Characteristics of included studies

The initial database search yielded a total of 12,875 articles (Fig. 1). After screening and excluding duplicates, 131 articles (91 case–control and 40 cross-sectional studies) were found to be eligible and are included in this systematic review and meta-analysis. There were 91 case–control studies involving 18,626 cases and 18,312 controls (Tables 1 and 2), and 40 cross-sectional studies with 19,994 participants (Tables 4 and 5).

Fig. 1

Flow diagram of the study design process

The odds ratio/pooled prevalence

The analysis based on case–control studies found a significant association between protozoan parasites and mental disorders (OR: 2.059, 95% CI 1.830–2.317) (Fig. 2, Table 2). The random-effects model for cross-sectional studies showed that the overall prevalence of protozoan parasites in patients with mental disorders was 0.252 (95% CI 0.189–0.320) (Fig. 3, Table 5). The heterogeneity was significant for both case–control (I2 = 74%; τ2 = 0.247; p < 0.001) and cross-sectional studies (I2 = 98%; τ2 = 0.063; p < 0.001) (Tables 2 and 5).

Fig. 2

Forest plot of odds rations for relationship between prevalence of protozoan parasites and mental disorders in case–control studies

Fig. 3

Forest plots for random-effects meta-analysis of the global prevalence of protozoan parasites among patients with mental health disorders based on cross-sectional studies (The boxes indicate the effect size of the studies (prevalence) and the whiskers indicate its confidence interval for corresponding effect size. There is no specific difference between white and black bars, only studies with a very narrow confidence interval are shown in white. In the case of diamonds, their size indicate the size of the effect, and their length indicate confidence intervals

Subgroup analysisThe odds ratio/pooled prevalence based on WHO regions

According to the WHO regions, our analyses of case–control studies revealed that the highest pooled OR was related to the Eastern Mediterranean Region (OR: 2.289, 95% CI 1.868–2.805) with heterogeneity (I2 = 79%; τ2 = 0.299; p < 0.001) (Table 2). The analyses of cross-sectional studies showed that protozoan parasitic infections were most prevalent in patients with mental disorders in the European Region (0.386, 95% CI 0.150–0.656) with heterogeneity (I2 = 98%; τ2 = 0.047; p < 0.001) (Table 5).

The odds ratio/pooled prevalence based on the type of the parasite

Subgroup analysis based on the type of protozoan parasite revealed the pooled OR of the higher risk of these parasites in patients with mental disorders in case–control studies (OR: 2.069, 95% CI 1.841–2.326) with heterogeneity (I2 = 73%; τ2 = 0.260; p < 0.001). However, the pooled OR of Cyclospora cayetanensis (4.719, 95% CI 1.352–16.474), followed by Cryptosporidium parvum (4.618, 95% CI 2.877–7.412) revealed a significantly higher risk of these parasites in patients with mental disorder compared to controls (Table 3).

The analysis based on cross-sectional studies showed that the pooled prevalence of different types of protozoan parasites was as follows: T. gondii (0.343, 95% CI 0.228–0.467) with heterogeneity (I2 = 99%; τ2 = 0.064; p < 0.001), Cryptosporidium spp. (0.087, 95% CI 0.005–0.242) with heterogeneity (I2 = 95%; τ2 = 0.032; p < 0.001), Blastocystis hominis (0.085, 95% CI 0.047–0.132) with heterogeneity (I2 = 93%; τ2 = 0.018; p < 0.001), Entamoeba coli (0.083, 95% CI 0.050–0.123) with heterogeneity (I2 = 92%; τ2 = 0.018; p < 0.001), Cystoisospora belli (0.076, 95% CI 0.022–0.156), Cyclospora cayetanensis (0.075, 95% CI 0.042–0.116), E. histolytica / dispar (0.064, 95% CI 0.011–0.151) with heterogeneity (I2 = 97%; τ2 = 0.067; p < 0.001), Giardia lamblia (0.062, 95% CI 0.040–0.088) with heterogeneity (I2 = 89%; τ2 = 0.008; p < 0.001), Dientamoeba fragilis (0.029, 95% CI 0.000–1.000) with heterogeneity (I2 = 95%; τ2 = 0.023; p < 0.001), Endolimax nana (0.029, 95% CI 0.006–0.067) with heterogeneity (I2 = 88%; τ2 = 0.017; p < 0.001), Iodamoeba butschlii (0.024, 95% CI 0.013–0.037) with heterogeneity (I2 = 53%; τ2 = 0.001; p < 0.001), Chilomastix mesnili (0.021, 95% CI 0.001–0.058) with heterogeneity (I2 = 77%; τ2 = 0.005; p < 0.001) (Table 6).

The odds ratio/pooled prevalence based on climatic variables

The estimates of pooled OR based on climatic variables showed that the highest rate was related to an annual precipitation range of 300–650 (OR: 2.297, 95% CI 1.700–3.103), humidity levels of 40–75% (OR: 2.194, 95% CI 1.917–2.510), annual rainfall of < 400 mm (OR: 2.256, 95% CI 1.844–2.761), and average temperatures of > 20 ℃ (OR: 2.281, 95% CI 1.850–2.811) (Table 2).

Moreover, our analyses of studies with cross-sectional design revealed that the highest pooled prevalence was observed for an annual precipitation of < 300 (0.270, 95% CI 0.192–0.355), humidity levels of < 40% (0.303, 95% CI 0.218–0.394), annual rainfall of < 400 mm (0.331, 95% CI 0.249–0.418), and the average temperatures of < 10 ℃ (0.448, 95% CI 0.398–0.499) (Table 5). The heterogeneity related to analyses based on each climatic variable in both study designs is presented in Tables 2 and 5.

