Analgesic management of postoperative pain associated with thoracic surgery remains a difficult clinical challenge. In the present study we used a thoracic muscle incision model to characterize pain-related behavior and changes in prostaglandin E2 (PGE2) in both thoracic cerebrospinal fluid (CSF) and incision site tissues. A deep muscle incision was made in the left thoracic region of rats anesthetized with isoflurane, propofol, or spinal bupivacaine. Thoracic CSF and incision site tissue concentrations of PGE2 were monitored for 6 h using microdialysis loop catheters. Postoperative pain-related behavior was assessed by recording exploratory locomotive activity. Thoracic muscle surgery decreased rearing and ambulation. Oral ketorolac or rofecoxib 3 mg/kg restored normal rearing and ambulation. Postoperative CSF PGE2 concentration increased most (threefold) with spinal anesthesia, and not at all with propofol. With surgery under isoflurane or spinal bupivacaine, presurgical oral administration of ketorolac or rofecoxib 3 mg/kg reduced postsurgical CSF PGE2 levels and tissue PGE2 levels. Intrathecal ketorolac (4 microg) reduced CSF PGE2 after surgery without affecting tissue PGE2 levels, whereas intrathecal L-745,337 (80 microg) did not reduce CSF PGE2. Thoracic surgical wounds increase pain-related behavior and CSF and tissue PGE2 levels, all of which can be attenuated by oral cyclooxygenase inhibitors.