- Email: [email protected]
Toxoplasma gondii seroprevalence and association with equine protozoal myeloencephalitis: A case–control study of Californian horses
The Veterinary Journal 224 (2017) 38–43
Contents lists available at ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Original Article
Toxoplasma gondii seroprevalence and association with equine protozoal myeloencephalitis: A case–control study of Californian horses K.E. Jamesa,* , W.A. Smitha , A.E. Packhamb , P.A. Conradb , N. Pusterlaa a b
Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California, 1 Shields Avenue, Davis, CA, 95616, USA Department of Pathology, Microbiology and Immunology, School of Veterinary Medicine, University of California, 1 Shields Avenue, Davis, CA, 95616, USA
A R T I C L E I N F O
Article history: Accepted 18 May 2017 Keywords: Equine protozoal myeloencephalitis IFAT Neospora hughesi Sarcocystis neurona Toxoplasma gondii
A B S T R A C T
While toxoplasmosis is not commonly considered a clinical disease of equines, previous seroprevalence studies have reported differing background rates of Toxoplasma gondii infection in horses globally. The objective of this study was to evaluate possible associations between T. gondii seroprevalence and clinical signs of equine protozoal myeloencephalitis (EPM) in horses. Using a case–control study design, 720 Californian horses with neurologic signs compatible with EPM were compared to healthy, nonneurologic horses for the presence of T. gondii antibodies (using indirect fluorescent antibody tests [IFAT]). Toxoplasma gondii seroprevalence among cases and controls was determined at standard serum cut-offs: 40, 80, 160, 320, and 640. At a T. gondii titre cut-off of 320, horses with clinical signs compatible with EPM had 3.55 times the odds of a seropositive test compared to those without clinical signs (P < 0.01) when adjusted for covariates. When restricted to the autumn season and at the same titre cut-off, an EPM suspect horse had 6.4 times the odds of testing seropositive to T. gondii, compared to non-neurologic horses. The association between high T. gondii titres and clinical signs compatible with EPM is potentially reflective of toxoplasmosis in equines. Serologic testing of cerebrospinal fluid and isolation of T. gondii in EPM suspect cases should be considered. Future studies investigating the relationship between T. gondii and EPM are warranted. © 2017 Elsevier Ltd. All rights reserved.
Introduction Toxoplasma gondii, an obligate intracellular apicomplexan parasite, is an important pathogen of both humans and warmblooded animals. While toxoplasmosis is not commonly considered a disease of equines, previous seroprevalence assays have been conducted globally over the past decade to determine the potential risk of T. gondii in horsemeat destined for human consumption (Tenter et al., 2000; Pomares et al., 2011). These studies have produced differing background rates of T. gondii infection in horses, with seroprevalences ranging from >50% in Egypt (Ghazy et al., 2007; Shaapan and Ghazy, 2007) to 6% in the United States (Dubey et al., 2014). While there is no validated immunoassay test for detecting T. gondii antibodies in equines, several dilution cut-offs to determine positive tests have been used throughout the literature, ranging from 1:6 for the modified agglutination test (MAT; Paştiu
* Corresponding author. E-mail address: [email protected] (K.E. James). http://dx.doi.org/10.1016/j.tvjl.2017.05.008 1090-0233/© 2017 Elsevier Ltd. All rights reserved.
et al., 2015) to 1:64 for the indirect fluorescent antibody test (IFAT; Fonseca de Araujo Valenca et al., 2015). Riskfactors for clinical infections in equine protozoal myeloencephalitis (EPM), a neurologic disease caused by two other apicomplexan protozoal parasites (Sarcocystis neurona and Neospora hughesi), are currently unknown. A biological model that can reproducibly transition an asymptomatic infection to clinical disease has not yet been developed (Saville et al., 2000). Experimental studies to induce EPM in horses with severe combined immunodeficiency (SCID) did not progress to infections of the spinal cord with associated signs of EPM, although infection was confirmed by demonstrating the presence of antibody in visceral tissues, (Sellon et al., 2004). However, healthy, immunocompetent horses can develop clinical signs and tissue damage consistent with EPM (as confirmed post-mortem using immunohistochemistry). This seems to suggest that an immune-mediated inflammatory process may be responsible for parasitic migration to the site of clinicaldisease, i.e. the nervous system (Lewis et al., 2014). In a recent study of Brazilian horses with high antibody titres for T. gondii (IFAT titres 64–1024), seropositive horses had elevated
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
humoral and cellular immune responses compared to seronegative horses (Do Carmo et al., 2015). Considering a comparative species approach, polyparasitism in wild marine mammals has been shown to modulate the severity of protozoal encephalitis, most notably co-infections by S. neurona and T. gondii, (Gibson et al., 2011). To help elucidate a possible role for T. gondii in the progression of EPM, the present study used a case–control study design to test the hypothesis that horses with neurologic signs compatible with EPM would have higher antibody titres to T. gondii compared to horses with no neurologic signs. Materials and methods Case selection Cases were defined as horses with at least one of the following clinical signs compatible with EPM: ataxia, muscle atrophy, head tilt, hypermetria, lameness, weakness, circling, mentation change. Sera were tested at the Immunology Laboratory of the William R. Pritchard Veterinary Medical Teaching Hospital (VMTH), School of Veterinary Medicine, University of California at Davis, for the presence of antibodies against S. neurona and N. hughesi using Sarcofluor (Conrad Protozoal Laboratory) and Neofluor (Conrad Protozoal Laboratory) IFAT during 2013. Study inclusion criteria included Sarcofluor and Neofluor serum results, availability of serum to be tested for the presence of T. gondii; horses must also have been resident in California. Equine health records contained breed, sex, use, and age data, but did not include any information on prior treatment for EPM. There were 5592 diagnostic submission requests for Sarcofluor and/or Neofluor from 1 January, 2013 to 31 December, 2013; this was the study population from which cases were selected based on study inclusion criteria. Control horse selection Controls were resident horses of California with no current neurologic signs during 2013. Otherwise, this population was comparable to the cases in terms of residence, availability of data on covariates, and distribution of demographic factors. This was originally a convenience sample of non-neurologic horses for a separate study of leptospirosis. Participating clinics enrolled voluntarily and were asked to sample approximately 110 horses/ practice. A 10 mL blood specimen was drawn from the jugular vein, and a questionnaire was completed for each horse to collect data on age, breed, primary use, and sex. Information regarding prior treatment for EPM was not collected. Serum from 5250 horses was collected across 18 states in September and October 2013. This constituted the study population from which the study sample of controls was selected. The presence of antibodies against S. neurona and N. hughesi was determined as previously described (Sarcofluor and Neofluor IFAT; James et al., 2017).
39
testing for T. gondii in equines, each sample was end-titrated to generate quantitative titres of antibodies directed to T. gondii. Two investigators were blinded to case or control status during the reading of the IFAT slides and read the slides independently. Statistical analysis All statistical analyses and data management were performed using Stata Statistical Software: Release 14 (StataCorp). Demographic frequency tables for the case and control populations were created to compare populations. Demographic factors included breed, sex, age, and use. Breed was divided into Quarter horse, Warmblood, Thoroughbred, Paint horse, Arabian, Draft horse, Pony/miniature, and other. Age was analysed as a 5-year increment categorical variable. Sex was categorised into male (gelding and stallion) and female. Use of animal was as follows: competition animals, resident farm animals, breeding animals, or other uses. Clinical factors, including S. neurona and N. hughesi seropositivity, were evaluated for each population. The S. neurona seropositive titre cut-off was set at 40; for N. hughesi a seropositive titre cut-off was set at 160. Bivariate logistic regression models between the demographic and potential co-infection factors with the clinical outcome were generated to determine possible covariates for a multivariable logistic regression model. Odds ratios (ORs; 95% confidence intervals, 95%CI) with P < 0.05 were considered statistically significant and were included in the multivariable logistic regression model in a forward stepping manner. Confounding and first order interaction of predictor variables was assessed for the clinical outcome in preliminary multivariable logistic regression models (Agresti, 2002). Toxoplasma gondii seroprevalence was determined for case and control populations at multiple titre cut-offs. Due to the lack of a generally accepted IFAT titre cut-off for determining seropositivity in equines, seroprevalence at reciprocal titres of 40, 80, 160, 320, and 640 were determined. Unadjusted effect measures for the association of T. gondii exposure at these titre cut-offs and the clinical outcome were created (unadjusted ORs). A multivariable logistic regression model was then created with the statistically significant demographic explanatory variables and the T. gondii exposure at the titre cut-offs. A further refinement of the statistical analysis included matching cases to the same season as controls to increase matching of the two groups. As described, all controls were selected in September and October 2013. In a secondary analysis, only cases with serum submitted to the UC Davis Immunology Laboratory from August to November were included, to investigate the effect of the autumn season. The demographic frequencies, bivariate associations, T. gondii exposure prevalences, and final multivariable logistic regression models were generated as described above. Results
Toxoplasma gondii exposure
Primary analysis results: All year 2013 cases and controls
To determine exposure status to T. gondii,T. gondii antibody titres were determined for cases and controls using IFAT, as previously described (Conrad et al., 1993; Miller et al., 2001). The T. gondii strain used in testing was a type 2 strain (ME49) available from American Type Cell Culture research foundation (ATCC). Tachyzoites were grown on monkey kidney cells (MA104) in 10% fetal bovine media, incubated at 37 C. Slide fixation techniques included a 10 min fix with formalin, followed by a phosphate buffered saline wash. However, instead of anti-ferret IgG, antihorse IgG (Jackson ImmunoResearch Laboratories) conjugated to fluorescein isothiocyanate (FITC) was used as the secondary antibody, diluted 1:100. Due to the lack of validated diagnostic
Of 5592 horses from the case population, 392 resided in California. Cases were not equally distributed throughout the calendar year and the majority of cases occurred in April–July (n = 146); however, this is not statistically different from August to November, in which there were 138 cases (P > 0.05). December– March had the fewest cases, with 105 cases total. Of the 5250 horses from the control population, 328 resided in California. Important risk factors from bivariate modeling (P < 0.05) were age and breed (Table 1), and were therefore included in the multivariable logistic regression model. Clinical characteristics for S. neurona and N. hughesi seropositivity also differed between the case and control populations.
