Hemilaminectomy, Diverticular Marsupialization, and Vertebral Stabilization for Thoracolumbar Spinal Arachnoid Diverticula in Five Dogs
ABSTRACT
Medical records (2004–2016) of five dogs with a thoracolumbar spinal arachnoid diverticula (SAD) that was diagnosed with stress myelography in four dogs and magnetic resonance imaging in three, and who had hemilaminectomy, diverticular marsupialization, and vertebral stabilization, were reviewed. Data on previous treatment, pre- and postoperative neurologic status, diagnostic findings, surgical techniques, and outcomes was retrieved. Follow-up clinical and radiographic evaluations were performed immediately; ∼1, 2, and 6 mo postoperatively; and at annual follow-up examinations.
The stress myelography demonstrated spinal cord dynamic compression in three of four dogs and change in size or shape of the SAD in all four. Two dogs who had SAD recurrences 4 and 13 mo after previous surgical dural fenestration, and one dog with no previous SAD treatment demonstrated long-term neurological improvement after vertebral stabilization (49, 77, and 126 mo). In two other dogs, recurrence of clinical signs was observed at the follow-up (8 and 12 mo).
This case series suggested that repetitive spinal cord injury from the dynamic lesion appears to be one potential cause of thoracolumbar SADs. In cases with dynamic lesions confirmed by stress myelography, vertebral stabilization with conventional techniques is indicated to prevent SAD recurrence.
Introduction
Spinal arachnoid diverticula (SAD) refers to a dilation of a focal area of the meninges in which cerebrospinal fluid (CSF) accumulates, with its pathophysiological mechanism suspected to be congenital (primary) or acquired (secondary).1 The pathophysiologic mechanism of congenital diverticula formation is uncertain, but abnormal development of the arachnoid membrane such as duplicating or splitting during the embryonic stage has been proposed, thereby allowing for its progressive expansion and secondary spinal tissue damage.1 It has been suggested that acquired arachnoid diverticula arises secondary to intervertebral disc disease, vertebral anomalies, spinal cord trauma, and inflammatory spinal diseases, although the mechanism has not been clearly confirmed.1,2 In addition, scarring, inflammation, or fibrin accumulation may have a role in diverticula formation.1
In previous studies of SAD, the most commonly affected area was the thoracolumbar region in small-to-medium–breed dogs.1 There was a tendency for affected dogs to have concurrent spinal disorders, such as intervertebral disc extrusion and/or protrusion, vertebral canal stenosis, and/or malformation at the same or adjacent vertebral level of SAD.3
Recurrences of SAD have frequently occurred.1 The cause of SAD recurrence has not been fully determined. There have been no surgical techniques that have proved superior to others in reducing the rate of SAD recurrence.4 Our previous study on congenital thoracic vertebral anomaly included a dog (dog 1 in the present study) with SAD recurrence at the level of a kyphosis from butterfly vertebrae 4 mo after initial mini-hemilaminectomy and dural fenestration by the referring veterinarian.2 The dog improved and maintained the ability to ambulate without neurological abnormalities after a second surgery in which a hemilaminectomy, marsupialization, and vertebral stabilization were performed.2 Vertebral anomalies are often associated with vertebral instability, which can cause chronic irritation of the arachnoid membrane.2 Intervertebral disc degeneration also results in disc-associated dynamic compression, disc herniation, or vertebral instability.5
We hypothesized that repetitive spinal cord injury as a result of dynamic lesion is one potential cause of acquired SAD formation, and that vertebral stabilization may prevent SAD recurrence in such cases.2 Given the suspected role of dynamic lesion in SAD formation and recurrence, we report the results of vertebral stabilization in a series of five dogs with SAD.
Materials and Methods
Medical records of dogs presented to the Aikawa Veterinary Medical Center who had thoracolumbar SADs and underwent a hemilaminectomy, marsupialization, and vertebral stabilization during the period of 2004–2016 were reviewed retrospectively. Data on age, breed, gender, body weight, pre- and postoperative neurologic status, diagnostic findings, surgery, and outcomes was retrieved. Follow-up neurologic and radiographic evaluations were performed at ∼1, 2, and 6 mo. Long-term outcomes were assessed by means of clinical examination or owner telephone interviews.
