A Controlled Prospective Study of Toxoplasma gondii Infection in Individuals With Schizophrenia: Beyond Seroprevalence

Abstract

Toxoplasma gondii (TG) infection has been reported to be more frequent in schizophrenia. The interaction of the lifelong persisting parasite with the host's immune system involves T-cell/interferon-gamma–induced degradation of tryptophan and provides a challenge to the host well beyond a possible role in the etiology of schizophrenia. The hypothesis we tested in this study was that TG infection may be more frequent (serofrequency) and/or more intense (serointensity) in patients with schizophrenia or major depression compared with psychiatrically healthy controls. In addition, these measures are associated with the clinical course. We did a cross-sectional, prospective investigation of individuals with schizophrenia (n = 277) and major depression (n = 465) admitted to our department (2002–2005) and of healthy controls (n = 214), with all groups adjusted for age and geographic home region. Serofrequency was comparable between the groups, but serointensity was significantly higher in the patients. In individuals with schizophrenia, serointensity was significantly positively associated with C-reactive protein levels and leukocyte counts, and first-episode patients yielded significantly higher serotiters. Immunomodulatory medication was associated with decreased serotiters. In addition, the route of infection appears to differ between patients and controls. Thus, our results support increased host responses to TG infection in the patients, as well as increased titers in first-episode patients with schizophrenia; this may relate to the shifted T-helper 1/2 status described in these patients. Therefore, we suggest that TG infection, particularly in individuals with schizophrenia, is an important environmental factor in the interaction between psychiatric vulnerability, genetic background, immunomodulation, and the neurotransmitter systems.

Introduction

Chronic infection with the intracellular parasite Toxoplasma gondii (TG) is more frequent in individuals with schizophrenia than in psychiatrically healthy controls, as indicated in several studies from different countries.^1,2 Furthermore, first-episode patients might differ from patients with recurrent or chronic course by having more frequent TG infection and/or a more intense immune response.^1,2 However, to date, the results are not equivocal,^1,2 with subjects generally characterized as “psychiatric patients” being shown to be more frequently affected than healthy controls or nonpsychiatric patients.^3–6 A study on well-characterized psychiatric patients with distinct diagnoses other than schizophrenia has not yet been published. Moreover, studies with relevant additional data, such as the interrelationship with psychiatric symptomatology and course of the disorder, are still lacking.

Briefly, TG infection in humans takes place when infectious microcysts, typically in affected undercooked and raw meat, are ingested or through contamination with infected cat faeces.⁷ Because the infection is ubiquitous, the probability of becoming infected increases with age, apart from any particular high-risk behavior, as described before.

When TG infects an organism, it invades various cells⁸ and persists intracellularly, including in neurons and glia.^9–11 The host organism is not able to eradicate the infection.⁷ However, immunocompetent hosts control the chronic infection with a T-lymphocyte–driven defense.¹² All immunologic mechanisms involved have not yet been unraveled, but it is known that interferon-gamma (IFN-γ) and the enzyme indoleamine 2,3-dioxygenase (IDO) play a role.^13–17 Activated T-helper cells secrete IFN-γ, which induces IDO. This enzyme degrades the tryptophan that is needed for the tachyzoitic phase of TG. Consequently, activated parasites die by tryptophan depletion.¹³ The tryptophan degradation products that accumulate via the kynurenine pathway¹⁸ may result in excess dopaminergic tone. Thus, the host defense system might produce a lack of serotonin and an accumulation of dopaminergic activity. Psychiatrically, this suggests depressive and psychotic syndromes.^19–22 Therefore, this parasitic chronic infection, which shifts between silent and microactivated states²³ in conjunction with the host defense system, presents an attractive theoretical schema for increased serofrequencies of this infection in psychiatric patients with affective and psychotic syndromes.

We hypothesized that TG infection might be more frequent and/or more intense in patients with schizophrenia and in patients with major depression compared with age-adjusted psychiatrically healthy controls. We rated severity of the symptoms and the course of the disorder. In addition, we analyzed general inflammatory measures, took a careful medication history, and queried the subjects specifically about behaviors associated with a greater risk of TG infection.

