Acute kidney injury (AKI) is a complex syndrome typically classified into strict categories. Alternatively, it may be more accurate to consider it as an intermediate event between an initiating cause and its outcome. Therefore, we investigated the burden of clinical scenarios associated with dialysis-requiring AKI (AKI-D) using latent class analysis (LCA) and examined the etiological spectrum and clinical phenotypes across different life stages. We analyzed 17,158 AKI-D patients from 170 medical facilities in Rio de Janeiro, Brazil (2002-2012). Utilizing survival curves and mixed-effects Cox regression for survival estimation, LCA classified patients based on clinical characteristics and outcomes, focusing on etiological variation over the human lifespan. The median age was 75 (IQR 59-83). Infections were the most common cause (44.2%), particularly community-acquired pneumonia (23.8%). Cardiovascular issues, especially ischemic heart disease (9.0%) and acute heart failure (8.1%), were also significant. LCA identified four distinct patient classes with varying clinical and outcome profiles. Class 1 patients were younger (median age 66), predominantly male, with lower ICU admission and higher rates of community-acquired AKI (60.8%). They had the lowest mortality (39.5%) and highest recovery rates. Class 2 had intermediate mortality (67.4%) and the highest comorbidity burden (mean Charlson score: 3.39). Classes 3 and 4 had the highest mortality (82.8% and 78.6%), requiring more mechanical ventilation and vasopressor use. Class 3 had a high prevalence of sepsis (92.7%) with lower comorbidities, while Class 4 had high chronic heart disease (76.3%) and perfusion factors (79.4%). Despite high mortality, Class 3 recovered better than Class 2 and 4. Survival analyses revealed diverse outcomes across etiological groups, with liver-related conditions being the most severe. This study highlights the complexity of AKI and the utility of LCA in revealing its clinical heterogeneity. It underscores distinct etiological trends across ages, suggesting future research should integrate clinical profiles with advanced diagnostics to understand AKI's clinical and molecular phenotypes throughout life.
Read full abstract