Unpacking KDIGO Guidelines: Prioritizing and Applying Exposures and Susceptibilities for AKI in Clinical Practice

<b>Background/Objectives:</b> Acute kidney injury (AKI) is a significant global health issue with a high morbidity and mortality. The Kidney Disease: Improving Global Outcomes (KDIGO) guidelines identify various exposures and susceptibilities as risk factors for AKI. However, the predictive significance of these factors in heterogeneous emergency department (ED) populations remains unclear. We hypothesized that assessing KDIGO-listed exposures and susceptibilities for AKI, alone and in combination, would provide an insight into their predictive value for AKI. Furthermore, we investigated whether adding biomarkers, plasma neutrophil gelatinase-associated lipocalin (pNGAL) and C-reactive protein (CRP), could enhance AKI risk prediction. <b>Methods:</b> Data were analyzed from the prospective longitudinal "NGAL study" conducted at North Zealand University Hospital in Denmark. A total of 344 ED patients were included, with AKI diagnosed using KDIGO's creatinine-based criteria. Patient data, including medical history, exposures, and susceptibilities, were extracted and analyzed. Predictive performance was evaluated using a receiver operating characteristic (ROC) analysis on individual and combined risk factors. Additional models incorporated pNGAL and CRP to assess their impact on prediction accuracy. <b>Results:</b> Individual exposures and susceptibilities showed a poor predictive performance, with nephrotoxic drugs and advanced age demonstrating the highest sensitivity but a low positive predictive value (PPV). Combining multiple risk factors improved AKI prediction, with models clustering into those optimizing sensitivity or PPV. The inclusion of pNGAL significantly enhanced predictive performance, achieving the highest combined sensitivity and PPV. Although less than pNGAL, CRP also improved prediction, while requiring fewer variables than pNGAL-inclusive models. <b>Conclusions:</b> No individual KDIGO-listed exposure or susceptibility could reliably predict AKI in the ED setting. Combining multiple exposures and susceptibilities improved the predictive accuracy, but the models excelled either at screening or confirmation, not both. The addition of pNGAL and CRP significantly enhanced AKI prediction, emphasizing the need for biomarker integration in risk stratification models. These findings highlight the limitations of clinical parameters alone and underscore the importance of a multifaceted approach to AKI risk assessment.