METHOD

All patients who were admitted to inpatient units of our department between 2002 and 2005 and who were diagnosed clinically with schizophrenia or major depression were analyzed for their serotiters to TG infection from blood drawn routinely at admission (including analysis of C-reactive protein [CRP] and leukocyte count). The patients were also rated by experienced psychiatrists (Wilms and Erdag) for the severity of their symptomatology using the German version of the Clinical Global Impression rating scale (CGI), item 1 (0–7)²⁴; this was then collapsed to an arbitrary rating scale with 1 for none or mild (1 or 3 in CGI), 2 for moderate (4 in CGI), and 3 for severe symptoms (5 and 6 in CGI) because items 0, 2, and 7 of CGI had not been marked. In addition, the patients were asked to complete a short questionnaire on risky eating habits (now or ever consuming raw or undercooked meat at least several times) and close and risky cat contacts (now or ever cat ownership, cat in the same household, playing closely with cats, cleaning cat litters). When a patient was discharged, the same experienced psychiatrists analyzed the patient's records of the admission and all records, if any, of prior admissions. Diagnoses were assigned according to International classification of Diseases, 10th Revision (ICD-10)²⁵ using ICD-10 checklists.²⁶ The course of the disorder for schizophrenia and depression was rated by an arbitrary rating scale, with 0 for first episode, 1 for recurrent course, and 2 for chronic course. In addition, patients with schizophrenia were rated for first episode, second episode, recurrent episodes, and chronic course. Patients with acute or chronic infectious or inflammatory processes otherwise specified and with any major medical problems involving the immune system were excluded. Nonpsychiatrically affected controls were recruited from the same geographic region as the patients who are admitted to our department, which provides regional sectorized psychiatric care. The controls had blood drawn for TG serology, and they completed the questionnaires on risk behavior. They were screened for psychiatric disorders and major medical problems as well as for present or recent history of infection or inflammation. This study was approved by the local Ethics Committee of the Faculty of Medicine at the Christian-Albrechts-University, Kiel. All participants gave informed consent to take part in the study.

Psychotropic medication use was categorized for statistical analysis as follows: regular vs on-demand use, antipsychotic medication by potency,²⁷ typical vs atypical, chemical structure, and all regular medication used by any patients by immunomodulatory potency according to human in vivo data (yes = amitriptylline,²⁸ carbamazepine,²⁹ clozapine,²⁸ lithium,^28,29 mirtazapine,³⁰ and olanzapine²⁸; no = haloperidol,²⁸ paroxetin,²⁸ and venlafaxin³⁰; and unknown/possible immunomodulation = amisulpride, flupenthixol, levomepromazine, melperone, perazine, perphenazine, pipamperone, promethazine, quetiapine, risperidone, and ziprasidone, with respect to antipsychotics).

TG serology was done by microbiologists (Däubener and Stoltenberg) using the internationally established and widely used Toxo-Spot IF (bioMerieux, Lyon, France) according to the instruction of the supplier. Conventionally, for the detection of acute TG infections, the titers are defined as follows: <1:16 negative, no infection; 1:16–1:256 positive immune reaction suggestive of chronic infection; >1:256 positive, immune reaction suggestive of recent infection. However, we were interested in differentiating low and high titers in our subjects on the basis of microbiological and immunological expertise. Thus, we collated the individual serotiters as follows: <1:16 no contact with TG; 1:16–1:64 TG contact, low-titer response; and ≥1:128 TG contact, intense individual antibody reaction, high-titer response.

For statistical analysis, we used cross-tables and correlational analyses (SPSS for Windows 8.0). In order to test whether certain additional variables had an impact on our results, we tested these variables by cross-tables and correlational analyses with our variables of interest. If there were no significant results, we assumed that the additional variables did not have a significant effect on our results. The level of significance (2 tailed) was set at P ≤ .05. No adjustment of the error probabilities for multiple testing was performed, because of the explorative nature of the study.

