https://orcid.org/0000-0002-3412-1033
Figures
Abstract
Background
Toxoplasma gondii, a cosmopolitan protozoan parasite causes toxoplasmosis in humans and many species of domestic and wild animals. T. gondii instigates significant economic losses in sheep and goat farming industry and can lead to abortion, stillbirth, congenital malformations and neonatal losses. The objective of this protocol is to evaluate worldwide seroprevalence of T. gondii exposure in goats using Bayesian hierarchical meta-analysis and geographic information system (GIS).
Methods
A comprehensive literature search will be conducted using search engines, including Web of Science, ScienceDirect, Scopus, PubMed, ProQuest, EMBASE, PROSPERO Register and, Google Scholar without date and language restrictions. The authors search for cross-sectional studies that determine the seroprevalence of T. gondii in goats. Two reviewers will independently screen, selected studies; also, they will extract data, and assess the risk of bias. In case(s) of disagreement, a consensus will be reached with the help of a third author. The Bayesian hierarchical meta-analysis will use to estimate country and worldwide true seroprevalence of T. gondii, which is consist of the sensitivity and specificity of the applied serological assays. The obtained data will be used to identify country-level risk factors associated with T. gondii exposure using GIS in the ArcGIS software.
Citation: Amouei A, Mizani A, Hanafi-Bojd AA, Jafari-Koshki T, Sarvi S, Aghayan SA, et al. (2023) Global epidemiology and spatial distribution of Toxoplasma gondii in goats: Protocol for a systematic review and Bayesian hierarchical meta-analysis. PLoS ONE 18(12): e0290012. https://doi.org/10.1371/journal.pone.0290012
Editor: Bibi Razieh Hosseini Farash, Mashhad University of Medical Sciences, ISLAMIC REPUBLIC OF IRAN
Received: July 8, 2023; Accepted: July 31, 2023; Published: December 29, 2023
Copyright: © 2023 Amouei et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Deidentified research data will be made publicly available when the study is completed and published.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
T. gondii is one of the most well studied coccidian parasites and causes widespread infections in almost all warm-blooded animals including humans and livestock. Parasite also plays a considerable zoonotic role and has both veterinary and medical importance worldwide [1, 2]. The infection is usually asymptomatic in humans but can also cause severe complications in immunocompromised individuals and abortion. Parasite may lead to abortion in ruminants, cause huge economic losses to the livestock industry and facilitate the transmission of infectious and parasitic disease to humans, due to the consumption of infected raw or undercooked meat or milk [3–5].
Goats are one of the main sources of meat borne infection [6] and play an essential role in the economy of some countries [7] since they are significant sources of their products (meat and milk). Toxoplasmosis is globally recognized as a food-borne disease and a public health problem. Researchers reported the occurrence of toxoplasmosis in humans and goats, which was linked to the consumption of raw goat milk and milk products [8, 9]. Based on data obtained in some regions, the seroprevalence of T. gondii can be as high as 77% in goats [4]. It is estimated that 30% to 63% of human infections have been exposed to T. gondii [8]. In Europe, among 14 foodborne diseases, toxoplasmosis has the highest human incidence and in the USA was suggested as one of three pathogens (together with Salmonella and Listeria) [10, 11]. However, human infections from different regions of the world depend on prevalence the of parasite in animals and eating habits [12].
Numerous reports from epidemiological surveys have been conducted on animal toxoplasmosis, although globally, no exhaustive documented data are available for the prevalence of goat toxoplasmosis. The aims of the current meta-analysis, therefore, were, (i) to estimate the true prevalence of T. gondii among goats in the world in a Bayesian framework, (ii) to assess its association with several risk factors and to display the distribution of T. gondii exposure in different countries using GIS.
Materials and methods
Study design
This systematic review will be conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol (PRISMA-P) guidelines [13] when reporting the findings (S1 Checklist). The current study protocol was registered with PROSPERO (https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020107928) [14]. The systematic review with meta-analysis will be managed following the recommendations of the Cochrane Collaboration Handbook of Systematic Reviews [15], as a systematic, transparent and reproducible method to investigate of scientific evidence.
