Materials and Methods

In a previous study,⁸ results of neurologic and radiologic examinations of 60 working GSDs were reported. Thirty-three of these 60 dogs were included in the present study. The remaining 27 dogs were no longer available for study because they had died (n = 15) or been lost to follow-up (12). Dogs included in the present study consisted of 28 males (all sexually intact) and 5 females (4 sexually intact and 1 spayed).

The 33 dogs were reexamined approximately 3 years after the initial examination. Mean ± SD age at the time of initial examination⁸ was 3.5 ± 1.7 years, and mean age at the time of examination for the present study was 7.0 ± 1.7 years. Mean time between examinations was 41 ± 8.0 months. All dogs were regularly used as patrol and attack dogs at a Swiss police unit during the time between examinations.

For the present study, owners answered a written questionnaire asking for information about the general health status of their dog and any concurrent disorders diagnosed by their veterinarian. With respect to potential lumbosacral disease, owners were specifically asked to indicate whether their dogs were able to perform their expected duties without any restrictions. Owners were also asked to indicate whether their dogs had had any signs of pain, lameness, or weakness of the pelvic limbs or tail.

Each dog underwent a neurologic examination that included careful assessments of gait, posture, anal and tail tone, and quality of conscious proprioception and spinal reflexes (ie, patellar, cranial tibial, and withdrawal reflexes). Lumbosacral hyperesthesia was evaluated by means of deep palpation, lumbosacral hyperextension, and hyperextension of the tail. Gait and posture were graded as normal (grade 0) or as mildly (grade 1), moderately (grade 2), or severely (grade 3) abnormal. Anal tone, tail tone, conscious proprioception, spinal reflexes, and response to lumbosacral palpation and manipulation were scored as normal (grade 0) or abnormal (grade 1). A presumptive diagnosis of DLSS was made in dogs with compatible clinical signs, with reproducible signs of lumbosacral pain being the minimal requirement, following exclusion of orthopedic and other neurologic disorders that might mimic the condition. In all dogs, the examination was performed by the same board-certified neurologist (FS) who had performed neurologic examinations of the dogs for the previous study.⁸

Following the neurologic examination, all dogs underwent survey radiography of the lumbosacral junction. For this radiographic evaluation, dogs were sedated with medetomidine (5 to 10 μg/kg [2.3 to 4.5 μg/lb], IV) and propofol to effect. Atipamezole (5 to 10 μg/kg, IV) was administered at the end of the radiographic evaluation to antagonize the effects of medetomidine.

Two radiographic views centered on the lumbosacral junction were obtained.^a One was obtained with the dog in lateral recumbency and the femurs perpendicular to the vertebral column; the other was obtained with the dog in dorsal recumbency, with the stifle joints abducted and the X-ray beam angled cranially by 7° to 10° so that it was parallel to the lumbosacral end-plates. Radiographic positions were the same as those obtained for the previous study,⁸ and radiographs obtained for the present study were compared with those obtained 3 years earlier. Radiographs were assessed by a board-certified radiologist (GS) for evidence of lumbosacral transitional vertebrae, spondylosis deformans, spondylarthrosis, and lumbosacral misalignment.^13,14 Spondylosis deformans was scored as absent (grade 0) or present (grade 1; presence of bony spurs, ossification cores between vertebral bodies, or bridging spurs).¹⁵ Spondylarthrosis was also scored as absent (grade 0) or present (grade 1; osteophytes, sclerosis, enlargement of the facet, and widening of the articular space). Lumbosacral misalignment was measured at the level of the cranial endplate of S1 as the distance between 2 lines drawn parallel to the dorsal cortices of the bodies of L7 and S1. Width of the L7-S1 intervertebral space was measured as the distance between the endplates of the vertebral bodies at the axial center of the disk along a line dividing the vertebral body of S1 into equal halves. If present, the dorsoventral extension of any area of mineralization within the L7-S1 disk was measured. The height of the spinal canal was measured at the level of the caudal endplate of L7 and the cranial endplate of S1. The height of the body of L7 was measured in the middle of L7 and at the level of the caudal endplate. The axial widths of L7 and S1 endplates were measured, and a score of 0 (≤ 2 mm) or 1 (> 2 mm) was assigned. Shape and contour of the cranial endplate of S1 were evaluated, with particular attention paid to any dorsal flattening of S1 (grade 0, no flattening evident; grade 1, flattening evident). Contour of the cranial endplate was graded as normal (grade 0) or abnormal (grade 1; rough or irregular contour). The presence of opacification or bony fragments within the spinal canal was recorded as absent (grade 0) or present (grade 1). The width of the L7-S1 intervertebral disk space was normalized by dividing the measured distance by the height of the spinal canal at the level of the caudal endplate of L7. The radiologist was blinded to previous radiographic findings and to results of the neurologic examination.

