The purpose of this study was to describe the clinical features at onset and outcome and the diagnostic approach in subjects with bronchiolitis obliterans-organizing pneumonia (BOOP). Over a 7-year period we observed 78 cases of biopsy-proven bronchiolitis obliterans-organizing pneumonia, in which well documented clinical and radiographic data were available. The final diagnosis of BOOP was validated when patients presented: (i) negative microbiological analysis on BAL fluid; (ii) a well documented improvement either spontaneous, or after steroid treatment or (iii) cases with progressive respiratory failure and increasing radiographic shadows, an open lung biopsy or autopsy that excluded other entities. There were 42 males and 36 females; the mean age was 61+/-12 years (range 12-85 years). Forty-two (54%) patients were current smokers, 25 (32%) had never smoked and 11 (14%) were ex-smokers. The clinical pattern at presentation of BOOP was more frequently similar to classical acute or sub-acute infectious pneumonia. Fever (63%), dyspnoea (58%) and dry cough (53%) were the typical symptoms on admission. A flu-like syndrome preceeding BOOP was observed in 21 cases (27%). Inspiratory crackles (78%) were the most typical finding at physical examination. However, 13% of the patients were asymptomatic and an abnormal chest X-ray film was the reason for seeking medical attention. Radiographically the most frequent pattern of BOOP was a unilateral consolidation (44%) with lower field predominance. A migratory behaviour was present in 22% of the cases. High-resolution computed tomographic (HR-CT) scan when performed, was more sensitive in detecting ground glass infiltrates, sub-pleural or peri-bronchovascular distribution or the presence of nodules or cavitation. Most patients (68%) were classified as having idiopathic BOOP. However, the same clinical-roentgenological pattern was observed in patients after radiotherapy for ductal breast carcinoma (6%), in collagen-vascular diseases (6%), related to drugs (9%), to infections serologically documented (4%), and to graft vs. host disease (4%). Four patients (all of whom had idiopathic BOOP) presented a rapid progressive respiratory failure needing mechanical ventilation. In another two cases respiratory failure appeared after a long period during which patients experienced exertional dyspnoea and low grade fever. BAL profile was characterized by lymphocytosis with a reduction of the CD4/CD8 ratio, associated with a slight increase of neutrophils and eosinophils and scattered mast cells. However in two cases we had an increased CD4/CD8 ratio and in one case the presence of a significant 12% of polyclonal B cells. In a few cases atypical (cytokeratin-positive cells) epithelial cells were detected: these cells were constantly present in the BAL fluid of patients with rapidly progressive respiratory failure. From the diagnostic point of view this series documents that transbronchial lung biopsy (coupled with BAL) can be the first diagnostic step. However, therapy can be started on the basis of BAL data (when a characteristic morphological and phenotypical profile is evident) in cases in which the clinical presentation is suggestive and a biopsy cannot be made. Most patients showed a rapid and good response to steroid therapy. However, three patients died (4%) in spite of steroid therapy (two cases) and steroid and cyclophosphamide therapy (one case). In conclusion, although clinical findings, chest X-ray film and CT Scan findings usually suggest the diagnosis a definite confirmation requires transbronchial lung biopsy and BAL and, less frequently, open lung biopsy.