40
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
Table 1 Demographic characteristics of horses sampled from California in 2013 and bivariate associations with equine protozoal myeloencephalitis (EPM) outcome status. Explanatory variables
Case outcome status: Horses with neurological signs compatible with EPM (n = 392) n (%)
Control outcome status: Horses with no neurological signs (n = 328)
Unadjusted odds ratio (95% confidence intervals)
Demographic characteristics Sex Female Male Unknown
124 (31%) 255 (65%) 13 (4%)
118 (36%) 199 (60%) 11 (4%)
1.0 (Reference) 1.21 (0.88–1.68) –
Age 1–5 years old 6–10 years old 11–15 years old 16–20 years old Over 20 years Unknown
92 (24%) 109 (27%) 74 (19%) 47 (12%) 47 (12%) 23 (5%)
31 (9%) 96 (30%) 106 (32%) 52 (16%) 36 (11%) 7 (2%)
1.0 (Reference) 0.38b (0.23–0.63) 0.24b (0.14–0.38) 0.30b (0.17–0.53) 0.44b (0.24–0.80) –
Breed Quarter horse Warmblood Thoroughbred Paint Arabian Draft horse Pony Other Unknown
91 (23%) 68 (17%) 95 (24%) 20 (5%) 23 (6%) 12 (3%) 5 (1%) 65 (17%) 13 (3%)
57 (17%) 116 (35%) 29 (8%) 22 (6%) 34 (10%) 19 (6%) 13 (4%) 37 (11%) 1 (<1%)
1.0 (Reference) 0.32b (0.21–0.50) 1.79a (1.06–3.03) 0.50a (0.25–0.99) 0.37b (0.20–0.69) 0.35a (0.16–0.76) 0.21b (0.07–0.62) 0.96 (0.57–1.62) –
Use Competition Ranch/farm Breeding Other Unknown
138 (35%) 159 (40%) 7 (1%) 1 (<1%) 87 (22%)
184 (56%) 121 (37%) 11 (4%) 0 (0%) 12 (4%)
1.0 (Reference) 1.07 (0.79–1.45) 0.52 (0.19–1.37) – –
Clinical characteristics Sarcocystis neurona serum titre <40 40
302 (77%) 90 (23%)
66 (20%) 262 (80%)
1.0 (Reference) 0.07b (0.05–0.11)
Neospora hughesi serum titre <160 160
386 (98%) 6 (2%)
185 (56%) 143 (44%)
1.0 (Reference) 0.02b (0.01–0.05)
a b
P < 0.05. P < 0.01.
Seropositivity for S. neurona at a reciprocal titre of 40 was highly associated with control group status (bivariate OR, 0.07; 95%CI, 0.05–0.11); similarly, seropositivity for N. hughesi at a reciprocal titre of 160 was associated with control group status (bivariate OR, 0.02; 95%CI, 0.01–0.05). These were considered important risk factors to include in a multivariable logistic regression model (P < 0.05; Table 1). Toxoplasma gondii prevalences at the specified titres are presented in Table 2. Significant adjusted ORs (adjusted for age,
breed, S. neurona seropositivity, and N. hughesi seropositivity) were only seen at a T. gondii titre cut-off of 320; cases had 3.55 times the odds of a T. gondii titre of 320 compared to controls. While all T. gondii cut-off titres were analysed, the cut-off titre of 320 generated the highest likelihood of clinical signs in horses. The full results of the multivariable logistic model at a T. gondii cut-off titre of 320 are presented in Table 2 and in the Appendix: Supplementary Tables S1 and S2. A T. gondii titre of 320 (OR, 3.55; 95%CI, 1.41–9.0) was associated with suspect EPM clinical status.
Table 2 Prevalence of Toxoplasma gondii serum titres in case horses with neurological signs compatible with equine protozoal myeloencephalitis (n = 392) and control horses with no neurological signs (n = 328), and Toxoplasma gondii titre odds ratios (OR) associated with case status at various serum titre cut-offs. Toxoplasma gondii serum titre Case 40 80 160 320 640 a b
OR titre (95% confidence intervals)
Prevalence
241 (62%) 121 (31%) 78 (20%) 49 (13%) 18 (5%)
Control 239 (73%) 107 (33%) 54 (16%) 16 (5%) 5 (2%)
Unadjusted OR b
0.59 (0.43–0.83) 0.92 (0.66–1.28) 1.26 (0.84–1.89) 2.7b (1.52–5.35) 3.10b (1.10–10.8)
Adjusted for statistically significant covariates (P < 0.05): age, breed, Sarcocystis neurona serum titre 40, and Neospora hughesi serum titre 160. P < 0.01.
Adjusteda OR 0.89 (0.58–1.38) 1.52 (0.96–2.44) 1.77 (0.98–3.22) 3.55b (1.4-9.0) 4.12 (0.70–24.6)
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
41
Table 3 Seasonally restricted demographic characteristics and bivariate associations with equine protozoal myeloencephalitis (EPM) outcome status in horses sampled from California in August–November, 2013. Explanatory variables
Case outcome status: Horses with neurological signs compatible with EPM (n = 139) n (%)
Control outcome status: Horses with no neurological signs (n = 328)
Unadjusted odds ratio (95% CI)
Demographic characteristics Sex Female Male Unknown
45 (32%) 88 (63%) 6 (5%)
118 (36%) 199 (60%) 11 (4%)
1.0 (Reference) 1.15 (0.74–0.45) –
Age 1–5 years old 6–10 years old 11–15 years old 16–20 years old Over 20 years Unknown
34 (24%) 40 (29%) 23 (17%) 16 (11%) 19 (14%) 7 (5%)
31 (9%) 96 (30%) 106 (32%) 52 (16%) 36 (11%) 7 (2%)
1.0 (Reference) 0.38b (0.21–0.70) 0.20b (0.11–0.38) 0.28b (0.13–0.59) 0.48 (0.23–1.01) –
Breed Quarter horse Warmblood Thoroughbred Paint Arabian Draft horse Pony Other Unknown
29 (21%) 28 (20%) 28 (20%) 6 (4%) 12 (8%) 5 (3%) 1 (<1%) 24 (17%) 6 (4%)
57 (17%) 116 (35%) 29 (8%) 22 (6%) 34 (10%) 19 (6%) 13 (4%) 37 (11%) 1 (<1%)
1.0 (Reference) 0.39b (0.21–0.71) 1.57 (0.81–3.06) 0.44 (0.16–1.20) 0.57 (0.26–1.25) 0.43 (0.15–1.25) 0.12a (0.02–.99) 1.0 (0.54–2.05) –
Use Competition Ranch/farm Breeding Other Unknown
46 55 2 0 36
(33%) (39%) (1%) (0%) (25%)
184 (56%) 121 (37%) 11 (4%) 0 (0%) 12 (4%)
1.0 (Reference) 1.02 (0.68–1.52) 0.41 (0.08–1.88) – –
Clinical characteristics Sarcocystis neurona serum titre <40 40
104 (75%) 35 (25%)
66 (20%) 262 (80%)
1.0 (Reference) 0.08b (0.05–0.14)
Neospora hughesi serum titre <160 160
136 (98%) 3 (2%)
185 (56%) 143 (44%)
1.0 (Reference) 0.03b (0.01–0.09)
a b
P < 0.05. P < 0.01.
