Spontaneous Spinal Subdural Hematoma of Intracranial Origin
Spontaneous Spinal Subdural Hematoma of Intracranial Origin
Nontraumatic spontaneous spinal SDHs are rare and can cause compression of the spinal cord or cauda equina. Cases have been reported in ages 1 through 79 years, with almost half presenting between ages 45 and 60 years. There is no sex difference, and most occur in the thoracolumbar and lumbar regions. More than half of these cases are due to anticoagulant use and iatrogenic procedures, such as lumbar punctures and peridural anesthesia. However, in a study done by Domenicucci et al., a significant number of patients had other predisposing factors: 19% had disorders of hemostasis, such as leukemia, hemophilia, thrombocytopenia, cryoglobulinemia, and polycythemia; 4% had an intradural tumor; 4% had a ruptured arteriovenous malformation; and 14% had no identifiable cause. The exact mechanism of nontraumatic spontaneous spinal SDHs is unclear because the spinal subdural space is relatively avascular and lacks bridging veins found in the intracranial space.
Two origins of hemorrhage in the spinal subdural space have been proposed. In the first theory, rapid changes in intra-abdominal and thoracic pressures in combination with a lag in cerebrospinal fluid pressure generates a pressure gradient that causes subarachnoid vessels to rupture and dissect into the subdural space. Blood in the subarachnoid space is then cleared by cerebrospinal fluid, leaving only blood in the subdural space. The risk of vessel rupture and dissection is further increased with iatrogenic procedures, disorders of hemostasis, and use of anticoagulants. In the second theory, intracranial hemorrhage in the subdural space extends and migrates as a result of gravity to the spinal subdural space, causing compression of the spinal cord. Continuity of these two spaces has been demonstrated by postmortem studies in which injection of air into the subdural spinal space was found later in the subdural intracranial space. This continuity is further supported by MRI evidence of hemorrhagic collections connecting these two spaces. Our patient reported no trauma or invasive procedures, was not on anticoagulation, and his laboratory values did not reveal any hemostatic abnormalities. Therefore, we postulate migration of intracranial subdural blood into the lumbar spine led to the spinal SDH.
Spinal SDHs present first as acute, severe localized back pain followed by radicular pain radiating to the trunk or upper or lower extremities. A review of case reports reveals that the radicular pain is most commonly bilateral. Within hours or, less commonly, days, varying degrees of neurologic symptoms develop, including paresthesias, sensory loss, weakness, and cauda equina compression causing bowel or bladder dysfunction and paraplegia. In traumatic spinal SDHs, symptoms develop immediately after the injury. However, in case reports describing spontaneous spinal SDH concomitant with intracranial SDH, the time to development of symptoms ranges from a few days to 1 week for acute spinal SDH and up to 3 weeks for chronic spinal SDH. We hypothesize this delay may be the time period during which the effect of gravity causes intracranial blood to migrate to the more gravity-dependent thoracolumbar spinal subdural space.
Nontraumatic spontaneous intracranial SDHs are most commonly caused by bleeding due to shearing forces on bridging veins or venous sinuses. Patients with cerebral atrophy, low intracranial pressure, coagulopathies, or alcoholic hepatopathy are at increased risk of bridging vein rupture. Other less common causes include aneurysm or arteriovenous malformation rupture, hypertensive hemorrhage into the subdural space, dura mater tumor hemorrhage, and cortical artery rupture.
