Epidural thoracic fentanyl infusions provide effective preoperative analgesia after thoracotomy; however, side effects can limit the effectiveness of this technique. This study evaluates epinephrine as an adjunct to continuous thoracic epidural fentanyl infusions after thoracotomy. Thirty-eight patients were studied in a prospective, randomized, double-blind trial comparing fentanyl alone to fentanyl with epinephrine 1:300,000. Epidural infusion rates were titrated to equivalent pain relief using a visual analog scale. With the addition of epinephrine, there was a significant reduction in fentanyl requirements (0.82 +/- 0.07 vs 1.19 +/- 0.11 micrograms.kg-1.h-1, P = 0.005, repeated-measures analysis of variance) and in plasma fentanyl concentrations (steady state: 0.91 +/- 0.13 vs 1.65 +/- 0.23 ng/mL, P = 0.007, repeated-measures analysis of variance). There were no differences in pain scores, side effects, spirometry, patient satisfaction scores, or hemodynamic variables. This study demonstrates that adding epinephrine 1:300,000 to continuous thoracic epidural infusions decreases fentanyl requirements titrated for effective analgesia. The reduction in fentanyl requirements was associated with reduced fentanyl plasma concentrations.