Abstract

BACKGROUND: Lung function is altered by infection and rejection in patients who undergo heart-lung transplantation. The sensitivity, specificity, and positive/negative predictive values (PPV and NPV) of lung function for the detection of allograft dysfunction in these patients were measured. METHODS: Thirty three patients who underwent heart-lung transplantation were followed for a mean period of 16.3 months. On 123 occasions functional measurements were obtained at the time a transbronchial biopsy specimen and/or bronchoalveolar lavage fluid was taken, which were used as gold standards. Optimal sensitivity (the value for which sensitivity equals specificity) was computed for each functional test from receiver-operator characteristic (ROC) curves. RESULTS: Acute rejection was present on 31 occasions and infection on 36 occasions; 56 samples were normal. Infection and rejection were accompanied by airflow obstruction, a rise in the slopes of the alveolar plateaus for nitrogen, hexafluoride sulphur and helium (SN2, SSF6, and SHe), and a decrease in the difference between SSF6 and SHe (delta S), total lung capacity (TLC), and lung transfer factor (TLCO). Optimal sensitivities for SHe, mid forced expiratory flow (FEF25-75), TLC, and forced expiratory volume in one second (FEV1) were 68%, 67%, 66%, and 60%, respectively; they were not different for infection and rejection and did not change over the study period. For infection and rejection together, PPV ranged from 72% to 88% and NPV from 27% to 52% according to the functional test and the postoperative period considered. CONCLUSIONS: Indices of ventilation distribution, FEF25-75, and TLC have the best optimal sensitivity for the diagnosis of infection and rejection after heart-lung transplantation. The high PPV of pulmonary function in detecting allograft dysfunction observed in this study suggests that a diagnostic procedure should be performed whenever one or more functional tests deteriorate; conversely, the low NPV indicates that a stable pulmonary function does not rule out allograft dysfunction.