Diabetes mellitus is no independent risk factor for perioperative mortality following hepatic resection

For patients with concomitant diabetes mellitus an increased perioperative mortality and morbidity in hepatic resections has repeatedly been described. Other studies, however, demonstrated equal outcome data in diabetic and non-diabetic patients. As patient populations were selected for underlying disease, conflicting results may reflect patient selection criteria rather than impact of diabetes mellitus on outcome measures. Therefore, a multivariate analysis in a largely unselected patient population has been performed to determine the independent prognostic value of diabetes mellitus in liver surgery. From a prospective database 633 adult patients undergoing hepatic resection without preceding major abdominal surgery or chemotherapy have been identified. Besides diabetes mellitus, demographic data, variables expressing the functional reserve of the liver, and parameters of surgical technique were analyzed for their impact on mortality and morbidity. 75 patients were diabetic (11.8%) and 96 hepatic resections (15.2%) were performed in cirrhotic patients. In the univariate analysis, concomitant diabetes was associated with an increased mortality compared to all non-diabetic patients (10.7% vs. 5.3%, p = 0.047). Diabetic patients, however, were also significantly older and presented a higher prevalence of liver cirrhosis. Multivariate modeling finally identified only age, albumin, cirrhosis, extent of surgery, and era of surgery as independent variables with an impact on perioperative mortality. Overall, complications were detected in diabetic and nondiabetic patients with a comparable frequency (44% vs. 36%, p = 0.179). Also, the length of in-hospital stay did not significantly differ between both groups (18.5 +/- 1.7 vs. 17.7 +/- 1.0 days, p = 0.119). Rates of postoperative renal impairment, prolonged ascites or pneumonia, however, were higher in diabetics than in other patients. Following established cardiopulmonary and surgical selection criteria, diabetes mellitus is not an independent risk-factor for perioperative mortality in hepatic resections. Although the overall postoperative morbidity was not different in diabetic and non-diabetic patients, a specific pattern of complications has been identified, mandating particular attention in the postoperative course of diabetic patients.