Seroprevalences of Toxoplasma gondii and Neospora caninum infections in Jordanian women who had a recent spontaneous abortion

This cross-sectional study investigates Toxoplasma gondii and Neospora caninum among 445 recently spontaneously aborted (RSA) Jordanian women using ELISA and indirect fluorescent antibody (at a cut-off value of 1/200) tests, respectively . The type of hospital, age, cat and dog contacts, raw and barbecued meat and wild plant consumption, number of abortions, and stillbirths were tested as independent variables using univariate and multivariate logistic regression analyses. The true seroprevalences were 22.1% for T. gondii- IgG, 22.7% for N. caninum -IgG, 2.6% for T . gondii -IgM, 10.6% for N. caninum -IgM, 0% for T. gondii -IgG and IgM, 6.7% for N. caninum -IgG and IgM, and 4.6% and 0% for both parasite IgG and IgM, respectively. T. gondii -IgM-seropositivity was associated with the number of abortions with odds ratios (OR) of 2.4 and eating barbecued meat (OR = 0.12). N. caninum -IgG-seropositivity was associated with having a dog in the house (OR = 2.6), and with stillbirth (OR = 0.1). N. caninum -IgM was associated with visiting a private-hospital (OR = 2.7). RSA Jordanian women are equally exposed to both parasites with significantly ( p < 0.05) higher seroprevalence of N. caninum -IgM compared to T. gondii- IgM suggestive of active infections among RSA women in Jordan.


Introduction
Toxoplasma gondii and Neospora spp. are closely related obligate intracellular protozoan parasites, belonging to the phylum Apicomplexa, family Sarcocystidae. There are two known Neospora species namely N. caninum and N. hughesi that cause neosporosis (Dubey et al., 2007).
T. gondii is a major parasite that infects all the warm-blooded animals including Man and domestic animals, causing toxoplasmosis (Montoya & Liesenfeld, 2004). Felines are the definitive and reservoir hosts for T. gondii, with more infection in feral than in domestic cats (Montoya & Liesenfeld, 2004). The minimum annual losses to the USA caused by T. gondii infection in man is estimated at $ 392 million (Roberts et al., 1994).
Lymphadenitis is the most common manifestation of human acute toxoplasmosis (Montoya & Liesenfeld, 2004). The acute phase is followed by a subclinical chronic phase. However, congenital fetal involvement, and abortion are the most important documented results (Torgerson & Mastroiacovo, 2013).
Seroprevalences of T. gondii-IgG and IgM in women have been reported from all over the world. The highest seroprevalence of anti-T. gondii-IgG was 84.3% in Mayotte Island (Julvez et al., 1994) while one of the lowest was 4.7% in India (Mohan et al., 2002). The seroprevalences of anti-T. gondii-IgM varied between 20.7% in repeatedly aborted females in Mexico (Galván Ramírez et al., 1995), and 0% among the Thai women (Sukthana, 1999). Globally, the reported seroprevalences of T. gondii-IgG and IgM among recently spontaneously aborted (RSA) women ranges between 0% to 46.1% and 0% to 18.4 (Nayeri et al., 2020) respectively. In contrast to T. gondii-IgG, T. gondii-IgM studies in the Middle East and in Jordan are very few (Qublan et al., 2002;Nimri et al., 2004).
In northern Jordan, the risk factors for seropositivity to T. gondii-IgG were studied in undergraduate female university students (Obaidat et al., 2015), pregnant (Jumaian, 2005;Nimri et al., 2004) and repeatedly aborted women (Abdel-Hafez et al., 1986;Qublan et al., 2002). Age, rural residence, consumption of undercooked meat and soil contact were found to be associated with seropositivity.
N. caninum is a well-known parasite of domestic and wild animals that utilizes a wide range of animals as intermediate hosts (Khan et al., 2020). Cattle, sheep, horses, goats, foxes, deer, buffaloes, and camels are its natural intermediate hosts and, in addition, cats, mice, rats, gerbils, and monkeys are experimental intermediate hosts, and many wild animal reservoirs (Gondim, 2006). Dogs and other canids have been identified as both intermediate and definitive hosts of this parasite (Gondim et al., 2004). Dogs are likely to become infected by the ingestion of contaminated tissue, including aborted fetuses, dead animals/birds, or placentas and shed oocysts in their faces, which can serve as a major source of infection for other species. In the dog, inflammatory lesions of skeletal muscles and the central nervous system predominate (Dubey et al., 2009). N. caninum is considered as a significant cause of abortion in ruminants, especially cattle Anderson et al. (1991). Similarly, experimental infection of pregnant monkeys with N. caninum resulted in transplacental transmission of N. caninum (Barr et al., 1994). Humans are exposed to two N. caninum infective stages; oocyst and tissue cyst. The oocysts are found in the contaminated environment, and the tissue cysts are located in food animal tissues (improperly cooked meat) and N. caninum-DNA was found in cow's milk, including colostrum (Milne et al., 2006;Moskwa et al., 2007). However, there is no conclusive evidence that lactogenic transmission of N. caninum occurs in nature (Dubey et al., 2007).
Recently, N. caninum DNA was detected in the umbilical blood of two women in Brazil (Duarte et al., 2020b) raising the concern about the zoonotic potential of N. caninum.
The seroprevalence of both anti-T. gondii and anti-N. caninum IgG and IgM antibodies in RSA women have not been reported in Jordan nor have the seroprevalences of those antibodies been reported anywhere in the World for N. caninum. This cross-sectional study compares the true seroprevalences of T. gondii and N. caninum specific IgG and IgM among RSA women in Jordan and investigates some risk factors for T. gondii and N. caninum seropositivity.

