Partial data from the unpublished study were published as a conference abstract in a journal supplement.27 The authors provided additional (raw) data and written permission to use these data in the meta-analysis (D. Cobia, PhD, written communication, April 9, 2020).
A random-effects model was used. Data markers represent standardized mean differences (SMDs), and horizontal lines represent 95% CIs. The diamond represents the pooled summary estimate, and the horizontal extremes of the diamond represent its 95% CI. The arrows pointing right on the forest plot represent the upper limit of the 95% CI exceeding 1.0. Data on overall heterogeneity are available in eTable 5A in the Supplement. Unpublished (raw) data were requested from the authors of several studies because the available published results were not usable for meta-analysis; these data were used in the calculations with written permission from the authors of the following studies: Cobia et al27 (D. Cobia, PhD, April 9, 2020), El-Hadidy et al28 (M. A. El-Hadidy, MD, written communication, May 11, 2020), Gajewski et al29 (P. D. Gajewski, PhD, written communication, April 3, 2020), Guenter et al30 (W. Guenter, PhD, written communication, April 30, 2020), and Torniainen-Holm et al32 (M. Torniainen-Holm, PhD, written communication, May 8, 2020).
A random-effects model was used. Data markers represent standardized mean differences (SMDs), and horizontal lines represent 95% CIs. The diamond represents the pooled summary estimate, and the horizontal extremes of the diamond represent its 95% CI. The arrow pointing right on the forest plot represents the upper limit of the 95% CI exceeding 1.0. Data on overall heterogeneity are available in eTable 5B in the Supplement. Unpublished (raw) data were requested from the authors of several studies because the available published results were not usable for meta-analysis; these data were used in the calculations with written permission from the authors of the following studies: Cobia et al27 (D. Cobia, PhD, written communication, April 9, 2020), Gajewski et al29 (P. D. Gajewski, PhD, April 3, 2020), and Guenter et al30 (W. Guenter, PhD, written communication, April 30, 2020).
A random-effects model was used. Data markers represent standardized mean differences (SMDs), and horizontal lines represent 95% CIs. The diamond represents the pooled summary estimate, and the horizontal extremes of the diamond represent its 95% CI. The arrow pointing right on the forest plot represents the upper limit of the 95% CI exceeding 1.0. Data on overall heterogeneity are available in eTable 5D in the Supplement. Unpublished (raw) data were requested from the authors of several studies because the available published results were not usable for meta-analysis; these data were used in the calculations with written permission from the authors of the following studies: Cobia et al27 (D. Cobia, PhD, written communication, April 9, 2020), El-Hadidy et al28 (M. A. El-Hadidy, MD, written communication, May 11, 2020), Gajewski et al29 (P. D. Gajewski, PhD, written communication, April 3, 2020), Guenter et al30 (W. Guenter, PhD, written communication, April 30, 2020), Nimgaonkar et al31 (V. J. Nimgaonkar, MD, PhD, written communication, May 12, 2020), and Torniainen-Holm et al32 (M. Torniainen-Holm, PhD, written communication, May 8, 2020).
