MRI of a Split Cord Malformation in a German Shepherd Dog

Introduction

Spinal dysraphism is a general term that encompasses a broad category of abnormalities of the spine, spinal cord, and associated structures that occur as a result of a defect early in embryologic development of the neural tube. Those abnormalities can involve the spinal cord, vertebrae, soft tissue, skin, and other embryologically related structures overlying the affected vertebral segments. Split cord malformations, also known as diastematomyelia and diplomyelia, are forms of spinal dysraphism. Split cord malformation refers to a longitudinal separation of a portion of the spinal cord. There is a continuous spectrum of this abnormality ranging from a partially cleft cord in a single dural tube at one end of the spectrum to a completely duplicated cord within dual dural tubes with an intervening bony spur at the other end of the spectrum.¹

Sagittal splitting of the spinal cord and other occult spinal dysraphic conditions are uncommon in humans; however, with the advent of advanced neuroimaging techniques, the condition is being recognized with increased frequency.^2,3 Historically, computed tomography myelography was used extensively in imaging spinal dysraphic conditions and remains superior in demonstrating the bony abnormalities that can be a component of the disease process.^4–6 More recently, advances in MRI technology have allowed for better evaluation of the soft-tissue structural abnormalities that are often associated with dysraphic conditions. MRI of split cord malformation is highly effective and is described by some as the current modality of choice for imaging spinal dysraphic conditions and associated common lesions that are not bony deformities.^2,3,6

Spinal dysraphic conditions are also recognized in veterinary medicine, with spina bifida being among the more commonly diagnosed abnormalities. Other spinal dysraphic conditions have been described in both human and veterinary medicine. Myeloschisis, intradural lipomas, meningomyelocele, and tethered cord syndrome have been documented in the Weimaraner, English bulldog, and Manx-type cats.^7–11 Although rare in animals, split cord malformations have been documented and are most frequently reported in calves.^7,12,13 The diagnosis of a split cord malformation was made in the German shepherd dog described in this report based on the appearance of the spinal cord using MRI and is supported by patient history and clinical signs. To the authors’ knowledge, there is no documentation of a canine with split cord malformation with any imaging modality.

Case Report

A 9 yr old spayed female German shepherd dog was referred to the Advanced Veterinary Medical Imaging facility for an MRI of the pelvis and the thoracic and lumbar spine. The dog's history included chronic, life-long problems with difficulty urinating. The owner described the problems, which included pollakiuria, stranguria, and dysuria characterized by voiding only small amounts of urine since being adopted at approximately 3 mo of age. The dog reportedly had previous episodes of an inability to urinate that necessitated catheterization within the last 2 yr; however, the underlying cause remained undiagnosed. Episodes of bacterial urinary tract infections that responded to routine antibiotic therapy occurred intermittently over the past several years. No urinary incontinence had been reported. In addition, the abnormal behavior of walking while defecating and dropping small amounts of stool was described. Previous episodes of an inability to urinate responded to urinary catheterization and oral prazosin therapy.

The dog was evaluated by its primary care veterinarian for an acute onset of abdominal distention, weakness, and vomiting. Physical examination findings included tachycardia, tachypnea, and abdominal distension due to a very large urinary bladder that could not be manually expressed. Diagnostic studies included a complete blood count (CBC), serum biochemical profile, and two orthogonal radiographs of the abdomen. Radiographic findings included a markedly distended urinary bladder without evidence of radiopaque cystic calculi, mass lesions, or any other abnormalities indicative of a mechanical obstruction. There were no radiographic abnormalities of the spine identified. CBC results identified a mild lymphopenia (0.4×10³/μL; reference range, 0.5–4.9×10³/μL) and thrombocytopenia (136×10³/μL; reference range, 175–500×10³/μL). Remaining values on the CBC were within normal limits. Serum biochemical profile results identified an elevated albumin (4.5 g/dL; reference range, 2.2–3.9 g/dL) and mildly elevated blood urea nitrogen (32 mg/dL; reference range, 7–27 mg/dL). Creatinine and the remaining blood chemistry values were within normal limits.

