Studies of T. gondii Antibodies In Schizophrenia

Antibody studies provide among the strongest evidence linking T. gondii   to schizophrenia. This is in spite of the fact that the standard assays currently in use to measure T. gondii antibodies are far from perfect. Studies in our laboratory have suggested that the standard assays may miss many Toxoplasma-infected cases i.e., that many cases that are reported as negative are really positive if a more sensitive assay was available. There is also evidence that T. gondii antibody titres decrease over time, as discussed below.

A recent and promising advance in our laboratory has been the development of an antibody that targets the bradyzoite cyst form of T. gondii, the MAG-1 antibody. The usual assay to detect T. gondii uses an antibody against the tachyzoite, the actively replicating form of the parasite.  However it is the cyst form in the brain that is thought to be the source of most psychiatric symptoms.  In our laboratory the MAG-1 assay in rodent studies has been shown to strongly correlate with the number of brain cysts in the infected host.29

The first study of antibodies against T. gondii  carried out on individuals with psychoses was done by Kozar in Poland in 1953 using a skin test. This was followed by other studies in East Germany in 1956, Czechoslovakia in 1957, Bulgaria in 1962, and Russia in 1962. In the early 1980s, Chinese researchers became aware of this research and subsequently became the leading researchers on the prevalence of T. gondii  antibodies in schizophrenia.

Since Kozar’s original study, there have been approximately 100 additional studies. A 2007 review of 42 of these studies included a meta-analysis of 23 of them in which the odds ratio of having T. gondii  antibodies with a diagnosis of schizophrenia was OR 2.73. In other words, if a person has been infected with T. gondii, he/she has a 2.7 times greater chance of having schizophrenia than if the person had not been infected.1

A 2012 updated meta-analysis of 38 such studies included 6,067 individuals with schizophrenia and 8,715 controls. The meta-analysis also reported an odds ratio of OR 2.71 (1.93–3.80).2 These studies were done in 14 different countries; thus, the finding of increased antibodies against T. gondii  in individuals with schizophrenia has been remarkably consistent geographically and over half a century.

A 2015 meta-analysis of 42 studies of individuals with schizophrenia, with considerable overlap of previous meta-analysis, reported and odds ratio of OR 1.81 (1.51-2.16) (p<0.0001).3  After controlling for publication bias the OR (1.43) remained significant.  These researchers also examined the 12 studies which assessed T.gondii serointensity and reported that this measure was also significantly associated with schizophrenia (OR 1.85, 1.11-3.10; P = 0.019).

The above meta-analyses focused mostly on IgG antibodies to T. gondii.  In 2015 a meta-analysis was published looking at IgM antibodies in studies of acute psychosis.  Sixteen studies with 2,353 patients and 1,707 controls were included.  There was a significant association of T. gondii IgM antibodies and acute psychosis (OR 1.68, 1.23-2.27, p = 0.001).4

There is some clear evidence from these antibody studies that the increase in antibodies is not secondary to anti-psychotic medication. The study by Leweke et al. in Germany assessed antibodies to T. gondii  in 36 individuals with schizophrenia who had never been treated, 10 who had had past treatment, and 39 receiving current treatment. The level of both serum and CSF antibodies to T. gondii was highest in the never-treated patients, intermediate in those treated in the past, and lowest in those receiving current treatment.5

In another interesting study using German patients with schizophrenia who had never been treated, it was reported that those individuals who had antibodies to T. gondii, compared to those who did not have antibodies, were more likely to have selected inflammatory markers (interleukin-6 receptor; transforming growth factor alpha; angiotensin converting enzyme).

In another interesting study involving a community study of 5,906 participants in Finland, seropositivity to T. gondii was significantly associated with self-reported psychotic-like symptoms (from the CIDI) [OR 1.77, p = 0.001). However, an association with diagnosed schizophrenia did not reach statistical significance [OR 1.45].26

Overall, of the approximately 100 studies done to date, approximately 80 of them reported a positive association between either having antibodies to T. gondii or having a higher titre of such antibodies and having a diagnosis of schizophrenia or related condition. The largest negative study published to date included 368 affected adults (295 schizophrenia; 58 schizoaffective; 15 schizophreniform) and 282 controls solicited by mail. The authors concluded that since their study was negative then “publication bias and the influence of confounders” likely would explain the positive studies.7

One reason for the discrepant antibody results may be the length of time the study individuals have been sick.  A large study of 1,481 individuals with schizophrenia, BP disorder, and major depression and 571 matched controls reported that the individuals with significantly elevated antibodies to T. gondii had had a recent onset (within two years) of their psychosis.27  This finding is also consistent with studies suggesting that the level of IgG antibodies to T. gondii decreases over time; see “Do antibodies to T. gondii remain detectable over many years” below.

