Interleukin 18 (IL-18) has been proposed as a biomarker for the early detection of acute kidney injury (AKI), but a broad range of its predictive accuracy has been reported. Meta-analysis of diagnostic test studies. Various clinical settings of AKI, including after cardiac surgery, after contrast infusion, in the emergency department, or in the intensive care unit. Prospective studies that investigated the diagnostic accuracy of IL-18 level to predict AKI. Increasing or increased urinary IL-18 excretion. The primary outcome was AKI development, mainly based on serum creatinine level (definition varied across studies). The other outcome was in-hospital mortality. We analyzed data from 23 studies and 7 countries involving 4,512 patients. Of these studies, 18 could be included in the meta-analysis. Across all settings, the diagnostic odds ratio (DOR) for urinary IL-18 level to predict AKI was 4.22 (95% CI, 2.90-6.14), with sensitivity and specificity of 0.58 and 0.75, respectively. The area under the receiver operating characteristic curve (AUROC) of urinary IL-18 level to predict AKI was 0.70 (95% CI, 0.66-0.74). Subgroup analysis showed the DOR/AUROC of urinary IL-18 was 5.32 (95% CI, 2.92-9.70)/0.72 (95% CI, 0.68-0.76) in cardiac surgery patients and 3.65 (95% CI, 1.88-7.10)/0.66 (95% CI, 0.62-0.70) in intensive care unit or coronary care unit patients. After stratification for age, IL-18 level had better diagnostic accuracy in children and adolescents versus adults: 8.12 (95% CI, 3.79-17.41)/0.78 (95% CI, 0.75-0.82) versus 3.31 (95% CI, 2.28-4.80)/0.66 (95% CI, 0.62-0.70). There was no significant difference in predictive performance of urinary IL-18 level among various times. Various clinical settings; different definition of AKI and serum creatinine level as the reference standard test for the diagnosis of AKI. Urinary IL-18 is a useful biomarker of AKI with moderate predictive value across all clinical settings.
Read full abstract