Toxoplasma gondii seroprevalence among pregnant women in Africa: A systematic review and meta-analysis

Background Toxoplasmosis is a serious endemic zoonotic disease caused by the protozoan parasite Toxoplasma gondii. Toxoplasma infection during pregnancy can result in congenital transmission and serious fetal and neonatal complications. This systematic review and meta-analysis aimed to assess the pooled seroprevalence of T. gondii infection and its determinants among pregnant women in African countries. Methods All articles reporting the seroprevalence of toxoplasmosis among pregnant women in African countries and published from 2010 to 2023 were searched using various databases. The pooled prevalence of toxoplasmosis was calculated using a random-effect model. The variation between the included studies was assessed using a funnel plot and I2 heterogeneity statistics. To identify the sources of heterogeneity, sub-group analysis was further conducted by country, diagnostic method, and sub-African region. The association of prevalence rates with the socio-economic level and geoclimatic parameters was also explored. Results In total, 29,383 pregnant women from 60 articles were included for analysis. The pooled T. gondii seroprevalence was 42.89% with high heterogeneity (I2 = 99.4%, P < 0.001). Sub-group analysis revealed variation by country (ranging from 2.62% in Namibia to 80.28% in Congo), diagnostic method used (from 8.66% in studies using a rapid diagnostic test to 55.69% in those using an agglutination test), and sub-African region (from 4.14% in regions of Southern Africa to 53.96 in Central Africa). Cat ownership (OR = 1.58) and the consumption of raw meat (OR = 1.50) and raw vegetables (OR = 1.48) had a statistically significant combined effect on T. gondii seroprevalence. No association was found between T. gondii prevalence and the level of income of the country or geoclimatic parameters. Conclusion The prevalence of toxoplasmosis infection among pregnant women in Africa is high, particularly in Central and Eastern Africa. The determinants of prevalence are multifactorial. Therefore, efforts should be made to increase the awareness of women concerning the risk factors for toxoplasmosis.


Introduction
Toxoplasmosis is an endemic zoonotic disease caused by Toxoplasma gondii, a protozoan parasite with a worldwide distribution [1,2].Cats and other felids are the definitive hosts of the parasite and humans, as well as virtually all warm-blooded animals, are intermediate hosts.They are contaminated by the ingestion of food or water containing oocysts spread in the environment through cat feces [2].Intermediate hosts undergo a systemic infection that can be controlled by an efficient Th1 immune response, but not eradicated, leading to the lifelong persistence of encysted parasites, mainly in the muscles, brain, and eyes.Therefore, the infection can also be acquired through the ingestion of contaminated meat if eaten undercooked [2][3][4].When primary infection occurs in a pregnant women, T. gondii can be transmitted through the placental barrier, resulting in congenital infection of varying severity, depending on the stage of gestation, or loss of the fetus [2,5].
An estimated one-third of the world population is infected with T. gondii, which is responsible for 1.2 million disability-adjusted life years (DALYs) each year [6].The seroprevalence of toxoplasmosis among pregnant women varies greatly between geographical areas as a result of combined factors, including the climate, socio-economic level and standard of living, quality of the drinking water, and proximity with cats or other felids.There is a trend towards a decreasing prevalence in European countries, such as France [7], likely related to modifications in dietary habits or food hygiene or handling.In Africa, a previous meta-analysis estimated that more than half of African pregnant women are chronically infected with Toxoplasma [8] but the selection criteria and inclusion period were not clear.Several studies from African countries reported high variability in T. gondii seroprevalence among pregnant women, ranging from below 5 to 80% [9,10].In a recent Egyptian review, Abbas et al. [11] reported a seroprevalence ranging from 4 to 71% among pregnant women, depending on the geographical location.Interestingly, it was as high as 27 to 100% among women with a history of recent abortion, suggesting that the burden of toxoplasmosis could be higher than suspected.In addition, epidemiological studies have shown that 95% of cats are seropositive for T. gondii, suggesting massive contamination of the environment with oocysts.
Although there are limited country-and sub-continent-level data about T. gondii seroprevalence among pregnant women in Africa, there have been no studies to assess its predictors across this continent.Knowing the seroprevalence and predictors of T. gondii infection could help health authorities to have a better understanding of the level of risk according to local factors and design targeted intervention measures to reduce fetal sequelae related to the infection.This systematic review and meta-analysis aimed to assess the seroprevalence of Toxoplasma gondii infection and its predictors among pregnant women in Africa.

