Bilateral medial medullary syndrome following anterior screw fixation of type 2 odontoid fracture, a case report of two patients

Abstract

Introduction

Medial medullary syndrome (MMS) has not been reported after anterior screw fixation of an odontoid type 2 fracture.

Case presentation

We report on two cases who suffered from an unstable type 2 odontoid fracture following a skiing and a domestic fall accident. Prior to anterior screw fixation surgery both patients presented without neurologic deficits but postoperatively developed a bilateral MMS, including an incomplete tetraparesis, impaired sensation of position and movement as well as tactile discrimination and paralysis of the tongue muscle with deviation to the paralyzed side. MRI showed a typical heart-shaped ischaemic lesion in the medial medulla bilaterally. The search for aetiologic factors was uneventful in both patients except for severe atherosclerosis.

Discussion

Due to the close proximity of the ischaemic area to the surgical site, we here propose the perioperative mechanical manipulation of the upper cervical spine during surgery of patients with atherosclerosis as a new aetiology for MMS.

Introduction

Odontoid fractures account for approximately 15–20% of all cervical spine injuries [1]. They represent the most frequent injuries of the upper cervical spine with a peak incidence in the elderly [2]. Anterior screw fixation is a standard procedure for the surgical fixation of unstable type two odontoid fractures [3, 4]. A number of complications associated with this procedure have already been reported in the literature mainly occurring in the geriatric population. They include screw loosening with loss of reduction due to osteoporosis [5], postoperative dysphagia [6], and lower union rates compared to posterior C1/C2 fusion [7]. Neurological deficits, however, are uncommon. According to the limited number of reports in the literature they either result from an acute trauma or from an additional trauma in the presence of a mobile odontoid nonunion [8, 9]. Perioperative neurological complications of anterior screw fixation of odontoid fractures have not been reported in the literature so far.

We here report two cases of bilateral medial medullary syndrome (MMS) following anterior screw fixation of type 2 odontoid fractures and discuss the possible pathogenesis of perioperative mechanical manipulation of the upper cervical spine during surgery as a new aetiology for MMS.

Case presentation

Case 1

A 57-year-old man presented with cervical neck pain following a skiing accident to the outpatient department. At first presentation, the patient had no neurological deficits. A cervical CT scan showed a posteriorly displaced type 2 odontoid fracture with a fracture of the spinous process of C2 (Fig. 1a). The patient was admitted for surgery on the same day. During anaesthetic induction, the patient developed anisocoria with left-sided mydriasis. Anaesthesia was interrupted and a CT angiography was performed immediately. No abnormal findings could be detected. There was neither a vascular pathology, in particular no pathology of the basilar artery, nor further displacement of the fracture. After weaning the patient showed no neurologic deficits. The anisocoria had resolved as well. Surgery was postponed followed by an uneventful intraoperative course two days later. Reduction of the posteriorly displaced fracture was performed by applying gentle traction followed by inclination of the head under fluoroscopic control. There was no reclination or rotation of the head. The cervical spine was immobilized with gentle traction using a halo ring, which was fixed to a custom-made immobilization device. Anterior odontoid screw fixation was performed without intraoperative surgical complications using two cannulated 4.0 mm cancellous lag screws.

Fig. 1. It shows representative images preoperatively and postoperatively.

a Case 1—Preoperative CT scan. Displaced type 2 odontoid fracture. b Case 1—Postoperative CT scan. Good reduction and correct screw positioning. c Case 1—MRI after weaning failure. DWI reveals a restricted diffusion with high DWI signal in the anterior medulla (left) and a high T2 signal intensity (middle). The ischaemic lesion is in close proximity to the surgical site and the vertebral artery. DWI diffusion-weighted imaging.

The postoperative course was complicated by a prolonged weaning phase requiring reintubation twice. A CT scan of the head and the cervical spine showed good reduction of the fracture and correct screw position (Fig. 1b). The cerebral CT scan was unremarkable. Therefore, an additional cerebral and cervical MRI was performed. It revealed an acute bilateral medial medullary ischaemia (Fig. 1c). A tracheotomy was performed and a PEG tube (percutaneous endoscopic gastrostomy) was placed resulting in successful weaning. Clinically the patient presented with a mainly left-sided tetraparesis. The cranial nerves were intact except for a deviation of the tongue to the right side indicating unilateral hypoglossal nerve palsy. Pyramidal signs were positive bilaterally. ISNCSCI (International standards for neurological classification of spinal cord injury) grading; upper extremity left (UEL) = 1, lower extremity left (LEL) = 1, sensory (light touch and pin prick) = not testable, upper extremity right (UER) = 12, lower extremity right (LER) = 12, sensory = not testable (patient was canulated). Voluntary anal contraction= yes, ASIA C (American Spinal Injury Association-Classification).

