Establishment of the nomenclature
The occurrence of a progressive pelvic limb ataxia and paresis in the older German shepherd dog (GSD) had been recognised for many years. Originally the clinical signs were attributed to the presence of ossified
plaques in the dura mater which commonly developed at the cervical and lumbar enlargements, usually on the ventral and ventrolateral aspects of the spinal cord. In extreme cases a bony tube formed around the cord
(Morgan, 1968). This ossification, which was relatively common in older dogs, was called chronic ossifying pachymeningitis or spinal dural ossification.
These plaques could not have accounted for the clinical signs seen in chronic degenerative radiculomyelopathy (CDRM). Morgan (1968) showed that dural ossification was a relatively common finding in middle to old age
dogs of large breeds but occurrence did not necessarily coincide with signs of pelvic limb ataxia or weakness. In his study, approximately two thirds of all dogs over two years of age had evidence of dural
ossification. Of the 52 cases which showed dural ossification only 12 had clinical signs of abnormal gait or hindquarter weakness; a further three dogs had clinical signs but no evidence of dural ossification. The
position of the majority of plaques at the cervical and lumbar enlargements would have resulted in a lower motor neurone lesion (LMN) whereas CDRM involved an upper motor neurone lesion (UMN) (Griffiths and Duncan,
1975a). In addition, the occurrence of plaques at the cervical enlargement should have caused concurrent thoracic limb and pelvic limb dysfunction, which did not occur.
The condition was first described as degenerative myelopathy by Averill (1973). When they found degenerating fibres in the lumbar dorsal nerve roots of a number of affected dogs, Griffiths and Duncan (1975a) proposed
the name chronic degenerative radiculomyelopathy (CDRM). German shepherd dog myelopathy was another term used to describe the condition. In this account the disease will be referred to as CDRM.
This disease had been described in a number of large breeds and large breed crosses such as German shepherd dog (GSD), Irish setter, Collie cross (Averill, 1973), Rough collie, Rhodesian ridgeback and Labrador cross
(Griffiths and Duncan, 1975a). Most cases reported had been in the GSD. Of the studies completed to date, the percentage of affected dogs which were GSD ranged from 56% (9/16) (Griffiths and Duncan, 1975a) to 82%
(18/22) (Averill, 1973). These relatively small studies did not give an accurate picture of the actual occurrence within breeds and precise information regarding incidence was difficult to obtain. Experts in the
field (Griffiths, personal communication, 1995) suggested that about 90% of affected dogs are GSDs. In addition, anecdotal evidence (GSD breeders, personal communication, 1996) suggested that up to 20% of the total
GSD population may develop clinical signs suggestive of CDRM at some stage in their lives.
Prior to 1973 the clinical signs were ascribed to the presence of dural bone plaques or multiple small intervertebral disc protrusions (Morgan, 1968). Averill (1973) mentioned the similarity with vitamin B12
deficiency in man, noting that in man the disease initially involved only the thoracic cord, affected all funiculi and was slowly progressive unless appropriately treated. There was however a fundamental difference
between humans and dogs in that the human disease initially exhibited a perivenular pattern and was usually multifocal, whereas CDRM in the dog was topographically continuous. Levels of methylmalonic acid in
affected dogs showed no abnormalities, which suggested that normal vitamin B12 metabolism was occurring. Williams et al. (, 1984) also investigated the possible involvement of vitamin B12. They found that all six
affected dogs tested had bacterial overgrowth in their duodenal juice. This abnormality was not associated with any histological changes in jejunal biopsies. These dogs, however, had marked biochemical abnormalities
which included the following changes to brush border enzymes: leucyl-2-naphthylamidase was decreased, g -glutamyl transferase was increased. Marked increases in activities of an endoplasmic reticular enzyme,
tris-resistant a -glucosidase and the lysosomal enzymes N-acetyl-b -glucosaminidase and acid phosphatase were also seen. The significance of these changes must be questioned as this study did not include a control
population of unaffected dogs. Vitamin B12 levels were subnormal in only three of the six dogs.