Sample size determination and sampling
As the expected seroprevalences had a wide range, a seroprevalence of 50% was used. As a 95% confidence level and a 5% absolute value precision are required according to Thrusfield (2007), the appropriate sample size is 384. During the period January to July 2003, a total of 445 blood samples were collected from RSA women, 7-10 days after abortion.

Hospitals and subjects
Three hospitals (2 public and 1 private) were employed. The two public hospitals were Al-Bashir, which is the largest public hospital in Jordan, located in Amman serving patients from central Jordan especially those of poor socio-economic standard, and Al Hussein Hospital, the fifth largest public hospital, located in Al-Salt City. Hiba is a private maternity hospital located in Amman serving mainly patients from the Christian community. The women were recruited from the outpatient clinics of the hospitals. A qualified nurse approached them and explained the objectives of the study. Of the women approached 48 did not want to have their blood drawn; in which cases other eligible women were approached and recruited.
The definition of a case of miscarriage was the spontaneous end of a pregnancy after the first 6 weeks and prior to 20 weeks of gestation. Pregnancy losses after 20 weeks are stillbirths.

Serology
Blood samples were collected, and sera were separated and stored at -20° C until testing at the end of the collection period.
Serum samples were tested for anti-T. gondii-IgG and IgM antibodies using Enzyme Immunoassay test kits (Biocheck, Inc, 837 Cowan Rd. Burlingame, CA 94010). According to the manufacturer, the IgG test sensitivity and specificity are 95% and 97.3% respectively and the IgM test sensitivity and specificity are 95.5% and 96.5% respectively.
Serum samples were tested for anti-N. caninum-IgG and IgM. Slides spotted with whole N. caninum (NC-1 strain) tachyzoites were purchased from VMRD, Inc., (Pullman, WA.). Cattle positive sera and anti-cattle sera labelled with fluorescein were used until positive human sera were found, after which the positive human sera were used. A cut-off titer of 1/200 was considered positive (Tranas et al., 1999;Dubey et al., 2009). The sample was considered positive when the fluorescent reactions involved all the periphery of the tachyzoites (Paré et al., 1998). The sensitivity and specificity of this test were 100% and 98% respectively when animal sera were used (Packham et al., 1998).

Data collection and ethical considerations
Consenting volunteer RSA women briefed on the objectives of the study, were interviewed using a questionnaire in Arabic which had been pilot tested. They were interviewed by a qualified nurse during the outpatient visit and at the time of blood collection. Information collected included name, type of hospital, age, place of residence, cat and dog in the house or garden/yard, feeding habits (eating raw meat, barbecued meat, or raw wild plants), 4/11 Seroprevalence of N. caninum and Toxoplasma gondii in women number of miscarriages and stillbirths, pregnancy age at the time of abortion. The study protocol was approved by the concerned committees at Jordan University of Science and Technology. All identifying information was kept confidential.