Neurologic Grading
Dogs were assigned a grade (G) of G0–G5 based on the severity of the neurologic dysfunction, in which G0 represents a clinically normal dog, G1 represents thoracolumbar paraspinal pain without neurologic deficits, and G2 represents ambulatory paraparesis. G2 was further classified as mild, moderate, and severe on the basis of the subjective assessment of the severity of ataxia; mild represents ambulatory paraparesis in which the dog can walk and run but the clinician can detect a slight to mild ataxia or conscious proprioception deficit; moderate represents ambulatory paraparesis with ataxia; and severe represents ambulatory paraparesis with obvious ataxia in which the dog has the ability to stand up and take several steps (>5 m) without support. G3 represents nonambulatory paraparesis. G4 represents paraplegia with intact or positive deep nociception in pelvic limbs and tail. G5 represents paraplegia with absent deep nociception in both pelvic limbs and tail. The assessment of the tail was conducted for dogs with a tail.6,7
Anesthetic Protocol
The anesthetic protocol included premedication with glycopyrrolatea (0.01 mg/kg intramuscularly [IM]), and morphine hydrochlorideb (0.5 mg/kg IM), and preoxygenation for 10 min before induction with midazolamc (0.3 mg/kg IV) and propofold (2 mg/kg IV, to effect). Anesthesia was maintained with isofluranee (1.5–2.0%) in oxygen with the aid of a ventilator. Cefazolin sodiumf (22 mg/kg IV) was administered 20 min before surgery and 90 min thereafter during surgery. On recovery from anesthesia, dogs were closely monitored and additional morphine hydrochlorideb (0.5 mg/kg IM) was administered as needed for analgesia.
Imaging and Myelographic Studies
Survey radiographic projections of the spine obtained under anesthesia were evaluated and the abnormal findings were recorded.
Lumbar myelography and/or MRI were used to diagnose SAD in all dogs, and a stress myelographic study was performed to definitively diagnose dynamic lesion.5 For lumbar myelography, including the stress study, iohexolg was injected into the subarachnoid space at the L5-6 interspace using a spinal needle. The amount of contrast agent administered was determined by observing for filling of the entire spinal subarachnoid space using fluoroscopy. Dynamic spinal cord compression was assessed by applying manual extension and flexion to the spine with the dog positioned in lateral recumbency in cases with inconclusive spinal cord compression in the neutral position. To evaluate dynamic compression, pre- and post-stress spinal cord height at the intervertebral disc spaces were measured from lateral radiographic projections and the percentage of reduction of spinal cord height was recorded. Change in size or shape of SADs on the lateral stress myelographic study was recorded. Ventrodorsal and oblique projections of the myelographic study were also evaluated.
MRI was also performed in three dogs using a 0.4 Tesla scannerh. T2-weighted images were acquired using a fast spin-echo sequence at a repetition time of 2000–2500 ms, and an echo time of 112–120 ms. T1-weighted spin-echo images were acquired at a repetition time of 405–406 ms and an echo time of 13 ms before and after administration of 0.2 mL/kg body weight gadodiamidei. SADs were demonstrated as a hyperintense fluid-filled cystic region on T2-weighted images that were hypointense on T1-weighted images. The images were also evaluated for the presence of other spinal cord disease.
Surgical Procedure
A hemilaminectomy was performed using nitrogen-driven surgical burs. After identifying the fluid-filled cystic region in the subarachnoid space, marsupialization was performed. Unilateral vertebral stabilization was performed using positively threaded profile pinsj with polymethylmethacrylate (PMMA) bone cementk. Pins were driven into the vertebral body. The extent of the vertebral stabilization and the sizes and number of pins used were based on the number of affected discs, patient size, and age. After hemilaminectomy, a subcutaneous fat graft was applied to the defect to prevent adhesions. Cefazolinf powder (250 mg) was added to 20 g of the PMMA polymer powder before mixing with the liquid monomer. Copious saline flush was used to minimize thermal injury to the spinal cord and adjacent tissue during PMMA polymerization. The dorsal spinal fascia and subcutaneous tissue layers were closed in a continuous pattern using polydioxanonel and the skin apposed with nylon.
