A review of our recent experience with spinal epidural abscess (SEA) reveals several important changing concepts in diagnosis, etiology, management, and outcome. All cases of SEA seen by our service from August 1975 to July 1989 were reviewed retrospectively, and 29 patients were identified (19 men and 10 women, aged 13-78 years). Abscesses were located in the lumbar (n = 21), thoracic (n = 7), and cervical (n = 1) epidural spaces. Gram-positive organisms were the infectious agent in 72% of the cases, and Staphyloccus aureus was the sole agent in 45% other agents were Gram-negative aerobes (n = 5), a Gram-negative anaerobe (n = 1), Mycobacterium tuberculosis(n = 1), and Sporotrichum schenckii(n = 1), the last occurring in a young woman with acquired immune deficiency syndrome. Seventeen patients had concomitant extraspinal infections. Diagnosis was confirmed by myelography, computed tomography, or magnetic resonance imaging. All patients underwent operative decompression and debridement; 2 required a second procedure for persistent infection. The most common operative findings were pus and granulation tissue in the epidural space (52%); the preoperative course correlated poorly with operative findings. The wound was closed primarily in 84% of cases. Postoperative intravenous antibiotic courses for the bacterial abscesses ranged from 1.5 to 6 weeks (median, 2 weeks), followed by antibiotics given orally for 0 to 6 weeks. Two patients died perioperatively. Neurological outcome was good in 21 patients and fair in 6 (mean follow-up, 1.4 years). Over the last 50 years the spectrum of organisms causing SEA has broadened, and the distinction between acute and chronic SEAs has minimal clinical significance. In addition, magnetic resonance imaging has come to have an adjunctive diagnostic role. Treatment by operative debridement, primary wound closure, and short courses of antibiotics given intravenously and orally has a consistently good result, and prognosis has markedly improved.

Reports differ on which nerve fibers are affected by radiofrequency lesions made in peripheral nerves, some stating that primarily the myelinated delta and unmyelinated C fibers are destroyed, others stating that the destruction affects all sizes of nerve fibers and both myelinated and unmyelinated fibers. This study was designed to confirm one of those two findings, and to study the role that different temperatures might play in determining which fibers are affected. Radiofrequency lesions (85 degrees C for 2 minutes) were made in dogs by placing a temperature-monitored electrode into the lumber intervertebral foramina. The dogs were killed at intervals up to 6 weeks after rhizotomy, and the lesions were studied by light and electron microscopy. In all lesions, there was a total loss of unmyelinated fibers and a nearly total loss of myelinated fibers. In other dogs, 2-minute lesions were made at 45 degrees, 55 degrees, 65 degrees, and 75 degrees C, and the lesions examined 1 week later. Again, all sizes and all types of fibers were destroyed.

There is general agreement that aggressive management and monitoring of the patient with closed head injury with control of intracranial pressure (ICP) will improve patient survival and eventual outcome. Conversely, there is little agreement on the value of surgical craniectomy for increasing intracranial volume and subsequently decreasing ICP in these same patients. This study examines 115 patients with severe closed head injuries (Glasgow Coma Score 8 or less) seen at the North Carolina Baptist Hospital between July 1, 1983, and April 1, 1987. All 115 patients were started on a regimen of head elevation, fluid restriction, chemoparalysis, and hyperventilation at PCO2 25-30 torr. Fifty-seven patients failed to respond to that therapy and were given mannitol. Twenty-seven of these still failed to respond; 24 were placed in a pentobarbital coma therapy group and 3 underwent subtemporal decompression. Of the 24 patients in pentobarbital coma, 17 failed to respond, 7 of whom underwent subtemporal decompression and 10 of whom were not operated on. Of all 10 patients undergoing subtemporal decompression, 7 (70%) responded with an average reduction in ICP of 34% (+/- 19.5% SD). Of the 10, 4 died (40%), in contrast with a mortality of 82.4% among patients in pentobarbital coma without subtemporal decompression. These data strongly suggest that subtemporal decompression can be beneficial in patients with medically intractable elevations of ICP.