Results

Characteristics of the Patients and Controls

We included the following subjects in our study: 277 patients with schizophrenia (mean age 37.4 ± 12.0 years [range 18.7–92.1], of which 24% [n = 67] were older than 45 years); 465 with unipolar major depression (mean age 46.0 ± 15.5 years [range 18.2–92.3], of which 48% [n = 221] were older than 45 years); and 214 healthy controls (mean age 38.9 ± 13.3 years [range 18.4–72.7], of which 26% [n = 55] were older than 45 years). For statistical reasons, the larger the numbers per group, the more difficult the matching becomes for one variable between the groups. Therefore, it was not possible to age match all the groups in our study. However, in order to deal with microbiologically relevant age effects with respect to TG serology in the between-group analyses, we created subsets in each group (over 45 years of age, and 45 years of age and under) because 45 years of age was a turning point in TG serofrequency in our control group (figure 1). In the following sections, all results are reported separately for these 2 age groups.

Fig. 1.

Percentage of Toxoplasma gondii–Positive Subjects Among All Controls Depending on Age.

Gender and smoking habits did not significantly interfere with the data analyses to follow.

Data Analysis

With respect to serofrequency, the only significant difference between the diagnostic groups and controls was in the individuals older than 45 years of age for titers ≥128. As depicted in table 1, the frequency of TG infection (ie, titer <16 vs titer ≥16) did not differ significantly between the diagnostic groups and controls, whereas there was a significant group difference when the comparison was done by titer categories. There are significantly more individuals with serotiters ≥128 in both patient groups compared with the controls. Thus, the intensity of the antibody response, ie, serointensity, is higher in the patients than in the controls. The questionnaires on cat contact and eating habits show that patient groups and controls reported similar frequency of these risk behaviors (data not shown). However, when analyzing interrelationships between serofrequency and the items in the questionnaires, we found that serofrequency is significantly associated with the eating of raw meat in all patient groups, whereas in the controls there was an effect of close and risky cat contact (table 2).

Table 1.

Serofrequency and Serointensity Between Diagnostic Groups and Controls (age > 45 y)

Table 2.

Eating Habits in all Patients' Groups and Risky Cat Contact in the Healthy Controls

	Patients				Controls
	N	Raw Meat − Percentage	Raw Meat + Percentage	N	Cat Contact − Percentage	Cat Contact + Percentage
Titer < 16	322	54	46	120	47	53
Titer 16-64	235	38	62	72	43	57
Titer ≥ 128	67	43	57	11	9	91
Titer ≥ 16	302	39	61	83	39	61

Note: Chi-square test on eating habits by serofrequency in both patients' groups: P = .001 with significant interrelationships between consumption of raw meat and seropositivity. Chi-square test on risky cat contact by serofrequency in all healthy controls: P = .05 with significant interrelationships between risky cat contact and seropositivity.

Further analyses of the schizophrenia group revealed a significant positive correlation between the titer categories and the number of leukocytes (rho = 0.130, P = .043 by Spearman correlation analysis). Moreover, high titers were significantly associated with high CRP values (table 3). With respect to the clinical course, first-episode patients were significantly more likely to have high antibody titers. Among the high–antibody titer patients, there are significantly more with first episode or chronic course vs recurrent course (table 4). These effects were not seen in seronegative or low–antibody titer patients. Whereas age did not interfere with these results, the use of antipsychotic medication with immunomodulatory or putative immunomodulatory effects did play a role. Using such medication was significantly correlated with lower antibody titers (rho = −0.118, P = .024 by Spearman correlation analysis). Other characteristics of the medication status did not have significant effects in our analysis.

Table 3.

CRP Values and Titer Categories in the Group “Schizophrenia”

	N	CRP < 6 [mg/l] Percentage	CRP ≥ 6 [mg/l] Percentage
Titer < 16		147	88	12
Titer 16-64	75	77	23
Titer ≥ 128	22	64	36
Titer ≥ 16	97	74	26

Note: Chi-square test on interrelationships between titer categories and CRP values: P = .005, with significant interrelationships between high CRP values and titer ≥ 128.