eMethods 1. Search Strategy
eMethods 2. Aggregation of Results
eTable 1. Aggregation of Tests in Cognitive Domains
eTable 2. Insufficient Reported Outcome Measures
eTable 3. Newcastle-Ottawa Scale Quality Assessment: Male and Female Participants Combined
eTable 4. Newcastle-Ottawa Scale Quality Assessment: Male and Female Participants Separated
eTable 5. Results From Meta-analyses and Subgroup Analysis
eTable 6. Results From Meta-regression Analyses
eTable 7. Meta-regression Analysis Dichotomized by Mean Age
eFigure 1. Forest Plot for Short-term Verbal Memory
eFigure 2. Forest Plot for Processing Speed Without Ene et al Study (Low Quality)
eFigure 3. Forest Plot for Executive Functioning Without Ene et al Study (Low Quality)
eFigure 4. Forest Plot for Executive Functioning Without Nimgaonkar et al Study
eFigure 5. Funnel Plot for Processing Speed
eFigure 6. Funnel Plot for Working Memory
eFigure 7. Funnel Plot for Short-term Verbal Memory
eFigure 8. Funnel Plot for Executive Functioning
eFigure 9. Meta-regression Analysis of Processing Speed: Mean Age
eFigure 10. Meta-regression Analysis of Processing Speed: Study Quality
eFigure 11. Meta-regression Analysis of Processing Speed: Seropositivity Cutoff
eFigure 12. Meta-regression Analysis of Working Memory: Mean Age
eFigure 13. Meta-regression Analysis of Working Memory: Seropositivity Cutoff
eFigure 14. Meta-regression Analysis of Short-term Verbal Memory: Mean Age
eFigure 15. Meta-regression Analysis of Executive Functioning: Mean Age
eFigure 16. Meta-regression Analysis of Executive Functioning: Study Quality
eFigure 17. Meta-regression Analysis of Executive Functioning: Seropositivity Cutoff
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
de Haan L, Sutterland AL, Schotborgh JV, Schirmbeck F, de Haan L. Association of Toxoplasma gondii Seropositivity With Cognitive Function in Healthy People: A Systematic Review and Meta-analysis. JAMA Psychiatry. 2021;78(10):1103–1112. doi:10.1001/jamapsychiatry.2021.1590
Is Toxoplasma gondii seropositivity in otherwise healthy people associated with alterations in cognitive function?
In this systematic review and meta-analysis of 13 studies comprising 13 289 healthy individuals, a modest but significant association was observed between T gondii seropositivity and impaired performance on cognitive tests in all analyzed domains (processing speed, working memory, short-term verbal memory, and executive functioning).
This study’s findings suggest that, given the high global prevalence of T gondii infection, the consequences of these associated cognitive impairments for global mental health could be substantial.
The parasite Toxoplasma gondii has been associated with behavioral alterations and psychiatric disorders. Studies investigating neurocognition in people with T gondii infection have reported varying results. To systematically analyze these findings, a meta-analysis evaluating cognitive function in healthy people with and without T gondii seropositivity is needed.
To assess whether and to what extent T gondii seropositivity is associated with cognitive function in otherwise healthy people.
A systematic search was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline. A systematic search of PubMed, MEDLINE, Web of Science, PsycInfo, and Embase was performed to identify studies from database inception to June 7, 2019, that analyzed cognitive function among healthy participants with available data on T gondii seropositivity. Search terms included toxoplasmosis, neurotoxoplasmosis, Toxoplasma gondii, cognition disorder, neuropsychological, and psychomotor performance.
Studies that performed cognitive assessment and analyzed T gondii seroprevalence among otherwise healthy participants were included.
Data Extraction and Synthesis
Two researchers independently extracted data from published articles; if needed, authors were contacted to provide additional data. Quantitative syntheses were performed in predefined cognitive domains when 4 independent data sets per domain were available. Study quality, heterogeneity, and publication bias were assessed.
Main Outcomes and Measures
Performance on neuropsychological tests measuring cognitive function.
The systematic search yielded 1954 records. After removal of 533 duplicates, an additional 1363 records were excluded based on a review of titles and abstracts. A total of 58 full-text articles were assessed for eligibility (including reference list screening); 45 articles were excluded because they lacked important data or did not meet study inclusion or reference list criteria. The remaining 13 studies comprising 13 289 healthy participants (mean [SD] age, 46.7 [16.0] years; 6586 men [49.6%]) with and without T gondii seropositivity were included in the meta-analysis. Participants without T gondii seropositivity had favorable functioning in 4 cognitive domains: processing speed (standardized mean difference [SMD], 0.12; 95% CI, 0.05-0.19; P = .001), working memory (SMD, 0.16; 95% CI, 0.06-0.26; P = .002), short-term verbal memory (SMD, 0.18; 95% CI, 0.09-0.27; P < .001), and executive functioning (SMD, 0.15; 95% CI, 0.01-0.28; P = .03). A meta-regression analysis found a significant association between older age and executive functioning (Q = 6.17; P = .01). Little suggestion of publication bias was detected.
