Delayed Paraplegia Following Correction of Severe Thoracolumbar Kyphotic Deformity by Posterior Vertebral Column Resection.
Jing Li,Guo-hua Lv,Xiao-bin Wang,Bing Wang,Chang Lu
DOI: https://doi.org/10.1111/j.1757-7861.2009.00066.x
2010-01-01
Orthopaedic Surgery
Abstract:Decompensation in the sagittal plane of the spine leads to specific complaints such as pain, cardiopulmonary compromise, limitation of the activities of daily living and neurological impairment. Along with the development of new instrumentation and spinal surgical techniques, correction of severe and rigid thoracolumbar kyphosis is now commonly being performed1. Vertebral column resection represents a new level for treating sharp angular kyphotic deformity, but on the other hand this technique is often associated with significant bleeding and has an increased risk of neural injury2. The etiology of neurologic injury during surgical correction of kyphosis is thought to be vascular, hypotensive, neural element stretch, or direct trauma to the neural elements. Typically, neurological deficits caused by correction of severe thoracolumbar kyphotic deformity occur acutely during surgery or immediately after surgery. Delayed paraplegia, due to various reasons, has been reported by a few researchers in scoliosis surgery1, 3-7. To the authors' knowledge, there is no report in the literature describing delayed post-operative paraplegia following vertebral column resection for thoracolumbar kyphosis deformity. In the current study, the authors report a rare case of delayed paraplegia following correction of a 93° kyphosis in a 44-year-old woman with Scheuermann's disease. A 44-year-old woman, who had had progressive worsening of thoracolumbar kyphosis from the age of 16 years, presented with a severe kyphosis accompanied by severe leg and back muscle fatigue and pain over the past three years. The patient had intermittent claudication due to back and leg muscle fatigue and pain and had received no prior medical care. General and neurological examination revealed that she had good muscle tone, normal sensation and reflexes. Her head and neck exhibited a forward posture. There was no bowel or bladder dysfunction. Upright anteroposterior and lateral radiographs demonstrated a 93° kyphosis from T8 to L2 with mild coronal scoliosis (Fig. 1). Preoperative extension radiographs revealed a fixed kyphosis. A total spine magnetic resonance imaging (MRI) failed to demonstrate any parenchymal signal change or intraspinal abnormality that might complicate correction of her kyphosis. Preoperative testing revealed a mild impairment of ventilatory function, with a forced vital capacity (FVC) of 2.3 L, forced expiratory volume in one second (FEV1) of 1.8 L, and total lung capacity (TLC) of 3.3 L. The corresponding preoperative percentages of the predicated normal values were 78% for FVC, 69% for FEV1, and 76% for TLC. (A) Upright anteroposterior radiographs showing mild coronal scoliosis. (B) Lateral radiographs demonstrating a 93° kyphosis from T8 to L2. The patient was placed in a prone position and a posterior midline incision made. Pedicle screws were placed bilaterally in the proximal four (T8–T11) and the distal three levels (L2–L4). Coronal and lateral radiographs were performed immediately after screw insertion to confirm acceptable placement of the screws. The patient then underwent posterior vertebral resection of T128, and Ponte Smith–Petersen facetectomies from T9 to L29. The deformity was gradually corrected with repeated additional compression and shortening of the vertebral column along a precontoured rod. Finally, autologous bone was inserted from the posterolateral side to fill the resection gap and provide anterior arthrodesis. During the operation, 24.5 unit (4900 ml) of allogenic red blood cells, and 550 ml of fresh frozen plasma were transfused to compensate for the intraoperative blood loss of 4500 ml. The mean arterial pressure was maintained at 60–70 mm Hg. The total operation time was 10 h, and the course of anesthesia was uneventful. Surgery was completed without intraoperative complications. Intraoperative neurophysiologic monitoring of the spinal cord was not available, and an intraoperative wake-up test revealed no motor impairment of the patient's lower extremities. The patient retained full motor strength and normal deep reflexes in her lower extremities. As the time taken to regain consciousness and commence spontaneous respiration after anesthesia was prolonged, the patient was admitted to the Intensive Care Unit (ICU) and maintained on a ventilator. It was not possible to take postoperative X-rays, but motor strength, pinprick sensation and deep reflex status the next morning were found to be stable. The patient's heart rate, blood oxygen saturation, and body temperature were almost normal. Postoperative hypoxemia was not noted. However, during the first day in the ICU, the estimated blood lost into her wound drainage bag was 6010 ml. The total volume of transfused fluid was 17 903 ml, consisting of 18.5 unit (3700 ml) of