Acute kidney injury (AKI) is one of the most frequent and critical complications in patients admitted to intensive care units (ICUs). AKI consists of a rapid decline in kidney function, which can develop within a few hours or days. It is often asymptomatic or presents with non-specific symptoms, making timely diagnosis particularly challenging, especially in patients with comorbidities, where clinical signs are often subtle or masked by other conditions.
Epidemiological data from multicenter studies confirm that over 50% of ICU patients develop AKI. This condition is associated with increased mortality, a higher incidence of multi-organ complications, and a significant extension of hospital stay (Hoste et al., 2015). These findings highlight the importance of frequent clinical evaluations and careful monitoring of renal function to detect even small but clinically relevant changes.
The diagnosis of AKI is primarily based on the 2012 KDIGO criteria. These define AKI as an increase in serum creatinine of at least 0.3 mg/dL within 48 hours, or a reduction in urine output to less than 0.5 mL/kg/h for at least six hours. However, serum creatinine has important diagnostic limitations. Its fluctuations can occur late and may not immediately reflect actual functional damage, which can exceed 50% before being detected (Waikar et al., 2009).
These limitations have led to the exploration of more sensitive and timely approaches, such as novel biomarkers, predictive algorithms, and advanced decision-support tools. Among the primary causes of AKI in intensive care are sepsis and the use of nephrotoxic drugs. Sepsis is implicated in approximately 50% of cases among critically ill patients, significantly contributing to clinical severity and worsening prognosis (Takeuchi et al., 2025). The use of nephrotoxic drug is also a significant additional risk factor (Ehrmann et al., 2019).
In this complex clinical context, artificial intelligence is emerging as a promising and innovative tool for managing AKI. AI-based predictive systems capable of analyzing large sets of clinical data in real time can anticipate the risk of AKI even before clear clinical manifestations appear, enabling timely and targeted interventions.
A concrete example of this evolution is the U-Care Renal Platform, a technological solution developed to promptly identify the risk of AKI in critically ill patients. A multicenter retrospective study published in PLOS ONE validated this platform in over 16,000 ICU patients across Italy, the Netherlands, and the United States. The predictive model achieved an area under the ROC curve (auROC) of 0.884 in the development cohort. In the Italian external validation, the value was even higher (0.911). Furthermore, the system anticipated stage 2/3 AKI diagnosis by approximately 14 hours compared to traditional clinical methods (Alfieri et al., 2023). This time advantage may be crucial for preventing disease progression, reducing the need for dialysis, minimizing complications, and lowering healthcare costs.
The U-Care Renal Platform also integrates easily with electronic health record systems, ensuring continuous, consistent monitoring. This diagnostic standardization promotes more equitable and coordinated care, facilitating the shared management of critically ill patients.
The combination of advanced monitoring, traditional biomarkers, and innovative predictive tools is shaping a new paradigm in AKI management. This proactive approach significantly improves clinical outcomes and enables intensive care medicine to evolve toward a more personalized model, focused on anticipating risk rather than reacting to advanced-stage damage.
Discover how the U-Care Renal Platform can help your ICU prevent AKI promptly and effectively: visit our website or contact us for a personalized demo.