The World Journal of Biological Psychiatry, 2005; 6(Suppl 2): 44
/48

LECTURE

Childhood meningitis increases the risk for adult schizophrenia*

ANDRÉ L. ABRAHAO1, ROBERTO FOCACCIA2 & WAGNER F. GATTAZ1

1
Departament of Psychiatry, Faculty of Medicine, University of Sao Paulo, Sao Paulo, Brazil, and 2Instituto de Infectologia
Emı́lio Ribas, Sao Paulo, Brazil
World J Biol Psychiatry Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 11/07/14

Abstract
Objective: We investigated the hypothesis that a meningitic infection in childhood may increase the risk of a psychiatric
disorder in adulthood. Method: We conducted a follow-up study of 190 individuals affected by a meningitis infection the
first 4 years of life, during an epidemic in São Paulo, Brazil, between 1971 and 1974. As a control group, we investigated
156 siblings of the meningitis patients who were not affected by meningitis at childhood. Results: In the 190 cases of
meningitis, we found eight (4.2%) cases of schizophrenia against none in the controls, and 40 (21.0%) cases of life
occurrence of psychotic symptoms compared to 12 (7.6%) cases in the control group (P B/0.001). We found no differences
between the two groups regarding the occurrence of other psychiatric disorders and of neurological soft signs. Conclusion:
Meningitis during childhood significantly increased the risk of schizophrenia in particular in adulthood, and of psychosis in
general.
For personal use only.

Key words: Meningitis, childhood infection, schizophrenia, risk factor

Introduction (O.R./7.8) and mood disorders with psychotic

symptoms (O.R./7.7) in adulthood.
In the last century, several authors discussed the
There is number of studies also showing the
participation of infectious diseases in the risk for
influence of pre-natal infections increasing the risk
schizophrenia. Kraepelin considered that ‘infections
for psychiatric illnesses. Influenza, rubella, herpes
in the years of development might have a causal
virus and cytomegalovirus have been studied and
importance’ for schizophrenia (Wright et al. 1999). positive findings have been shown (Brown and Susser
Menninger, in 1928, in a review of the studies of his 1999). Influenza, the most studied prenatal infec-
time about infectious diseases and psychiatric illness, tion, shows some contradictory findings (reviewed by
related that cases of acute meningitis, encephalitis, Wright et al. 1999). Brown et al. (2001) found that
typhoid fever, recurrent fever, typhus, influenza, about 21% of patients who had presented clinical
malaria, tuberculoses, gastrointestinal infections manifestations of congenital rubella have developed a
and septicemia were frequently followed by an acute schizophrenic spectrum disorder in adulthood.
and transitory episode of a schizophreniform In the city of São Paulo, Brazil, there was an
syndrome or a schizophrenic disease in its more meningococcal meningitis epidemic from 1971 to
specific sense. Rantakallio et al. (1997), in a study 1974 (Iversson 1976). Most of the cases (90%) were
comparing individuals with a central nervous system admitted to the Emilio Ribas Hospital, a public and
infection during childhood with a control group university hospital, specialized in infectious diseases.
followed-up until age 27 years, found an increase of In the present study, we investigated the prevalence
schizophrenia in the infection group (O.R. /4.2). of psychiatric diagnoses in general in a sample of
Leask et al. (2002), in an investigation of data from adults who, during the epidemic period, were
medical school examinations, reported that child- admitted up to the age of 4 years to this hospital
hood meningitis increased the risk of schizophrenia with the diagnosis of meningitis.

Correspondence: Wagner F. Gattaz, M.D., Full Professor of Psychiatry & Director of the Laboratory of Neuroscience, Department of
Psychiatry, Faculty of Medicine, University of Sao Paulo, Rua Dr. Ovidio Pires de Campos 785, 05403-010 Sao Paulo, SP, Brazil. Tel:
/5511 3069 8010. E-mail [email protected]
*Part of these results has already been published in the Eur Arch Psychiatry Clin Neurosci 254:23
/6, 2004.

