Resective surgery for liver tumor: a multivariate analysis of causes and risk factors linked to postoperative complications

In spite of accurate selection of patients eligible for resection, and although advances in surgical techniques and perioperative management have greatly contributed to reducing the rate of perioperative deaths, stress must be placed on reducing the postoperative complication rates reported to be still as high as 50%. This study was designed to analyze the causes and foreseeable risk factors linked to postoperative morbidity on the grounds of data derived from a single-center surgical population. From September 1989 to March 2005, 287 consecutive patients, affected either with HCC or liver metastasis, had liver resection at our department. Among the HCC series we recorded 98 patients (73.2%) in Child-Pugh class A, 32 (23.8%) in class B and 4 in class C (3%). In 104 colorectal metastases, 71% were due to colon cancer, 25% rectal, 3% sigmoid, and 1% anorectal. In 49 non-colorectal metastases, 22.4% were derived from breast cancer, 63.2% gastrointestinal tumors (excluding colon) and 14.4% other cancers. We performed 80 wedge resections, 77 bisegmentectomies and/or left lobectomies, 74 segmentectomies, 22 major hepatectomies, 20 left hepatectomies, and 14 trisegmentectomies. The in-hospital mortality rate in this series was 4.5%, and the morbidity rate was 47.7%, because of pleural effusion (30%), hepatic abscess (25%), hepatic insufficiency (19%), ascites (10%), hemoperitoneum (10%), or biliary fistula (6%). The variables associated with the technical aspects of the surgical procedure that were responsible for the complications were: a Pringle maneuver length more than 20 minutes (P=0.001); the type of liver resection procedure, including major hepatectomy (P=0.02), left hepatectomy (P=0.04), trisegmentectomy (P=0.04), bisegmentectomy and/or left lobectomy (P=0.04); and a blood transfusion of more than 600 ml (P=0.04). The evaluation of causes and foreseeable risk factors linked to postoperative morbidity during the planning of surgical treatment should play the same role as other factors weighed in the selection of patients eligible for liver resection.