Eligibility criteria
The eligibility criteria for the study were determined using the PICOS classification (Population, Intervention, Comparison, Outcome, and Study design) as a tool to guide the research and adjust the search strategy, as described in Table 1 [16]. No language and date restrictions were applied in this study.
Search strategy
To assess the seroprevalence of T. gondii goats from a global perspective, six databases, namely the PubMed, Scopus, ScienceDirect, Web of Science, EMBASE, ProQuest, PROSPERO Register, and the Internet search engine Google Scholar, will be searched by two authors (AA and AM) of a trained team for relevant studies published before June 2022. All databases will be searched using the keywords terms presented in Table 2 in all fields. Keywords will use individually or combined with a list of all of them and without any language restrictions. The authors systematically scrutinized a manual search of the reference lists of the identified reports and reviews.
Study selection
All cross-sectional studies will be included if they meet the inclusion criteria. Two reviewers of the research team will independently review the retrieved papers. All search records will be imported to EndNote software V.X7.0.1 software. Following the removal of duplicate entries, each of these studies will be screened, by reading the title and the contents of the abstract or the full-text to identify papers that described studies. A data extraction sheet will then be drawn and data will be recorded in the selection sheet. Discrepancies will be resolved with the provision for arbitration from a third reviewer within the review team. Study selection will be performed according to the PRISMA flowchart (Fig 1) [17].
Data extraction
Data from the retrieved studies will be collectd independently by two reviewers, using a piloted form which will be checked by another investigator. Abstracts will not be included in the meta-analysis for risk factors. Finally, data will be transfered to a Microsoft Excel software (version 2016; Microsoft, Redmond, WA, USA). The following information will be used (S2 Checklist): title, first author, year of publication, continent, country, total sample size, number of positive samples, gender, age group (< 1 year and ≥ 1 year), detection method, the presence or absence of cats on the farm, system of management (extensive, intensive, and semi-intensive), sensitivity, specificity, as well as geographic and natural climatic conditions for each country (latitude, longitude, altitude, mean annual temperature, relative humidity, and annual rainfall). Data will also be collected for both minimally adjusted and maximally adjusted risk estimates, if available (presence of cat, farm management age, and gender adjustment). The inclusion criteria are: 1) studies published until June 10th, 2021, 2) cross-sectional studies that estimate the prevalence of toxoplasmosis in goats, 3) original research studies, 4) papers with available full texts, and 5) studies with information on the total sample size and positive samples. The exclusion criteria are: 1) studies with no exact information about the sample size and the diagnostic criteria, 2) descriptive, letters to the editors, and review articles, 3) studies conducted in other animal and human models.
Assessment of bias
A formal assessment for the risk of bias in the included reports will have limited utility given the lack of an appropriate assessment tool in animal prevalence studies. Although quality checklists have been developed for meta-analyses studies of disease prevalence, many aspects of the available checklist are not straightly related to the present survey question [18]. Therefore, study quality will be assessed based on 10 key factors formulated as questions which will score 3, 2, or 1 based on a simple scale system (3 for “yes”, 2 for “no”, or 1 for “unsure”) [19]. These questions are as follows:
Q1: Is the purpose of the study defined well?
Q2: Is the target population (time and location) defined well?
Q3: Are the inclusion and exclusion criteria defined well?
Q4: Is the sampling method (random sampling) specified well?
Q5: Is the sample size adequate?
Q6: Is the working method described well?
Q7: Is an appropriate diagnostic method used?
Q8: Are the subgroups divided well?
Q9: Is a proper analysis does well?
Q10: Are the effects of confounders removed well?
A quality score will determine by rescaling the sum of scores of each eligible paper between 10 to 30 points with a low-quality report earning a score of ≥15. Quality assessment will be completed independently by two assessors, and a table of quality score computation for each eligible paper will be addressed in the S2 Checklist.