ISSN 1562-2975 print/ISSN 1814-1412 online # 2005 Taylor & Francis

DOI: 10.1080/15622970510030063
 Childhood meningitis increases the risk for adult schizophrenia 45

Material and methods Statistics

Finding the subjects For data analysis, the following statistics were used,
as needed: Kolmogorov
/Smirnov test, Breslow
/
The database used was the file of microfilmed
Day Test,Mann
/Whitney test, Chi-square, t -test,
medical registers from Emı́lio Ribas Hospital.
From these registers, we obtained the names of the Kappa test, test for the difference of signs and
patients, of their parents and the patients’ age when stratified analysis. The data are presented as
admitted. Then a search was done in the database average9/standard deviation.
of the telephone company, on the web site (www.
telefonica.net.br), for the name of the patient and of Results
his parents. When a subject was found, he was
informed about our study and invited to an interview Individuals with meningitis in childhood had a
at the Institute of Psychiatry of the University of Sao higher prevalence of schizophrenia and of psychoses
Paulo Medical School. We also asked to the inter- in general, as well as more neurological disorders
and higher comorbidity rates, compared to their
World J Biol Psychiatry Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 11/07/14

view one sibling of the patient, preferably older, but

aged as close as possible to the patient, and who had siblings without childhood meningitis (Tables II, III
not been affected by meningitis in childhood. The and IV).
siblings composed the control group. The cases Since deafness may increase the risk for psychoses
(meningitis) and the controls (siblings) received (Altshuler and Sarlin 1969), we compared the
R$50 (9/U$20) each, as a compensation for coming diagnoses between three groups: controls, cases
to the interview. with any hearing deficit and cases without hearing
deficit. Basically, all the differences shown before
are maintained here, with a predominance of psy-
Interviews
choses in general in the meningitis cases (Table V).
A standard psychiatric diagnostic interview was There was no association between neurological
For personal use only.

undertaken based on the ICD-10 Checklist (Janca diagnoses and psychosis. If we exclude from the
and Hiller 1996), and the presence of neurological analyses cases and controls with neurological diag-
soft signs was tested with the neurological evaluation noses, the difference remains as before, with more
scale from Buchanan and Heinrichs (1989). psychosis and more schizophrenia in the meningitis
group (Table VI).
Reliability of the diagnoses
Since only one psychiatrist has interviewed all Soft signs
patients, and he was aware whether they were cases There were no differences in the neurological soft
or siblings, we tested the reliability of the diagnoses. signs between cases (7.49/4.1) and controls (7.89/
Four psychiatrists received a written resume of the 4.2). However, soft signs ratings were higher in
clinical histories of all individuals. The resumes were
individuals with psychoses (8.99/3.8) than in indi-
codified, not identifying who was case or sibling. All
viduals without psychoses (7.39/4.2, P B/0.01).
four doctors showed an agreement in their diag-
noses, with the main researcher higher than 80% in
the Kappa test. Age of meningitis
The age at the time of meningitis infection was lower
Sample description in cases with psychotic symptoms (21.99/13.2
When the microfilmed files were analysed, 4951 months) than in cases without psychotic symptoms
registers were found from patients admitted with (27.49/15.3 months, P B/0.05).
meningitis, aged 4 or less, from January 1970 to
December 1975. Up to now, 1,890 files had Table I. Socio-demographic characteristics of cases and controls.
telephone searches, and 361 (18.1%) individuals
Cases (n/190) Siblings (n/156)
were localized. Of these, 190 (52.6%) cases and 156
(82.1%) controls agreed and came to the interview. Men 88 (46.3%) 56 (35.9%)
Table I presents the demographic and social Women 102 (53.7%) 100 (64.1%)/
economical data of cases and controls. No remark- Age (years) 29.29/1.6 30.09/5.9
Income (in Reais) R$ 948.009/943.00* R$ 739.009/828.00
able difference was found, except that (interestingly)
Years at school 11.59/3.7 11.39/4.0
in the cases income was significantly higher than in
their siblings. *P B/0.05, /P B/0.10.
 46 A. L. Abrahao et al.
Table II. General prevalence of neurological and psychiatric Table IV. Neurological disorders.
disorders.
Siblings
Cases Siblings Case (n /190) (n/156)
(n/190) (n/156)
No neurological diagnosis 143 (75.2%)*** 147 (94.2%)
All psychiatric disorders 117 (61.5%) 93 (59.6%) Deafness (partial and total) 18 (9.4%)*** 1 (0.6%)
All neurological disorders 47 (24.7%)*** 9 (5.7%) Motor deficit 9 (4.7%)** 0 (0%)
Headache NOS 12 (6.3%) 7 (4.5%)
***P B/0.001. Epilepsy 6 (3.1%) 1 (0.6%)
Visual deficit 1 (0.5%) 2 (1.2%)
Other neurological disorders 2 (1.0%) 0 (0%)
Discussion (visual deficit related to
meningitis, anosmia,
Our main finding was that meningitis in childhood nystagmus cross-eye and
increased the risk for psychosis in adulthood by 4.6 tremors)
times, and for schizophrenia in particular by 4.4
**P B/0.01; ***P B/0.001.
World J Biol Psychiatry Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 11/07/14

