Noriyoshi Takei

Relationship between in utero exposure to influenza epidemics and risk of schizophrenia in Denmark

Several recent epidemiological studies suggest that exposure to influenza during gestation increases the risk of later developing schizophrenia. Inconsistency exists, however, particularly in studies that have examined the relationship between the prevalence of influenza and the monthly number of schizophrenic births, over many years. Our sample (N = 9462) was obtained from a Danish computerized case register, and consisted of schizophrenia patients born between 1915 and 1970, and first admitted to Danish psychiatric hospitals between 1971 and 1991. The study sample was chosen to represent incidence cases to allow us to calculate the population attributable risk fraction (PAF). The temporal correlation of fluctuations in the prevalence of influenza and fluctuations in the monthly number of preschizophrenic births was examined using a Poisson regression analysis. Exposure to influenza 4 months prior to birth (i.e., about the 6th month of gestation) was significantly associated with an increased risk of later schizophrenia, especially for narrowly defined schizophrenia. The number of schizophrenic births was found to have risen by 12% (95% confidence interval: 1-24%) for every 100,000 cases of influenza in the 4th month before birth. Our model indicates the PAF to be 1.4%, that is, only 1.4% of the whole schizophrenic sample is attributed to prenatal exposure to influenza. Although maternal exposure to influenza during midgestation is not a major risk factor for schizophrenia, the elucidation of its causal mechanism may open the avenue to understanding the neurodevelopmental origins of the disease.

Genes, viruses and neurodevelopmental schizophrenia

Recent neuroimaging and neuropathological studies suggest a developmental origin for schizophrenia. Some cases may, therefore, be caused by a genetic defect in the specification of brain development. Early environmental hazards such as obstetric complications, and maternal exposure during pregnancy to influenza epidemics, have also been found to increase the risk of later schizophrenia. The relationship between the prevalence of influenza and birth date has been found more consistently for female than male schizophrenics. Female schizophrenia is also associated with a higher risk of schizophrenia in first degree relatives. This raises the question of whether part of the genetic predisposition to schizophrenia may comprise an abnormal reaction to maternal influenza.

Small head circumference at birth in schizophrenia.

The growing evidence for neurodevelopmental basis to schizophrenia has focused attention on the prenatal development of individuals who later develop the illness. Several previous studies have shown reduced birth weight (BW) in schizophrenics and one recently reported smaller birth head circumference (BHC). The current study compared 67 DSM-III-R schizophrenics and a general population group of 1640, using information obtained from contemporaneous birth records. When gestational age and gender were controlled for, no significant difference was found in BW between the schizophrenics and the comparison population. However, the preschizophrenics showed significantly smaller BHC for gestational age, suggestive of slower fetal brain growth.

Age–period–cohort analysis of the incidence of schizophrenia in Scotland

Studies examining a possible decline in the incidence of schizophrenia over the last two to three decades have paid little attention to the possible role of birth cohort effects. We collected data on a Scottish national sample of all schizophrenic patients, admitted for the first time between 1966 and 1990 (N = 11348; male = 6301). In an Age-Period-Cohort analysis, a full model, incorporating three factors, had a substantially better fit to the data than other models (especially, an Age-Period model), providing clear evidence of the presence of a cohort effect. After adjustment for the effects of age and period, there was a 55% reduction in the rate of schizophrenia in men and a 39% fall in the number of women over the 50-year birth period from 1923 to 1973. The marked decline in the first admission rates observed in Scotland cannot, however, be attributed entirely to this cohort effect. Rather, a greater proportion of the declining first admission rates (88%) is ascribed to the period effect (i.e. artefactual or causally related cross-sectional effects). Nevertheless, the fact that a birth-cohort effect accounts for part of the declining incidence, suggests that causal environmental factors operating early in life have been diminishing in intensity.