Table 4.

Serointensity and Clinical Course in the Group “Schizophrenia”

Further analysis of the major depression group showed significant effects for medication only. As can be seen in table 5, TG seronegative patients were more likely to have had no antidepressants, whereas the medium antibody titer patients were most likely to have been medicated with antidepressants. In addition, medium antibody titer patients were more likely to have had tri/tetracyclic antidepressants (TCA), whereas seronegative individuals were evenly balanced in the use of TCAs vs selective serotonin reuptake inhibitors (SSRIs). The seropositive individuals had been administered TCAs significantly more often than SSRIs; monaminoxidase (MAO) inhibitors were not used. Immunomodulatory medication was more frequent in the medium antibody titer patients, as depicted in table 6. The use of antipsychotic and mood-stabilizing medication did not show any significant interrelationship with the TG antibody titer data.

Table 5.

Use of Antidepressants and Titer Category in the Group “Major Depression”

	N	No AD Percentage	SSRI Percentage	TCA Percentage
Titer < 16	221	61\53	22\51	17\33
Titer 16-64	192	46\35	21\43	33\55
Titer ≥ 128	50	60\12	12\6	28\12
Titer ≥ 16	242	49\47	19\49	32\67

Note: Chi-square test on the interrelationships between use of antidepressants and titer categories: P = .002 with significant interrelationships between seronegativity and no antidepressant as well as between seropositivity and the use of TCA's; percent figure preceding \ refers to the line and percent figure following \ refers to the column. No AD = no antidepressant medication, SSRI = selective serotonin reuptake inhibitor as antidepressant medication, TCA = tri-/tetracyclic antidepressant medication.

Table 6.

Use of Immunomodulatory Medication and Titer Category in the Group “Major Depression”

	N	IM− Percentage	IM+ Percentage	IM? Percentage
Titer < 16	222	61\56	11\33	28\43
Titer 16-64	192	44\35	22\56	34\45
Titer ≥ 128	50	48\10	18\12	34\12
Titer ≥ 16	242	45\44	22\67	34\57

Note: Chi-square test on the interrelationships between the use of immunomodulatory medication (IM− = no, IM+ = yes, IM? = questionable immunomodulation; for the individual drugs, please, see Materials and Methods) and the titer categories: P = .005 with significant interrelationships between the use of immunomodulatory medication and titer 16-64; percent figure preceding \ refers to the line and percent figure following \ refers to the column.

Discussion

We have presented the first controlled, age-adjusted, prospective, large study of TG infection in psychiatric patients with schizophrenia and major depression diagnosed according to ICD-10 and all recruited in the same geographic region and during the same period. We found that serofrequency (titer <1:16 vs titer ≥1:16) did not differ between groups but that serointensity (<1:16 vs 1:16–1:64 vs ≥1:128) did differ between patients and controls, with higher serointensity in the patients. The analyses of interrelationships within each group between the TG antibody titers and further immunologic or psychopathologic parameters revealed significant results for the schizophrenia group only. High antibody titers were associated with higher inflammatory responses and with first-episode status.

The unique meta-analysis by Torrey and coworkers³¹ of all published studies included only well-described and methodologically sound studies and calculated an odds ratio of average 2.73. There was no systematic effect of the size of the groups and the results. The number of studies with control groups matched by any means to the patient groups is low, ie, 4 of 23, of which 3 dealt with first-episode patients only. For these 4 studies, the odds ratios range between 2.19 and 5.45. We cannot explain these discrepancies between different studies. However, we suggest for consideration the following issue: matching of the comparison groups is important with respect to age and geographic region of residence (particularly urban vs rural) for the likelihood of acquiring the infection.³² We matched for both of these criteria, whereas none of the previously published studies did so. From our experience, controlling for age is essential. Slight differences in mean age, particularly disequilibrium between the microbiologically relevant age blocks, may shift the results. If we had compared the groups without adjusting for age, seropositives would have been significantly more frequent in the oldest group of patients with major depression. Thus, we think that studies that do not carefully adjust for age between groups must be considered with caution. Furthermore, studies that report “increased antibody levels” in patients must be carefully evaluated because increased antibody levels are not identical to increased frequency of seropositives.^33,34