The urinary bladder was catheterized using a polypropylene catheter and employing a rigid otoscope for visual guidance under sedation with IV medetomidine^a (0.01 mg/kg) and butorphanol^b (0.21 mg/kg). An abnormal band of tissue was visualized along the right side of the vaginal canal during the procedure. Catheterization was successful and the urinary bladder was drained. The urine was grossly hematuric and was submitted for urinalysis. Urinalysis abnormalities reported included an elevated pH 7.5 (reference range, 5.5–7.0), elevated protein (3+; reference range, negative), presence of occult blood (3+; reference range, negative), >50 red blood cells (RBCs)/high-power field (HPF; reference range, 0–3/HPF), and bacteriuria (>100 cocci/HPF; reference range, 0/HPF).

Amoxicillin trihydrate/clavulanate potassium^c (12.98 mg/kg per os [PO] q 12 hr) and carprofen^d (3.46 mg/kg PO q 24 hr) were prescribed. The dog was then referred to veterinary specialists.

The dog was evaluated by a board-certified internal medicine specialist 3 days after being examined and treated by the primary care veterinarian. The owner reported that the dog had shown a significant improvement in both attitude and energy level since being treated by the primary care veterinarian; however, over the previous 24 hr, she began exhibiting frequent posturing and stranguria without voiding any urine. Physical exam findings included normal vital signs; however, the abdomen was distended and uncomfortable on palpation. Those signs were again attributed to a large urinary bladder. Musculoskeletal and neurologic abnormalities consisted of a slight plantigrade stance; mild pelvic limb muscle atrophy with mild, bilateral proprioceptive deficits; weakness; and mild ataxia in the pelvic limbs. Cranial nerves and forelimb neurologic assessments were normal. An in-house serum biochemical profile and CBC were performed. Serum biochemistry noted resolution of the previously elevated blood urea nitrogen, and all remaining measured values were within normal limits. The CBC identified a leukopenia (3.87×10³ WBC/μL; reference range, 5.50–16.90×10³ WBC/μL). The lymphopenia and thrombocytopenia persisted, with 0.43×10³ lymphocytes/μL (reference range, 0.50–4.90×10³/μL) and 98×10³ platelets/μL (reference range, 175–500 ×10³/μL). Urine was collected via cystocentesis and submitted for urine culture.

Sedation was achieved with the IV administration of butorphanol^b (0.21 mg/kg) for an abdominal ultrasound. Ultrasonography noted a mild to moderate decrease in the definition of the corticomedullary junction in both kidneys, and the urinary bladder appeared thin-walled, with echogenic fluid and sediment. No calculi, mass lesions, or evidence of an anatomic defect in either the urinary bladder or trigone were reported. The remainder of the sonographic study was unremarkable. General anesthesia was induced with the IV administration of propofol^e (4.8 mg/kg), and the dog was maintained under anesthesia with inhaled isoflurane^f and oxygen for cystoscopy. The previously described abnormal and thickened band of tissue on the right lateral aspect of the vaginal canal was visualized, but was perceived to be an incidental finding because it was not causing obstruction of either the external urethral orifice or vaginal outflow tracts. The urethra was evaluated and devoid of strictures, calculi, or mass lesions. Thus, the cause for the inability to urinate was not determined by this study. After the procedure, the urethra was catheterized and the bladder was drained of approximately 1 L of dark brown, foul-smelling urine. Surgical placement of a cystostomy tube was recommended and the procedure was performed that same day by a board-certified veterinary surgeon.