Statistical analysis—Descriptive statistics (mean and SD) were computed. The Wilcoxon signed rank test was used to compare clinical and radiographic findings from the original examination with findings for the present study. Forward and backward multiple logistic regression analysis was performed to test for associations between radiographic signs observed during the original examination and diagnosis of DLSS. All analyses were performed with standard software.^b,c Values of P < 0.05 were considered significant.

Results

Owners of 13 of the 33 (39%) dogs reported problems potentially associated with a lumbosacral disorder. This included 11 (33%) dogs with mild lameness or weakness of the pelvic limbs, 10 (30%) dogs with difficulties jumping, and 7 (21%) dogs with signs of lower back pain. Owners of 22 of the 33 (67%) dogs reported that their dogs were able to perform their duties without any restrictions, and owners of an additional 4 (12%) dogs reported that their dogs were able to perform restricted duties (ie, dogs were used on patrol but were not used or trained for attack duties). Reasons for restricted duties were reported as DLSS in 3 dogs and thoracolumbar disk disease in 1. Seven dogs had been excluded from active duty because of DLSS (3 with confirmed and 3 with suspected DLSS) or other reasons (1).

In 15 of the 33 (45%) dogs, a presumptive diagnosis of DLSS was made on the basis of results of the neurologic examination. In all 15, signs of pain were elicited by means of digital palpation of the lumbosacral area and hyperextension of the lumbosacral junction. Additional signs included abnormalities of gait in 8 dogs (24%; grade 1 in 2 dogs, grade 2 in 3 dogs, and grade 3 in 3 dogs) and presence of mild neurologic deficits, including proprioceptive deficits, in 5 (15%) dogs. In 11 of these 15 dogs, a diagnosis of DLSS had also been made at the time of the original examination 3 years earlier, and in 9 of the 11, compression of the cauda equina had subsequently been confirmed by means of myelography, epidurography, computed tomography, magnetic resonance imaging, or surgery. Of the 11 dogs in which DLSS had been diagnosed at the time of the original examination, 4 were on unrestricted duty, 2 were on restricted duty, and 5 were excluded from active duty because of DLSS.

Musculoskeletal abnormalities unrelated to DLSS that were identified included fibrosis of the semitendinosus muscle (n = 2 dogs), thoracolumbar disk disease (1), rupture of the cranial cruciate ligament (2), forelimb lameness secondary to elbow arthrosis (4), and pelvic limb lameness secondary to hip dysplasia (7). No neurologic abnormalities unrelated to DLSS were identified.

Several dogs with clinical signs of DLSS had undergone medical or surgical treatment in the time between the original and present examinations. Nonsteroidal anti-inflammatory drugs had been administered intermittently to 3 dogs (2 with a good response and 1 without any response) and continuously to 1 (with a good response). One dog had been treated with a long-acting corticosteroid every third month with satisfactory clinical results.

Surgical decompression of the cauda equina and nerve roots was performed in 2 dogs by means of a standard dorsal laminectomy. The outcome in 1 was excellent initially, but the dog was restricted from duties 1 year after surgery and had reproducible signs of pain on lumbosacral palpation and extension. The other dog was euthanatized 6 months after surgery because of poorly specified reasons unrelated to DLSS; a necropsy was not performed.

Thirteen of the 15 dogs with DLSS had radiographic abnormalities of the lumbosacral junction, as did 14 of the 18 dogs without clinical signs of DLSS. Radiographic abnormalities included mineralization of the L7-S1 intervertebral disk (n = 12), sclerosis and thickening of the lumbosacral endplates (4), spondylosis deformans (25), and mineralization within the spinal canal (7). Lumbosacral transitional vertebrae were seen in 2 dogs, and lumbosacral misalignment was present in 16 dogs. Mean ± SD dorsoventral extent of mineralization within the L7-S1 disk was 3.1 ± 6.1 mm. Mean normalized width of the L7-S1 intervertebral disk space (ie, width of the disk space divided by height of the spinal canal) was 0.4 ± 0.05. Mean height of the spinal canal at the level of the caudal endplate of L7 was 11.4 ± 1.23 mm, and mean height of the spinal canal at the level of the cranial endplate of S1 was 7.6 ± 1.08 mm. Mean lumbosacral misalignment was 0.44 ± 1.04 mm. Mean height of the body of L7 at the level of the caudal endplate was 23.3 ± 1.46 mm.

