Abdominal mesenteric traction (MT) results in decreased mean arterial pressure (MAP), systemic vascular resistance (SVR) and increased cardiac output (CO). This response is induced by a considerable release of prostacyclin (PGI2). Precipitous falls in systemic arterial pressure related to central and/or autonomic nervous reflex arcs also have been described during operations on the upper abdominal viscera. Those hypotensive responses to visceral traction appear to be transmitted along afferent fibres contained within the splanchnic nerves. We investigated the influence of supplementary thoracic epidural anaesthesia on mesenteric traction response during major abdominal surgery. METHODS. With the approval of the Human Investigation Review Board we studied 40 patients scheduled for major abdominal surgery (infrarenal aortic, gastrointestinal and pancreatic surgery) according to a prospective, randomized double-blinded protocol. Patients were randomized to two different anaesthetic regimens. Patients in group 1 received general anaesthesia (GA n = 20) with 0.1-0.15 mg/kg midazolam and 10 micrograms/kg fentanyl prior to skin incision. Maintenance included 65% nitric oxide in oxygen and 0.1 mg increments of fentanyl as required. Group 2 patients (EA n = 20) underwent a combined technique of dose-reduced general anaesthesia and supplementary continuous, thoracic epidural anaesthesia (bupivacaine 0.25%, sensory blockade T4 to L1-3). In both anaesthesia groups ibuprofen (400 mg i.v.) or a placebo equivalent was administered 15 min before the induction of anaesthesia. MT was applied in a uniform fashion. Baseline values preceded the incision of the peritoneum. Further assessments followed 5, 15 and 30 min after MT. The plasma concentrations of 6-keto-PGF1 alpha (stable metabolite of PGI2), TXB2 (stable metabolite of thromboxane), PGF2 alpha, KH2-PGF2 alpha (stable metabolite of PGF2 alpha) were determined by radioimmunoassay. At all assessments we recorded systolic and diastolic blood pressure, heart rate and measured arterial blood gases. Statistical analyses were performed using three-factor ANOVA for repeated measurements after log(x) transformation. A P-value of less than 0.05 was considered significant when the Bonferroni-Holm adjustment was applied. RESULTS. Patients with supplementary epidural anaesthesia demonstrated lower systolic (P = 0.0001) and diastolic (P = 0.006) blood pressure than those in the GA group. Nevertheless, in untreated patients in the EA and GA group there was a significant decrease of about 20-30% in systolic and diastolic blood pressure (P = 0.0001) after mesenteric traction. Irrespective of the anaesthetic procedure, paO2 (P = 0.0001) decreased after mesenteric traction in the placebo group. The control patients in the GA group exhibited a more pronounced increase in heart rate after MT. After traction on the mesentery a significant 20- to 30-fold increase in 6-keto-PGF1 alpha plasma concentrations occurred in the placebo group: GA group 1950/58 (5 min), 1574/59 (15 min) 858/66 (30 min) ng/l, P < 0.0001; EA group: 2002/106 (5 min), 2955/107 (15 min) 1807/70 (30 min) ng/l, P < 0.0001, for placebo vs ibuprofen. There was no statistically significant difference between the two anaesthetic procedures used. In ibuprofen-pretreated patients haemodynamics and paO2 values were stable, while 6-keto-PGF1 alpha plasma concentrations remained within the normal range. CONCLUSION. Our data clearly indicate that the mesenteric traction response consists in relevant haemodynamic alterations and a significant decrease of paO2. Stable haemodynamics and paO2 following cyclooxygenase inhibition signify an action mediated by prostacyclin. Deafferentation of the splanchnic nerves by supplementary thoracic epidural anaesthesia did not influence either prostacyclin release or the decrease in blood pressure and paO2 after traction on the mesentery root...