Analysis of patient selection and external validity in the Swedish contribution to the COLOR trial.

Objective: The colon cancer laparoscopic or open resection (COLOR) trial is an international, randomised controlled trial comparing outcomes of open and laparoscopic surgery for colon cancer. The main purpose of this study was to determine representability by comparing included and nonincluded patients in the participating Swedish centres.

Design: At eight centres, which included 391 of the 422 Swedish patients, a local database search was performed to identify retrospectively all patients (n = 2,384) who underwent surgery for colon cancer during the inclusion period, and data was retrieved from medical records.

Results: Four hundred fifty-six patients were randomised, 65 of whom were excluded post randomisation (group 2), leaving 391 patients in the study (group 1). For 1,566 patients, valid exclusion criteria were found (group 3). Thus, 362 patients were eligible but not included (group 4). Relative to group 1, patients in group 4 had a significantly higher American Society of Anaesthesiologists (ASA) score, more advanced tumour stage and difference regarding the resections performed. Results showed that 1470 patients (62%) could be calculated as feasible for laparoscopic colon resection (LCR) in a clinical, nontrial situation.

Conclusions: The study population in the Swedish part of the COLOR trial was representative of the eligible population with the exception of comorbidity, where those actually included had less severe comorbidity than the nonincluded but eligible patients. In Sweden, 50-60% of colon cancer patients can be operated on by laparoscopy.