Association between Central Nervous System Infections during ...

International Journal of Epidemiology © International Epidemiological Association 1997

Vol. 26, No. 4 Printed in Great Britain

Association between Central Nervous System Infections during Childhood and Adult Onset Schizophrenia and Other Psychoses: A 28-Year Follow-Up PAULA RANTAKALLIO,* PETER JONES,** JUHA MORING† AND LENNART VON WENDT‡ Rantakallio P (Department of Public Health Science and General Practice, University of Oulu, Aapistie 1, FIN-90220 OULU, Finland), Jones P, Moring J and von Wendt L. Association between central nervous system infections during childhood and adult onset schizophrenia and other psychoses: A 28-year follow-up. International Journal of Epidemiology 1997; 26: 837–843. Background. Maternal exposure to influenza epidemics during pregnancy may increase the risk of schizophrenia in the offspring. We investigated the association between central nervous system (CNS) infections defined prospectively up to the age of 14, and later onset of schizophrenia and other psychoses in the 1966 birth cohort in Northern Finland, which covers 96% of all births in the area during that year. Methods. Data regarding CNS infections were collected 1966–1980. Registered diagnoses of psychoses in 1982–1993 were validated on DSM-III-R criteria. Results. Out of 11 017 subjects, 145 had suffered a CNS infection during childhood, 102 of them a viral infection, 76 had DSM-III-R schizophrenia and 53 some other psychosis. Four cases of schizophrenia had suffered viral CNS infection and two cases of other psychosis bacterial infection. When neurological abnormalities and father’s social class were adjusted odds ratio (OR) of schizophrenia after viral CNS infection was 4.8 (95% confidence intervals [CI] : 1.6–14.0); the other significant risk factors being intelligence quotient (IQ) ,85, perinatal brain damage and male sex but not epilepsy. Similarly adjusted OR of other psychoses was 6.9 (95% CI : 1.4–32.8) after bacterial CNS infection; the other significant risk factors being IQ ,85 and severe hearing defect. Two of the five viral infections were caused by Coxsackie B5 during an epidemic in which 16 neonates were infected together. Conclusions. Central nervous system infections during childhood clearly carried an increased risk of adult onset schizophrenia or other psychoses, viral infections being important for schizophrenia, particularly Coxsackie B5 during the newborn period. Keywords: psychosis, schizophrenia, viral central nervous system infection, bacterial central nervous system infection, neurological handicaps, school performance

of the main aetiological hypotheses for schizophrenia alongside a genetic effect, whereby infections, especially viral infections, may attenuate fetal brain development. 6 ‘Slow’ or latent viral infections have also been hypothesized to explain the relatively long period between the occurrence of the noxious agent and the onset of the disease.7 Experimental work in this field has been based on measurements of antibodies and immunofunctions in psychiatric patients, but no epidemiological studies have been presented regarding the association in individuals between verified CNS infection during childhood and adult onset of psychosis at the individual level. The purpose of the present work was to investigate the possibility of an association between childhood CNS infection, and adult onset psychosis up to the age

The finding that psychosis may arise during the acute phase of central nervous system (CNS) infection was first published 70 years ago1 and many reports have been published thereafter.2 Several authors have also found an association between putative maternal exposure to viral infection during pregnancy and the subsequent development of adult-onset schizophrenia in the offspring.3–5 These findings have given rise to one

* Department of Public Health Science and General Practice, University of Oulu, Aapistie 1, FIN-90220 OULU, Finland. ** Department of Psychiatry, University of Nottingham, UK. † Department of Psychiatry, University of Oulu, Finland. ‡ Helsinki University Central Hospital for Children’s and Adolescents’ Pediatric Neurology.

837

838

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

of 28 years in the Northern Finland birth cohort for 1966, in which CNS infection had been recorded prospectively.8 This is the first such investigation in a large, general population birth cohort with laboratory confirmation of infection and validation of individual diagnoses.