Conclusions and Relevance
The study’s findings suggested that T gondii seropositivity was associated with mild cognitive impairment in several cognitive domains. Although effect sizes were small, given the ubiquitous prevalence of this infection globally, the association with cognitive impairment could imply a considerable adverse effect at the population level. Further research is warranted to investigate the underlying mechanisms of this association.
Toxoplasma gondii is an intracellular parasite that produces quiescent infection in approximately 30% of humans worldwide. Toxoplasmosis prevalence varies depending on geographic location and increases with age. The course of the infection is usually asymptomatic or occasionally symptomatic with nonspecific symptoms.1 Reproduction of T gondii is possible in the intestines of felids only, but a wide range of intermediate hosts is known to carry the infection.2 Hosts acquire toxoplasmosis by ingesting oocysts or cysts of the parasite via contaminated water or food. The T gondii parasite is able to permeate the blood-brain barrier and can settle as a quiescent infection in muscle, brain, and liver tissue.3-6
Toxoplasmosis may have consequences for the behavior of intermediate hosts. Infected rodents have exhibited more risk-taking and impulsive behavior compared with uninfected rodents,7,8 and infected mice have exhibited impaired reaction times,9 reduced learning capacity,10 and decreased motor performance.11,12 However, these observations have not been consistently reported in all studies.13
Several observational studies have reported neurocognitive changes associated with toxoplasmosis in humans; however, effect sizes and directions varied.14-17 Meta-analyses have suggested an association between T gondii and neuropsychiatric disorders, particularly schizophrenia.18-20 Although meta-analytic findings have not found an association between T gondii and attention-deficit/hyperactivity disorder,21 1 recent case-control study did report an association.22 Furthermore, exposure to T gondii has been associated with an increased number of motor vehicle crashes and suicide attempts.23
To identify potential adverse consequences of T gondii seropositivity, it is important to evaluate whether the highly prevalent T gondii parasite is associated with cognition in humans and, if so, to examine which specific cognitive functions are impaired, to what extent they are impaired, and whether participant or study characteristics have consequences for any association between seropositivity and cognitive performance. The findings of the literature thus far have not conclusively answered these questions. Therefore, we conducted a systematic review and meta-analysis to examine whether T gondii seropositivity was associated with alterations in cognitive function among otherwise healthy people.
This systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guidelines.24,25 The study was registered on PROSPERO (CRD42020154860). A systematic search of PubMed, MEDLINE, Web of Science, PsycInfo, and Embase was performed by a trained researcher (A.S.) to identify studies from database inception to June 7, 2019, that analyzed cognitive function among healthy participants with available data on T gondii seropositivity. Search terms included toxoplasmosis, Toxoplasma gondii, cognition disorder, neuropsychological, and psychomotor performance (the full search strategy is available in eMethods 1 in the Supplement).
Titles, abstracts, and full-text articles were independently screened by 2 researchers (Lies H. and J.S.), and differences in study selection were resolved by consensus. Reference lists of all reviews pertaining to toxoplasmosis and brain involvement were hand searched for eligible studies. In cases of overlapping participants, the largest cohort was included. Studies were included if (1) they enrolled a group of healthy individuals among whom antibodies to T gondii were assessed, (2) neuropsychological tests were used to assess cognitive function, and (3) the means and SDs of the cognitive test scores were described, calculable, or made available by the authors. Studies were excluded if participants (1) were immunocompromised or had severe complications associated with toxoplasmosis, (2) had congenital toxoplasmosis, or (3) had a psychiatric disorder. Corresponding authors were approached to request additional data if needed.
