The first case describing a spinal epidural hematoma was reported by Jackson in 1869 (Jackson 1869). Since then, more than 600 patients receiving surgical decompression of spontaneous spinal epidural hematomas (SSEH) have been reported (Groen and Ponssen 1990) and 112 pediatric cases (Yoneyama et al. 1998). In the last few years, several reports have been published describing the conservative treatment of SSEHs (Sagar and Hassan 2010; Riaz et al. 2008; Ross et al. 1987; Silber 1996; Sokolowski et al. 2008). The increase in the reported SSEHs is most likely due to increased use of MRI in establishing the radiological diagnosis (Oh and Lingaraj 2008).
Most SSEHs present with acute symptoms of severe pain with rapid signs of spinal cord and/or root compression. It is rare that patients present with slow progressive and chronic symptoms or intermittent relapsing radiculopathy that may mimic a prolapsed disc or spinal synovial cyst. In a review of the literature, Groen reported isolated root compression symptoms in 9 % of the patients (Groen 2004). In 67 % of patients, symptoms of spinal cord compression was present, cauda equina symptoms occurred in 2 % of cases while combined spinal cord and cauda equina symptoms were present in 5 % of the reported cases (Garzia et al. 1999).
It has been proposed that the internal vertebral venous plexus is the most likely source of SSEHs (Adamson et al. 2004; Hancock et al. 1997; Groen et al. 1997; Wait et al. 2005). Most hematomas are located posterior to posterolateral in the spinal canal.
The two regions of the spinal cord most predisposed to SSEHs are the cervicothoracic and thoracolumbar junctions. A purported weakness of the epidural venous plexus in these locations has been hypothesized as the reason for this preference (Adamson et al. 2004; Hancock et al. 1997; Jackson 1869; Wait et al. 2005).
In order to explain the more benign clinical course in some patients with SSEH, several mechanisms have been reported. Some argue that immediate replacement therapy in patients with coagulopathy prevents progression of the hematoma, allowing for relief of symptoms and neurological signs (Groen et al. 1997; Riaz et al. 2008). In contrast, Connelly et al. propose that spinal hematomas caused by coagulopathy can be managed conservatively because the hematoma remains liquid for a longer time than with normal clotting, allowing the liquid to dissipate in the spinal epidural space (Kebaish and Awad 2004). The findings in our case are contradictory to the proposal by Connelly et al. (Kebaish and Awad 2004). This may be due to serial bleeding adding more volume to the hematoma over time, which may lead to spinal cord and/or root compression symptoms. Inamasu et al. suggested that a liquefied hematoma may leak through the intervertebral foramen, which would lead to spontaneous decompression of the spinal cord (Wagner et al. 1996; Inamasu et al. 2000). If this was the case, a higher incidence of root compression symptoms would be expected. However, a previous report revealed that less than 10 % of patients diagnosed with SSEH had root symptoms (Oh and Lingaraj 2008). Furthermore, most hematomas are located posterior and not lateral in the spinal cord.
The proposal by Inamasu et al. is however interesting, with respect to our case report, in which L5-radiculopathy was a dominating presenting symptom. Furthermore, the MRI performed on our patient revealed a space occupying structure quite lateral on the left side at level L4–L5 which would likely cause compression of the left L5-root. Additionally, during the operation the organized hematoma was pursued laterally to where the left L5-root entered the intervertebral foramina. The patient also exhibited clinical symptoms of a left L5-root compression and had a positive Lasuegé on neurological examination.
Spreading of a hematoma may be possible in the acute phase following a bleed before blood clotting has occurred or in the subacute to chronic phases when the organised hematoma may, at least partially, become liquefied. Thus, patients on anticoagulation/platelet treatment or with an endogenous coagulopathy may experience a delayed clot formation allowing the liquefied hematoma to spread. However, the same mechanisms that delay clot formation may also promote further bleeding, which increases the space-occupying hematoma.
In a previous report in which patients with coagulopathy were excluded from the study (Kreppel et al. 2003), an analysis of SSEHs conservatively managed was compared to SSEHs surgically decompressed. The mean length of the hematoma along the spinal epidural space was significantly higher in those patients in which conservative management was applied, compared to the operated group (Oh and Lingaraj 2008). This finding may support the hypothesis that the liquefied hematoma may spread in the epidural space. An expanding hematoma may cause a release of the dural filaments crossing the epidural space, a process which is hypothesized to facilitate the spread of the haematoma, which results in a decompression of the dural sac (Kreppel et al. 2003).