SERIES AND METHOD The original data consisted of 12 058 live births from the two northernmost provinces of Finland, Oulu and Lapland, and covered 96% of all children born in the region in 1966.9 A wide variety of social and biological characteristics of the mother and family were recorded during the pregnancies in 1965 and 1966,9 and the later social status of the family and the children’s health and school performance were determined by means of another questionnaire in 1980 and 1981, when the children were 14 years old. Only 14 children (1.2 per 1000) had been lost to follow-up by that age.10 The series considered here consists of all those alive at the age of 16 and still living in Finland, 11 017 in all. From 1041 lost cases, 284 had died and 757 emigrated. Most of the latter had emigrated with their families during the late 1960s because of the high unemployment rate in Finland at that time. All hospital records regarding clinical diagnoses of psychosis in the Finnish Hospital Discharge Register (FHDR) were collected from various hospitals up to 1993. The DSM-III-R criteria based on Diagnostic and Statistical Manual of Mental Disorders (DSM) were applied to the information in the case records in order to identify schizophrenia and other psychoses. The data were rechecked against clinical records by two senior researchers.11 Individuals fulfilling the DSM-III-R criteria for psychosis were further classified into two groups for the purposes of our analyses: (i) narrowly defined, ‘nuclear’ DSM-III-R schizophrenia (295.1–295.3, 295.6 and 295.9, n = 76), and (ii) other psychoses (291–293, 296, 298, 299, n = 36), and ‘spectrum’ schizophrenia (295.4, 295.7, 297.1, 301.2, n = 18). Data concerning CNS infections up to the age of 14 were collected in several ways between 1966 and 1980, the most important sources being records of admissions to the four children’s hospitals in the area from 1966 to 1972 and the FHDR thereafter until 1980. An additional 12% of the cases were identified from other sources, mainly the records of neurological outpatient clinics.8 Several possible confounding factors were studied; father’s social class, perinatal brain damage, mental retardation, childhood epilepsy and hearing defects. Social class was classified by father’s occupation I + II being professional, III skilled workers, IV unskilled

workers, and farmers as a separate class. Children were considered to have perinatal brain damage if they had an Apgar score of zero at one min. or less than five at 15 min., convulsions during the neonatal period, or a diagnosis of asphyxia, brain injury, or intraventicular haemorrhage, but did not have CNS malformation, chromosomal aberrations, or hereditary CNS degeneration.12 Mental retardation was operationally defined in terms of school performance up to the age of 14 years and intelligence quotients (IQ) were added to our files for those with learning difficulties;13 an IQ of below 85 being the class limit in this analysis. The criteria for childhood epilepsy and severe hearing defect up to the age of 14 are described in the earlier work.14,15 Data on school performance at the age of 16 were taken from the National Register of Applicants for Intermediate Education. Final educational status and data on employment status in 1990, at the age of 24 years, were obtained from the Central Statistical Office of Finland. Fisher’s exact test was used to examine the statistical significances of differences in proportions, and otherwise the risk of psychotic diseases was expressed in terms of a cumulative incidence ratio with 95% confidence intervals (CI). Attributable fraction and population attributable risk we defined according to Basic Epidemiology.16 Multiple logistic regression was used to examine the association between CNS infection and schizophrenia versus other psychoses when the possible confounding factors were adjusted.

RESULTS There were 174 verified CNS infections in 167 children,8 of whom 145 were alive and still living in Finland at the age of 16 years (91, 64%, males); 102 of them had a viral infection. By the age of 27 there were 76 cases of schizophrenia (51 men, 25 women) and 53 cases of other psychosis including schizophrenia spectrum disorders (35 men, 18 women). The cumulative incidence of schizophrenia was 0.7% (72/10 872) among those not exposed in childhood to CNS infections and 2.8% (4/145) among those who were so exposed, the corresponding figures for other psychoses being 0.5% and 1.4%. Thus, the cumulative incidence ratios were 4.2 (95% CI : 1.5–11.3) for schizophrenia and 2.9 (95% CI : 0.7–12.1) for other psychosis. The population attributable risk for schizophrenia following CNS infection was 4% (95% CI : 1.9–4.8%), and the attributable fraction of CNS infection for schizophrenia is 76% (95% CI : 35–91%) assuming causality. All infected cases of schizophrenia had suffered from a viral infection, Coxsackie B5 in two cases and Adenovirus 7 and Mumps virus in one case each (Table 1).

839

CHILDHOOD CNS INFECTIONS AND ADULT ONSET PSYCHOSES

TABLE 1 Main clinical data on six patients with schizophrenia and other psychosis who suffered from central nervous system (CNS) infection during childhood in Northern Finland, 1966 birth cohort Patients

Schizophrenia female Schizophrenia male Schizophrenia male Schizophrenia male Other psychosis male Other psychosis male a

Perinatal adverse factors

Type of CNS infection

Age at CNS infection

Neurological impairment up to age 14

Age at onset of psychosis (years)

Neonatal convulsions

Coxsackie B5

9 days

–

20

Neonatal convulsions

Coxsackie B5

10 days

–

27

RDSa birthweight 2050 g 34 gestational weeks –

Viral encephalitis Adeno 7

13 months

–

20

Mumps encephalitis

87 months

–

21

Bacterial meningitis Haemophilus influenzae Bacterial meningitis Neisseria meningitides

5 months

Epilepsy

25

33 months

Deafness

16

– –

Respiratory distress syndrome.

