Thoracic spinal cord injury after surgical removal of a ruptured cerebellar arteriovenous malformation in a patient in the Concorde position: illustrative case

Abstract

BACKGROUND

Thoracic spinal cord injury after posterior cranial fossa surgery in younger patients is a rare complication. There have been reports of this complication in tumor and spine fields but not in vascular surgery.

OBSERVATIONS

A 22-year-old-man experienced cerebellar arteriovenous malformation rupture, and the malformation was surgically removed with the man in the Concorde position. After surgery, the man had severe paraplegia, and a thoracic spinal cord injury was diagnosed.

LESSONS

In younger patients, cervical hyperflexion in the Concorde position can cause thoracic spinal cord injury even in surgery for cerebrovascular disease.

There have been reports of thoracic spinal cord injury after posterior cranial fossa surgery in younger patients.^1–6 This injury is a rare complication whose cause is not clearly understood. Among the reported cases, surgery has been conducted for posterior cranial fossa tumors and spinal lesions, but there have been few, if any, reports of such complications in surgery for treating cerebrovascular disease. We report a case of thoracic spinal cord injury after the removal of a ruptured cerebellar arteriovenous malformation (AVM).

Illustrative Case

A 22-year-old man (height 162 cm, weight 48 kg, body mass index 18.2 kg/m²) with no medical history experienced the sudden onset of headache and nausea and was transferred to our hospital. Computed tomography showed a left cerebellar hemorrhage on the dorsal side (Fig. 1A). Cerebral angiography showed a Spetzler-Martin grade I cerebellar diffuse-type nidus AVM comprising a left posterior inferior cerebellar artery feeder and two cortical vein drainers (Fig. 1B and C). The draining veins ran upward and anastomosed with the left transverse sinus. Spinal vessels were not involved. Surgical removal was performed on the third hospital day. After general anesthesia was induced, the patient was set in the Concorde position, and his head was fixed with a Mayfield three-pin clamp. In setting the Concorde position, two finger breadths were kept between chin and neck. The motor evoked potentials (MEPs) were not monitored, because the AVM was located on the dorsal surface of the cerebellum. The AVM was removed via a midline suboccipital approach. Surgery lasted 632 minutes. The diffuse-type nidus was carefully dissected, which was confirmed by intraoperative angiography. The patient’s systolic blood pressure was low at 80–100 mm Hg for most of the anesthesia time. Estimated blood loss was 500 ml, but blood transfusion was not performed, because the patient was not anemic. Normal cerebral vessels around the AVM were not damaged.

FIG. 1.

Head computed tomography showing a cerebellar hemorrhage (A). Cerebral angiography showing a cerebellar AVM. Distal cortical branches of the left posterior inferior cerebellar artery are the feeder (arrow), and two cortical veins are the drainer (arrowhead) and enter the left transverse sinus (B and C).

After surgery, the patient was paraplegic immediately upon awakening. A bilateral sensory disturbance below the T7 level and a neurogenic bladder were seen. Sacral sparing was not observed. These were findings of a thoracic spinal cord injury of American Spinal Injury Association (ASIA) Impairment Scale grade A. Thoracic T2-weighted magnetic resonance imaging (MRI) showed a long hyperintensity signal in the spinal cord (Fig. 2A and B). Follow-up MRI the next day showed a prolonged T2 hyperintensity area, which exhibited diffusion-weighted imaging hyperintensity and apparent diffusion coefficient hypointensity (Fig. 2C–E). Spinal angiography and computed tomography angiography were performed, but no abnormal spinal vessel was detected. The artery of Adamkiewicz originated from the left eighth intercostal artery. We diagnosed a thoracic spinal cord infarction with edema. Steroid pulse therapy, osmotic diuretic therapy, and hyperbaric oxygen therapy were performed. Three weeks after surgery, the T2 hyperintensity on MRI had disappeared. Despite repeated spinal angiography after the disappearance of spinal cord edema, the cause of the thoracic spinal cord injury was not determined. Six months after surgery, there was focal T2 hyperintensity at the T6 level and atrophy of the peripheral spinal cord, and thus it was believed that the T6 level was the ischemic core (Fig. 2F). Although long-term rehabilitation was performed, the spinal cord injury remained at ASIA Impairment Scale grade B.

FIG. 2.

T2-weighted MRI 2 days after surgery showing hyperintensity in the thoracic spinal cord. Hyperintensity was seen mainly in the gray matter (A and B). Diffusion-weighted imaging showing hyperintensity and apparent diffusion coefficient showing hypointensity (arrows) and thus indicating thoracic cord infarction (C–E). T2-weighted imaging showing focal transverse hyperintensity at the T6 vertebral level approximately 6 months after onset (F).

