Abstract
Objectives
The aims of the study were to estimate Toxoplasma gondii seroprevalence in pet cats in Norway and to evaluate risk factors for seropositivity. Additionally, serum biochemistry and haematological variables for T gondii seropositive and seronegative cats were compared.
Methods
A convenience sample of surplus sera submitted to the Central Laboratory, Norwegian University of Life Sciences, was collected. The samples were from healthy cats and cats with a variety of diseases. Analyses for IgG antibodies to T gondii were performed with a commercial direct agglutination test, with 1:40 as the threshold value. For risk factor analysis a logistic regression model of the relationship between predictors and the outcome was applied.
Results
One hundred and ninety-six of 478 cats were seropositive for T gondii, and the estimated seroprevalence in the study sample was 41.0% (95% confidence interval 36.6–45.4). Compared with domestic cats, pedigree cats had reduced risk for Toxoplasma seropositivity (odds ratio [OR] 0.42). Males had increased risk (OR 1.63) compared with females. The effect of age was highly significant, and an increase in the cats’ age across the interquartile range (IQR; 52–160 months/4–13 years of age) doubled the risk of Toxoplasma seropositivity (OR 2.11). The risk for Toxoplasma seropositivity among cats living in Oslo was significantly reduced (OR 0.51) when compared with the rest of Norway.
Conclusions and relevance
Pet cats in Norway appear to be commonly exposed to T gondii. Signalment and geographical region influenced the odds of Toxoplasma seropositivity, whereas health status did not.
Introduction
Toxoplasma gondii is a protozoan parasite with global distribution. Definitive hosts are domestic cats (Felis catus) and other felids, which shed environmentally resistant oocysts to the environment. A wide range of mammals and birds can serve as intermediate hosts, in which asexual parasite reproduction leads to formation of tissue cysts. Both intermediate and definitive hosts can be infected by ingestion of oocysts or tissue cysts, or via vertical transmission.1,2
Ingestion of tissue cysts is reported to be the route of infection in up to 97% of naive cats. 3 Most cats infected with T gondii have no noticeable clinical signs of infection.3–5 In feline clinical toxoplasmosis, clinical signs such as fever, anorexia, dehydration, depression, apathy, tachypnoea, dyspnoea, icterus, cutaneous nodules and various ocular manifestations may occur.3–8 A variety of serum biochemical and haematological abnormalities have been associated with feline clinical toxoplasmosis, but none is specific for the disease. 5 In some cases, feline clinical toxoplasmosis may be fatal, and is reportedly most severe in transplacentally infected kittens. Affected kittens may be stillborn, or they may die before weaning, usually of pulmonary or hepatic disease.4,5,9 Kittens with congenital infection frequently develop ocular toxoplasmosis with chorioretinitis or uveitis. 10 The clinical signs in older cats may result from spread of tachyzoites after primary infection or be due to reactivated latent infection following immunosuppression. 4
Toxoplasmosis is also an important public health issue and considered a major zoonosis. T gondii infection is usually asymptomatic in immunocompetent humans; however, fatal toxoplasmosis may occur in immunocompromised individuals or in cases of vertical transmission. 11
Environmental contamination by oocysts from cats is a crucial factor in the transmission of the parasite to meat producing animals and humans. The shedding of faecal oocysts is transient (1–2 weeks), whereas seroconversion ensues after active infection and shedding. Worldwide, T gondii has an estimated seroprevalence of 30–40% in domestic cats. 2 In Scandinavia, seroprevalences of 62% and 42% have been reported in Denmark and Sweden, respectively,12,13 whereas Kapperud found a seroprevalence of 24% among 87 Norwegian cats. 14 In a recent study from Finland, a seroprevalence of 48.4% was reported. 8
The aim of the present study was to estimate T gondii seroprevalence in pet cats in Norway and to evaluate risk factors for seropositivity. The hypothesis was that signalment, health status and living region influenced the odds of seropositivity.
Additionally, serum biochemistry and haematological variables for T gondii seropositive and seronegative cats were compared.
