Background and methods: We reviewed and correlated the histologic and clinical records for the 1027 transbronchial biopsies performed, as clinically indicated, in 313 heart and lung transplant recipients in the Harefield Transplant Unit from 1988 through 1991. Three pieces of lower lobe or radiologically abnormal lung were routinely sent for histologic diagnosis. Clinical diagnoses of rejection and infection were based on symptomatologic, radiologic, and bacteriologic findings and response to appropriate therapy. Standard histopathologic technology and diagnostic criteria were used, including the Working Formulation for the standardization of nomenclature in the diagnosis of heart and lung rejection grading.
Results: Rejection was the most common finding (22.2%) and showed good clinicopathologic correlation. With unequivocal histologic features of rejection (Working Formulation grade A1 or above), specificity (clinical agreement with biopsy diagnosis) was 93.1% and sensitivity (clinical rejection confirmed by transbronchial biopsy) was 61%. Sensitivity increased to 77% if unsatisfactory specimens were excluded. Possible/probable rejection only was reported in 83 specimens; there were technically unsatisfactory, showed only minimal perivascular infiltrates, or had infiltrates limited to one vessel; 71% of these did have clinical rejection. Infection, excluding opportunistic, was reported in 18.5% of biopsy specimens; specificity was 70.5% and sensitivity 51.3% (both rising by 9%), with unsatisfactory specimens excluded. Histologic features of both rejection and infection were seen in 47 transbronchial biopsy specimens (4.7%). Where both components appeared definite specificity was 66.7%, but where either had been doubtful the clinical diagnosis was most often rejection. Sensitivity was also 66.7%. Cytomegalovirus inclusions were identified in 12.1% of biopsy specimens, with specificity of 91% and sensitivity of 83.5%. Sensitivity (88%) and specificity (100%) were both high for the 17 cases with pneumocystis infections. Sensitivity for the 25 transbronchial biopsy specimens from fungal infections was only 20%. Sensitivity was also poor (27.7%) in obliterative bronchiolitis, although specificity was 75%. Almost a third of transbronchial biopsy specimens from patients with obliterative bronchiolitis were unsatisfactory. Pneumonitis was the only change noted in 68 biopsy specimens. Most correlated with clinical status, but 26.5% were from patients with active rejection. Nonspecific changes or no significant pathologic condition was seen in 278 transbronchial biopsy specimens; over a third of these were from patients with clinical rejection (17.7%) or infection (18%) and 6.5% were from obliterative bronchiolitis cases. Excluding 78 technically unsatisfactory specimens reduced the proportion of false negative findings in rejection and infection by 6% and 4%, respectively.
Conclusions: We found that transbronchial biopsies consisting of three adequate pieces of lung parenchyma correlated well with clinical rejections and infections other than fungal but was of limited value in confirming a diagnosis of obliterative bronchiolitis or fungal infection.