Postoperative Care and Follow-Up
Postoperative radiographs were obtained immediately after surgery. All dogs were hospitalized for 4–7 days. Postoperative neurological status was evaluated daily for any evidence of neurological deterioration compared with the preoperative neurological grade. All dogs were re-examined 10–14 days postoperatively at the time of suture removal. Monthly or bimonthly reexaminations and radiographic evaluations were conducted by the authors for 3–6 mo, and annually thereafter. For dogs who were not re-examined by the authors for more than 12 mo from the previous examination, annual follow-up was made by telephone interview with the owner or referring veterinarian.
Results
Signalment
Five dogs met the inclusion criteria (Table 1). The breeds represented included a French bulldog, pug, Italian greyhound, shih tzu, and Shiba Inu. There were two males and three females (one spayed female). The mean age was 72.8 mo (median = 72 mo, range = 13–113 mo). The median body weight was 9.0 kg (range = 4.3–9.6 kg).
Patient’s History and Presenting Clinical Signs
Three dogs had a history of ambulatory paraparesis (G2) associated with SAD and were surgically treated by dural fenestration of SAD at other hospitals or our hospital. Dog 1 had previous surgery at another hospital 7 mo prior to presentation to the authors and had initially improved neurologically. However, her clinical signs deteriorated to a moderate G2 4 mo after the initial surgery. Dog 4 had previous surgery in the authors’ hospital 51 mo previously, and her clinical signs improved to mild G2 but gradually deteriorated to nonambulatory paraparesis (G3) 6 mo after the initial surgery. Dog 5 had surgery at another hospital 33 mo previously and had improved neurologically, but her clinical sign deteriorated to moderate G2 13 mo after the initial surgery.
The remaining two dogs had no history of previous surgery. One dog (dog 2) had a 1 yr history of G2 that progressed to G3 1 mo prior to surgical intervention, and one dog (dog 3) had an acute onset of moderate-to-severe G2, back pain, and urinary incontinence 1 mo prior to surgical intervention.
Imaging Findings
On plain radiography, vertebral luxation was not observed in any dogs. Butterfly vertebrae and malalignment on T 8-10 with moderate kyphosis in dog 1 and spondylosis deformans on T 12-13 in dog 5 were observed. Three dogs underwent MRI before (dogs 3 and 4) or after (dog 2) myelography. A fluid-filled SAD in the dorsal area of spinal cord, which was hyperintense on T2-weighted images, hypointense on T1-weighted images, and displayed no contrast enhancement after administration of IV gadolinium, was observed in all dogs. Dogs 2 and 4 also had an intramedullary focal hypointense region that was isointense to CSF at the level of SADs on T1-weighted images and was hyperintense on T2-weighted images. The central canal appeared to be involved in this parenchymal lesion and communicated with SAD (Figure 1).
Citation: Journal of the American Animal Hospital Association 55, 2; 10.5326/JAAHA-MS-6762
Myelography performed with the spine in a neutral position revealed the presence of SAD in four dogs. One dog (dog 2) had a diffuse region of the subarachnoid space that filled with contrast but was not conclusive for a SAD (Figure 2); thus, the stress myelographic study was not performed, and the dog underwent MRI. On the stress myelographic study performed on four dogs, dynamic changes in the size or shape of the cystic region were observed when the spine was flexed or extended. A reduction in post-stress spinal cord height was seen in three dogs (dogs 3–5), with a height reduction ranging from 8.0 to 11.0% (mean = 9.5%). The post-stress spinal cord compression was observed while positioned in extension (dog 3), flexion (dog 4), and both flexion and extension (dog 5; Figure 3).
Citation: Journal of the American Animal Hospital Association 55, 2; 10.5326/JAAHA-MS-6762
Citation: Journal of the American Animal Hospital Association 55, 2; 10.5326/JAAHA-MS-6762
CSF Analysis
Lumbar CSF was collected in three dogs before myelography (dogs 2–4). CSF analysis identified a mild increase in protein concentration (51 mg/dL, range < 45 mg/dL) in dog 3 and normal protein concentration in the others. The total nucleated mononuclear cell count was within the normal range in all three dogs (range < 5 WBC/μL).