Patient Informed Consent

The necessary patient informed consent was obtained in this study.

Discussion

Observations

We report a case of a thoracic spinal cord injury after the surgical removal of a cerebellar AVM with the patient in the Concorde position. Such a serious complication is rare and generally not well known. Yahanda et al.⁶ reviewed 13 case reports of spinal cord injury after posterior cranial fossa surgery. In their review, the sitting position tended to result in quadriplegia, and the prone position tended to result in paraplegia. Patients in these cases ranged from 4 to 45 years of age. There have been similar complication reports for posterior cranial fossa tumors and spinal disease but few reports for cerebrovascular disease.

Although a definitive cause of the complication was not identified in most cases, cervical hyperflexion due to the positional setting was the mechanism proposed for paralysis. Regarding the present case, we make three comments pertaining to the cause of the thoracic spinal cord injury. First, the long-time invasive operation and venous stasis due to the Concorde position may have advanced hypercoagulation, resulting in spinal cord infarction. Regarding the surgical time, Maduri et al.² reported the same complication for a surgical time of 310 minutes. Our case had a longer surgical time of 632 minutes; therefore, a shorter procedure may have been better. Second, an enlarged spinal epidural venous plexus may have compressed the spinal cord. It is known that the spinal epidural venous plexus is enlarged under high abdominal pressure due to collateral circulation of the inferior vena cava. The pressure within the spinal epidural venous plexus is in equilibrium with the intracranial pressure.⁷ As an example, for the head-down position, when the cranial part of the AVM is dissected, the cerebrospinal fluid is reduced, and the lower intracranial pressure may induce spinal epidural venous congestion.⁸ Third, and we believe most important, cervical hyperflexion in young patients may induce spinal cord stretching and hypoperfusion. The clinical course of the present case is similar to that of surfer’s myelopathy in terms of thoracic spinal cord injury in young patients. Surfer’s myelopathy is a nontraumatic spinal cord injury that is seen mainly in thoracic lesions. Although the cause of surfer’s myelopathy is described as hyperextension in novice surfers, it has also been reported to occur in other cervical extension situations.⁹ Similarly, because hyperextension affects the spinal cord, we think that hyperflexion can cause remote thoracic spinal cord injury. In the Concorde position, because the spinal cord is arched by cervical flexion and the body position, spinal hypoperfusion is likely to occur at the most stretched lesion, which would be a lesion of the upper thoracic spinal cord. A cervical spinal cord injury sustained in the sitting position may have a similar mechanism.⁶ In addition, the patient in the present case was short in stature. We consider that a short stature and thus a short spinal cord may increase the likelihood of spinal stretching. Because discs, cervical ligaments, and other tissues of younger patients may be more stretchable than those of older patients, the complication may occur only in younger patients. Therefore, when posterior cranial fossa surgery is performed for younger patients in the Concorde position, hyperflexion in the Concorde position should be avoided as much as possible.

Finally, we consider what measures other than appropriately positioning the body should be taken to avoid thoracic spinal cord injury. In past reports, most patients have had severe neurological symptoms immediately after surgery.^1–6 However, it is unclear when these spinal cord injuries occurred. In the case reported by Tong et al.,¹⁰ posterior cervical decompression for craniocervical stenosis was performed. The MEPs showed no response in the lower limbs prior to skin incision, and postoperative thoracic spinal cord injury appeared at the T2 level. Remote spinal cord damage may occur from the time that the head and body positions are set, and the application of the MEPs to the lower limbs should thus be prepared in posterior cranial fossa surgery, with the Concorde position for younger patients even if the target lesion does not relate to the motor nerve.

Lessons

Even in cerebrovascular disease, thoracic spinal cord injury relating to the adoption of the Concorde position can occur. In younger patients, cervical hyperflexion in the Concorde position may induce thoracic spinal cord stretching and injury because surrounding tissues such as discs and ligaments are flexible. Measuring the lower-limb MEPs may be useful in avoiding complications.

Acknowledgments

We thank Edanz for editing a draft of this paper.

Author Contributions

Conception and design: Ishikawa, Ohtake, Nakamura. Acquisition of data: Ishikawa, Ohtake, Watanabe. Analysis and interpretation of data: Ishikawa, Endo, Ohtake. Drafting the article: Ishikawa, Ohtake. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Ishikawa. Statistical analysis: Ishikawa, Nakamura. Administrative/technical/material support: Ishikawa, Endo. Study supervision: Watanabe, Nakamura.