No first order interaction terms were statistically significant at P < 0.05, and confounding factors such as age and breed were adjusted for in the multivariable logistic regression model. Secondary analysis results: Seasonal restriction (August–November, 2013) Of the 392 cases included in this study, 139 met the seasonal criteria matching the collection season of the controls (August–
November, within one month of the September–October control selection period). The seasonal case and control population demographics were similar to the 1-year case and control populations, as presented in Table 3. Breed, age and S. neurona and N. hughesi seropositivity, were statistically significant in the bivariate analysis (P < 0.05), and were therefore included in the multivariable logistic regression analysis. Toxoplasma gondii seroprevalence at the titre cut-offs described in the primary analysis are presented in Table 4. When the association between T.
Table 4 Seasonally restricted prevalence of Toxoplasma.gondii serum titres in case horses with neurological signs compatible with equine protozoal myeloencephalitis (n = 392) and control horses with no neurological signs (n = 328), and T. gondii titre odds ratios (OR) associated with case status at various serum titre cut-offs. Toxoplasma gondii serum titre
40 80 160 320 640 a b c d
Prevalence
OR (95% confidence intervals)
Casea
Controla
Unadjusted OR
90 (65%) 52 (37%) 35 (25%) 24 (17%) 7 (5%)
239 (73%) 107 (33%) 54 (16%) 16 (5%) 5 (2%)
0.68 (0.43–1.08) 1.23 (0.80–1.90) 1.71c (1.02–2.83) 4.07d (1.98–8.48) 3.43c (0.91–14.0)
Adjustedb OR 1.20 (0.67–2.08) 2.15d (1.21–3.82) 2.79d (1.38–5.63) 6.40d (2.27–18.0) 6.53 (0.92–46.4)
Only horses sampled from California in August–November, 2013 were included in the analysis. Adjusted for statistically significant covariates (P < 0.05): age, breed, Sarcocystis neurona serum titre 40, and Neospora hughesi serum titre 160. P < 0.05. P < 0.01.
42
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
gondii and the outcome was adjusted for breed, age, and S. neurona and N. hughesi seropositivity, significant associations with the outcome were determined at T. gondii titre cut-offs of 80, 160, 320, and 640. At a titre cut-off of 320, cases had 6.40 times the odds of being T. gondii seropositive compared to controls. No first order interaction terms were statistically significant at P < 0.05, and confounding factors such as age and breed were adjusted for in the multivariable logistic regression model. Appendix: Supplementary material contains tables describing the risk factors associated with the multivariable logistic regression model of the highest statistically significant titre cut-off for T. gondii (320). The other T. gondii cut-offs showed similar patterns. Warmbloods had lower odds than Quarter horses for being cases (OR, 0.36; 95%CI, 0.20–0.95), while 11–15 year old horses had lower odds than the reference category of 1–5 year olds for being cases (OR, 0.31; 95%CI, 0.13–0.74). Sarcocystis neurona and N. hughesi seropositive horses had greater odds of being controls than cases. Toxoplasma gondii seropositivity was associated with case status; cases have 6.4 times the odds of controls of being seropositive at a titre cut-off of 320. Discussion In this study, horses with higher titres to T. gondii (80, 160, 320) were more likely to have clinical signs compatible with EPM than healthy, non-neurologic horses. While there was no evidence that co-infections by T. gondii and S. neurona/N. hughesi were required for clinical signs of EPM to develop, the association between T. gondii seropositivity and clinical EPM suggested that T. gondii was associated with neurologic signs in the study population of Californian horses. When clinical outcome was restricted to only those cases and controls collected in autumn, this association became even stronger, and horses with a T. gondii serum titre of 320 had >6 times the odds of clinical signs compatible with EPM as horses with T. gondii serum titres <320. The cases and control horses in this study were similar in sex and use distributions, and there were no statistically significant differences between the two potential clinical outcomes. However, there were differences in the breed and age distributions. Cases were more likely to be younger, aged 1–5 years, compared to controls. This could be due to the fact that 1–5 yearold horses are at greatest risk for EPM and could have clinical signs that would potentially require diagnostic testing (Saville et al., 2000). In terms of breed distribution, cases were also more likely to be Thoroughbred, while controls were more likely to be Warmbloods. This is in agreement with previous studies that reported Thoroughbreds and Quarter horses were more likely to have EPM than other breeds, while Warmbloods had notably lower rates of clinical disease (Fayer et al., 1990; Pusterla et al., 2014). Perhaps the most interesting difference between the case and control population was the seroprevalence of S. neurona and N. hughesi as determined by IFAT. While the titre cut-offs used in this study are used primarily for healthy horses (Duarte et al., 2004a,b), to allow comparison between the two populations, these cut-offs were also applied to horses with neurologic signs. While this may have overestimated the prevalence of clinically relevant infections of the two parasites, it determined background infection rates in the two populations. Controls were much more likely to be seropositive for both parasites at the defined cut-offs (40 for S. neurona, 160 for N. hughesi; Pusterla et al., 2014). This could be due to several factors, most notably a limitation in our study design, which used serodiagnostics only. Because the control group was older and contained were more Warmbloods, higher seroprevalence of S. neurona and N. hughesi were expected, as older age and Warmblood breed have previously been associated with higher
rates of seropositivity for these two parasites (Saville et al., 2000; James et al., 2017). When the inclusion criteria for cases was limited to the season that matched the controls, the association between T. gondii and the clinical outcome was stronger and exhibited more of a dose response (higher titre cut-offs were associated with greater odds of clinical outcome). In the late summer/autumn season (August– November), horses with T. gondii titres of 320 had 6.4 times the odds of being a case than a control. Since there was no evidence that co-infection played a role in the development of clinical signs, this association could indicate T. gondii-related neurologic disease. In horses, concurrent infections with seasonal viral and bacterial infections, and/or seasonal fluctuations in immune function, could precipitate the activation of Toxoplasma cysts, and could contribute to seropositivity due to increased anti-T. gondii antibody activity. In other species, there is evidence that increased infection rates in the later part of the calendar year could be due to peak periods of oocyst shedding in domestic kittens and cats during late summer and autumn, and the accumulation of oocysts in the environment during dry weather (Schares et al., 2008; Herrmann et al., 2010; Schares et al., 2016). As cats are the definitive host of T. gondii, environmental oocyst contamination increases with local cat population density, and the initial infection can lead to substantial oocyst shedding (Dabritz et al., 2007; Vanwormer et al., 2013a, 2016). Oocysts build up in the environment over dry weather, and are washed from land to sea during the initial rainfalls of the wet season, leading to seasonal loading of oocyst contamination during drier months. Cat reproduction cycles could also affect the environmental burden of viable oocysts. The seasonal load of T. gondii oocysts shed by cats could increase if dry seasons are prolonged, followed by late wet seasons, a trend that has been occurring in California. As such, T. gondii infection in horses could become more common in other parts of the year (VanWormer et al., 2013b; Schares et al., 2016; VanWormer et al., 2016). Limitations in data collection for this study did not allow for an investigation of whether season was important in case status, but future studies should aim to collect cases and controls throughout the year to determine any seasonality in equine infection. Although there is no evidence in this study that co-infections by T. gondii and S. neurona/N. hughesi were critical for the development of the neurologic signs of EPM, previous studies of the relationship between protozoal myeloencephalitis and polyparasitism in other species suggests that co-infections should not be ignored in the development of EPM (Gibson et al., 2011). The association between high T. gondii titres (>160) is potentially suggestive of toxoplasmosis in equines. Future studies of possible links between T. gondii infection and neurologic abnormalities in equines should include expansion of serologic testing to include cerebrospinal fluid samples, as well as isolating T. gondii from cases of EPM at necropsy using culture, immunohistochemistry, or PCR. Conclusions Further studies should be based on confirmed serum titre results using serum: CSF titre ratios (ante-mortem) and/or necropsy and direct identification of T. gondii (post-mortem), if possible, and should include cases where clinical signs and lesions are compatible with EPM but there is no evidence of S. neurona or N. hughesi infection. The strong association between T. gondii serum titre and signs compatible with EPM demonstrated in this study could be augmented by changes to climate and ecosystems which favour T. gondii oocyst shedding from the definitive host (domestic and wild felids) and enhance the viability of the oocyst in the environment. Although this study produced no evidence that co-infections by T. gondii and S. neurona/N. hughesi were associated with more severe signs of EPM, the association between
