Perioperative nutritional support in patients undergoing hepatectomy for hepatocellular carcinoma.

This prospective, randomized, controlled trial from the University of Hong Kong evaluated the efficacy of perioperative parenteral nutrition (PN) in patients requiring hepatectomy for primary hepatocellular carcinoma. From September 1990 through June 1993, 150 consecutive patients with resectable hepatocellular carcinoma were randomly assigned to receive either perioperative PN (n = 75), in addition to usual oral diet, or to no additional therapy (oral diet alone without PN; n = 75). Excluding patients with metastatic disease, a total of 64 patients in the perioperative PN group (39 with associated cirrhosis, 18 with chronic active hepatitis, and 7 without associated liver disease) were compared to 60 control patients (33 with cirrhosis, 12 with chronic active hepatitis and 15 with no associated liver disease). PN was started 7 days before hepatic resection and continued for 7 days after operation in the experimental patients. The PN consisted of standard micronutrients, dextrose, lipid emulsion (containing 50 percent of lipid as medium-chain triglycerides, MCT) and amino acids enriched in branched-chain amino acids (BCAA, 35 percent of PN protein intake), and provided = 1.5 g protein/kg/day and 30 kcal/kg/day. PN was administered via a superior vena cava Broviac catheter cycled over 12 hours each evening preoperatively, and as a 24 hour infusion during the postoperative week. Control patients received only 5 percent dextrose in normal saline postoperatively, with volume and sodium content similar to the experimental PN-treated patients. All patients studied (experimental and control) received 25 grams of albumin intravenously for 5 days postoperatively, and all were allowed to consume enteral diet as tolerated throughout the entire study period. Preoperative assessment included standard anthropometric indices, serum chemistries and proteins, indocyanine green clearance (an index of hepatic function), hand grip strength, and immune function tests (serum immunoglobulin concentrations and peripheral lymphocyte stimulation by phytohemagglutinin). Postoperative assessment included the same preoperative indices (chemistries measured from days 1 to 8 post-operatively), and overall postoperative mortality and morbidity during the hospitalization. Morbidity indices included both infectious complications and non-infectious complications (eg, pleural effusion, ascites, renal failure, hepatic coma). The two groups of patients were similar in age, sex, total and percent weight loss, hepatic carcinoma stage, incidence of cirrhosis, and other preoperative indices. However, a higher percentage of patients in the PN group had abnormal preoperative hepatic function by indocyanine green clearance (67 vs 47%, p = 0.03). The proportion undergoing major hepatectomy and other important intraoperative factors were similar between groups. No significant difference in postoperative hospital mortality occurred between groups (PN 8% vs control 15%; p = 0.30), and PN use did not change hand-grip strength, skin-fold thickness or midarm circumference. However, a significant beneficial effect of PN on hospital morbidity was observed Perioperative PN use was associated with a significant reduction in the overall postoperative morbidity rate (PN group 34% vs control group 55%; p = 0.02). This difference was mainly due to a significant reduction in infectious complications (PN 17% vs control 37%; p = 0.01), and in the need for diuretic drugs to control ascites (PN 25% vs control 50%; p = 0.004). There were no differences between groups in serum immunoglobulins or lymphocyte response to mitogens. There was less deterioration of liver function with PN as measured by the change in the rate of indocyanine green clearance (PN group -2.8% loss vs control group -4.8% loss; p = 0.05). The attenuation of hepatic function loss with PN occurred despite a significant rise in serum transaminase values from days 5 to 8 postoperatively. PN therapy was also associated with le