A prospective, randomised study of 82 patients having postoperative epidural analgesia was performed to determine whether the tunnelling of an epidural catheter influences its migration. Tunnelling of the catheter subcutaneously for a distance of 5 cm reduced the incidence of inward migration of 1 cm or more (p < 0.01) compared to a standard method of fixation with a transparent adhesive dressing. This effect was more marked if the epidural catheter was sited in the thoracic rather than the lumbar area. Sixty two percent (n = 26) of tunnelled catheters remained within 0.5 cm of their original position compared to 38% (n = 16) of non-tunnelled catheters, although this difference was not statistically significant. Outward catheter migration was not reduced by subcutaneous tunnelling.