Maternal exposure to influenza and paranoid schizophrenia

Abstract

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 between in utero exposure to influenza and subsequent development of paranoid schizophrenia we hypothesised was not supported by our data.

Introduction

A majority of epidemiological studies from different countries in both hemispheres have reported a relationship between maternal exposure to influenza during gestation and the development of schizophrenia in the offspring. Ten studies have shown a positive association between prenatal exposure to influenza and schizophrenia, three have produced equivocal results, while three found no such association (reviewed by Wright et al., in press). Two recent studies have further shown that this relationship is stronger in more strictly defined schizophrenia. In a Danish study (Takei et al., 1996), the association was more conspicuous in those with narrowly defined schizophrenia, while national data from the Netherlands (Takei et al., 1995b) also demonstrated a more marked association with typical schizophrenia.

However, the latter two studies used a crude distinction between `typical' and `atypical' schizophrenics and did not examine further whether any specific subtype of schizophrenia was responsible for this association. The effect of prenatal exposure to influenza in the development of schizophrenia appears more readily detectable in females (Takei et al., 1993Takei et al., 1994Takei et al., 1995a). Furthermore, Machon et al. (1995)claimed that schizophrenics who were exposed to influenza in the second trimester have significantly more symptoms of delusions of jealousy, delusions of reference and suspiciousness, than do schizophrenics exposed to influenza in the first or third trimester. Taken with the evidence that female schizophrenics more frequently manifest a paranoid picture than males (Goldstein et al., 1990), these findings suggest that paranoid schizophrenia may be the subtype responsible for the association previously found between prenatal exposure to influenza and narrowly defined schizophrenia. Therefore, we decided to test the hypothesis that the risk increasing effect of in utero exposure to influenza is associated particularly with the paranoid schizophrenia subtype.