Inflammatory changes in acute respiratory distress syndrome (ARDS) were evaluated using bronchoalveolar lavage (BAL) to sample inflammatory cells of the lower respiratory tract. Currently, no reliable technique exists for harvesting large numbers of viable alveolar macrophages. We constructed a modified bronchoalveolar lavage (MBAL) cannula using a commercially available retrograde coronary sinus reperfusion cannula with a large central lumen for rapid fluid instillation, a distal balloon for bronchial seal, and distal side ports to prevent a high-pressure jet. In normal sheep, mean aspirate return was significantly higher using MBAL compared with bronchoscopy (69 ± 4 mL vs. 25 ± 1 mL;p≤ 0.001). We then compared macrophage yield using bronchoscopy and MBAL using their lethal dose 40% (LD40) smoke/burn sheep model of ARDS. Total macrophage recovery was significantly higher (more than 80 times) using MBAL compared with bronchoscopy (16.9 ± 3.6 × 106macrophages vs. 0.20 ± 0.06 × 106macrophages; p ≤ 0.009). Likewise, the number of macrophages per milliliter of aspirate was significantly higher with MBAL compared with bronchoscopy (0.46 ± 0.16 × 106macrophages/mL vs. 0.009 ± 0.003 × 106macrophages/mL;p≤ 0.05). We conclude the MBAL cannula enables harvest of significantly higher quantities of macrophages in smoke/burn-injured sheep, facilitating serial assessment of the pathophysiology of ARDS.