Percutaneous drainage of pancreatic collections has recently been advocated as a means of diagnosis of bacterial contamination, for temporizing unstable patients, and as definitive treatment in itself. In order to assess its efficacy, the role of percutaneous drainage of infected pancreatic fluid collections was retrospectively reviewed by a single surgical practice. Seventeen patients were treated over a 5-year period from 1987 to 1992. All patients admitted or referred with a diagnosis of infected peripancreatic fluid collection were included in the review. The group consisted of eleven males and six females; mean age was 55.2 years (range 28 to 70). Patients were stratified into one of two groups based on initial treatment modality. Group A consisted of eight patients treated initially with percutaneous drainage as presumed definitive management. Eight patients in Group B were treated initially with surgical debridement and drainage. APACHE II scores on admission were 5.62 +/- 3.66 for Group A and 9.12 +/- 3.87 for Group B (N.S.). Mean hospital stay was 100 days (range 13-311) for Group A and 71 (range 25-149) for Group B (N.S.). Despite initial percutaneous drainage, six of eight (75%) patients in Group A required operative debridement because of clinical deterioration. APACHE II scores in this subset went from 6.83 +/- 3.43 to 9.83 +/- 5.04 (N.S.) despite a total of 18 preoperative percutaneous procedures (2.25 per patient; range 1-7). The number of complications for this group totaled 15. Five of the six patients with positive cultures from their initial aspiration failed percutaneous drainage.(ABSTRACT TRUNCATED AT 250 WORDS)