After screening studies based on the eligibility criteria, two investigators (MD and BE) independently extracted relevant data. Any discrepancies were resolved through discussion with a third author. The extracted data encompassed both quantitative and qualitative information and were systematically organized into Word tables and an Excel spreadsheet. The final dataset comprised 104 articles, including details on: Primary author’s name, Publication and study year, Geographical location, Study animal species (host), Sample size and number of positives, diagnostic methods, data collection methodologies, Ethical considerations, prevalence and principal findings of the parasite type.

In this review, the quality of the included studies was evaluated using the AMSTAR-2 tool (Supplementary material 2), as the methodological rigor of the studies directly influences the reliability of the meta-analysis results. Two investigators (MD and AB) independently assessed study quality, applying the 15-item AMSTAR-2 checklist, which covers both randomized and nonrandomized trials of health interventions. Some items are applicable to both types of studies, while others are specific to either randomize or nonrandomized trials. Since this meta-analysis focused on nonrandomized (observational) studies, only the relevant items were used. The quality assessment is integral in determining the confidence in the results and ensuring that the conclusions drawn from the meta-analysis are based on studies with sound methodologies. Regarding data extraction, two investigators (MD and BE) independently extracted relevant data from the selected studies. In the case of any discrepancies, the investigators resolved disagreements through discussion and consultation with a third author (AB), ensuring consensus and consistency in the data.

The random effects model, which employs the restricted maximum likelihood method (REML) to compute within- and between-study variability, was employed to estimate the aggregated prevalence and 95% confidence intervals. This model was utilized to conduct a comprehensive meta-analysis (overall effect size (ES) or often represented as a percentage or proportion), assess heterogeneity, and determine the weight of each study. Furthermore, graphs and tables were employed to illustrate the prevalence status of the primary parasitic zoonosis on the basis of the geographical distribution of animal species, the type of parasites, and the study years.

To estimate the prevalence, we extracted data from the number of events and the total number of samples. These data were then subjected to proportional meta-analysis via the “metaprop” function of the “meta” package version 4.1.3-0 in Balduzzi et al. (46) R statistical software. Pooled proportions of prevalence were estimated via logit transformation via a logistic-normal random effects regression model. Subgroup analysis was conducted via the mixed effect logistic regression model.

The Cochran’s Q test (indicated by the p value), τ2 (representing between-study variance), and the inverse variance index (I²) were employed to assess the sources of heterogeneity. The I² index, as elucidated by Higgins and Thompson (47), was calculated to indicate low, moderate, and high levels of heterogeneity, corresponding to I² values of 25, 50, and 75%, respectively. Heterogeneity was deemed statistically significant if the I² value exceeded 50%, and the Q test yielded a p value less than 0.10. τ² (tau-squared) measures the variance between studies. In simple terms, it tells us how much the true effects differ from study to study. I² (I-squared) indicates the percentage of variability across studies that is due to heterogeneity rather than chance. For example, if I² = 50%, this means that 50% of the variation between study results is due to true differences rather than random sampling error. By using these indices, we were able to identify the sources and extent of heterogeneity in the meta-analysis.

The extent of study heterogeneity was evaluated through a forest plot diagram, which depicted the weights, effect magnitudes, and 95% confidence intervals for between-study variability. A subgroup analysis of the prevalence of the primary prevalence of toxoplasmosis, cryptosporidiosis, bovine cysticercosis and cystic echinococcosis was performed, considering factors such as publication year, study location or region, species of study animals, and methods of detection.

To assess publication bias, funnel plot diagrams and Egger’s regression test were used. A funnel plot is a scatter plot of the effect sizes from individual studies against their sample sizes. If there is no publication bias, the plot should resemble a symmetrical funnel. Egger’s regression test is a statistical method that helps determine whether the funnel plot is symmetrical. If the plot is asymmetrical (i.e., the studies are skewed in one direction), it suggests potential publication bias, meaning that smaller or less significant studies may be missing from the analysis. A p-value from the test greater than 0.05 typically indicates that the funnel plot is symmetric, suggesting no significant publication bias.

