Potential risk factors associated with seropositivity for Toxoplasma gondii among pregnant women and HIV infected individuals in Ethiopia: A systematic review and meta-analysis

Background Toxoplasma gondii is an obligate intracellular and neurotropic apicomplexan protozoan parasite infecting almost all warm-blooded vertebrates including humans. To date in Ethiopia, no systematic study has been investigated on the overall effects of potential risk factors associated with seropositivity for Toxoplasma gondii among pregnant women and HIV infected individuals. We intended to determine the potential risk factors (PRFs) associated with seropositivity for Toxoplasma gondii from published data among pregnant women and HIV infected individuals of Ethiopia. Methodology An systematic review of the previous reports was made. We searched PubMed, Science Direct, African Journals Online, and Google Scholar for studies with no restriction on the year of publication. All references were screened independently in duplicate and were included if they presented data on at least two risk factors. Meta-analysis using the random or fixed-effects model was made to calculate the overall effects for each exposure. Results Of the 216 records identified, twenty-four reports met our eligibility criteria, with a total of 6003 individuals (4356 pregnant women and 1647 HIV infected individuals). The pooled prevalences of anti-Toxoplasma gondii antibodies were found at 72.5% (95% CI: 58.7% - 83.1%) in pregnant women and 85.7% (95% CI: 76.3% - 91.8%) in HIV infected individuals. A significant overall effect of anti-Toxoplasma gondii seropositivity among pregnant women (p < 0.05) was witnessed with age, abortion history, contact with cats, cat ownership, having knowledge about toxoplasmosis, being a housewife and having unsafe water source. Age, cat ownership, and raw meat consumption were also shown a significant effect (p < 0.05) to anti-Toxoplasma gondii seropositivity among HIV infected individuals. Conclusions This review showed gaps and drawbacks in the earlier studies that are useful to keep in mind to design accurate investigations in the future. The pooled prevalence of anti-Toxoplasma gondii antibodies was found to be higher among pregnant women and HIV infected individuals. This suggests that thousands of immunocompromised individuals (pregnant women and HIV infected patients) are at risk of toxoplasmosis due to the sociocultural and living standards of the communities of Ethiopia. Appropriate preventive measures are needed to reduce the exposure to Toxoplasma gondii infection. Further studies to investigate important risk factors are recommended to support the development of more cost-effective preventive strategies.


Unfunded studies
Enter: The author(s) received no specific funding for this work.

INTRODUCTION
Toxoplasmosis is a food and waterborne protozan infection with high prevalence and distribution in all warmblooded vertebrates including humans caused by Toxoplasma gondii. It causes a considerable public health impact with higher burden in developing countries. Globally, it is estimated to infect approximately 30% of the human population [1]. Infection in healthy individuals is usually asymptomatic, but in immunodeficient patients, it can lead life-threatening encephalitis [2][3][4][5]. Further, infection of pregnant women may result in miscarriage or spontaneous abortion or congenital infection that may cause severe pathological effects [5,6]. It also results in personality shifts, epilepsy, bipolar disorder, suicide attempt, car accident and schizophrenia [7,8].
The seroprevalence of toxoplasmosis has been reported ranging from 1% to 100% in humans from various parts of the globe and also shows significant variations within a country or between communities [9]. In Africa, the infection rate ranges from 3% to 92.5% in women of reproductive age and pregnant women [2,6,10]. In Ethiopia, the infection rate estimates of 18.5-96.3% have been reported in the different risk groups of the population [2,11,12]. Various potential factors: sociodemographic and biological factors, lifestyle, climate, socioeconomic status and health care promotion have been documented to effect host-pathogen interaction [13][14][15]. Thus, control options for toxoplasmosis must rely on the vigorous evidence of the risk factors contributing to its circulation among hosts [16]. However, the relative effects of each noted factors have not been fully summarized, except their variations from area to area [13]. Therefore, this review was made to determine the seroprevalence and potential risk factors (PRFs) of Toxoplasma infection in Ethiopian pregnant women and HIV infected individuals.

Search strategy
Literature searching was performed following the PRISMA guideline (Sup 1 File) and its checklist (STROBE) (Sup2 File) [17,18] Moreover, unpublished thesis manuscripts were also accessed from Ethiopian Universities.

Inclusion criteria
We used the following inclusion criteria to confirm the eligibility of the searched papers: (1) original research articles and thesis; (2) cross-sectional studies that were reported: seroprevalence and risk factors; (3) studies with full texts; (4) targeted study population: pregnant women and HIV infected individuals; (5) studies with serological tests; (6) studies that provided the total sample size and the outcome of interest; and (7) studies published in English language. Each paper that did not meet the above-mentioned criteria was excluded.

