Post-hepatectomy biliary leaks: analysis of risk factors and development of a simplified predictive scoring system.

Abstract

Most studies identifying risk factors for post-hepatectomy biliary leaks (PHBLs) have relatively small proportions of major hepatectomies. A simplified predictive score to identify high risk patients is necessary in order to investigate the efficacy of mitigation strategies. A retrospective analysis of a prospectively maintained database of liver resections from a high-volume cancer center was performed. Multivariate regression was utilized for identification of risk factors and development of the predictive score. A total of 862 patients underwent a curative hepatic resection over 10years, of whom 146 (16.9%) developed a biliary leak; 85 (9.86%), 52 (6.03%), and 9 (1.04%) patients had a grade A, B, and C leak respectively. A biliary-enteric anastomosis [OR 5.1 (95% CI 2.45-10.58); p < 0.001], a central [OR 4.33 (95% CI 1.25-14.95); p = 0.021] or an extended hepatectomy [OR 4.29 (95% CI 1.52-12.12); p = 0.006], liver steatosis [OR 2.28 (95% CI 1.09-4.77); p = 0.027], and blood loss of ≥ 2000mL [OR 2.219 (95% CI 1.15-4.27); p = 0.017] were identified as independent predictors of a clinically significant biliary leak and were assigned 1 point each to develop the biliary leak score. Clinically significant biliary leaks were seen in 11 (2.79%), 20 (6.38%), 19 (15.4%), 9 (56.3%), and 1 (100%) patients with scores of 0, 1, 2, 3, and 4 respectively (p < 0.001). A biliary-enteric anastomosis, a central or extended hepatectomy, liver steatosis, and blood loss ≥ 2L combined result in a simple predictive score for clinically significant biliary leaks.