Risk factors for liver abscess formation after hepatic chemoembolization

PURPOSE: To assess the frequency and risk factors for liver abscess after hepatic chemoembolization. MATERIALS AND METHODS: The authors performed retrospective analysis of 397 chemoembolization procedures in 157 patients. All patients received prophylactic intravenous antibiotics before the procedure and 5 days of oral antibiotics after the procedure. The association between abscess formation and risk factors was determined with use of chi(2) analysis and the Fisher exact test and expressed as an odds ratio. RESULTS: Liver abscess occurred in seven of 157 patients (4.5%) after eight of 397 procedures (2.0%) at a mean of 19 d +/- 7 after chemoembolization. No patients had neutropenia. Organisms isolated reflected intestinal flora. Six patients required percutaneous drainage for 35 d +/- 29. The seventh patient required drainage for the remainder of life as a result of a nonhealing biliary fistula. Three of 24 patients with neuroendocrine tumors had abscesses (12.5%; odds ratio, 4.6; 95% CI, 0.96-22.1; P =.07), as did three of 14 patients with gastrointestinal sarcomas (21%; odds ratio, 9.5; 95% CI, 1.9-47.8; P =.016), and one of two with pancreatic adenocarcinoma. Six of the seven patients with abscesses underwent a Whipple procedure before chemoembolization. Only one patient with a history of a Whipple procedure did not develop an hepatic abscess. In the absence of a bilioenteric anastomosis, abscess occurred in only one of 150 patients (0.7%), or one of 383 procedures (0.3%). The odds ratio for liver abscess among patients with a bilioenteric anastomosis was 894 (95% CI, 50-16,000; P <.0001). CONCLUSION: Earlier bilioenteric anastomosis is the major determinant of liver abscess formation after hepatic chemoembolization. The prophylaxis regimen used failed to prevent abscess formation in patients with earlier bilioenteric anastomosis.