Materials and methods
Study sample
A convenience sample of surplus sera submitted to the Central Laboratory, Norwegian University of Life Sciences, Oslo, Norway, during the autumn of 2009 was collected. The samples were from healthy cats and from cats with a variety of diseases, and from all over the country. After biochemical analyses, surplus sera were stored at −70 ºC. Data collected from submission forms for the requested analysis included information about the cat’s breed, sex, reproductive status, living region and the indication (a synopsis of the clinical history and findings) for the analysis. Also, results from serum biochemical analyses, and any haematological analyses performed, were collected.
Serum biochemical and haematological analyses
Serum biochemical analyses were carried out using a Siemens Advia 1650 chemistry analyser, whereas haematological samples were assayed with a Siemens Advia 120 Hematology System (Siemens AG Healthcare Sector).
Serological testing for T gondii
With an expected seroprevalence of 40%, an accepted deviation of the true prevalence of 5% and a confidence level of 95% the sample size necessary to estimate the seroprevalence was calculated to be 369 (OpenEpi, http://www.openepi.com/v37/menu/OE_Menu.htm).
Surplus serum samples were thawed at room temperature and analysed for IgG antibodies to T gondii with a commercial test kit (Toxo-Screen DA; bioMerieux) following the manufacturer’s instructions. This test is a direct agglutination test with whole tachyzoites as antigen and treatment of sera with 2-mercaptoethanol to inactivate IgM antibodies. Sera were diluted to 1:40, which was set as threshold value, and samples with visible agglutination at this dilution were interpreted as positive, following the manufacturer’s instructions. All inconclusive samples, together with 22 randomly selected positive samples, were titrated by testing twofold dilutions from 1:40 to 1:1280. Results obtained with the test were expressed as antibody titre, that is the reciprocal of the highest dilution at which the agglutination was visible after 5–15 h of incubation at room temperature.
Statistical analysis
Stata 12 (Stata Corp) was used for all analyses. Descriptive statistics are reported for demographic and health status data, and serum biochemical and haematological variables.
Serum biochemistry and haematological variables in Toxoplasma seropositive and seronegative cats
Distribution of serum biochemistry and haematological variables were assessed graphically by histograms. Non-normally distributed data were analysed using Wilcoxon rank sum test and normally distributed data were analysed using Student’s t-test. A P value ⩽0.05 was considered significant.
Risk factor analysis
The dependent variable, T gondii seropositivity, was a dichotomous variable, and a logistic regression model of the relationship between predictors and the outcome was applied. Associations between the dependent variable and the predictor variables were first screened with univariable logistic regression. Categorical variables with too few observations were amalgamated when biological or logical new categories were possible to make. The linear relationship between the continuous variable age in months and the logit of the outcome was assessed by lowess curves.
The variables were tested for collinearity by Goodman and Kruskal’s gamma for ordinal and dichotomous variables and pair-wise correlations for continuous vari-ables. Associations >0.7 or <–0.7 were considered evidence of collinearity. Variables with a P value ⩽0.20, provided there was no collinearity between them, were then considered for further analysis in a multivariable logistic regression model to assess the relationship with Toxoplasma seropositivity. When collinearity was detected between two predictors, the predictor with fewest missing data was selected for further analyses.
The multivariable logistic regression model was constructed using forward selection. Predictor variables were retained in the model when the P value was <0.05. Potential confounding and intervening variables were considered after constructing a causal diagram. Changes of >20% in the coefficients in the model with the potential confounder present were used as additional indication of confounding. A variable was considered intervening if adding it removed the entire effect of another variable and if the intervening variable lay on the causal path between the factor and the outcome. Intervening variables were excluded from the final model. Interactions between biological plausible significant predictors were tested by adding an interaction term to the final model, and the interaction term was retained if P <0.01.
The multiple Wald test and the likelihood ratio test were used to evaluate differences between categories of categorical variables. The Stata command lincom was used to conduct contrasts among each category of categorical predictors.
Model evaluation
The final model was evaluated as described in the literature. 15
Results
Study sample
In total, 478 cats were included. The distribution of cats related to regions was as follows: ‘Østlandet’ (n = 355; 74.3%), ‘Vestlandet’ (n = 68; 14.2%), ‘Nord-Norge’ (n = 37; 7.7%), ‘Trøndelag’ (n = 8; 1.7%) and ‘Sørlandet’ (n = 10; 2.1%). The distribution of cats related to each of the 19 counties in Norway is reported in Table 1.