Data analysis
The data entered into Microsoft Excel will be analyzed using the STATA statistical software (version 14; Stata Corp, College Station, TX, USA). The collected data will be analyzed using the Bayesian hierarchical meta-analysis model as follows. First, the number of seropositive cases in the i-th country, xi, grouped according to the continents, will be assumed to follow a binomial distribution with a country-specific apparent seroprevalence APi. Logit-transformed seroprevalence in each country will be assumed to follow a normal distribution with a common mean over the pertinent continent, that is, logit (APi) ~ Normal (θj, σw2), where i represents the country and j represents its continent and σw2 is the variance of seroprevalence over the continent. We will assume normally distribute hyperpriors for the continent mean seroprevalence, θj ~ Normal (θ0, σb2) in which θ0 is the worldwide mean and σb2 is the variance of seroprevalence of T. gondii among continents. After the estimation of the Bayesian posteriors, the logit-transformed missing seroprevalence of T. gondii for study k in a continent, logit (θk*), will be estimated using the posterior predictive distributions. That is, the missing data will be generated from a normal distribution as .
The next step will include the estimation of the true seroprevalence of each study (TPi). To conduct this analysis, data on sensitivity (Se) and specificity (Sp) of commercial kits used in these studies will be collected. This information will be considered in their contribution to estimating the true seroprevalence, TPi, by the equation below.
In the Bayesian setting, due to their possible data range, Beta-distributed priors will be used for Se and Sp [20]. As before, a normal distribution will be considered for logit-transformed true seroprevalence; , with normal hyperpriors
.
In all models, a Cauchy distribution will be assumed for variances. All models will be run in the R-interface of Stan Bayesian modeling language [21]. The convergence of models will be assessed by inspecting history, autocorrelation plots, R-hat and the number of effective sample sizes. Posterior estimates of parameters will be reported with 95% credible intervals.
Spatial epidemiology
Data from this the systematic review will be used to identify country-level risk factors associated with T. gondii exposure using GIS. In the ArcGIS software, the spatial distribution of this parasite will be mapped. The MaxEnt model (MaxEnt software version 3.4.1 [22] will be used to predict the distribution of T. gondii. The following six climatic factors and four environmental variables will be used:
Altitude as the Elevation above the sea level (m), BIO1 as the Annual Mean Temperature (°C), BIO2 as the Mean Diurnal Range [Mean of monthly (max temp—min temp) (°C)], BIO3 as the Isothermality (BIO2/BIO7) (×100), BIO7 as the Temperature Annual Range (BIO5-BIO6) (°C), BIO12 as the Annual Precipitation (mm).
The climatic variables and altitude will be downloaded from the two WorldClim website at a resolution of 1 km2. A set of global climate layers with a spatial resolution of about 1 km2 will be download from the online free WorldClim platform (https://www.worldclim.org). WorldClim (version 2) will be used for average monthly climate-data (minimum, mean, and maximum temperature and precipitation). All points’ coordinates will be manually put into the nasa.gov system (https://power.larc.nasa.gov/data-access-viewer/) and inter-annual values will be extracted for precipitation, humidity and temperature.
In total 80% of collection points will be used randomly by MaxEnt for model training and 20% will be kept for testing the results. Finally, the output of the model will identify the environmental suitability for the presence of parasite. Jackknife analysis of the MaxEnt model will be used to find which climatic and/or environmental factors have the greatest effect on the distribution of Toxoplasma.
Discussion
T. gondii is a protozoan parasite that infects a wide range of vertebrate hosts, such as birds, wild and domestic animals and humans. Among domestic animals, goats are the most susceptible host to T. gondii. Infection caused this parasite can lead to pre-term deliveries, abortion, weak newborns, and death in young or adult animals and decreased milk production, diarrhea and hair opacity among symptomatic animals [23].