Study protocol and eligibility
This systematic review and meta-analysis is reported using the preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) reporting checklist.
All cross-sectional and cohort observational studies conducted in various countries of Africa on the prevalence of toxoplasmosis among pregnant women published in the English language from January 1, 2010, to March 30, 2023, were included.

Sources of information and search strategies
We searched through various databases to retrieve pertinent findings about the prevalence of toxoplasmosis infection among pregnant women (PubMed, Researchgate, Africa Journal Online (AJOL), and Europe PMC using the following Mesh terms ("Toxoplasmosis" OR "Toxoplasma gondii"), AND ("Pregnant Women" OR "Prenatal care"), AND ("Africa") or combining with every individual African country (S1 File).

Study selection, data extraction, and quality assurance
Duplicated articles were removed using Endnote X8 citation manager.Initially, all retrieved articles were screened against the eligibility criteria via careful reading of the article title and abstract using the CoCoPoP (Condition, Context, and Population) framework.The condition (Co) was the prevalence of T. gondii and was the main outcome of interest, the study population (Pop) was pregnant women, and the context (Co) was a study conducted in all African countries.Articles fulfilling the inclusion criteria were included and reviewed.The eligibility of each study was assessed by two independent investigators.The inter-rater agreement between the two reviewers was rated using the "kappa index" tool [12] and the result showed almost perfect agreement between them (observed agreement (Po) = 0.94, kappa cofficient (k) = 0.832) (S2 File).Discrepancies between the two investigators were resolved by consensus after the paper was reviewed by a third reviewer.
The reviewers independently collected data from all eligible publications using a standardized form.The following data were captured and recorded using Microsoft Excel 2016: author, year of publication, country, study subject, study design, sample population, number of cases (participants with an outcome), diagnostic tool/method, and prevalence.The quality of the included studies and the risk of bias were critically evaluated using the revised Joanna Briggs Institute (JBI) eight-item critical appraisal tool to assess the quality of observational studies (analytical cross-sectional studies) [13] (S3 File).The quality ranking of each full extracted article was scored as low (� 3 of 8), medium (between 3 and 6 of 8), or high (� 6 of 8).To reduce the risk of bias, we only included articles with higher quality assessment scores.

Strategy of data analysis
Data were imported to Stata version 15.1 from a Microsoft Excel sheet for analysis.The pooled prevalence of toxoplasmosis, with the 95% confidence interval (95%CI), was determined using a random effect model.Heterogeneity between the included studies was examined using I 2 statistics.Heterogeneity is considered when the p-value of I 2 is < 0.05.I 2 statistics results of 25, 50, and 75% represent low, moderate, and high heterogeneity, respectively.An Egger regression asymmetry test for meta-analysis was used to assess publication bias.An Egger test pvalue < 0.05 shows the presence of publication bias.Subgroup analysis and meta-regression were performed to verify the sources of heterogeneity.Moreover, sensitivity analysis was carried out to ensure the robustness of the result and to determine the influence of a single study.To assess the influence of a single study, the pooled prevalence was calculated leaving one study out at a time.
Data on potential predictor variables such as residence, cat ownership, the consumption of raw meat or raw vegetables, the drinking of raw milk, contact with soil, a history of blood transfusion, and HIV status were recorded and analyzed.

Results
In total, 189 articles were initially identified from various electronic databases, of which 74 were removed due to duplication.After screening using the article title and abstract, 82 articles were selected for full retrieval.Twenty-two were further excluded due to the inability to retrieve the full text of the article, the use of a language other than English, or an inadequate target population, outcome, or diagnostic method.Finally, 60 articles were included for analysis (Fig 1 ).

Characteristics of included studies
In total, 60 full document articles published between January 1, 2010, and March 30, 2023, from 19 different African countries were included in this systematic review and meta-analysis [10,.Approximately half (48.4%) of the included studies were from three countries: Ethiopia (14 studies), Nigeria (9 studies), and Tanzania (6 studies).In terms of the sub-regions of Africa, most of the included articles were from East Africa (28 studies) or West Africa (17 studies).In total, 29,383 pregnant women were included.An average of 490 pregnant women were included in each study, ranging from 90 in a Nigerian study [64] to 5,578 in one from Morocco [21].Most studies (87.1%) used a cross-sectional study design.The presence of anti-Toxoplasma gondii antibodies was assessed by enzyme-immunoassay (EIA) (48 studies), agglutination tests (AT) (10 studies), or rapid diagnostic tests (RDTs) (2 studies) (Table 1).

