Non-inferiority of creatinine excretion rate to urinary L-FABP and NGAL as predictors of early renal allograft function

BACKGROUND We evaluated accuracy of urinary liver type fatty acid-binding protein (L-FABP) for prediction of early allograft function and compared it to neutrophil gelatinase associated lipocalin (NGAL), diuresis and urinary creatinine excretion rate (UCr). METHODS Urine samples from 71 consecutive patients were taken 4, 10, 24 and 48 h after transplantation. We classified recipients into two groups: immediate graft function (IGF), with more than 70% reduction of serum Cr at 7th day post-transplant, and delayed graft function (DGF)/slow graft function (SGF) group (DGF--the need for hemodialysis procedure in the first week, SGF--less than 70% reduction of serum Cr in the first week). RESULTS Thirty-one recipients had IGF and 40 had DGF/SGF. L-FABP was only useful 48 h post-transplant with ROC AUC of 0.85 (95% C.I. 0.74-0.92); NGAL 24 h post-transplant had ROC AUC of 0.82 (0.7-0.91). Sensitivity, specificity, PPV and NPV for prediction of DGF/SGF with L-FABP > 9.5 mg/mmol Cr and NGAL > 33.1 μg/mmol Cr were: 86, 80, 83 and 83% (L-FABP), and 68, 93, 91, and 73% (NGAL). The difference in urine output between the groups was largest 4 h post-transplant (p = 0.001), later on the difference diminished. There were no significant differences in ROC AUC between L-FABP at 48 h, NGAL at 24 h, urine output at 4 h and UCr excretion rate at 10 h post-transplant. UCr < 0.56 mmol/h 10 h post-transplant predicted DGF/SGF with 94% sensitivity, 84% specificity, 89% PPV and 91% NPV, ROC AUC was 0.9. Classification tree with urine output 4 h and UCr 10 h post-transplant accurately predicted 89% of outcomes. When L-FABP or NGAL were added, the prediction was accurate in 92 or 90%, respectively. CONCLUSIONS L-FABP is comparable to NGAL for prediction of first week allograft function, however UCr and diuresis were non-inferior.