Abstract
Background: Frailty is characterized by an increased vulnerability and a decline in physiological reserve. Frailty has been previously linked to cortisol concentrations and blunted diurnal cortisol secretion. Our objective was to determine the association of urine steroid metabolome and its diurnal variation with frailty and prefrailty in older adults. Methods: Cross-sectional study of community-dwelling adults ≥ 50 years. Participants with adrenal disorders, end-stage renal or liver disease, on exogenous steroids or drugs affecting steroid metabolism were excluded. All participants completed day and night separate urine collection. Frailty was assessed using a phenotype model (weight loss, exhaustion, grip strength, low physical activity, and slow walking pace). Participants were characterized as frail if they met at least three criteria, pre-frail if they fulfilled one or two criteria, and robust if no criteria were met. Urine samples were analyzed with the liquid-chromatography, high-resolution, accurate-mass mass spectrometry for 25 urine steroid metabolites. Results: Of 119 participants, 60 (50.4%) were women, without sex differences in age or education status. On frailty assessment, 5 (4.2%) participants were frail and 33 (27.7%) were prefrail, with equal sex distribution. Urine steroid metabolome analysis demonstrated 21/25 steroids were higher in men vs women. In an age adjusted model, presence of prefrailty or frailty was associated with a higher ratio of total cortisol metabolites/total androgen metabolites (TCM/TAM) in men (estimate 0.64, P-value= 0.0004), but not in women. In men, after adjusting for age, among cortisol metabolites, lower day to night ratio of 5α-Tetrahydrocortisol (estimate -0.36, P-value= 0.0419) and β-Cortol (estimate -0.35, P-value= 0.0238) were associated with frail or prefrail phenotype. After adjusting for age, higher ratio of TCM/TAM was associated lower gait speed in men (estimate -1.2, P-value= 0.046) and women (estimate -3.9, P-value= 0.012); and lower hand grip strength in men (estimate -0.04, P-value= 0.046) but not in women. Conclusion: We showed that a higher glucocorticoid to androgen ratio and a flattened circadian steroid variation were associated with presence of frail or prefrail phenotype in men. Further studies should examine the role of steroid metabolism and HPA axis impairment, and the associated sex differences, in the functional decline in aging population.