A CBC and in-house serum biochemistry were rechecked the following day. Serum biochemistry remained normal. The CBC noted normalization of both the total WBC count (10.38×10³/μL)and lymphocytes (0.88×10³/μL. The thrombocytopenia was persistent, but had improved (148×10³/μL). The dog was discharged to continue oral amoxicillin trihydrate/clavulanate potassium^a as previously described and tramadol (75mg PO q 12 hr)^g. The urine culture resulted in positive bacterial growth and identification of Pseudomonas spp. and Acinetobacter spp. Antibiotic sensitivity testing was performed and the Pseudomonas spp. showed significant resistance to 10 of the 16 antibiotics tested. There was intermediate sensitivity to three antibiotics (piperacillin, gentamycin, and tobramycin) and sensitivity to three (tetracycline, imipenem, and amikacin). The Acinetobacter spp. showed significant resistance to 12 of the 20 antibiotics tested. It demonstrated intermediate sensitivity to two antibiotics (piperacillin and cephalexin) and sensitivity to six antibiotics (amoxicillin trihydrate/clavulanate potassium, cefovecin, cefpodoxime, imipenem, amikacin, and tobramycin). Doxycycline^h (5.19mg/kg PO q 12 hr for 4 wk) was prescribed based on those results.

MRI was recommended to further evaluate the spinal cord, and the owner was referred to the author’s facility for the procedure. The dog presented approximately 4 wk after initially being examination by the primary care veterinarian for the current problem. Anesthesia was induced with IV propofol^e and maintained with isoflurane^f and oxygen. The spine was imaged from the second thoracic vertebra (T2) to the third lumbar vertebra (L3) and from L3 to the third sacral vertebra (S3) with a 1.5 Tesla superconducting MRI systemⁱ using a phased array spine coil. The examination included sagittal T2-weighted (2 mm slice thickness, 0 mm gap), sagittal short τ inversion recovery (3 mm slice thickness, 0 mm gap), dorsal short τ inversion recovery (3 mm slice thickness, 0 mm gap), transverse T2-weighted (4 mm slice thickness, 0 mm gap), and transverse T1-weighted (4 mm slice thickness, 0 mm gap) sequences. T1-weighted sequences were obtained both prior to and after the IV administration of gadopentetate dimeglumine^j (100 mg/kg).

The sagittal T2-weighted image (Figure 1) showed a central, intramedullary hyperintensity along the majority of the spinal cord, more intense than ordinarily observed in normal dogs. The hyperintense signal was not evident in the transverse plane and was consistent with a truncation artifact. A truncation artifact can occur in MRI as a result of differences in the phase and frequency encoding steps during image acquisition, resulting in errors with image data manipulation and image reconstruction. They can appear as either bright or dark lines that are parallel to borders of abrupt intensity change (e.g., when going from bright the cerebrospinal fluid to the dark spinal cord on a T2-weighted image). Truncation artifacts can result in the appearance of conditions such as a dilated central canal or syringohydromyelia in an image when it either does not exist or is not present in orthogonal views. No evidence of the split cord malformation was apparent in the sagittal plane.

View Figure

• Download as Powerpoint
• Download Figure

The transverse T2-weighted image (Figure 2) obtained at the level of the midbody of T7 readily demonstrated sagittal splitting with a dorsoventrally oriented, midsagittal hyperintense band separating the spinal cord into two approximately equal hemicords. The split cord malformation remained persistently evident in the transverse T2-weighted image (Figure 2) obtained at the level of the caudal third of T9. Those images illustrated the continuity of the spinal cord defect. The segment of the spinal cord where sagittal splitting is most evident extends from the level of the cranial aspect of the vertebral body of T6 caudally to the level of the cranial aspect of the vertebral body of T10. Overall, the length measured 9.2 cm, with a margin of error of 0.4 cm due to slice thickness. The area of the widest separation between the two hemicords was at the level of the vertebral body of T7. The transverse measurement of the spinal cord at the point of greatest separation measured 0.62 cm, with a separation of 0.10 cm between the two hemicords. The spinal cord regained a normal, united appearance for only a short distance before a second area of sagittal splitting of the spinal cord was identified. The visibly normal segment of the spinal cord extended approximately one vertebral body in length, from the cranial aspect of the vertebral body of T5 to the cranial aspect of the vertebral body of T6. The normal appearing spinal cord had a transverse width of 0.52 cm. The second region of sagittal splitting was identified, extending from the level of the cranial aspect of the vertebral body of T2 caudally to the level of the cranial aspect of the vertebral body of T5. The spinal cord was not imaged cranial to T2. The dorsoventrally oriented intervening space separating the hemicords in this cranial segment was less well defined and appeared very narrow. Pre- and postcontrast T1-weighted images obtained at the midbody of T7 showed only an indistinct, mildly hypointense and nonenhancing vertical band between the two hemicords (Figure 3). No evidence of contrast enhancement was identified. A fibrous septum and visible evidence to support tethering of the cord to the dura were not identified. There were no other vertebral abnormalities noted throughout the region of the spine that was imaged.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

