Comparison of indirect and modified agglutination tests for detection of antibodies to Toxoplasma gondii in domestic cats

Abstract

Toxoplasma gondii is a protozoan parasite with worldwide distribution. The accurate detection of this zoonotic agent in cats and other hosts has public health importance. Blood samples from 89 domestic cats were tested for antibodies to T. gondii using 2 commercial agglutination test kits, an indirect (IHAT; Toxo-HAI FUMOUZE; Fumouze Diagnostics) and a modified (MAT; Toxoscreen DA; bioMérieux) agglutination test. Antibodies were found in 16 of 89 (18%) cats by the IHAT and in 23 of 89 (26%) cats by the MAT, with an overall agreement between the 2 serologic tests of 92% (κ = 0.77; i.e., substantial agreement beyond chance). Considering the MAT as the gold standard, the IHAT showed perfect relative specificity (100%) and lower relative sensitivity (70%). The suboptimal sensitivity of the IHAT limits its use in epidemiologic studies in cats.

Toxoplasma gondii, the etiologic agent of toxoplasmosis, is a well-studied parasite, given its medical, veterinary, and public heath importance.⁵ Cats (domestic and wild) are crucial in the epidemiology of toxoplasmosis because they are the only hosts that can excrete environmentally resistant oocysts in feces.⁶ Cats can excrete millions of oocysts in a very short period (2–3 d) after primary infection, but only 1% of cats in a population have been found to be excreting oocysts at any given time because cats that have excreted oocysts become seropositive and generally do not excrete oocysts again.⁶ Therefore, testing cat feces for oocysts for epidemiologic studies is not rewarding. Given that seropositive cats are likely to have already excreted oocysts, serologic surveys for the detection of antibodies to T. gondii in these animals are helpful to assess the degree of environmental contamination, as well as to determine the potential risk of infection in distinct geographic areas.¹⁵

Several serologic tests are available for the detection of antibodies to T. gondii, although most of them were originally developed to detect antibodies in human sera. One of these commercial tests, the modified agglutination test (MAT; Toxoscreen DA; bioMérieux, Marcy-l’Étoile, France), has undergone additional evaluation for the serologic detection of T. gondii in naturally and experimentally infected cats.^1,6-8,10 All evidence indicates that none of the feline pathogens cross-react with T. gondii. Antibodies to T. gondii have not been detected in a 1:10 dilution of sera in more than 1,000 cats raised in captivity; although some of them were infected with Cystoisospora felis (JP Dubey, unpublished observations). The MAT is simple to perform, the antigen is stable at 4°C for many months, and it does not need species-specific reagents. However, MAT kits are expensive and not available in many countries. Another commercial test (indirect hemagglutination test [IHAT]; Toxo-Hai FUMOUZE; Fumouze Diagnostics, Levallois-Perret, France) has been used to detect T. gondii antibodies in livestock.^2,20 A comparison of 6 commercial agglutination tests found good correlation using human sera¹⁹; however, there is no validation of the IHAT kit with feline sera. Previously, another IHAT kit (TPM-Test; Wampole Laboratories, Cranbury, NJ) was found insensitive to detect antibodies to T. gondii in experimentally infected cats.¹⁰

We compared the results obtained by the IHAT and MAT for detection of T. gondii antibodies in sera of naturally exposed domestic cats (Felis catus). We evaluated the agreement between the 2 serologic tests and the relative sensitivity and specificity of the IHAT.

Between August 2016 and February 2017, a convenience set of blood samples from 79 domestic cats aged 1 mo to 18 y were collected from 13 of the 18 districts of mainland Portugal. Blood was sent to a private veterinary laboratory by veterinary medical centers for serologic detection of T. gondii infection. All of the samples were collected for diagnostic purposes; no ethical approval was required according to Portuguese legislation for the protection of animals (Law 92/1995 and Decree-Law 113/2013). Blood samples were centrifuged at 1,500 × g for 10 min, with sera separated and stored at 4°C until further analysis (2–3 d). Control serum samples (n = 10; 5 positive and 5 negative) from the United States were also tested. The positive controls were from experimentally infected cats. The negative controls included 2 cats sampled before experimental infection; 2 other cats infected with Hammondia hammondi, a parasite most closely related to T. gondii, but without serologic cross-reactivity in the definitive host, the cat; and one additional seronegative cat.¹²

All serum samples were first tested for immunoglobulin (Ig)M and IgG antibodies to T. gondii using a commercial IHAT (Toxo-Hai FUMOUZE) according to the manufacturer’s instructions. Briefly, serum samples were added to a U-bottom microtiter plate with phosphate-buffered saline and a suspension of ovine red blood cells sensitized with T. gondii antigens. We tested serum samples at 2 dilutions, 1:80 and 1:160. Well contents were homogenized by lateral tapping on the edges of the plate for 2 min. Then, the plate was incubated for 2 h at room temperature (18–25°C), under protection from dehydration and vibration. Positive and negative controls were incorporated in each plate test, as well as reagent and serum controls. The test was considered positive when a layer of agglutinated erythrocytes covered >50% of the well’s diameter, and negative when a compact button or a ring covered <50% of the well’s diameter. The results obtained with the IHAT were expressed as an antibody titer (i.e., the reciprocal of the highest dilution at which agglutination was still visible after 2 h incubation at room temperature). A cutoff titer of 80 (8 IU/mL in relation to a WHO international reference serum) was chosen to maximize both sensitivity and specificity of the test, according to the manufacturer’s instructions.

