A randomized trial of lumbar epidural injections for the treatment of acute sciatic pain was carried out. All the patients had unilateral sciatica for less than six months and had never previously been treated in hospital for their backs. The treatments used were 20 ml normal saline, 80 mg Depomedrone in normal saline made up to 20 ml, 20 ml, 20 ml 0.25% bupivacaine solution and needling with a standard Touhy injection needle into the interspinous ligament but with no injection. The patients improved at about the same rate for all four treatments. Overall, 75% of the patients improved or were cured. As some of these were in the group treated by needling of the interspinous ligament, it seems likely that the epidural injections administered achieved effects partially as a placebo and partially by virtue of the natural history of the acute sciatic syndrome.