CORRELATION BETWEEN BANFF CLASSIFICATION, ACUTE RENAL REJECTION SCORES AND REVERSAL OF REJECTION

The Banff classification of acute rejection is based on histologic gra des and scores for borderline changes, glomerular, vascular, interstit ial and tubular lesions. We reviewed 56 episodes of acute rejection oc curring in 44 kidney allograft recipients (30 cadaveric and 14 living donor transplants), comparing Banff classification to degree of revers ibility of rejection. Rejection reversal was defined as complete if se rum creatinine returned less than or equal to 25% of baseline, partial if creatinine was > 25% to < 75% of baseline, and irreversible if cre atinine was less than or equal to 75% of baseline or graft loss occurr ed. Eight biopsies were classified as borderline (SUM score 1.6 +/- 0. 5), 14 grade I (SUM score 3.3 +/- 0.4), 19 grade II (SUM score 4.2 +/- 0.3), and 15 grade III (SUM score 8.5 +/- 0.4). SUM distinguished bor derline and grade III rejections, but not grades I and II. Clinically, grade and SUM score correlated with rejection reversal. Complete reve rsal of rejection occurred in 93% of patients with grade I rejection, while 47% of patients with grade III had irreversible rejection. The m ean SUM for complete reversal was 3.9 +/- 0.34 and was different from SUM of partial (6.0 +/- 0.86) and irreversible (8.5 +/- 0.93), P < 0.0 06. Meanwhile, vascular scores were similar for rejections with comple te (0.9 +/- 0.2) or partial (1.0 +/- 0.4) reversal, but significantly higher in those with irreversible rejection (3.0 +/- 0.4, P < 0.000). Likewise, mean scores for tubulitis and interstitial inflammation were significantly higher for irreversible rejection. Resolution of reject ion by steroids was correlated to low vascular score (steroid sensitiv e 0.65 +/- 0.25 vs. steroid resistant 1.42 +/- 0.18, P < 0.01), and lo w SUM score (steroid sensitive 3.7 +/- 0.5 vs. steroid resistant 5.22 +/- 0.43, P < 0.04). Neither scores for tubulitis nor interstitial cel lular inflammation were predictive of steroid sensitivity. These data demonstrate that Banff scoring has clinical relevance in predicting re jection reversal and has implications to first-line therapy of rejecti on episodes.