The following data were collected: year of publication, study design, inclusion and exclusion criteria, sample size, age presented in means and SDs, male to female ratio, assay used for T gondii detection, IgG threshold for T gondii seropositivity, and cognitive outcomes presented in means and SDs. Both cross-sectional and longitudinal studies were included; for longitudinal studies, the baseline assessments were used. Some studies enrolled a healthy population sample, whereas in other studies, patients with psychiatric disorders were compared with healthy individuals. Only data concerning healthy participants were extracted.
Data (summary estimates) were extracted from published articles, and 2 researchers (Lies H. and J.S.) requested additional (raw) data from the corresponding authors of selected studies26-32 when the available published results were not usable for meta-analysis. Data provided were used in the meta-analysis with written permission from the authors of the following studies: Cobia et al27 (D. Cobia, PhD, written communication, April 9, 2020), El-Hadidy et al28 (M. A. El-Hadidy, MD, written communication, May 11, 2020), Gajewski et al29 (P. D. Gajewski, PhD, written communication, April 3, 2020), Guenter et al30 (W. Guenter, PhD, written communication, April 30, 2020), Nimgaonkar et al31 (V. J. Nimgaonkar, MD, PhD, written communication, May 12, 2020), and Torniainen-Holm et al32 (M. Torniainen-Holm, PhD, written communication, May 8, 2020). Partial data from the Cobia et al27 study were published as a conference abstract in a journal supplement. Unpublished data from Cobia et al27 and other studies28-32 were used in the following calculations: mean age; male to female ratio; seropositivity to seronegativity ratio; overall SMD of processing speed, working memory, short-term verbal memory, and executive functioning; and corresponding sensitivity analyses, meta-regression analyses, and funnel plots.
Cognitive tests were classified into corresponding cognitive domains according to the Compendium of Neuropsychological Tests: Administration, Norms, and Commentary of Strauss et al,33 Neuropsychological Assessment by Lezak et al,34 and the MATRICS Consensus Cognitive Battery part 135 (eMethods 2 and eTable 1 in the Supplement). When a study used multiple neuropsychological tests pertaining to the same cognitive domain, we included only the data from the test that was applied most frequently across the included studies to minimize heterogeneity.
We calculated standardized mean differences (SMDs) by dividing mean differences by SDs. Effect directions of all available tests were adjusted so that a positive SMD represented favorable cognitive performance. The significance threshold was set at 2-sided P = .05.
The quality of the included studies was assessed by 2 researchers (Lies H. and J.S.) using the case-control form of the Newcastle-Ottawa Scale for quality assessment (score range, 0-9 stars, with 0 indicating lowest quality and 9 indicating highest quality).36 Differences in judgment of quality were discussed with a third researcher (A.S.) and resolved by consensus.
A random-effects model was applied because heterogeneity was expected. A minimum of 4 studies was required per domain to allow data to be pooled for the meta-analysis. An overview of outcome measures that could not be pooled in the meta-analysis because they were reported in an insufficient number of studies is available in eTable 2 in the Supplement. Publication bias was assessed using the Egger test (with a significance threshold of 1-sided P < .10) and visual inspection of funnel plots. The Duval and Tweedie trim-and-fill method37 was applied if suggestions of publication bias were detected.
The I2 statistic was used to measure heterogeneity.38 If heterogeneity was detected, the following moderating factors were assessed: mean age, sex, type of test, cutoff values for seropositivity, study sample (population sample or healthy control group), and study quality. Subgroup analyses (for categorical variables) or meta-regression analyses (for continuous variables) were performed to assess the variable’s consequences for heterogeneity. Comprehensive Meta-analysis, version 3 was used for the analyses.39
The systematic search yielded 1954 total records. After removal of 533 duplicates, an additional 1363 records were excluded based on a review of titles and abstracts. A total of 58 full-text articles were assessed for eligibility (including reference list screening); 45 articles were excluded because they lacked important data or did not meet study inclusion or reference list criteria. Two of those articles26,40 reported outcomes in which fewer than 4 studies were available and were therefore excluded. Thirteen studies16,27-32,41-46 met eligibility criteria and were included in the meta-analysis (Figure 1).