Publication bias
The funnel plot and bias coefficient (Egger's test; b = 35.2%,95%CI: 9.6-60.9,P = 0.008) for studies published on T. gondii seroprevalence among pregnant women in Africa confirmed the presence of publication bias and a small-study effect.The results of trim analysis also revealed the presence of publication bias (Q = 14.162 on 61 degrees of freedom (p = 1.000) and Moment-based estimate of between-study variance = 0.000), with a pooled estimate of 3.299, 95%CI: 2.826-3.773,p-value < 0.001) (Fig 3).

Sub-group analysis
Sub-group analysis was conducted using the following variables: country, sub-African region, publication period, serological method, and type of study design.Sub-group analysis by country showed the highest mean seroprevalence to be observed in Congo (80.28%), followed by 64.28% in Ethiopia, whereas the lowest was 2.62% in Namibia and 5.84% in Zambia (Table 2).The results of country-level I 2 statistics showed the presence of heterogeneity between studies of the same country, except for the studies included in Sudan (I 2 = 65.1%,P = 0.057).On the other hand, the sub-group analysis by Sub-African region showed the highest prevalence of

Sensitivity analysis
We further investigated the sources of heterogeneity in the pooled T. gondii seroprevalence in Africa and ensured the robustness of the data by performing a sensitivity analysis.The removal of each study one by one and subsequent calculation of the pooled prevalence showed the seroprevalence of toxoplasmosis among pregnant women to vary between 42 [9].Based on this analysis, we concluded that our estimation of the pooled prevalence was robust and showed no evidence in favor of a single-study influence (S4 File).
Of the 19 countries with documented seroprevalence among pregnant women from 2010 onwards, six had a seroprevalendce of < 25%, five of 25% to 45%, and eight of > 45% (Table 3 and Fig 4).Half of the countries with a Toxoplasma prevalence > 45% were low-income countries (Table 3).However, 4/6 countries with a seroprevalence < 25%, were also low-income countries, showing that poverty cannot solely explain the risk of Toxoplasma exposure.
We found no association between Toxoplasma seroprevalence and geographical features, such as the type of climate, average elevation, or annual rainfall observed in the countries (Table 4).

Factors associated with T. gondii seroprevalence
The meta-analysis was performed on 38 of the 60 included articles with eight identified potential predictor variables.The results show that pregnant women who own a cat (OR = 1.584,  ) were more likely to be seropositive for Toxoplasma, with an overall effect, P < 0.05 (Table 5).

