McMahon BA, Galligan M, Redahan L, Martin T, Meaney E, Cotter EJ, Murphy N, Hannon C, Doran P, Marsh B, Nichol A, Murray PT: Biomarker Predictors of Adverse Acute Kidney Injury Outcomes in Critically Ill Patients: The Dublin Acute Biomarker Group Evaluation Study. Am J Nephrol DOI 10.1159/000500231
The use of serum and/or urinary biomarkers for the prediction of the occurrence and/or severity of AKI among hospitalized (mostly intensive care units [ICU]) is of great and growing interest. However, relatively few large, prospective, observational trials have been conducted that examine the specific value of such biomarkers for the risk of AKI.
McMahon and colleagues in the Dublin Acute Biomarker Group Evaluation (DAMAGE) study sought to remedy this knowledge gap. Subjects (n=669) with acute critical illness admitted to an ICU were followed sequentially with a panel of urinary biomarkers (creatinine, cystatin-calbumin, NGAL, KIM-1, alpha and pi GST, L-FABP – not including the Nephrocheck® panel of TIMP-2/IGF-BP7) and for the development and severity of AKI (by KDIGO criteria) over 2 – 7 days and for a 30 day composite outcome of renal replacement therapy (RRT) or death. Of the original cohort, 38% developed AKI within 7 days of ICU admission (41% of whom had stage 3 AKI and 31% met the 30 days composite end-point).
Importantly, no single biomarker was capable of predicting AKI with an area under the curve (AUC) of >0.70. Clinical models alone (APACHE-II, gender, serum creatinine, urine output, BUN) were modestly predictive of AKI and outcomes, particularly for stage 3 AKI, but the biomarkers added no statistically significant improvement of AUC. Combinations of NGAL/albumin seemed to perform best for predicting severe AKI with an AUC of 0.90.
The take home message (for me) is that clinical modelling appears adequate for prediction of AKI and outcomes in a heterogeneous collection of patients with critical illness admitted to an ICU, and that testing of urinary biomarkers adds little (caveat: the FDA approved Nephrocheck® test was not evaluated in this study) to this prediction. Many factors probably account for these disappointing results. Further studies are needed, using broad panels of biomarkers and more homogeneous patient cohorts to explore the potential value of biomarker profiles in ICU patients at high risk of developing AKI.
Quoted Karger Article