Data extraction
Initially, articles were screened based on their titles and abstracts following the predefined inclusion criteria.
Then, articles that seem potential for eligibility were selected and downloaded for full text. The searched articles were reviewed and abstracted carefully by two independent reviewers to prove the eligibility. Disagreements between reviewers were settled by discussion. For each article, the following information were extracted: first author, publication year, study year, location, study design, sampling method, sample size, study subject, diagnostic test, potential risk factors (exposure), number of positive and negative samples. The study effect size, odds ratio and their corresponding confidence intervals were also calculated from the extracted data. Study searching strategies and exclusion criteria are presented in detail in Fig 1.

Meta-analysis
The meta-analysis was performed according to the protocol noted by Foroutan-Rad et al. [19] and Khalkhali et al. [20]. In brief, data on pregnant women and HIV infected individuals were analyzed separately. The seroprevalence and corresponding 95% confidence intervals (CIs) of toxoplasmosis were calculated for each study. The pooled seroprevalence estimate was computed according to Barendregt et al. [21]. Also, meta-analyses were made to determine the effects of each identified PRF if at least two included studies reported data on the same risk factor. OR and its respective 95% CI were calculated for each risk factor. We used pooled odds ratios (OR) as a measure of effect to assess the overall effects of each PRF. Forest plot was employed to present the outcomes of meta-analysis. Cochran's Q-statistics and inverse variance index (I 2 ) were computed to determine the heterogeneity and inconsistency among studies, respectively [2]. Galbraith plot was also constructed to assess the heterogeneities of study level estimates. We used random-effect model (if the p-value of the Q test was < 0.05 and I 2 was >50%) or fixed-effect model to pool the estimations in accordance with the heterogeneity result. Besides, we used random-effect model for risk factor assessment if we had five and above studies to accept the previous cut-off points [23]. Further, funnel plot, Egger's and Begg's tests [24] were performed to assess small study effects and publication bias. STATA software version 16 (StataCorp, College Station, TX, USA) was used for meta-analyses.

Search results and characteristics
We retrieved 216 reports for Toxoplasma infection in pregnant and HIV infected individuals. Of them, only 24 were found to be eligible for the inclusion criteria and data extraction (Fig 1). All selected studies were performed between 2007 and 2019 using a cross-sectional study design with convenient, systematic and simple random sampling procedures. Further, the studies were conducted using latex agglutination test(LAT), Enzyme-linked immunosorbent assay (ELISA) and Enzyme immune assay(EIA) to detect Toxoplasma antibodies (Table 1 and Table 2). The total sample size of the included studies was 6003 individuals (4356 pregnant women and 1647 HIV infected individuals). Of them, 4243 were found to be seropositive for Toxoplasma infection. The overall apparent infection rate in pregnant women and HIV infected individuals was 66% and 84%, respectively.

Meta-analysis and bias assessment
The random effect model with inverse variance procedures showed an overall pooled seroprevalence of 72  In our review, Funnel plot observations (Fig 5, 6

Potential risk factors with T. gondii infection in pregnant women
Nineteen potential risk factors (PRF) were identified and meta-analysis was made on thirteen of the fifteen included papers (Table 3). Of them, seven were shown a statistically significant effect on Toxoplasma infection (test for overall effect, p < 0.05) with higher odds of outcome: "history of abortion" (OR: 1.52), "age ≥ 35 years" (OR: 2.93), "contact with cat" (OR: 1.50), "cat ownership" (OR: 2.35), "knowledge about toxoplasmosis" (OR: 0.13), "being house wife" (OR: 1.58) and "unsafe water source" (OR: 1.55) as shown in Table 3 and Sup 3.  (Table 3). Forest plots, funnel plots and single weight of each publication contributing to the overall risk factors are presented in Sup 3 File.

Potential risk factors with T. gondii infection in HIV infected individuals
Thirteen potential risk factors were identified and meta-analysis was performed on seven of the included ten papers (Table 4). Of them, three were shown a statistically significant combined effect on Toxoplasma infection (test for overall effect, p < 0.05): "age ≥25 years" (OR: 3.087), "cat ownership" (OR: 4.34) and "raw meat and "being female" (OR: 1.16) ( Table 2). Besides, four risk factors: having "blood transfusion experience", having "knowledge about toxoplasmosis", "religion", and having "unsafe drinking water source" were not appeared to influence the odds of infection in HIV infected individuals (Table 4). Forest plots, funnel plots and single weight of each publication contributing to the overall risk factors are presented in Sup 4 File.

