University of Michigan
Department of Pediatrics
Evidence-Based Pediatrics Web Site

Question

• In children with sepsis, does steroid therapy improve length of stay and mortality?

Clinical Bottom Lines

1. When adjusting for the clinical acuity of the patient as measured by PRISM II score, urinary IL-18 (UIL-18) has a weak independent predictive value (OR 3.7 unadjusted v. 1.3 adjusted) for acute kidney injury (AKI) or mortality (OR 1.3) in critically ill children.  The PPV (20-27%) is too low to be useful for predicting AKI, and the NPV (80-82%) is not sufficiently high enough to reliably be used to rule out AKI.
2. When adjusting for PRISM II score, UIL-18 does have a moderate, statistically significant ability to predict the severity of AKI - PPV 80%.
3. In a subgroup of non-septic patients, there is a strong association between UIL-18 and acute kidney injury (OR 5.23, 95%CI 1.61-16.84) for AKI an its severity.  However, the article did not evaluate PPV or NPV for this subgroup.
4. At this time, UIL-18 is not well established enough as a biomarker for AKI to inform clinical decision making in critically-ill children.   

Summary of Key Evidence

1. Study was designed to address the ability of UIL-18 to predict onset of AKI, severity of AKI and mortality.  Subgroup analysis of non-septic patients was also performed
2. Enrollment criteria: age 1 month - 21 years, admission to PICU, mechanical ventilation and placement of urinary catheter.  Exclusion criteria: ESRD, anuria, or renal transplant.
3. 150 patients were enrolled and 137 had urine samples available for analysis.
4. Peak, first, and day-of AKI onset UIL-18 samples were used for mortality prediction, AKI prediction and AKI severity analysis respectively. 

Additional Comments

• The same group of patients in the above study were used to establish and validate the pRIFLE classification scheme in another article⁴: risk=CrCl 25%, injury=CrCl 50%; failure CrCl 75% or < 35mL/min/1.73m<; loss > 4wks duration; end stage > 3 months duration.²
• The same group was also used to examine the role of urinary NGAL and AKI, which found that it may be a good predictor of the onset of AKI, but not the severity. Although they examined the relationship with PRISM II scores, they did not adjust for them in their analysis, and so the confounding nature of illness severity could not be reliably assessed.³
• It is more likely that a panel of serum and urine tests, including S/U-NGAL, UIL-18, serum cystatin C, and UKIM-1, which are sequentially expressed at various stages of AKI, will emerge as useful for timing the initial insult, assessing the duration/severity, and differentiating between the types of etiologies of AKI, but we are still a ways off before this enters the realm of clinical usefulness.²

Citation

1. Washburn KK, et al. Urinary interleukin-18 is an acute kidney injury biomarker in critically ill children. Nephrol Dial Transplant 2008;23:566-72.
2. Nguyen MT, et al. Biomarkers for early detection of acute kidney injury. Pediatr Nephrol 2007.
3. Zapitelli M, et al. Urine neutrophil gelatinase-associated lipocalin is an early marker of acute kidney injury in critically ill children: a prospective cohort study. Critical Care. 2007.
4. Akcan-Arikan A, et al. Modified RIFLE criteria in critically ill children with acute kidney injury. Kidney International. 2007;71:1028-35.

CAT Author: Krishna Rao, MD

CAT Appraisers: James Gurney, PhD

Date appraised: March 19, 2008