|Home | About | Journals | Submit | Contact Us | Français|
We report a rare case of postoperative paraplegia in a patient with carcinoma of the oesophagus following oesophagectomy. Neurological deficit was characterised by loss of sensation from the spinal level T2 down to T6, together with flaccid paraparesis of both lower extremities. This was initially thought to be secondary to an epidural haematoma and a magnetic resonance scan was arranged. This suggested changes consistent with spinal cord infarction from D2 to D6 with no haematoma or abscess. This severe complication was most probably caused by embolisation from an atherosclerotic plaque of the thoracic aorta, as the nature of the surgery requires aortic manipulation. This may have consequently led to occlusion of a significant part of the spinal blood supply. Even though anterior spinal artery syndrome is a well known problem in the operative management of thoracic aortic aneurysms, this complication is extremely rare after oesophagectomy.
The incidence of paraplegia following thoracotomy for non-vascular operations is estimated at 0.08%,1 with only five cases been reported between 1966 and 2001 after oesophageal resection.1 The five cases described were patients in whom the tumours were located in the mid to distal oesophagus, with the level of spinal injury being located between T6 and L11. This is the area commonly fed by the great radicular artery of Adamkiewicz. Our case is unique as the level of spinal injury was located between T2 and T6.
A man presented with a history of haematemesis and meleana. An oesophagogastroduodenoscopy was performed on admission which revealed mucosal irregularity at 35 cm and a biopsy confirmed the diagnosis of squamous cell carcinoma. The patient subsequently had an endoscopic ultrasound scan (EUS) which demonstrated blurring of the mucosal layers with a single para-oesophageal lymph node but no other lymphadenopathy.
The patient was then referred to the cardiothoracic services for assessment of operability. Further investigations included a computed tomography (CT) scan and a positive emission tomography (PET) body scan. These confirmed an eccentric primary lesion posterior to the left atrium representing the primary lesion with no evidence of metastasis. An echocardiogram demonstrated no left atrial abnormality and in particular no thrombus, and an electrocardiogram (ECG) confirmed the patient was in sinus rhythm.
Before surgery the patient was optimised in view of his significant past medical history of diabetes and renal failure, along with having thromboembolic deterrent stockings (TEDs) and prophylactic doses of enoxaparin.
For the purposes of postoperative analgesia a thoracic epidural catheter was placed with the patient awake in the anaesthetic room. At this stage no neurological disturbance was observed. A left thoracophrenotomy was then performed under general anaesthesia (GA). After exploring the abdominal cavity, the oesophagus was mobilised and the stomach was prepared as a substitute. The right gastro-epiploic artery was preserved to supply the stomach tube. The remaining vessels were ligated and the inferior pulmonary ligament lymph node was biopsied. Extensive lymph node dissection was not performed. After resection of the distal third of the oesophagus, a stapled anastomosis was performed above the aortic arch. Haemostasis was attained without having to ligate the intercostal or spinal arteries and without the use of oxidised cellulose.
Postoperatively the patient was extubated and transferred to the cardiothoracic critical care unit where his renal function deteriorated over 48 h, necessitating continuous veno-venous haemofiltration (CVVH).
On the fourth postoperative day it was noted that the patient could not move his lower limbs. Neurological assessment of the lower limbs revealed a flaccid tone, power grade 0, absent knee and ankle reflexes, and loss of sensation.
An initial suspicion of an epidural haematoma was made and a magnetic resonance imaging (MRI) scan of the spine was arranged. The MRI suggested changes consistent with spinal cord infarction extending from D2 through to D6, with no epidural haematoma or abscess noted (figs 1 and and2).2). The primary cause was therefore thought to be thromboembolic.
Unfortunately the patient’s condition continued to deteriorate with increasing respiratory compromise and no improvement in renal function. This was discussed with the family and, given the poor prognosis, it was decided that it would not be advisable to escalate treatment. The patient died 24 h later.
