Thoracic Epidural Infusion Complicated by Epidural Compartment Syndrome

A 94-yr-old woman incurred severe blunt injuries as a restrained driver in a motor vehicle accident. When she was transferred to our institution, she was spontaneously ventilating and had been placed in full spinal precautions. In the emergency department, the trauma team placed bilateral chest tubes for hemothoraces secondary to multiple, bilateral rib fractures. Despite this, the patient remained hemodynamically stable.

Further examination showed both upper and lower extremity fractures, but results of an examination of the abdomen and spine were normal. The patient's Glasgow Coma Scale score was 15. In the emergency department, she was noted to be able to move all four extremities. Due to impending respiratory insufficiency, the patient was intubated and mechanically ventilated in the emergency department. She was transferred, in critical condition, to the intensive care unit.

On hospital day 2, the Acute Pain Service was consulted and asked to consider placing an epidural catheter to facilitate ventilator weaning. On examination, the patient was able to communicate nonverbally and to move her lower extremities on command. A review of physical examination, laboratory (including coagulation), and radiographic findings failed to produce any contraindications to thoracic epidural catheter placement. Informed proxy consent was obtained from the patient's adult child, and a thoracic epidural catheter was placed.

At the time of catheter placement, the patient was alert and responsive to verbal commands. She reported having difficulty taking deep breaths or coughing, secondary to pain. The T5–T6 vertebral interspace was accessed via  a right paramedian approach, using the standard loss-of-resistance technique with a 17-gauge Tuohy needle. An epidural catheter (Arrow FlexTip^®Catheter; Arrow International, Reading, PA) was advanced easily to a depth of approximately 5 cm. Subsequently, aspiration of the catheter yielded a negative result, and a test dose of 1.5% lidocaine with epinephrine, 1:100,000, was nonreactive. The catheter was then loaded with a bolus dose of 2% lidocaine and fentanyl, resulting in a reduction of reported pain. The patient verified a sensory block to cold temperature (although the exact thoracic dermatomes were indeterminate). She was able to move her toes and flex her left leg; her right leg was splinted. She had a moderate decrease in systolic blood pressure, which responded to ephedrine. An infusion of 0.1% bupivacaine with 10 μg/ml hydromorphone was started at 8 ml/h. Later in the evening, the infusion rate was increased to 10 ml/h to improve pain control. The patient was noted to be able to wiggle her toes on request at that time.

On postprocedure day 2, the patient continued to be weaned from mechanical ventilation while receiving light sedation. The epidural solution continued at an infusion rate of 9 ml/h. Routine examination by the Acute Pain Service was deferred because the patient was sleeping; however, the nursing and trauma teams reported that, when aroused, the patient was oriented and comfortable. At 10:00 pm, the patient's nurse noted that the patient was unable to move her toes on command, and that the epidural infusion was decreased at that time. However, for unknown reasons, the physicians were not notified. The following morning, the nurses noted that the patient was unable to move her legs. They discontinued the infusion, and the Acute Pain Service was notified.

The Acute Pain Service physicians examined the patient and found that she demonstrated flaccid paralysis in the lower extremities and loss of rectal tone. The epidural catheter was aspirated without result and was removed intact. An emergency magnetic resonance image (MRI) of the spine was obtained. Neurosurgical consultation was obtained.

The radiologist interpreted the MRI as demonstrating “extensive posterior epidural complex fluid collection probably representing a mixture of fluid and blood products extending from C6 to T7. The epidural collection significantly displaced the cord anteriorly.” (fig. 1). The patient was taken emergently to the operating room to evacuate the fluid via  laminectomy at the T2–T3 level. Puncture of the ligamentum flavum resulted in a substantial volume of clear fluid squirting out of the epidural space. The fluid was described as clear, with the appearance of saline or aqueous medication, but a specimen was not obtained. The surgeon reported that the fluid was under pressure at the time of evacuation. There was no evidence of a hematoma or abscess. A tracheostomy was performed following the procedure.

Over the next several days, the patient regained increasing sensation in both feet but continued to lack motor function in the lower extremities. She was eventually weaned from the ventilator to a tracheostomy mask. At the time of discharge to a skilled nursing facility, 6 weeks later, the patient had regained only the ability to wiggle her toes. At follow-up 1 yr later, her motor function had improved, but she had not regained full strength in her lower extremities. She remains unable to ambulate.

The spinal epidural space extends from the base of the skull to the sacrococcygeal membrane. It has complicated direct communications with the paravertebral space and indirect communications with the cerebrospinal fluid. When viewed in the anteroposterior radiographic orientation, the epidural space appears to have a saw-toothed shape and has been demonstrated in cadaver studies to be segmented. 6However, this segmentation is not believed to be of major clinical significance in most cases. In normal situations, solution infused into the epidural space is able to egress via  the neuroforaminae.

In this case, the patient's advanced age and concomitant osteoarthritic changes likely prevented the egress of infusate through the neuroforaminae. There was no evidence of the more typical space-occupying lesions responsible for compressive myelopathy, such as a hematoma or abscess, nor was there any radiologic evidence of anterior spinal artery syndrome. The etiology of this patient's compressive myelopathy appears similar to that of a compartment syndrome, in which the pressure within an anatomic space exceeds the capillary pressure required to perfuse adjacent tissue.

Because of this patient's debilitated state, it was difficult to clearly determine her baseline neurologic status and to document changes in that status over the course of her care. It is probable that a more rapid diagnosis of her compressive myelopathy could have reduced this patient's neurologic morbidity. After the patient failed to improve following the reduction in the flow rate of the epidural local anesthetic, a more thorough neurologic evaluation revealed a progressive deterioration in neurologic status. However, because her examination was complicated, the communications required to initiate her neurosurgical therapy were delayed by several hours. Had there been better communication between the nurses and physicians involved in her care and/or more frequent examinations performed by physicians, this delay might have been obviated.

This case demonstrates a previously undescribed clinical diagnosis—epidural compartment syndrome. Because of anatomic considerations, elderly patients may be more at risk for this disorder. Epidural compartment syndrome should enter into the differential diagnosis of acute neurologic deficit in a patient receiving continuous epidural infusion, particularly in patients who are unable to deliver a history. Furthermore, this case stresses the importance of effective nurse–physician communication, especially in light of alterations in neurologic examination findings.