Background. There is some evidence that epidural analgesia (EDA) reduces tumour recurrence after breast and prostatic cancer surgery. We assessed whether EDA reduces long-term mortality after colorectal cancer surgery. Methods. All patients having colorectal cancer surgery between January 2004 and January 2008 at Linkoping and Orebro were included. Exclusion criteria were: emergency operations, laparoscopic-assisted colorectal resection, and stage 4 cancer. Statistical information was obtained from the Swedish National Register for Deaths. Patients were analysed in two groups: EDA group or patient-controlled analgesia (PCA group) as the primary method of analgesia. Results. A total of 655 patients could be included. All-cause mortality for colorectal cancer (stages 1-3) was 22.7% (colon: 20%, rectal: 26%) after 1-5 yr of surgery. Multivariate regression analysis identified the following statistically significant factors for death after colon cancer (Pless than0.05): age (greater than72 yr) and cancer stage 3 (compared with stage 1). A similar model for rectal cancer found that age (greater than72 yr) and the use of PCA rather than EDA and cancer stages 2 and 3 (compared with stage 1) were associated with a higher risk for death. No significant risk of death was found for colon cancer when comparing EDA with PCA (P=0.23), but a significantly increased risk of death was seen after rectal cancer when PCA was used compared with EDA (P=0.049) [hazards ratio: 0.52 (0.27-1.00)]. Conclusions. We found a reduction in all-cause mortality after rectal but not colon cancer in patients having EDA compared with PCA technique.