The 13 studies16,27-32,41-46 included in the meta-analysis comprised 13 289 healthy participants (mean [SD] age, 46.7 [16.0] years based on 8 studies29,31,32,41-44,46 reporting complete age data; 6586 men [49.6%]); of those, 3006 participants (22.6%) had antibodies against T gondii (Table). All studies used the enzyme-linked immunosorbent assay to detect T gondii antibodies.
Data eligible for meta-analysis comprised 4 cognitive domains: processing speed, working memory, short-term verbal memory, and executive functioning (eMethods 2 in the Supplement). None of the included studies received fewer than 5 stars during the quality assessment. Eleven studies controlled for 1 or more important factor. All 13 studies16,27-32,41-46 used independent validation of cases and ascertained outcomes using a validated test. Detailed results of the quality assessment are available in eTable 3 and eTable 4 in the Supplement.
A total of 9 studies27-30,32,41,42,45,46 reported data on processing speed that could be aggregated. In these studies, processing speed was measured by the Trail Making Test Part A, the Serial Reaction Time Test, or go/no-go reaction time tests. An overall SMD of 0.12 (95% CI, 0.05-0.19; P = .001) was found, indicating a small but statistically significant association between T gondii seropositivity and decreased processing speed (Figure 2; eTable 5A in the Supplement). The Egger test (intercept, −0.40; P = .21) and funnel plot detected no publication bias (eFigure 5 in the Supplement). Heterogeneity was low (I2 = 13%; P = .32). No significant consequences for heterogeneity by the available moderating factors were found (eFigures 9-11 in the Supplement). A sensitivity analysis, in which the study with the lowest quality27 was excluded, did not significantly change the results (SMD, 0.13; 95% CI, 0.07-0.19; P < .001) (eFigure 2 in the Supplement).
Six studies16,27,29,30,43,44 provided results from the Wechsler Adult Intelligence Scale or the Wechsler Intelligence Scale for Children digit span test measuring working memory. An overall SMD of 0.16 (95% CI, 0.06-0.26; P = .002) was found, which represented a small but significant association between T gondii seropositivity and working memory impairment (Figure 3; eTable 5B in the Supplement). The Egger test (intercept, −0.01; P = .50) and funnel plot did not detect any publication bias (eFigure 6 in the Supplement). There was no evidence of heterogeneity (I2 = 0%), and all included studies were of fair quality. Therefore, no sensitivity analyses were performed, and meta-regression analyses revealed no significant change in results (eFigure 12 and eFigure 13 in the Supplement).
Five studies16,29,32,42,43 provided results from the Auditory Verbal Learning Test, the California Verbal Learning Test, or the Verbal Learning and Memory Test measuring short-term verbal memory. A statistically significant association was found between T gondii seropositivity and verbal memory impairment, with an SMD of 0.18 (95% CI, 0.09-0.27; P < .001) (eFigure 1 and eTable 5C in the Supplement). The Egger test (intercept, 1.10; P = .09) and funnel plot suggested some publication bias (eFigure 7 in the Supplement). The Duval and Tweedie trim-and-fill method imputed 1 study to the right of the mean. This result suggested some publication bias, causing the unadjusted effect size to appear more convincing compared with the real effect size. However, the adjusted effect size remained significant (SMD, 0.17; 95% CI, 0.06-0.29). Heterogeneity was moderate (I2 = 22%; P = .28), and an analysis of moderating factors did not provide an explanation for this heterogeneity (eFigure 14 in the Supplement).