Surgery
All dogs underwent surgery for hemilaminectomy, marsupialization, and vertebral stabilization. Vertebral stabilization of two vertebrae was performed in two dogs (dogs 3 and 4), three vertebrae in one dog (dog 5), five vertebrae in one dog (dog 1), and six vertebrae in one dog (dog 2). The stabilization was performed with four to eight pins (1.2–2.4 mm) and PMMA in all dogs (Figure 4).
Citation: Journal of the American Animal Hospital Association 55, 2; 10.5326/JAAHA-MS-6762
Outcome
Dog 1 maintained the ability to ambulate immediately after surgery, and her neurological deficits resolved by 11 mo postoperatively. The dog retained normal ambulation until 84 mo postoperatively, when she had signs of moderate paraparesis. The dog remained ambulant with mild paraparesis at the time of last follow-up (126 mo). Dog 2 initially improved to a moderate G2 postoperatively but lost the ability to ambulate 8 mo postoperatively. The dog’s neurological status was unimproved when the dog died of hepatic cancer at 37 mo postoperatively. Dog 3 improved to a mild G2 and remained at this grade until the dog died of unrelated diseases 77 mo postoperatively. Dog 4 improved to a moderate G2 postoperatively, but then lost the ability to ambulate 12 mo postoperatively. No improvement was observed at the last follow-up (81 mo). Dog 5 lost the ability to ambulate postoperatively but regained the ability to ambulate within 1 mo. The dog was neurologically normal at the last follow-up (49 mo).
In all five dogs, follow-up radiographs taken on subsequent examinations were unremarkable.
Discussion
In the present study, we identified SAD in five small-to-medium–sized dogs with clinical signs of urinary incontinence and slowly progressive paraparesis ranging from moderate ambulatory to nonambulatory paraparesis. Three dogs temporally improved after previous dural fenestration of SAD, but all three dogs had a recurrence of clinical signs 4–13 mo after the initial surgery and subsequently had second stabilization surgery. Myelography and MRI demonstrated the presence of SAD in the dorsal aspect of thoracic spinal cord. The location of SAD was similar to those in previous reports.1,3 The stress myelography was used to assess the dynamic lesion associated with the SAD. In four dogs, stress myelography demonstrated a dynamic change in the size or shape of SAD in all dogs and dynamic spinal cord compression in three dogs. SAD was observed at the same level of kyphosis from butterfly vertebrae (dog 1) and spondylosis deformans (dog 5). Although dog 1 did not show detectable dynamic compression, vertebral stabilization was performed based on SAD change in size and the potential vertebral instability in the kyphotic segment of the malformed spine.2
Stress myelography may be the only way to assess the dynamic lesion. Although it is somewhat subjective and not a very sensitive method of diagnosing the dynamic lesion, the dynamic spinal cord compression was exclusively diagnosed in the affected segment but not in the adjacent segments, suggesting this method may have some diagnostic value.5
Stress myelography has not been routinely used in the diagnosis of SAD because its pathophysiological mechanism relating to dynamic lesion has not been confirmed. The tendency for the cases in the present study to have concurrent dynamic lesion may suggest that the dynamic lesion was related to SAD formation in some cases, and it may have been overlooked in previous reports in which stress myelography was not performed.
