目的探讨尿N-乙酰-β-D-氨基葡萄糖苷酶(urine N-Acetyl-β-D-glucosaminidase,uNAG)联合血肌酐(serum creatinine，sCr)在成人重症患者中预测急性肾损伤(acute kidney injury，AKI)的临床价值。方法前瞻性入选2015 年9 月～2016 年1 月入住南方医科大学附属小榄医院重症医学科(intensive care unit,ICU)的成人重症患者共124 例。按照2012 年改善全球肾脏病预后组织(kidney disease:improving global outcomes，KDIGO)的AKI 诊断标准(KDIGO 标准)，将患者分为非AKI组和AKI 组，比较各组间uNAG 及sCr 水平,运用受试者工作特征曲线(receiver operating characteristic curve,ROC 曲线)评价uNAG、sCr 及两者联合后对AKI 诊断的预测能力。结果①124 例重型患者中有34 例发生AKI，AKI 发生率为27.4%。②AKI 组患者uNAG(Z=-6.579，P＜0.001)和sCr(Z=-5.160，P＜0.001)水平显著高于非AKI 组。③15 例患者发展为重症AKI(12.1%)，重症AKI 组uNAG 和sCr 水平均明显高于非AKI 组(uNAG：Z=-5.353，P＜0.001；sCr：Z=-4.846，P＜0.001)和轻症AKI 组(uNAG：Z=-1.994，P=0.046；sCr：Z=-2.880，P＜0.004)，差异有统计学意义。④uNAG 及sCr 预测AKI 的曲线下面积(area under the curve,AUC)分别为O.884 和0.808，两者联合预测AKI 的AUC 为0.933，明显高于单一指标(Delong法，uNAG+sCr 比uNAG：Z=1.979，P=0.048；uNAG+sCr 比sCr：Z=3.270，P=0.001)。⑤整个研究队列的死亡率为9.7%，肾脏替代治疗率为4.0%，AKI 组的肾脏替代治疗率明显高于非AKI 组(χ2=7.238，P=0.020)，住ICU 时间明显更长(Z=-2.547，P=0.011)。结论uNAG 是预测重症患者发生AKI 的敏感指标，其与sCr联合后可进一步提高预测重症患者发生AKI 的效能。

Objective To explore the clinical value of urinary N-Acetyl-β-D-glucosaminidase (uNAG) associated with serum creatinine (sCr) in predicting acute kidney injury (AKI) in adult critically ill patients. Methods In this prospective study, we enrolled 124 adult critically ill patients who admitted to the mixed ICU of Xiaolan Hospital of Southern Medical University during September 2015 to January 2016. According to the Kidney Disease Improving Global Outcomes (KDIGO) criterion, the patients were divided into non- AKI group and AKI group (including mild AKI and severe AKI)． The receiver operating characteristic (ROC) curve and the area under curve (AUC) were used to evaluate the capabilities of the biomarkers in predicting AKI. Results Of the 124 patients, 34 patients (27.4%) developed AKI. The levels of uNAG and sCr were significantly higher in AKI group than in non-AKI group (P＜0.001)．In this cohort, 15 patients (12.1%) developed severe AKI. The levels of uNAG and sCr were significantly higher in severe AKI patients than in non- AKI and mild AKI patients (P＜0.05). The AUC value was higher in combined uNAG and sCr (0.933 & 0.946) than uNAG or sCr alone in predicting AKI and severe AKI. In-hospital mortality was 9.7% and renal replacement therapy rate was 4.0%. AKI group had higher renal replacement therapy rate and longer ICU stay than non AKI group (P＜0.05). Conclusions uNAG is a sensitive biomarker for predicting AKI in adult critically ill patients. When uNAG combined with sCr, AKI prediction can be further improved.