Potential risk factors analysis
We observed significantly higher odds of Toxoplasma infection in pregnant women and HIV infected individuals as the study subjects' age increases. This agrees with studies that reported higher Toxoplasma seroposetivity in an older groups than youngs [5,16,54,83,87]. Increased seroprevalence in older groups was also confirmed [9,15,52,64,[88][89][90]. Studies included in this review were also considered age as a potential factor [26,29,33,43] This might be explained by the assenting interaction between an increase in age, with a prolonged risk of exposure to Toxoplasma oocysts over time, with divers transmission route, lack of community awareness and a long-lived immune response [15,91]. This could also be due to that older groups could have a longer period of exposure to any of the risk factors [60,92].
Toxoplasma seropositivity has shown an association with history of abortion (p < 0.05). This is in accordance with the previous reports [60,93,94]. Being cat ownership, close contact with cats and abundance of cats are also considered to be important drivers for Toxoplasma infection [13,54,90]. Domestic cats are important sources of viable oocysts for environmental contamination [5,90]. Cats after having a single viable tissue cyst or bradyzoite can release millions of viable oocysts and augment the likelihood of pathogen transmission to the risk groups [19,83]. In this review, ownership of cats was found with significantly higher combined odds ratio in both study targets. Contact with cats was also found with significantly higher odds ratio in pregnant women. This suggests that ownership of cats or close contact with cats in couple with frequent exposure to cat faeces, cats' litter box management way or neglect of preventive measures (hand washing or wearing gloves) could increase the risk of infection to appreciable level [13,15,19,37,77,90].
Toxoplasma infection in high-risk groups such as pregnant women, HIV infected individuals and cancer patients has been reported to be influenced by educational status and knowledge about toxoplasmosis [54,90]. Sun et al. [95] and Liu et al. [77] also stated that individuals with no knowledge and formal education are more likely to acquire Toxoplasma infection. Other authors suggest that health education as cost-effective intervention [13,77,90]. This is supported by our findings in which literacy and health knowledge are considered as protective factors. Similar findings also reported by other authors [40,41,77,83]. This might be attributed to that those with low/no formal education status may have less hygienic practice and they are more likely to acquire Toxoplasma infection [54,77]. Lack of basic information about toxoplasmoses such as a source of infection, hygiene, raw meat/vegetable consumption, transmission route, and ignorance of the disease can maintain the risk of infection.
Previous reports have pointed out that Toxoplasma seroprevalence is varied among various working groups.
Certain working groups have more contact than others with Toxoplasma directly or indirectly. The present meta-analysis reveals that housewives are a more vulnerable group with significantly higher Toxoplasma seroprevalence. Our results are in line with those studies done in Saudi Arabia and the United States [72,96,97].
Habitually, housewives spend more time in taking care of pets, cooking and tasting food at home during meal preparation, handling and chopping meat without wearing gloves, cleaning and washing vegetables, and engaging in gardening, mainly in rural regions [15,83].
Published reports have documented that unhygenic drinking water is a considerable risk factor for Toxoplasma infection in human and animals [13,90]. In our review, unsafe drinking water has also shown a significant association with Toxoplasma infection in pregnant women but not in HIV infected individuals. This might be due to the contradicting findings of the included studies and wide confidence intervals. Hence, we considered the absence of tap water as an indicator of unsafe drinking water quality. However, this might not be necessarily true.
The microbiological quality of well or spring water could be satisfactory for the majority of the time unless contamination events happen by flooding and contaminated individuals/animals [98,99].
Higher odds of Toxoplasma infection were confirmed in pregnant women with risk factors like raw meat and vegetable consumption; though it was not significant. Similarly, raw meat consumption habit has shown significantly higher odds of Toxoplasma infection in HIV infected individuals (p = 0.01). This is consistent with reports across the globe [34,77,90,[100][101][102]. The consumption of raw meat with tissue cysts and vegetables contaminated with sporulated oocysts are considerable sources of infection [77]. The risks of acquiring Toxoplasma via contaminated meat with tissue cysts and vegetable with oocysts vary with cultural and feeding habits in various communities [13].
This systematic and meta-analysis has certain limitations, including: (1) all studies were used cross-sectional study design, this may bias our risk factor analysis, (2) all studies that exploring seroprevalence/ factors were drawn from limited samples of participants which may not represent the national seroprevalence, (3) lack of related risk factors evaluated by the majority of papers, (4) the number of eligible studies used for each risk factor analysis were small; so the estimates and the predictive values of the risk factors were assessed accordingly; it might decrease the power of meta-analysis, (5) limited or lack of studies in many regions of the country and (6) lack of a standard questionnaire to gather suitable data in the country. In

Ethical approval
None sought

Funding
This work not supported by any funding agency.

Competing interests
We declare that no conflict of interests.

Availability of data and materials
All data generated during this study are included as supplementary information.

PRISMA Flow Diagram used for Study Searching process
`````````````