The Ever-Changing Paradigm of Acute Kidney Injury Definition: Does Duration and Trajectory of Acute Kidney Injury Add Another Dimension?*

The field of pediatric critical care nephrology has undergone a clinical and research revolution over the past 2 decades which has changed our "nihilistic" perspective to that of realism. Prior to 2004, more than 30 definitions were used to define acute kidney injury (AKI) which resulted in different prevalence and outcomes in varied settings dependent on the definition used. Great strides have been made in our ability to diagnose and stage AKI using a standardized definition which culminated in the establishment of the Kidney Disease Improving Global Outcomes (KDIGO) AKI criteria (1). Application of a standardized definition and use of non-AKI comparison groups in multicenter AKI studies like Assessment of Worldwide AKI Renal Angina Index and Epidemiology has allowed for better understanding of the association of AKI with short-term outcomes (2). AKI in these critically sick children was independently associated with increased mortality (which increased with increasing severity of AKI), length of PICU stay, and duration of mechanical ventilation. Based on these high-quality studies, it is now obvious that patients are not dying "with" AKI but are dying "from" AKI. In this issue of Pediatric Critical Care Medicine, Alobaidi et al (3) discuss the importance of duration of AKI after admission to PICU as an important factor in determining the prognosis of critically sick children. The authors classified children with AKI persisting less than 48 hours as "transient," whereas AKI that lasted more than 48 hours after PICU admission was termed as "persistent." They found that mortality was associated with AKI status—lowest mortality in those with no AKI (1.8%) followed by those with transient AKI (5.4%) and highest with persistent AKI (17.6%). The authors concluded that transient and persistent AKI have different clinical phenotypes associated with different outcomes. Independent of the degree of severity of AKI, persistent AKI was associated with worse outcomes. Even when AKI was transient, the adverse outcomes were worse than in children without AKI. Most children in the study who were later labeled as having "persistent AKI" actually developed AKI within 24 hours of PICU admission. This information conveys a very clear message that identification of AKI or those at risk of AKI (renal angina index, novel biomarkers) in the first 24 hours of admission to PICU is crucial to initiate mitigating strategies such as preventing fluid overload and avoiding or minimizing nephrotoxic drugs (4). Secondary outcomes, for example, mechanical ventilation, vasoactive support, and hospital free-days at 28 days, were worse with increased duration of AKI. Do these findings indicate a paradigm shift in the way we define and stage AKI currently? In other words, do we need to add duration of AKI to the existing KDIGO criteria to define and stage AKI? To address these questions, recommendations from the AKI, and renal recovery: consensus report of the acute disease quality initiative suggests that AKI definition and staging can be further refined by incorporating both the magnitude of rise of creatinine and the duration of AKI after admission to the PICU (5). Various population studies indicate that duration of AKI may be an additional or an independent marker of poor prognosis in patients with AKI across various clinical settings. Han et al (6) concluded that the impact of AKI duration (divided into three tertiles of increasing duration) was higher than that of AKI staging using the KDIGO guidelines. When severity of AKI staging and duration were combined, the impact remained significantly higher than that using the KDIGO stage alone. Similar associations between mortality and duration of AKI were demonstrated in critically sick patients with sepsis-associated acute respiratory distress syndrome who developed AKI (7). Longer duration of AKI within each KDIGO stage of AKI was significantly associated with higher graded rate of mortality in patients undergoing both cardiac and noncardiac surgery (8,9). Due to the retrospective nature of their study, Alobaidi et al (3) do not comment about the evolution of AKI during the PICU stay. Some patients with higher stage AKI could have had full recovery or partial recovery to lower stage AKI or lower stage AKI could have progressed to severe AKI over the course of PICU stay. Sanchez-Pinto et al (10) demonstrated that new or progressing AKI after admission to PICU had higher mortality than those with unchanging or nonprogressive AKI. Patients who retained the same degree of AKI during PICU admission had higher mortality than those patients whose AKI either resolved or decreased in severity. This can have important implications as resolution or nonrecovery of AKI can have different short-, medium-, and long-term outcomes as shown in the study by Kellum et al (11) who identified five distinct recovery phenotypes on the basis of the clinical course over the first week after AKI manifestation—sustained early reversal, late reversal after day 7 of AKI, early reversal with one or more relapses but with ultimate recovery, relapsing AKI without recovery, and finally no reversal at all. All these phenotypes had different outcomes with best outcomes for early sustained reversal of AKI and worst outcomes for nonrecovery. Trajectory of AKI after the episode of AKI in PICU can distinguish the risk of long-term adverse outcomes with greater proportion of nonresolving AKI at risk of complications (12). The authors have limited their analysis to stage 2/3 AKI and compared with patients with no AKI, this might lead to a misconception that stage-1 AKI carries no significance. In fact, stage-1 AKI might lead to sustained exposure with significant loss of nephron mass. Although patients with stage-1 AKI had a higher proportion of transient AKI, they were not followed up to see if they progressed to severe AKI or had sustained pattern of recovery. The new report by Alobaidi et al (3) also does not give us information about whether AKI relapsed after initial recovery. Relapses of AKI after apparent recovery are common and associated with many-fold increase in long-term mortality as compared to early sustained recovery. Therefore, we cannot label a patient who recovers for 24 hours from AKI as "permanently" recovered as most relapses occur after 48–72 hours of apparent recovery. This can have clinical bedside significance for reinitiation of nephrotoxic drugs and fluid management which were altered due to obvious AKI. Similarly, there are patients who might not recover from AKI at PICU or hospital discharge but can have late recovery from AKI. Although the authors describe the outcomes up till PICU discharge (based on serum creatinine and urine output), renal recovery or follow-up after PICU discharge was not evaluated which precludes us from any meaningful conclusion of the medium-long–term outcomes of AKI in critically ill children especially relapse, late recovery, or persisting nonrecovery. AKI could recover completely during this time (late recovery) or can recover and relapse (relapsing AKI). It may, however, not recover at all entering into a spectrum of chronic kidney disease. Different patterns of AKI recovery will be extremely informative to risk stratify patients and organize appropriate investigations (serum creatinine, proteinuria, blood pressure measurements) and follow-up consultation by a clinical nephrologist for those who are at high risk for adverse long-term outcomes after an episode of AKI in the PICU (13,14). Unfortunately, there are no definitive treatment strategies for the management of AKI. This deficit could be because various trials have coupled transient AKI with persistent AKI, so the treatment signal gets obscured. Since persistent or nonresolving AKI seems to be the risk factor for long-term adverse outcomes, future therapeutic clinical trials might target this particular group. Can clinical risk factors or biomarkers at admission or during the PICU stay predict the development of nonresolving AKI to target patient-specific diagnostic and therapeutic interventions? There are current gaps in knowledge in long-term outcomes in children, and future research is needed to develop evidence-based guidelines for follow-up of children who survive their episode of AKI in PICU. This report by Alobaidi et al (3) of a large, multicenter cohort of critically sick children prompts the clinician to think beyond the usual parameters of serum creatinine and urine output in prognosticating AKI by including the duration of AKI which might serve as a trigger to redefine AKI in the future. Longer duration of AKI has a distinct clinical phenotype which might predict complex hospital course with increased mortality. This might help us to risk stratify patients and initiate early interventions which could potentially modify the course and outcome in children with AKI (3). AKI research consortia like the acute dialysis quality initiative (now with dedicated pediatric component) encourage researchers to continue their quest to finetune definition and management of AKI, develop better risk stratification models, and further understand risk factors for long-term adverse renal and nonrenal outcomes.