Both intracranial and spinal SDHs are surgical emergencies. MRI is the recommended method of imaging for spinal SDH, as it more accurately evaluates the size and extent of hemorrhage and differentiates subdural from epidural hematoma. Non-contrast head CT is recommended for intracranial SDH, although brain MRI can also evaluate presence of SDH. Patients with large or acute intracranial SDH or with progressive neurologic deterioration should be treated with craniotomy and evacuation of the hematoma. Patients with spinal SDH who have stable, minimal sensorimotor deficit without cauda equina symptoms can be conservatively managed, while surgical decompression is recommended for those with progressive and severe sensorimotor impairment and cauda equina symptoms. Spinal SDH outcomes are related to the severity, location, and timing of surgical decompression. Prognosis is poor for severe neurologic deficit, such as paralysis, for cervical or thoracic lesions, and for symptoms longer than 3 months. Prognosis is good for minimal neurologic impairment, for lumbar lesions, and when there is surgical intervention before permanent paralysis. There are several case reports of concomitant intracranial and spinal SDHs with stable neurologic examinations managed conservatively with good recovery, including our patient.
A review of the literature indicates that most patients with concomitant intracranial and spinal SDH have symptoms of both headache and back pain. However, it might be possible that we are underestimating the incidence of this condition if imaging studies are done of only the head or the spine. Wong et al. presented a case report of a patient who had intracranial SDH but an asymptomatic spinal SDH found on screening MRI. Kokubo et al. performed spinal MRI screening on 168 patients with intracranial SDH and found that 2 patients had concomitant intracranial and spinal SDH. Both had direct head and spinal trauma and neither had neurologic deficits. The case report and study suggest that screening for spinal SDH in patients with intracranial SDH in patients without back pain or sensorimotor deficits would likely not change management, as these asymptomatic SDHs would most likely do well with observation. Although asymptomatic spinal SDH can exist with intracranial SDH, the management of these patients without any symptoms of back pain or sensorimotor deficits would still be observation. However, if a patient has suspected or confirmed intracranial SDH and has developed new-onset back pain or any neurologic deficits, then MRI imaging of the spine should be considered to exclude the diagnosis of spinal SDH. Symptomatic patients might progress and need screening for hemostatic abnormalities and interventions such as surgical decompression and anticoagulant reversal.
Conversely, we can also ask whether finding a spinal SDH necessitates imaging of the brain. No recommendations exist regarding this question. Some case reports describe patients in whom the spinal SDH was found before the intracranial SDH. Imaging of the brain was prompted by patient confusion or complaints of increased headache, nausea, or vomiting. In two of these cases, brain imaging might have been delayed because of the chronic nature of headache and lack of confusion or cranial nerve deficits. In all cases, the intracranial SDH was treated with craniotomy. Therefore, when you diagnose or are suspicious of spinal SDH, it seems prudent to identify if the patient also has headache, nausea, or vomiting. If intracranial symptoms are present, CT or MRI should be included in the workup. In patients with spinal SDH but without intracranial symptoms, there are no recommendations regarding brain imaging. Jibu et al. describe a patient who had spinal SDH with an asymptomatic intracranial SDH found on screening brain MRI. However, this patient's intracranial SDH was managed conservatively and the patient did well.
In patients with intracranial SDH, mortality rates are higher for those on anticoagulation than those who are not. For such patients, a panel of experts on stroke, neurologic intensive care, and hematology recommend urgent reversal of warfarin, but do not come to a consensus as to the best reversal agent. Three experts recommend prothrombin complex concentrates (PCC) only, two recommend recombinant factor VIIa (rFVIIa) only, one expert recommends rFVIIa with fresh frozen plasma (FFP), and one expert recommends either PCC or FFP. There are no randomized controlled trials evaluating the clinical outcomes of different reversal strategies for patients on warfarin with intracranial hemorrhage. Vitamin K should be given as well, although it does not normalize the international normalized ratio (INR) for 6 to 24 h. FFP can also take several hours to normalize the INR in addition to requiring 2 to 4 L volume for full reversal. PCCs normalize INR faster than FFP, but there are no data showing an effect on patient outcomes. Finally, rFVIIa has been shown to decrease the INR in minutes. However, rFVIIa is expensive, has a short half-life, and, after its administration, the INR might not be accurately followed to guide treatment. Given that even small amounts of blood can cause permanent neurologic injury in the spinal subdural space, most reported cases of patients with symptomatic spinal SDH on anticoagulation also received anticoagulation reversal.