According to the PRISM 2020 flow chart (Figure 1), an initial search identified 3,581 articles across various electronic databases. After excluding 74 articles due to duplicates and non-English records, 3,507 articles remained. Of these, 1,227 were eliminated based on a review of titles and abstracts. A total of 2,280 articles were then sought for retrieval, but 1,896 could not be obtained. Out of the 384 articles assessed for eligibility, 200 were excluded for various reasons. Ultimately, 104 studies were included in the meta-analysis.

This systematic review included 104 studies focusing on four major zoonotic parasitic diseases affecting domestic ruminants in Ethiopia: bovine cysticercosis, cystic echinococcosis, toxoplasmosis, and cryptosporidiosis. All of the investigations employed an observational study design.

The current systematic review includes a total of 34 articles on cysticercosis in cattle, with each article presenting one or multiple examinations of parasites (14). These articles were published between 2008 and 2023. Notably, the individual studies included in this systematic review reported a minimum sample size of 234 in Northeast Ethiopia (48) and a maximum sample size of 4,456 (49) in the Amhara Region in Northwest Ethiopia. The overall number of animals sampled throughout the years of the study was 25,065, of which 3,076 tested positive for bovine cysticercosis. The majority of the data collection methods utilized by observations conducted at abattoirs and surveys conducted through questionnaires. A comprehensive overview of the investigated studies is provided in Table 1.

This systematic review provides a comprehensive analysis of 37 studies on cystic echinococcosis in domestic ruminants, including cattle, sheep, and goats. In this systematic review, the largest sample size examined in a study was 5,194 animals, whereas the smallest study analyzed only 246 animals. A total of 28,024 animals were examined and among them, 7,277 were positive for cystic echinococcosis. Table 2 provides a comprehensive summary of the reviewed studies.

In the present study, 14 articles involving a total of 6,957 animals were examined, with 2,889 of those animals diagnosed as positive for toxoplasmosis. According to Teshale et al. (32), the central region of Ethiopia exhibited the highest prevalence rate of 74.9% at the individual study level. In comparison, Esubalew et al. (33) reported a prevalence rate of 70.5% in the Amhara region, Northwest Ethiopia. With respect to the species of host animals, Esubalew et al. (33) reported a high prevalence of toxoplasmosis (70.5%) in the Amhara region in sheep, followed by goats (74.9%), as indicated by Teshale et al. (32), and cattle (59%), as reported by Maru et al. (50), in the Amhara region. Among the included studies, five did not state ethical considerations. The details of the articles are summarized in Table 3.

A total of 19 studies were discovered for the estimation of the pooled prevalence of cryptosporidiosis. The majority of the studies are located in the southern region of Ethiopia, specifically within the Oromia region. The study subjects included calves, kids, lambs, and adult sheep and cattle, which consisted of a total of 6,972 and 959 cryptosporidiosis infections (event). Notably, most of the literature has focused on calves and lambs. The most frequently isolated species were C. bovis, C. andersoni, C. ubiquitum, and ryane. The southwest region of Ethiopia, specifically the Oromia regional state, recorded a high prevalence rate (24%) of cryptosporidiosis with calves documented by Gashaw et al. (51) (Table 4).

The diagnostic approaches varied across the four parasitic diseases, with post-mortem meat inspection, serological tests, microscopy, and molecular techniques being the primary methods. Post-mortem meat inspection was the gold standard for diagnosing bovine cysticercosis and cystic echinococcosis, as both diseases were primarily identified through abattoir-based surveys (Tables 1, 2). The presence of cystic lesions in infected organs served as a key diagnostic indicator. Since data were collected from slaughtered animals, ethical approval was generally not required for these studies. For toxoplasmosis, serological tests were the primary diagnostic tool, with ELISA, Latex Agglutination Test (LAT), and Direct Agglutination Test (DAT) commonly used across studies (Table 3). These methods detect specific antibodies in blood samples, making them suitable for screening live animals. In several studies, serological tests were supplemented with questionnaire surveys to assess risk factors, transmission dynamics, and disease prevalence in different livestock populations.

