Distinct morphological features of acute tubular injury in renal allografts correlate with clinical outcome.

Acute tubular injury (ATI) is common in renal allografts and is related to inferior long-term allograft function. However, it is unknown which of the morphological features of ATI can predict outcome and how they should be graded. Here, we examine features of ATI systematically in protocol biopsies and biopsies for cause to define the most predictive features. Analyses included 521 protocol biopsies taken at 6 wk, 3 mo, and 6 mo after transplantation and 141 biopsies for cause from 204 patients. Features of ATI included brush border loss, tubular epithelial lucency, flattening, pyknosis, nuclei loss, and luminal debris, each graded semiquantitatively. Additional immunohistochemical stainings were performed for markers of cell injury (neutrophil gelatinase-associated lipocalin), cell death [cleaved caspase-3, fatty acid-coenzyme A ligase 4 (FACL4)], and proliferation (Ki-67). Interobserver reliability was good for pyknosis, flattening, and brush border loss and poor for lucency, nuclei loss, and luminal debris. In protocol biopsies between 6 wk and 6 mo, the degree of ATI remained virtually unchanged. Biopsies for cause had generally higher injury scores. Deceased donor source, delayed graft function, ganciclovir/valganciclovir treatment, and urinary tract infection correlated with ATI. The degree of pyknosis, flattening, and brush border loss correlated best with impaired allograft function. FACL4 expression was observed in areas of ATI. Only patients with Ki-67 expression showed stable or improved allograft function in the longitudinal assessment. Reliable assessment of ATI is possible by semiquantitative grading of tubular epithelial cell brush border loss, flattening, and pyknosis. Examination of Ki-67 expression can help determine the potential for recovery from this damage.