times, whereas no changes were found in the risk for

other psychiatric disorders.
An interesting aspect of our study is that the (lipopolysacccharide, LPS) to pregnant female rats
control group was formed by siblings of the patients,
disrupts the prepulse inhibiton (PPI) of the acoustic
who did not have had meningitis in early child-
startle reflex in the offspring, and this effect could be
hood. So, to a considerable extent, both groups
reversed by antipsychotic drugs. This finding is of
were very well matched regarding aspects such as
interest because PPI is an accepted animal model for
genetic, socio-cultural, economical and nutritional
schizophrenia.
backgrounds. Thus, the difference between our
The prevalence of psychiatric disorders in our
groups of cases and controls was, basically, the
study, 60.3% in siblings and 62.4 in controls, is
presence or absence of meningitis in the first four
above what is found in epidemiological studies done
For personal use only.

years of life.
Our findings are similar to the studies of post- in representative samples of the population in Brazil
natal infections increasing the risk for psychosis (Andrade et al. 1999 reported 45.6%) and in the
(Rantakallio et al. 1997; Leask et al. 2002). Our USA, where Robins and Regier (1990), in the
cases with psychotic symptoms showed more neu- Epidemiological Catchment Area Study (ECA),
rological soft signs as compared to the siblings report 32% of total psychiatric morbidity. In our
group, and this also agrees with the literature study, it is likely that the call to a medical interview
(Brown et al. 2001; Leask et al. 2002). in the University Hospital may have selected a
The mean age of meningitis was lower in the cases sample in higher need of medical-psychiatric care,
that presented psychotic symptoms than in cases therefore with a higher morbidity, among those who
without psychotic symptoms, suggesting that vulner- answered the call. Nevertheless, our figures are
ability to psychosis may be increased by earlier similar to those reported by Brown et al. (2001),
insults during the maturation of the brain. However, who found 58.5% overall psychiatric morbidity in
the mechanisms by which it happens are not yet the rubella study.
clarified. In animal experiments, Borrell et al. (2002) Some studies pointed to an association between
found that the injection of a bacterial endotoxin deafness and increased risk for psychosis (Altshuler

Table III. Psychiatric diagnosis grouped.

Cases (n/190) Siblings (n /156)

Without any diagnoses 73 (38.5%) 63 (40.4%)

Anxiety disorders (F 40.1, F 40.2, F 41.0, F 41.1, F 41.2, F 42.9) 71 (37.3%) 55 (35.2%)
Personality disorders (F 60.3, F 60.5) 8 (4.2%) 3 (1.9%)
Alcohol and drugs abuse (F 10.1, F 10.2, F 12.1, F 14.2) 14 (7.3%) 12 (7.6%)
Mood disorder without psychotic symptoms (F 32.0, F 32.1, F 32.2, 51 (26.8%) 38 (24.3%)
F 33.0, F 33.1, F 33.2, F 34.0 F 34.1 e F 31.0)
Mood disorder with psychotic symptoms (F 32.3, F 33.3)(2) 13 (6.8%) 6 (3.8%)
Schizophrenia (1) (F 20.0, F 20.5, F 20.6) 8 (4.2%)*** 0 (0%)
Other psychosis (3) (F 29, F 10.5, F 22.0) 21 (12.1%)** 4 (2.8%)
All psychoses 40 (21.0%)*** 12 (7.6%)
Comorbidity 52 (30.0%)** 20 (14.2%)

**P B/0.01; ***P B/0.001.

 Childhood meningitis increases the risk for adult schizophrenia 47
Table V. Psychiatric diagnosis in cases with and without hearing deficit.