Distribution of serum samples, and Toxoplasma seronegative and seropositive serum samples in relation to geographical regions (n = 5) and Norwegian counties (n = 19)
Three hundred and eighty-four (80.3%) cats were domestic cats and 94 (19.7%) were pedigree cats. One hundred and eighty-three (38.3%) cats were females and 295 (61.7%) were males. For 12 (2.5%) cats reproductive status was not reported, whereas 229 (47.9%) cats were reportedly intact and 237 (49.6%) were neutered. One hundred and five (22.5%) female cats were intact and 74 (15.9%) were spayed, whereas 124 (26.6%) male cats were intact and 163 (35.0%) were castrated. Mean ± SD age was 108.3 ± 61.7 months (range 4.0–245.0 months). When divided into age categories (life stage classification 16 ), eight (1.7%) cats were kittens (0–6 months), 75 (15.7%) were juniors (7 months to 2 years), 75 (15.7%) were adults (3–6 years), 95 (19.9%) were mature (7–10 years), 118 (24.7%) were seniors (11–14 years) and 57 (11.9%) were geriatric (15 years+). Age was not reported for 50 (10.5%) cats.
Regarding health status, 41 (8.6%) cats were categorised as clinically healthy (eg, preanaesthetic survey, health control, including senior/geriatric health control), 112 (23.4%) as chronically ill (eg, illness of several weeks duration, significant weight loss) and 263 (55.0%) as acutely ill (eg, illness of days duration, up to a maximum of 2 weeks). None of the cats had a history of suspected toxoplasmosis. In 62 (13.0%) cats the health status was not possible to categorise owing to lack of information in the submission form.
Seroprevalence
One hundred and ninety-six cats were seropositive for T gondii and thus the estimated seroprevalence in the study sample was 41.0% (95% confidence interval [CI] 36.6–45.4). Of the 22 positive samples titrated by testing two-fold dilutions from 1:40 to 1:1280, 18 (81.8%) had a titre of 1280 (95% CI 59.7–94.8), one (4.5%) had a titre of 640 (95% CI 0.12–22.84), two (9.1%) had a titre of 320 (95% CI 1.12–29.16) and one (4.5%) had a titre of 160 (95% CI 0.12–22.84).
Serum biochemistry and haematological variables in Toxoplasma seropositive and seronegative cats
Descriptive statistics for biochemical and haematological variables are reported in Table 2. Regarding total protein and globulin, statistically significant differences between Toxoplasma seropositive and seronegative cats (P values of 0.02 and 0.005, respectively), with higher median values in positive cats, were found. The number of platelets was also significantly higher in seropositive cats (P = 0.004).
Serum biochemical and haematological variables reported as median values with ranges (minimum–maximum) in Toxoplasma seronegative and seropositive cats. Data were analysed using Wilcoxon rank sum test and a P value ⩽0.05 was considered significant
Reference intervals for adult cats from the Central Laboratory, Norwegian University of Life Sciences
AST = aspartate aminotransferase; ALT = alanine aminotransferase; ALP = alkaline phosphatase; CK = creatine kinase; Hgb = haemoglobin; WBC = white blood cells
Risk factor analysis
Living region was initially evaluated by county (n = 19) and subsequently by region (n = 5). Finally, the effect of living region was evaluated as Oslo County vs the rest of the Norwegian counties represented in the study sample. The latter subcategorisation of living region was performed to add enough power to the analyses to detect potential differences between urban and rural regions.
The unconditional odds ratio (OR), P value, and 95% CI for the variables breed, sex, reproductive status, age, living region and ‘health category’ are outlined in Table 3. After unconditional screening the variables breed, sex, age and living region were selected for model building. No collinearity was detected among selected variables.