Discussion
Nontraumatic spontaneous spinal SDHs are rare and can cause compression of the spinal cord or cauda equina. Cases have been reported in ages 1 through 79 years, with almost half presenting between ages 45 and 60 years. There is no sex difference, and most occur in the thoracolumbar and lumbar regions. More than half of these cases are due to anticoagulant use and iatrogenic procedures, such as lumbar punctures and peridural anesthesia. However, in a study done by Domenicucci et al., a significant number of patients had other predisposing factors: 19% had disorders of hemostasis, such as leukemia, hemophilia, thrombocytopenia, cryoglobulinemia, and polycythemia; 4% had an intradural tumor; 4% had a ruptured arteriovenous malformation; and 14% had no identifiable cause. The exact mechanism of nontraumatic spontaneous spinal SDHs is unclear because the spinal subdural space is relatively avascular and lacks bridging veins found in the intracranial space.
Two origins of hemorrhage in the spinal subdural space have been proposed. In the first theory, rapid changes in intra-abdominal and thoracic pressures in combination with a lag in cerebrospinal fluid pressure generates a pressure gradient that causes subarachnoid vessels to rupture and dissect into the subdural space. Blood in the subarachnoid space is then cleared by cerebrospinal fluid, leaving only blood in the subdural space. The risk of vessel rupture and dissection is further increased with iatrogenic procedures, disorders of hemostasis, and use of anticoagulants. In the second theory, intracranial hemorrhage in the subdural space extends and migrates as a result of gravity to the spinal subdural space, causing compression of the spinal cord. Continuity of these two spaces has been demonstrated by postmortem studies in which injection of air into the subdural spinal space was found later in the subdural intracranial space. This continuity is further supported by MRI evidence of hemorrhagic collections connecting these two spaces. Our patient reported no trauma or invasive procedures, was not on anticoagulation, and his laboratory values did not reveal any hemostatic abnormalities. Therefore, we postulate migration of intracranial subdural blood into the lumbar spine led to the spinal SDH.
Spinal SDHs present first as acute, severe localized back pain followed by radicular pain radiating to the trunk or upper or lower extremities. A review of case reports reveals that the radicular pain is most commonly bilateral. Within hours or, less commonly, days, varying degrees of neurologic symptoms develop, including paresthesias, sensory loss, weakness, and cauda equina compression causing bowel or bladder dysfunction and paraplegia. In traumatic spinal SDHs, symptoms develop immediately after the injury. However, in case reports describing spontaneous spinal SDH concomitant with intracranial SDH, the time to development of symptoms ranges from a few days to 1 week for acute spinal SDH and up to 3 weeks for chronic spinal SDH. We hypothesize this delay may be the time period during which the effect of gravity causes intracranial blood to migrate to the more gravity-dependent thoracolumbar spinal subdural space.
Nontraumatic spontaneous intracranial SDHs are most commonly caused by bleeding due to shearing forces on bridging veins or venous sinuses. Patients with cerebral atrophy, low intracranial pressure, coagulopathies, or alcoholic hepatopathy are at increased risk of bridging vein rupture. Other less common causes include aneurysm or arteriovenous malformation rupture, hypertensive hemorrhage into the subdural space, dura mater tumor hemorrhage, and cortical artery rupture.
Both intracranial and spinal SDHs are surgical emergencies. MRI is the recommended method of imaging for spinal SDH, as it more accurately evaluates the size and extent of hemorrhage and differentiates subdural from epidural hematoma. Non-contrast head CT is recommended for intracranial SDH, although brain MRI can also evaluate presence of SDH. Patients with large or acute intracranial SDH or with progressive neurologic deterioration should be treated with craniotomy and evacuation of the hematoma. Patients with spinal SDH who have stable, minimal sensorimotor deficit without cauda equina symptoms can be conservatively managed, while surgical decompression is recommended for those with progressive and severe sensorimotor impairment and cauda equina symptoms. Spinal SDH outcomes are related to the severity, location, and timing of surgical decompression. Prognosis is poor for severe neurologic deficit, such as paralysis, for cervical or thoracic lesions, and for symptoms longer than 3 months. Prognosis is good for minimal neurologic impairment, for lumbar lesions, and when there is surgical intervention before permanent paralysis. There are several case reports of concomitant intracranial and spinal SDHs with stable neurologic examinations managed conservatively with good recovery, including our patient.