Cryptosporidiosis was predominantly diagnosed using microscopy (modified Zeehl–Nielsen), which detects oocysts in fecal samples. Some studies employed PCR and nested PCR to achieve molecular characterization of Cryptosporidium species, including C. bovis, C. andersoni, C. ubiquitum, and C. ryanae (18, 52, 53) as shown Table 4. However, molecular diagnostics were used in only a limited number of studies. Additionally, questionnaire surveys were commonly employed as a supplementary diagnostic tool across all four parasitic diseases. These surveys provided valuable insights into risk factors, transmission patterns, and disease awareness among livestock owners. By integrating observational, serological, molecular, and questionnaire-based approaches, the studies provided a comprehensive assessment of zoonotic parasitic diseases in Ethiopia.

This meta-analysis of the included studies provides an in-depth assessment of the pooled prevalence, heterogeneity, and potential sources of variability for each parasitic disease. The findings highlight significant variations in prevalence rates across regions, host species, and diagnostic methods, emphasizing the need for subgroup analyses to better understand disease distribution and impact. The substantial heterogeneity were observed in this meta-analysis is likely driven by several factors, including variations in study design, host species, diagnostic techniques, and geographic distribution. The heterogeneity test confirmed significant differences between species, suggesting that biological and ecological factors play a major role in prevalence estimates.

Regional differences in prevalence are likely influenced by climate, altitude, and management practices. Warmer, humid environments may enhance oocyst survival and transmission, while variations in animal husbandry (e.g., free grazing vs. confined feeding) impact exposure risk. Differences in veterinary healthcare infrastructure further affect disease detection and control.

Heterogeneity also arises from methodological variations among studies. Differences in sampling strategies (random, convenience, stratified), population characteristics, and study periods contribute to variability. Sample sizes ranged from small farms to large epidemiological studies, affecting prevalence estimates. Study designs varied, with some focusing on controlled farm settings and others on free-grazing animals, leading to inconsistencies in exposure risk. Additionally, the use of different diagnostic assays influenced reported prevalence rates.

The following sections summarize the key results for each disease, incorporating heterogeneity measures, statistical findings, and publication bias assessments.

The estimated pooled prevalence was 38% (95% CI: 29–48%, Figure 2), indicating a relatively high burden of toxoplasmosis in domestic ruminants. However, significant heterogeneity was observed (I² = 98%; τ² = 1.1286; p < 0.01), suggesting variations in study design, host species, or diagnostic methods across the included studies. The Galbraith plot confirmed high inter-study heterogeneity, as nearly 95% of the studies fell outside the confidence interval (Figure 3). Further subgroup analyses based on host species and geographic location are essential to understand which animal groups or regions experience higher prevalence rates. Despite the heterogeneity, no significant publication bias was detected, as indicated by Egger’s regression asymmetry coefficient (b = 2.38; 95% CI: −2.19, 4.43; p > 0.05). The funnel plots (Figure 4) also demonstrated a symmetrical distribution, reinforcing the robustness of the meta-analysis findings.