When clinical signs observed during the original and present examinations were compared, there were significant increases in scores for gait (P = 0.011; higher scores were indicative of more severe gait abnormalities), quality of conscious proprioception (P = 0.025; scored as 0 [normal] or 1 [abnormal]), and response to lumbosacral palpation and manipulation (P = 0.002; scored as 0 [normal] or 1 [abnormal]). Scores for anal tone, tail tone, and spinal reflexes were not significantly different between observation times.

Similarly, when radiographic signs observed during the original and present examinations were compared, significant differences included an increase in the mean dorsoventral extent of areas of mineralization within the L7-S1 disk space (P ≤ 0.001), a decrease in mean normalized width of the L7-S1 intervertebral disk space (P ≤ 0.001), increases in scores for spondylosis deformans (P < 0.001; scored as 0 [absent] or 1 [present]) and mineralization within the spinal canal (P = 0.025; scored as 0 [absent] or 1 [present]), a decrease in mean height of the spinal canal at the level of the caudal end-plate of L7 (P = 0.004), and an increase in the mean height of the spinal canal at the level of the cranial end-plate of S1 (P = 0.001). No significant differences were found between observation times with regard to dorsal flattening of S1, contour of the cranial endplate of S1, axial widths of the L7 and S1 endplates, lumbosacral misalignment, spondylarthrosis, or height of the body of L7 at the level of the caudal endplate.

Multiple logistic regression analysis did not identify any radiographic signs that could be used to predict the development of DLSS.

Discussion

The most common and typically earliest finding in dogs with DLSS is signs of pain during palpation and hyperextension of the lumbosacral junction. Although not specific for DLSS, this test is easy to perform and highly sensitive, with positive results in 91% to 100% of affected dogs.^1,2,16 Most often, pain arises as a result of compression of the nerve roots of the cauda equina, although other potential sources of pain include the lumbosacral disk and the articular facets. Although the diagnosis of DLSS was not confirmed by means of advanced diagnostic imaging in all dogs in the present study, the presumptive diagnosis was made on the basis of repeated neurologic examinations and by exclusion of other diseases that may mimic the disease.

Although radiographic signs suggestive of lumbosacral degeneration have been shown to be unreliable for the diagnosis of DLSS in previous reports,^9,10 the predictive value of radiography has not been assessed previously. In an era of advanced diagnostic imaging, the role of biplanar radiography for the diagnosis of spinal disorders becomes more and more marginal, owing to its poor ability to visualize soft tissue and foraminal abnormalities. However, studies on spinal disorders have demonstrated the value of survey radiography despite these limitations. For instance, in a prospective study¹⁷ of clinically normal Doberman Pinschers with radiographic signs of cervical spondylomyelopathy, 71.8% developed clinical signs of the disorder within the following 5 years, indicating the predictive value of screening radiography. In a recent human study,¹⁸ bi-planar radiography was able to distinguish early stages of lumbar disk degeneration, whereas magnetic resonance imaging could only detect advanced degeneration, and it was concluded that radiography was a cost-effective method for detecting early disk degeneration. In the present study, significant progression of various radiographic signs suggestive of lumbosacral degeneration was identified. However, none of the radiographic signs was significantly associated with clinical signs of DLSS. This finding was supported by the fact that dogs without clinical signs of DLSS had similar increases in radiographic signs of lumbosacral degeneration. Thus, our findings support the suggestion that survey radiography is of low diagnostic value not only for the contemporaneous diagnosis of DLSS but also for prospective identification of dogs at risk of developing the disease.