Eight studies16,27-32,42 reported results from the Trail Making Test Part B, verbal fluency tests, or clock drawing tests measuring executive functioning. An association between T gondii seropositivity and worse executive functioning was observed, with an SMD of 0.15 (95% CI, 0.01-0.28; P = .03) (Figure 4; eTable 5D in the Supplement). Both the funnel plot and the Egger test (intercept, −0.10; P = .46) did not detect publication bias (eFigure 8 in the Supplement). Significant heterogeneity was found (I2 = 63%; P = .008). An analysis of moderating factors found no significant associations between heterogeneity and study design, sex, type of test, cutoff values, or study quality (eFigures 4, 16, and 17 in the Supplement). A sensitivity analysis, in which the study with the lowest quality27 was excluded, did not substantially change the findings (SMD, 0.15; 95% CI, 0.004-0.29; P = .04) (eFigure 3 in the Supplement).29 A meta-regression analysis of mean age revealed significantly greater effect sizes as age increased [Q = 6.17; R2 = 81%; P = .01) (eTable 6 and eFigure 15 in the Supplement).
We conducted 1 post hoc sensitivity analysis that excluded data from Nimgaonkar et al31 because this study reported that some participants may have provided their blood samples 1 year after study entry (when baseline cognition had already been assessed), which presented a risk of bias. This analysis did not substantially change the results (SMD, 0.11; 95% CI, 0.03-0.18; P = .005) (eFigure 4 in the Supplement).
To our knowledge, this meta-analysis is the first to evaluate the association between T gondii seroprevalence and cognitive performance in otherwise healthy individuals. The results suggested that T gondii seropositivity was significantly associated with worse cognitive function in all analyzed domains, which comprised processing speed, working memory, verbal short-term memory, and executive functioning. Although the extent of the associations was modest, the ubiquitous prevalence of the quiescent infection worldwide (approximately 30%)1 suggests that the consequences for cognitive function of the population as a whole may be substantial, although it is difficult to quantify the global impact.
Infection with T gondii has thus far been associated with several behavioral changes and psychiatric disorders.18-20,23 Furthermore, a recent meta-analysis47 found a marginally significant association between T gondii seroprevalence and Alzheimer disease. These findings have mainly been associated with the manipulation hypothesis, which asserts that the T gondii parasite invades the brain of warm-blooded animals and alters their behavior to render them easier prey for felids, thereby increasing the likelihood of the parasite completing its life cycle.7-12 The findings of the current meta-analysis are consistent with this hypothesis, as worse cognitive function was observed in otherwise healthy people with T gondii seropositivity. However, because these associations were largely derived from cross-sectional studies, it remains uncertain whether cognitive impairment may be causally linked to T gondii infection. Reverse causation (ie, people with more cognitive or psychiatric problems acquire infection at a higher rate) or another factor (eg, poverty) that may increase the likelihood of both T gondii infection and cognitive or psychiatric problems cannot be excluded. Nevertheless, the association between T gondii infection and cognitive impairment in otherwise healthy people and the potential for increased morbidity and mortality rates among humans is concerning.18-20,23 It is therefore important to further explore this association by examining consistency, temporality, dose response, experimental association, and the possibility of a biological explanation.48
The findings of this meta-analysis suggested that T gondii seropositivity was consistently associated with worse cognitive function in several domains. Moreover, there was little suggestion of publication bias. As expected, some heterogeneity between studies was found. A meta-regression analysis of executive functioning by mean age revealed a positive association between mean age and the extent of cognitive function deficits, which may suggest that cognitive impairment increases as exposure increases. Dichotomizing for mean age did not yield any new insights (eTable 7 in the Supplement).