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
T. gondii serum titres and neurological abnormalities consistent with EPM should be examined further, especially considering the worldwide distribution of T. gondii. Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper. Acknowledgements This study was supported in full by the Center for Equine Health, University of California at Davis. We wish to thank Dr. Eva Tamez-Trevino and the Immunology Laboratory of the William R. Pritchard Veterinary Medical Teaching Hospital, School of Veterinary Medicine, University of California at Davis and Zoetis, Inc. for their assistance in sample collection. Appendix: Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.tvjl.2017.05.008. References Agresti, A., 2002. Building and applying logistic regression models, Categorical Data Analysis. Second Ed. John Wiley and Sons, Inc., Hoboken, NJ, USA, pp. 211–266. Conrad, P.A., Barr, B.C., Sverlow, K.W., Anderson, M., Daft, B., Kinde, H., Dubey, J.P., Munson, L., Ardans, A., 1993. In vitro isolation and characterization of a Neospora spp. from aborted bovine foetuses. Parasitology 106, 239–249. Dabritz, H.A., Miller, M.A., Atwill, E.R., Gardner, I.A., Leutenegger, C.M., Melli, A.C., Conrad, P.A., 2007. Detection of Toxoplasma gondii-like oocysts in cat feces and estimates of the environmental oocyst burden. Journal of the American Veterinary Medical Association 231, 1676–1684. Do Carmo, G.M., Da Silva, A.S., Klauck, V., Pazinato, R., Moura, A.B., Duarte, T., Duarte, M.M., Bochi, G.V., Moresco, R.N., Stefani, L.M., 2015. Immunological response and markers of cell damage in seropositive horses for Toxoplasma gondii. Comparative Immunology, Microbiology and Infectious Diseases 38, 9–13. Duarte, P.C., Conrad, P.A., Barr, B.C., Wilson, W.D., Ferraro, G.L., Packham, A.E., Carpenter, T.E., Gardner, I.A., 2004a. Risk of transplacental transmission of Sarcocystis neurona and Neospora hughesi in california horses. Journal of Parasitology 90, 1345–1351. Duarte, P.C., Daft, B.M., Conrad, P.A., Packham, A.E., Saville, W.J., MacKay, R.J., Barr, B. C., David Wilson, W., Ng, T., Reed, S.M., et al., 2004b. Evaluation and comparison of an indirect fluorescent antibody test for detection of antibodies to Sarcocystis neurona, using serum and cerebrospinal fluid of naturally and experimentally infected, and vaccinated horses. Journal of Parasitology 90, 379–386. Dubey, J.P., Ness, S.L., Kwok, O.C., Choudhary, S., Mittel, L.D., Divers, T.J., 2014. Seropositivity of Toxoplasma gondii in domestic donkeys (Equus asinus) and isolation of T. gondii from farm cats. Veterinary Parasitology 199, 18–23. Fayer, R., Mayhew, I.G., Baird, J.D., Dill, S.G., Foreman, J.H., Fox, J.C., Higgins, R.J., Reed, S.M., Ruoff, W.W., Sweeney, R.W., et al., 1990. Epidemiology of equine protozoal myeloencephalitis in North America based on histologically confirmed cases. A report. Journal of Veterinary Internal Medicine/American College of Veterinary Internal Medicine 4, 54–57. Fonseca de Araujo Valenca, S.R., Barreto Valenca, R.M., Pinheiro Junior, J.W., Feitosa de Albuquerque, P.P., Souza Neto, O.L., Mota, R.A., 2015. Risk factors of occurrence of Toxoplasma gondii among horses in the state of Alagoas Brazil.
43
Acta Parasitologica/Witold Stefanski Institute of Parasitology, Warszawa, Poland 60, 707–711. Ghazy, A.A., Shaapan, R.M., Abdel-Rahman, E.H., 2007. Comparative serological diagnosis of toxoplasmosis in horses using locally isolated Toxoplasma gondii. Veterinary Parasitology 145, 31–36. Gibson, A.K., Raverty, S., Lambourn, D.M., Huggins, J., Magargal, S.L., Grigg, M.E., 2011. Polyparasitism is associated with increased disease severity in Toxoplasma gondii-infected marine sentinel species. PLoS Neglected Tropical Diseases 5, e1142. Herrmann, D.C., Pantchev, N., Vrhovec, M.G., Barutzki, D., Wilking, H., Frohlich, A., Luder, C.G., Conraths, F.J., Schares, G., 2010. Atypical Toxoplasma gondii genotypes identified in oocysts shed by cats in Germany. International Journal for Parasitology 40, 285–292. James, K.E., Smith, W.A., Conrad, P.A., Packham, A.E., Guerrero, L., Ng, M., Pusterla, N., 2017. Seroprevalence of anti-Sarcocystis neurona and anti-Neospora hughesi antibodies among healthy equids in the United States. Journal of the American Veterinary Medical Association 250, 1291–1301. Lewis, S.R., Ellison, S.P., Dascanio, J.J., Lindsay, D.S., Gogal Jr., R.M., Werre, S.R., Surendran, N., Breen, M.E., Heid, B.M., Andrews, F.M., et al., 2014. Effects of experimental Sarcocystis neurona-induced infection on immunity in an equine model. Journal of Veterinary Medicine 2014, 239495. Miller, M.A., Sverlow, K., Crosbie, P.R., Barr, B.C., Lowenstine, L.J., Gulland, F.M., Packham, A., Conrad, P.A., 2001. Isolation and characterization of two parasitic protozoa from a Pacific harbor seal (Phoca vitulina richardsi) with meningoencephalomyelitis. The Journal of Parasitology 87, 816–822. , P., Rosenthal, B.M., Oltean, M., Villena, I., Paştiu, A.I., Györke, A., Kalmár, Z., Bolfa Spînu, M., Cozma, V., 2015. Toxoplasma gondii in horse meat intended for human consumption in Romania. Veterinary Parasitology 212, 393–395. Pomares, C., Ajzenberg, D., Bornard, L., Bernardin, G., Hasseine, L., Darde, M.L., Marty, P., 2011. Toxoplasmosis and horse meat France. Emerging Infectious Diseases 17, 1327–1328. Pusterla, N., Tamez-Trevino, E., White, A., VanGeem, J., Packham, A., Conrad, P.A., Kass, P., 2014. Comparison of prevalence factors in horses with and without seropositivity to Neospora hughesi and/or Sarcocystis neurona. The Veterinary Journal 200, 332–334. Saville, W.J., Reed, S.M., Morley, P.S., Granstrom, D.E., Kohn, C.W., Hinchcliff, K.W., Wittum, T.E., 2000. Analysis of risk factors for the development of equine protozoal myeloencephalitis in horses. Journal of the American Veterinary Medical Association 217, 1174–1180. Schares, G., Vrhovec, M.G., Pantchev, N., Herrmann, D.C., Conraths, F.J., 2008. Occurrence of Toxoplasma gondii and Hammondia hammondi oocysts in the faeces of cats from Germany and other European countries. Veterinary Parasitology 152, 34–45. Schares, G., Ziller, M., Herrmann, D.C., Globokar, M.V., Pantchev, N., Conraths, F.J., 2016. Seasonality in the proportions of domestic cats shedding Toxoplasma gondii or Hammondia hammondi oocysts is associated with climatic factors. International Journal for Parasitology 46, 263–273. Sellon, D.C., Knowles, D.P., Greiner, E.C., Long, M.T., Hines, M.T., Hochstatter, T., Tibary, A., Dame, J.B., 2004. Infection of immunodeficient horses with Sarcocystis neurona does not result in neurologic disease. Clinical and Diagnostic Laboratory Immunology 11, 1134–1139. Shaapan, R.M., Ghazy, A.A., 2007. Isolation of Toxoplasma gondii from horse meat in Egypt. Pakistan Journal of Biological Sciences 10, 174–177. Tenter, A.M., Heckeroth, A.R., Weiss, L.M., 2000. Toxoplasma gondii: from animals to humans. International Journal for Parasitology 30, 1217–1258. VanWormer, E., Carpenter, T.E., Singh, P., Shapiro, K., Wallender, W.W., Conrad, P.A., Largier, J.L., Maneta, M.P., Mazet, J.A., 2016. Coastal development and precipitation drive pathogen flow from land to sea: evidence from a Toxoplasma gondii and felid host system. Scientific Reports 6, 29252. Vanwormer, E., Conrad, P.A., Miller, M.A., Melli, A.C., Carpenter, T.E., Mazet, J.A., 2013a. Toxoplasma gondii, source to sea: higher contribution of domestic felids to terrestrial parasite loading despite lower infection prevalence. Ecohealth 10, 277–289. VanWormer, E., Fritz, H., Shapiro, K., Mazet, J.A., Conrad, P.A., 2013b. Molecules to modeling: Toxoplasma gondii oocysts at the human-animal-environment interface. Comparative Immunology, Microbiology and Infectious Diseases 36, 217–231.
