mass, and remained relatively unchanged over time,” Vandenput and colleagues wrote in the Journal of Bone and Mineral Research.

Study limitations include the fact that fall events were based on self-report, and results were based on single serum sex steroid measurements. Furthermore, the bioavailable sex steroid levels were calculated and not measured directly, which is the preferred method.

“Both falls and bone strength parameters are independent predictors of fracture risk in older men,” noted the authors in the Journal of Bone and Mineral Research. “With the present data, we propose that low serum testosterone influences fracture risk via an increased risk of falls, whereas low estradiol might increase fracture risk mainly through reduced bone strength,” the authors concluded.

Share the knowledge