Operative mortality after hepatic resection: are literature-based rates broadly applicable?

INTRODUCTION: Literature-based data on mortality after hepatectomy may be misleading, as poor outcomes are less likely to be published. The objective of the current study was to compare published vs public, nationally available mortality rates after hepatic resection.
MATERIALS AND METHODS: A systematic MEDLINE review was conducted to identify reports of hepatectomy outcome between January 1998-December 2004. Data were analyzed to calculate literature-based mortality rate and then compared with population-based mortality rate for hepatectomy using the Nationwide Inpatient Sample (NIS) dataset.
RESULTS: Twenty-three publications fulfilled screening criteria. The studies included 7,073 patients who had undergone hepatic resection (46.1% within USA vs 53.9% outside USA). Most patients were male (58.6%) with median age of 56 years. Indications for hepatic resection included hepatocellular carcinoma (47.7%), metastatic disease (34.3%), or other (18.1%). Cirrhosis was present in 23.2% of patients; 46.9% patients underwent either a hemi-hepatectomy or extended resection. The literature-based mortality rate was 3.6% (US centers only, 2.8%). Analysis of NIS revealed 11,429 hepatectomy cases. After controlling for gender, age, extent of hepatectomy, hepatocellular cancer diagnosis, and presence of cirrhosis, the adjusted NIS-based perioperative mortality rate for hepatectomy was 5.6% (95% CI, 5.0-6.2%). The relative mortality after hepatectomy was 1.6-fold higher based on population-based data compared with reports from the literature (P<0.05).
CONCLUSION: Actual population-based mortality rates for major liver resections may be higher than those reported in the literature. Informed consent should reflect actual local and national mortality rates rather than selective reports from the literature.