Baujat diagnostic plots (54) was used to detect studies that overly contribute to the heterogeneity in a meta-analysis. The plot shows the contribution of each study to the overall heterogeneity (as measured by Cochran’s Q) on the horizontal axis and its influence on the pooled effect size on the vertical axis. Accordingly, a study by author ID number 5 which denotes Sheferaw and Abdu (1) was found to be an influential study graphically which contributed overly to the overall heterogeneity, though it did not affect the pooled effect size significantly as shown in the leave-one-out meta-analysis result (Figure 5). High heterogeneity (I² = 98%) suggests further subgroup is needed to explore sources of variability. Finally, a leave-one-out analysis was conducted to evaluate the influence of a single study on the overall effect size estimate. In this analysis, a specific study was excluded, and a meta-analysis was performed on the remaining (n−1) studies. If the confidence interval of the study did not encompass the overall effect size estimate, it was deemed that the study had a substantial impact on it. In this study, the overall effect size estimate was 0.38, and it fell within the confidence intervals of all the studies. As a result, excluding one study did not have a substantial impact on the overall effect size estimate (Figure 2). Similarly, sensitivity analyses have been conducted in a comparable manner for toxoploamosis disease studies to assess the impact of individual studies on the overall pooled prevalence estimates. For example, in the study by Abebe et al. (55), the graphical sensitivity analysis indicated that this study had a notable influence on the reported prevalence of cryptosporidiosis, suggesting that its inclusion may have contributed to variations in prevalence estimates. Likewise, in the study by Higgins and Thompson (47), sensitivity analysis revealed that this study was influential in the context of cystic echinococcosis, demonstrating a similar pattern of influence.

However, despite these individual studies showing some degree of influence in graphical representations, their exclusion did not significantly alter the overall pooled prevalence estimates. This suggests that while certain studies may introduce variability at a localized level, the robustness of the meta-analysis remains intact, and the general conclusions drawn from the pooled data are not substantially affected. These findings reinforce the stability of the prevalence estimates and highlight the reliability of our analytical approach in accounting for potential outliers.

For cryptosporidiosis, 19 studies were analyzed, revealing a pooled prevalence of 14% (95% CI: 12–16%). The heterogeneity was substantial (I² = 86%; τ² = 0.1910; p < 0.01), indicating that study-level differences significantly influenced the estimated prevalence rates. The forest plot (Figure 6) illustrates variations in prevalence across studies, while the funnel plot (Figure 7) suggests that the studies were evenly distributed without asymmetry, indicating no significant publication bias. The differences in diagnostic techniques (microscopy vs. molecular methods) and host species (calves, lambs, and adult ruminants) may contribute to the heterogeneity. Additionally, environmental factors such as hygiene, management practices, and seasonal variations could also influence prevalence rates. The Egger regression test (b = 1.2953; CI: 0.1792, 1.4114; p = 0.2007) further confirmed the absence of statistically significant publication bias in the meta-analysis.

The overall pooled prevalence was 9% (95% CI: 7–17%: Figure 8), highlighting a moderate burden of bovine cysticercosis in Ethiopia. The analysis showed high between-study variability (τ² = 0.861; H² = 59; I² = 98.31%), reinforcing the need for subgroup analyses based on host characteristics and geographic factors. The funnel plot (Figure 9) suggested asymmetry, indicating the possibility of publication bias, which was confirmed by Egger’s regression test (b = 0.3229; 95% CI: 0.1792, 0.4114; p = 0.02812). This suggests that smaller studies may be underreported or that there is selective reporting of significant results, potentially influencing the pooled prevalence estimate.

The pooled prevalence was estimated at 25% (95% CI: 29–48%: Figure 10), indicating a substantial disease burden among domestic ruminants. High heterogeneity (τ² = 0.7104; I² = 98.0%) was observed, likely due to differences in study locations, sample sizes, and diagnostic approaches. The funnel plot (Figure 11) showed no asymmetry, suggesting that smaller studies were not systematically underrepresented. Similarly, Egger’s test (b = −2.58; 95% CI: 0.17–0.31; p = 0.08075) confirmed that publication bias was not statistically significant. The high prevalence and widespread distribution of cystic echinococcosis underscore its public health and economic impact, reinforcing the need for improved disease control strategies in Ethiopia.

A subgroup analysis of each of the four parasitic diseases (toxoplasmosis, cryptosporidiosis, bovine cysticercosis and cystic echinococosis) was conducted on the basis of year, species, and region. All of the analyses of exploratory outcomes showed considerable heterogeneity (I² > 71). Significant statistical heterogeneity in the subgroup analysis reveals a likely interaction among exploratory variables.

