Toxoplasma gondii infection in domestic and wild felids as public health concerns: a systematic review and meta-analysis

Felidae as definitive hosts for Toxoplasma gondii play a major role in transmission to all warm-blooded animals trough oocysts dissemination. Therefore the current comprehensive study was performed to determine the global status of T. gondii infection in domestic and wild felids aiming to provide comprehensive data of interest for further intervention approaching the One Health perspective. Different databases were searched by utilizing particular key words for publications related to T. gondii infecting domestic and wild feline host species, worldwide, from 1970 to 2020. The review of 337 reports showed that the seroprevalence of T. gondii in domestic cats and wild felids was estimated in 37.5% (95% CI 34.7–40.3) (I2 = 98.3%, P < 0.001) and 64% (95% CI 60–67.9) (I2 = 88%, P < 0.0001), respectively. The global pooled prevalence of oocysts in the fecal examined specimens from domestic cats was estimated in 2.6% (95% CI 1.9–3.3) (I2 = 96.1%, P < 0.0001), and that in fecal samples from wild felids was estimated in 2.4% (95% CI 1.1–4.2) (I2 = 86.4%, P < 0.0001). In addition, from 13,252 examined soil samples in 14 reviewed studies, the pooled occurrence of T. gondii oocysts was determined in 16.2% (95% CI 7.66–27.03%). The observed high rates of anti-T. gondii antibodies seroprevalence levels and oocyst excretion frequency in the felids, along with soil (environmental) contamination with oocysts may constitute a potential threat to animal and public health, and data will result of interest in further prophylaxis programs.

www.nature.com/scientificreports/ to a serious Canadian outbreak 14 , but viable oocysts were observed in contents of the intestine of a wild trapped cougar (Felis concolor vancouverensis) and in a fecal pile in close proximity to the reservoir 15 . It is important to highlight that the sporulated oocysts are very resistant and can remain viable and infective for more than 1 year in favourable conditions 11,[16][17][18] . In this regard, a recent paper reviewed the environmental pathways by which T. gondii can infect animals and people mostly driven by water, soil or contaminated fresh produce or seafood 19 . Toxoplasma gondii antibodies have been largely found in cats worldwide, and the seroprevalence degree increases with the age of the cat, suggesting postnatal transmission of T. gondii 20 . It is assumed that postnatal sero-conversion in cats is linked to oocysts excretion episodes. The life style of cats influences the occurrence of T. gondii infections since feral cats that hunt for their food will present higher rates than domestic cats with limited access to parasites 21 . Seroprevalence level varied among continents, countries and even cities, linked to many possible environmental factors influencing these variations. As an example, in an urban population of 301 domestic cats in Lyon, France 20 , the anti-T. gondii seroprevalence was only 18.6%, approximately half the prevalence in other surveys in Europe 22,23 . The control of rodents in the area and feeding of cats by people were considered as protective factors limiting infections. On the other hand, a low income and poor sanitation were not the determining factors for low seropositivity to T. gondii in cats in Durango, Mexico 24 . Since a high density of felines (specially domestic cats) increases the risk of infection and T. gondii prevalence in intermediate hosts, a gradient of prevalence rate of infection has been demonstrated depending on the anthropization degree of the environment 25,26 .
Nearly up to 30% of the world's human population has had contact with the parasite evidenced by the presence of anti-T. gondii antibodies; while T. gondii infections are usually asymptomatic, they can lead to harmful effects, especially in congenital cases and immunocompromissed persons 27,28 . Humans become primarily infected mostly via oral ingestion of viable tissue cysts present in raw or undercooked meat and oocysts contaminating water or foodstuffs 6,8,29 . Nowadays, comprehensive local studies are still necessary to determine the source attribution of human infections; this constitutes an interesting challenge that should be approached under the One Health perspective.
To date, different surveys have been focused on domestic and wild felids in order to determine aspects as seroprevalence rates of anti-T. gondii antibodies, frequency of oocysts excretion and soil presence worldwide [30][31][32][33][34][35][36] , but with a certain degree of variance among studies. A systematic review recently assessed the seroprevalence of T. gondii in felids from 1967 to 2017 with a search strategy restricted to articles in English 37 . So that, the present investigation was aimed to determine the global frequency of T. gondii infections in domestic cats and wild felids, the occurrence of T. gondii-like oocysts shedding, and the frequency of oocysts in soil; such information will be useful to implement further measures aiming to reduce animal and human infections under a One Health perspective.