The MAT, which detects IgG antibodies, was performed using a commercial kit (Toxoscreen DA; bioMérieux), according to the manufacturer’s instructions, with sera added to a U-bottom microtiter plate and diluted at 1:20 and 1:40. Positive and negative controls supplied with the kit were included in each test plate. The well contents were homogenized using the vibrator, covered with a self-adhesive sheet, and then incubated for 18 h at room temperature (18–25°C). A positive reaction exhibits agglutination of the tachyzoites covering ≥50% of the well’s diameter. In a negative reaction, a compact button or a ring covered <50% of the well’s diameter. The results obtained with the MAT were expressed as an antibody titer (i.e., the reciprocal of the highest dilution at which agglutination was still visible after 5–18 h incubation at room temperature). A cutoff titer of 20 (2 IU/mL in relation to a WHO international reference serum) was chosen to maximize both sensitivity and specificity of the test.^10,15

In order to compare the 2 serologic tests, the MAT was considered the gold standard. Agreement beyond chance between the results from the 2 tests was calculated using the Cohen kappa coefficient (κ). Values of κ were interpreted as follows: 0.0–0.20 = slight agreement; 0.21–0.40 = fair agreement; 0.41–0.60 = moderate agreement; 0.61–0.80 = substantial agreement; 0.81–1.0 = almost perfect agreement. Sensitivity and specificity values of the IHAT were also assessed. Discordant results between the MAT and IHAT were defined as false-negatives or false-positives by IHAT. Statistical analyses were conducted using WinEpi (http://www.winepi.net/) with 95% confidence intervals (CIs).

In accordance with the established cutoff value, 16 of the 89 domestic cats (18%; CI: 10–26%) were seropositive to T. gondii by the IHAT: 1 had a titer of 80, and 15 had a titer ≥160. In contrast, 23 of the 89 cats (26%; CI: 17–35%) were seropositive by the MAT: 2 had a titer of 20, and 21 had a titer ≥40 (Table 1).

Table 1.

Comparison of antibody titers to Toxoplasma gondii obtained by the IHAT and MAT in sera from domestic cats (n = 89).

IHAT titer	MAT titer
	<20	20	≥40	Total
<80	66	2	5	73
80	0	0	1	1
≥160	0	0	15	15
Total	66	2	21	89

IHAT = indirect agglutination test; MAT = modified agglutination test.

The global agreement between the 2 tests was 92% (82 of 89). Of the 73 negative results by the IHAT, 7 were positive by the MAT (i.e., there were 7 of 23 [30%] false-negative results by the IHAT). On the contrary, none (0%) of the results was interpreted as false-positive by IHAT, given that all of the results positive by IHAT were also positive by MAT (Table 2).

Table 2.

Distribution of domestic cats (n = 89) according to the results obtained by the IHAT and MAT for Toxoplasma gondii antibodies in juvenile (1–11 mo) and adult cats (1–18 y).

Age group	IHAT (cutoff titer: 80)	MAT (cutoff titer: 20)
		Positive	Negative
Juvenile	Positive	7	0
	Negative	4	25
Adult	Positive	8	0
	Negative	2	30
Unknown	Positive	2	0
	Negative	0	11

IHAT = indirect agglutination test; MAT = modified agglutination test.

Considering the MAT as the gold standard test, the 2 serologic tests had substantial (κ = 0.77) agreement beyond chance in detecting antibodies to T. gondii in domestic cats. The relative sensitivity and relative specificity of the IHAT were 70% (CI: 51–88%) and 100%, respectively.

Seropositivity by MAT (26%) was higher than by IHAT (18%), with an observed agreement between the 2 serologic tests of 92%. The 7 discordant results between the 2 methods were negative by IHAT and positive by MAT, which resulted in 30% false-negative results by IHAT. Of these 7 cats, 2 had an antibody titer of 20 by MAT, and the remaining 5 had a titer ≥40. The occurrence of false-negative results by IHAT may be influenced by several factors, such as the analytical sensitivity of the tests, different cutoffs, or different types of antigens. Whole tachyzoites are used in the MAT, whereas soluble antigens are used in the IHAT. In a recent infection, antibodies may not yet be detected by the test (e.g., cats experimentally infected with T. gondii usually develop antibody titers detectable by IHAT later than by MAT).¹⁰ Similarly, in subclinical infection, the antibody titer may decrease to undetectable levels.¹⁴ In these cases, the use of a high cutoff value makes it difficult to detect seropositive animals.¹⁸ Additionally, some authors have suggested that the failure of the IHAT to consistently detect IgM antibodies specific to T. gondii in feline serum is also a result of the antigens or their concentration used to coat the red blood cells used in the test.¹⁴

Considering the MAT as the gold standard, the IHAT showed optimal relative specificity (100%) and lower relative sensitivity (70%). Our results were consistent with a previous study in experimentally infected cats, which also suggested that the IHAT was less sensitive for detection of T. gondii antibodies than the MAT,¹⁰ but with tests that were different from the commercial ones used in our report. Comparison of the diagnostic ability of the IHAT and MAT in naturally infected domestic animals has focused on sheep, pigs, and chickens.^3,4,11,17 Reports in those species showed that IHAT is a test with lower sensitivity, but with high specificity in the detection of the infection, as in our study.

The Cohen kappa coefficient (κ) remains the most commonly used measure, mainly because it provides a measure of agreement beyond what would be expected by chance. In our study, the κ value showed that the 2 serologic tests had a substantial (κ = 0.77) agreement beyond chance in detecting antibodies to T. gondii in domestic cats.

In T. gondii infections, the accuracy of different methods is usually assessed by comparing their results to those from either bioassays in cats or mice, or other serologic tests of expected high sensitivity and specificity.¹⁷ Among the many serologic tests available for the detection of T. gondii antibodies, the MAT has shown high sensitivity in several host species, is simple to perform, does not require special equipment or species-specific reagents, its antigen is stable for months, and reagents are available commercially.^9-11,13,16