Is the Timing of Renal Replacement Therapy Initiation Associated With Mortality in Critically Ill Children?

The presence of acute kidney injury (AKI) and fluid overload can adversely affect outcomes in children with critical illness. Continuous renal replacement therapy (CRRT) is one therapeutic modality that can improve outcomes in these patients. However, the trigger when to initiate this therapy is not known. There are studies suggesting that degree of fluid overload may be such a trigger but this does not seem to provide the entire answer. In adults, some studies have suggested that early initiation of CRRT in critically ill patients can improve outcomes compared to late initiation, although there is a paucity of evidence in children. The authors of this study sought to assess the effect timing of CRRT has on mortality in critically ill children.

In this observational retrospective cohort study, the authors studied the medical histories of children admitted to a pediatric intensive care unit (PICU) who received CRRT from 2001 to 2009. After applying exclusion criteria, they analyzed demographic and physiologic information on 190 subjects.  They subsequently divided the subjects into two groups – survivors and nonsurvivors – in hopes of delineating the effect that CRRT timing might have on mortality. In addition, they compared outcomes on those patients who received early (less than or equal to 5 days after PICU admission) versus late (greater than 5 days after PICU admission).

In this study, the authors found that the nonsurvivors were sicker on admission (with higher Pediatric Index of Mortality 2 scores – or PIM), tended to have underlying oncologic diseases, and were more often admitted to the PICU with a diagnosis of sepsis. However, interestingly, the pediatric Risk-Injury-Failure-Loss of Function-End Stage Criteria (pRIFLE), a measure of renal dysfunction, was higher on admission in the survivor group, which also tended to be started on CRRT earlier than the nonsurvivor group. The nonsurvivor group had worse pRIFLE scores upon initiation of CRRT, and tended to be more fluid overloaded than the survivors. Still another difference between the two groups was the reason for CRRT initiation, as the survivor group was started on CRRT secondary to fluid overload primarily while the nonsurvivor group was started on this therapy because of acute renal failure and fluid overload. The magnitude of effect on mortality of both fluid overload status and timing of CRRT was also studied and determined to be independent of each other.

In addition to looking at survivor/nonsurvivor characteristics, the authors compared outcomes and physiologic features between those patients started on early versus late CRRT. The early CRRT patients were noted to be on vasopressors more often, had worse pRIFLE scores on admission, but were less fluid overloaded  and had decreased mortality overall. Again, like in the nonsurvivor group, the late CRRT group had worse pRIFLE scores on initiation of CRRT.

In this work, the authors showed that, like fluid overload, timing of the initiation of CRRT is an independent predictor of mortality in critically ill children. This supports the findings noted in other studies that suggested that temporally defined initiation of CRRT provided better outcomes than biomarker defined initiation. Interestingly, the authors make the observation that the nonsurvivors had better renal function on admission to the PICU, which could suggest that the renal injury suffered by the nonsurvivors, although more lethal than the one seen in the survivors, may not be clearly delineated by pRIFLE. Their suggestion that a search for a more reliable biomarker of AKI prior to starting a prospective trial on the benefits of early CRRT is a good one. Until this study is done, however, the decision on whether to start a critically ill child on CRRT early in the course of a PICU stay will remain a difficult one.

This Concise Critical Appraisal is authored by SCCM member Daniel E. Sloniewsky, MD. Each installment highlights journal articles most relevant to the critical care practitionerDaniel Sloniewsky is an associate professor in the Department of Pediatrics at the Stony Brook Long Island Children’s Hospital in Stony Brook, NY, where he is board certified in pediatrics and pediatric critical care. He completed his fellowship training at Children’s Memorial Hospital and Northwestern University in Chicago. His major interests are in acute pediatric pulmonary disease, transfusion medicine and ethics. He is also actively involved in resident education, Pediatric Advanced Life Support and Pediatric Fundamental Critical Care Support instruction.

eNewsletter Issue: