Spinal epidural abscess (SEA) was first described in the medical literature in 1761 and represents a severe, generally pyogenic infection of the epidural space requiring emergent neurosurgical intervention to avoid permanent neurologic deficits. Spinal epidural abscess comprises 0.2 to 2 cases per 10,000 hospital admissions. This review intends to offer detailed evaluation and a comprehensive meta-analysis of the international literature on SEA between 1954 and 1997, especially of patients who developed it following anesthetic procedures in the spinal canal. In this period, 915 cases of SEA were published. This review is the most comprehensive literature analysis on SEA to date. Most cases of SEA occur in patients aged 30 to 60 years, but the youngest patient was only 10 days old and the oldest was 87. The ratio of men to women was 1:0.56. The most common risk factor was diabetes mellitus, followed by trauma, intravenous drug abuse, and alcoholism. Epidural anesthesia or analgesia had been performed in 5.5% of the patients with SEA. Skin abscesses and furuncles were the most common source of infection. Of the patients, 71% had back pain as the initial symptom and 66% had fever. The second stage of radicular irritation is followed by the third stage, with beginning neurological deficit including muscle weakness and sphincter incontinence as well as sensory deficits. Paralysis (the fourth stage) affected only 34% of the patients. The average leukocyte count was 15,700/microl (range 1,500-42,000/microl), and the average erythrocyte sedimentation rate was 77 mm in the first hour (range 2-50 mm). Spinal epidural abscess is primarily a bacterial infection, and the gram-positive Staphylococcus aureus is its most common causative agent. This is true also for patients who develop SEA following spinal anesthetics. Magnetic resonance imaging (MRI) displays the greatest diagnostic accuracy and is the method of first choice in the diagnostic process. Myelography, commonly used previously to diagnose SEA, is no longer recommended. Lumbar puncture to determine cerebrospinal fluid protein concentrations is not needed for diagnosis and entails the risk of spreading bacteria into the subarachnoid space with consequent meningitis; therefore, it should not be performed. The therapeutic method of choice is laminectomy combined with antibiotics. Conservative treatment alone is justifiable only for specific indications. Laminotomy is a therapeutic alternative for children. The mortality of SEA dropped from 34% in the period of 1954-1960 to 15% in 1991-1997. At the beginning of the twentieth century, almost all patients with SEA died. Parallel to improvements in the mortality rate, today more patients experience complete recovery from SEA. The prognosis of patients who develop SEA following epidural anesthesia or analgesia is not better than that of patients with noniatrogenic SEA, and the mortality rate is also comparable. The essential problem of SEA lies in the necessity of early diagnosis, because only timely treatment is able to av...