Spinal Epidural Hematoma Detected by Lumbar Epidural Puncture

We report a case of an asymptomatic spinal epidural hematoma (SEH) after epidural anesthesia that was detected by a second lumbar epidural puncture when the patient presented for a second vascular procedure.

Case Report

A 75-yr-old, 60-kg man presented with left foot ischemic pain. His past medical history was significant for a previous aortobifemoral bypass, old myocardial infarction, renal insufficiency, and a cerebrovascular accident with residual left lower extremity weakness. Medications included pentoxifyl-line 400 mg three times a day and one coated aspirin (ASA) daily. Laboratory examination was remarkable for blood urea nitrogen levels of 98 mg/dL and creatinine levels 1.7 mg/dL. Prothrombin time, partial thromboplastin time, and platelet count were within normal limits. Arteriography revealed arterial occlusive disease at the level of the popliteal artery trifurcation. The patient was taken to the operating room for a femoral-distal bypass to the posterior tibial artery under epidural anesthesia. Epidural puncture using a loss of resistance to saline technique was performed without difficulty at the L3-4 interspace with an 18-gauge Tuohy needle by an experienced anesthesiologist. Aspiration for blood and cerebrospinal fluid was negative. Aspiration of an epidural catheter placed through the Tuohy needle 3 cm into the epidural space was also negative. Five hours later, 5000 units of heparin was administered intravenously. Protamine reversal of the heparin was not administered ^[1]. Dextran 40, 100 mg, was administered, and a continuous infusion of 20 mL/h was begun 7 h after epidural placement and continued for 72 h. The epidural catheter remained in place for 24 h postoperatively for pain management. ASA therapy was resumed on Postoperative Day 1.

Coagulation variables were within normal limits at the time of removal of the epidural catheter. The patient remained neurologically intact and did not complain of back or radicular pain, paresthesia, or urinary retention. On the 10th postoperative day, he again complained of left foot pain. Repeat arteriography revealed flow-limiting stenosis above the distal anastomosis. The patient was brought to the operating room for revision of the stenotic graft. An epidural anesthetic was once again planned. Coagulation variables were within normal limits. When an 18-gauge Tuohy needle was inserted at the L3-4 level, black, tarry, liquefied blood returned under pressure from the epidural space. After removal of approximately 10 mL of old blood, no additional blood returned passively or with gentle aspiration. The needle was withdrawn, and the anesthetic technique was converted to a spinal anesthetic at the L2-3 interspace. The surgery proceeded uneventfully. Computerized tomography (CT) without contrast was performed postoperatively. The CT scan was equivocal for the presence of epidural hematoma. Subsequent magnetic resonance imaging (MRI) revealed epidural hematoma at interspace L2-4. The patient remained neurologically intact and was discharged on the sixth postoperative day.

Discussion

Asymptomatic SEH is a rare incidental finding at autopsy or laminectomy ^[2-11]. The diagnosis of SEH is usually made from the clinical presentation and radiographic imaging. SEH detected during lumbar epidural puncture has not been previously reported. In our patient, there was no clinical suspicion of SEH, and the diagnosis was not considered until a liquefied clot was evacuated from the epidural space at the time of repeat lumbar epidural puncture.

Spinal epidural hematoma may have a spontaneous, traumatic, idiopathic, or iatrogenic origin (Table 1) ^[12-26]. Compressive SEH with associated neurological deficits is more likely to occur within the cervical or thoracic spinal canal, as the spinal cord occupies a greater percentage of the canal diameter in this region ^[12,26]. Our case illustrates the ability of the lumbar epidural space below L2 to accommodate hemorrhage without producing neurologic symptoms. With the spinal cord normally terminating opposite to the L1-2 disc space, compressive SEH within the lumbar spinal canal below that level can cause a cauda equina syndrome. Injury to the cauda equina can present with back pain progressing to sensory changes, motor weakness, and bowel and bladder deficits ^[12]. It is believed that cauda equina injuries have a better prognosis for neurological recovery than upper spinal levels, but permanent motor and sensory loss can occur if SEH is left untreated ^[27,28].

