Intraaxial Spinal Cord Hemorrhage Secondary to Atlantoaxial Subluxation in a Dog

Introduction

Atlantoaxial (AA) subluxation is encountered most often in small-breed dogs, particularly Yorkshire terriers, Chihuahuas, and miniature poodles;^1–3 however, any breed can be affected.^4,5 The clinical presentation may be acute or chronic, with clinical signs ranging from mild cervical pain or tetraparesis to tetraplegia, respiratory compromise, and death.² Atlantoaxial subluxation results in caudal and dorsal displacement of the axis in relation to the atlas and flexion of the AA joint, which may cause compression and concussion of the spinal cord.² The typical cause of AA subluxation is loss of the ligamentous intervertebral support, which often coincides with a congenital malformation of the dens.^6–8 Trauma to the cranial cervical vertebral column may also result in AA subluxation secondary to fracture of the dens or ligamentous rupture.⁹

Atlantoaxial subluxation is often diagnosed radiographically.⁶ The typical radiographic features include: increased distance between the spinous process of the axis and the dorsal arch of the atlas; dorsal displacement of the body of the axis into the vertebral canal; and absence, agenesis, or dorsal deviation of the dens.^2,4,8 When not obvious, flexion of the neck will often make the subluxation more apparent; however, flexion can lead to further spinal cord injury.² Computed tomography (CT) or CT combined with myelography can be used to identify spinal cord compression and provide accurate evaluation of the dens.² Magnetic resonance imaging (MRI) can be used to evaluate the vertebrae and to identify spinal cord compression. Magnetic resonance imaging is particularly sensitive for the identification of intraaxial spinal cord abnormalities. To date, no study has detailed the MRI appearance of AA subluxation in dogs. In this report, the MRI findings and management of a dog with AA subluxation are detailed. In addition to the AA subluxation, abnormalities were observed within the spinal cord that were compatible with intraaxial spinal cord hemorrhage.

Case Report

A 1-year-old, 3.5-kg, spayed female, toy poodle was presented for acute-onset tetraplegia and severe neck pain. Although the owners did not observe a traumatic incident, the dog was unsupervised prior to being found recumbent and vocalizing. The dog had no previous medical problems.

Physical examination revealed only tachycardia and tachypnea, considered to be a response to pain. On neurological examination, the mentation was normal. The dog was tetraplegic except for occasional voluntary movement of the right pelvic limb. Postural reactions were absent in all four limbs. Myotatic and withdrawal reflexes were normal in all four limbs. The cutaneous trunci reflex was normal. Muscular tone was increased in all four limbs. The dog would vocalize in pain with minimal handling and with any movement of the neck. Neuroanatomical diagnosis was consistent with a first through fifth cervical vertebrae (C₁ through C₅) spinal cord lesion. Differential diagnoses included AA subluxation, vertebral fracture, diskospondylitis, meningomyelitis, and intervertebral disk disease.

Diagnostic testing included a complete blood count, serum biochemical profile, and urinalysis, all of which were normal. A lateral radiograph of the cervical vertebral column was obtained in a neutral position and demonstrated dorsal displacement of the second cervical vertebra (C₂). The dorsal displacement of the body of the axis in relation to the atlas created an increased distance between the dorsal arch of the atlas and the spinous process of the axis. These radiographic findings were consistent with AA subluxation. Ventrodorsal and oblique radiographs were not performed. Given the severity of the neurological deficits, MRI was performed to assess the compression and parenchymal changes of the spinal cord secondary to the AA subluxation. Under anesthesia, MRI of the cervical vertebral column was performed using a 3.0T MR unit^a with an extremity (knee) coil. The dog was placed in dorsal recumbency with the neck in a neutral to slightly extended position. The following pulse sequences were performed: T1-weighted fluid-attenuated inversion recovery (T1W FLAIR), T2-weighted (T2W), T2*-weighted (T2*W), T2-weighted fluid-attenuated inversion recovery (T2W FLAIR), and two-dimensional multiple-echo recalled gradient echo (2D-MERGE). The T1W FLAIR images were obtained after intravenous (IV) administration (0.2 mL/kg) of gadopentetate dimeglumine.^b Images were obtained in the sagittal, axial, and dorsal planes.