Methods
Search strategy. The review process exactly followed the protocol suggested by the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines (Supplementary data: PRISMA/ STROBE) 38 . We retrieved published studies from the databases of MEDLINE (via PubMed) (https:// www. ncbi. nlm. nih. gov/ pubmed/), Scopus (https:// www. scopus. com/), Web of Science (https:// www. webof nowl edge. com/) and CAB Abstracts (https:// www. cabi. org/ AHPC) with no restriction on language from Jan 1, 1970, to Dec 31, 2019. Search terms included a combination of Medical Subject Heading terms (MeSH) and free-text words in titles, abstracts and full texts.
The systematic search for PubMed accomplished using several Medical Subject Heading terms (Table S1). In addition, Scopus, Web of Science and CAB Abstracts were searched using the same strategy (Supplementary DATA). The Google Scholar search engine was used for checking the search strategy. The reference lists of all included articles and relevant reviews were hand searched for potentially eligible literature. In addition, authors and experts in the field were consulted to aid in the identification of relevant conference abstracts related to Toxoplasma and toxoplasmosis. Sometimes, we have had to contact the authors for raw data collection 39 , especially in old literature.

Selection of studies.
Initial screening by manuscript titles and abstracts was performed independently by two researchers (KHN and EA), that also assessed the full texts of all potentially relevant studies and applied inclusion criteria. Discrepancies when detected, were resolved after constructive discussion (AD, MZ and MTR).
The studies providing data on the seroprevalence of T. gondii in domestic or wild felids, frequency of oocyst excretion in felids, and those reporting soil contaminations with oocysts were included. On the other hand, studies meeting the epidemiology of T. gondii in non-feline hosts, studies where cat faecal samples were collected from the ground, and data from each animal was not independently retrievable, experimental studies, articles that only presented the final result and did not provide the raw data, or those without definite sample size, abstracts presented in congresses without full text, and case-control studies and clinical trials that could not report a correct estimate of prevalence were excluded. Any duplicated research was also excluded.
Quality assessment. The standard Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist was used 40 . In present study, articles were evaluated as low quality: less than 16.5, moderate quality: 16.6-25.5, and high quality: 25.6-34; the articles included in the meta-analysis presented acceptable quality.
Data extraction. After comprehensive examination of selected articles, the following data were extracted: the first author's last name, publication year, country, feline scientific names, keeping status of felids, sample size, www.nature.com/scientificreports/ source of soil samples, beginning and end date of study implementation, the number of the positive and negative cases, cut-off, age groups, as well as information about diagnostic tools. Data were extracted separately if two different populations had been studied. All extracted data from each study were entered into an Excel spreadsheet.  41 . In addition, potential publication bias was explored using Funnel plot and Egger's test.

Results
The initial database search retrieved 14,870 publications. First screening enabled us to exclude 14,441 studies not meeting the inclusion criteria. Altogether, 429 studies were retained for further investigation. In a secondary assessment, another 34 documents were excluded because of one of the following reasons: review articles, follow up studies or case series reports, and papers with insufficient data, or data from each animal were not independently retrievable. Eventually, 395 studies which met our eligibility criteria and evaluated soil contamination with T. gondii oocysts (n = 14), and T. gondii infection in domestic cats (serology, n = 268; oocyst in feces, n = 112) and wild felids (serology, n = 69; oocyst in feces, n = 15) during five decades were retained for analysis ( Fig. 1). Potential publication bias in the conducted studies regarding the prevalence of T. gondii infection in domestic cats and wild felids, Toxoplasma-like oocysts shedding, and frequency of oocysts in soil are shown using Funnel plot and Egger's test (Fig. 2).  Table 2). Heterogeneity was, however, very low I 2 = 79.9%. Retrieved information of the eligible studies based on the of anti-Toxoplasma antibodies prevalence in domestic cats and wild felids are summarized in Tables S2 and S3 (Tables S2, S3). Amongst the reviewed studies, just one investigation applied PCR method for detection of T. gondii DNA in stray cats using blood samples 42 (Table S2).  (Table S4, S5). The global pooled prevalence of oocysts in the fecal examined specimens from domestic cats was estimated in 2.6% (95% CI 1.9-3.3) (I 2 = 96.1%, P < 0.0001) ( Table 3). The highest and the lowest prevalence rates were detected in Africa (9.8%, 95% CI 2.4-21.5) (I 2 = 94.1%, P < 0.0001), and North America (0.9%, 95% CI 0.5-1.3), respectively. Heterogeneity was, however, very low I 2 = 50% (Tables 1, 3). The most used methodology for detection of oocysts was microscopy (99 studies) which was followed by molecular (16 studies) and mouse bioassay (10 studies) methods (Table S4). Some studies combined two techniques for detection of oocysts in feline feces. The highest prevalence was related to the molecular detection method (6.5%, 95% CI 3.7-10) (I 2 = 92.1%, P < 0.0001), followed by bioassay (2.8%, 95% CI 0.6-6.4) (I 2 = 95.1%, P < 0.0001), and microscopy (2.1%, 95% CI 1.4-2.8) (I 2 = 96.3%, P < 0.0001) ( Table 3).