Does prenatal exposure to influenza in mice induce pyramidal cell disarray in the dorsal hippocampus

Epidemiological studies point to an association between prenatal exposure to influenza and later schizophrenia. Such studies are consistent with neuropathologic reports demonstrating cytoarchitectural abnormalities in the hippocampus and parahippocampal gyrus suggestive of second trimester developmental anomalies. The hypothesis that prenatal exposure to influenza in the second trimester may induce hippocampal pyramidal cell disarray in mice was investigated. Between days 9-16 of pregnancy, 35 Balb/c mice were intranasally inoculated with either a mouse-adapted or non mouse-adapted pool of Influenza A/Singapore/1/57 (H2N2), and 10 controls were inoculated with normal saline. Offspring were sacrificed on day 21 postpartum. Microscopic examination of the CA1-CA2 junctional areas in the offspring of mice exposed to influenza failed to demonstrate excess pyramidal cell disarray when compared with influenza-free, age matched controls. There was evidence that disarray was greater among those exposed on day 13 of pregnancy. Analyses of the data by sex and severity of maternal infection failed to reveal any significant effects.

Maternal exposure to influenza and risk of schizophrenia: A 22 year study from The Netherlands

We investigated any effect of prenatal exposure to influenza during gestation on subsequent risk of schizophrenia using a national sample from The Netherlands. Dates of births of all Dutch-born schizophrenia (ICD-9) patients (n = 10,630) admitted to hospitals for the first time between 1970 and 1992 were examined in relation to the occurrence of influenza epidemics between 1947 and 1969. As a measure of prevalence of influenza, the number of deaths from influenza per month in The Netherlands was used. A Poisson regression analysis revealed that an increase in the prevalence of influenza 3 months prior to birth was followed by an increase in births of preschizophrenics, although this fell outside statistical significance (p = .11). However, the effect became marked in typical schizophrenics (n = 4726), but not in less typical cases (n = 5389). For typical schizophrenics, the parameter estimate derived from the regression model indicates that there was a 10% increase (95% confidence interval: -1 to 22%) in preschizophrenic births for every 500 deaths from influenza 3 months before birth.

Seasonality of birth in schizophrenia: the effect of regional population density

It has been suggested that winter birth in urban areas may be associated with an increased risk of schizophrenia. Data on 4139 patients with a hospital diagnosis of schizophrenia were obtained from a survey of public departments of psychiatry in metropolitan France. Place of birth was classified according to the administrative subdivision of French territory into 95 départements. Density of population in the decade and in the département of birth was identified using successive census data. A 20% excess (OR, 1.20; 95% CI, 1.03-1.40; p = 0.02) of winter births (January-March) was found among patients born in highly densely populated areas (> 136 inhabitants/km2) compared to those born in other areas. These results suggest that seasonal early environmental risk factor(s) linked to schizophrenia predominantly operate in urban areas.

PRENATAL EXPOSURE TO INFLUENZA EPIDEMICS AND RISK OF MENTAL RETARDATION

This study was undertaken to determine whether prenatal exposure to influenza epidemics increases the risk of mental handicap. The monthly birth frequencies of 827 first-admission individuals (mean age at admission 13 years) with a primary diagnosis of non-specific mental retardation, discharged from psychiatric hospitals in England and Wales, were examined in relation to the monthly death rates from influenza over the period 1953–1980. The relative risk of developing mental handicap when exposed to influenza epidemics during mid-gestation was assessed by a generalized linear model. Increased death rates from influenza, a measure of prevalence of the infection, were significantly associated with an increase in births of mentally handicapped individuals 6 months later. For every 1000 female deaths from influenza there was a 17% increase in births of mentally handicapped individuals 6 months later. Maternal exposure to influenza at approximately the third to fourth month of gestation may be risk factor for developing mental handicap.

Maternal exposure to influenza and paranoid schizophrenia

Previous research has suggested that schizophrenics exposed to influenza in the second trimester have more delusions of jealousy, delusions of reference and suspiciousness. We therefore hypothesised that the risk-increasing effect of in utero exposure to influenza would be particularly demonstrable in paranoid schizophrenia. We studied patients with an ICD diagnosis of schizophrenia in England and Wales who were born each month between 1923 and 1965 (N = 17,247. Chi-square test for trend showed that an increase in influenza exposure level during the fifth month of gestation was accompanied by an increase in the proportion of patients with paranoid schizophrenia. However, logistic regression analysis including sex, seasonality and birth period in the model resulted in the loss of any significant association between in utero exposure to influenza and the development of paranoid schizophrenia, the loss of this significance being mainly accounted for by birth period. Therefore, the association in utero exposure to influenza and subsequent development of paranoid schizophrenia we hypothesised was not supported by our data.