In the sagittal plane images, dorsal subluxation of the vertebral body of the axis in relation to the atlas was noted. The spinous process of the axis was difficult to visualize, which made it difficult to appreciate the increased distance between the spinous process of the axis and the dorsal arch of the atlas. However, ventrally the distance between the body of C₁ and the dens appeared (subjectively) enlarged [Figure 1]. The dens also appeared (subjectively) hypoplastic in the sagittal and dorsal plane images. The ventral and dorsal subarachnoid space was attenuated in the sagittal and axial plane images. In the sagittal plane, the spinal cord was deviated dorsally and was dorsoventrally compressed by the cranial aspect of the axis and dens as a result of the AA subluxation. An intraaxial, linear spinal cord lesion was identified extending from the midbody of the atlas to the caudal body of the axis in the dorsal funiculus.

The lesion was hypointense (signal void) on T1W FLAIR, T2W, T2W FLAIR, T2*W, and 2D MERGE images [Figures 2A, 2B]. The signal void was most evident in the sagittal and axial planes. On T2W and T2W FLAIR images, the hypointense spinal cord lesion was surrounded by an area of hyperintensity [Figure 3]. Contrast enhancement was not noted. In addition to AA subluxation, the imaging findings were consistent with intraaxial spinal cord hemorrhage and edema at the level of the AA articulation.¹⁰ Following MRI, cerebrospinal fluid was collected from the fifth and sixth lumbar (L₅ to L₆) intervertebral site. Cerebrospinal fluid analysis was normal.

Based on the MRI findings and the dog’s severe neurological deficits, the prognosis for full neurological recovery was considered to be guarded; therefore, the dog was treated conservatively for 7 days, limiting exacerbation of the concussive injury while monitoring for neurological improvement to enable a more accurate prognostic assessment prior to pursuing surgical stabilization of the joint. Conservative management consisted of external coaptation of the head and neck from the rostral mandible to the cranial thorax, as described previously.¹¹ In addition, the dog was administered buprenorphine^c (0.005 mg/kg IV q 8 hours), IV fluid therapy,^d and prednisone^e (0.5 mg/kg per day per os). Neurological deficits gradually improved, and 8 days after initial presentation, the dog had voluntary motor function in all four limbs but remained nonambulatory.

With the improvement suggestive of a more favorable prognosis, surgical stabilization was performed via a modified ventral fixation technique using cortical screws,^f Kirschner wires,^g and polymethylmethacrylate.^h,12 Postoperative radiographs disclosed adequate alignment and fixation. Five days postoperatively, the dog was able to ambulate with minimal assistance. Seven days postoperatively, on the day of discharge, the dog was ambulatory but had a severe proprioceptive ataxia and tetraparesis. Fourteen days postoperatively, the dog was ambulatory with a mild proprioceptive ataxia. Two months postoperatively, the neurological examination was normal.

Discussion

Treatment of AA subluxation consists of either conservative management or surgical stabilization. Although numerous risk factors affecting the prognosis of AA subluxation have been assessed, only age at onset and duration of clinical signs have been associated with prognosis.¹³ Specifically, affected dogs <2 years of age and with a duration of clinical signs <10 months were associated with a positive outcome following surgical stabilization.¹³ Similarly, when treated conservatively, affected dogs with a duration of clinical signs <30 days were significantly more likely to have a positive outcome.¹¹ The dog reported here initially was treated conservatively with external coaptation to temporarily stabilize the AA joint until the dog showed signs of improvement. Conservative management was selected initially in order to more accurately predict whether a positive outcome could be expected following surgical stabilization.