A temporal association could not be inferred from the present study’s findings. However, 1 longitudinal study reported that T gondii seropositivity preceded more rapid decreases in executive functioning and Mini-Mental State Examination scores.31
Although the present study was not able to evaluate whether serointensity had potential implications for the extent of association between T gondii infection and cognition, an association between T gondii seropositivity and cognitive impairment among adults was noted in a previous study, and anxiety and depression in pregnant women were observed as anti–T gondii titers increased.49-51 However, 1 study reported a more substantial association between T gondii seropositivity and reaction time among participants with lower titers, and another study reported no association with anxiety and depression.52,53 Evaluating the association of serointensity with the severity of cognitive impairment is hampered by varying or unknown cutoff plasma IgG values for T gondii seropositivity. In the sensitivity analysis, the cutoff values of the different studies did not alter the overall results (eTable 7 in the Supplement).
Studies of rodents found inhibition of neuronal functioning, suggesting neuronal pathogenicity of the T gondii parasite.2,54,55 Infection with T gondii was followed by behavioral and cognitive changes, including impaired reaction time, motor performance, memory, and learning.7-12 However, the consistency and specificity of these findings have been debated.13,56
Several neurobiological mechanisms underlying the association between T gondii seropositivity and cognition may provide a plausible explanation at a biological level. Mice with T gondii infection have exhibited increased dopamine release.57-59 Moreover, T gondii infection has been reported to be associated with increased dopaminergic release in neurons in vitro.59,60 A critical range of dopamine turnover for optimal cognitive function has been found, with excessive dopamine turnover being associated with cognitive impairment.61-63 Dysregulated dopamine may also have implications for neuronal plasticity in the hippocampus, a brain structure that is indispensable for memory function and spatial orientation.64,65 Reduced memory capacity in otherwise healthy older adults with T gondii seroprevalence was reported.29,49
The tryptophan pathway is another possible mechanism by which T gondii seroprevalence could alter cognition. As a defense mechanism to infection with the T gondii parasite, hosts rapidly break down tryptophan, which is necessary for duplication of the parasite; this breakdown produces increased levels of kynurenine and quinolinic acid.66 Increased levels of these substances could alter neurotransmitter signaling, and seropositivity and higher levels of kynurenic acid have been associated with suicide attempts.67-69 Higher rates of neurotoxic effects and impulsive behavior associated with increased levels of dopamine, kynurenine, and quinolinic acid together have been reported.70 In addition, the T gondii parasite enters most hosts through the gastrointestinal tract and initiates a proinflammatory immune response that activates complement component 1q, which also functions in the brain as a selective marker for clearance of excessive synapses. It is conceivable that increased expression of complement component 1q is associated with excessive clearance of synapses.71
Infection with T gondii has also been associated with the presence of dysbiotic intestinal flora in mice,72 which could increase the permeability of the gut-blood barrier; this increased permeability has been associated with psychiatric disorders and altered behavior.73,74 In the central nervous system of mice, T gondii initiates microglial activation and an increased number of leukocytes interacting with the cerebral capillary endothelium. Infected mice were less capable of vasodilatation compared with uninfected mice, suggesting vascular dysfunction.75 The eventual consequences of neurotransmitter disruption (direct implications) or neurodegeneration (indirect implications) may include cognitive impairment and susceptibility to psychiatric disorders.41,76,77
Further research is needed given the substantial prevalence of T gondii infection worldwide and the consistent modest association between T gondii seropositivity and less favorable cognitive function. Future studies may examine potential moderating factors, such as socioeconomic status, strain type, serointensity, possible coinfections, and duration of infection. Moreover, development of improved laboratory methods is needed not only to distinguish strain types78 but also to detect bradyzoites and tachyzoites to assess the stage of infection and measure T-cell–based immunity.
More studies of individuals from non-Western countries and more longitudinal studies are needed to investigate causality and pathophysiological mechanisms. It may also be valuable to perform studies that focus on the improvement of anti-Toxoplasma medications and the development of a vaccine because current treatments for complicated T gondii infection have been associated with several adverse effects, limited therapeutic benefits, and drug resistance.29,79 Further research into alternative antiparasitic drugs is warranted.
