A frequent dilemma is discerning the likelihood of pneumonia and the need for empiric antibiotic therapy in liver transplant recipients with pulmonary infiltrates in the intensive care unit (ICU).We performed a prospective, observational study of consecutive liver transplant recipients developing pulmonary infiltrates in the ICU.Of 90 consecutive liver transplant patients in the ICU over a 3-year period, 44% (40) developed pulmonary infiltrates. The etiologies were pneumonia (38%, 15 of 40), pulmonary edema (40%, 16 of 40), atelectasis (10%, 4 of 40), adult respiratory distress syndrome (8%, 3 of 40), contusion (3%, 1 of 40), and unknown (3%, 1 of 40). Pneumonia was due to methicillin-resistant Staphylococcus aureus in 27% (4 of 15), Pseudomonas aeruginosa (27%, 4 of 15), invasive aspergillosis (20%, 3 of 15), and Enterobacter cloacae, Serratia marcescens, Pneumocystis carinii pneumonia, and unknown (7%, 1 of 15) in one each. None of the patients had cytomegalovirus or herpes simplex virus pneumonia. Seventy-five percent of methicillin-resistant Staphylococcus aureus and all Aspergillus pneumonias, but only 14% of the Gram-negative pneumonias, occurred within 30 days of transplantation. Twenty-seven percent of the pneumonias occurred >365 days after transplantation; all of these were in patients with recurrent viral hepatitis C virus or hepatitis B virus, disseminated posttransplant lymphoproliferative disorder, or late rejection. Of patients with pneumonia, 87% were ventilated and 40% had bacteremia. Clinical pulmonary infection score (Pugin score) >6 (73% vs. 6%, P = 0.0001), abnormal temperature (73% versus 28%, P = 0.005), and creatinine level >1.5 mg/dl (80% versus 50%, P = 0.05) were predictors of pneumonia versus other etiologies of pulmonary infiltrates. Overall mortality in patients with pulmonary infiltrates was 28% (11 of 40); pneumonia as etiology (P = 0.06), creatinine level >1.5 mg/dl (P = 0.028), higher blood urea nitrogen (P = 0.017), and worse APACHE neurological score (P = 0.04) were predictors of poor outcome.Our data have implications not only for identifying pneumonia as a potential cause of pulmonary infiltrates, but for the likely etiology of the pneumonia and thus the selection of empiric antibiotic therapy in critically ill liver transplant recipients. Pugin score >6 in patients with pulmonary infiltrates warrants antimicrobial therapy. Early onset within 30 days after transplantation raises the spectra of aspergillosis.