Discussion
Africa is a continent with a multicultural community that has highly varying living conditions and feeding habits, a high prevalence of communicable diseases, including HIV/AIDS, and most often a low and traditional agriculture-dependent economy level, which can favor the seroprevalence of toxoplasmosis.This systematic review and meta-analysis is the first to assess the predictive factors for T. gondii chronic infection among pregnant women in Africa.We did not report the pooled prevalence of acute infection because of the unavailability of clearly stated data (IgM or IgG avidity results) in most of the included studies.Nonetheless, the estimation of the prevalence of chronic infection is a widely recongnized indicator of parasite circulation in a given population.We observed a 42% pooled seroprevalence of T. gondii infection.The I 2 statistics results showed the presence of heterogeneity and Egger's test and trim analysis revealed the presence of publication bias.However, based on the sensitivity analysis, the value for the pooled seroprevalence is robust and high, suggesting a potentially high impact on pregnancy outcomes due to the possible transmission of infection to the fetus.A high level of congenital toxoplasmosis has already been reported for Northern Africa [74][75] but it is still widely unrecognized in sub-Saharan Africa.Recent epidemiological studies reported recombinant and atypical Toxoplasma genotypes in severe cases of congenital toxoplasmosis in Tunisia [76][77], as well as in breeding animals [78], showing the circulation of virulent genotypes.Little is known about the distribution of circulating genotypes in sub-Saharan Africa and their virulence.To date, a virulent genotype, Africa-1, has been identified in a few immunocompetent travelers returning to France from Senegal, the Ivory Cost, and the Republic of Congo presenting with fever, myalgia, and neurological signs [79].These observations suggest that severe congenital cases are likely to occur in sub-Saharan Africa.Knowledge acquired from South America showed that virulent atypical strains can be responsible for fetal loss at any stage of pregnancy and for severe sequelae [80].In Africa, the presence of virulent genotypes may also be associated with a high burden of disease.In the absence of documentation, these cases are probably not diagnosed.A major study by Mercier et al. [81] showed that nearly half of Toxoplasma strains isolated from domestic animals in Gabon were identified as Africa-1 or Africa-3 genotypes, both highly virulent in mice, with a lethal outcome in > 90% of animals.Their phylogenetic analysis concluded that African and American strains share a common ancestor.Thus, given the high observed pooled seroprevalence, it would be of great interest to better evaluate the burden of toxoplasmosis in humans in Africa.Similar seroprevalence estimates have been reported in systematic reviews conducted on pregnant women in Iran (43%) [82], the general population in Iran (39.3%) [83], and pregnant women who had a history of abortion worldwide (43%) [84].A similar pooled seroprevalence estimate of toxoplasmosis has also been reported in a systematic review and meta-analysis conducted on HIV-infected pregnant women worldwide (45.7%) [85].Noteworthy, in this latter study, 10 out of the 14 included studies originated from African countries, which could explain the very similar prevalence, as that observed in our study on non-immunocompromised women.Indeed, the immune background, particularly HIV infection in not known to influence the seroprevalence rate [86].Our finding is higher than the overall worldwide prevalence of toxoplasmosis among pregnant women (33.8%, 95%CI: 31.5% -35.9%) [87].In particular, it is higher than prevalence rates reported in Asia (28.5% in Saudi Arabia [88], 29.9% in Sri Lanka [89], 10% in Japan [90]) and Europe (31% in France [7] and 9% in Norway [91]).Several factors, such as the climate, standard of living, quality of the drinking water, cooking habits of meat, and density of felids, could account for these differences in seroprevalence.Few data are available regarding the prevalence of toxoplasmosis in felids in Africa, and even fewer in wild felids, compared to Europe or America.However, two recent meta-analyses [92,93] reported that the highest T. gondii pooled seroprevalences in domestic cats worldwide were observed in Australia (52% and 66.6%, respectively) and in Africa (51% and 55.7%, respectively).Interestingly, the highest rate of seroprevalence in wild felids was observed in lions (87.6%) [93].Besides, the highest rate of Toxoplasma oocysts detection in feces from domestic cats (9.8%) was reported from African studies [93].Although these data need to be consolidated, they indicate that the high prevalence in African felids could support the overall high seroprevalence observed in pregnant women on this continent.
We observed that the Central and East African regions had the highest pooled T. gondii seroprevalence, whereas the lowest was reported in the South African region.Moreover, the sub-group analysis conducted by country showed four countries from Central and East Africa (Congo, Gabon, Ethiopia, and Eritrea) to be among the top five countries with the highest pooled seroprevalence.The possible reason for the highest seroprevalence in Central and East Africa may be due to the frequency of contact with domestic or stray cats or other felids, and poor environmental hygiene [94].Geoclimatic conditions in Central Africa, which has a tropical warm humid climate and small temperature range, may also favor long-term survival of oocysts in the environment, as observed in South America, where a high seroprevalence has been documented [95].Some authors have also postulated that climate could affect the life cycle and population dynamics of prey species, thereby providing favorable conditions for cat reproduction and increasing numbers of young, naive definitive hosts susceptible to T. gondii infection [96].