Treatment of SAD has historically involved diverticulum fenestration and marsupialization.1 The larger studies of surgically treated SAD involving 15, 12, and 11 dogs reported the overall short-term successful outcome in 29 of 38 dogs.4,8,9 The first study reported long-term (>1 yr follow-up) outcome in 12 of 15 dogs, with a successful outcome in 8 (66%) and a median follow-up time of 28 mo (range = 14–72 mo).4 Two other studies reported success, with 8 of 12 (66%) (6–30 mo follow-up) and 7 of 11 (64%) dogs (1–36 mo follow-up) having good functional recovery, and none of the dogs developed recurrence of clinical signs.8,9 In the first study, recurrence of clinical signs was observed in 3 of 12 dogs (25%) 14, 18, and 26 mo after surgery, respectively, suggesting the recurrence rate may have been underestimated because of the limited follow-up in previous reports.1,4 In all surgically treated dogs in these three reports, only 3 dogs with good outcome followed longer than 3 yr (45, 63, and 72 mo).4,8,9
In the present study, vertebral stabilization was performed with hemilaminectomy and marsupialization in all five dogs, in which the dynamic lesion was confirmed in four dogs and unconfirmed in one dog. Three dogs with preoperative G2 (dogs 1, 3, and 5) improved postoperatively and maintained the ability to ambulate for a long term (follow-up 49, 77, and 126 mo). On the other hand, two dogs with preoperative G3 improved to G2 postoperatively but had signs of recurrence and became G3 8 mo (dog 2) and 12 mo (dog 4) after surgery. In these two dogs, no abnormalities were observed on follow-up X-rays, but follow-up computed tomography or MRI evaluation was not performed. Preoperative MRI of these two dogs revealed indistinct changes in the spinal cord adjacent to SAD that was hyperintense on T2-weighted images, hypointense on T1-weighted images, and demonstrated no contrast enhancement. Conversely, preoperative MRI of the dog with G2 demonstrated that SAD was not associated with any intramedullary changes. This finding may suggest that an SAD that has evidence of an adjacent intramedullary lesion on MRI may have a worse prognosis. In previous studies, the presence of syringomyelia/hydromyelia concurrent with SAD has been associated with a negative prognosis.4,10,11 Therefore, preoperative MRI may be helpful in evaluating concurrent diseases and predicting prognosis, especially when the dog’s neurological grade is G3 or worse. Additional MRI sequences such as the Fluid-Attenuated Inversion Recovery or Half-Fourier-Acquired Single-shot Turbo spin Echo may have provided more information of the concurrent intrathecal lesions.12
Long-term neurological improvement after vertebral stabilization was achieved in three dogs (median follow-up = 77 mo), two of whom had a history of postoperative SAD recurrences after conventional surgery within 13 mo. Therefore, the dynamic lesion might have existed as an underlying cause, and vertebral stabilization might have prevented SAD recurrence in these cases. In contrast, two dogs had recurrent neurological sign within 12 mo postoperatively, suggesting that not all SAD cases respond to the vertebral stabilization.
This study, being retrospective in nature, does not include control cases and does not fully evaluate the dynamic lesion in one dog and cause of recurrent neurological sign in two dogs. Therefore, it cannot be conclusively stated that dynamic lesion played a role in the development of disease in all dogs in whom neurological improvement was observed. In addition, dynamic lesion was subjectively evaluated by the authors using stress myelography, and it was uncertain whether any dynamic lesion existed prior to SAD formation. Additional studies involving a larger number of dogs are needed to further evaluate the role of dynamic lesion in the development of SAD.
Conclusion
This case series suggested the repetitive spinal cord injury from dynamic lesion appears to be one potential cause of thoracolumbar SAD. In cases with dynamic lesion confirmed by stress myelography, vertebral stabilization with conventional techniques is indicated to prevent the risk of SAD recurrence.
T1-weighted images at the level of SAD. (A) Dog 3 had a focal hypointensity in the region of SAD that was noted to be compressing the spinal cord from the dorsal aspect. (B) Dog 2 and (C) Dog 4 had a focal intramedullary hypointense region indistinctly communicating with SAD. This finding may suggest that SAD infiltrated into the spinal cord and/or communicated with the central canal, or that there is concurrent syringomyelia/hydromyelia or other parenchymal disorders. SAD, spinal arachnoid diverticula.
Dog 2 had diffuse contrast enhancement from the level of T6 to T13 that was not conclusive for an SAD. SAD, spinal arachnoid diverticula.
Dog 5. (A) Neutral myelogram demonstrating a contrast-filled diverticular lesion in the dorsal subarachnoid space at T12 to T13. (B) Stress myelogram with dorsal extension. (C) Flexion showing moderate spinal cord compression due to dynamic compression. Change in SAD size and shape, and reduction in spinal cord height (11%) were noted. SAD, spinal arachnoid diverticula.
Dog 3. (A) Intraoperative view and (B) postoperative radiograph, showing six positively threaded profile pins laterally inserted into the vertebral bodies of T12 and T13.
Contributor Notes
