Clinical data-driven classification of pre-frailty reveals sex-specific patterns – Data from the Berlin Aging Study II (BASE-II)

Abstract

Frailty is a geriatric condition with multidimensional consequences that strongly affect older adults’ quality of life. The lack of a universal standard to describe, diagnose, and treat frailty further complicates this situation. Nowadays, multitudinous frailty assessment tools are applied depending on the regional and clinical context, adding complexity by increasing heterogeneity in the definition and characterization of frailty. Better insights into the causes and pathophysiology of frailty and its early stages are required to establish strong and accurately tailored treatment rationales for frail patients. We analysed participants aged 60 and above using cross-sectional biochemical and survey data from the Berlin Aging Study II (BASE-II, N = 1512, pre-frail=470, frail=14), applying machine-learning techniques to investigate determinants of physical frailty measured by Fried et al.’s 5-item frailty phenotype. Our findings highlight new prognostic sex-specific biomarkers of pre-frailty (the early stage of frailty) with possible clinical applications, enriching the current sex-agnostic diagnostic scores with easy monitorable physical and physiological characteristics. Low appendicular lean mass and high fat composition in men, or vitamin D deficiency and high white blood cell counts in women, emerged as strong indicators of the respective pre-frailty profiles. Because the number of fully frail individuals was extremely small (n = 14, <1 %), our findings should be interpreted as reflecting predictors of pre-frailty, not of frailty itself. We conclude that understanding the development of frailty remains a complex challenge, and that sex-specific differences must be considered by clinical geriatricians and researchers.