The above findings suggested that CDRM may occur due to the abnormal absorption of some other nutrient(s) or that the biochemical abnormalities occurred secondary to the neurologic dysfunction. A consideration of
nutritional factors which may be involved in CDRM must include vitamin E. Williams et al. (, 1985) investigated the serum tocopherol concentrations in a group of seven GSDs with CDRM and concluded that serum
tocopherol concentrations were slightly lower in affected dogs (0.5-8.6mg/l) compared to healthy controls (4.4-29.4mg/l). These authors suggested that the enteropathy they saw, and the associated reduction in serum
tocopherol levels, may have arisen secondary to CDRM. Alternatively, the enteropathy may have been the primary lesion responsible for malabsorption of tocopherol and subsequent development of CDRM. It had been shown
that deficiency of vitamin E caused neuropathology in other species; for example, degenerative myelopathy/myeloencephalopathy in the horse and ataxia with vitamin E deficiency (AVED) in man. AVED patients were known
to respond to the administration of high doses of vitamin E (DiDonato, 1995). In addition, loss of axons and myelin sheaths, especially in the dorsal columns (Pentschew and Schwarz, 1962; Nelson et al., 1981;
Southam et al., 1991), had been associated with chronic vitamin E deficiency in rats and rhesus monkeys. As vitamin E was known to work as a free radical scavenger, therefore protecting neuronal cell membranes from
peroxidation, its absence could lead to increased membrane fragility and ultimately neuronal cell death.
Several familial neuronal degenerations such as Freidriech`s ataxia, hereditary spastic ataxia and hereditary spastic paraplegia existed in man. These were the result of neuronal atrophy or abiotrophy in which the
cell body and proximal nerve fibre remained intact whilst the distal axon degenerated. Although the actual mechanisms were poorly understood, Averill mentioned a possible similarity to CDRM in 1973, suggesting that
this process should be considered in CDRM. If correct, he suggested, neuronal loss or atrophy in several important locations such as the spinal ganglia (SpG), red nucleus, rostral colliculus and vestibular nucleus
should be present (Averill, 1973).
Griffiths and Duncan (1975a) considered the neuropathology indicative of a "dying back", or distal axonopathy, phenomenon such as occurred in various human neurodegenerative conditions and experimental
neurological diseases. Central "dying back" processes had been implicated in conditions such as amyotrophic lateral sclerosis and some of the spinocerebellar disorders (Griffiths and Duncan, 1975a).
Griffiths and Duncan considered that no proof of a hereditary factor had been put forward but explained the greater occurrence in large breed dogs to be the result of differential susceptibility due to the longer
nerve fibres present compared with small dogs. This did not, however, explain the particular breed predilection seen in CDRM. In studies on equine laryngeal hemiplegia Duncan et al. (1974) showed a positive
correlation between nerve fibre length and susceptibility to degeneration. Braund (1978) argued that the distribution of lesions in CDRM did not fit the expected pattern of a "dying back" disease. The
severity of lesions in the thoracic segments which suggested selective vulnerability of thoracic cord (Averill, 1973; Braund and Vandevelde, 1978) in CDRM remained unexplained despite having been found in viral,
nutritional and idiopathic disorders in other species.
Waxman et al. (1980b) considered the pathology of CDRM to be indicative of a chronic demyelinating disease rather than an acute inflammatory response. As the immune system was known to play an important role in
chronic demyelinating diseases in other species such as multiple sclerosis (MS) in man and experimental allergic encephalomyelitis (EAE) in rats (Bernard and Kerlero de Rosbo, 1992; Hartung and Rieckmann, 1997) the
authors investigated the response to thymus-dependent mitogens in seven normal and six CDRM affected dogs. An abnormal T cell function was found in the peripheral blood cells of affected dogs, leucocytes displayed
an impaired response to the T cell mitogens concanavalin A (Con A) and phytohaemagglutinin P (PHA). Serum or plasma obtained from dogs with CDRM was not suppressive for normal cells and neither was normal
proliferative function restored when peripheral blood cells from affected dogs were cultured in normal serum or plasma, which suggested that any abnormal function of T cells was not due to the presence of serum
inhibitory factors. Whether or not this was related to the disease process of CDRM was not known. The results of this study did not explain why depressed lymphocyte responsiveness occurred in dogs with CDRM, nor did
it explain why the immunological deficit was seen only in the peripheral blood leucocyte compartment. It was suggested that an autoimmune lymphocyte population may have been activated during the course of CDRM.