Contents lists available at ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Original Article
Toxoplasma gondii seroprevalence and association with equine protozoal myeloencephalitis: A case–control study of Californian horses K.E. Jamesa,* , W.A. Smitha , A.E. Packhamb , P.A. Conradb , N. Pusterlaa a b
Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California, 1 Shields Avenue, Davis, CA, 95616, USA Department of Pathology, Microbiology and Immunology, School of Veterinary Medicine, University of California, 1 Shields Avenue, Davis, CA, 95616, USA
A R T I C L E I N F O
Article history: Accepted 18 May 2017 Keywords: Equine protozoal myeloencephalitis IFAT Neospora hughesi Sarcocystis neurona Toxoplasma gondii
A B S T R A C T
While toxoplasmosis is not commonly considered a clinical disease of equines, previous seroprevalence studies have reported differing background rates of Toxoplasma gondii infection in horses globally. The objective of this study was to evaluate possible associations between T. gondii seroprevalence and clinical signs of equine protozoal myeloencephalitis (EPM) in horses. Using a case–control study design, 720 Californian horses with neurologic signs compatible with EPM were compared to healthy, nonneurologic horses for the presence of T. gondii antibodies (using indirect fluorescent antibody tests [IFAT]). Toxoplasma gondii seroprevalence among cases and controls was determined at standard serum cut-offs: 40, 80, 160, 320, and 640. At a T. gondii titre cut-off of 320, horses with clinical signs compatible with EPM had 3.55 times the odds of a seropositive test compared to those without clinical signs (P < 0.01) when adjusted for covariates. When restricted to the autumn season and at the same titre cut-off, an EPM suspect horse had 6.4 times the odds of testing seropositive to T. gondii, compared to non-neurologic horses. The association between high T. gondii titres and clinical signs compatible with EPM is potentially reflective of toxoplasmosis in equines. Serologic testing of cerebrospinal fluid and isolation of T. gondii in EPM suspect cases should be considered. Future studies investigating the relationship between T. gondii and EPM are warranted. © 2017 Elsevier Ltd. All rights reserved.
Introduction Toxoplasma gondii, an obligate intracellular apicomplexan parasite, is an important pathogen of both humans and warmblooded animals. While toxoplasmosis is not commonly considered a disease of equines, previous seroprevalence assays have been conducted globally over the past decade to determine the potential risk of T. gondii in horsemeat destined for human consumption (Tenter et al., 2000; Pomares et al., 2011). These studies have produced differing background rates of T. gondii infection in horses, with seroprevalences ranging from >50% in Egypt (Ghazy et al., 2007; Shaapan and Ghazy, 2007) to 6% in the United States (Dubey et al., 2014). While there is no validated immunoassay test for detecting T. gondii antibodies in equines, several dilution cut-offs to determine positive tests have been used throughout the literature, ranging from 1:6 for the modified agglutination test (MAT; Paştiu
* Corresponding author. E-mail address: [email protected] (K.E. James). http://dx.doi.org/10.1016/j.tvjl.2017.05.008 1090-0233/© 2017 Elsevier Ltd. All rights reserved.
et al., 2015) to 1:64 for the indirect fluorescent antibody test (IFAT; Fonseca de Araujo Valenca et al., 2015). Riskfactors for clinical infections in equine protozoal myeloencephalitis (EPM), a neurologic disease caused by two other apicomplexan protozoal parasites (Sarcocystis neurona and Neospora hughesi), are currently unknown. A biological model that can reproducibly transition an asymptomatic infection to clinical disease has not yet been developed (Saville et al., 2000). Experimental studies to induce EPM in horses with severe combined immunodeficiency (SCID) did not progress to infections of the spinal cord with associated signs of EPM, although infection was confirmed by demonstrating the presence of antibody in visceral tissues, (Sellon et al., 2004). However, healthy, immunocompetent horses can develop clinical signs and tissue damage consistent with EPM (as confirmed post-mortem using immunohistochemistry). This seems to suggest that an immune-mediated inflammatory process may be responsible for parasitic migration to the site of clinicaldisease, i.e. the nervous system (Lewis et al., 2014). In a recent study of Brazilian horses with high antibody titres for T. gondii (IFAT titres 64–1024), seropositive horses had elevated
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
humoral and cellular immune responses compared to seronegative horses (Do Carmo et al., 2015). Considering a comparative species approach, polyparasitism in wild marine mammals has been shown to modulate the severity of protozoal encephalitis, most notably co-infections by S. neurona and T. gondii, (Gibson et al., 2011). To help elucidate a possible role for T. gondii in the progression of EPM, the present study used a case–control study design to test the hypothesis that horses with neurologic signs compatible with EPM would have higher antibody titres to T. gondii compared to horses with no neurologic signs. Materials and methods Case selection Cases were defined as horses with at least one of the following clinical signs compatible with EPM: ataxia, muscle atrophy, head tilt, hypermetria, lameness, weakness, circling, mentation change. Sera were tested at the Immunology Laboratory of the William R. Pritchard Veterinary Medical Teaching Hospital (VMTH), School of Veterinary Medicine, University of California at Davis, for the presence of antibodies against S. neurona and N. hughesi using Sarcofluor (Conrad Protozoal Laboratory) and Neofluor (Conrad Protozoal Laboratory) IFAT during 2013. Study inclusion criteria included Sarcofluor and Neofluor serum results, availability of serum to be tested for the presence of T. gondii; horses must also have been resident in California. Equine health records contained breed, sex, use, and age data, but did not include any information on prior treatment for EPM. There were 5592 diagnostic submission requests for Sarcofluor and/or Neofluor from 1 January, 2013 to 31 December, 2013; this was the study population from which cases were selected based on study inclusion criteria. Control horse selection Controls were resident horses of California with no current neurologic signs during 2013. Otherwise, this population was comparable to the cases in terms of residence, availability of data on covariates, and distribution of demographic factors. This was originally a convenience sample of non-neurologic horses for a separate study of leptospirosis. Participating clinics enrolled voluntarily and were asked to sample approximately 110 horses/ practice. A 10 mL blood specimen was drawn from the jugular vein, and a questionnaire was completed for each horse to collect data on age, breed, primary use, and sex. Information regarding prior treatment for EPM was not collected. Serum from 5250 horses was collected across 18 states in September and October 2013. This constituted the study population from which the study sample of controls was selected. The presence of antibodies against S. neurona and N. hughesi was determined as previously described (Sarcofluor and Neofluor IFAT; James et al., 2017).
39
testing for T. gondii in equines, each sample was end-titrated to generate quantitative titres of antibodies directed to T. gondii. Two investigators were blinded to case or control status during the reading of the IFAT slides and read the slides independently. Statistical analysis All statistical analyses and data management were performed using Stata Statistical Software: Release 14 (StataCorp). Demographic frequency tables for the case and control populations were created to compare populations. Demographic factors included breed, sex, age, and use. Breed was divided into Quarter horse, Warmblood, Thoroughbred, Paint horse, Arabian, Draft horse, Pony/miniature, and other. Age was analysed as a 5-year increment categorical variable. Sex was categorised into male (gelding and stallion) and female. Use of animal was as follows: competition animals, resident farm animals, breeding animals, or other uses. Clinical factors, including S. neurona and N. hughesi seropositivity, were evaluated for each population. The S. neurona seropositive titre cut-off was set at 40; for N. hughesi a seropositive titre cut-off was set at 160. Bivariate logistic regression models between the demographic and potential co-infection factors with the clinical outcome were generated to determine possible covariates for a multivariable logistic regression model. Odds ratios (ORs; 95% confidence intervals, 95%CI) with P < 0.05 were considered statistically significant and were included in the multivariable logistic regression model in a forward stepping manner. Confounding and first order interaction of predictor variables was assessed for the clinical outcome in preliminary multivariable logistic regression models (Agresti, 2002). Toxoplasma gondii seroprevalence was determined for case and control populations at multiple titre cut-offs. Due to the lack of a generally accepted IFAT titre cut-off for determining seropositivity in equines, seroprevalence at reciprocal titres of 40, 80, 160, 320, and 640 were determined. Unadjusted effect measures for the association of T. gondii exposure at these titre cut-offs and the clinical outcome were created (unadjusted ORs). A multivariable logistic regression model was then created with the statistically significant demographic explanatory variables and the T. gondii exposure at the titre cut-offs. A further refinement of the statistical analysis included matching cases to the same season as controls to increase matching of the two groups. As described, all controls were selected in September and October 2013. In a secondary analysis, only cases with serum submitted to the UC Davis Immunology Laboratory from August to November were included, to investigate the effect of the autumn season. The demographic frequencies, bivariate associations, T. gondii exposure prevalences, and final multivariable logistic regression models were generated as described above. Results
Toxoplasma gondii exposure
Primary analysis results: All year 2013 cases and controls
To determine exposure status to T. gondii,T. gondii antibody titres were determined for cases and controls using IFAT, as previously described (Conrad et al., 1993; Miller et al., 2001). The T. gondii strain used in testing was a type 2 strain (ME49) available from American Type Cell Culture research foundation (ATCC). Tachyzoites were grown on monkey kidney cells (MA104) in 10% fetal bovine media, incubated at 37 C. Slide fixation techniques included a 10 min fix with formalin, followed by a phosphate buffered saline wash. However, instead of anti-ferret IgG, antihorse IgG (Jackson ImmunoResearch Laboratories) conjugated to fluorescein isothiocyanate (FITC) was used as the secondary antibody, diluted 1:100. Due to the lack of validated diagnostic
Of 5592 horses from the case population, 392 resided in California. Cases were not equally distributed throughout the calendar year and the majority of cases occurred in April–July (n = 146); however, this is not statistically different from August to November, in which there were 138 cases (P > 0.05). December– March had the fewest cases, with 105 cases total. Of the 5250 horses from the control population, 328 resided in California. Important risk factors from bivariate modeling (P < 0.05) were age and breed (Table 1), and were therefore included in the multivariable logistic regression model. Clinical characteristics for S. neurona and N. hughesi seropositivity also differed between the case and control populations.
40
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
Table 1 Demographic characteristics of horses sampled from California in 2013 and bivariate associations with equine protozoal myeloencephalitis (EPM) outcome status. Explanatory variables
Case outcome status: Horses with neurological signs compatible with EPM (n = 392) n (%)
Control outcome status: Horses with no neurological signs (n = 328)
Unadjusted odds ratio (95% confidence intervals)
Demographic characteristics Sex Female Male Unknown
124 (31%) 255 (65%) 13 (4%)
118 (36%) 199 (60%) 11 (4%)
1.0 (Reference) 1.21 (0.88–1.68) –
Age 1–5 years old 6–10 years old 11–15 years old 16–20 years old Over 20 years Unknown
92 (24%) 109 (27%) 74 (19%) 47 (12%) 47 (12%) 23 (5%)
31 (9%) 96 (30%) 106 (32%) 52 (16%) 36 (11%) 7 (2%)
1.0 (Reference) 0.38b (0.23–0.63) 0.24b (0.14–0.38) 0.30b (0.17–0.53) 0.44b (0.24–0.80) –
Breed Quarter horse Warmblood Thoroughbred Paint Arabian Draft horse Pony Other Unknown
91 (23%) 68 (17%) 95 (24%) 20 (5%) 23 (6%) 12 (3%) 5 (1%) 65 (17%) 13 (3%)
57 (17%) 116 (35%) 29 (8%) 22 (6%) 34 (10%) 19 (6%) 13 (4%) 37 (11%) 1 (<1%)
1.0 (Reference) 0.32b (0.21–0.50) 1.79a (1.06–3.03) 0.50a (0.25–0.99) 0.37b (0.20–0.69) 0.35a (0.16–0.76) 0.21b (0.07–0.62) 0.96 (0.57–1.62) –
Use Competition Ranch/farm Breeding Other Unknown
138 (35%) 159 (40%) 7 (1%) 1 (<1%) 87 (22%)
184 (56%) 121 (37%) 11 (4%) 0 (0%) 12 (4%)
1.0 (Reference) 1.07 (0.79–1.45) 0.52 (0.19–1.37) – –
Clinical characteristics Sarcocystis neurona serum titre <40 40
302 (77%) 90 (23%)
66 (20%) 262 (80%)
1.0 (Reference) 0.07b (0.05–0.11)
Neospora hughesi serum titre <160 160
386 (98%) 6 (2%)
185 (56%) 143 (44%)
1.0 (Reference) 0.02b (0.01–0.05)
a b
P < 0.05. P < 0.01.
Seropositivity for S. neurona at a reciprocal titre of 40 was highly associated with control group status (bivariate OR, 0.07; 95%CI, 0.05–0.11); similarly, seropositivity for N. hughesi at a reciprocal titre of 160 was associated with control group status (bivariate OR, 0.02; 95%CI, 0.01–0.05). These were considered important risk factors to include in a multivariable logistic regression model (P < 0.05; Table 1). Toxoplasma gondii prevalences at the specified titres are presented in Table 2. Significant adjusted ORs (adjusted for age,
breed, S. neurona seropositivity, and N. hughesi seropositivity) were only seen at a T. gondii titre cut-off of 320; cases had 3.55 times the odds of a T. gondii titre of 320 compared to controls. While all T. gondii cut-off titres were analysed, the cut-off titre of 320 generated the highest likelihood of clinical signs in horses. The full results of the multivariable logistic model at a T. gondii cut-off titre of 320 are presented in Table 2 and in the Appendix: Supplementary Tables S1 and S2. A T. gondii titre of 320 (OR, 3.55; 95%CI, 1.41–9.0) was associated with suspect EPM clinical status.
Table 2 Prevalence of Toxoplasma gondii serum titres in case horses with neurological signs compatible with equine protozoal myeloencephalitis (n = 392) and control horses with no neurological signs (n = 328), and Toxoplasma gondii titre odds ratios (OR) associated with case status at various serum titre cut-offs. Toxoplasma gondii serum titre Case 40 80 160 320 640 a b
OR titre (95% confidence intervals)
Prevalence
241 (62%) 121 (31%) 78 (20%) 49 (13%) 18 (5%)
Control 239 (73%) 107 (33%) 54 (16%) 16 (5%) 5 (2%)
Unadjusted OR b
0.59 (0.43–0.83) 0.92 (0.66–1.28) 1.26 (0.84–1.89) 2.7b (1.52–5.35) 3.10b (1.10–10.8)
Adjusted for statistically significant covariates (P < 0.05): age, breed, Sarcocystis neurona serum titre 40, and Neospora hughesi serum titre 160. P < 0.01.
Adjusteda OR 0.89 (0.58–1.38) 1.52 (0.96–2.44) 1.77 (0.98–3.22) 3.55b (1.4-9.0) 4.12 (0.70–24.6)
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
41
Table 3 Seasonally restricted demographic characteristics and bivariate associations with equine protozoal myeloencephalitis (EPM) outcome status in horses sampled from California in August–November, 2013. Explanatory variables
Case outcome status: Horses with neurological signs compatible with EPM (n = 139) n (%)
Control outcome status: Horses with no neurological signs (n = 328)
Unadjusted odds ratio (95% CI)
Demographic characteristics Sex Female Male Unknown
45 (32%) 88 (63%) 6 (5%)
118 (36%) 199 (60%) 11 (4%)
1.0 (Reference) 1.15 (0.74–0.45) –
Age 1–5 years old 6–10 years old 11–15 years old 16–20 years old Over 20 years Unknown
34 (24%) 40 (29%) 23 (17%) 16 (11%) 19 (14%) 7 (5%)
31 (9%) 96 (30%) 106 (32%) 52 (16%) 36 (11%) 7 (2%)
1.0 (Reference) 0.38b (0.21–0.70) 0.20b (0.11–0.38) 0.28b (0.13–0.59) 0.48 (0.23–1.01) –
Breed Quarter horse Warmblood Thoroughbred Paint Arabian Draft horse Pony Other Unknown
29 (21%) 28 (20%) 28 (20%) 6 (4%) 12 (8%) 5 (3%) 1 (<1%) 24 (17%) 6 (4%)
57 (17%) 116 (35%) 29 (8%) 22 (6%) 34 (10%) 19 (6%) 13 (4%) 37 (11%) 1 (<1%)
1.0 (Reference) 0.39b (0.21–0.71) 1.57 (0.81–3.06) 0.44 (0.16–1.20) 0.57 (0.26–1.25) 0.43 (0.15–1.25) 0.12a (0.02–.99) 1.0 (0.54–2.05) –
Use Competition Ranch/farm Breeding Other Unknown
46 55 2 0 36
(33%) (39%) (1%) (0%) (25%)
184 (56%) 121 (37%) 11 (4%) 0 (0%) 12 (4%)
1.0 (Reference) 1.02 (0.68–1.52) 0.41 (0.08–1.88) – –
Clinical characteristics Sarcocystis neurona serum titre <40 40
104 (75%) 35 (25%)
66 (20%) 262 (80%)
1.0 (Reference) 0.08b (0.05–0.14)
Neospora hughesi serum titre <160 160
136 (98%) 3 (2%)
185 (56%) 143 (44%)
1.0 (Reference) 0.03b (0.01–0.09)
a b
P < 0.05. P < 0.01.
