state of renal function, GFR and CrCl cannot be measured reliably due to the rapidly changing parameters, especially in AKI critically ill patients. In this case, the so-called kinetic GFR has been proposed [22]. Some interesting initial applications of this new estimation have been provided in clinical practice even if further validation is currently awaited [23]. Moreover, the capacity of the kidney to respond to physiological demands – the so-called RFR – is another aspect of renal function that requires careful attention and precise monitoring [8]. As a matter of fact, GFR is not constant, but depending on physiological conditions, can change, being dynamic; and like the heart during exercise, GFR can increase to a certain maximum level [8]. The ability of the kidney to increase the GFR under specific stimuli (protein load) is an important property of renal physiology. The KDIGO criteria based on isolated SCr measurements cannot be a valuable system for the assessment of RFR that, in turn, could provide a more sensitive way of assessing renal function [8]. Further research on RFR is currently considered a priority on AKI agenda in order to verify its diagnostic and prognostic value in critically ill patients [1].
Additional Limitations of KDIGO Criteria for AKI Diagnosis
• AKI is a very heterogeneous syndrome caused by many different pathologies isolated or combined such as hypovolemia, hypoperfusion, hypotension, ischemia, inflammation, sepsis, nephrotoxic drugs, and contrast media. Then, different specific “endotypes” including hepatorenal, cardiorenal, nephrotoxic, and sepsis-associated syndromes have been identified and these conditions all have specific and dedicated pathophysiologies and, sometimes, treatments [24]. In this context, different AKI identified by the same KDIGO stage may have significantly different clinical courses and therapeutic approaches.
• There is no gold standard to validate the KDIGO definition or other AKI markers: even if it was not conceived with this initial aim, the KDIGO definition and criteria has been validated by the association of AKI diagnosis/severity with mortality and other hard outcomes [7, 25, 26].
• Finally, the KDIGO definition of AKI is weak for the evaluation of recovery from renal injury. Some recent studies considered renal recovery as achieved when AKI, defined by KDIGO, is definitely reversed [26, 27]. However, as previously focused, conditions as sarcopenia and/or weight gain due to fluid overload, may cause an overestimation of renal recovery.
An Outlook to the Future: AKI Biomarkers and Subclinical AKI
In order to overcome some of the limitations of SCr and UO, a variety of functional and damage biomarkers have been identified and validated in humans [28, 29] as tools to diagnose AKI earlier than traditional tests and prognosticate outcomes. Despite their potential to detect the development of AKI prior to an SCr rise (subclinical AKI), no clear benefit on patients’ outcome has been demonstrated so far. Among them, Cystatin C, Neutrophil Gelatinase-Associated Lipocalin, and markers of cell cycle arrest (tissue inhibitor of metalloproteinase-2 and insulin-like growth factor binding protein-7) are now commercially available and their role in future AKI diagnosis and classification systems still has to be evaluated. For example, Cystatin C is less affected by muscle mass and may outperform SCr for the detection of decreased GFR in critically ill patients [21]. The association of (tissue inhibitor of metalloproteinase-2) and (insulin-like growth factor binding protein-7; Nephrocheck, Astute Medical, San Diego, CA, USA) showed good performance in identifying patients at risk of AKI after cardiac surgery with a threshold value >0.3. These patients have taken advantage from the close application of “KDIGO bundles” consisting of optimization of volume status and hemodynamics, avoidance of nephrotoxic drugs, and preventing hyperglycemia [30]. Definitely, positive clues from the literature are coming up but whether the use of biomarkers may impact clinical decisions and contribute to better prevention and management of AKI remains to be proven.
AKI Definition and Classification in Children and Neonates
In 2007, the RIFLE criteria were adapted for use in children (pediatric RIFLE or pRIFLE) [31] (Table 1). In that pivotal study, the essential modifications to the original RIFLE criteria were the introduction of percentage decrease of estimated CrCl based on Schwartz formula (instead of change of absolute creatinine levels/GFR) and slightly modified UO criteria in order to meet the pediatric low changes of creatinine levels and different UO flows. Despite the relatively low number of enrolled children (n = 150), extensive information was already described by this observation: the majority of pediatric AKI (pAKI) episodes occurred in the first ICU admission week and the incidence was very high (up to 80%). About one third of patients attained the pAKI diagnosis by both UO and SCr, whereas the remaining patients only achieved 1 of the 2 criteria (including 18% by UO criteria and 47% by SCr). PAKI was associated with the length of stay in the hospital and ICU. Worse pRIFLE class I or F, after adjustment for severity of illness score, was associated with a significantly increased risk of mortality. All these results were further validated in several studies. In 2015, data from more than 14,000 hospitalized children were analyzed [32] and pAKI diagnosis according to pRIFLE, AKIN, and KDIGO criteria were reported. AKI incidences across the cohort according to pRIFLE, AKIN, and KDIGO were 51.1%, 37.3%, and 40.3%: pRIFLE showed the highest sensitivity (especially in detection of class R patients) and AKIN was the most specific in the most severe class. Mortality was increased in AKI children according to all definitions (pRIFLE, 2.3%; AKIN, 2.7%; KDIGO, 2.5%; versus no AKI [0.8–1.0%]). Even if the clinical relevance of definitions differences in diagnosing and staging of AKI is unclear, there currently is a general consensus to apply KDIGO classification for pAKI. Under these premises, the recent Assessment of Worldwide AKI and Renal angina Epidemiology trial was conducted on more than 4,000 patients (see chapter on pAKI) [33].
A slightly different pathway has been covered by the neonatal AKI (nAKI) definition. In this peculiar population, renal physiology may significantly vary with respect to older patients and the application of pRIFLE or KDIGO may lead to inadequate staging. Problems in the diagnosis of AKI in neonatal ICUs are essentially related to the difficulty on achieving a daily creatinine measurement (due to blood sparing), the uncertainty on normal creatinine levels (mostly influenced by maternal creatinine), variable GFR (secondarily to “immature” kidney function, especially in preterm babies) and, consequently, the absence of baseline/reference values [34]. Furthermore, the need for exactly timed diaper weight is needed to measure UO in case the bladder catheter is not in place. Finally, the pRIFLE urinary criteria appear significantly restrictive because newborn and premature urine flow has to be considered normal only when higher than 1.5 mL/kg/h. The reason for this finding might be identified in the increased neonatal body water content