Studies of antibodies in individuals prior to the onset of schizophrenia

In a study of military personnel, serum specimens were available from periods of up to 11 years prior to the onset of schizophrenia. The serum of 180 individuals with schizophrenia and 532 matched (3:1) controls were assessed for IgG antibodies to T. gondii  and other infectious agents. Among those with schizophrenia, significantly increased levels of antibodies were seen prior to the onset of illness (hazard ratio = 1.24, p<0.01), maximal in the 6 months prior to onset but seen as early as 3 years prior to the onset.8 An attempt was made to replicate this finding, using the same methodology, in a larger military cohort of 855 individuals with schizophrenia and 1,165 controls. The results failed to replicate the original finding (hazard ratio – 1.06, p=0.27).9

In another study, antibodies against T. gondii  were assessed in 105 young individuals who were thought to be at “ultra-high risk” for developing schizophrenia because of their early symptoms and behavior. Among the 105, 18 had antibodies to T. gondii, and “being Toxoplasma-positive was significantly associated with more severe positive psychotic symptoms, and more severe psychiatric symptoms in general”.10  However, a study of a larger group of Dutch adolescents who were assessed for “psychic experiences” reported no correlation between antibodies to T. gondii  and self-reports of “psychic experiences”.11

Another study of antibodies in individuals prior to the onset of schizophrenia was carried out in Denmark using their national case register. Among 45,609 women who gave birth between 1992 and 1995 (at which time T. gondii  antibodies were assessed), 80 developed schizophrenia during the following 13–16 years, indicating a relative risk of 1.68 (0.77–3.46). “When the mothers were classified according to IgG level, only those with the highest IgG levels had a significantly higher risk of schizophrenia spectrum disorder”.12

A study in Wuhan, China, assessed antibodies to T. gondii  in 7,126 students entering Wuhan University for their first year. At the time of entering the university, 707 (9.9%) had IgG and 113 (1.6%) had IgM antibodies to T. gondii. The students were then followed for the next four years and all those who developed a psychotic disorder were noted. At the end of that time, 84 students (1.2%) had developed a psychotic disorder. Among those with psychosis, 18 (21.4%) had IgG antibodies and 9 (10.7%) had IgM antibodies to T. gondii  at the time they entered the university. Regarding T. gondii  as a risk factor for psychosis, 18 of the 707 (2.6%) students with IgG antibodies developed psychosis compared to 66 of the 6,419 (1.0%) students without IgG antibodies. Similarly, 9 of the 113 (8.0%) students with IgM antibodies developed psychosis compared to 104 of the 7,013 (1.5%) students without IgM antibodies (unpublished study).


Studies of T. gondii antibodies in the mothers of children who are later diagnosed with schizophrenia

Five such studies have been carried out, 3 of which used sera collected from the mother late in the third trimester and 2 of which used sera from the newborn in the first week of life.  Antibodies found in newborn children came from the mother, so all 5 studies were basically measuring the same thing.