In a follow-up study Waxman et al. (1980a) showed that CDRM dogs had circulating suppressor cells associated with the impaired peripheral blood response. These findings had not been linked directly to the disease
process but they could have indicated either an immunodeficiency problem where depressed T cell function allowed an infectious agent to enter the central nervous system (CNS) or where the suppressor cell population
was activated secondarily by the host to regulate an autoimmune event. Barclay and Haines (1994a) studied a small group of dogs (five affected and one unaffected) and suggested that an immune-mediated spinal cord
destruction may be occurring, because they found immunohistochemical evidence of immunoglobulin and a component of complement in the spinal cords of affected dogs which was not present in the normal dog. Clemmons
(1992) suggested that, despite the lack of direct evidence for CDRM being caused by an infectious agent, it was possible that the disease was the result of a combination of genetic predisposition and exposure to an
infectious agent. He reported in 1992 that workers were investigating whether a canine retrovirus infection or "scrapie-like" prion organism was involved, however, neither of these possibilities had been
proven (no references were cited on the Clemmons paper).
Undoubtedly, CDRM has a complex aetiology which probably involves several different factors but most authors agreed there was almost certainly a genetic factor due to the very high incidence of the disease in one
Age of Onset
All accounts to date conceded that there was great variation in age of onset, the youngest reported case was five years old, the eldest 14 years while the majority were about nine years old at first presentation.
Clinical signs of CDRM had been well documented (Averill, 1973; Griffiths and Duncan, 1975a; Braund and Vandevelde, 1978; Clemmons, 1989; Clemmons, 1992). The dogs showed a slowly progressive pelvic limb ataxia and
paresis with loss of proprioception. Initially they scuffed the middle two toes of one or both hind feet. Subsequently they developed problems with circling and with stairs, especially going down, they often
scuffed, misjudged distances and showed hypermetria (ataxia in which intended movements overreach the intended goal). The dogs were often affected asymmetrically although both pelvic limbs were usually involved.
Disuse muscle atrophy developed over the trunk and hindquarters several months after disease onset. With time, prolonged scuffing resulted in excoriation and ulceration of the feet. Eventually the disease resulted
in marked paraparesis, the dogs could not rise without assistance and would pull themselves along with their thoracic limbs. A degree of faecal and urinary incontinence would often develop late in the disease. Dogs
maintained beyond this stage could show thoracic limb signs. It had been reported that brain stem involvement eventually occurred (Clemmons, 1992) which could result in a number of signs including asymmetrical
tetraparesis, cranial nerve abnormalities and altered mental status. The clinical course was very variable ranging over six to 12 months (Clemmons, 1992) from onset of the disease to complete pelvic limb
paraparesis. The clinical signs were inexorably progressive and whilst they may have stabilised for short periods of time improvement did not occur.
On neurological examination conscious proprioception as evidenced by paw position sense, reflex stepping and sway test was severely affected either uni- or bilaterally which depended, to some extent, on the duration
of disease. Pedal reflex was usually unaffected while the patellar reflex varied in response to the disease. Hyperreflexia of the patellar reflex, with or without clonus, had been reported (Averill, 1973; Waxman et
al., 1980b; Clemmons, 1992). Hyporeflexia and loss of the patellar reflex had also been reported (Griffiths and Duncan, 1975a). The panniculus reflex remained intact, as did the perineal reflex, until late in the
It was not possible to diagnose CDRM definitively in life. A presumptive diagnosis could be made based on typical history and clinical signs. Other differential diagnoses such as disc disease, lumbosacral disease or
neoplasia could be ruled out by further investigations which included haematology, serum chemistry profiles, urinalysis, CSF analysis, spinal radiographs and myelography. Some authors had reported an unspecified
increase in CSF protein and abnormality in electrophoretic pattern (Clemmons, 1992; Clemmons et al., 1995) in dogs with CDRM. Further diagnostic procedures such as electrophysiology, epidurography, discography,
computed tomography (CT) and magnetic resonance imaging (MRI) were indicated in specific cases. The results from all these tests were negative in CDRM as definitive diagnosis could be achieved only by histological
examination of the spinal cord.
