Nutritional-inflammatory, affective, sleep and cognitive correlates of self-reported frailty in maintenance hemodialysis: a cross-sectional study

Background Frailty is common in maintenance hemodialysis (MHD), but the associations of nutritional-inflammatory burden, depressive symptoms, sleep quality and cognition with frailty have rarely been estimated in the same patients, and the instruments used to measure them share item content with frailty screens. We estimated frailty prevalence and the exposure-specific associations of these domains with frailty severity. Methods In this cross-sectional study of 101 prevalent in-center MHD patients, frailty was assessed with the self-reported FRAIL scale. Exposures were the Malnutrition-Inflammation Score (MIS), Beck Depression Inventory (BDI), Pittsburgh Sleep Quality Index (PSQI), Standardized Mini-Mental State Examination (SMMSE), C-reactive protein (CRP), ferritin and albumin. Each exposure was analyzed in a separate proportional-odds model of the three-level frailty category, with an adjustment set derived from a causal diagram (age, sex, education, diabetes and dialysis vintage; plus CRP for ferritin and albumin). Post hoc sensitivity analyses removed item content shared between the exposures and the FRAIL scale. Results Frailty was present in 40 (39.6%) and pre-frailty in 33 (32.7%). Per 1-SD higher exposure, adjusted odds ratios (ORs) for a higher frailty category were 3.46 (1.96–6.11) for MIS, 3.16 (1.83–5.48) for BDI, 2.23 (1.45–3.44) for PSQI and 1.94 (1.27–2.97) for log-transformed CRP. Associations for SMMSE (0.63 (0.37–1.08)), ferritin (1.34 (0.90–2.01)) and albumin (0.90 (0.55–1.46)) were imprecise and compatible with no association. Removing overlapping items attenuated but did not abolish the associations: 1.55 (1.05–2.29) for a reduced MIS against a FRAIL score without the weight-loss item, 2.13 (1.35–3.34) for the cognitive-affective BDI subscale against a FRAIL score without the fatigue item, and 1.62 (1.09–2.41) for PSQI without daytime dysfunction. Conclusions Nearly three-quarters of prevalent MHD patients were frail or pre-frail. Nutritional-inflammatory burden, depressive symptoms, poor sleep quality and systemic inflammation were each associated with greater self-reported frailty, and these associations were only partly explained by shared questionnaire content. Because the design is cross-sectional, the direction of these associations cannot be established; longitudinal studies with performance-based frailty measures are needed.