Aseptic Suppurative Meningitis in Juvenile Boxer Dogs: Retrospective Study of 12 Cases

Introduction

Meningitis, meningomyelitis, and meningoencephalomyelitis may be of infectious (e.g., viral, bacterial, fungal, or parasitic) or noninfectious origin.^1,2 Noninfectious and immune-mediated causes of these disorders are numerous and include granulomatous meningoencephalomyelitis, “little white shaker” dog syndrome, necrotizing meningoencephalitis of pugs, necrotizing leukoencephalitis of Yorkshire terriers, and steroid-responsive meningitis-arteritis of the beagle, Bernese mountain dog, and boxer.^3–6

Steroid-responsive meningitis-arteritis (also described as canine meningeal polyarteritis, juvenile polyarteritis syndrome, and aseptic suppurative meningitis) is a common cause of meningitis in young dogs.^7–10 The diagnosis is based on an aseptic suppurative cerebrospinal fluid (CSF) in a young adult dog (i.e., <2 years of age).^2,5,10 Steroid-responsive meningitis-arteritis occurs in medium- and large-breed dogs. Beagles, Bernese mountain dogs, and boxers are predisposed.^11–14 The most common clinical form is an acute onset of neck pain and fever with an elevated CSF nucleated cell count.¹⁵ A chronic form with less severe clinical signs also occurs, has a mixed population of cells on CSF analysis, and may be harder to diagnose.¹⁵ Histopathology of affected dogs shows a meningeal polyarteritis, which is how the name “meningo-arteritis” originated.

The purpose of this study was to evaluate the clinical findings, prevalence, and prognosis of steroid-responsive meningitis-arteritis in the boxer breed.

Materials and Methods

Medical records of 12 boxers diagnosed with aseptic suppurative meningitis between June 1997 and March 2002 were reviewed. Steroid-responsive meningitis-arteritis was suspected when a marked neutrophilic pleocytosis and elevated protein levels were found on CSF analysis in a young adult dog with clinical signs of meningitis.^1,2,5,10,15 Bacterial cultures of the CSF were not performed, because bacterial meningitis is uncommon in dogs, a large amount of CSF is needed for culture, and the sensitivity of CSF cultures is low.^1,5,16 However, bacteria were not identified on the CSF cytology of any dog included in the study.

For each dog, the age, sex, duration and type of clinical signs, general physical and neurological findings, results of blood and CSF analyses, treatments, and outcomes were recorded. Laboratory tests included a complete blood count, total protein, serum albumin concentration, and serum high-resolution agarose gel electrophoresis. With the dog in lateral recumbency and under general anesthesia using isoflurane,^a following diazepam^b premedication and thiopental^c induction, CSF was collected from the cerebellomedullary cistern. The CSF sample was immediately divided into two specimens. The CSF nucleated cell count was performed within 10 minutes after CSF collection. One drop of homologous serum was added to the first specimen to facilitate differential CSF cell counts after cytocentrifugation^d at 600 rpm for 10 minutes.¹⁷ A CSF nucleated cell count of >5 cells/μL was considered abnormal.¹⁸ The second CSF specimen was centrifuged for protein evaluation. The cells remaining after centrifugation and a serum sample for each dog were tested for canine distemper virus and Neospora caninum by indirect immunofluorescence or by polymerase chain reaction (PCR).

The CSF supernatant was assayed for total protein with a previously described dye-binding method using a pyrogallol red technique and electrophoresis.¹⁹ In order to assess the permeability of the blood-brain barrier, a paired electrophoresis of CSF and serum was performed in nine dogs. Agarose gel electrophoresis of serum and CSF was performed with a semiautomatic machine^e and high-resolution gels^f (without prior concentration). To evaluate blood-brain barrier permeability, albumin quotient was also calculated. The same albumin quantification was used for CSF and serum.²⁰ The albumin quotient was calculated as follows: CSF albumin concentration (mg/dL) divided by serum albumin concentration (g/L). The albumin quotient was considered abnormal when the value was >0.3.^20–22 On the centrifuged sample of CSF and on serum, immunoglobulin A quantification was performed using an enzyme-linked immunosorbent assay test. The normal values ranged from 0 to 0.2 μg/mL in CSF and from 10 to 100 μg/mL in serum.²³

Follow-ups were performed in all cases by examinations and phone calls to the owner or to the referring veterinarian. Examinations were performed weekly during the first month and then monthly until the treatment was discontinued. Adverse side effects (e.g., gastrointestinal signs, polyuria, polydipsia, skin lesions) were recorded. Reevaluations at 1 month were performed by the authors in 11 dogs, and one dog was rechecked by the referring veterinarian. Follow-up times for 11 dogs were >2 years (median 36 months) after treatment discontinuation. One dog was lost to follow-up at 14 months. Outcome was considered excellent if the dog returned to normal, good if the dog had residual neurological signs, and poor if residual neurological deficits compromised the animal’s quality of life.

Simple, descriptive statistical calculations were performed with a Microsoft Excel spreadsheet. Mean and median values were calculated for age at presentation, body temperature, white blood cell count, CSF nucleated cell count, as well as CSF and serum total protein concentration and albumin quotient.

Results

During the study period, boxers comprised 2% of all the canine cases seen at the hospital. During the same period, 30 cases of steroid-responsive meningitis-arteritis were diagnosed, and boxers represented 12 of those 30 cases. Other breeds diagnosed with the disorder included Labrador retriever (n=6), mixed-breed dog (n=4), and golden retriever (n=2). Boxers comprised 40% of all cases of steroid-responsive meningitis-arteritis. The boxers ranged in age from 5 to 29 months (median 9.5 months; mean 11.6 months). Five were females (two spayed), and seven were males (two castrated).

The durations of signs prior to presentation ranged from 5 days to 4 months (median 1 month; mean 1.1 months). Dogs were referred for fever (n=6), neck pain (n=4), lethargy (n=1), and gait abnormality (n=1). Upon physical examination, fever (>39°C) was found in 10 dogs, and four of those were lethargic. Four dogs had only neck pain, four dogs had delayed proprioception, and two dogs had both cervical hyperesthesia and delayed proprioception [Table 1]. Gait abnormalities were actually observed in four dogs during the neurological examination and ranged from mild ataxia to paraparesis and hypermetria. Reflexes were abnormal (i.e., decreased patellar reflexes) in two dogs. Cranial nerves were normal in all dogs except one. In this last dog, deficits were compatible with central vestibular syndrome (i.e., head tilt to the left, vertical nystagmus, left-sided proprioception deficits).

Complete blood counts revealed leucocytosis (white blood cell count >13 × 10⁹/L; reference range 6 to 13 × 10⁹/L) in eight dogs, and three of the differential cell counts were compatible with a stress leukogram. The CSF cell counts ranged from 14 to 10,000 cells/μL (median 780; mean 874) [Table 2]. The CSF pleocytosis was considered mild (5 to 100 cells/μL) in two dogs, moderate (100 to 1000 cells/μL) in five dogs, marked (1000 to 5000 cells/μL) in three dogs, and very marked (>5000 cells/μL) in two dogs. In 10 dogs, the major cell type (i.e., >70% of all cells counted) was the neutrophil. In two cases, the cell type was mixed. Both of these dogs had been on anti-inflammatory doses of methylprednisolone for >1 month when the CSF was collected.

An elevated CSF protein concentration (>0.3 g/L) was noted in 11 dogs (median 0.95 g/L; mean 0.99 g/L). Paired, high-resolution agarose gel electrophoresis of CSF and serum was performed in nine cases. The albumin quotient was elevated (median 4.82; mean 9.26) in all nine cases. Another electrophoretic abnormality was a spike in the alpha-2 globulin region, both in serum and CSF (n=6). In three dogs, elevations in immunoglobulin A were detected in both CSF (0.9, 3.4, and 9.2 μg/mL, respectively) and serum (142, 174, and 382 μg/mL, respectively). All serum (n=12) and CSF (n=12) samples were negative for canine distemper virus and Neospora caninum. Four serum and CSF samples were also tested for Ehrlichia canis and Borrelia burgdorferi and were negative.

At the time of presentation, nine dogs were already receiving either nonsteroidal or steroidal anti-inflammatory drugs. After presentation and diagnostic testings, treatment was standardized for all dogs. The treatment consisted of one injection of dexamethasone^g (0.3 mg/kg intravenously [IV]) followed by prednisolone^h at 1 mg/kg orally (PO) q 12 hours for 5 days; then 1 mg/kg PO q 24 hours for 10 days; then 0.5 mg/kg PO q 24 hours for 2 weeks; and finally, 0.5 mg/kg PO q 48 hours for 1 month. If relapse occurred as the treatment was tapered, prednisolone was increased to 1 mg/kg PO q 12 hours. If the clinical response was insufficient, azathioprineⁱ was added at 2.2 mg/kg PO q 24 hours for 1 month, then 2.2 mg/kg PO q 48 hours for 1 more month. Clindamycin^j (8 mg/kg PO q 12 hours) was also given, pending the CSF and laboratory test results.

The durations of prednisolone treatment after diagnosis ranged from 8 to 40 weeks, with a median time of 10 weeks [Table 1]. Clinical signs resolved in seven dogs within 12 weeks. For these dogs, clinical signs had been present for <1.5 months prior to presentation. No relapses occurred during the tapering period or during the 2 years of follow-up (i.e., after the treatment was discontinued). Five dogs required a longer duration of treatment, ranging from 16 to 40 weeks. All of these dogs had clinical signs for >1 month (1 to 4 months) prior to presentation, and all four relapsed during treatment. The longest duration of treatment was in the youngest dog (5 months old at presentation) with central vestibular syndrome (case no. 10). In this dog, azathioprine (2.2 mg/kg PO q 24 hours for 1 month and q 48 hours for 1 month) was given with prednisolone. For the four other cases, an increase in the prednisolone dosage was sufficient to control the clinical signs (e.g., fever, reluctancy to walk, neck pain). These five dogs did not show any signs of relapse >2 years (median 36 months) after the treatment was discontinued.

