The objective of this study was to clarify conflicting reports of the sensitivity and specificity of bronchoalveolar lavage or bronchial washings for diagnosing invasive pulmonary aspergillosis. The study was a retrospective review of 300 consecutive patients in a tertiary referral centre subjected to 343 fiberoptic bronchoscopic procedures for the evaluation of pulmonary infiltrates. Classification of paired fungal culture and cytologic examination of bronchoalveolar lavage or bronchial washing fluid according to clinical, radiographic, histological and autopsy evidence of invasive pulmonary aspergillosis. One-hundred and fifteen deaths occurred, with a 58% autopsy rate. A diagnosis of invasive pulmonary aspergillosis was made in 21 immunosuppressed patients with 16 deaths. Bronchoalveolar lavage cytology showed aspergillus in 19 specimens (invasive pulmonary aspergillosis in 16), cultures yielded aspergillus in 41 (invasive pulmonary aspergillosis in ten), with both tests positive in nine. Cytology sensitivity was 64.0%, specificity 99.1%, positive predictive value 84.2%, and negative predictive value 97.2%. Culture sensitivity was 40.0%, specificity 90.3%, positive predictive value 24.4%, and negative predictive value 95.0%. Concordant cytology and culture sensitivity was 32.0%, specificity 99.7%, positive predictive value 88.9%, and negative predictive value 94.9%. In conclusion, when characteristic hyphae are visualized in bronchoalveolar lavage specimens from immunosuppressed patients with compatible clinical data, it is advisable to treat for presumptive invasive pulmonary aspergillosis.