We hypothesized that the neurological deficits were due primarily to spinal cord concussion and the associated hemorrhage and edema rather than the ongoing spinal cord compression or vertebral instability. In AA subluxation, vertebral instability is a chronic lesion (despite the frequent acute presentation of affected animals) that is often the result of minor trauma leading to concussive injury of the spinal cord.^2,13 In the case reported here, external coaptation was initially utilized to eliminate instability and prevent continued concussive injury to the spinal cord. By doing so, it was believed that resolution of the hemorrhage and edema in the spinal cord would occur, clinical improvement could be observed, and the need for an immediate stabilization surgery would be eliminated. Such case management is contraindicated in severely affected dogs with acute intervertebral disk herniation, in which time to surgical intervention is correlated with return to function.¹⁴ Likewise, in cervical vertebral fractures, the shorter the time to surgical intervention the better the prognosis.¹⁵ In both of these latter disease processes (i.e., intervertebral disk herniation and cervical vertebral fracture), compressive and concussive forces occur acutely and lead to secondary spinal cord injury.¹⁶ Immediate decompressive surgery is advisable in cases in which acute compressive injury underlies the pathophysiology of the spinal cord injury.

Acute traumatic injury to the spinal cord leads to pathological changes in the spinal cord that may not respond to surgery alone.¹⁷ In chronic diseases such as caudal cervical spondylomyelopathy (Wobbler syndrome), vertebral instability leads to chronic pathological changes in the spinal cord including demyelination, neuronal loss, and gliosis.¹⁸ As a result, though prompt surgical intervention should be strongly considered in severely affected dogs with chronic vertebral instability, the need for immediate surgical intervention is unknown.¹⁹ While definitive recommendations cannot be ascertained from a single case, initial conservative treatment of spinal cord hemorrhage and edema secondary to AA subluxation in the severely affected dog reported herein provided the clinicians valuable prognostic information. A positive outcome was predicted prior to pursuing surgical stabilization.

In this case, the intraaxial spinal cord lesion was observed in the spinal cord that was hypointense (signal void) on spin-echo and gradient-echo sequences. Signal voids on spin-echo sequences are associated with gas, cortical bone, calcification, fibrous tissue, metallic implants, fast-flowing blood, and blood breakdown products.¹⁰ Based on the acute onset of neurological deficits and the presence of the AA subluxation and consequent spinal cord compression, the hypointensity in the dorsal funniculus was assumed to be associated with blood breakdown products. Findings on both of the gradient-echo sequences (T2*W and 2D MERGE) supported the presumptive diagnosis of intraaxial spinal cord hemorrhage.

Similar to other pulse sequences, gradient-echo sequences utilize a radiofrequency (RF) pulse for the conversion of longitudinal magnetization into transverse magnetization. Immediately after the RF pulse, protons that are in phase begin to dephase as a result of four effects: spin-spin interactions, chemical shift effects, magnetic inhomogeneities, and magnetic susceptibility differences.²⁰ Spin-spin interactions refer to the effect one proton has on a neighboring proton.²¹ The chemical shift effect is the difference in the precessional frequency of water protons compared to fat protons.²⁰ Magnetic inhomogeneities reflect that the main magnetic field is never uniform. Finally, magnetic susceptibility refers to the effect a material or tissue has on the main magnetic field.²²

Most tissues in the body are diamagnetic, essentially nonmagnetic, and they weaken the main magnetic field.²¹ Paramagnetic substances (such as the gadolinium-containing MRI contrast agents) become magnetized in an external magnetic field, and they strengthen the main magnetic field.²¹ Ferromagnetic materials, such as iron, are permanently magnetic and tend to greatly increase the main magnetic field.²¹ When two tissues of greatly different magnetic susceptibilities are adjacent (such as the bone/fat interface of the vertebrae and epidural fat), strong local variations in the magnetic field are created, resulting in a heterogeneous magnetic field causing an artifact (magnetic susceptibility artifact).²³ When dephasing is the result of only spin-spin interactions, it is referred to as T2 decay.²⁴ However, when dephasing is the result of all four effects, it is referred to as T2* (T2 star) decay.²³

Unlike spin-echo sequences, gradient-echo sequences utilize gradients to rephase protons after the application of the initial RF pulse. Since gradients do not correct the effects of chemical shift, magnetic field inhomogeneities, and magnetic susceptibility, all four factors contribute to dephasing in gradient-echo images. While these sequences are inherently prone to artifacts, this creates an advantage for hemorrhagic lesions because of the paramagnetic and ferromagnetic nature of blood breakdown products.²⁵ The paramagnetic and ferromagnetic nature of blood breakdown products results in increased local dephasing, causing magnetic susceptibility artifacts that ultimately lead to signal void on T2*-weighted images.¹⁰