Clemmons (1992) suggested, on the basis that he considered dogs with CDRM to have three to 10 times more circulating immune complexes than normal dogs, that it may be possible to develop serum markers for the
specific diagnosis of CDRM. From preliminary work, he suggested the presence of an 85kDa antigen in dogs with CDRM but not in those with other neurological disorders. However, to date, no other authors have
mentioned such a possibility.
The emphasis in CDRM research over the last 25 years had been to map accurately the neuropathology and investigate the possible aetiologies involved. It had been considered an untreatable disease (Braund, 1987).
Clemmons (1989) was the first author to suggest a therapy which may slow down the rate of deterioration in affected dogs. Glucocorticoids, used for the treatment of inflammatory processes or to suppress supposed
immune responses, had been used by veterinary surgeons but without any success. Clemmons suggested the use of other immunosuppressive drugs, initially he tried cyclophosphamide and azothiaprine without success.
Immunostimulants such as levamisole were also tried but appeared to speed up the rate of progression of the disease. The use of intravenous dimethylsulphoxide (DMSO), which has a number of unproven possible
activities including the scavenging of free radicals, and intramuscular cobra venom, had also been tried unsuccessfully.
High doses of vitamin E (2000IU/day), high potency B vitamin complex and epsilon aminocaproic acid (EACA) had all been used as treatments (Clemmons, 1989; Clemmons, 1992) although their efficacy appeared
questionable. EACA has anti-protease activity, Clemmons considered that it would therefore be helpful in CDRM as it would block the final step in the inflammatory pathway thus helping to prevent tissue destruction.
All authors agreed that maintenance of regular exercise and optimal body weight seemed beneficial to affected dogs.
There was no further evidence which suggested that any of the therapies suggested by Clemmons were beneficial in the treatment of CDRM which was still considered untreatable.
The pathology was initially described by Averill (1973) and subsequently by others, all of whom agreed that the lesion involved mainly the spinal cord white matter and affected both the axon and myelin sheath. As
previously discussed, before 1973 it was thought that the clinical signs seen in CDRM were attributable to the occurrence of spinal dural ossification. Averill reported irregular oval plaques formed from thin
cancellous bone which contained normal bone marrow on the ventral aspect of the dura mater. These did not impinge on the intervertebral space and from their position could not have resulted in the upper motor
neurone (UMN) lesions seen in CDRM. Averill also found varying amounts of spondylosis, these varied from small bony spurs on the vertebral bodies to large, hard, wedge-shaped structures which extended underneath the
intervertebral space. The incidence of spondylosis was not significantly different between the affected and control groups of dogs and in addition the sites did not correspond with the spinal cord pathology. The
spinal cord changes were consistent for all affected dogs. On gross examination of the cord the leptomeninges were slightly thickened, histologically they exhibited slight fibrosis. The white matter ascending and
descending tracts were affected to varying degrees throughout the cervical, thoracic and lumbar segments (mid to caudal thoracic segments were the most severely affected). Only the fasciculus proprius appeared to be
unaffected. Averill considered these areas to consist of loss of axons and myelin sheaths with an associated astrocytic hypertrophy and hyperplasia. The only grey matter changes noted were a possible loss of ventral
horn neurones and an occasional mild perivascular fibrosis of the arterioles in this region. He noted that none of the degenerative changes appeared to extend into either the brain or the peripheral nerves. There
were no consistent visceral lesions present in affected dogs.
The pathology was described in more detail by Griffiths and Duncan (1975a). Again they described the presence of degenerative lesions in the spinal cords of all dogs examined but with variations in distribution and
severity, these lesions were considered to be typical of Wallerian-type degeneration and affected both the axon and myelin sheath. They noted that any observed clinical asymmetry correlated with the pathological
asymmetry. Loss of axons and myelin sheaths was seen in the dorsal funiculi, specifically in the fasciculus gracilis, of cervical and cranial thoracic segments. Occasional axon and myelin sheath loss was also noted
in the dorsal funiculi of lumbar and sacral cord segments. In the lateral funiculus the most consistently affected area was the corticospinal tract which was most severely affected in the caudal thoracic segments.
