Introduction: Concerns of neurological complications following thoracic epidural catheterisation are not supported in the literature1. The benifits of thoracic epidural analgesia2 appear to outweigh the risks. However, there is a paucity of randomized controlled trials comparing two popular techniques for managing severe postoperative pain1. Objectives: The purpose of this study was to demonstrate a difference between Thoracic Epidural Opioid-Local Anaesthetic Analgesia (TEOLAA) and Lumbar Epidural Morphine Analgesia (LEMA) with respect to the duration of post-operative ventilation and the quality of post-operative analgesia in patients having a thoracoabdominal esophagectomy. Methods: This was a randomized controlled, observer blinded study. Twenty two patients scheduled for an elective thoracoabdominal esophagectomy were randomized to a TEOLAA or a LEMA group. The TEOLAA group had a T6-T8 and the LEMA group had a L2-L4 epidural catheter inserted prior to induction of general anesthesia. The TEOLAA group received a combined epidural-general anesthetic. Both groups were limited lo a maximum fentanyl dose of 5 mcg.kg'Mr'. Criteria for trachea! extubation ( VC 10 ml.kg'1; MIP < -30 cm H2O; respiratory rate > 6 and < 20 b.p.m.; oxygen saturation > 95% on FiO2 1.0 ; temperature > 34.5°C and eye opening on command) were assessed by a blinded observer and the time of trachéal extubation was recorded. Postoperatively, the TEOLAA group received PCEA 0.125% bupivicaine and fentanyl Smcg.mT1, with a continuous infusion at 10ml.h"', and a bolus of 9 ml q 30 min to a maximum of 3 boluses in 6 hours. The LEMA group received PCEA morphine 0.2mg.ml'', with a continuous infusion of 0.6 mg.h'1, and bolus of 1.8 mg q 30 min to a maximum of 3 boluses in 6 hours. Early extubation was defined as trachéal extubation within 4 hours postoperatively. Visual analogue pain scores (VAPS) at rest ( static VAPS, SVAPS) and with movement (Dynamic VAPS, DVAPS) were recorded at 1 h, 6 h, 12 h, 18 h and 24 h post-extubation. Failure of the epidural protocol (FEP) was defined as a request by the patient for additional analgesia, despite a functioning epidural catheter as assessed by the response to 10ml 1.5%lidocaine. Results: There were no demographic differences between the groups. Eleven patients in the TEOLAA and 10 patients in the LEMA group were available for analysis. Time to extubation was significantly less in the TEOLAA group (p= 0.01). However, at 4 hours postoperatively, the difference was not significant (p=0.09). The rate of FEP was significantly higher in the LEMA group (p=0.01), with a 50% FEP at 6 hours post-extubaion. There were no FEP in the TEOLAA group. SVAPS and DVAPS were significantly better in the TEOLAA group (p < 0.01). The difference between SVAPS and DVAPS in the TEOLAA group was statistically significant (p = 0.03), but clinically unimportant. However, the differences in the LEMA group were both statistically (p < 0.01) and clinically important. Discussion: Earlier time to extubation and superior pain control was demonstrated in the TEOLAA group. The benefits of earlier extubation and superior pain control need further randomized control trials to evaluate their influence on patient outcome, length of 1CU stay and cost reduction.