Thus, the cumulative incidence ratio of schizophrenia was even higher among those with a past viral CNS infection; 5.9 (95% CI : 2.2–16.0). Two cases of other psychoses had suffered from bacterial meningitis caused by Haemophilus influenza and Neisseria meningitides. Only one of these six cases of CNS infection was female, belonging to the schizophrenic group (Table 1). Of particular interest are the two cases of Coxsackie B5 meningitis during the neonatal period who developed adult schizophrenia. These cases formed part of an epidemic in which 17 newborn cases of Coxsackie B5 were identified within one month from July to August 1966, all except one occurring in the same maternity unit.17 Of the cases, 16 belong to the present series, one having emigrated, giving an incidence of adult onset schizophrenia of 2/16 (12.5%) among those with neonatal meningitis caused by Coxsackie B5. The differences between this figure and the incidences in the others, 74/11 001, 0.7% is statistically significant (P = 0.005). Antibody responses were studied by a neutralization test and all 15 cases studied, including both with schizophrenia, showed high Coxsackie B5 antibody titres.17 In none of the cases was intrauterine infection suspected, although one of the mothers had markedly increased antibody titres (as did eight other mothers exposed to epidemics). All the infants with Coxsackie B5 meningitis had a typical clinical picture of serous meningitis; the maximum body temperature varied from 38.0 to 40.0°C, and 15 from 17 had clinical symptoms such as nuchal rigidity, irritability, whimpering etc, and the cases with later schizophrenia did not differ from the

others (cases 5 and 6 in the original paper).8 A clinical follow-up examination was carried out on all the 17 children at the age of 6 months and for 15 children at the age of 12 months including both cases with later onset schizophrenia. 18 The neurological status was concluded to be normal at that time for all cases. In Tables 2 and 3 the rates of schizophrenia and other psychoses are stratified by potential confounders, paternal social class and neurological abnormalities. Neurological abnormalities were more common in the total group with schizophrenia and other psychoses than among the controls. The risk for schizophrenia among those with childhood epilepsy15 was 3.2 and that of other psychoses 3.4 fold, the corresponding figures among those with mental retardation (IQ ,85) being 5.6 and 5.7. No severe hearing defects were associated with the schizophrenia but the risk of other psychoses was increased in this group. Table 4 shows the association (adjusted odds ratios [OR]) between schizophrenia and viral CNS infection during childhood as well as the other risk factors. The association between schizophrenia and CNS infection was significant (OR = 4.8, 95% CI : 1.6–14.0) as well as between IQ ,85 and perinatal brain damage. When all CNS infections were taken into account the OR was 3.3 (95% CI : 1.2–9.6). Similar analysis for other psychoses indicated that association with all CNS infection was not significant (OR = 2.4, 95% CI : 0.5–10.5). Only variables IQ ,85 and severe hearing defect had a significant association with other psychoses. When only bacterial infections were included the association

840


TABLE 2 Number of cases and incidence of schizophrenia by social class and neurological abnormalities and central nervous system (CNS) infection: Northern Finland, 1966 birth cohort CNS infection No Number all

Paternal social class I + II III IV farmers Epilepsy no yes IQ ,85 no yes Cerebral palsy no yes Severe hearing defect no yes Perinatal brain damage no yes

Yes Number all

Incidence of schizophrenia (number) %

Number all

0.7 0.6 0.7 0.6

39 50 42 14

(–) – (3) 6.0 (1) 2.4 (–) –

2611 3547 2798 2061

10 696 176

(68) 0.6 (4) 2.3

130 15

(4) 3.1 (–) –

10 826 191

(72) 0.7 (4) 2.1

10 595 276

(62) 0.6 (10) 3.6

137 8

(4) 2.9 (–) –

10 732 285

(66) 0.6 (10) 3.5

10 824 48

(70) 0.6 (2) 4.2

139 6

(4) 2.9 (–) –

10 963 54

(74) 0.7 (2) 3.7

10 829 43

(72) 0.7 (–) –

139 6

(4) 2.9 (–) –

10 968 49

(76) 0.7 (–) –

10 750 122

(68) 0.6 (4) 3.3

135 10

(2) 1.5 (2) 20.0

10 885 132

(70) 0.6 (6) 4.5

2572 3497 2756 2047


Total

(19) (22) (18) (13)