allogenic red blood cells, and 1700 ml of fresh frozen plasma. Because of her severe blood loss, the patient had prolonged hypotension, and a low hemoglobin and hematocrit, her hemoglobin decreasing from 11.7 g/L to 5.6 g/L, and hematocrit from 39.9% to 18.2%. The extent of blood loss was unusual and active bleeding was thought to be unlikely. Initially we used a hemostat, a vasopressor agent (dopamine) and fluid replacement to manage the bleeding and increase the patient's arterial pressure. Thirteen hours later (the next morning), the mean arterial pressure had increased (Fig. 2) and the amount of blood in her wound drainage bag had markedly decreased, so emergency surgical revision was not undertaken. Blood pressure fluctuations in the first 25 h after surgery. Twenty-five hours after surgery, the ICU physicians noticed that the patient had developed paraplegia. Physical examination revealed a flaccid paraparesis (0/5) in both lower extremities with absence of knee and ankle reflexes. The patient's had no sensation below the umbilicus level. By this time, her hemoglobin had decreased to 5.6 g/L and her hematocrit was 18.2%. High-dose methylprednisolone (30 mg/kg bolus followed by 5.4 mg·kg−1·h−1 for 23 h) was administered without delay and emergency surgical revision undertaken. No intraspinal hematoma or hardware malposition was identified, however the spinal cord was grossly swollen and the dura looked redundant between T11 and L1. The laminae of T11 and L1 were resected to decompress the spinal cord. An X-ray film taken three days after revision surgery confirmed acceptable placement of the screws with correction of the hyperkyphosis to 32° (Fig. 3). MRI failed to demonstrate any compromise of the spinal canal, but showed an area of hyper-intensity below the T6 vertebra on T2-weighted images, indicating a medullary lesion with spinal ischemia (Fig. 4A). X-ray films after revision surgery confirming acceptable placement of the screws with correction of the hyperkyphosis to 32°. (A) MRI showing an area of hyperintensity below the T6 vertebra on T2-weighted images. (B) Six months after the operation, the high-intensity signal had partially disappeared. The patient returned for final follow-up 6 months after surgery. Her neurological status had improved slightly, with persisting paraplegia below the3 inguinal level. MRI revealed that the high-intensity signal had disappeared (Fig. 4B). Neurological deficits represent the other side of the coin of surgical correction of spinal deformity. The many reasons for such defects include factors both intrinsic and extrinsic to the spinal cord and column itself 1, 10. Scoliosis or kyphosis with congenital spinal abnormalities, such as syringomyelia, are the main candidates for this kind of complication. The severity and rigidity of the curve, as well as the degree of kyphotic deformity, also contribute to the risk of neurological complications during corrective procedures10, 11. Vertebral column resection has been described for tumors, spondyloptosis, excision of hemivertebrae and spinal deformities. While this procedure has made the treatment of serious and rigid spinal deformities possible, it has concomitantly increased the risk of neurological complications2. An important factor determining the degree of risk is the specific level of kyphosis correction. The distinctive vascular anatomy of the spinal cord in the thoracic region contributes to a high risk of neurological lesion when operating in this area. Compared with the cervical spine, the thoracic spine has a less liberal blood supply. Therefore, both in terms of absolute quantity and volume per square centimeter, the blood supply to the thoracic nervous tissue is inferior, which contributes to the potential for ischemic cord injury in this region12. Another vascular characteristic of the spinal cord is the artery of Adamkiewicz, or the major anterior segmental medullary artery, which originates from intercostal or lumbar vessels near the intervertebral foramen between the T7 and L4 levels. Although the concept of the major anterior segmental medullary artery has been challenged and disputed, it is extremely important in regards to spinal cord supply, as it is the principle blood supply to the lower anterior two-thirds of the spinal cord12-14. When a neurologic deficit is induced during spinal distraction, it is thought to be caused by occlusion or spasm of this artery15. When a neurological deficit occurs following scoliosis surgery, it may be acute or delayed in development. Spinal cord injury may be caused by corrective distraction measures, ischemia or canal compromise (by implants or hematoma). Delayed neurological deficits several hours after spinal surgery are uncommon, and are usually attributable to the latter two reasons. The etiology of spinal cord ischemia is multifactorial and includes vascular division, traction, arterial or venous embolization, and segmental vessel ligation. In addition, systemic hypotension, anemia, or local edema can exacerbate the effects of poor perfusion of the