Physical Function Phenotypes in Chronic Kidney Disease: A Hierarchical Cluster Analysis

Background: Poor physical function is common in chronic kidney disease (CKD) and is associated with adverse outcomes; however, patterns of impairment are heterogeneous, and underlying mechanisms remain poorly understood. Traditional single-variable classification approaches may not fully capture this heterogeneity, potentially missing distinct at-risk subgroups. Hypothesis: We hypothesized that pattern-based phenotyping would identify clinically meaningful subgroups with different physiological profiles. Methods: Patients with CKD stage 3-4 (n=72, 62± 9 yr, 78% male, eGFR: 45±13 ml/min/1.73m 2 ) completed a 4-domain physical function assessment (handgrip strength, gait speed, chair-stand time, and static balance), body composition and volume status (extracellular water/total body water (ECW/TBW)) by bioimpedance, and vascular stiffness (pulse pressure, augmentation index (AIx). Hierarchical clustering was applied to z-standardized 4-domain for physical function scores to identify multi-variable patterns. Construct validity was evaluated through Spearman correlations and principal components analysis (PCA). To further characterize phenotypes, the associations between individual physical functional measures and clinical risk factors were assessed. Results: Three phenotype clusters emerged; 1) High-functioning (n=38, 53%, 60± 9 yr, 18% female) demonstrated better performance across all functional domains compared to other clusters (p< 0.05); 2) Lower-limb Mobility Impairment (n=6, 8.3%, 54± 15 yr, 17% female) exhibited markedly slower chair-stand performance (p< 0.05) with preserved gait and balance, relatively to the High-functioning; 3) Multi-domain Impairment (n=28, 39%, 66±6 yr, 29% female) showed reduced performance across all domains (p< 0.05 for all), with the older age (p=0.006), elevated pulse pressure (p=0.024), ECW/TBW (p=0.017) and albumin-to-creatine ratio (p=0.059) and reduced skeletal muscle mass percent (p=0.067), compared to the High-functioning. Functional measures demonstrated distinct constructs (Spearman correlations: all |r|< 0.4). PCA revealed two components explaining 74% of variance: PC1 (47%, general function) and PC2 (27%, mobility vs. stability). Grip strength was associated with age, skeletal muscle mass percent, pulse pressure, AIx, ECW/TBW; gait speed with ECW/TBW (p< 0.001 for all); balance with age (p=0.003) and pulse pressure. Conclusion: Pattern-based phenotyping identified distinct physical function phenotypes in CKD with different clinical profiles. While the High-functioning phenotype maintained functional capacity across all domains, the Multi-domain Impairment phenotype displayed global functional deficits accompanied by volume overload, vascular dysfunction, and low muscle mass. The Lower-limb Mobility Impairment phenotype exhibited domain-specific dysfunction ofimpaired lower-extremity power despite preserved gait speed and balance, representing a dissociation pattern undetectable by single-variable classification. Pattern-based multi-domain assessment provides comprehensive characterization of physical function profiles in CKD and may inform targeted intervention strategies. This abstract was presented at the American Physiology Summit 2026 and is only available in HTML format. There is no downloadable file or PDF version. The Physiology editorial board was not involved in the peer review process.