In a controlled crossover study, "identical" surgical procedures, the prophylactic removal of bilateral non-erupted 3rd molar teeth, were performed on two separate occasions in 24 healthy patients. Prior to each procedure, either betamethasone 9 mg (Celeston Chronodose) or placebo was administered intramuscularly, in a randomized fashion. Objective and subjective assessments were recorded for paired comparison of the post-operative course, including swelling, pain, trismus, local temperature, bleeding, wound-healing and preference for treatment. In 23 patients, less swelling occurred when betamethasone was given pre-operatively. The mean reduction on the 3rd and 6th post-operative days was 55% (p less than 0.001) and 69% (p less than 0.001), respectively. Pain assessments (visual analogue scale) were significantly lower after the corticosteroid injection; mean response: 1st evening 17 vs 56 mm, 2nd evening 5 vs 37 mm, and 3rd evening 2 vs 13 mm. No significant correlation between the steroid-induced reduction in swelling and pain could be made. This may indicate that dissociation may exist between pain and other inflammatory events like swelling. No clinically apparent infection or other disturbance of wound-healing was noted after corticosteroid administration. This treatment course was preferred by 23 of the 24 patients.