between other psychosis and bacterial CNS infections was statistically significant (OR = 6.8, 95% CI : 1.4–32.8). Five out of the six psychotic cases who had suffered a CNS infection were in the appropriate class for their age at a regular school at the age of 14; the case with bacterial meningitis and deafness being at a school for the deaf. The marks on school reports at the age of 16 were significantly poorer in both psychiatric groups than among the controls, but the cases with CNS infection did not do any worse than the unexposed cases. Similarly, the total number of years of schooling was on average 0.8 years lower among the psychotic cases than in those without psychiatric illness, and those cases who had suffered CNS infection showed a similar effect. The employment status of the cases with CNS infection was similar to that of the other cases of schizophrenia and other psychosis, one being employed at the age of 25, one studying, three on a sickness pension benefit and one unemployed.

DISCUSSION Population studies of the total incidence of CNS infections are very rare and the methodological differences


(19) (25) (19) (13)

0.7 0.7 0.7 0.6

between these studies makes comparison difficult.8 The annual incidence of bacterial CNS infection among children under 14 in this series, 33.3/100 000, was clearly higher than in a cross-sectional Finnish study for 1978, 19/100 000 for children under 15. The main explanation —besides the methodological differences between the studies—for the different incidence figure is that a vaccination campaign was conducted in Finland in 1974– 1976.8 The mortality from bacterial CNS infections in our series was over three times that recorded in the later Finnish series (14.5% versus 4%).8 The main explanation for this is that infancy in the cross-sectional study occurred 12 years later than infancy in our cohort study by which time medical care resources against bacterial infections had greatly advanced; incidence of bacterial CNS infection during the first year of life is about five times higher than during later childhood. Because South-North differences in infant mortality have almost vanished since end of 1960s we can assume that our figures represent the whole country. Further Finnish infant mortality figures have, since the late 1950s, been clearly and constantly lower than, for example, in Great Britain so that we can assume that our figures

841

CHILDHOOD CNS INFECTIONS AND ADULT ONSET PSYCHOSES

TABLE 3 Number of cases and incidence of other psychosis by social class and neurological abnormalities and central nervous system (CNS) infection: Northern Finland, 1966 birth cohort CNS infection No Number all

Paternal social class I + II III IV farmers Epilepsy no yes IQ ,85 no yes Cerebral palsy no yes Severe hearing defect no yes Perinatal brain damage no yes

Yes Number all


Number all

0.3 0.6 0.4 0.5

39 50 42 14

(–) – (–) – (2) 4.8 (–) –

2611 3547 2798 2061

10 696 176

(50) 0.5 (2) 1.1

130 15

(1) 0.8 (1) 6.7

10 826 191

(51) 0.5 (3) 1.6

10 595 277

(45) 0.4 (7) 2.5

137 8

(2) 1.5 (–) –

10 732 285

(47) 0.4 (7) 2.5

10 824 48

(52) 0.5 (–) –

139 6

(2) 1.4 (–) –

10 963 54

(54) 0.5 (–) –

10 829 43

(51) 0.5 (1) 2.3

139 6

(1) 0.7 (1) 16.7

10 968 49

(51) 0.5 (2) 4.1

10 824 48

(52) 0.5 (–) –

139 6

(2) 1.4 (–) –

10 963 54

(54) 0.5 (–) –

2572 3497 2756 2047


Total

(8) (21) (12) (11)

TABLE 4 Adjusted odds ratios (OR) with their 95% confidence intervals (CI) for schizophrenia estimated by the multiple logistic regression model: Northern Finland, 1966 birth cohort Variables

OR

95% CI

Paternal social class I + II IV farmers male sex viral CNS infection IQ ,85 perinatal brain damage cerebral palsy epilepsy severe hearing defect

1.06 0.93 0.93 1.83 4.80 4.77 4.55 0.64 1.13 –

0.58–1.93 0.51–1.70 0.47–1.82 1.13–2.97 1.65–14.0 2.21–10.3 1.72–12.1 0.11–3.77 0.33–3.82 –

are roughly representative for the birth cohorts in Europe in 1960s. The difficulties encountered in comparing incidences of viral infections in different regions are still more evident, since milder infections obviously get less attention. We assumed that about 30% of CNS viral infections had remained unrecorded.8


