Utility of fiberoptic bronchoscopy in neutropenic patients admitted to theintensive care unit with pulmonary infiltrates

Objective: To analyze the impact of fiberoptic bronchoscopy and bronchoalve olar ravage (BAL) on guiding the treatment and intensive care unit (ICU) cl inical outcome in neutropenic patients with pulmonary infiltrates admitted to the ICU. Design: Prospective collection of data. Setting: Medical ICU in a teaching hospital. Patients: During a 6-yr period, we analyzed the results of 93 fiberoptic br onchoscopies plus BALs performed in 93 consecutive neutropenic ICU patients . We separated the patients into two groups according to the cause of neutr openia (high-dose chemotherapy [n = 41] or stem cell transplantation [SCT; n = 521]). Results: Of the 93 BALs, 53 were performed to evaluate diffuse infiltrates and 42 were performed on mechanically ventilated patients. Forty-nine perce nt of BALs (46 patients) were diagnostic, with a significantly better yield in ICU patients with high-dose chemotherapy-induced neutropenia (26 of 41 BALs). The number of cases of proven infectious pneumonia was significantly higher in this group of ICU neutropenic patients. In patients who underwen t SCT, diffuse infiltrates were statistically correlated with a negative re sult of BAL. Twenty-six patients who underwent diagnostic BALs changed ther apy. Sixteen complications (17%) occurred with only two intubations, The ov erall mortality rate in the ICU and the mortality rate in mechanically vent ilated neutropenic patients were 71% and 93%, respectively. In neutropenic patients who underwent SCT, the mortality rate was statistically higher in patients in whom no diagnosis was established. Patients who had a diagnosti c BAL that changed therapy did not have an increased probability of surviva l compared with patients who had a BAL that did not change therapy. Conclusions:The use of routine diagnostic BAL in ICU neutropenic patients w ith pulmonary infiltrates is difficult to establish, even ii BAL is helpful in the management of these critically ill patients. BAL in our ICU neutrop enic patient population had an acceptable overall diagnostic yield (49%), w hich was higher in ICU patients with chemotherapy-induced neutropenia. Neve rtheless, in the ICU, if BAL had a low complication rate, it had infrequent ly led to changed treatment and was not associated with improved patient su rvival.