No first order interaction terms were statistically significant at P < 0.05, and confounding factors such as age and breed were adjusted for in the multivariable logistic regression model. Secondary analysis results: Seasonal restriction (August–November, 2013) Of the 392 cases included in this study, 139 met the seasonal criteria matching the collection season of the controls (August–
November, within one month of the September–October control selection period). The seasonal case and control population demographics were similar to the 1-year case and control populations, as presented in Table 3. Breed, age and S. neurona and N. hughesi seropositivity, were statistically significant in the bivariate analysis (P < 0.05), and were therefore included in the multivariable logistic regression analysis. Toxoplasma gondii seroprevalence at the titre cut-offs described in the primary analysis are presented in Table 4. When the association between T.
Table 4 Seasonally restricted prevalence of Toxoplasma.gondii serum titres in case horses with neurological signs compatible with equine protozoal myeloencephalitis (n = 392) and control horses with no neurological signs (n = 328), and T. gondii titre odds ratios (OR) associated with case status at various serum titre cut-offs. Toxoplasma gondii serum titre
40 80 160 320 640 a b c d
Prevalence
OR (95% confidence intervals)
Casea
Controla
Unadjusted OR
90 (65%) 52 (37%) 35 (25%) 24 (17%) 7 (5%)
239 (73%) 107 (33%) 54 (16%) 16 (5%) 5 (2%)
0.68 (0.43–1.08) 1.23 (0.80–1.90) 1.71c (1.02–2.83) 4.07d (1.98–8.48) 3.43c (0.91–14.0)
Adjustedb OR 1.20 (0.67–2.08) 2.15d (1.21–3.82) 2.79d (1.38–5.63) 6.40d (2.27–18.0) 6.53 (0.92–46.4)
Only horses sampled from California in August–November, 2013 were included in the analysis. Adjusted for statistically significant covariates (P < 0.05): age, breed, Sarcocystis neurona serum titre 40, and Neospora hughesi serum titre 160. P < 0.05. P < 0.01.
42
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
gondii and the outcome was adjusted for breed, age, and S. neurona and N. hughesi seropositivity, significant associations with the outcome were determined at T. gondii titre cut-offs of 80, 160, 320, and 640. At a titre cut-off of 320, cases had 6.40 times the odds of being T. gondii seropositive compared to controls. No first order interaction terms were statistically significant at P < 0.05, and confounding factors such as age and breed were adjusted for in the multivariable logistic regression model. Appendix: Supplementary material contains tables describing the risk factors associated with the multivariable logistic regression model of the highest statistically significant titre cut-off for T. gondii (320). The other T. gondii cut-offs showed similar patterns. Warmbloods had lower odds than Quarter horses for being cases (OR, 0.36; 95%CI, 0.20–0.95), while 11–15 year old horses had lower odds than the reference category of 1–5 year olds for being cases (OR, 0.31; 95%CI, 0.13–0.74). Sarcocystis neurona and N. hughesi seropositive horses had greater odds of being controls than cases. Toxoplasma gondii seropositivity was associated with case status; cases have 6.4 times the odds of controls of being seropositive at a titre cut-off of 320. Discussion In this study, horses with higher titres to T. gondii (80, 160, 320) were more likely to have clinical signs compatible with EPM than healthy, non-neurologic horses. While there was no evidence that co-infections by T. gondii and S. neurona/N. hughesi were required for clinical signs of EPM to develop, the association between T. gondii seropositivity and clinical EPM suggested that T. gondii was associated with neurologic signs in the study population of Californian horses. When clinical outcome was restricted to only those cases and controls collected in autumn, this association became even stronger, and horses with a T. gondii serum titre of 320 had >6 times the odds of clinical signs compatible with EPM as horses with T. gondii serum titres <320. The cases and control horses in this study were similar in sex and use distributions, and there were no statistically significant differences between the two potential clinical outcomes. However, there were differences in the breed and age distributions. Cases were more likely to be younger, aged 1–5 years, compared to controls. This could be due to the fact that 1–5 yearold horses are at greatest risk for EPM and could have clinical signs that would potentially require diagnostic testing (Saville et al., 2000). In terms of breed distribution, cases were also more likely to be Thoroughbred, while controls were more likely to be Warmbloods. This is in agreement with previous studies that reported Thoroughbreds and Quarter horses were more likely to have EPM than other breeds, while Warmbloods had notably lower rates of clinical disease (Fayer et al., 1990; Pusterla et al., 2014). Perhaps the most interesting difference between the case and control population was the seroprevalence of S. neurona and N. hughesi as determined by IFAT. While the titre cut-offs used in this study are used primarily for healthy horses (Duarte et al., 2004a,b), to allow comparison between the two populations, these cut-offs were also applied to horses with neurologic signs. While this may have overestimated the prevalence of clinically relevant infections of the two parasites, it determined background infection rates in the two populations. Controls were much more likely to be seropositive for both parasites at the defined cut-offs (40 for S. neurona, 160 for N. hughesi; Pusterla et al., 2014). This could be due to several factors, most notably a limitation in our study design, which used serodiagnostics only. Because the control group was older and contained were more Warmbloods, higher seroprevalence of S. neurona and N. hughesi were expected, as older age and Warmblood breed have previously been associated with higher
rates of seropositivity for these two parasites (Saville et al., 2000; James et al., 2017). When the inclusion criteria for cases was limited to the season that matched the controls, the association between T. gondii and the clinical outcome was stronger and exhibited more of a dose response (higher titre cut-offs were associated with greater odds of clinical outcome). In the late summer/autumn season (August– November), horses with T. gondii titres of 320 had 6.4 times the odds of being a case than a control. Since there was no evidence that co-infection played a role in the development of clinical signs, this association could indicate T. gondii-related neurologic disease. In horses, concurrent infections with seasonal viral and bacterial infections, and/or seasonal fluctuations in immune function, could precipitate the activation of Toxoplasma cysts, and could contribute to seropositivity due to increased anti-T. gondii antibody activity. In other species, there is evidence that increased infection rates in the later part of the calendar year could be due to peak periods of oocyst shedding in domestic kittens and cats during late summer and autumn, and the accumulation of oocysts in the environment during dry weather (Schares et al., 2008; Herrmann et al., 2010; Schares et al., 2016). As cats are the definitive host of T. gondii, environmental oocyst contamination increases with local cat population density, and the initial infection can lead to substantial oocyst shedding (Dabritz et al., 2007; Vanwormer et al., 2013a, 2016). Oocysts build up in the environment over dry weather, and are washed from land to sea during the initial rainfalls of the wet season, leading to seasonal loading of oocyst contamination during drier months. Cat reproduction cycles could also affect the environmental burden of viable oocysts. The seasonal load of T. gondii oocysts shed by cats could increase if dry seasons are prolonged, followed by late wet seasons, a trend that has been occurring in California. As such, T. gondii infection in horses could become more common in other parts of the year (VanWormer et al., 2013b; Schares et al., 2016; VanWormer et al., 2016). Limitations in data collection for this study did not allow for an investigation of whether season was important in case status, but future studies should aim to collect cases and controls throughout the year to determine any seasonality in equine infection. Although there is no evidence in this study that co-infections by T. gondii and S. neurona/N. hughesi were critical for the development of the neurologic signs of EPM, previous studies of the relationship between protozoal myeloencephalitis and polyparasitism in other species suggests that co-infections should not be ignored in the development of EPM (Gibson et al., 2011). The association between high T. gondii titres (>160) is potentially suggestive of toxoplasmosis in equines. Future studies of possible links between T. gondii infection and neurologic abnormalities in equines should include expansion of serologic testing to include cerebrospinal fluid samples, as well as isolating T. gondii from cases of EPM at necropsy using culture, immunohistochemistry, or PCR. Conclusions Further studies should be based on confirmed serum titre results using serum: CSF titre ratios (ante-mortem) and/or necropsy and direct identification of T. gondii (post-mortem), if possible, and should include cases where clinical signs and lesions are compatible with EPM but there is no evidence of S. neurona or N. hughesi infection. The strong association between T. gondii serum titre and signs compatible with EPM demonstrated in this study could be augmented by changes to climate and ecosystems which favour T. gondii oocyst shedding from the definitive host (domestic and wild felids) and enhance the viability of the oocyst in the environment. Although this study produced no evidence that co-infections by T. gondii and S. neurona/N. hughesi were associated with more severe signs of EPM, the association between