spinal cord1, 7. To the authors' knowledge, 15 cases of delayed neurological deficit following uneventful surgery or spinal deformity have previously been described in the literature. Although the causes of such deficits are potentially related to many factors, ischemia induced by hypotension is the most commonly reported7, 11, 16-20, followed by postoperative haematoma3, 6, 10, vascular origin1, 4, anemia10, instrumentation compression10, and hypokalaemia5. MRI of the spine is the best option for localizing the exact cause of paraplegia. The options for treatment in these cases are: (i) high-dose methylprednisolone; (ii) blood transfusion; (iii) vasopressor agents and fluid replacement; (iv) emergency decompression; (v) removal of the instrumentation. Most spinal surgeries are assisted by a state of induced hypotension to minimize blood loss, but unchecked and persistent hypotension may lead to inadequate perfusion of the spinal cord, resulting in spinal cord ischemia. Both in experimental animals21 and during aortic surgery in humans22, 23 profound hypotension has been documented to cause neurological deficits. The incidence of paraplegia due to hypotension after surgery for correction of spinal deformity has not been well documented. One case of paraplegia caused by significant intraoperative hypotension during anterior spinal fusion surgery has been reported16, and in another case intraoperative neurophysiologic monitoring data disappeared17, but in the other cases reported, the onset was somewhat delayed7, 11, 18-20. Taylor et al. have shown a clear and reversible relationship between spinal cord function and postoperative hypotension7. In the case reported by these authors, the postoperative blood pressure fell to the same level as during surgery, causing loss of spinal cord function. Once normal arterial pressure had been restored by transfusion, neurologic function recovered rapidly without any surgical revision. The authors proposed that hypovolemia with consequent reduced cardiac output were the main factors causing the delayed postoperative paraplegia. Johnston et al. also reported two cases of delayed paraplegia resulting from neural ischemic and edema after correction of complex spinal deformities with sublaminar wiring19. Sublaminar segmental spinal instrumentation was used in both cases, and both patients had permanent paraplegia at the last follow-up. Othman et al. reported a case of correction of Scheuermann's kyphosis with T4–L2 pedicle screw placement and multilevel Smith-Peterson osteotomies17. During the operative procedure, the patient's mean arterial pressure dropped precipitously to 50 mm Hg, and percutaneous neurogenic mixed evoked potentials (NMEP) and somatosensory evoked potentials (SSEP) were acutely lost bilaterally. Once normal blood pressure had been restored, the SSEP and NMEP data returned. A wake-up test performed after deformity correction indicated no neurological deficit. Cervellati et al. reported a case of severe kyphoscoliosis complicating spinal muscular atrophy type 210. The patient suddenly became paraplegia 26 h after correction of the deformity using the Hartshill technique. Spinal cord ischemia, with a significant decrease in hemoglobin to 5.6 g/L, was the main cause of the neurological deficit. MRI revealed a high-intensity signal which indicated a medullary lesion secondary to cord ischemia. The patient retained full motor strength and normal deep reflexes in the lower extremities for about 25 h after correction of kyphosis. Postoperatively, she had massive blood loss with prolonged hypotension, low hemoglobin and hematocrit. Although a dopamine infusion was maintained throughout most of the first postoperative day together with transfusion of fluid, the mean arterial blood pressure still ranged from 56 mm Hg to 68 mm Hg over the first 13 h. Her hemoglobin decreased from 11.7 g/L to 5.6 g/L, and hematocrit from 39.9% to 18.2%. Thus we deduce that the cause of the delayed neurological deficits in our case was cord ischemia caused by hypotension and anemia. Immediate revision surgery and sequential MRI confirmed spinal cord ischemia, edema and infarction above the instrumentation level. Mechanical obstruction from the hardware or a developing hematoma was ruled out. Our case illustrates that delayed ischemia following complicated spinal surgery can lead to serious irreversible neurological deficits. Altogether, during or after surgery for correction of deformity there are many potential risks to the spinal cord which can lead to neurological compromise. Such neurological deficits are usually found intraoperatively or immediately after surgery. Several-hour delayed paraplegia is uncommon. This case report suggests that hypovolemia is indeed an important causative mechanism for delayed neurologic deficits in spinal deformity correction surgery especially with the vertebral column resection technique. One should be aware of the importance of postoperative observation of spinal deformity cases in order to detect any impending neurologic deficit early.