  • A Danish study assessed newborn sera from neonatal blood spots in 71 individuals who later developed early-onset schizophrenia (onset prior to age 19); 115 who later developed schizophrenia-like disorders; 258 who later developed affective disorders; and 684 controls. The only group with elevated T. gondii antibodies compared to the controls was those with schizophrenia (p = .045; OR 1.8).13
  • A Swedish study assessed newborn sera from neonatal blood spots in 47 individuals who later developed schizophrenia; 151 who later developed non-affective psychoses (e.g. delusional disorders, schizotypal), and 524 controls. Those with schizophrenia, but not those with other non-affective psychoses, had significantly elevated T. gondii antibodies compared to controls (OR 2.0). The difference was greatest for the individuals with the highest antibody level (OR 3.2).14 A follow-up study reported that the infections were most common among neonates who had evidence of immune system dysfunction, as measured by their acute-phase proteins.15
  • An American study (Kaiser birth cohort) assessed third trimester sera in 63 individuals who later developed schizophrenia, schizoaffective disorder, delusional disorder, psychotic disorder NOS, and schizotypal personality disorder, and 123 matched controls. Among the schizophrenia spectrum group 40% were positive for T. gondii antibodies compared to 24% of the controls. The individuals in the schizophrenia spectrum group with the highest antibody titers, but not those with moderate titers, were significantly elevated compared to the controls (p= .05, OR 2.6).16
  • An American study (Collaborative Perinatal Project) assessed third trimester sera in 27 individuals who later developed schizophrenia (n =13) or other psychoses (psychoses NOS; brief psychosis; bipolar with psychotic features, schizophreniform), and 54 controls. The level of T. gondii antibodies did not differ between cases and controls.17
  • An American study (larger collection of Collaborative Perinatal Project, including the cases used by Buka et al) assessed third trimester sera in 219 individuals who developed schizophrenia (n = 106) or other psychoses including affective psychoses (n = 88), and 618 controls. T. gondii antibody was detected in 42% of the cases and 35% of the controls. The assessment was done by the strain of the T. gondii antibody. For Type I, the cases (n = 27) differed significantly from the controls (n = 44) ( p = .03; OR 1.9). For the other strains and the cases as a whole there were no significant differences from the controls.18


In conclusion, 4 out of the 5 studies reported significant elevated antibodies in the maternal and newborn sera of individuals who later developed strictly-defined schizophrenia but not affective psychoses.  The one negative study included 27 cases, of which only 13 had schizophrenia.

Do antibodies to T. gondii  remain detectable over many years?

It has been widely assumed that once a person is exposed to T. gondii  they will remain antibody-positive for life. However, no long-term study has been done on this question in humans. In 1941, Dr. Albert Sabin reported that in his experiments on monkeys, “it has been observed that convalescent monkeys may lose their antibodies as early as six weeks after infection”.19 Studies of rodents have also demonstrated waning immunity. One study noted that “in serially congenital transmission in mice, parasite detection by serology wanes despite the detectable presence of parasites” [by PCR] In fact, the waning of immunity after about 8 weeks following T. gondii infection in mice was noted as early as 1943.20  Similarly, research on mice in 1964 reported examples of mice definitely infected with T. gondii but seronegative.31 There are also suggestions from human studies that seropositivity is not lifelong; in one study, the mean duration of seropositivity was 40 years.21  There are at least two other studies of adult toxoplasmosis reporting conversions from T. gondii  seropositivity to seronegativity22-23 and one study showing a similar change in cases of congenital toxoplasmosis.24 Serial antibody studies from our own laboratory also have recorded numerous individuals whose T. gondii  antibody status has switched from positive to negative over time. A study of blood donors in Scotland similarly reported that 6/184 (3.3%) seroreverted from positive to negative over 4 years.30

In an attempt to assess the stability of antibodies to T. gondii, they were measured every 3 months for a year in 60 individuals with schizophrenia.  During that time there was no significant decrease in antibody levels.25

Related to this issue is whether the toxoplasma tissue cysts persist over the lifetime of the host, as has been widely assumed, and what relationship this has to the persistence of IgG antibodies.