The subanalysis of toxoplasmosis prevalence across different animal species revealed substantial variability, which we now interpret more clearly. The studies were categorized into three groups based on animal species: goats (n = 13), sheep (n = 11), and cattle (n = 3). The pooled prevalence estimates for toxoplasmosis varied across these species, with sheep showing the highest prevalence at 41% (95% CI: 26–57%; I² = 98%), followed by goats at 39% (95% CI: 28–51%) and cattle at 28% (95% CI: 8–63%) (Figure 12). Despite the high level of within-study variability across all species (I² = 98%), the heterogeneity test revealed statistically significant differences (Q = 180.28; D.F. = 18; p < 0.001). However, no significant difference in the number of studies between species was observed (Q = 0.8; D.F. = 2; p = 0.7855), indicating that the observed variation in prevalence is not due to the number of studies per group but likely reflects other factors, such as ecological or methodological influences.

The regional variability analysis also provided important insights. We found similar levels of variability (I² = 98%) across the Amhara, AA, and Oromia regions. The pooled prevalence for toxoplasmosis was highest in the Amhara region (53%), followed by AA (44%), Tigray (26%), Oromia (35%), and SNNP (23%) (Figure 13). The regional heterogeneity test (Q = 8.21; D.F. = 4; p < 0.00841) indicated statistically significant differences, suggesting that regional factors, such as climate and management practices, may influence the prevalence of toxoplasmosis.

Furthermore, subgroup analysis based on diagnostic techniques (LAT, DAT, and ELISA) revealed considerable heterogeneity (I² > 86%) across all methods (Figure 14). The highest degree of heterogeneity (I² = 98%) was observed in studies using the DAT and ELISA diagnostic methods. The pooled prevalence was highest in studies using the LAT method (58%), followed by DAT (35%) and ELISA (23%). The subgroup difference test showed a statistically significant group effect (Q = 23.49; D.F. = 2; p < 0.0001), suggesting that the choice of diagnostic method plays a significant role in the variation in prevalence estimates. We have emphasized the potential clinical and ecological implications of these findings and noted that the variability observed may reflect underlying biological, environmental, or methodological factors that were not fully addressed in the analysis. Regional factors such as climate, management practices, and differences in diagnostic methods could contribute to the observed discrepancies in prevalence estimates. We also acknowledge that further research is needed to explore the biological, ecological, and methodological factors that contribute to these variations. This additional research would help provide a more comprehensive understanding of the true prevalence of toxoplasmosis and its implications for livestock health and management.

The subanalysis of cryptosporidiosis prevalence across different publication years revealed the highest prevalence between 2017 and 2019 (16%) compared to other periods (14% for 2020–2023 and 12% for 2004–2016) (Figure 15). Statistically significant differences were found between the time periods (Q = 8.69; D.F. = 2; p = 0.001), with high heterogeneity (I² = 89% for 2017–2019, I² = 81% for 2020–2023, and I² = 89% for 2004–2016). The variation in prevalence could reflect improvements in detection methods over time, but further investigation is needed to understand the driving factors.

For the animal species analysis, bovines had the highest pooled prevalence (16%), followed by ovines (11%) and caprines (8%) (Figure 16). The heterogeneity was highest in ovine (I² = 93%), followed by bovines (I² = 78%) and caprines (I² = 71%). The significant subgroup effect (Q = 33.3; D.F. = 2; p < 0.001) indicates that species differences play a key role in prevalence rates, and specific factors, such as immune responses and management practices, likely contribute to this variability.

The age category subanalysis showed more variability in adults (I² = 91%) than in young animals (I² = 83%). The pooled prevalence for young animals (15%) was slightly higher than for adults (14%) (Figure 17), suggesting that age-related differences might influence susceptibility to cryptosporidiosis. These findings emphasize the need for further exploration of factors such as species, age, and time period to better understand cryptosporidiosis prevalence and improve control strategies.