The serological test used to diagnose T. gondii infection in the included studies was recorded, as this could introduce bias depending on its performance [97].Indeed, latex agglutination is known to yield a certain number of false positive results, whereas the sensitivity of certain RDTs can be lower than that of EIA, as already observed in field studies [98].However, the prevalence estimates obtained in Ethiopian studies using AT and EIA did not significantly differ (median seroprevalence of 80.55% and 69.60%, respectively, p = 0.74).Interestingly, the sole two studies using RDTs, performed in Nigeria and Zambia, showed the lowest prevalence rates (11.7% and 5.8%, respectively) [22,43] observed in central African countries.Both countries have a humid tropical climate and a low or lower-middle income.Toxoplasma seroprevalence was, thus, expected to be high.The study performed in Nigeria by Akubuilo et al. [22] contrasted with other Nigerian studies using EIA, as Dairo et al. [37], Nasir et al. [57], and Oboro et al. [52] reported a seroprevalence of 35.9%, 48.9%, and 65.6%, respectively.This supports the hypothesis of poor sensitivity of the RDT used by Akubuilo et al. [22].The RDT used in the study performed in Zambia [43] showed poor efficiency in a study comparing three RDTs for Toxoplasma diagnosis [99], which could also explain such an unexpectedly low seroprevalence.However, no other data are available from Zambia for other patient groups, and thus these data still need to be consolidated.
In this systematic review and meta-analysis, we confirmed that pregnant women owning a cat had higher odds of seropositivity for toxoplasmosis than those who did not, as observed in previous individual studies from various countries [100][101][102][103][104], including Ethiopia [105,106].Following primary infection, young cats can spread several hundred thousand oocysts per day for three weeks, and thus play a crucial role in T. gondii transmission through direct contact or the handling of litter.In addition, stray cats also massively contaminate the environment, water, and vegetables [1,2].Although experimental infection of cats has shown that re-infection with a same genotype strain did not lead to re-shedding of oocysts in feces, infection with other genotypes did.In Africa, cats are mostly free-ranging or feral animals, thus are likely to encounter multiple genotypes of T. gondii via hunting various infected preys.The same applies even more likely to wild felids, ensuring multiple re-infections and re-sheeding of oocysts in the environement [96].
Another important risk factor was the consumption of raw meat, which is currently practiced in Eastern Africa.A similar finding was reported in studies conducted in Saudi Arabia, the Philippines, and Brazil [107][108][109].Consistent with this finding, a systematic review and meta-analysis of case-controlled studies showed that the consumption of raw meat is a risk factor significantly associated with acute T. gondii infection in humans (OR = 3.44 (1.29-9.16)[110].In other studies, not only eating raw meat but also contact with raw meat was shown to be associated with a higher risk of acquiring Toxoplasma infection [106].
In accordance with the results of previous studies from other continents [104,109,[111][112][113], we found that the consumption of raw vegetables had a significant overall effect on Toxoplasma gondii seropositivity among pregnant women, who had a 47.6% higher risk of acquiring T.gondii infection.However, the risk associated with raw vegetable consumption was not unanimously recognized, as observed in Iran [82,104], as well as in African countries (9, 10, 14, 18, 19, 24, 26, 28, 31-33, 42, 47, 55, 59).This discrepancy could be due to the type of agriculture, the hygiene of the food market and food handling, the washing habits of vegetables and quality of the water supply, soil contamination by felids, and the density of cats.
This study confirms that the seroprevalence of toxoplasmosis in Africa is high.Recent epidemiological data from animals showed that the Africa-1 genotype predominates in Benin [114] and Gabon [81], meaning that similar strains are likely to infect humans and could be responsible for congenital infection.A retrospective study conducted in Benin on pregnant women estimated that the Toxoplasma seroconversion rate was 3.4% [16], which is more than 10-fold higher than that estimated among French pregnant women [7].Further studies should be undertaken in African countries to evaluate the rate of infected newborns, for example by serological screening at birth, as already carried out in Colombia [115].The incidence and burden of toxoplasmosis in African countries is currently unknown, whether in the setting of pregnancy or among HIV-infected patients, as biological confirmation is rarely obtained [116].A call for raising awareness concerning toxoplasmosis and fungal diseases was made several years ago [117], but actions are still pending to better diagnose toxoplasmosis and characterize its true burden.If a high incidence rate is confirmed, serological screening and counseling of all pregnant women for toxoplasmosis infection as a part of antenatal care would be crucial to minimize congenital transmission and its complications.
We were unable to identify new predictors for acquiring toxoplasmosis infection other than those already reported, i.e., contact with cats and the consumption of raw meat and unwashed vegetables.Therefore, efforts should be made towards prevention programs targeting women of child-bearing age, raising awareness of the clinical signs and possible consequences for the fetus, and focusing on hygienic prevention measures to avoid infection, such as the proper handling and cooking of meat, washing of vegetables and hands, and management of domestic cats [2].We believe that our study should pave the way for epidemiological studies on humans and encourage policymakers to fund prevention programs.

Table 1 . Characteristics of the 60 studies included in the systematic review and meta-analysis to estimate the seroprevalence of toxoplasmosis among pregnant women in Africa. First author, year Reference no Country Study design Sample size Number of seropositive women Prevalence (%) Diagnostic tool Quality score
toxoplasmosis to be in Central Africa (53.96%, 95%CI: 37.30-70.63)andthelowest in the South African region (4.14%, 95%CI: 0.98-7.29).As indicated by overlapping of the 95%CI, there was a significant difference between the South African region and other sub-African regions (Table2).