This study has strengths. The aggregation of neuropsychological test results into predefined cognitive domains reduced the risk of chance findings. Furthermore, most of the included studies were of fair to high quality, suggesting a low risk of bias, and the evidence of publication bias was low.
This study also has limitations. First, the study used only cross-sectional data; therefore, causality could not be established. The findings did not allow us to examine whether the association between T gondii seropositivity and cognition was altered by prolonged exposure to the parasite, recent exposure in which the IgG immune response just began, or reverse causation.
Second, heterogeneity could be only partially explained. Some potential moderating factors, such as the strain of T gondii and genetic variation in hosts, could not be evaluated. Toxoplasma strain type 1 and strain types with divergent alleles appear to be the most virulent and might underlie the association with schizophrenia, depression, or anxiety.51,78,80 Positivity for the RHD gene may offer protection against Toxoplasma-induced motor performance impairment in humans.45,78,81 Some polymorphisms of the MMP-9 gene have been associated with the development of neurological disorders. One particular polymorphism of MMP-9 has been found more frequently among patients with schizophrenia and T gondii seropositivity.82 Many other genes are reported to be involved in susceptibility to infection or immune response to T gondii.83 The studies included in the current meta-analysis did not examine strain types or genetic factors among participants with seropositivity, hindering exploration of these associations.
Third, coinfection is another possible moderating factor that could not be evaluated. The presence of multiple neurotropic infections has been associated with schizophrenia, presumably because coinfections may moderate the immune response more substantially.84,85 Other potential moderating factors that could not be evaluated were serointensity and socioeconomic status. In addition, most included studies were performed in Western countries and comprised a predominantly White study population. Our findings are therefore not necessarily representative of the global population.
Quiescent infection with T gondii is highly prevalent and may be associated with cognitive function, including processing speed, working and verbal memory, and executive functioning, in healthy people. Based on the present findings and those of previous meta-analyses18-20,23 examining the association of T gondii seropositivity with motor vehicle crashes, suicide attempts, and the prevalence of psychiatric disorders, public health programs to prevent T gondii infection are warranted. These programs might, at a minimum, consist of hygienic measures, especially after human contact with contaminated sources.56 Hygienic measures are often already undertaken to prevent other infectious diseases. However, these measures are not sufficient to prevent quiescent T gondii infection. It may be wise to consider further research into the development of a vaccine against infection with T gondii in either humans or felids.29
Accepted for Publication: May 6, 2021.
Published Online: July 14, 2021. doi:10.1001/jamapsychiatry.2021.1590
Corresponding Author: Arjen L. Sutterland, MD, Department of Psychiatry, Amsterdam UMC, University of Amsterdam, Meibergdreef 5, 1105 AZ Amsterdam, the Netherlands (email@example.com).
Author Contributions: Dr Lies de Haan had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Lies de Haan, Sutterland, Schotborgh, Lieuwe de Haan.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Lies de Haan, Sutterland, Lieuwe de Haan.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Lies de Haan, Sutterland, Schirmbeck.
Administrative, technical, or material support: Lies de Haan, Schotborgh.
Supervision: Sutterland, Schirmbeck, Lieuwe de Haan.
Conflict of Interest Disclosures: None reported.
Additional Contributions: We thank Cynthia P. Wyman, MS, and Derin Cobia, PhD (Brigham Young University), Mohamed A. El-Hadidy, MD (Mansoura University [Egypt]), Patrick D. Gajewski, PhD (Leibniz Research Centre for Working Environment and Human Factors [Germany]), Wojciech Guenter, PhD (Nicolaus Copernicus University [Poland]), Vishwajit J. Nimgaonkar, MD, PhD (University of Pittsburgh School of Medicine), and Minna Torniainen-Holm, PhD (National Institute for Health and Welfare [Finland]). These authors shared their unpublished data or provided us with additional data on their published studies,26-32 thereby contributing to a more reliable meta-analysis.