Data analysis
Data were stored in a database and analyzed using SPSS, version 10.0. The true seroprevalence and the 95% confidence interval (CI) were calculated according to Rogan & Gladen (1978). For both parasites true seroprevalence, the smallest sensitivity and specificity were employed. Univariate and multivariable analyses were performed. The dependent variable was parasite-Ig-seropositive status, coded as 0 (negative) or 1 (positive). A total of nine variables were tested. Screening of the significant variables to be used in the final logistic regression was conducted using univariable analysis. One variable at a time was tested for associations with N. caninum/T. gondii-seropositivity using the Chi-square test. For ordered categorical variables, the Chi-square test for trends was employed. Fishers exact test was used when the expected frequency was less than five. A multivariable analysis was then conducted starting with all factors that had a p ≤ 0.05 or an OR ≤ 0.3 or an OR ≥ 3.0 in the univariable analysis. Only those factors that remained in the final models were presented. A p-value of < 0.05 was considered statistically significant. Odds ratios and its 95% confidence intervals were calculated.

Results
RSA women were aged between 15 and 45 years with quartiles of: Q1 = 24 years, Q2 = 27 years and Q3 = 32 years. Table 1 summarizes the true seroprevalences results. The true seroprevalence of N. caninum-IgM was significantly (p < 0.05) higher than that of T. gondii-IgM. Figure 1 summarizes the number of N. caninum-IgG and IgM positive sera and their titers up to 1/1000. Tables 2 and 3 summarize the univariate analysis results. Two variables, eating barbecued meat and number of abortions were associated with T. gondii-IgM seropositivity with both univariate and logistic regression analyses ( Table 2 and 4) and none of the studied variables were associated with T. gondii-IgG seropositivity with both      univariate and logistic regression analyses. Three variables, namely, age group, dog contact and had a stillbirth were associated with N. caninum-IgG seropositivity with univariate analysis (Table 3). After stepwise selection, the final model had dog contact in the house and had a stillbirth (Table 4). Only one variable, the type of hospital, was associated with N. caninum-IgM-seropositivity with both univariate and logistic regression analyses (Table 3 and 4).

