Bronchoalveolar lavage (BAL) is a practicable procedure establishing the etiology of pneumonia. In patients with neutropenia, empirical antimicrobial treatment is mandatory immediately after diagnosis of infection, usually before results of BAL are available. We evaluated the impact of BAL on treatment and outcome of pneumonia in immunocompromised patients with a special regard to neutropenia. Bronchoscopy with BAL was performed in 58 episodes of clinical documented pneumonia in patients with hematological malignancies (88%) or solid tumors (12%), in 30 cases patients had neutropenia, in 28 cases patients had no neutropenia. In 93% of cases, BAL was performed under empirical antimicrobial treatment. BAL fluid was cultivated for bacteria, fungi, and tested for Pneumocystis jirovecii and cytomegalovirus (CMV). BAL revealed positive bacterial results in 67% of cases. Gram-positive microorganisms were detected in 95% of positive BAL results, gram-negative microorganisms in 23%, mixed bacterial cultures occurred in 41%. Positive fungi cultures were found in 59%. P. jirovecii was detected in 5% of cases tested and CMV in 8%. There was no significant difference between neutropenic and non-neutropenic patients. BAL results directed a change of therapy in only six of 58 episodes (5%). Overall mortality related to pneumonia was 16%. In this patient setting, the yield of BAL rarely has a significant influence on treatment and outcome of pneumonia. The early beginning of antimicrobial treatment reduces the diagnostic yield of BAL. In patients with pneumonia during neutropenia, its use should be well considered.