Discussion
Felids as final host play an irreplaceable role for T. gondii life cycle that exclusively yield and excrete oocysts in their faeces, contaminating soil, water and food [6][7][8]10 . According to our findings, 37.5% of domestic cats showed exposure to T. gondii and 2.6% were actively shedding T. gondii or T. gondii-like oocysts. Similarly, the worldwide seroprevalence of Toxoplasma in domestic cats had been previously estimated at levels of 30-40% 1,43 . Based on the results, the highest value of seroprevalence in domestic cats was observed in Australia, followed by Africa, Europe, Central/South America, North America and Asia; although the number of studies for Australia (n = 6) and Africa (n = 15) were relatively low, whereas only one study in USA included 12,628 animals. The lowest prevalence was observed in Asia (28.3%), nevertheless, most studies have been conducted in Asian countries (91 studies). The number of surveys was higher in USA (North America; 34 studies), followed by Brazil (South America; 30 studies). Our investigation identified a number of countries without any data on T. gondii infection in cats, emphasizing the need for further studies in this field. According to our findings, 64% of nondomestic cats showed evidence of exposure to T. gondii and 2.4% were actively shedding T. gondii oocysts. Accordingly, the highest sero-prevalence of T. gondii in different wild felids were in the following order: lion www.nature.com/scientificreports/ colocolo). Although in general it should be considered that the number of studied nondomestic cats were lower compared to domestic cats, and keeping status of wilds cats, in captive (n = 2,492) or free ranging (n = 2,949), is probably associated with how they are fed and how they become infected. Albeit such findings also highlights the importance of serological and isolation studies on T. gondii infecting their prey (ungulates, birds, etc.) using validated methodologies for bias reducing.  www.nature.com/scientificreports/ The considerable gap between the prevalence of oocysts in feces and positive serum antibodies can be explained by the fact that infected felids can shed T. gondii oocysts for a short period (10-15 days), shortly after primoinfection, and then they become seropositive indefinitely. As, one important point, the activation of humoral immunity and antibodies production prevent from re-shedding of oocysts; new excretion episodes can occur when severe immunosuppression appears 29,44 . The short period of oocyst shedding and the low prevalence rate of felids which actively excrete oocysts, have led some authors to discuss that direct contact with felids should not be considered as a risk factor for human infection 7,29,45 . A systematic review in Iran showed that humans with history of close contact with cats presented a higher T. gondii seroprevalence rate compared to those without contact 46 . In the study conducted by Jones and colleagues 29 in the USA, exposure to kittens was statistically linked to T. gondii infection. In the study conducted in different European centers 45 , infections in pregnant women were attributed to the consumption of undercooked or cured meat products and soil contact in the 30-63% and 6-17% of cases respectively, but contact with cats was not identified as a risk factor. Similarly, another study showed that contact with cats is not related to infections, while the ingestion of raw or undercooked meat highly increased the risk of infection 47 .  48 . Contaminated soil has been demonstrated as an important source for infection for humans and animals 13,19 . It has been shown that gardening and occupations in contact with soil increases the risk of T. gondii infection 49 , as previously seen 50 . In a follow-up study of the toxoplasmosis outbreak during 1977 in Georgia 51 , after 25 years, among 37 individual (exposure to an indoor horse arena), 14 equestrian were tested, that three (21%) were found to have toxoplasmic retinochoroiditis lesions. Based on the observations is possible that cat feces containing the  www.nature.com/scientificreports/ organism were most likely stirred up when horses ran on the dirt floor, and were inhaled or ingested by riders and observers. Based upon number of studies conducted in different European centers, contact with soil or vegetables or fruit presumably contaminated with soil were highly associated to T. gondii infection in pregnant women 45,[52][53][54] . Investigation on sentinels (i.e., molluscs) for environmental contamination 55 and also the infection source attribution by using specific tests 56 will be of great interest for integration with data compilation in definitive and intermediate susceptible hosts.