There was a single, small (0.3 cm), sharply marginated, nonenhancing, uniformly T2 hyperintense, cystic appearing structure within the neural canal associated with the left articular facet joint of L4 and L5 typical for a degenerative spinal cyst (Figure 4). Compared with the appearance of the split cord in Figure 2, the spinal cord appeared normal at this level, and the cystic structure was suspected to be an incidental finding. Additional incidental findings identified include decreased signal of the nucleus pulposus on T2-weighted images (Figure 5), consistent with either intervertebral disc degeneration or dehydration, and numerous sites of dorsal protrusion or bulging of the discs. Two areas of focal, mild to moderate spinal cord compression were identified at the intervertebral disc space between L1 and L2 (not shown) and between L3 and L4 (Figure 5) secondary to chronic appearing intervertebral disc protrusions.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The dog was reexamined by the internal medicine specialist approximately 10 days after the MRI. The dog was reported by the owner to be active, alert, comfortable, and acting herself again. The urinary bladder was being managed with the cystostomy tube. A serum biochemical profile, CBC, and urinalysis were performed. Serum biochemistry results were within the normal limits. The CBC revealed a leukopenia (3.9×10³WBC/μLand neutropenia (2,223 segmented neutrophils/μL; reference range, 3,000–11,500/μL). The previously identified lymphopenia and thrombocytopenia had resolved, and results were within normal limits. Urinalysis noted trace blood and RBCs (10–15 RBCs/HPF). No bacteria identified and remaining values were within normal reference range.

Approximately 2–3 mo after the MRI, the owners began to notice increased weakness in the hind limbs, difficulty jumping, and the hind limbs occasionally crossed over when running. The weakness persistent for approximately 6 mo longer before the owners began noticing significant and progressive decline in neurologic function of the hind limbs. Physical and neurologic examinations at that time noted a weak gait in the hind limbs; conscious proprioceptive deficits bilaterally; occasional crossing over of the limbs; and frequent falling over. Hind limb mobility continued to decline over the next month, and the dog developed fecal incontinence. The continued decline in neurologic function and difficulties associated with the long-term management of the urinary bladder with the cystostomy, as well as the use of in-dwelling catheters and recurrent urinary bacterial infections contributed to the dog’s overall declining quality of life. The owners elected humane euthanasia. The dog was euthanized approximately 10 mo after the being diagnosed with a split cord malformation via MRI. A necropsy was not performed.

Discussion

There has been significant debate regarding nomenclature and classification criteria of split cord dysraphic conditions. Some have proposed the term, “split cord malformation,” while others retain the traditional term, “diastematomyelia” when referring to splitting of the spinal cord into hemicords.^1,14–16 In brief, split cord malformation type I (diastematomyelia type I) is defined as two hemicords each within its own dural tube, separated by an osteocartilaginous median septum. In contrast, split cord malformation type II (diastematomyelia type II) is defined as two hemicords housed within a common dural tube without an intervening rigid median septum, but the cord may be separated by a nonrigid, fibrous or fibrovascular median septum.¹⁵ Three variants of split cord malformation type II have been identified: presence of an intervening fibrous septum, absence of a septum, and partial cord splitting. Absence of the septum is the most common occurrence in type II.¹ In comparison with diastematomyelia, the term diplomyelia refers to a true duplication of the spinal cord. Both conditions are considered split cord malformations.¹⁵ For purposes of diagnosis and discussion in this article and to remain consistent with similar conditions described in the recent veterinary literature, the authors have elected to apply the classification system as described by Pang (1992), which favors use of the term split cord malformation.^7,15,17