  1. Torrey EF, Bartko JJ, Lun Z-R, et al. Antibodies to Toxoplasma gondii  in patients with schizophrenia: a meta-analysis. Schizophr Bull. 2007;33:729–736.
  2. Torrey EF, Bartko JJ, Yolken RH. Toxoplasma gondii : meta-analysis and assessment as a risk factor for schizophrenia. Schizophr Bull. 2012;38:642-647.
  3. Sutterland, AL, et al. Beyond the association. Toxoplasma gondii in schizophrenia, bipolar disorder, and addiction: systematic review and meta-analysis. Acta Psychiatrica Scandinavica. 2015; 1-19.
  4. Monroe JM, et al. Meta-analysis of anti-Toxoplasmsa gondii IgM antibodies in acute psychosis. Schiz Bull. 2015; 41: 989-998.
  5. Leweke FM, Gerth CW, Koethe D, et al. Antibodies to infectious agents in individuals with recent onset schizophrenia. Eur Arch Psychiatry Clin Neurosci. 2004;254:4–8.
  6. Hayes LN, et al. Inflammatory molecular signature associated with infectious agents in psychosis. Schiz Bull. 2014; 40(5):963-72.
  7. de Witte LD et al. The association between antibodies to neurotropic pathogens and schizophrenia: a case-control study. Nature Schizophrenia. 2015; 1: 2334-265X/15.
  8. Niebuhr DW, Millikan AM, Cowan DN, et al. Selected infectious agents and risk of schizophrenia among U.S. military personnel. Am J Psychiatry. 2008;165:99–106.
  9. Li Y, Weber NS, Fisher JA, et al. Association between antibodies to multiple infectious and food antigens and new onset schizophrenia among US military personnel. Schizophr Res. 2013;151:36-42.
  10. Amminger GP, McGorry PD, Berger GE, et al. Antibodies to infectious agents in individuals at ultra-high risk for psychosis. Biol Psychiatry. 2007;61:1215–1217.
  11. Wang H, Yolken RH, Hoekstra PJ, et al. Antibodies to infectious agents and the positive symptom dimension of subclinical psychosis: the TRAILS study. Schizophr Res. 2011;129:47–51.
  12. Pedersen MG, Stevens H, Pedersen CB, et al. Toxoplasma infection and later development of schizophrenia in mothers. Am J Psychiatry. 2011;168:814-821.
  13. Mortensen, PB et al. Toxoplasma gondii as a risk factor for early-onset schizophrenia: analysis of filter paper blood samples obtained at birth. Biol Psychiatry 61:688-693, 2007.
  14. Blomstrom, A, et al. Maternal antibodies to infectious agents and risk for non-affective psychoses in the offspring—a matched case-control study. Schizophr Res. 140(1-3):25-30, 2012.
  15. Blomstrom, A., et al. Influence of maternal infections on neonatal acute phase proteins and their interaction in the development of non-affective psychosis. Transl Psychiatry. 2015. doi: 10.1038/tp.2014.142.)
  16. Brown, AS, et al. Maternal exposure to toxoplasmosis and risk of schizophrenia in adult offspring. Am J Psychiatry. 162: 767-773, 2005.
  17. Buka, SL, et al. Maternal infections and subsequent psychosis among offspring. Arch Gen Psychiatry. 58: 1032-1037, 2001.
  18. Xiao, J. et al. Serological pattern consistent with infection with type I Toxoplasma gondii in mothers and risk of psychosis among adult offspring. Microbes Infect; 11:1011-1018, 2009.
  19. Sabin AB. Toxoplasmic encephalitis in children. JAMA. 1941;116:801–807.
  20. Weinman D. Chronic toxoplasmosis, Journal Infectious Diseases. 1943; 73: 85-91.
  21. Van Druten H, Van Knapen F, Reintjes A. Epidemiologic implications of limited-duration seropositivity after Toxoplasma infection. Am J Epidemiol. 1990;132:169–180.
  22. van der Veen J, Polak MF. Prevalence of Toxoplasma antibodies according to age with comments on the risk of prenatal infection. J Hyg (Camb). 1980;85:165–174.
  23. Konishi E. Annual change in immunoglobulin G and M antibody levels to Toxoplasma gondii  in human sera. Microbiol Immunol. 1989;33:403–411.
  24. Koppe JC, Kloosterman GJ. Congenital toxoplasmosis: long-term follow-up. Paediatrie und Paedologie. 1982;17:171–179.
  25. Yuksel P, Kocazeybek B, Ozdemir A, et al. Stability of Toxoplasma gondii: Antibody levels in schizophrenia. Schizophrenia Research 2017
  26. Lindgren M, et al. The association between toxoplasma and the psychosis continuum in a general population setting. Schizophrenia Research. 2018; 193: 329-335.
  27. Yolken RH, Torrey EF, Dickerson F. Evidence of increased exposure to Toxoplasma gondii in individuals with recent onset psychosis but not with established schizophrenia. PLOS Neglected Tropical Diseases.  11(11): e0006040.
  28. Rougier S, Montoya JG, Peyron F. Lifelong persistence of Toxoplasma cysts: A questionable dogma? Trends in Parasitology. 2017; 22: 93-101.
  29. Xiao J, et al. Toxoplamsa gondii: Biological parameters of the connection to schizophrenia. Schizophrenia Bulletin. 2018; 44:983-992.
  30. Burrells A, et al. The prevalence and genotypic analysis of Toxoplasma gondii from individuals in Scotland, 2006-2012. Parasit Vectors. 2016;9: 324.
  31. Jacobs L, et al. The occurrence of Toxoplasma infection in the absence of demonstrable antibodies. Proceedings of the First International Congress of Parasitology. [Milan: Pergmanon Press, 1966] 176-177.