The renal medulla in acute renal allograft rejection: comparison with renal cortex

A retrospective cohort study was undertaken to evaluate the diagnostic value of the renal medulla in acute renal allograft rejection (ARAR). One hundred and ninety-five biopsies from 98 patients were randomly selected out of 565 transplant biopsies. Biopsies were graded blindly from Grade 0 (no rejection) to Grade 3 (severe rejection) using standard criteria; ARAR was confirmed by a fall in all cases of mean serum creatinine concentration from 0.331 ± 0.182 to 0.184 ± 0.079 mmol/l, with anti-rejection therapy. In the 43 biopsies which contained both cortex and medulla, the ARAR grades and the intensities of mononuclear cell, plasma cell, polymorphonuclear cell and eosinophil infiltrates, and of interstitial oedema and haemorrhage, were similar in cortex and medulla (Spearman's Rank Correlation r=0.55−0.81, P < 0.001 ). The sensitivity, specificity and overall accuracy of medullary changes in predicting ARAR changes in the cortex were 77%, 100% and 38%, respectively. Acute vascular rejection changes could not be compared between renal cortex and renal medulla because of the anatomical differences between cortex and medulla. Further evaluation of ARAR in the all 195 biopsies, of which 188 had cortical tissue and 50 had medullary tissue, showed no significant differences in histological features (P > 0.05), except for more cortical biopsies with plasma cells (29%) than medullary biopsies with plasma cells (10%; P < 0.02). It is concluded that: (1) ARAR histological changes are similar in cortex and medulla; (2) the predictive value of ARAR medullary changes for cortical rejection changes has low sensitivity (77%) and high specificity (100%). It is suggested that a predominantly normal medullary renal biopsy in suspected rejection should be repeated to obtain cortical tissue.