Initially the patient was treated with heparin 4000IE subcutaneously per day, which was increased to 6000IE twice a day complemented by a thrombocyte aggregation inhibitor (acetylsalicylic acid aspirin) considering a vascular endothelial thrombotic aetiology. Neurologic rehabilitation was initiated with successful decannulation 3 months after trauma. At discharge from the rehabilitation centre 7 months after trauma the patient presented with residual left-sided hemiparesis with no distal motor function (ASIA D, ISNCSCI: UEL = 8, LEL = 6, UER = 19, LER = 17). He was mobilized in a wheelchair and was capable of walking a few steps with a cane and suffered from recurrent aspiration pneumonias. The search for aetiologic factors was uneventful except for atheromatous plaques in the ascending aorta (grade IV) and the carotid artery.

Case 2

An 82-year-old man suffered a domestic fall accident and was admitted with cervical neck pain. At first presentation, no neurological deficits were reported. A CT scan showed a type 2 odontoid fracture (Fig. 2a). Flexion-extension fluoroscopy in the awake patient showed instability of the fracture with posterior displacement of the odontoid in extension. Surgery was delayed for eight days due to a preexisting acenocoumarin medication following a pulmonal artery embolism three years ago. Bridging of anticoagulation was performed with heparin 4000IE subcutaneously twice per day. An anterior surgical approach was chosen due to the severe obesity of the patient impeding prone positioning. Reduction of the fracture, intraoperative immobilization of the cervical spine and anterior screw fixation were performed in the same manner as in the first case.

Fig. 2. It displays preoperative as well as postoperative CT and MRI scans.

a Case 2—Preoperative CT scan. Type 2 odontoid fracture. b Case 2—Postoperative CT scan. Unsatisfactory reduction and misplacement of the left screw. c Case 2—Postoperative MRI DWI cannot be read due to artefacts of the misplaced left screw (left). No T2 signal intensity hyperintensity, but artefacts are visible (middle). Screw artefacts in T2 weighted imaging show the close proximity between the surgical site and the medulla oblongata (right). DWI = diffusion-weighted imaging. d Case 2—Revision surgery Intraoperative fluoroscopic control of bilateral anterior transarticular C1/C2 fixation. e Case 2—Revision surgery. Postoperative CT control after bilateral anterior transarticular C1/C2 fixation. f Case 2—MRI after revision surgery. DWI reveals a restricted diffusion with high DWI signal in the anterior medulla (left) and a high T2 signal intensity in the axial (middle) and the sagittal view (right).

Postoperatively, during the weaning process, the patient presented with a right-sided hemiparesis. A CT scan showed an unsatisfactory reduction and a mispositioned left screw (Fig. 2b). Cerebral and cervical MRI did not show any signs of haematoma, direct compression of the brain stem or dissection of cerebral and/or cervical arteries. Diffusion-weighted MRI (DWI), could not be evaluated due to metal artifacts (Fig. 2c). Revision surgery was performed immediately and included screw removal and anterior transarticular C1/C2 fixation (Fig. 2d). A postoperative CT scan after revision showed good reduction of the fracture and adequate screw positions (Fig. 2e). After revision surgery however the neurologic condition had worsened and the patient could not be weaned and presented with a mainly right-sided tetraparesis. A second MRI was performed two days later and revealed a subacute medial bilateral medullary ischaemia. No dissection of any cerebral arteries could be detected (Fig. 2f).

A tracheotomy was performed and a PEG tube was placed. After gaining consciousness the patient presented with a nearly complete tetraplegia (ASIA B, ISNCSCI: UEL = 3, LEL = 0, UER = 0, LER = 0, sensory both sides: not testable, voluntary anal contraction=no). Swallowing was insufficient and aspirations led to recurrent pneumonia with intensive care unit treatment. His general health condition worsened and therefore best supportive care management was initiated. The patient died three months after the accident. The search for aetiologic factors revealed diabetes mellitus type II, obesity and increased cholesterol levels. Cerebral MRI showed vascular leukoencephalopathy (Fazekas III).

Discussion

We here report two cases of postoperative bilateral medial medullary stroke following anterior screw fixation of a type 2 odontoid fracture. Perioperative acute ischaemic stroke is a frequent complication in cardiac and vascular surgery. While the rate is lower in noncardiac and nonvascular surgery, it is not an infrequent complication in these cases neither [10]. Risk factors include age, atrial fibrillation, diabetes, vascular co-morbidities, immobilization and preexisting anticoagulation [11].

The presented cases differ from the typical cases of perioperative acute ischaemic stroke. First, MMS is a rare type of ischaemic stroke and second, the ischaemia occurred in close proximity to the surgical site questioning the MMS aetiology. As mentioned, bilateral MMS is a very rare syndrome [12] and medullary infarction in general represents less than 1% of all posterior circulation strokes. MMS results from a paramedian infarction of the medulla oblongata secondary to occlusion of the vertebral or anterior spinal artery branches due to an atherothrombotic event [13, 14]. MMS typically shows a rapid neurologic progression with a nadir within less than 4 h. The development of neurologic symptoms in our cases, however, is unclear due to their perioperative onset. The clinical presentations match the classical MMS with contralateral hemiparesis, a contralateral impaired sensation of position and movement as well as tactile discrimination and an ipsilateral paralysis of the tongue muscle with deviation to the paralyzed side when the tongue is protruded. The temporarily present anisocoria in the first case however remains unclear and might represent a topical contamination (e.g., atropine) during anaesthetic induction.