Controls Cases with hearing deficit Cases without hearing deficit

(n/141) (n/18) (n /155)

Without psychiatric diagnostic 54 (38.3%) 9 (50.0%) 56 (36.1%)

Anxiety disorder 51 (36.1%) /2 (11.1%)* 60 (38.7%)
Personality disorders 3 (2.1%) 0 5 (3.2%)
Alcohol and drugs 12 (8.3%) 1 (5.6%) 13 (8.4%)
Mood without psychotic symptoms 34 (24.1%) 5 (27.8%) 44 (28.4%)
Mood with psychotic symptoms 4 (2.8%) 1 (5.6%) 12 (7.7%)
Schizophrenia 0 1 (5.6%) 7 (4.5%)*
Other psychosis 4 (2.8%) 1 (5.6%) 20 (12.9%)**
All psychoses 7 (5.0%) /3 (16.7%) 35 (22.6%)***
Co morbidity 20 (14.2%) 2 (11.1%) 50 (32.3%)***

/P B/0.1; *P B/0.05; **P B/0.01; ***P B/0.001.

World J Biol Psychiatry Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 11/07/14

and Sarlin 1969). However, in our study the excess women in controls does not contribute to the lower
of psychosis in the meningitis group is maintained prevalence of psychosis in this group. Besides, the
if we left off the cases with partial or full deaf- stratified analysis showed that the variation of sex
ness from the analyses (Table V). The differences does not contribute to the differences in the occur-
are also maintained after the exclusion from the rence of psychotic diseases in cases and controls.
analysis of the cases with a neurological diagnosis And, finally, comparing only the female individuals
(Table VI). we also found more psychoses in women with
In the control group there was a non-significant meningitis than in women without meningitis during
dominance of women, compared to the group of childhood (Table VII).
For personal use only.

cases. The onset of schizophrenia in women is about Taken together, our findings add to the body of
3 years later then men. In a representative sample, literature showing that infectious diseases that may
Häfner et al. (1999) report that the average age of affect the brain in childhood do increase the risk for
appearance of the first negative symptoms of the psychosis in adults. The clarification of the mechan-
disease was about 22 years in men and 25 in women, isms by which this occurs could shed more light in
and the age at the first psychotic symptoms was 26.7 the understanding of environmental influences in the
and 30.9 years, respectively. However, our sample risk for psychosis.
has an average age of 30.09/5.9 years old, indicating
that most of the controls had already passed the risk
Statement of interest
age for disease onset. Therefore, the excess of
The authors have no conflict of interest with any
commercial or other associations in connection with
Table VI. Psychiatric diagnosis in cases and controls without the submitted article.
neurological disorders.

Cases without Table VII. Psychiatric diagnosis in women.

neurological
Controls disorders Siblings Cases
(n/147) (n/143) (n /100 women) (n/102 women)

Without psychiatric 58 (39.4%) 54 (37.7%) Without psychiatric 30 (30.0%) 26 (25.5%)

diagnostic diagnostic
Anxiety disorder 50 (34.0%) 52 (36.3%) Anxiety 40 (40.0%) 53 (51.9%)
Personality disorders 0 5 (3.5%)* Personality disorders 3 (3.0%) 3 (2.9%)
Alcohol and drugs 10 (6.8%) 9 (6.3%) Alcohol and drugs 4 (4.0%) 6 (5.9%)
Mood without 28 (19.0%) 25 (17.5%) Mood without psychotic 35 (35.0%) 39 (38.2%)
psychotic symptoms symptoms
Mood with psychotic 6 (4.1%) 9 (6.3%) Mood with psychotic 5 (5.0%) 11 (10.8%)
symptoms symptoms
Schizophrenia 0 7 (4.9%)** Schizophrenia 0 3 (2.9%)
Other psychosis 6 (4.0%) 16 (11.2%)* Other psychosis 4 (4.0%) 15 (14.7%)*
All psychoses 12 (8.1%) 29 (20.3%)** All psychosis 9 (9.0%) 27 (26.4%)**
Co morbidity 23 (15.6%) 44 (30.7%)** Co morbidity 16 (16.0%) 38 (37.2%)**
*P B/0.05; **P B/0.01. *P B/0.05; **P B/0.01; ***P B/0.001; /P B/0.1.
 48 A. L. Abrahao et al.

