Complications after liver surgery: a benchmark analysis

Background: The best achievable short-term outcomes after liver surgery have not been identified. Several factors may influence the post-operative course of patients undergoing hepatectomy increasing the risk of post-operative complications. We sought to identify risk-adjusted benchmark values [BMV] for liver surgery. Methods: The National Surgery Quality Improvement Program (NSQIP) database was used to develop Bayesian models to estimate risk-adjusted BMVs for overall and liver related (post-hepatectomy liver failure [PHLF], biliary leakage [BL]) complications. A separate international multi-institutional database was used to validate the risk-adjusted BMVs. Results: Among the 11,243 patients included in the NSQIP database, the incidence of complications, PHLF, and BL was 36%, 5%, and 8%, respectively. The risk-adjusted BMVs for complication (range, 16-72%), PHLF (range, 1%-20%), and BL (range, 4%-22%) demonstrated a high variability based on patients characteristics. When tested using an international database including nine institutes, the risk-adjusted BMVs for complications ranged from 26% (Institute-4) to 43% (Institute-1), BMVs for PHLF between 3% (Institute-3) and 12% (Institute-5), while BMVs for BL ranged between 5% (Institute-4) and 9% (Institute-7). Conclusions: Multiple factors influence the risk of complications following hepatectomy. Risk-adjusted BMVs are likely much more applicable and appropriate in assessing "acceptable" benchmark outcomes following liver surgery.