Non-operative treatment of spontaneous spinal epidural hematomas: a review of the literature and a comparison with operative cases

Objective. To identify factors that favour spontaneous recovery in patients who suffered a spontaneous spinal epidural hematoma (SSEH). Methods. The literature was reviewed reguarding non-operative cases of SSEH (SSEHcons). Sixty-two cases from the literature and 2 of our own cases were collected, focusing on sex, age, medical history, position of the hematoma, segmental distribution and length of the hematoma, diagnostic imaging, neurological condition and outcome. Those data were analysed and compared with the data from a literature review of 474 cases operated on because of a SSEH (SSEHoper). Results. The mean length of the hematoma was significantly higher in SSEHcons, compared to SSEHoper (5.4 versus 4.2 vertebral segments; [standard error of the difference (SED) is 0.38 vertebral segments; 95% confidence limits for the difference are 0.45 to 1.95]). Also after exclusion of patients with coagulopathy, mean length of the hematoma was significantly higher in SSEHcons (4.7 versus 3.9 vertebral segments [SED is 0.39 vertebral segment; 95% confidence limits for the difference are 0.04 to 1.56]). Neurological signs and symptoms in SSEHcons were significantly less severe (P<0.005) and diagnosis was based on Magnetic Resonance Imaging (MRI) in the majority of cases (P<0.0005), when compared to SSEHoper. All other patient characteristics showed no correlation with spontaneous recovery. Conclusion. The recent increase of publications of SSEHcons has to be explained by the introduction of MRI in daily medical practice. As a result, more patients with a mild or benign clinical course are being diagnosed. In earlier times those patients would have escaped medical attention. The mean length of the hematoma in SSEHcons appears to be significantly higher compared to SSEHoper. This suggests that spontaneous regression of neurological symptoms may result from decompression of the neural structures by spreading of the (liquid) hematoma along the spinal epidural space in the early stages after haemorrhage. Based on the present review, there appear to be no factors which promote conservative treatment in SSEH. In the majority of cases with SSEH, the mainstay of treatment will remain surgical decompression of the neural structures and removal of the hematoma. The decision for conservative treatment has to be based on the severity of the neurological deficit and on the clinical course. Retrospectively, the length of the hematoma seems to give a clue to the spontaneous recovery which occurs in some cases of SSEH. Nevertheless, hematoma-length can not be used as a guide to treatment.