Diagnosis of split cord malformation is considered relatively straightforward both on transverse (axial) and dorsal (coronal) MRIs, but may be difficult to appreciate on sagittal images. The two hemicords can be seen clearly by the outlining of the T2-weighted hyperintense signal originating from the surrounding and intervening subarachnoid space. If present, fibrous septa can be identified on T2-weighted images as thin, hypointense stripes interposed between the two hemicords.¹ Split cord malformation in humans is most commonly located in the lower thoracic through lumbar spine, but has also been identified in other areas of the vertebral column. The length of the affected section of the spinal cord in a split cord malformation type II lesion is variable, defined in one study as 0.5–2 vertebral levels with a mean of 1.17 when measured by using the number of vertebral bodies the split segment spans on CT imaging.¹⁷ Split cord malformations can be accompanied by various other developmental abnormalities, including scoliosis, kyphoscoliosis, hemivertebra, block or butterfly vertebra, widened interpediculate distance, rib abnormalities, low-lying cord, thickened filum terminale, meningomyelocele, hydromyelia, lipomas, and dermoid cyst. Associated vertebral anomalies are usually milder in split cord malformation type II, and a split cord can also exist as the only identifiable abnormality.^1,6,17

The variance of type and severity of dysraphic states in humans leads to a wide array of clinical signs and symptoms. Commonly associated conditions include pain, atrophy of leg musculature, weakness or spasticity, proprioceptive deficits, decrease in pain and temperature appreciation in the limbs, chronic constipation or bowel incontinence, and neurogenic bladder. A full spectrum of lower urinary tract symptoms and urodynamic abnormalities can be seen in this condition, and this may reflect the variable sites of neural disruption.¹⁸ Lesions can produce abnormalities varying from no loss of nerve function to extensive upper or lower motor neuron dysfunction.¹⁹ Symptoms may include increased urinary frequency, increased urgency, feeling of incomplete voiding, postvoid dribbling, poor voluntary control, and urge or stress incontinence. Neurogenic bladder conditions are static from birth in some patients, and only recent in onset in others. Repeated urinary tract infections, likely due to incomplete voiding of the urinary bladder, is a frequent consequence.^17–19

It is important to note that many of the same clinical symptoms that occur due to split cord malformations in humans can be attributed to tethering of the spinal cord to the intervening bony or fibrous septum or to the overlying dural tube, which is tethered cord syndrome. Tethering causes traction on the spinal cord, resulting in progressive neurologic deterioration due to metabolic derangement.¹ Clinical symptoms may include neurogenic bladder with urinary retention, urinary incontinence, decreased perianal sensation, numbness in lower extremities, spastic gait, and lower extremity pain.²⁰ Tethering of the spinal cord may not be identified with diagnostic imaging, but imaging alone cannot rule out either its presence or absence. In one human study, tethering of the cord was identified during surgical procedures in 18 of 18 cases of split cord malformation type II despite no specific evidence suggesting the presence of a tethering lesion with MRI.¹⁷ Given the similarities between the symptoms of tethered cord syndrome, split cord malformation and the clinical signs in the German shepherd dog described herein, tethering of the spinal cord may be a source of, or may be contributing to, the split cord malformation disease process. Definitive proof of tethering of the spinal cord and confirmation of the imaging diagnosis of split cord malformation type II would necessitate necropsy for gross and histopathologic evaluation of the spinal cord.

Micturition disorders are common in neurologic disorders, with spinal injury among the most common cause in dogs and cats. The micturition reflex is a complex integration of parasympathetic, sympathetic, and somatic pathways extending from the sacral segments of the spinal cord to the cerebral cortex. The primary component of micturition is the detrusor reflex, and the coordinated act of urination results from the completion of a brain stem/spinal cord reflex arc that is the detrusor reflex.^21,22 A simplified review of the detrusor reflex is summarized here; however, a more extensive and detailed discussion of neuroanatomy and physiology is beyond the scope and intent of this article. The detrusor reflex begins as sensory discharge from bladder distension and ascends the spinal cord to the pontine reticular formation in the brain stem. Integration of this sensory discharge occurs here and is necessary for the detrusor reflex to be coordinated and sustained long enough for bladder evacuation. Integration leads to a motor discharge back down the spinal cord and peripherally to the bladder, causing contraction of the detrusor muscle and relaxation of the urethral sphincter.²²