Unconditional odds ratio (OR), P value and 95% confidence interval (CI) for the risk factors for Toxoplasma seropositivity in cats in Norway (n = 478)
Reproductive status was missing in 12 observations
Age was missing in 50 observations
All other counties represented in the study sample
Information on health status was missing in 62 observations
Results (OR, P value and 95% CI) from the final multivariable model are shown in Table 4. Compared with domestic cats, pedigree cats had reduced risk for Toxoplasma seropositivity (OR 0.42), and male cats had increased risk (OR 1.63) when compared with females. The effect of age was highly significant (Table 4), and an increase in the cats’ age across the interquartile range (52–160 months/4–13 years of age) doubled the risk of Toxoplasma seropositivity (OR 2.11). The risk for Toxoplasma seropositivity among cats living in Oslo was significantly reduced (OR 0.51) when compared with the rest of Norway. The variable reproductive status was added to the final model and tested for a potential confounding effect, but there were no changes in estimated effects of the other variables in the model. Hence, reproductive status was excluded. The following interactions were tested: breed × age, breed × living location and age × living location. None of the tested interactions were significant.
Results from the final multivariable logistic regression model estimating the effects of risk factors for Toxoplasma seropositivity in cats in Norway. The odds ratio (OR), P value and 95% confidence interval (CI) for the significant risk factors are presented. Information about age was missing in 50 cats, and total number of observations in this analysis was 428
All other counties represented in the study sample
Model evaluation
The final model showed reasonably good fit; the Hosmer–Lemeshow goodness of fit test was not significant (X2 = 6.34, 8 degrees of freedom; P = 0.61). Outlying observations with influence on the model were not found. The predictive ability of the model was moderate to low, area under the receiver operating characteristics curve was 0.69.
Discussion
In the present study, an estimated seroprevalence of 41.0% was found, indicating that the cats in the study sample were commonly exposed to T gondii. The reported seroprevalences against T gondii in domestic cats worldwide range from 4.8% to 97.4%.17,18 To the best of our knowledge, only one Toxoplasma seroprevalence study in Norwegian cats has previously been published. 14 Using the Sabin–Feldman dye test, Kapperud reported a prevalence of 24.1% (95% CI 15.6–34.5), which is significantly lower than the prevalence reported in the present study. Also, although tested in a rather small subsample, the majority of cats (81.8%) in the present study had high titers (1280). Similarly, in a recent study in Finnish cats, 75.0% of the samples had high titers (⩾4000). 8 As clinically healthy cats reportedly can have titres >10,000 as long as 6 years after being experimentally infected, a single high IgG titre does not necessarily indicate recent infection.19,20
Several factors were identified as risk factors for seropositivity in the present study. When domestic cats were compared with pedigree cats, the odds of Toxoplasma seropositivity was doubled and significantly higher in domestic cats. This is consistent with the results obtained in the studies by Voillare et al and Lopes et al,21,22 although the latter study failed to show a statistically significant difference. The most plausible explanation is differences related to the cat’s lifestyle with more access to outdoor environment, and thus hunting, among domestic cats. In a recent Swedish study, examining cat-breeding catteries, 70% kept their cats strictly indoors. 23 Although not all pedigree cat owners are breeders, a Norwegian survey also found that a higher proportion of pedigree cats were kept strictly indoors when compared with domestic cats (unpublished observation). Felines are very susceptible to infection with T gondii; a single bradyzoite from a tissue cyst can be sufficient to establish infection. 24 As they are much less susceptible to infection through indigestion of oocysts, feeding of raw or undercooked food and predation of intermediate hosts, like rodents and birds, are major risk factors for infection.
In the present study, the odds of seropostivity was significantly higher in male cats compared with female cats when controlling for breed, age and living region. There are conflicting results in the literature regarding the effect of sex on Toxoplasma seroprevalence,17,21,25–34 possibly related to differences in study design, study population, sample size and statistical method. Afonso et al found that feral male cats were more often infected with Toxoplasma than females, 35 and the authors suggested that because adult males are larger and heavier than females, they are expected to consume more prey items and to be more often exposed to infection. One could speculate whether this also could be valid for cats in our study sample. However, the effect of sex could also be a confounder for an unmeasured factor. A possible confounding effect of reproductive status, for example owing to more extensive roaming by intact cats, was not observed in the present study.
As reported in many previous studies, it was shown that increasing age increases the odds of seropositivity to T gondii.21,22,27,30,32–34,36–38 This is related to increased exposure time of adult cats to the infective forms of T gondii.
Comparing cats from Oslo County with cats from outside Oslo showed that cats from Oslo had halved odds of seropositivity, reflecting increased risk of acquiring toxoplasmosis in other potentially less urbanised parts of Norway. Differences in seroprevalences may be a reflection of the survivability of T gondii oocysts in the environment related to regional climatic conditions.21,39 However, the subcategorisation of living region in Oslo County vs ‘the rest of Norway’ was performed to add enough power to the analyses to detect potential differences between urban and rural regions, and the reported difference is most likely related to the cat’s lifestyle. 28 As reported by Clancy et al, 40 a Norwegian survey also found that owners living in urban areas were more likely to keep their cats strictly indoors (unpublished observation). Although outdoor access is unlikely to have a direct effect, it is strongly associated with hunting behaviour. In the study by Opsteegh et al, 38 the largest fraction of T gondii infections (35%) in the study sample could be attributed to hunting.
In a study from the USA, the seroprevalence in clinically ill cats with a suspicion of toxoplasmosis was 31.6%. 21 In the present study, odds of seropositivity between ill and healthy cats did not differ significantly. Toxoplasma seropositivity between acutely ill and chronically ill cats also did not differ. A clinical suspicion of toxoplasmosis was not raised in any of the cats included in the present study, but, owing to the relatively unspecific clinical signs of this disease, one cannot exclude the possibility that some of the ill cats had clinical toxoplasmosis. In cats with toxoplasmosis, a history of acute illness lasting about 1 week is reportedly common. 8
Although a variety of clinical pathological abnormalities has been associated with feline clinical toxoplasmosis, Jokelainen et al reported that in four cats with generalised toxoplasmosis the main biochemistry abnormality detected was markedly elevated alanine aminotransferase (ALT). 8 Also, in cats with positive T gondii IgM antibodies, ALT and alkaline phosphatase (ALP) activity was increased when compared with cats with positive T gondii IgG antibodies, or Bartonella species or haemoplasma infection, 41 whereas Akhtardanesh et al found no significant difference in biochemical and haematological findings between Toxoplasma seropositive and seronegative cats. 42 In the present study, median values for ALT, ALP, aspartate aminotransferase and total bilirubin in serum were higher in Toxoplasma seronegative cats, although a large range of values for most analytes was detected in both groups. The biological significance of the higher globulin levels and platelet numbers observed in the Toxoplasma seropositive cats is uncertain. Higher globulin levels could merely reflect concomitant illness and higher mean age (data not shown). Platelet clumping leading to decreased platelet numbers, is often observed in feline blood samples, even in healthy cats. 43 In the study by Spada et al any correlation between T gondii IgG seropositivity and haematological abnormalities was not detected. 44
Possible limitations of the present study should be discussed. Firstly, owing to the study design, only limited background information of the cats was available. Secondly, selection of a convenience sample might have introduced a bias, as the majority of cats were non-healthy. Including sera sampled at veterinary clinics may have limited inclusion of farm cats kept for rodent control, and thus the prevalence could be an underestimate because outdoor access and hunting behaviour could be less present in the study sample than in the study population. Furthermore, in spite of the Central Laboratory being a nationwide laboratory, large differences in the number of samples received from different regions were found; thus, a potential participation bias was introduced with regional differences influencing the overall estimation.
Conclusions
Pet cats in Norway appear to be commonly exposed to T gondii. Breed, sex, age and geographical region were identified as risk factors for seropositivity to T gondii in the study sample. In several previous studies, living indoors and/or no feeding with raw food is found to be protective. Prevention of T gondii infections in cats would decrease contamination of the environment with oocysts, which has implications for human health.
Footnotes
Acknowledgements
We would like to thank Biomedical Laboratory Scientist Elisabeth Bleken for her assistance with the samples, and Professor Hans Jørgen Larsen for providing suitable laboratory facilities and equipment.
Conflict of interest
The authors do not have any potential conflicts of interest to declare.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.