Any other fibre loss was scattered and not severe enough to constitute an actual tract involvement. The only grey matter change reported was the presence of increased numbers of astrocytes in the lateral and dorsal
horns of grey matter in the lumbar and caudal thoracic segments. Some affected dogs had an occasional SpG neurone which exhibited central chromatolysis. No changes were seen in peripheral nerves or muscles of
affected dogs. There was no indication in this paper that the brains had been examined. As the present study identified significant brain pathology this illustrated the necessity for carrying out a complete
The report by Braund and Vandevelde (1978) confirmed most of the previous findings, however these authors did not observe any pathological changes in the SpG. They concluded that this may have been due to the
clinical and pathological variations in the dogs studied. In addition to previous findings they reported marked axon and myelin loss in the white matter which surrounded the ventromedian fissure. Again they noted
that the thoracic cord was the most severely affected region, although both cervical and lumbar segments were also involved.
Clemmons (1989) reported widespread demyelination and axon loss in affected areas of the spinal cord white matter with an associated increase in astrocyte numbers and increase in density of small vascular elements.
In the thoracic cord he saw vacuolation in all funiculi, with swollen axons and eosinophilic globules, which represented degenerate and regenerate axons. A number of the vacuoles seen contained macrophages and
astrocytes. Clemmons reported finding similar lesions, very occasionally, throughout the white matter areas of the brain (1989).
Other degenerative myelopathies of dogs
A number of diseases which had clinical and pathological similarities with CDRM had been recognised in other dog breeds. They were all characterised clinically by a progressive pelvic limb ataxia and paresis. On post
mortem examination the spinal cord white matter exhibited degeneration of both axons and myelin sheaths, often with an associated gliosis. The age of onset amongst this group of diseases was, however, very variable.
In 1973 an ataxia with an autosomal recessive mode of inheritance was reported in Jack Russell terriers (Hartley and Palmer, 1973) in Britain and in Sweden. A progressive pelvic limb incoordination and dysmetria with
over-protraction of the thoracic limbs commenced at three to six months of age. Affected dogs showed a characteristic "dancing" gait. On pathological examination the two populations varied, although both
had focal symmetrical loss of myelinated fibres in the peripheral dorsolateral white matter. In the more severely affected English cases a widespread Wallerian-type degeneration was noted throughout the CNS with
degenerative changes in the central auditory pathway and peripheral nerves. It had been suggested that the differences may be due to the different time spans of disease.
A number of authors have reported a condition in a variety of hounds which had been variously named hound ataxia (Palmer and Medd, 1981; Palmer et al., 1984; Palmer and Medd, 1988) or spinal myelinopathy (Sheahan et
al., 1991). This condition occurred in the adult animal, two to seven years old, with affected animals exhibiting pelvic limb weakness and incoordination. Most cases had a restricted panniculus localising to
mid-thoracic to cranial lumbar segments. Conscious pain sensation remained intact. On pathological examination of the CNS vacuolar degeneration was present in lateral and ventral funiculi throughout the length of
the spinal cord. Astrocytic proliferation and myelin debris were common. Changes in the spinal cord grey matter were uncommon, but chromatolysis was observed in the lateral vestibular nucleus and in thoracic
neurones in segments T7 and T8 in two cases out of sixteen (Palmer et al., 1984). In all other cases the pathology observed in the brain stem was indicative of tract degeneration. The changes suggested a primary
myelinopathy, unlike CDRM which appeared to have simultaneous degeneration of both the axon and myelin sheath. The occurrence of this disorder amongst foxhounds coincided with a change in diet to include a high
proportion of ruminant stomachs (tripe). Subsequent alteration to include a much higher proportion of meat had resulted in decreased incidence of the disease. Thus, a clear association had been established between
the occurrence of hound ataxia and the feeding of ruminant stomachs which possibly resulted in decreased methionine levels and increased methionine synthetase activity in serum. It had been previously suggested that
neurological deficits in subacute combined degeneration in man could have resulted from a methyl group deficiency associated with decreased levels of methionine and a concomitant decrease in levels of
S-adenosylmethionine. The suggestions for the aetiology of hound ataxia have never been confirmed, although the incidence had reduced dramatically since the diet was altered to include more meat.