Imaging is a critical issue in the assessment of cerebral ischaemia with MRI representing the gold standard. MRI without DWI may be unremarkable within the first 12 h after the onset of clinical symptoms and a CT scan even up to 24 h [15]. Accordingly, DWI is important for an early diagnosis of medullary ischaemia. The second case, however, highlights some shortcomings of DWI in the assessment of ischaemia (Fig. 2c). In this case, MRI with DWI was performed ~5 h after surgery. The misplaced screw however was in close proximity to the brain stem and the artifacts lead to an illegible DWI signal (Fig. 2c). The MRI in the first case (Fig. 1c) and the second MRI after screw removal in the second case (Fig. 2d) then showed a typical heart-shaped lesion of restricted diffusion with a high DWI and low ADC signal as well as high T2 signal intensity indicating bilateral medullary ischaemia. We therefore recommend for patients with weaning failure or new neurologic deficits undergoing anterior screw fixation of a type 2 odontoid fracture to perform MRI with DWI to rule out ischaemia as CT scans might be negative.

The aetiology of medullary infarction is predominantly caused by an occlusion of the paramedian branches of the vertebral artery and/or the anterior spinal artery due to a vertebral artery dissection, vasculitis or cardiac embolic occlusion. Especially in the elderly population arteriosclerosis represents a considerable risk factor for medullary ischaemia [16]. The aetiology in our two cases is not clear. No acute vascular pathology (e.g., dissection), no cardioembolic events and no medullary haemorrhage were detected. Considering the close proximity of the surgical site and the medulla oblongata it is reasonable to assume that local mechanical manipulation contributed to the medullary ischaemia. At first glance, it might be speculated that the screw mispositioning (Fig. 2b) as depicted in the second case caused a local thrombotic occlusion. We therefore performed an immediate revision surgery with screw removal. However, MRI in the second case did not reveal any signs of direct mechanical impact on the medulla such as haematoma, oedema or direct compression by the screw (Fig. 2c) and there was no screw malpositioning in the first case. As MMS is a rare finding and has not been reported following anterior screw fixation of type 2 odontoid fractures and at the same time screw malpositioning is not unusual at all, we retrospectively do not think that this was the cause for the stroke in the second case.

We hypothesise that the perioperative mechanical manipulation of the upper cervical spine for intubation, positioning of the patient and reduction of the fracture may have led to local atherothrombotic occlusions of small branches of the vertebral artery and/or the anterior spinal artery, which cannot be directly visualized by conventional MRI due to resolution constraints. Additionally, the continuous cervical traction applied by rigid perioperative immobilization of the cervical spine may be an additional risk factor. Importantly, both patients were treated according to the internationally applied reduction and immobilization protocol for the treatment of displaced odontoid fractures: The cervical spine was immobilized during the waiting period for surgery in both cases using a cervical collar in order to reduce unintentional flexion and extension of the cervical spine without further traction. During surgery cervical immobilization was performed using a Halo device. Reduction of the posteriorly displaced fractures were performed by applying gentle traction followed by inclination of the head under fluoroscopic control. There was no reclination or rotation of the head. Reduction was performed under fluoroscopic control in a lateral view, which showed no signs of overtraction at any time.

Both patients suffered from severe atherosclerosis (case one; grade IV plaques of the ascending aorta and case 2; vascular leukoencephalopathy (Fazekas III) and it could be shown in an animal study that the tensile properties of atherosclerotic arteries are impaired under mechanical manipulation, which makes especially small vessels prone for injuries during mechanical manipulation [17]. It was also shown that the vulnerability of atherosclerotic plaques is increased during cervical rotatory manipulation [18]. In a large survey, it could be shown that chiropractic manipulation especially in the cervical spine is associated with a risk of neurological complications such as strokes and myelopathies mainly in the posterior circulation [19]. These studies support our hypothesis of atherothrombotic occlusions of small branches of the vertebral artery during mechanical manipulation of the upper cervical spine.

In conclusion, we here propose the perioperative mechanical manipulation of the upper cervical during anterior screw fixation of a type 2 odontoid fracture as a new aetiology for bilateral MMS. Importantly, these findings should not challenge the indication or surgical procedure for fixation of type II odontoid fracture in patients with severe atherosclerotic diseases. These two cases should increase the awareness of physicians for this rare and severe complication. In cases of weaning failure or development of new neurologic deficits following mechanical manipulation of the upper cervical spine, MRI including DWI might help to identify the underlying aetiology (haematoma versus interstitial edema due to screw malpositioning versus ischaemia) and help guide further treatment decisions (e.g., anticoagulation/antiplatelet therapy).

Author contributions

MZ, AS, BP, and DK took care of the patient. MZ, MN, and DK drafted and wrote the manuscript. MZ, AS, BP, MN, and DK critical review and consent on the final version of the manuscript.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41394-021-00462-x.