In the present investigation, the prevalence of soil contamination was highly variable in the selected studies, which might be influenced by the soil characteristics and the number of infected animals in the area 57 . The included studies also reported highly heterogeneous results regarding the prevalence of cat infections, which could be due to the different risk factors, to note: sex, age, climates, study periods, cat breeds, living conditions and diagnostic methods as well as other unrecognized confounding factors. Based on our results, T. gondii seroprevalence in cats (Felis domesticus) in different countries oscillated from less than 10% in Thailand, Taiwan and Angola to more than 70% in Qatar, and Ethiopia. This can partly be explained by the different environmental conditions among the countries 58 . It has been shown that cat infections present higher occurrence in warm, moist and low altitude regions, maybe linked to oocysts sporulation and survival of T. gondii oocysts in such latitudes 34 . Similarly, T. gondii seroprevalence in pigs was associated with lower geographical latitude and higher mean annual temperature 59 , fact that may suggest high environmental contamination with T. gondii sporulated oocysts. It seems to be clear that oocysts shedding by cats constitute the essential element for sustainment of the parasite in the environment, this was demonstrated when extremely low seroprevalence of T. gondii (0.9%) was detected in feral pigs from a remote island lacking cats in the USA 60 . Furthermore, the time period of study might influence the results, as the infection rate is higher in autumn, winter 22 , and rainy years 20 . One may consider breed as a variable factor for cat T. gondii infection. It has been shown that Toxoplasma seroprevalence is highly variable in different cat breeds from 18.8% in Burmese cats to 60% in Persian cats 35 . Even though, a high occurrence rate of T. gondii infection in cat may be attributed to some important factors including: uncontrolled food and access to contaminated sources, wandering outdoor, humid and temperate climate; and cat abundance. Furthermore, stray cats have been shown to have a higher seroprevalence compared to pet cats, which can be explained by more access to contaminated source and outdoor living 31,34,61 . Additionally, pet cats with an outdoor access are also at an increased risk of infection compared to those kept indoor 32,35 . It has been reported that rural cats show a higher seroprevalence rate of T. gondii compared to urban ones 62 . www.nature.com/scientificreports/ Furthermore, the different diagnostic methods used to detect T. gondii antibodies and oocysts could influence the results. While the different techniques used for anti-T. gondii antibodies detection showed comparably good diagnostic performance, most of the studies aiming to detect T. gondii oocysts employed less reliable microscopic methods, which might result in false positives, as oocysts and sporocysts of some other coccidia (e.g. Hammondia hammondi, Besnoitia darlingi) may resemble those of T. gondii 48 . It shows the necessity of testing environmental and fecal samples by using specific-PCR aided with amplicon sequencing for identity confirmation 63 .
Felids as key elements in the epidemiology of toxoplasmosis should be considered as a potential threat to animal and public health, due potential oocysts contamination of the environment; such information is still missing in several worldwide locations, so further epidemiological investigations on final hosts would be of special interest for evaluating the status of T. gondii infection and risk assessment implementations. Further investigations based on QMRA approaches 64,65 combining raw data in Felidae with those from the environmental side and those from susceptible hosts will complement the One Health puzzle in defined areas.
In present meta-analysis, it is shown that about one-third of domestic and non-domestic cats have been exposed to T. gondii, and globally about 1 in 50 cats are actively shedding T. gondii or T. gondii-like oocysts. In addition, 16.2% of the soil samples examined were contaminated with T. gondii-like oocysts informing on a broad environmental distribution. Felids are the only final host of T. gondii and play a major role in its life cycle, therefore measures aiming to reduce environmental contamination with T. gondii oocysts will be of major interest, and a One Health perspective covering human, animal and environmental health should be taken into account.

Data availability
The data that supports the findings of this study are available in the supplementary material of this article. www.nature.com/scientificreports/