Misalignment of the lumbosacral junction (ie, ventral subluxation of S1 relative to L7) deserves special comment. Lumbosacral misalignment has been considered to be a sign of instability by several authors,^19,20 whereas others found no significant differences between affected and unaffected dogs.^8,10 Schmid and Lang¹⁰ found that ventral subluxation of > 4 mm was strongly suggestive of an abnormal lumbosacral junction, but Suwankong et al²⁰ suggested that a lumbosacral step as small as 2 mm may be clinically relevant. However, because dogs in these previous studies were clinically abnormal or were not followed up clinically, statements about the predictive value of this finding were not possible. In the present study, 16 dogs had lumbosacral misalignment, with the degree of misalignment ranging from 1 to 3 mm. When radiographs obtained for the present study were compared with the original radiographs for these dogs, no progression in the misalignment was noted radiographically and a lumbosacral step was not associated with clinical signs of DLSS. Thus, results of the present study indicate that the presence of a lumbosacral step on a lateral radiographic view does not represent a relevant clinical finding in regard to the diagnosis or possible subsequent development of DLSS.

The association between radiographically apparent spondylosis deformans and clinical signs of DLSS is controversial in the veterinary literature. Spondylosis deformans and type II disk disease were suspected to be associated in a recent study,²¹ and higher rates of spondylosis at sites of type II disk protrusion have been found.²² In contrast, our results confirm the finding of Scharf et al⁸ that radiographically apparent spondylosis deformans is not consistently associated with clinical signs of lumbosacral disease. Spondylosis was a common radiographic finding in dogs with DLSS in the present study, but it was not associated with clinical signs nor did it predict future development of the disease.

The number of dogs with clinical signs of DLSS increased from 37% to 46% over the 3 years from the previous⁸ to the present study. Significant progressions in gait abnormalities and conscious proprioceptive deficits reflected the progressive nature of the disease, and our findings are in accordance with the findings of other in-vestigators^1,23 that older dogs are more commonly and more severely affected than younger ones. Of the 33 GSDs reevaluated for the present study, 22 were able to perform unrestricted duty. Four dogs were in restricted duty, 6 dogs were excluded from duty because of DLSS, and 1 dog was excluded from duty for other reasons. All told, 9 of the 33 (27%) dogs were excluded from duty or on restricted duty because of DLSS. In a previous study,³ 19.4% of GSDs in military working service were found to have pathologic lesions of spinal cord and cauda equina disease. The higher incidence in the present study may be explained by the fact that slightly to moderately affected dogs were included.

Six of the dogs in the present study with clinical signs of DLSS were able to perform unrestricted duties. Five of these dogs only had signs of pain during the neurologic examination, but 1 had more severe neurologic deficits. As suggested previously,²³ milder forms of DLSS do not necessarily interfere with working capabilities in working GSDs because of their temperament and, possibly, higher pain tolerance.

Our data must be interpreted with some caution as imaging modalities such as epidurography, computed tomography, or magnetic resonance imaging were not used to demonstrate compression of the cauda equina nerves in all cases. The radiographic diagnosis was made solely on the basis of survey radiographs, and the role of survey radiography was to rule out other diseases that could cause clinical signs similar to those seen with DLSS. Although presumptive, the diagnosis was strengthened by several facts. Non-neurologic causes of similar clinical signs, such as degenerative joint disease of the hip or stifle joint, could be excluded on the basis of results of repeated physical and radiographic examinations. Neurologic diseases that mimic DLSS, such as the lower motor neuron form of degenerative myelopathy and vertebral or spinal neoplasia, would have become overt during the period of surveillance and therefore could be excluded with near certainty. An important consideration that could not be ruled out was disk protrusion at a site other than the lumbosacral junction. In addition, because the 2 studies spanned a maximum of 3.5 years, it is possible that some dogs with radiographic abnormalities that did not have any current clinical signs of DLSS will develop the disease later on. Therefore, our statement that plain radiography is of limited predictive value for DLSS may prove to be incorrect for surveys performed over longer periods of time.

In conclusion, it is tempting to interpret radiographic signs of lumbosacral degeneration as a diagnostic criterion for DLSS, as the contrary will be difficult to prove in the absence of lifelong radiographic studies of dogs with and without these abnormalities. However, it can be stated with great certainty that radiographic signs of lumbosacral degeneration may precede, appear simultaneously with, or even develop after the onset of clinical signs. Thus, they are not helpful as predictors for development of DLSS. Further efforts such as a combination of radiography and magnetic resonance imaging or identification of genetic markers suggestive of a hereditary predisposition for DLSS may be necessary to identify dogs that are at risk of developing the disease during their life span. Degenerative lumbosacral stenosis is an important reason for premature termination of the working career of GSDs, but this study suggests that slightly to even moderately affected dogs may be used in police service for months and even years.