Long-Term Consequences of Acute Kidney Injury in Children

AKI has become a common complication in hospitalized children (1,2). It has garnered much attention of late because of its association with higher mortality, longer hospital and intensive care unit (ICU) lengths of stay, and prolonged mechanical ventilation courses (1). Although AKI has traditionally been thought to be a self-limited phenomenon, recent data in adults and children suggest that it is also associated with long-term chronic morbidity (3–7). In adults, AKI has been linked with higher risk for hypertension, cardiovascular disease, and CKD (3,4,8); in children, AKI has been associated with CKD, proteinuria, and hypertension (5,6). Despite this, it has been challenging to establish a definitive causal link between AKI and the subsequent development of chronic sequelae. This has been particularly challenging in children (7).

In this issue of the Clinical Journal of the American Society of Nephrology, Hessey et al. have taken a unique investigative approach to the subject (9). Their study is a retrospective analysis of 2041 children (mean age upon admission, 6.5±5.8 years) admitted to two tertiary care ICUs based in Montreal, Canada. The prevalence of AKI among these children was 22%. This is consistent with the recently published Assessment of Worldwide AKI, Renal angina, and Epidemiology (AWARE) study, which found an AKI prevalence of 27% across a multicenter, multinational cohort of children receiving critical care, indicating that their findings are likely to be externally valid (1). Hessey et al. defined AKI according to the Kidney Diseases: Improving Global Outcomes (KDIGO) consensus AKI guidelines, which is a tremendous strength (10). Historically, pediatric AKI outcomes research has been plagued by inconsistent application of diverse definitions, making extrapolations and comparisons difficult. At this point, the use of a consensus AKI definition should be considered a prerequisite for studies such as this (7). Additionally, the authors excluded patients undergoing cardiac surgery. As they describe, these children have a unique AKI physiology and healthcare utilization pattern (9); this decision makes it easier to extrapolate their findings to the general AKI survivor cohort.

The most unique aspect of their study, however, is their choice of outcome. Rather than focus on a single sequela or specific disease process, they decided to assess longitudinal healthcare utilization. In many ways, this represents an excellent universal and general proxy for chronic morbidity; patients with a greater disease burden are more likely to require medical care and will have higher resource utilization rates. Their findings are impressive and support the prevailing belief that AKI is associated with long-term medical sequelae. Among the cohort of children who were admitted to the ICU and survived, those who developed AKI were more likely to be hospitalized again during the 5 years after discharge from their index admission. This remained true even after adjusting for demographic, medical, and social differences/confounders, which is important because some have suggested that AKI is merely a proxy for severity of illness and underlying disease burden. The fact that the adjusted risk of rehospitalization remained higher should at least mitigate concerns about this relationship. The study also found that 5 years after discharge, children who developed AKI had more physician visits than those who did not experience AKI. This was true even if when visits to a nephrologist were excluded which suggests that the higher resource utilization seen in AKI survivors is not limited to the management of kidney disease. Again, the increased risk remained after adjusting for potential confounders and intergroup differences.

Although not the primary focus of their analysis, the study also provides interesting insight into follow-up care for children who experience AKI. Although definitive recommendations regarding ideal post-AKI follow-up are limited, the KIDGO guidelines do suggest that AKI survivors be re-evaluated 3 months after the onset of AKI to assess for resolution and/or the development of CKD (10). The study by Hessey et al. found that children who experienced AKI were more likely than those without AKI to see a nephrologist, which is at first reassuring. However, a more in-depth analysis reveals the dearth of follow-up in this cohort; fewer than one in five AKI survivors saw a nephrologist within a year of discharge and only one in four saw one within 5 years. Although more research is necessary to determine which patients are most likely to benefit from longitudinal care, studies such as this should highlight the relationship between AKI events and the subsequent need for long-term medical attention.

Despite these strengths, the retrospective nature of the study, concern for ascertainment and provider bias, and reliance upon administrative outcome data limit the authors ability to identify a causal relationship between AKI and an increased need for longitudinal medical care. The authors conclude that AKI is either associated with long-term morbidity or, at a minimum, is a marker for it; they accurately identify the challenge inherent in distinguishing between a true causal effect and an association. Regardless, the authors’ study adds to the growing body of literature that raises concern regarding the development of long-term sequelae in AKI survivors. Additionally, the analysis possesses several important characteristics that should serve as exemplars for future work. First is the use of a consensus AKI definition. KDIGO is an ideal choice because it can be used in both adults and children and it represents a harmonization of previously proposed consensus AKI criteria. Second is the need to include non-AKI patients as comparators; although many studies have demonstrated a high prevalence of complications among AKI survivors, to accurately assess the independent effect of AKI, it is of paramount importance to include a non-AKI comparator cohort. Finally, it is imperative that outcomes are standardized and comparable. In this study, although administrative data were utilized, rehospitalizations and physician visits were rigorously defined. Likewise, studies which examine specific diseases, such as CKD, cardiovascular disease, proteinuria, or hypertension, should use standard consensus definitions (7). It is the adherence to consistent approaches such as these that will allow us to definitively characterize the relationship between AKI and chronic morbidity.

Disclosures

None.