Discussion
The current results showed a significantly (p < 0.05) higher true seroprevalence of N. caninum-IgM compared to T. gondii-IgM but similar with the respect of IgG against both parasites. A portion of the sample had only N. caninum-IgM antibodies, a second had both IgM and IgG and a third had IgG only (Table 1), which is suggestive of active infections at different stages during pregnancy in RSA women.
In northern Jordan, the apparent seroprevalence of T. gondii-IgG is 57% in repeatedly aborted women (Abdel-Hafez et al., 1986) was similar to the 54% finding in high-risk women who had abnormal pregnancies (Nimri et al., 2004). However, the current results showed lower apparent seroprevalence among RSA women ( Table 1).
Analogous with the current results (Table 1), the seropositivity of T. gondii-IgM in pregnant women in northern Jordan was reported at 1.5% (2 of 132) (Qublan et al., 2002) and in Central Jordan at 2.7% (4 of 148) in high-risk women who had abnormal pregnancies (Nimri et al., 2004).
Most of previous N. caninum seroprevalence studies employed IFAT and reported the results at different titers, which impacts the validity of the seropositivity comparison. A cutoff titer of 1/50 in the IFAT was considered fitting to avoid cross-reactivity between N. caninum and T. gondii in animal studies (Lobato et al., 2006;Silva et al., 2007). A cutoff titer of 1/160 for IFAT was found to give the greatest sensitivity and specificity values for N. caninum (Packham et al., 1998) in cattle and a titer of 1/50 was recommended for dogs (Dubey et al., 2009). For this study, a cut-off titer of 1/200 was considered positive (Tranas et al., 1999).
By contrast, the search for anti-N. caninum-IgM took two attempts using IFAT. One yielded no seropositivity in normally delivered umbilical cord blood (Duarte et al., 2020b) and another in pregnant women reported 5.9% seropositivity at a 1/100 titer (Duarte et al., 2020a).
The reported high seroprevalences in Jordan is a reflection of the heavy environmental contamination with the infective stages, as well as the vertical transmission of the parasite in livestock (Dubey et al., 2007) which has been proved experimentally in nonhuman primates (Barr et al., 1994) but not yet in women.
However, European studies from Denmark (Petersen et al., 1999) and the UK were negative for the N. caninum test (Hemphill & Gottstein, 2000). The low N. caninum seropositivity in Europe may be explained by differences with other countries reporting higher seropositivity rates. For example, an association between N. caninum seropositivity and climate was found in livestock, in Europe (Julvez et al., 1994;Rinaldi et al., 2005). In Jordan, which encompasses six climatic zones and a high seroprevalence of Neospora spp., only the cool temperate rainy climatic zone was associated with reduced small ruminant flock-level seroprevalence of N. caninum (Abo-Shehada & Abu-Halaweh, 2010) and with horse titers <1/400 (Abu-Halaweh et al., 2020).
In addition, seroprevalences of parasites transmitted from dog to Man are influenced by locally adopted measures aimed at controlling stray dogs, reducing parasitic infections in owned dogs, and mitigating environmental contamination with the parasites. Such measures included, dog testing and treatment, the use of parasite free and cooked dog feed and proper canine waste disposal. Each of those measures play a role in mitigating the spread of infection in Man and animals and are reflected in the frequencies of parasites transmitted from dog to Man, including N. caninum.
Globally, most previous studies concentrated on consumption of raw/undercooked meat as an enhancing risk factor for T. gondii seropositivity. None studied properly cooked meat as an important food, representing a source of needed nutrients for the pregnant woman, and very few studies investigated risk factors for T. gondii-IgM seropositivity.
As consuming blood is prohibited in Islam, most Jordanians eat well cooked meat and only 6% (28 of 445) of the examined women ate raw or undercooked meat. This feeding habit is expected to reduce the transmission of the studied parasites through eating meat. In Jordan, eating barbecued meat reflects high socio-economic status and can be confounded with malnutrition and education level. On the other hand, well cooked meat provides nutrients needed by pregnant women and its consumption is associated with reduced T. gondii-IgM seropositivity. The association between nutritional status and resistance to protozoal parasites is well established and reported with regards to other protozoal infections. For example, malnutrition was reported as a risk factor for severe visceral leishmaniasis (Cerf et al., 1987). This finding emphasizes the need of pregnant women for good balanced nutrition to safeguard their own health and that of their unborn children.
To date, as far as is known, only three risk factors for N. caninum-IgG seropositivity in pregnant women were reported. These are: eating Senegalese barbecued meat (OR = 3.5) (Coulibaly et al., 2020) and the presence of domestic animals (OR = 2.3), and dogs (OR = 2.0) [11] (Duarte et al., 2020b).
Epidemiological evidence for the role of the dog in neosporosis, was found before the discovery of the dog as a definitive host for this parasite. This was an association between seropositivity of N. caninum in cattle and the presence of dogs (Bartels et al., 1999). As with N. caninum seropositivity in livestock hosts, e.g., cattle and small ruminants (Abo-Shehada & Abu-Halaweh, 2010), and the cord blood of women (Duarte et al., 2020b), N. caninum-IgG seropositivity in RSA women was associated with the presence of dogs in the house (OR = 2.6).
N. caninum-IgG seropositivity is associated negatively (OR= 0.1) with fetal loss during the second half of pregnancy. A significant inhibitory effect on the intracellular multiplication of the N. caninum in cells was demonstrated in vitro using sheep serum which had anti-N. caninum antibodies (Omata et al., 2005). N. caninum-IgM antibodies are detectable within two weeks of infection, while N. caninum-IgG requires several weeks, peaking six months later (Khan et al., 2020).
N. caninum-IgM seropositivity was associated with the type of hospital visited, as the women who used the private hospital had high odds of N. caninum seropositivity. The private hospital included in this work was utilized by Christian women, and women of a high socio-economic standard who could afford the private hospital costs and may have closer contact with pet dogs unlike those of the Muslim population in Jordan by whom dogs are perceived as unclean. Additionally, the low number of hospitals tested may have biased the results.

Conclusions
Jordanian RSA women are equally exposed to both N. caninum and T. gondii with a significantly (p < 0.05) higher true seroprevalence of N. caninum-IgM compared to that of T. gondii, which is suggestive of active infections. Despite sero-evidence of N. caninum-IgG and IgM and molecular evidence in the blood of two human umbilical cords, N. caninum has yet to be isolated from human tissue. The zoonotic potential and the role of N. caninum in human pregnancy needs to be investigated.
The current findings suggest that the aborted human placenta is a suitable tissue for isolating N. caninum from Man and bring the zoonosis verdict closer than ever. The causal relationship between N. caninum infection and spontaneous abortion can be explored by utilizing a large cohort study of pregnant women infected with N. caninum compared to controls. An avidity ELISA was developed to discriminate between acute and chronic antibody titers in individual cattle (Björkman et al., 1999). There is a need to develop a similar assay for use in humans.