In the present case, 2D-MERGE was used to evaluate the spinal cord. As with other multiple-echo gradient-echo sequences, the 2D-MERGE acquires multiple echoes with four different echo times, and they are averaged together, decreasing magnetic susceptibility artifacts.²⁶ Increasing the bandwidth also reduces chemical shift and magnetic susceptibility artifacts.²⁷ Two-dimensional multiple-echo gradient-echo sequences use large bandwidths to minimize magnetic susceptibility artifacts; this adversely affects image quality of the vertebral column.²⁶ However, the magnetic susceptibility artifacts related to hemorrhage remain apparent.

In veterinary medicine, the use of T2*W imaging is limited.^10,28–34 While T2*W sequences are utilized in the evaluation of hemorrhagic lesions, other substances such as ferritin, calcium, metallic material (surgical implants or metallic particles postoperatively from surgical drills), and air also result in a marked loss of signal.³⁵ Based on the signalment and diagnosis of AA subluxation, the signal void within the cranial cervical spinal cord was assumed to be hemorrhage.

In acute spinal cord injury (SCI) in humans, MRI of the spinal cord has largely supplanted the use of other imaging modalities.³⁶ Several imaging characteristics have been observed with acute SCI in humans. In acute SCI, hyperintensity on T2W images within the spinal cord is thought to be edema, whereas hypointense lesions typically represent hemorrhage.^37–39 As seen in the dog reported here, a hypointensity surrounded by peripheral hyperintensity in the spinal cord on T2W images likely represents hemorrhage with associated peripheral edema.³⁷ With acute SCI, the findings of spinal cord edema, swelling, and hemorrhage are associated with the severity of the neurological impairment in humans.^{37,38,40–42} In one study, intraaxial spinal cord hemorrhage was always associated with absence of sensory and motor function;⁴¹ however, in other studies, intraaxial spinal hemorrhage was also observed in conjunction with less severe neurological deficits.^42,43 Similarly, spinal cord edema and swelling also were observed in cases with preserved sensory or motor function.^41,42 Although the severity of clinical signs associated with hemorrhage in the spinal cord has not been evaluated in dogs, intraaxial spinal cord hyperintensity on T2W images may influence the prognosis in dogs with intervertebral disk herniations.^44,45

In humans with acute SCI, a poor prognosis has been correlated with spinal cord abnormalities observed with MRI.^41–43,46 Specifically, a poor prognosis has been associated with intraaxial hemorrhage in the spinal cord.^41–43,47 However, the effect of spinal cord edema or swelling on prognosis in humans with acute SCI remains unclear.^42,43 Dogs lacking nociception and having a hyperintensity in the spinal cord on T2W images had a postoperative success rate of 33%.⁴⁴ Furthermore, dogs lacking nociception and having a hyperintensity length >3 times the length of the second lumbar (L₂) vertebra had only a 10% postoperative success rate compared to a 100% postoperative success rate in dogs lacking nociception without a hyperintensity in the spinal cord.⁴⁴ In the dog presented here, the intraaxial spinal cord hemorrhage did not have a negative impact on outcome. Postoperatively, the dog recovered the ability to ambulate, and at the 2-month postoperative follow-up, the neurological examination was normal.

Conclusion

The signalment, history, and neurological signs of the dog in this report were consistent with AA subluxation. Magnetic resonance imaging evaluation of the cervical vertebral column correlated with the clinical and radiographic findings; however, noting the dorsal displacement of the body of the axis in relation to the body of the atlas was more helpful in the diagnosis of AA subluxation than observing the distance between the dorsal arch of the atlas and spinous process of the axis. In addition to observing the AA subluxation, intraaxial spinal cord pathology was noted. Findings were compatible with intraaxial spinal cord hemorrhage. The T2*W images were helpful in identifying a signal void in the spinal cord that was compatible with hemorrhage.⁴⁸ In the future, MRI may be utilized in dogs suspected of AA subluxation in order to assess the spinal cord for evidence of edema or hemorrhage. Finally, initial conservative management of AA subluxation in select cases of severely affected dogs may provide valuable prognostic information prior to pursuing surgical stabilization.