Clustering analysis of multi-site electronic health records reveals distinct subphenotypes in stage-1 acute kidney injury.

Abstract

Background: Acute Kidney Injury (AKI) can adversely affect multiple organ systems, including the heart, brain, and immune system. Stage 1 AKI (AKI-1), although mild in clinical presentation, constitutes a substantial subset of AKI patients with heterogeneous outcomes, warranting further investigation into its subphenotypes.

Methods: We performed clustering analysis on seven-day serum creatinine (SCr) trajectories preceding AKI-1 onset in 53,565 AKI-1 patients (aged 18-89 years; 55.57% male) across eight academic hospitals. Each AKI-1 patient was matched to a non-AKI counterpart to evaluate how different AKI-1 subphenotypes influence clinical indicators and outcomes.

Results: Three distinct AKI-1 subphenotypes are identified. Patients in Subphenotype C (n = 5,378; 10.0%) exhibit a higher proportion of abnormal values across clinical indicators compared to those in Subphenotypes A (n = 27,049; 50.5%) and B (n = 21,138; 39.5%). Subphenotype C is associated with significantly higher odds ratios (ORs) for in-hospital, 30-day, and one-year all-cause mortality relative to Subphenotypes A and B. Conversely, Subphenotype B exhibits a higher susceptibility to developing chronic kidney disease (CKD) within one year after discharge following AKI-1, compared to both Subphenotypes A and C, after adjustment for baseline SCr levels. All AKI-1 subphenotypes are associated with significantly elevated risks of all-cause mortality and the need for dialysis or renal replacement therapy (RRT) compared to their respective non-AKI counterparts.

Conclusions: This study reveals substantial heterogeneity in clinical indicators and outcomes within AKI-1. Future research focusing on these subphenotypes may pave the way for more personalized and targeted interventions for patients with AKI-1.