Table 1: Potential Causes of Spinal Epidural Hematoma

The incidence of SEH in a large case series of 850,000 patients after central neuroaxial blockade (CNB) is estimated to be 1:150,000 ^[29], although that review does not specify whether these patients received anticoagulant therapy. The safe use of epidural anesthesia during anticoagulant therapy has been documented in reports over the past 15 years ^[30,31]. However, in reviews of published cases of SEH involving a CNB technique, the most constant risk factor identified in 60%-80% of patients was impaired coagulation at the time of initiation of CNB or epidural catheter removal ^[30,31]. We chose epidural anesthesia for our patient because of his medical history and to facilitate postoperative pain management. Our patient received ASA therapy prior to initiation of the first epidural anesthetic and a dextran infusion at the time of withdrawal of the epidural catheter. Considered independently, the incidence of SEH after CNB during either ASA or dextran therapy is very low ^[31-35]. In our case, the combination of ASA therapy, heparinization, and dextran therapy may have increased the risk of SEH after epidural anesthesia. Conceivably, our patient's chronic renal insufficiency may have resulted in abnormal platelet function, further contributing to this risk.

Several factors were considered prior to performing a spinal anesthetic for the second operative procedure after a resolving clot was detected during a lumbar epidural puncture. The patient was asymptomatic prior to evacuating the liquefied clot within the epidural space, and perioperative anticoagulation was not anticipated. Furthermore, the greatest risk of producing bleeding into the epidural space occurred at the time of epidural puncture rather than during the less traumatic subarachnoid puncture ^[31]. Postoperative radiographic imaging confirmed that these procedures did not produce acute hemorrhage within the epidural space.

Evidence of the liquefied clot during the lumbar epidural puncture prompted us to seek a radiographic diagnosis. The superiority of MRI in the diagnosis of subacute spinal epidural hematoma compared with CT imaging was evident in this case. Because of its multiplanar imaging capability and its ability to date hematomas by the unique magnetic resonance signal characteristics of various ages of hemorrhage, MRI allows a more specific diagnosis of spinal hematomas ^[36-39]. CT was chosen as the initial imaging method because MRI was unavailable. Axial CT images from L2-3 through L4-5 did not convincingly demonstrate the epidural hematoma, even in retrospect. Since the hematoma was not acute when imaged with CT, the blood no longer appeared dense (white), which diminished its conspicuity (Figure 1). Axial (Figure 2) and parasagittal MRI images revealed increased signal intensity on both T1- (Figure 3a) and T2- (Figure 3b) weighted images within the dorsal epidural compartment from L2 through L4, consistent with an epidural hematoma. A normal MRI image (Figure 4) is included for comparison. The increased signal intensity on both T1- and T2-weighted images is consistent with subacute hematoma ^[26,36], as was suggested by the appearance of the liquefied clot evacuated from the epidural space and the history of a previous lumbar epidural puncture 10 days prior to MRI. In contrast, acute hemorrhage (between 24 and 72 hours) is usually heterogeneously hypointense on weighted images ^[36]. Intravenous administration of gadopentetate dimeglumine (gadolinium) is useful in the evaluation of spinal tumor or infection but is not necessary for the detection of spinal canal hemorrhage ^[40] and was not administered in this case.

Figure 1: Axial computed tomography image at the L2 level demonstrates gas within the epidural space near the site of lumbar puncture.

Figure 2: Axial T1-weighted magnetic resonance image demonstrates dorsolaterally located epidural blood (arrows).

Figure 3: Sagittal T1-(A) and T2-(B) weighted magnetic resonance images demonstrate hyperintense epidural blood (arrows).

Figure 4: "Normal" lumbar spine magnetic resonance image. Sagittal T1-weighted image. At the laminar levels (L), the thecal sac lies against the laminae, with negligible intervening fat signal. Between laminae (arrowheads), the dorsal epidural fat has straight borders and does not appear to bulge focally into the spinal canal. Note midline disc herniations at L4-5 and L5-S1 (arrows).

We chose observant management for this patient with an SEH at the L2-4 region because of the absence of neurological dysfunction and its location below the termination of the spinal cord. When the diagnosis of SEH is made or strongly suspected, immediate neurosurgical consultation should be obtained. An emergent MRI scan followed by surgical evacuation of the SEH is recommended for those patients with a neurological deficit. Neurological recovery is related to the severity of preoperative deficits and the rapidity of surgical decompression ^[13]. The neurosurgical consultant can assist in determining the most appropriate treatment for the neurologically intact patient with a SEH below the level of termination of the spinal cord based on the size and nature of the hematoma.

This case demonstrates several important points: 1) epidural hematomas may be asymptomatic and may occur more often than is realized, 2) MRI is the diagnostic modality of choice, and 3) conservative therapy is effective in patients without neurological deficits.