E mail address eivind gottlieb vedi
E-mail address: [email protected] (E. Gottlieb-Vedi).
association is weaker for other gastrointestinal cancer resections . Studies examining how hospital volume influences long-term survival are fewer and the findings are less conclusive, with some studies supporting a prognostic role [11e14], and others not [12,15e20]. Yet, because short-term survival following cancer sur-gery has improved greatly during the last decades, long-term sur-vival has become an increasingly important outcome in surgical research [21e23]. To facilitate the decision-making whether and how to centralize gastrointestinal cancer surgical procedures, there is an advantage of directly comparing the prognosis after these procedures in the same study. Yet, to the best of our knowledge, no original study has assessed hospital volume in relation to long-term survival in all types of gastrointestinal cancers.
This study aimed to assess how hospital volume influences long-term survival after elective surgery in all gastrointestinal cancers in
0748-7983/© 2019 Elsevier Ltd, BASO ~ The Association for Cancer Surgery, and the European Society of Surgical Oncology. All rights reserved. r> Please cite this article as: Gottlieb-Vedi E et al., Annual hospital volume of surgery for gastrointestinal cancer in relation to prognosis, European Journal of Surgical Oncology, https://doi.org/10.1016/j.ejso.2019.03.016
2 E. Gottlieb-Vedi et al. / European Journal of Surgical Oncology xxx (xxxx) xxx
a well-defined population with long and complete follow-up. The H 89 2HCl was a prognostic benefit of higher annual hospital vol-ume of surgery for more complex procedures (esophageal and pancreatic resections) and a smaller benefit for other procedures.
Materials and methods
This was a nationwide Swedish population-based cohort study examining how annual hospital volume (exposure) influences disease-specific 5-year mortality (main outcome) and all-cause 5-year mortality (secondary outcome) in patients who have under-gone surgery for any of eight gastrointestinal cancer locations separately. The cancer locations belonged to three main groups: upper gastrointestinal cancers (esophagus and stomach), hepato-pancreatico-biliary cancers (liver, pancreas, and bile ducts), and lower gastrointestinal cancers (small bowel, colon, and rectum). The operations were conducted between January 1, 2005 and December 31, 2013 and the follow-up ended on December 31, 2016 for disease-specific mortality and on April 30, 2018 for all-cause mortality (latest available updates). Thus, not all patients were followed up for 5 years. Data were retrieved from high-quality nationwide Swedish health data registries budding contain informa-tion on cancer, other diagnoses, surgical procedures, and mortality (described below). Each participant's data were linked between registries using the personal identity numbers assigned to each Swedish resident at birth or immigration . The study was approved by the Regional Ethical Review Board in Stockholm, Sweden (2015/1916-31/1).
The Swedish Cancer Registry was used to identify all patients with a gastrointestinal cancer diagnosis in 2005e2013 according to the 7th edition of the International Classification of Diseases (ICD-
7) (Supplementary Table 1) . Only histologically verified inva-sive carcinomas were included (Supplementary Table 2). Patho-logical tumor stage data were recorded according to the TNM classification of the Union for International Cancer Control (UICC), and were available from 2005 onwards (Supplementary Table 2) . Validation studies have shown the Cancer Registry to be 96% complete for newly diagnosed cancers , and tumor stage data to be above 98% accurate for operated esophageal cancer .
The Swedish National Patient Registry provided data on all elective surgical resections conducted in 2005e2013 among gastrointestinal cancer patients. The surgical procedures were defined according to the Swedish version of the Nordic Medico-Statistical Committee Classification of Surgical Procedures (NCSP-S) Version 1.9 (Supplementary Table 1) . Information regarding operating hospital, year of surgery, and medical comorbidities (according to the most recent and well-validated Charlson Co-morbidity Index scoring system) was also retrieved [30,31]. The Patient Registry has 100% positive predictive values for esophageal cancer surgery , and between 85 and 95% for comorbidities .
The Swedish Cause of Death Registry provided information about date and causes of death in the cohort until the dates pre-sented above (Design) . The Cause of Death Registry is above 98% complete for causes of death and 100% complete for date of death .