K.E. James et al. / The Veterinary Journal 224 (2017) 38–43
T. gondii serum titres and neurological abnormalities consistent with EPM should be examined further, especially considering the worldwide distribution of T. gondii. Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper. Acknowledgements This study was supported in full by the Center for Equine Health, University of California at Davis. We wish to thank Dr. Eva Tamez-Trevino and the Immunology Laboratory of the William R. Pritchard Veterinary Medical Teaching Hospital, School of Veterinary Medicine, University of California at Davis and Zoetis, Inc. for their assistance in sample collection. Appendix: Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.tvjl.2017.05.008. References Agresti, A., 2002. Building and applying logistic regression models, Categorical Data Analysis. Second Ed. John Wiley and Sons, Inc., Hoboken, NJ, USA, pp. 211–266. Conrad, P.A., Barr, B.C., Sverlow, K.W., Anderson, M., Daft, B., Kinde, H., Dubey, J.P., Munson, L., Ardans, A., 1993. In vitro isolation and characterization of a Neospora spp. from aborted bovine foetuses. Parasitology 106, 239–249. Dabritz, H.A., Miller, M.A., Atwill, E.R., Gardner, I.A., Leutenegger, C.M., Melli, A.C., Conrad, P.A., 2007. Detection of Toxoplasma gondii-like oocysts in cat feces and estimates of the environmental oocyst burden. Journal of the American Veterinary Medical Association 231, 1676–1684. Do Carmo, G.M., Da Silva, A.S., Klauck, V., Pazinato, R., Moura, A.B., Duarte, T., Duarte, M.M., Bochi, G.V., Moresco, R.N., Stefani, L.M., 2015. Immunological response and markers of cell damage in seropositive horses for Toxoplasma gondii. Comparative Immunology, Microbiology and Infectious Diseases 38, 9–13. Duarte, P.C., Conrad, P.A., Barr, B.C., Wilson, W.D., Ferraro, G.L., Packham, A.E., Carpenter, T.E., Gardner, I.A., 2004a. Risk of transplacental transmission of Sarcocystis neurona and Neospora hughesi in california horses. Journal of Parasitology 90, 1345–1351. Duarte, P.C., Daft, B.M., Conrad, P.A., Packham, A.E., Saville, W.J., MacKay, R.J., Barr, B. C., David Wilson, W., Ng, T., Reed, S.M., et al., 2004b. Evaluation and comparison of an indirect fluorescent antibody test for detection of antibodies to Sarcocystis neurona, using serum and cerebrospinal fluid of naturally and experimentally infected, and vaccinated horses. Journal of Parasitology 90, 379–386. Dubey, J.P., Ness, S.L., Kwok, O.C., Choudhary, S., Mittel, L.D., Divers, T.J., 2014. Seropositivity of Toxoplasma gondii in domestic donkeys (Equus asinus) and isolation of T. gondii from farm cats. Veterinary Parasitology 199, 18–23. Fayer, R., Mayhew, I.G., Baird, J.D., Dill, S.G., Foreman, J.H., Fox, J.C., Higgins, R.J., Reed, S.M., Ruoff, W.W., Sweeney, R.W., et al., 1990. Epidemiology of equine protozoal myeloencephalitis in North America based on histologically confirmed cases. A report. Journal of Veterinary Internal Medicine/American College of Veterinary Internal Medicine 4, 54–57. Fonseca de Araujo Valenca, S.R., Barreto Valenca, R.M., Pinheiro Junior, J.W., Feitosa de Albuquerque, P.P., Souza Neto, O.L., Mota, R.A., 2015. Risk factors of occurrence of Toxoplasma gondii among horses in the state of Alagoas Brazil.
43
Acta Parasitologica/Witold Stefanski Institute of Parasitology, Warszawa, Poland 60, 707–711. Ghazy, A.A., Shaapan, R.M., Abdel-Rahman, E.H., 2007. Comparative serological diagnosis of toxoplasmosis in horses using locally isolated Toxoplasma gondii. Veterinary Parasitology 145, 31–36. Gibson, A.K., Raverty, S., Lambourn, D.M., Huggins, J., Magargal, S.L., Grigg, M.E., 2011. Polyparasitism is associated with increased disease severity in Toxoplasma gondii-infected marine sentinel species. PLoS Neglected Tropical Diseases 5, e1142. Herrmann, D.C., Pantchev, N., Vrhovec, M.G., Barutzki, D., Wilking, H., Frohlich, A., Luder, C.G., Conraths, F.J., Schares, G., 2010. Atypical Toxoplasma gondii genotypes identified in oocysts shed by cats in Germany. International Journal for Parasitology 40, 285–292. James, K.E., Smith, W.A., Conrad, P.A., Packham, A.E., Guerrero, L., Ng, M., Pusterla, N., 2017. Seroprevalence of anti-Sarcocystis neurona and anti-Neospora hughesi antibodies among healthy equids in the United States. Journal of the American Veterinary Medical Association 250, 1291–1301. Lewis, S.R., Ellison, S.P., Dascanio, J.J., Lindsay, D.S., Gogal Jr., R.M., Werre, S.R., Surendran, N., Breen, M.E., Heid, B.M., Andrews, F.M., et al., 2014. Effects of experimental Sarcocystis neurona-induced infection on immunity in an equine model. Journal of Veterinary Medicine 2014, 239495. Miller, M.A., Sverlow, K., Crosbie, P.R., Barr, B.C., Lowenstine, L.J., Gulland, F.M., Packham, A., Conrad, P.A., 2001. Isolation and characterization of two parasitic protozoa from a Pacific harbor seal (Phoca vitulina richardsi) with meningoencephalomyelitis. The Journal of Parasitology 87, 816–822. , P., Rosenthal, B.M., Oltean, M., Villena, I., Paştiu, A.I., Györke, A., Kalmár, Z., Bolfa Spînu, M., Cozma, V., 2015. Toxoplasma gondii in horse meat intended for human consumption in Romania. Veterinary Parasitology 212, 393–395. Pomares, C., Ajzenberg, D., Bornard, L., Bernardin, G., Hasseine, L., Darde, M.L., Marty, P., 2011. Toxoplasmosis and horse meat France. Emerging Infectious Diseases 17, 1327–1328. Pusterla, N., Tamez-Trevino, E., White, A., VanGeem, J., Packham, A., Conrad, P.A., Kass, P., 2014. Comparison of prevalence factors in horses with and without seropositivity to Neospora hughesi and/or Sarcocystis neurona. The Veterinary Journal 200, 332–334. Saville, W.J., Reed, S.M., Morley, P.S., Granstrom, D.E., Kohn, C.W., Hinchcliff, K.W., Wittum, T.E., 2000. Analysis of risk factors for the development of equine protozoal myeloencephalitis in horses. Journal of the American Veterinary Medical Association 217, 1174–1180. Schares, G., Vrhovec, M.G., Pantchev, N., Herrmann, D.C., Conraths, F.J., 2008. Occurrence of Toxoplasma gondii and Hammondia hammondi oocysts in the faeces of cats from Germany and other European countries. Veterinary Parasitology 152, 34–45. Schares, G., Ziller, M., Herrmann, D.C., Globokar, M.V., Pantchev, N., Conraths, F.J., 2016. Seasonality in the proportions of domestic cats shedding Toxoplasma gondii or Hammondia hammondi oocysts is associated with climatic factors. International Journal for Parasitology 46, 263–273. Sellon, D.C., Knowles, D.P., Greiner, E.C., Long, M.T., Hines, M.T., Hochstatter, T., Tibary, A., Dame, J.B., 2004. Infection of immunodeficient horses with Sarcocystis neurona does not result in neurologic disease. Clinical and Diagnostic Laboratory Immunology 11, 1134–1139. Shaapan, R.M., Ghazy, A.A., 2007. Isolation of Toxoplasma gondii from horse meat in Egypt. Pakistan Journal of Biological Sciences 10, 174–177. Tenter, A.M., Heckeroth, A.R., Weiss, L.M., 2000. Toxoplasma gondii: from animals to humans. International Journal for Parasitology 30, 1217–1258. VanWormer, E., Carpenter, T.E., Singh, P., Shapiro, K., Wallender, W.W., Conrad, P.A., Largier, J.L., Maneta, M.P., Mazet, J.A., 2016. Coastal development and precipitation drive pathogen flow from land to sea: evidence from a Toxoplasma gondii and felid host system. Scientific Reports 6, 29252. Vanwormer, E., Conrad, P.A., Miller, M.A., Melli, A.C., Carpenter, T.E., Mazet, J.A., 2013a. Toxoplasma gondii, source to sea: higher contribution of domestic felids to terrestrial parasite loading despite lower infection prevalence. Ecohealth 10, 277–289. VanWormer, E., Fritz, H., Shapiro, K., Mazet, J.A., Conrad, P.A., 2013b. Molecules to modeling: Toxoplasma gondii oocysts at the human-animal-environment interface. Comparative Immunology, Microbiology and Infectious Diseases 36, 217–231.