Percutaneous cholecystostomy (PC) has been proposed as a method of biliary decompression in critically ill patients with acute cholecystitis. We evaluated the efficacy of PC in this setting. The charts of 33 critically ill patients (mean age 52, range 5-87) who underwent PC for suspected acute cholecystitis were retrospectively examined. Univariate analysis was performed to identify which patients might benefit from PC. PC was technically successful in all patients with no direct mortality or major complications. Failure to improve within 24 hours was associated with increased mortality (P = 0.02). A total of 22/33 patients improved, 17/33 survived, and 8/33 required surgery. PC delayed definitive operation in two patients. Cholelithiasis was associated with surgical intervention (P = 0.01) but not increased mortality. Favorable prognosticators for survival included gallbladder dilatation (P = 0.01), pericholecystic fluid (P = 0.01), and absence of a pulmonary artery catheter (P = 0.02). Predictors of improvement included gallbladder nonvisualization on hepatobiliary scan (P = 0.047), positive bile cultures (P = 0.017), and initial drainage of < / = 100 cc (P = 0.009). Age, laboratory data, the use of total parenteral nutrition, and intubation did not predict outcome. Nine positive bile cultures prompted antibiotic changes in five cases. Finally, PC was less expensive than open cholecystostomy ($1620 versus $3155). PC is a safe, cost-effective, minimally invasive procedure that has diagnostic and therapeutic value in critically ill patients with acute cholecystitis. The involvement of a general surgeon is important to ensure that those patients who do not improve within 24 hours receive early surgical intervention and provide long-term definitive care for those patients with cholelithiasis.