Our study is the first that also systematically collected information on the putative routes of infection. Patients and controls did not differ with respect to the frequency of these risk factors, but they did differ with respect to the interrelationship between these risk factors and seropositivity: the eating of raw meat in the patients and risky cat contact in the controls were associated with higher seropositivity. Thus, if this association points to the infectious route, our patients might have eaten infected raw meat more often, and the controls might have been hygienically less careful in their contacts with infected cats. On the basis of our data, we cannot further analyze this issue. However, if there is particular risk behavior in psychiatric patients, it would be useful to know this and to specifically intervene.

We found that serotiters were significantly higher in the 2 patient groups compared with the controls in all individuals older than 45 years. Moreover, the patients with schizophrenia demonstrated significant interrelationships between the titer category (serointensity) and the course of the disorder as well as with inflammatory variables and the use of immunomodulatory medications. Our results are consistent with previous studies that found increased serotiters in first-episode patients with schizophrenia.^33,34 As mentioned before, increased serotiters are not necessarily in the range of the microbiologically high titers suggestive of recent infection and, therefore, might rather support the more intense antibody response of the host.³⁵ Consequently, the higher antibody titers in the first-episode patients, regardless of age, might be based on the special condition of the patients. This is further supported by the additional interrelationship between high antibody titers and increased CRP levels and numbers of leukocytes as markers of inflammatory activity. There is a vast literature on immunologic peculiarities in individuals with schizophrenia.^1,36–40 In the light of TG infection, these data are of special interest.^1,36 As mentioned previously, one basic host defense strategy is through proinflammatory T-helper lymphocytes, ie, T_H1 cells.^12,13 It was described that individuals with schizophrenia with acute symptomatology display increased T_H1-associated cytokine responses, returning to control levels during successful antipsychotic treatment.⁴¹ Therefore, an infection such as TG might recurrently induce a T_H1 response that modulates the serotonin and dopamine neurotransmitter systems, leading to respective psychotic symptomatology. Antipsychotic treatment that reorganizes these neurotransmitter systems reduces the symptoms by modulating this common final pathway. Immunomodulating antipsychotics^28,37,42 in addition might become involved in the host defense against this infection. We found that the use of immunomodulating antipsychotics was associated with lower TG titers. We suggest that the cytokine pattern of increased T_H1 response in patients with acute symptoms might be caused by the proinflammatory response to infections such as TG. To our knowledge, there is no published study that investigated TG infection and T_H1 immune variables simultaneously. Such a study will be necessary to determine whether TG infections in individuals with schizophrenia are associated with increased T_H1 activation; it might also provide an explanation for the diverse findings in schizophrenia on the overactivation of T_H1 or T_H2 cytokines (for review of the T_H2 hypothesis of schizophrenia see Schwarz³⁹).

We found high TG titers in first-episode patients and an association between high titers and increased CRP and leukocyte values, as well as an association between the use of immunomodulating antipsychotics and low titers. One reason could be the use of immunomodulating antipsychotics from the first episode onward. On the other hand, because the high titers were seen in both first-episode and chronic-course patients, continuously high titers might favor a chronic course. Furthermore, a strong proinflammatory immune response to properly control TG microcysts could be counterproductive. One can speculate whether modulating this intense immune response might improve the clinical course in the patients with high titers and chronic course, as well as prevent the high-titer, first-episode patients from developing a chronic course.

Even if our data cannot answer all the questions and speculations presented, they represent the first detailed data on several important variables, and they support continued research on these questions. If TG infection is really able to induce a proinflammatory immune response that leads to dysregulated serotonin and dopamine neurotransmitter systems, to clinically relevant psychiatric symptoms, and perhaps even to the precipitation of schizophrenia in vulnerable subjects,⁴³ then a specific and effective treatment of this infection, coupled with a better understanding of how this vulnerability is defined, will be mandatory.