Diagnosis of pulmonary infections in immunocompromised patients.

In an immunocompromised patient with fever and pulmonary infiltrates, it frequently is difficult to decide which invasive procedure, if any, to use to obtain a definitive diagnosis. Because most lung infiltrates in immunosuppressed patients are caused by bacteria and sputum usually is readily available for examination, empiric therapy with potent, safe, broad spectrum, antibacterial drugs often is successful. Invasive procedures that prove a diagnosis may result in substantive changes in therapy in perhaps as few as 10 to 20 per cent of patients, and the procedure itself may harm the patient. In a unique study in which patients with acute pneumonitis without neutropenia were randomized to either empiric antibiotic treatment or treatment based on results of open lung biopsy, patients with open lung biopsy had a worse outcome, possibly related to morbidity of open lung biopsy. Furthermore, no diagnoses were provided by open lung biopsy that were not treated by the empiric regimen. A missed treatable disease may be tragic, however. A thoughtful clinician must evaluate each patient with careful consideration of the history in light of the underlying disease and its treatment, rapidity of clinical course, physical examination, and laboratory data, particularly the chest radiograph, sputum examination, and bleeding parameters. Fiberoptic bronchoscopy with washings and brushings is very safe; the addition of transbronchial biopsy adds diagnostic power at the price of some complications. Bronchoalveolar lavage is a very promising technique that probably will find widespread use. However, none of the foregoing techniques is completely sensitive. When no diagnosis is established and bronchoscopy studies are negative, open lung biopsy must be considered, especially when the chest radiograph or computed tomography scan suggests focal disease or lymphadenopathy. Needle aspiration can be used, particularly if local experience is favorable and lung disease is peripheral. When evaluating a procedure, local experience must be considered rather than reliance on published diagnostic yields and complication rates. New diagnostic and therapeutic developments may change decision analysis in the near future. At present, cultures for viruses and fungi and serologic techniques have little application at most medical centers, and decisions on data from invasive procedures pivot on interpretation of histology and smears. Development of assays for antigen (for example, Aspergillus) and rapid culture techniques (for example, cytomegalovirus and the shell vial method), coupled with new, effective antimicrobials, may demand maximum effort for a definitive diagnosis in every patient.