RESUMO
BACKGROUND: The Lung Rejection Study Group (LRSG) created a scheme for grading acute allograft rejection in 1990 and then revised it in 1996, but virtually no studies have evaluated the reliability of this formulation. This investigation assessed the reliability of the current LRSG system by determining inter- and intrareader agreement for grading transbronchial biopsy samples from lung transplant recipients. METHODS: Biopsy samples from a cohort of 204 recipients were reviewed and classified by a single pathologist who was blinded to original interpretations. The "A" and "B" rejection grades from this contemporary review were compared with original grades by the kappa statistic. RESULTS: For "A" grading, weighted kappa was 0.65 (95% confidence interval [CI] 0.60-0.70) for interreader agreement (n = 529 specimens) and 0.65 (95% CI 0.53-0.76) for intrareader agreement (n = 97 specimens). For "B" grading, weighted kappa was 0.26 (95% CI 0.14-0.39) for interreader agreement (n = 164 specimens) and 0.33 (95% CI 0.15-0.51) for intrareader agreement (n = 58 specimens). CONCLUSIONS: On the basis of the analysis of the LRSG scheme, "A" grades exhibit very good reliability, but "B" grades have only fair reliability, and steps to improve this shortcoming should be taken.