The proposed protocol has several strengths. Firstly, it will adhere to the guidelines, the studies published in multiple languages (if available), and several researchers will independently perform the intrinsic evaluation method. Secondly, using Bayesian hierarchical and GIS analysis, the true prevalence of T. gondii exposure in goats will be systematically reviewed regionally and globally. The Bayesian hierarchical model used in this study will allow to incorporate the sensitivity and specificity of the serological test.
Thirdly, we will undertake the geographical distribution of exposure in a cross-sectional serological study, including prediction modeling using GIS. The present study has some limitations as well: (i) the study design and between-study heterogeneity; (ii) the restriction of research to unpublished, ongoing, gray literature; (iii) the use of diverse diagnostic methods with variations in the sensitivity and specificity and iiii) the use of reports with limited information on some of risk factors. We also believe that these limitations will be justified by meta-regression analysis on the adjusted seroprevalence and a regression model implemented within a Bayesian hierarchical framework. Nevertheless, from a global perspective, our multidimensional approach report here will be very close to true parasite seroprevalence in goats.
The authors hope that the spatial distribution of Toxoplasma prevalence in meat-producing animal will be helpful in developing effective intervention strategies to reduce the burden of this zoonotic disease. In addition, we expect that using spatial epidemiology and GIS technology will map diseases to identify the known distribution of pathogens.
Acknowledgments
Authors express their gratitude on the Research Deputy and Technology and Toxoplasmosis Research Center (TRC), Mazandaran University of Medical Sciences, Sari, Iran, to approve the research project No. 14522.
References
- 1. Minuzzi CE, Pires Portella L, Brӓunig P, Antonio Sangioni, Ludwig A, Silva Ramos S, et al. Isolation and molecular characterization of Toxoplasma gondii from placental tissues of pregnant women who received toxoplasmosis treatment during an outbreak in southern Brazil. PLoS ONE. 2020; 15(1): 1–7. pmid:31999785
- 2. Stelzer S, Basso W, Benavides Silván J, Ortega-Mora LM, Maksimov P, Gethmann J, et al. Toxoplasma gondii infection and toxoplasmosis in farm animals: Risk factors and economic impact. Food Waterborne Parasitol. 2019; 12: 1–32. pmid:32095611
- 3. Cook AJ, Gilbert RE, Bufolano W, Zufferey J, Petersen E, Jenum PA, et al. Sources of Toxoplasma infection in pregnant women: European multicentre case-control study. European research network on congenital toxoplasmosis. BMJ. 2000; 321 (7254): 142–147. https://doi.org/10.1136/bmj.321.7254.142.
- 4. Tenter AM, Heckeroth AR, Weiss LM. Toxoplasma gondii: from animals to humans. Int. J. Parasitol. 2000; 30 (12–13), 1217–1258. https://doi.org/10.1016/s0020-7519(00)00124-7.
- 5. Lopes WDZ, Rodriguez JD, Souza FA, Santos TR, Santos RS, Rosanese WM, et al. Sexual transmission of Toxoplasma gondii in sheep. Vet Parasitol. 2013; 195 (1–2): 47–56. https://doi.org/10.1016/j.vetpar.2012.12.056.
- 6. Rasti S, Marandi N, Abdoli A, Delavari M, Mousavi SGA. Serological and molecular detection of Toxoplasma gondii in sheep and goats in Kashan, Central Iran. J Food Saf. 2018; 38 (2): 1–5. https://doi.org/10.1111/jfs.12425.
- 7. Jimenez-Martín D, García-Bocanegra I, Almería S, Castro-Scholten S, Dubey JP, Amaro-Lopez MA, et al. Epidemiological surveillance of Toxoplasma gondii in small ruminants in southern Spain. Prev Vet Med. 2020; 183: 1–5. https://doi.org/10.1016/j.prevetmed.2020.105137.
