Hospital variation and outcomes after repeat hepatic resection for colorectal liver metastases: a nationwide cohort study

Michelle R de Graaff,Joost M Klaase,Marcel den Dulk,Wouter W Te Riele,Jeroen Hagendoorn,N Tjarda van Heek,M Vermaas,Eric J Th Belt,Koop Bosscha,Gerrit D Slooter,Wouter K G Leclercq,Mike S L Liem,J Sven D Mieog,Rutger-Jan Swijnenburg,Ronald M van Dam,Cees Verhoef,Koert Kuhlmann,Peter van Duijvendijk,Michael F Gerhards,Paul Gobardhan,Peter van den Boezem,Eric R Manusama,Dirk J Grünhagen,Niels F M Kok,Dutch Hepato Biliary Audit Group, Collaborators,Hans Torrenga
DOI: https://doi.org/10.1016/j.hpb.2024.02.014
Abstract:Background: Approximately 70% of patients with colorectal liver metastases (CRLM) experiences intrahepatic recurrence after initial liver resection. This study assessed outcomes and hospital variation in repeat liver resections (R-LR). Methods: This population-based study included all patients who underwent liver resection for CRLM between 2014 and 2022 in the Netherlands. Overall survival (OS) was collected for patients operated on between 2014 and 2018 by linkage to the insurance database. Results: Data of 7479 liver resections (1391 (18.6%) repeat and 6088 (81.4%) primary) were analysed. Major morbidity and mortality were not different. Factors associated with major morbidity included ASA 3+, major liver resection, extrahepatic disease, and open surgery. Five-year OS after repeat versus primary liver resection was 42.3% versus 44.8%, P = 0.37. Factors associated with worse OS included largest CRLM >5 cm (aHR 1.58, 95% CI: 1.07-2.34, P = 0.023), >3 CRLM (aHR 1.33, 95% CI: 1.00-1.75, P = 0.046), extrahepatic disease (aHR 1.60, 95% CI: 1.25-2.04, P = 0.001), positive tumour margins (aHR 1.42, 95% CI: 1.09-1.85, P = 0.009). Significant hospital variation in performance of R-LR was observed, median 18.9% (8.2% to 33.3%). Conclusion: Significant hospital variation was observed in performance of R-LR in the Netherlands reflecting different treatment decisions upon recurrence. On a population-based level R-LR leads to satisfactory survival.
What problem does this paper attempt to address?