Study objective: To determine if performing bronchoalveolar lavage (BAL) from more than one lung segment augments the diagnostic yield in patients with pulmonary infection.
Patients and study design: Seventy-six patients who underwent bisegmental BAL for the diagnosis or exclusion of pulmonary infection were studied prospectively.
Main results: In patients with AIDS, the concentration of Pneumocystis carinii was higher in the upper lobes than in the middle lobes, regardless of whether the patients had been receiving pentamidine prophylaxis. In patients without HIV infection, the number of P carinii clusters was much lower. In 2 of 5 HIV-negative patients, P carinii organisms were detectable but in one of two lavage specimens, whereas in only 1 of 19 AIDS patients, P carinii organisms were not found in both lavaged segments. In bacterial pneumonia, BAL fluid from a segment of the radiologically most involved area had a much higher cell concentration, percentage of neutrophils, and concentration of bacteria than from the segment that was not or less severely involved on chest radiograph. In two of nine patients with AIDS and cytomegalovirus (CMV) pneumonia, cytopathogenic CMV effects were not found in both lavaged segments. In one of eight patients, mycobacteria could be cultured only from one of two radiologically involved segments.
Conclusion: An increase in the diagnostic sensitivity by performing BAL in two lung segments is limited to cases where P carinii pneumonia is a relevant consideration in immunocompromised patients without HIV infection. In bacterial pneumonia, BAL can be performed in a single radiographically involved lung segment without a loss in diagnostic sensitivity. Since our study population of patients with CMV pneumonitis, mycobacterial infections, and fungal infections was small, no reliable conclusions are possible and BAL of more than one lung segment seems justified until more information is available.