- 8. Meerburg BG, Van Riel JW, Cornelissen JB, Kijlstra A, Mul MF. Cats and goat whey associated with Toxoplasma gondii infection in pigs. Vector Borne Zoonotic Dis. 2006; 6 (3): 266–274. pmid:16989566
- 9. Sacks JJ, Roberto RR, Brooks NF. Toxoplasmosis infection associated with raw goat’s milk. J Am Med Assoc. 1982; 248 (14): 1728–1732. pmid:7120593
- 10. Olivier A, Herbert B, Sava B, Pierre C, John DC, Aline DK. Surveillance and monitoring of Toxoplasma in humans, food and animals: a Scientific Opinion of the Panel on Biological Hazards. EFSA J. 2007; 583: 1–64. doi.org/10.2903/j.efsa.2007.583.
- 11. Roghmann MC, Faulkner CT, Lefkowitz A, Patton S, Zimmerman J, Morris JG. Decreased seroprevalence for Toxoplasma gondii in Seventh Day Adventists in Maryland. Am J Trop Med Hyg. 1999; 60 (5): 790–792. https://doi.org/10.4269/ajtmh.1999.60.790.
- 12. Opsteegh M, Langelaar M, Sprong H, Den Hartog L, De Craeye S, Bokken G, et al. Direct detection and genotyping of Toxoplasma gondii in meat samples using magnetic capture and PCR. Int J Food Microbiol. 2010; 139 (3): 193–201. https://doi.org/10.1016/j.ijfoodmicro.2010.02.027.
- 13. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ. 2015; 349: 1–25. pmid:25555855
- 14. Daryani A, Amouei A, Mizani A, Nayeri T, Sarvi S, Alipour A, et al. Global epidemiology and spatial distribution of toxoplasmosis in goat: a systematic review and meta-analysis. PROSPERO 2020 CRD4202010792 Available from: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020107928.
- 15. Higgins J, Thomas J, Chandler J, Cumpston M, Li T, Page M, et al. Cochrane Handbook for Systematic Reviews of Interventions version 6.1 (updated September 2020). Available from: www.training.cochrane.org/handbook. Accessed 26 oct 2020.
- 16. Eriksen MB, Frandsen TF. The impact of patient, intervention, comparison, outcome (PICO) as a search strategy tool on literature search quality: a systematic review. J Med Libr Assoc. 2018; 106 (4): 420–431. pmid:30271283
- 17. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Syst Rev. 2021;10(1):1–11. https://doi.org/10.1186/s13643-021-01626-4.
- 18. Fanelli A, Battisti E, Zanet S, Trisciuoglio A, Ferroglio E. A systematic review and meta‐analysis of Toxoplasma gondii in roe deer (Capreolus capreolus) and red deer (Cervus elaphus) in Europe. Zoonoses Public Health. 2021; 68(3): 182–93. pmid:33164352
- 19. Hoy D, Brooks P, Woolf A, Blyth F, March L, Bain C, et al. Assessing risk of bias in prevalence studies: Modification of an existing tool and evidence of interrater agreement. J Clin Epidemiol. 2012; 65: 934–939. pmid:22742910
- 20. Deng H, Devleesschauwer B, Liu M, Li J, Wu Y, van der Giessen JW, et al. Seroprevalence of Toxoplasma gondii in pregnant women and livestock in the mainland of China: a systematic review and hierarchical meta-analysis. Sci Rep. 2018;8(1):1–10. https://doi.org/10.1038/s41598-018-24361-8.
- 21. Stan Development Team (2022). “RStan: the R interface to Stan.” R package version 2.21.7, https://mc-stan.org.
- 22. Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. Ecol modell. 2006; 190 (3–4): 231–59. https://doi.org/10.1016/j.ecolmodel.2005.03.026.
- 23. Batista SP, dos Santos Silva S, Sarmento WF, Silva RF, do Nascimento Sousa R, de Menezes Oliveira CS, et al. Prevalence and isolation of Toxoplasma gondii in goats slaughtered for human consumption in the semi-arid of northeastern Brazil. Parasitol Inter. 2022; 86: 1–4. pmid:34506948