POS-072 PRACTICAL EXPERIENCE WITH THE IMPLEMENTATION OF NGAL FOR AKI DIAGNOSIS IN A CLINICAL SETTING

Abstract

Kidney injury biomarkers for AKI have been shown to detect AKI earlier, predict progression and better discriminate etiology compared to serum creatinine. These biomarkers have been studied in various clinical trials, but their implementation and prediction capacity in standard clinical settings have rarely been reported. We describe our initial clinical experience during the implementation of the Neutrophil gelatinase-associated lipocalin (NGAL) in our academic medical centre. We prospectively audited the initial implementation and use of urine and plasma NGAL (uNGAL, pNGAL) as diagnostic biomarkers in hospitalised patients with newly diagnosed AKI. This novel biomarker was ordered only by the Nephrology Consult service. Questionnaires for initial evaluation of the suspected etiology and risk of progression were completed by the consulting renal team before and one day after usual diagnostic workup. NGAL results were available for the second follow-up assessment. An NGAL level of 150 ng/mL was used by the SVUH Clinical Biochemistry Laboratory as the cut-off value to discriminate the presence of tubular damage (range 25-3000 ng/mL). An attending nephrologist, not involved in the care of these patients, independently conducted a retrospective review using medical records, to determine the most likely etiology of AKI, and whether or not clinical progression occurred following initial renal consultation. Descriptive statistics and Mann-Whitney U tests were used for analyses. This audit was pre-approved by the SVUH Clinical Audit Committee. Between May and October 2020, 47 NGAL samples were measured (uNGAL, n=23; pNGAL, n=24) in 26 AKI patients, with a median age of 66.5 (IQR 55.3-81.0). Chronic kidney disease was present in 12 (46%) patients. Stage 1, 2 and 3 AKI occurred in 6 (23%), 11 (42%), and 9 (35%) patients, respectively. The median delay between admission and NGAL measurement was 4 (IQR 0-20) days. The creatinine level at diagnosis was similar for pre-renal AKI (median 224 (IQR 157-340) umol/L) and intra-renal AKI (median 247 (IQR 193-487) umol/L). Urine NGAL was 41 (IQR 29-280) ng/mL in the pre-renal group, compared to 1308 (IQR 175-1361) ng/mL for intra-renal AKI (p=0.08), while pNGAL were 251 (IQR 99-340) ng/mL and 420 (IQR 255-865) ng/mL (p=0.19), respectively. NGAL levels were higher in the subgroup of ischemic ATN (uNGAL: 1332 (IQR 888-1365) ng/mL; pNGAL: 495 (IQR 384-889) ng/mL) compared to toxic ATN (uNGAL: 178 (IQR 105-1389) ng/mL; pNGAL 208 (IQR 139-751) ng/mL). According to adjudication, the initial clinical impression of AKI etiology was incorrect in 3 patients (12%), while the initial prediction of progression vs. non-progression was incorrect in 8 patients (31%). In those 8 patients, NGAL values correctly predicted progression in 7 of them. We reported the early use of both uNGAL and pNGAL in clinical practice on the Nephrology consultation service in an academic medical centre. Based on our experience, we consider that NGAL, especially uNGAL can be an interesting tool for differential diagnosis and to predict AKI progression. The diagnostic accuracy of NGAL tests for these purposes may be improved by using a higher cut-off level to determine the etiology and prognosis of new AKI cases. Additional integration of NGAL tests into clinical algorithms for AKI management and follow-up audits are required.