The clinical signs associated with urination in the dog described in this report are consistent with what is commonly referred to as an upper motor neuron bladder. In such cases, the urinary bladder is large, turgid, and can be either very difficult or impossible to express manually. Upper motor neuron bladder dysfunction results in either the complete inability to urinate or an inability to empty the bladder by disruption of the detrusor reflex.²¹ Cortical or spinal lesions may abolish voluntary control of the sphincter and may produce increased sphincter tone, which increases outflow resistance. Typically, spinal cord lesions from T3 to L3 abolish the long-routed detrusor reflex and cause increased tone in the sphincter.²² In addition to the upper motor neuron deficits, chronic overdistension of the urinary bladder, resulting in detrusor atony and recurrent urinary tract infections, may have contributed to both the previous and recent episodes of acute decompensation in the case described here. Urodynamic and electrophysiologic studies are used in human medicine, but only infrequently in veterinary medicine, to evaluate functional integrity of the neuromuscular structures of the lower urinary system. Although they were not performed in this case, those diagnostic tests may have been useful to more definitively localize a neurologic lesion as a neuromuscular lesion, upper motor neuron, lower motor neuron, or a mixed lesion involving both upper and lower motor neuron dysfunction. Additional diagnostics that might have been considered in this case include a cerebrospinal fluid analysis and a serologic evaluation for tick-borne disease, but these were not performed. The cerebrospinal fluid analysis would have been useful to evaluate for inflammatory and infectious causes of the central nervous system as a cause of the dog’s clinical signs. With respect to hematologic abnormalities (that included lymphopenia and thrombocytopenia at the beginning of the case report), the additional differential diagnosis of tick-borne disease that can result in myelitis should also be included. Serology would be useful to either include or exclude tick-borne disease more definitively; however, the dog was treated with doxycycline and showed only brief improvement in general attitude, and an overall clinical decline persisted despite its use. The additional clinical signs associated with the pelvic limbs that were described may also be due to the split cord malformation because similar limb and gait abnormalities have been described in humans with this condition. Alternative causes of the clinical signs associated with the pelvic limbs, specifically degenerative myelopathy given the predisposition in German shepherd dogs, cannot be ruled out as the cause some or all of the clinical signs based on these imaging findings. It is unlikely that the described urinary problem can be attributed solely to degenerative myelopathy as the urinary issues were reported since the time of adoption, when the dog was very young. Although the urinary signs were noted throughout the dog’s life, it is unclear if the more subtle pelvic limb deficits were not recognized by the owner and were also present for the same duration or if they were more recent in onset.

Conclusion

The lack of anatomic abnormalities associated with either the urinary bladder or urethra and the absence of significant metabolic abnormalities led the attending veterinarian to pursue MRI of the spinal cord. The diagnosis of a split cord malformation was made based on the MRI, which identified the spinal cord lesion, the clinical signs, and history of chronic urinary abnormalities. The split cord lesion is a likely cause for the life-long history of dysuria and the lack of posturing to defecate. The split cord malformation may have also been a significant predisposing factor to the episodes of bladder atony and recurrent urinary tract infections. As the authors have previously reviewed, those clinical signs are similar to many of the documented clinical conditions and consequences of the presence of a split cord malformation in humans. It is unknown whether the split cord malformation played a role or contributed to the progressive neurologic decline in the dog described in this report; however, it is very unlikely that the more recent, severe, and rapidly progressive neurologic deterioration occurring after the MRI was performed are solely attributed to the presence of the split cord malformation. Rapid progression of clinical signs has not been documented in people with split cord malformations. Alternative differentials that are more common in geriatric dogs and dogs of this breed are more likely and must be considered as the primary cause for the acute, progressive decline in neurologic status and quality of life observed in this dog.