(8) (21) (14) (11)

0.3 0.6 0.5 0.5

The statistically significantly higher risk of schizophrenia among those who had suffered CNS infection during childhood, together with the fact that all the infections found among the schizophrenics were caused by viruses, supports the general hypothesis that virus infections are one aetiological factor for schizophrenia. The fact that less than 6% of the total group of schizophrenics had suffered a CNS infection is quite consistent with the notion of schizophrenia as a heterogeneous disorder resulting from an interaction between multiple factors, including the person’s genetic predisposition and various environmental effects.6 An infectious aetiology for schizophrenia, even though widely assumed, has not finally been proved. Following Mednick’s3 and others original claim that fetuses in the second trimester at the time of influenza epidemics had an elevated risk of schizophrenia, several other reports have been published, some of them supporting the finding, some not. 19 Many have pointed to abnormalities in immune function in cases of schizophrenia, but few findings have been replicated and most alterations in immune function have been modest in degree and non-specific.6 Three out of the four cases with viral CNS infection and later schizophrenia had also suffered adverse

842


neonatal events, two of them convulsions and one preterm birth, low birthweight and respiratory distress syndrome (Table 1). It has been suggested that obstetric complications and adverse perinatal events may increase the risk of schizophrenia20,21 and the finding in the Northern Finland birth cohort for 1966 supports that view,22 even if the question of the chain of causation remains open. Our results show that childhood CNS infection and perinatal brain damage both had independent associa-tions with schizophrenia. The 1966 birth cohort was prospectively carefully followed up for perinatal brain damage and neurological handicaps up to the age of 14 years.12–15 An IQ ,85 and perinatal brain damage were significant risk factors for schizophrenia but independently of these factors CNS infection was significant risk factor for schizophrenia. We have found one previous publication about the association between history of convulsions and schizophrenia; in the WHO Ten Country Study, patients with schizophrenia were two to three times more likely to have a history of childhood convulsive disorder than the rest of the study population.23 In our series schizophrenia was three times more likely if the subjects had a history of childhood epilepsy (Table 2), although after adjusting the other factors this association disappeared (Table 4). Even though of great theoretical interest, it is not so important from the practical point of view how the adverse effect is mediated. The important thing is to try to prevent CNS infection, especially among young children. Special attention in this respect should be paid to Coxsackie infections, which are still a problem in maternity units.24,25 Contrary to the situation in older children and adults, Coxsackie viruses in newborn babies can cause fulminant infection, sometimes fatal, or as it would appear from our series, severe, lifelong disease. In addition to the question of the specificity of the effect, the vulnerable period during which noxious events with respect to brain development may increase the risk of psychosis may be longer than that initially proposed by Mednick et al.3 It appears to span the first trimester, when malnutrition may have an effect,26 through the second trimester for influenza, and beyond the neonatal period in our series. Thus, we conclude that it is unlikely that there is a single, specific developmental lesion underlying psychosis, least of all if infection is the primary event.

ACKNOWLEDGEMENTS Financial support was received from the Medical Research Council of the Finnish Academy and from the Stanley Foundation.