References for the causes of schizophrenia. 4. Balance of the century,

Darmstadt: Steinkopff; Berlin: Springer. p 38.
Altshuler K, Sarlin M. 1969. Deafness and schizophrenia: A Iversson LB. 1974. Aspectos Epidemiológicos da meningite
family study. In: Rainer J, Altshuler K, Kallman F, editors. meningoccocica no municipio de São Paulo (Brasil) no perı́odo
Family and mental health problems in a deaf population. de 1968 a 1974. Rev Saude publ S. Paulo 10:1
/16.
Springfield, IL: Charles C. Thomas. pp 204
/213. Janca A, Hiller W. 1996. ICD-10 Checklists
/ A tool for
Andrade LH, Lólio CA, Gentil V, Laurenti R. 1999. Epidemio- clinicians. Use of the ICD-10 Classification of Mental and
logia dos transtornos mentais em uma área definida de Behavioral Disorders. Compr Psychiatry 37(3):180
/187.
captação da cidade de São Paulo, Brasil. Rev Psiq Clı́n 26(5, Menninger KA. 1928. The schizophrenic syndromes as a product
Edição Especial):257
/261. of acute infectious disease. Arch Neurol Psychiatry 20:464
/
Borrell J, Vela JM, Arevalo-Martin A, Molina-Holgado E, Guaza 481.
C. 2002. Prenatal immune challenge disrupts sensorimotor Leask SJ, Done J, Crow TJ. 2002. Adult Psychosis, common
gating in adult rats. Implications for the etiopathogenesis of childhood infections and neurological soft signs in a national
schizophrenia. Neuropsychofarmacology 26(2):204
/215. bith cohort. Br J Psychiatry 181:387
/392.
Brown AS, Susser ES. 1999. Plausibility of prenatal rubella, Lim KO, Beal DM, Harvey RL Jr, Myers T, Lane B, Sullivan EV,
influenza, and other viral infections as risk factors for schizo- Faustman WO, Pfefferbaum A. 1995. Brain dysmorphology in
phrenia. In: Susser ES, Brown AS, Gorman JM, editors. adults with congeniatal rubella plus schizophrenialike symp-
World J Biol Psychiatry Downloaded from informahealthcare.com by CDL-UC Santa Cruz on 11/07/14

Prenatal exposures in schizophrenia. Progress in psychiatry toms. Soc Biol Psyhciatry 37:764
/776.
series, no. 56. Washington and London: American Psychiatric Rantakallio P, Jones P, Moring J, Von Wendt L. 1997. Association
Press. pp 113
/133. between central nervous system infections during childhood
Brown AS, Cohen P, Harkavy-Friedman J, Babulas V, Malaspina and adult onset schizophrenia and other psychoses: A 28-year
D, Gorman JM. Prenatal rubella, premorbid abnormalities and follow-up. Int J Epidemiol 26(4):837
/843.
adult schizophrenia. Biol. Psychiatry 49:473
/486. Robins LN, Regier DA. 1990. Psychiatric disorders in America
/
Buchanan RW, Heinrichs DW. 1989. The Neurological Evalua- The Epidemiologic Cathment Area Study. New York: The Free
tion Scale (NES): a structural instrument for the assessment of Press. pp 329
/343.
neurological signs in schizophrenia. Psychiatry Res 27(3):335
/ Wright P, Takei N, Murray RM, Sham PC. 1999. Seasonality,
350. prenatal influenza exposure, and schizophrenia. In: Susser ES,
Häfner H, Maurer K, Löffler W, Heiden W, Stein A, Könnecke R, Brown AS, Gorman JM, editors. Prenatal exposures in schizo-
Hambrecht M. 1999. Onset and prodromal phase as determi- phrenia. Progress in psychiatry series, no. 56. Washington and
nants of the course. In: Gattaz WF, Häfner H, editors. Search London: American Psychiatric Press. pp 89
/112.
For personal use only.