I read with interest the commentary by Bradley1 on pancreatic adenocarcinoma. The studies comparing survival rates after resection and following palliative treatments are retrospective in design and imply that the various procedures have been self-selected according to the extent of the disease. Randomized studies should be required, but whenever a complete resection is technically feasible, it would be unethical to perform a palliative procedure. Therefore, multifactorial studies and statistical adjustments are useful to limit bias of nonrandomized retrospective comparisons. We performed such a multicentric study2 that included 3741 patients. The results indicated higher survival rates after resection than after palliative procedures.
Huguier M. Improving the Results of Pancreatoduodenectomy. Arch Surg. 2002;137(12):1447. doi: