Granulomatous meningoencephalomyelitis is an inflammatory disease of the CNS that has been identified in dogs of various ages, breeds, and sexes.1 Associated nervous system lesions may be focal or multifocal. The cause of GME remains unknown; however, infectious agents, toxicants, and environmental factors have been suggested. Because no causative infectious agent has been identified, it is considered likely that GME is an immune-mediated disease triggered by genetic or environmental factors or both.2,3
Environmental triggers have been explored for other immune-mediated diseases such as IMHA and SRMA in dogs. Some reported findings suggest a link between IMHA and infectious, neoplastic, or inflammatory processes or vaccine exposure.4 On the other hand, no association was found between SRMA and vaccination, season, geographic location, or dog sex in 1 study,5 but a significant association was identified between SRMA and specific breeds. It is therefore suspected that genetic factors may predispose an individual dog to the development of GME; however, specific factors may then be needed to trigger the onset of disease.6–10 The purpose of the study reported here was to determine whether sex, body weight, BCS, vaccination history, season of onset of clinical signs, and human population density in the dog's geographic area were significantly associated with a diagnosis of GME in dogs.
Materials and Methods
Animals
A case-control study design was used. The medical database at UW Veterinary Care, the veterinary teaching hospital at the University of Wisconsin-Madison, was searched to find all dogs that received a histologic diagnosis of GME from January 1, 2003, through December 31, 2014. Results of histologic examination were reviewed, and the diagnosis was confirmed by a board-certified veterinary pathologist (MEP) prior to inclusion in the study. The diagnosis of GME was made on the basis of standard diagnostic criteria, as described elsewhere.11 Dogs were included in the case group if histologic findings were consistent with GME and complete medical records were available for review.
The medical records for the same period were also searched to identify 3 control dogs matched to each case dog by age (within 12 months), breed, and year of the visit to the teaching hospital in which they qualified for inclusion. For mixed-breed dogs, control dogs were matched to case dogs on the basis of apparent predominant breed, as recorded. Control dogs were excluded if they had been evaluated for neurologic disease, had a history of neurologic disease for which a neurologist had been consulted (other than if they had a surgical or histologic diagnosis of intervertebral disk herniation as their only cause of neurologic disease), were receiving antiseizure or immunosuppressive medications, or died on arrival at the hospital and no necropsy was performed. If insufficient control dogs meeting these criteria were identified to match to a given case dog, more control dogs were identified by expanding the allowable age difference (to ≥ 12 months) or evaluation year.
Data collection
Data were obtained from the medical records for dogs in both groups regarding age, sex, body weight, BCS, and home address at initial diagnosis (for dogs with GME) or the visit of study inclusion (control dogs) as well as the date of the most recent vaccination prior to or including the visit of study inclusion. Body condition was recorded on a 9-point scale, with scores ≥ 6 considered overweight. If no BCS was recorded but subjective assessment data were available, dogs recorded as in good, fair, or excellent body condition were classified for the study as having a BCS of 5, dogs recorded as in thin body condition were classified as having a BCS of 3, and those considered overweight or having excess body habitus were classified as having a BCS of 7. For control dogs specifically, if no BCS was recorded at the visit of study inclusion, the BCS from previous visits was used provided no change in body weight had occurred since that time. The referring veterinary hospital was contacted to obtain any missing patient information, and unavailable data were omitted from statistical analysis.
For study purposes, the season of the visit of study inclusion (control dogs) or onset of clinical signs (case dogs) was classified as winter (December through February), spring (March through May), summer (June through August), or fall (September through November). If the date of onset of clinical signs was not recorded for case dogs, the date of necropsy diagnosis was used instead.
Human population density within the dog's geographic area was determined by conversion of the home address to longitude and latitude coordinates by use of an online resource.a The human population density for the 5 square miles surrounding these coordinates was then determined by use of an online population calculator.b The time between onset of clinical signs (case dogs) or date of visit (control dogs) and most recent vaccination was calculated and categorized as ≤ 1 month, ≤ 6 months, and ≤ 12 months from the onset of clinical signs.
Statistical analysis
Associations between case-control status and the various potential predictor variables (age, sex, body weight, BCS, time since last vaccination, human population density, sex, and season of visit or onset of clinical signs) were explored by means of conditional logistic regression, with matched-pair status as a stratum variable. The P values pertaining to these models were computed by use of the likelihood ratio test. Odds ratios were computed as estimates of effect size; for categorical predictor variables, the OR was calculated relative to an arbitrarily chosen referent category. Distributions of dogs in the case group were compared among seasons of onset of clinical signs by use of the Fisher exact test. Several conditional logistic regression models were fit that contained multiple predictor variables. However, because the results were similar to the results of univariate analyses, no results of multivariate modeling are shown. All statistical analyses were performed with the aid of statistical software.c Values of P < 0.05 were considered significant.
Results
Thirty-eight dogs with a histologic diagnosis of GME met the criteria for inclusion in the case group. Seven of these dogs were excluded because of incomplete medical records or disagreement with the initial diagnosis when reviewed for inclusion, leaving 31 dogs in the case group. Three control dogs were identified for each case dog, except for 1 case dog for which only 1 control dog could be identified, resulting in inclusion of 91 dogs in the control group.
Case dogs
Median age at onset of clinical signs of GME for the 31 dogs in the case group was 6.2 years (range, 0.8 to 13.6 years), and median body weight was 8.3 kg (18.3 lb; range, 1.3 to 43.2 kg [2.9 to 95.0 lb]). Median BCS on a 9-point scale for the 29 dogs for which it could be determined was 5 (range, 3 to 8). Twelve (39%) dogs were male (11 neutered and 1 sexually intact), and 19 (61%) dogs were female (all spayed). Dogs included 7 (23%) mixed-breed dogs, 4 (13%) Dachshunds, 2 (6%) Pugs, 2 (6%) Shih Tzus, 2 (6%) Labrador Retrievers, 2 (6%) Golden Retrievers, and 1 (3%) each of various other breeds or breed mixes. For the 24 case dogs with available information, median time from the date of last vaccination to the onset of clinical signs of GME was 212 days.
Control dogs
Median age of the 91 dogs in the control group was 7.0 years (range, 0.4 to 14.2 years), and median body weight was 10.4 kg (22.9 lb; range, 1.6 to 43.4 kg [3.5 to 95.5 lb]). Median BCS of the 81 dogs for which it could be determined was 6 (range, 3 to 8). Forty (44%) dogs were female (37 spayed and 3 sexually intact), and 51 (56%) were male (47 castrated and 4 sexually intact). For 10 control dogs, the age range for matching to case dogs needed to be expanded to allow matching by breed. For the 91 dogs with available information, median time from last vaccination to the visit of study inclusion was 235 days.
Statistical analysis
Month and season of onset of clinical signs of GME were fairly evenly distributed within the case group, with no unequal distribution identified among seasons (P = 0.48; Figure 1; Table 1). No significant difference was identified between case and control groups regarding age, body weight, BCS, or time of last vaccination. The odds of a GME diagnosis did not change with human population density (OR, 1.00; 95% confidence interval, 0.99 to 1.01; P = 0.11). Although females appeared more likely to have a GME diagnosis than males, this association was not significant (P = 0.06). This relationship did not differ after controlling for BCS, time of last vaccination, or population density.
Distributions of dogs with a diagnosis of GME (case dogs; black bars; n = 31) by month of onset of clinical signs and of matched control dogs (white bars; 91) by month of the visit that qualified them for study inclusion.
Citation: Journal of the American Veterinary Medical Association 254, 7; 10.2460/javma.254.7.822
Distributions of dogs with a diagnosis of GME (case dogs; black bars; n = 31) by month of onset of clinical signs and of matched control dogs (white bars; 91) by month of the visit that qualified them for study inclusion.
Citation: Journal of the American Veterinary Medical Association 254, 7; 10.2460/javma.254.7.822
Distributions of dogs with a diagnosis of GME (case dogs; black bars; n = 31) by month of onset of clinical signs and of matched control dogs (white bars; 91) by month of the visit that qualified them for study inclusion.
Citation: Journal of the American Veterinary Medical Association 254, 7; 10.2460/javma.254.7.822
Results of univariate analysis of putative categorical risk factors for a GME diagnosis in dogs.
| Factor | Proportion (%) of case dogs | Proportion (%) of control dogs | OR (95% CI) | P value |
|---|---|---|---|---|
| BCS | ||||
| > 5 | 15/29 (52) | 47/81 (58) | Referent | - |
| ≤ 5 | 14/29 (48) | 34/81 (42) | 1.38 (0.56–3.43) | 0.49 |
| Time since last vaccination (mo) | ||||
| ≤ 1 | 1/24 (4) | 10/91 (11) | 0.40 (0.05–3.46) | 0.36 |
| > 1 | 23/24 (96) | 81/91 (89) | Referent | - |
| ≤ 6 | 9/24 (38) | 35/91 (38) | 0.84 (0.33–2.16) | 0.72 |
| > 6 | 15/24 (62) | 56/91 (62) | Referent | - |
| ≤ 12 | 15/24 (62) | 69/91 (76) | 0.55 (0.20–1.51) | 0.25 |
| > 12 | 9/24 (38) | 22/91 (24) | Referent | - |
| Sex | ||||
| Male | 12/31 (39) | 51/91 (56) | Referent | - |
| Female | 19/31 (61) | 40/91 (44) | 2.35 (0.96–5.74) | 0.06 |
| Season of diagnosis | ||||
| Spring | 7/31 (23) | 30/91 (33) | 0.36 (0.10–1.30) | 0.12 |
| Summer | 5/31 (16) | 17/91 (19) | 0.53 (0.14–1.94) | 0.34 |
| Fall | 9/31 (29) | 15/91 (16) | Referent | - |
| Winter | 10/31 (32) | 29/91 (32) | 0.59 (0.20–1.74) | 0.34 |
CI = Confidence interval.
- = Not applicable.
Discussion
The purpose of the present study was to determine whether dog sex, body weight, BCS, time of last vaccination, season of onset of clinical signs, or human population density was significantly associated with a histologic diagnosis of GME in dogs. None of these variables, including the evaluated environmental factors, were associated with a GME diagnosis. A potential female predilection for meningoencephalomyelitis of unknown etiology and IMHA in dogs has been reported,1,4,12,13 but not for immune-meditated diseases such as SRMA.5 Although female dogs appeared more likely to have GME than male dogs in the present study, this association was not significant. Whether this lack of association was attributable to a low sample size and related low statistical power remains unknown. A sex predilection, if it existed, could suggest that the presence, or proportion, of various amounts of androgens may influence the development of immunemediated disease. The possibility of a complex interaction between sex hormones, X chromosomes, and the development of GME should be explored in future research.
Dogs with a BCS ≥ 6 had no greater odds of a GME diagnosis than dogs with a BCS ≤ 5 in the case-control study reported here. Although no causal associations can be identified through observational studies, this finding suggested that overweight and obesity (grouped together here) were not associated with GME. However, whether obesity specifically is associated with or contributes to GME remains to be determined.
Vaccination has been suggested as a trigger of autoimmune disease in dogs, although whether such an association exists is controversial. In 1 study,14 significantly more dogs developed IMHA within 1 month after vaccination than control dogs. Conversely, in a case-control study,5 vaccination within 6 weeks prior to onset of clinical signs of SRMA (case dogs) or the visit of study inclusion (control dogs) was not associated with an increase in the odds of the disease. Similarly, the present study revealed no association between the time of last vaccination, as categorized, and a GME diagnosis, providing no evidence to support client concerns about a temporal association between these 2 variables.
No association was identified between human population density and a GME diagnosis for the dogs of the present study. However, caution is warranted in interpreting this finding given that dogs in low-density areas may still have had exposure to higher-density areas and vice versa, confounding the results. Additional studies should be performed with dogs restricted to rural or urban environments to avoid this limitation.
No seasonal trend was identified in GME diagnoses in the present study, although in a previous study,13 more dogs had onset of IMHA during the spring and summer in California than during the fall and winter. Multiple sclerosis in humans and autoimmune encephalomyelitis in mice are immune-mediated CNS diseases that have been extensively investigated for potential environmental triggers. Exposure to UV radiation is suggested to be protective against development of both conditions owing to activation of regulatory T cells and induction of anti-inflammatory cytokines. Higher versus lower exposure to UV radiation is associated with a reduced incidence of multiple sclerosis and experimentally induced autoimmune encephalitis,15–17 and a significant decrease in the incidence of multiple sclerosis has been identified at low latitudes (ie, close to the equator).17 Although no significant association with season was identified for the dogs with GME in the present study, exploration of the amount or rate of exposure to UV radiation may be worthy of further investigation.
Limitations of the present study included those inherent to all retrospective studies, such as reliance on previously recorded data that may have been inaccurate or incomplete as well as the subjectivity of BCS recording, the small sample size, and the restricted geographic range of included dogs. Furthermore, the small numbers of dogs in various categories precluded inclusion of all predictor variables in 1 model, which would have been ideal. Although effort was made to ensure dogs with a diagnosis of GME truly had the disease, errors in diagnosis were possible as well.