A review of the literature indicates that most patients with concomitant intracranial and spinal SDH have symptoms of both headache and back pain. However, it might be possible that we are underestimating the incidence of this condition if imaging studies are done of only the head or the spine. Wong et al. presented a case report of a patient who had intracranial SDH but an asymptomatic spinal SDH found on screening MRI. Kokubo et al. performed spinal MRI screening on 168 patients with intracranial SDH and found that 2 patients had concomitant intracranial and spinal SDH. Both had direct head and spinal trauma and neither had neurologic deficits. The case report and study suggest that screening for spinal SDH in patients with intracranial SDH in patients without back pain or sensorimotor deficits would likely not change management, as these asymptomatic SDHs would most likely do well with observation. Although asymptomatic spinal SDH can exist with intracranial SDH, the management of these patients without any symptoms of back pain or sensorimotor deficits would still be observation. However, if a patient has suspected or confirmed intracranial SDH and has developed new-onset back pain or any neurologic deficits, then MRI imaging of the spine should be considered to exclude the diagnosis of spinal SDH. Symptomatic patients might progress and need screening for hemostatic abnormalities and interventions such as surgical decompression and anticoagulant reversal.
Conversely, we can also ask whether finding a spinal SDH necessitates imaging of the brain. No recommendations exist regarding this question. Some case reports describe patients in whom the spinal SDH was found before the intracranial SDH. Imaging of the brain was prompted by patient confusion or complaints of increased headache, nausea, or vomiting. In two of these cases, brain imaging might have been delayed because of the chronic nature of headache and lack of confusion or cranial nerve deficits. In all cases, the intracranial SDH was treated with craniotomy. Therefore, when you diagnose or are suspicious of spinal SDH, it seems prudent to identify if the patient also has headache, nausea, or vomiting. If intracranial symptoms are present, CT or MRI should be included in the workup. In patients with spinal SDH but without intracranial symptoms, there are no recommendations regarding brain imaging. Jibu et al. describe a patient who had spinal SDH with an asymptomatic intracranial SDH found on screening brain MRI. However, this patient's intracranial SDH was managed conservatively and the patient did well.
In patients with intracranial SDH, mortality rates are higher for those on anticoagulation than those who are not. For such patients, a panel of experts on stroke, neurologic intensive care, and hematology recommend urgent reversal of warfarin, but do not come to a consensus as to the best reversal agent. Three experts recommend prothrombin complex concentrates (PCC) only, two recommend recombinant factor VIIa (rFVIIa) only, one expert recommends rFVIIa with fresh frozen plasma (FFP), and one expert recommends either PCC or FFP. There are no randomized controlled trials evaluating the clinical outcomes of different reversal strategies for patients on warfarin with intracranial hemorrhage. Vitamin K should be given as well, although it does not normalize the international normalized ratio (INR) for 6 to 24 h. FFP can also take several hours to normalize the INR in addition to requiring 2 to 4 L volume for full reversal. PCCs normalize INR faster than FFP, but there are no data showing an effect on patient outcomes. Finally, rFVIIa has been shown to decrease the INR in minutes. However, rFVIIa is expensive, has a short half-life, and, after its administration, the INR might not be accurately followed to guide treatment. Given that even small amounts of blood can cause permanent neurologic injury in the spinal subdural space, most reported cases of patients with symptomatic spinal SDH on anticoagulation also received anticoagulation reversal.