Paraplegia after surgery of the thoraco-abdominal aorta has been well reported. However, it is rare after pulmonary, tracheal or pleural surgery,2,3 with the overall incidence estimated to be 0.08%.1 Only five cases of paraplegia have been reported between 1966 and 2001 after oesophageal resection.2 The five cases previously described were performed on men in the 58–85 year age range who had advanced atherosclerotic disease and risk factors such as diabetes mellitus, hypertension and obesity. The resected tumours in these cases were located in the mid to distal oesophagus, with the level of spinal cord injury being located between T6 and L1. This is the area commonly fed by the great radicular artery of Adamkiewicz. This case is unique as the level of injury was located higher up, between T2 and T6.
The spinal cord is supplied with blood by one anterior spinal artery and two posterior spinal arteries. As they descend down the spinal cord, they are fed by segmental arteries that arise from the aorta. There are more segmental arteries supplying the posterior spinal arteries than the anterior arteries. This accounts for anterior spinal artery syndrome being more frequently observed than the posterior spinal artery syndrome.
Most radicular arteries arise from the left extraspinal vessels. Very often, a single, large radicular artery is found. The radiculo-medullary artery (artery of Adamkiewicz) constitutes the largest and most important radicular artery. It enters the spinal canal in variable segments—T7 to T12 in 75%, T5 to T8 in 15%, and L1 to L4 in 10%—with it being on the right in 17% of cases.
There is variability in the number and size of the supplying arteries—that is, both longitudinal arteries and the anterior spinal artery. This results in significant variation in the diameter.
The variability in the blood supply makes it difficult to identify and spare the specific blood vessels as they arise from the aorta. This creates a risk that surgical trauma can result in spinal cord ischaemia. The pathophysiology of the ischaemic insult may not be very clear.
Mechanisms for interruption of the blood supply are numerous.4 These include the inadvertent ligation of the artery of Adamkiewicz, lateral flexion of the spine during surgery5,6 or attempts to control bleeding at the costovertebral angle by using oxidised cellulose or ligation of intercostals vessels.6,7 Stagnation of blood within the vessels supplying the spinal cord can also occur as a result of prolonged periods of hypotension. Physiological factors can add to the risk including atherosclerosis, polycythaemia, and anatomic changes of the spine (kyphoscoliosis), vascular malformations and the hypercoagulability associated with malignancy. An epidural haematoma can also cause extrinsic compression, though paraplegia has been reported following thoracotomy in patients not receiving a thoracic epidural.5
Paraplegia resulting from inadequate blood supply through the anterior spinal artery produces a clinical picture known as anterior spinal artery syndrome. This is characterised by loss of motor function with intact or, more commonly, impaired sensory function below the level of injury.
In our case, while on one lung ventilation there was no significant observed fall in the saturations, with the patient maintaining saturations of 100% throughout. In addition the operation was uneventful with no documented hypotensive episode. The patient recovered in an intensive care unit and required no inotropic support as he was maintaining a mean arterial pressure >80 mmHg. In this patient an epidural haematoma or abscess was among the differential diagnoses as a cause for the paraplegia; however, this was excluded by means of an MRI scan. The cause of this patient’s injury is not clear because the lesion appears to be very high for an injury to the artery of Adamkiewicz, and there was no reported ligation of intercostal or spinal arteries and no intraoperative use of oxidised cellulose to attain haemostasis. Embolisation from an atherosclerotic plaque is the likely cause as the nature of the surgery required aortic and potentially left atrial manipulation.
In conclusion, spinal cord lesions rarely occur after oesophageal resection and are thought to be due to an ischaemic insult. It is therefore important to identify preoperatively patients who are at increased risk of compromised blood flow to the spinal cord, such as elderly patients with known risk factors for atherosclerosis, to implement preventive measures that may help avoid this complication. These measures may include preoperative echocardiography, reduced aortic manipulation and thromboprophylaxis both pre- and postoperatively. However, the avoidance of hypoxia and hypotension at any stage of the procedure must be paramount.
Competing interests: None.