The odds ratio/pooled prevalence based on diagnostic method

In terms of case–control studies, the highest rate of OR was related to studies that utilized a combination of EIA and Western blot methods (OR: 5.818, 95% CI 1.490–22.715) (Table 2).

Regarding cross-sectional studies, the highest pooled prevalence was associated with the combination of direct smear & PCR (0.708, 95% CI 0.618–0.783) (Table 5).

The odds ratio/pooled prevalence based on mean age and income level

The analyses of cross-sectional studies showed that protozoan parasites were most prevalent in patients with a mean age ranging from 15–20 years old (0.664, 95% CI 0.594–0.724) (Table 5).

According to our estimates regarding income level, the pooled OR (2.190, 95% CI 1.823–2.632) with heterogeneity (I2 = 77%; τ2 = 0.263; p < 0.001) and pooled prevalence (0.285, 95% CI 0.201–0.378) of protozoan parasitic infections was found to be highest in patients in lower-middle income regions with heterogeneity (I2 = 98%; τ2 = 0.060; p < 0.001) (Tables 2 and 5).

The odds ratio/pooled prevalence based on the type of the mental disorder

In terms of case–control studies, the association between different mental disorders and protozoan parasitic infections was as follows: neurodevelopmental disorders (OR: 2.485, 95% CI 1.413–4.368), bipolar and related disorders (OR: 2.163, 95% CI 1.436–3.260), schizophrenia spectrum and other psychotic disorders (OR: 2.088, 95% CI 1.807–2.412), obsessive–compulsive disorder (OCD) and related disorder (OR: 1.893, 95% CI 0.908–3.947), depressive disorder (OR: 1.876, 95% CI 1.311–2.683), neurocognitive disorders (OR: 1.534, 95% CI 0.899–2.616), and trauma and stressor related disorders (OR: 1.757, 95% CI 0.493–6.261) (Fig. 2).

Moreover, the analysis of cross-sectional studies revealed that the pooled prevalence based on the type of mental disorders was as follows: 0.341 (95% CI 0.244–0.446) in neurodevelopmental disorder with heterogeneity (I2 = 97%; τ2 = 0.051; p < 0.001), 0.321 (95% CI 0.000–0.995) in bipolar and related disorders with heterogeneity (I2 = 99%; τ2 = 0.399; p < 0.001), 0.213 (95% CI 0.127–0.314) in schizophrenia spectrum and other psychotic disorders with heterogeneity (I2 = 98%; τ2 = 0.062; p < 0.001), 0.134 (95% CI 0.100–0.171) in depressive disorder with heterogeneity (I2 = 67%; τ2 = 0.001; p < 0.001), 0.119 (95% CI 0.082–0.170) in trauma and stressor-related disorders with, and 0.083 (95% CI 0.049–0.137) in neurocognitive disorders (Fig. 3).

Publication bias

Substantial publication bias was detected in case–control studies, as indicated by Egger’s funnel plot (t = 1.03, p = 0.306) and Begg’s test (t = 1.02, p = 0.308). Additionally, the Doi plot test revealed no asymmetry (LFK index: 0.43) (Fig. 4). Substantial publication bias was detected in cross-sectional studies, as indicated by Egger’s funnel plot (t = 2.24, p = 0.025) and Begg’s test (t = 3.07, p = 0.004). Additionally, the Doi plot test revealed a major asymmetry (LFK index: 2.74) (Fig. 5).

Fig. 4

Egger's funnel plot and Begg's funnel plot to assess publication bias in studies evaluating of protozoan parasites among patients with mental health disorders based on case—control studies (Colored circles represent each study. The middle line is the effect size and the other two lines are the corresponding confidence ranges) (A, B). Doi plot for the global prevalence of intestinal protozoan parasites among patients with mental health disorders. A Luis Furuya -Kanamori (LFK) index 0.43 indicates no asymmetry (C)

Fig. 5

Egger's funnel plot and Begg's funnel plot to assess publication bias in studies evaluating of protozoan parasites among patients with mental health disorders based on cross-sectional studies (Colored circles represent each study. The middle line is the effect size and the other two lines are the corresponding confidence ranges) (A, B). Doi plot for the global prevalence of intestinal protozoan parasites among patients with mental health disorders. A Luis Furuya—Kanamori (LFK) index 2.74 indicates major asymmetry (C)

Meta-regression

The results of the meta-regression analysis demonstrated that among all moderators, only annual rainfall, significantly affected the OR and the prevalence in studies with estimates of (slop = 0.6507, p < 0.0001) for a cross-sectional design and (slop = 0.8326, p < 0.0001) for a case–control studies (Figs. 6 and 7).

Fig. 6

A meta-regression graph for the global prevalence of protozoan parasites among patients with mental health disorders based on annual rainfall (A), and year of publication (B) in cross-sectional studies. The pink line is the regression line, which was plotted based on the intercept and the slope of the regression model. The different color bubbles represent the countries under study and their sizes indicate the effect size of each study

Fig. 7

A meta-regression graph for the global prevalence of protozoan parasites among patients with mental health disorders based on annual rainfall (A), and year of publication (B) in case—control studies. The pink line is the regression line, which was plotted based on the intercept and the slope of the regression model. The different color bubbles represent the countries under study and their sizes indicate the effect size of each study

QGIS3 map

To demonstrate the prevalence of protozoan parasites in patients with mental disorders in various geographical locations of the world, a map was created using QGIS3 software (https://qgis.org/en/site/) based on the included cross-sectional and case–control studies (Figs. 8 and 9).

Fig. 8

The prevalence of protozoan parasites among patients with mental health disorders based on cross-sectional studies in different geographical regions of the world

Fig. 9

The prevalence of protozoan parasites among patients with mental health disorders based on case—control studies in different geographical regions of the world