Uni-variable meta-regression was conducted with animal species (goat, sheep and cattle), and region, sample size (continuous variable) and diagnosis method used as categorical variables using the mixed effect model, in case of toxoplasmosis (Table 3). These variables were subjected to assessment to see a linear relationship with the dependent variable that is the effect sizes. Those variables with a p-values < 0.1 were used in the multivariable meta-regression analysis. Only region and diagnostic methods were found to be significant at p-values < 0.1 and therefore were included in the final multivariable analysis. The multivariable meta-regression model as a whole explained 21.96% amount of heterogeneity, while using diagnostic method and regional variability. Even when controlling for diagnostic methods compared to antibody detections (Table 5).

In case of cryptosporidiosis meta-regression was performed based on animal species (bovine, ovne and caprine), publication year (2004–2016, 2017–2019, and 2020–2023) and age category of the animals (Adult and Yung). Unfortunately, all the moderators were found to be significant at p-values < 0.1 at univariable meta-regression. Therefore, all of the three variables were included in multivariable meta-regression analysis (Table 6).

Subgroup analysis of bovine cysticercosis and cystic echinococcosis based on study region and publication year was performed. The results of the subanalysis, such as the heterogeneity (I²), proportion, 95% CI, and test of between-study heterogeneity (Q), are summarized in Table 7. The subgroup difference test suggested that there was a statistically significant subgroup effect (Q = 11.69; DF = 1; p = 0.0029), implying that the publication year category was significantly associated with the prevalence of bovine cysticercosis in cattle, whereas the subgroup difference test suggested that there was a statistically significant subgroup effect (Q = 11.69; DF = 1; p = 0.0029) in the study region and that the prevalence of cystic echinococcosis (Q = 6.67; df = 4; p = 0.023).

In this study, C. bovis and hydatid cysts were identified as the most important foodborne parasites at post-mortem examination at the slaughterhouse. The main organs affected by these parasites include the lungs, heart, liver, kidneys, and shoulder and masseter muscles. The included studies were categorized as (i) articles examining both parasites simultaneously in the same organ, (ii) studies treating only C. bovis in a single organ, and (iii) articles addressing only the distribution of cystic echinococcosis in a single organ (kidney).

In the present study, the lungs, heart and liver were affected by both C. bovis and cystic echinococcosis (same study). The overall pooled prevalence rates of both C. bovis co-infections with cystic echinococcosis of the lungs, liver and heart were 29, 20 and 8%, respectively (Figures 18–20). In the case of the lung, as shown in Table 6, a subanalysis was performed on the basis of the categorization of studies under C. bovis (eight articles) and cystic echinococcosis (only three articles), and the pooled prevalence was found to be 51 and 1%, respectively. Similarly, subanalyses of the included studies were based on parasite type, and the pooled prevalence rates of C. bovis and hepatic hydatosis were 8 and 36%, respectively (Table 6). In addition, subgroup analysis of studies based on parasite type revealed that the pooled prevalence of C. bovie with cystic echinococcosis in the heart was 14 and 3%, respectively (Table 6).

In our study, the kidney was the only organ affected only by cystic echinococcosis. Among a total of 37 studies, six reported infection of the kidney by cystic echinococcosis. The pooled proportion of cystic echinococcosis in the kidney was 2% (95% CI; 1–6%), with high inter-study heterogeneity (I² = 96%; p < 0.01) (Supplementary material 3).

In this meta-analysis, the highest levels of C. bovis were observed in the shoulder muscle, and the tongue was the second most susceptible organ. Regarding the anatomical distribution of the cysts, 32% of the affected organs were located in the shoulder region, with minimal heterogeneity observed in this proportion (Cochran’s Q = 11.21 and p value of 0.0474, I² statistic of 55%). The deltoid, 21% (I² = 83%; Q = 53.32, p = 0.000) of the tongue, 19% (I² = 66.3%, Q = 26, 72, p = 0.0016) and 6% (I² = 86.3%, Q = 46.52, p = 0.0001) of the tongue (Supplementary materials 4a–d). Among the organs examined, the diaphragm was least affected by C. bovis, as summarized in Table 8.