REFERENCES 1

Menninger K A. Influenza and schizophrenia. An analysis of post-influenzal ‘dementia precox,’ as of 1918, and five years later. Further studies of the psychiatric aspects of influenza. Am J Psychiatry 1926; 4: 469–529. 2 Nunn K P, Lask B, Cohen M. Viruses, neurodevelopmental disorder and childhood psychosis. J Child Psychol Psychiatry 1986; 27: 55–64. 3 Mednick S A, Machon R A, Huttunen M O, Bonett D. Adult schizophrenia following prenatal exposure to an influenza epidemic. Arch Gen Psychiatry 1988; 45: 189–92. 4 Murray R M, Jones P, O’Callaghan E, Takei N, Sham P. Genes, viruses, and neurodevelopmental schizophrenia. J Psychiatr Res 1992; 26: 225–35. 5 Takei N, Sham P C, O’Callaghan E, Glover G, Murray R M. Schizophrenia: increased risk associated with winter city birth—a case-control study in 12 regions within England and Wales. J Epidemiol Community Health 1995; 49: 106–09. 6 Kirch D G. Infection and autoimmunity as etiologic factors in schizophrenia: A review and reappraisal. Schizophr Bull 1993; 19: 355–70. 7 King D J, Cooper S J, Earle J A P et al. A survey of serum antibodies to eight common viruses in psychiatric patients. Br J Psychiatry 1985; 147: 137–44. 8 Rantakallio P, Leskinen M, von Wendt L. Incidence and prognosis of central nervous system infections in a birth cohort of 12 000 children. Scand J Infect Dis 1986; 18: 287–94. 9 Rantakallio P. Groups at risk in low birth weight infants and perinatal mortality. Acta Paediatr Scand 1969; 193 (Suppl.): 1–71. 10 Rantakallio P. The longitudinal study of the Northern Finland birth cohort of 1966. Paediatr Perinat Epidemiol 1988; 2: 59–88. 11 Isohanni M, Partanen U, Moring J, Jones P, Räsänen P, Koiranen M. The validity of diagnoses of first-onset schizophrenia in the Finnish Hospital Discharge Register: a comparison of clinical and scientific views in a national birth cohort sample. Soc Psychiatry Psychiatr Epidemiol (In Press). 12 Rantakallio P, von Wendt L, Koivu M. Prognosis of perinatal brain damage: A prospective study of a one year birth cohort of 12 000 children. Early Hum Dev 1987; 15: 75–84. 13 Rantakallio P, von Wendt L. Mental retardation and subnormality in a birth cohort of 12 000 children in Northern Finland. Am J Ment Defic 1986; 90: 380–87. 14 von Wendt L, Rantakallio P, Saukkonen A-L, Mäkinen H. Epilepsy and associated handicaps in a 1 year birth cohort in Northern Finland. Eur J Pediatr 1985; 144: 149–51. 15 Sorri M, Rantakallio P. Prevalence of hearing loss at the age of 15 in a birth cohort of 12 000 children from Northern Finland. Scand Audiol 1985; 14: 203–07. 16 Beaglehole R, Bonita R, Kjellström T. Basic Epidemiology. Geneva: World Health Organization, Macmillan/Eastern Press 1993. 17 Rantakallio P, Lapinleimu K, Mäntyjärvi R. Coxsackie B5 outbreak in a newborn nursery with 17 cases of serous meningitis. Scand J Infect Dis 1970; 2: 17–23. 18 Rantakallio P, Saukkonen A-L, Krause U, Lapinleimu K. Follow-up study of 17 cases of neonatal Coxsackie B5 meningitis and one with suspected myocarditis. Scand J Infect Dis 1970; 2: 25–28.

CHILDHOOD CNS INFECTIONS AND ADULT ONSET PSYCHOSES 19

Crow T J. Prenatal exposure to influenza as a cause of schizophrenia. There are inconsistencies and contradictions in the evidence. Br J Psychiatry 1994; 164: 588–92. 20 Lewis S W, Owin M J, Murray R M. Obstetric complications and schizophrenia: methodology and mechanisms. In: Schulz S C, Tamming C A (eds). Schizophrenia: A Scientific Focus. New York: Oxford University Press, 1988, pp. 56–68. 21 Buka S L, Tsuang M T, Lipsitt L P. Pregnancy/delivery complications and psychiatric diagnosis. A prospective study. Arch Gen Psychiatry 1993; 50: 151–56. 22 Jones P, Rantakallio P, Hartikainen A-L, Isohanni M, Sipilä P. Does schizophrenia result from pregnancy, delivery and perinatal complications? A 28 year study in the 1966 North Finland birth cohort. In: European Psychiatry. The Journal of the Association of European Psychiatrists 1996, 11 (Suppl. 4): 242S.

23

843

Jablensky A, Sartorius N, Ernberg G et al. Schizophrenia: manifestations, incidence and course in different cultures. A World Health Organization ten-country study. Psychol Med Monogr 1992; (Suppl. 4): 1–91. 24 Druyts-Voets E, Van Renterghem L, Gerniers S. Coxsackie B virus epidemiology and neonatal infection in Belgium. J Infect 1993; 27: 311–16. 25 Kaplan M H, Klein S W, McPhee J, Harper R G. Group B coxsackievirus infections in infants younger than three months of age: A serious childhood illness. Rev Infect Dis 1983; 5: 1019. 26 Susser E G, Lin S P. Schizophrenia after exposure to the Dutch Hunger Winter of 1944–1945. Arch Gen Psychiatry 1992; 49: 983–88.

(Revised version received October 1996)