Abstract
Frailty makes many older people vulnerable to external stressors. It predicts mortality, hospitalization and morbidity. Age-related physical decline and increased prevalence of frailty mirror androgen decline in older men. This study assessed the cross-sectional and longitudinal associations of oestradiol, bioavailable oestradiol, testosterone, bioavailabe testosterone and sex-hormone-binding globulin (SHBG) with frailty. A previous cross-sectional study found no association between testosterone and frailty. No longitudinal studies have reported on this association before.
Men (n = 1602) were randomly selected for sex hormone measures from the 5995 participants of the Osteoporotic Fractures in Men (MrOS) Study. Participants provided a previously stored (−70°C) sample from the baseline examination, which was analysed by gas chromatography/mass spectrometry for oestradiol and testosterone. Bioavailable values were derived using Sodegard's mass action equations. SHBG was measured on an Immulite chemiluminescence assay. Incomplete hormone results or medications led to the exclusion of 133 participants. The remaining 1469 had clinical examination (biometric data), whole body DEXA scan (skeletal mass, total body fat) and a self-reported questionnaire (Cardiovascular Health Study modified criteria for frailty). Cross-sectional baseline analysis was undertaken. A second visit (4 years later) assessed participants for frailty. Mortality data were included for the 160 participants who died before visit 2.
In the cross-sectional analysis, men in the lowest quartile of bioavailable testosterone had a 1.4-fold increased odds of greater frailty compared with those in the highest quartile after adjusting for age, body size, health status and medical conditions. In the age-adjusted longitudinal analysis, men in the lowest quartile of bioavailable testosterone had a 1.5-fold increased odds of greater frailty four years later, although adjustment for confounding factors reduced the strength of this association. The association between total testosterone and frailty status was non-significant. Oestradiol, bioavailable oestradiol and SHBG were not associated with frailty status at baseline or follow-up.
The study results indicate that low bioavailable testosterone is related to frailty status. However, a causal association can only be established by a randomized trial. The authors further suggest that frailty should be considered in testosterone supplementation trials.