Despite some long durations of prednisolone treatment (up to 40 weeks), the side effects observed were mild. They included polyuria, polydipsia, and some gastrointestinal signs (i.e., diarrhea and vomiting) that resolved with symptomatic treatment.⁹ Outcomes were considered excellent in 11 dogs at the end of treatment. The dog with central vestibular signs had a residual head tilt but no other general or neurological signs.

Discussion

During the study period, the prevalence of boxers presented to the hospital was two per 100 dogs. The prevalence of boxers among all cases of steroid-responsive meningitis-arteritis during the same period was 40%. Based on the number of affected boxers in this study, the breed appeared to be predisposed to steroid-responsive meningitis-arteritis, as has been reported in prior studies.^14,15 Results of the clinical laboratory and CSF analyses for the dogs of this report were similar to other studies.^5,14 In a prior study, data on three affected boxers revealed a possible familial relationship, but no familial relationship was found in the current 12 cases.¹⁴ Most of the dogs in the study reported here were <1 year of age (n=10), and all were <2.5 years old. These ages approximated those of affected beagles (3 to 18 months of age) and Bernese mountain dogs (3 to 28 months).^11,13 No sex predisposition was noted in these two breeds or in the boxers.^11–13

The clinical presentations of the dogs in this report were similar to those seen with beagles and Bernese mountain dogs and other cases of previously described steroid-responsive meningitis-arteritis.^8–10,14,24 In several of the cases reported here (n=6), a delayed proprioceptive response was noted, suggesting the presence of myelitis in association with the meningitis. Gait abnormalities like ataxia, paraparesis, and hypermetria have also been noted in other studies.^8,10,15 The central vestibular signs of one dog in this study were attributed to an aseptic suppurative meningoencephalitis of the brain stem. Although less common than in the spinal cord, lesions of the leptomeninges of the brain can occur with this disease.^7,25 Ten of the dogs exhibited the acute form of the disease and had >100 nucleated cells in their CSF. The two other dogs had no neck pain or fever, and their CSF results showed a mixed pleocytosis compatible with the chronic form of the disease.

In all cases, the albumin quotient was elevated, which was consistent with a blood-brain barrier disturbance and leakage of serum albumin into the CSF.^20,26 Even though not specific for the disease, a blood-brain barrier disturbance is likely with aseptic suppurative meningitis, because the leptomeninges are infiltrated with polymorphonuclear cells. The presence of an electrophoretic alpha-2 spike has been previously documented in the blood of dogs with aseptic suppurative meningitis.¹⁵ Alpha-2 spikes in the CSF and in the serum were found in six dogs (among the nine tested) and were compatible with leakage of inflammatory proteins (e.g., haptoglobulin) into the CSF.²⁷

Immunofluorescent and PCR assays were negative for canine distemper virus and Neospora caninum in all samples tested. In France, these diseases are the most common infectious agents known to induce central nervous system inflammation in the dog.^1,2,28 Although CSF bacterial cultures were not performed, complete and definitive recovery with immunosuppressive treatment helped to rule out bacterial meningomyelitis. The cause of aseptic suppurative meningomyelitis is still unknown. No infectious agents have been isolated in suspected or confirmed cases. Immunoglobulin A seems to play a major role in the pathogenesis of this disease.^23,29 Elevated Immunoglobulin A both in serum and CSF is consistent with the diagnosis of steroid-responsive meningitis-arteritis.^15,23,29 Immunoglobulin A quantification was elevated in both serum and CSF in three dogs of this report.

In the current study, complete resolutions were obtained in all cases. Based on these results, the prognosis for boxers with steroid-responsive meningitis-arteritis appears to be better than the prognosis for the two other commonly affected breeds.^5,12,13 The five dogs that needed the longest duration of treatment were the ones that had signs for >1 month before diagnosis. Most of the dogs that received shorter treatments (<12 weeks) had signs for <2 weeks. These results confirmed, as in previous studies, that better responses occur if the immunosuppressive treatment is initiated early in the disease.^9,15 In one case of this report, initiating a more aggressive immunosuppressive treatment by combining prednisolone and azathioprine helped to resolve signs of a relapse. Further investigation of this combination of drugs to prevent relapses is warranted.

A limitation of the current study was that no histopathological results were available. Without histopathological data, and because numerous immune-mediated inflammatory diseases of the central nervous system may be called “steroid responsive,” it was considered more appropriate to define this disease as aseptic suppurative meningitis based on the CSF findings.

Conclusion

Twelve boxers <2 years of age were diagnosed with aseptic suppurative meningitis after being presented with clinical signs of fever and cervical pain. All dogs but one were followed for ≥2 years and completely recovered on prednisolone therapy, with or without azathioprine. Based on the results of this study, the prognosis appears to be good to excellent in the boxer, especially if immunosuppressive treatment is initiated early in the course of the disease.

Acknowledgment

The authors thank Dr. Simon Platt and Dr. Simona T. Radaelli for their help and advice.