Bichsel and Vandevelde (1983) described three related Siberian huskies with a late-onset progressive pelvic limb ataxia. In the two cases which could be examined muscle tone and flexor reflexes in the pelvic limbs
were weak, whereas patellar reflexes were exaggerated. Pain sensation remained intact but proprioceptive positioning of the pelvic limbs was slow. These dogs all developed urinary incontinence as the disease
progressed. At post mortem all three dogs had similar changes which comprised of disseminated vacuolation of the white matter, sometimes associated with axonal swelling and necrosis. These changes were found at all
levels of the cord, especially in the ventral and lateral columns. Minimal gliosis was observed and there was no evidence of inflammation. The thoracic segments were the most severely affected, particularly in the
peripheral white matter. The changes were asymmetrical and were not attributable to specific tracts. The affected dogs were closely related, two littermates and their mother, which suggested a possible hereditary
basis to the disease.
A progressive pelvic limb ataxia had been reported in a five month old Pyrenean mountain dog (Wright and Brownlie, 1985). Diffuse axonal degeneration was noted in the dorsolateral, lateral and ventral white matter
extending from the medulla to the fifth lumbar segment. The most severely affected segments were C8 to The degenerative changes involved both the axon and myelin sheath. The SpG and the spinal cord grey matter were
unaffected. Changes found in the brain included axonal swelling and degeneration in the spinocerebellar tracts, caudal and middle cerebellar peduncles and the cerebellar folia. No attempt had been made to determine
the underlying cause of this disease.
In 1995 a degenerative myelopathy was reported in a ten year old Miniature poodle (Matthews and de Lahunta, 1985). A progressive pelvic limb ataxia and paresis developed with loss of proprioception, which initially
affected only one pelvic limb. As the disease progressed hypertonia and hyperreflexia became evident in both pelvic limbs. Post mortem examination of the spinal cord revealed extensive diffuse degeneration of the
white matter in all funiculi at all levels, which involved both axons and myelin. Again, the thoracic cord was the most severely affected. The spinal nerve roots were unaffected in this case. No cause was
A progressive neurological disorder had been reported in a young adult Ibizan hound (Summers et al., 1995). The ataxia was obvious in the pelvic limbs from the time of ambulation and progressed to involve the
thoracic limbs. The gait was spastic and dysmetric with overflexion and abduction which led to a distinctive bobbing, bouncing type of gait. Patellar reflexes were absent. On microscopic examination of the spinal
cord there was symmetrical degeneration of ascending and descending tracts in all funiculi and at all levels of the spinal cord although the thoracic cord was most severely affected. Unlike CDRM, the pathology
extended into specific peripheral nerves where there were dilated myelin sheaths which contained intact axons. This condition essentially represented a diffuse nervous system degeneration with myelopathy and
neuropathy which mainly affected axons. Pedigree analysis supported an autosomal recessive pattern of inheritance for this condition.
A degenerative myelopathy with some resemblance to CDRM had recently been reported in the young GSD (Longhofer et al., 1990). The two affected animals were six and seven months of age. They exhibited a progressive
pelvic limb ataxia and weakness. These dogs had very poor muscling of their hindquarters and normal conscious proprioception which distinguished them from dogs with CDRM. Pedal and patellar reflexes were intact but
one dog had clonus of the patellar reflex. The pathological findings were similar in both dogs and involved symmetrical diffuse degeneration of axons and myelin with mild gliosis. These changes occurred
predominantly in the lateral and ventral funiculi. The thoracic cord was most severely affected. Axonal degeneration was a prominent feature in this disease with many axonal spheroids and macrophages in affected
areas of white matter. The only changes seen in the brain involved occasional degenerating axons in cerebellar white matter. The SpG and spinal nerve roots were all normal. Due to the significant differences between
this condition and CDRM it was suggested that the pathogenesis may be unrelated. In addition, since the dogs were both thin and poorly muscled, a possible hereditary metabolic disorder was postulated.
There were other progressive neurodegenerative conditions which had been reported in a number of different dog breeds. These conditions were too dissimilar to CDRM to warrant discussion in this thesis. Examples of
these were hereditary myelopathy in the Afghan hound (Cockrell et al., 1973; Cummings and de Lahunta, 1978), leukoencephalomyelopathy in the Rottweiler (Gamble and Chrisman, 1984) and progressive axonopathy in the
Boxer (Griffiths, 1989).
Notes taken from extracts written by Pamela Elizabeth Jean Johnstone.