Natural history of pulmonary complications in children after bone marrow transplantation.

We sought, in children after bone marrow transplantation (BMT), (1) to determine the natural history and incidence of pulmonary complications, (2) to evaluate the diagnostic yield of fiberoptic bronchoscopy and bronchoalveolar lavage (BAL); and (3) to determine the effect of bronchoscopy with lavage on patient outcome. The study design was a retrospective review in a tertiary care university hospital of all children undergoing BMT over a 5-year period. Patients were separated into 2 study groups: children with and without pulmonary complications. Pulmonary complications were defined as new or persistent pulmonary infiltrates on chest radiograph or chest computed tomography scan, respiratory symptoms, hypoxemia, or hemoptysis. Three hundred sixty-three pediatric patients underwent BMT between January 1, 1995, and December 31, 1999. Ninety patients (25%) developed pulmonary complications and were evaluated with bronchoscopy and BAL. Patients with pulmonary complications had a higher mortality (65% versus 44%; P < .01). The median posttransplantation survival for children with pulmonary complications was 258 days, compared with 1572 days in patients without pulmonary complications. The incidence of pulmonary complications was increased in patients with allogeneic BMT (P < .01). The time-dependent onset of severe (grade III to IV) graft-versus-host disease increased the relative risk of pulmonary complications by 2.0 (95% confidence interval, 1.1-3.7; P = .02). Pulmonary complications increased the time-dependent relative risk of mortality by 3.5 (95% confidence interval, 2.5-4.8). The diagnostic yield of bronchoscopy with lavage was 46% in patients undergoing BAL. Diagnostic bronchoscopy did not enhance either 30- or 100-day survival. Pathogen identification did not decrease mortality (P = .45). Pulmonary complications occur in 25% of children undergoing BMT and increase the risk of death in the first year after BMT. Although pathogen identification does not confer a survival advantage, rigorous, prospective screening may allow for earlier identification of pathogens and thereby provide a benefit to this uniquely vulnerable population.