Acute kidney injury

Acute kidney injury (AKI) is the sudden reversible renal dysfunction, frequent especially in intensive care unit (ICU) patients, and typically multifactorial. Despite the progress in critical care medicine, AKI incidence has increased and the associated mortality rate has remained static and high. AKI is present in approximately 5% of all hospital and 35% of ICU admissions. AKI considerably increases morbidity, mortality and heath care expenditures, independently of other comorbidity. This even applies to ‘moderate’ stages of AKI.

Currently no effective therapy of AKI is available and efforts focused on early primary and secondary prevention, and early initiation of renal replacement therapy (RRT) for tertiary prevention. Rapid administration of “early goal directed therapy” and antimicrobial agents decrease the AKI rate in sepsis. The same applies to haemodynamic stabilization in shock, volume correction in hypovolaemia and measures in rhabdomyolysis. AKI needs to be detected as early as possible for prevention to be most effectively.

The recent AKIN definitions of AKI incorporate serum creatinine and urine output as markers to diagnose and stage AKI. However, these markers do not permit early detection of AKI. Recently, serum and urinary markers such as cystatin C, interleukin-18 (IL-18), kidney injury molecule-1 (KIM-1), and neutrophil gelatinase-associated lipocalin (NGAL) performed well in pilot trials with small, homogeneous cohorts as detection and prognostic markers of AKI. Often, these promising results could not be confirmed in larger multicenter trials. As AKI is multifactorial and heterogeneous in origin, it seems unlikely that one singular marker but rather a biomarker panel will be required to detect AKI early and to predict its outcome.

Mosaiques, together with the Department of Nephrology of the University Duisburg-Essen, identified urinary peptide markers predictive for AKI in urine samples obtained from indwelling bladder catheters of ICU patients who later developed AKI (maximum 5 days prior AKI) defined by a serum creatinine increase ≥ 50% in ≤ 48 hours or maintained normal kidney function. The statistically most significant markers were combined and validated on a blinded set of ICU patient samples.

A combination of 20 peptides allowed classification of a blinded test set of ICU patient samples (n=20) with 89% sensitivity and 82% specificity. In order to evaluate general applicability, the urinary proteomic model was further applied to the classification of single-void urine samples from hematopoietic stem cell transplanted (HSCT) leukemia patients of which 13 developed AKI after transplantation and 18 did not. Sensitivity and specificity values in this validation set were 94 and 82 %, respectively. Compared to the proteomic model, ROC curve analysis revealed poorer classification accuracy of cystatin C, IL-18, KIM-1 and NGAL with the respective AUC values being outside the statistical significant range.



In a subsequent study the diagnostic accuracy of the AKI predictive proteomic model was evaluated in patients after cardiac surgery from the Department of Internal Medicine, Nephrology and Hypertension of the Saarland University Medical Centre as another etiologic AKI group. One hundred and ten indwelling bladder catheter samples drawn from patients directly after cardiac surgery were analyzed in this study. Fifty-nine patients (54%) out of the study population developed AKI (R: 42.4%; I: 23.7%, F: 33.9%). Application of the AKI predictive proteomic model demonstrated an AUC in ROC analysis of 0.81. Compared to the proteomic model, ROC curve analysis revealed poorer classification accuracy of KIM-1 and NGAL with the respective AUC values of 0.57 and 0.63 being again statistically insignificant.



In contrast to the single AKI biomarkers cystatin C, IL-18, KIM-1 and NGAL, the proteomic marker pattern allowed accurate detection of AKI as early as 5 days in advance of serum creatinine irrespective of the patient population in whom AKI occurred.



Piedrafita A, Siwy J, Klein J, et al. A universal predictive and mechanistic urinary peptide signature in acute kidney injury. Crit Care. 2022; 26(1):344.

Metzger J, Mullen W, Husi H, et al. Acute kidney injury prediction in cardiac surgery patients by a urinary peptide pattern: a case-control validation study. Crit Care 2016; 20(1):157.

Metzger J, Kirsch T, Schiffer E, et al. Urinary excretion of twenty peptides forms an early and accurate diagnostic pattern of acute kidney injury. Kidney Int 2010; 78(12):1252-62.