Archdekin Ben, Sharma Atul, Gibson Ian W, Rush David, Wishart David S, Blydt-Hansen Tom D
Faculty of Health Sciences, Queen's University, Kingston, Ontario, Canada.
Department of Pediatrics and Child Health, Children's Hospital at Health Sciences Center, University of Manitoba, Winnipeg, Manitoba, Canada.
Pediatr Transplant. 2019 May;23(3):e13364. doi: 10.1111/petr.13364. Epub 2019 Feb 4.
Acute kidney injury (AKI) is a major concern in pediatric kidney transplant recipients, where non-alloimmune causes must be distinguished from rejection. We sought to identify a urinary metabolite signature associated with non-rejection kidney injury (NRKI) in pediatric kidney transplant recipients. Urine samples (n = 396) from 60 pediatric transplant participants were obtained at time of kidney biopsy and quantitatively assayed for 133 metabolites by mass spectrometry. Metabolite profiles were analyzed via projection on latent structures discriminant analysis. Mixed-effects regression identified laboratory and clinical predictors of NRKI and distinguished NRKI from T cell-mediated rejection (CMR), antibody-mediated rejection (AMR), and mixed CMR/AMR. Urine samples (n = 199) without rejection were split into NRKI (n = 26; ΔSCr ≥25%), pre-NRKI (n = 35; ΔSCr ≥10% and <25%), and no NRKI (n = 138; ΔSCr <10%) groups. The NRKI discriminant score (dscore) distinguished between NRKI and no NRKI (AUC = 0.86; 95% CI = 0.79-0.94), confirmed by leave-one-out cross-validation (AUC = 0.79; 95% CI = 0.68-0.89). The NRKI dscore also distinguished between NRKI and pre-NRKI (AUC = 0.82; 95% CI = 0.71-0.93). In a linear mixed-effects regression model to account for repeated measures, the NRKI dscore was independent of concurrent rejection, but there was a non-statistical trend for higher dscores with rejection severity. A second exploratory classifier developed to distinguish NRKI from clinical rejection had similar test characteristics (AUC = 0.81, 95% CI = 0.70-0.92, confirmed by LOOCV). This study demonstrates the potential of a urine metabolite classifier to detect NRKI in pediatric kidney transplant patients and non-invasively discriminate NRKI from rejection.
急性肾损伤（AKI）是小儿肾移植受者的一个主要问题，在这些受者中必须区分非同种免疫性病因与排斥反应。我们试图确定小儿肾移植受者中与非排斥性肾损伤（NRKI）相关的尿液代谢物特征。在肾活检时获取了60名小儿移植参与者的尿液样本（n = 396），并通过质谱对133种代谢物进行了定量分析。通过潜在结构判别分析对代谢物谱进行了分析。混合效应回归确定了NRKI的实验室和临床预测因素，并将NRKI与T细胞介导的排斥反应（CMR）、抗体介导的排斥反应（AMR）以及混合性CMR/AMR区分开来。将无排斥反应的尿液样本（n = 199）分为NRKI组（n = 26；ΔSCr≥25%）、NRKI前期组（n = 35；ΔSCr≥10%且<25%）和无NRKI组（n = 138；ΔSCr<10%）。NRKI判别分数（dscore）区分了NRKI组和无NRKI组（AUC = 0.86；95%CI = 0.79 - 0.94），留一法交叉验证证实了这一点（AUC = 0.79；95%CI = 0.68 - 0.89）。NRKI dscore还区分了NRKI组和NRKI前期组（AUC = 0.82；95%CI = 0.71 - 0.93）。在一个考虑重复测量的线性混合效应回归模型中，NRKI dscore与同时发生的排斥反应无关，但随着排斥反应严重程度的增加，dscore有升高的非统计学趋势。为区分NRKI与临床排斥反应而开发的第二个探索性分类器具有相似的测试特征（AUC = 0.81，95%CI = 0.70 - 0.92，经留一法交叉验证证实）。这项研究证明了尿液代谢物分类器在检测小儿肾移植患者NRKI以及无创区分NRKI与排斥反应方面的潜力。