Acknowledgments
The authors declare that there were no conflicts of interest. No funding was received for this study.
ABBREVIATIONS
| BCS | Body condition score |
| GME | Granulomatous meningoencephalomyelitis |
| IMHA | Immune-mediated hemolytic anemia |
| SRMA | Steroid-responsive meningitis arteritis |
Footnotes
Get latitude and longitude, LatLong.net. Available at: www.latlong.net. Accessed Dec 1, 2016.
Circular Area Profiles (CAPS), American Community Survey (ACS) version, Missouri Census Data Center, Columbia, Mo. Available at: mcdc.missouri.edu. Accessed Dec 1, 2016.
R: A language and environment for statistical computing, version 3.3.2.tar.gz, R Foundation for Statistical Computing, Vienna, Austria.
References
1. Granger N, Smith PM, Jeffery ND. Clinical findings and treatment of non-infectious meningoencephalomyelitis in dogs: a systematic review of 457 published cases from 1962 to 2008. Vet J 2010;184:290–297.
2. Uchida K, Park E, Tsuboi M, et al. Pathological and immunological features of canine necrotising meningoencephalitis and granulomatous meningoencephalitis. Vet J 2016;213:72–77.
3. Barber RM, Porter BF, Li Q, et al. Broadly reactive polymerase chain reaction for pathogen detection in canine granulomatous meningoencephalomyelitis and necrotizing meningoencephalitis. J Vet Intern Med 2012;26:962–968.
4. Talarico LR, Schatzberg SJ. Idiopathic granulomatous and necrotising inflammatory disorders of the canine central nervous system: a review and future perspectives. J Small Anim Pract 2010;51:138–149.
5. Rose JH, Kwiatkowska M, Henderson ER, et al. The impact of demographic, social, and environmental factors on the development of steroid-responsive meningitis-arteritis (SRMA) in the United Kingdom. J Vet Intern Med 2014;28:1199–1202.
6. Schrauwen I, Barber RM, Schatzberg SJ, et al. Identification of novel genetic risk loci in Maltese Dogs with necrotizing meningoencephalitis and evidence of a shared genetic risk across toy dog breeds. PLoS One 2014;9:e112755.
7. Barber RM, Schatzberg SJ, Corneveaux JJ, et al. Identification of risk loci for necrotizing meningoencephalitis in pug dogs. J Hered 2011;102(suppl 1):S40-S46.
8. Greer KA, Schatzberg SJ, Porter BF, et al. Heritability and transmission analysis of necrotizing meningoencephalitis in the Pug. Res Vet Sci 2009;86:438–442.
9. Gregory SG, Schmidt S, Seth P, et al. Interleukin 7 receptor α chain (IL7R) shows allelic and functional association with multiple sclerosis. Nat Genet 2007;39:1083–1091.
10. Lundmark F, Duvefelt K, Iacobaeus E, et al. Variation in interleukin 7 receptor α chain (IL7R) influences risk of multiple sclerosis. Nat Genet 2007;39:1108–1113.
11. Summers BA, Cummings JF, de Lahunta A. Veterinary neuropathology. St Louis: Mosby, 1995;110–111.
12. Muñana KR, Luttgen PJ. Prognostic factors for dogs with granulomatous meningoencephalomyelitis: 42 cases (1982–1996). J Am Vet Med Assoc 1998;212:1902–1906.
13. Kidd L, Rasmussen R, Chaplow E, et al. Seasonality of immunemediated hemolytic anemia in dogs from southern California. J Vet Emerg Crit Care (San Antonio) 2014;24:311–315.
14. Duval D, Geiger U. Vaccine-associated immune-mediated hemolytic anemia in the dog. J Vet Intern Med 1996;10:290–295.
15. Wang Y, Marling SJ, Beaver EF, et al. UV light selectively inhibits spinal cord inflammation and demyelination in experimental autoimmune encephalomyelitis. Arch Biochem Biophys 2015;567:75–82.
16. Davis GE, Lowell WE. Solar cycles and their relationship to human disease and adaptability. Med Hypotheses 2006;67:447–461.
17. Spelman T, Gray O, Trojano M, et al. Seasonal variation of relapse rate in multiple sclerosis is latitude dependent. Ann Neurol 2014;76:880–890.
