The serological evidence for maternal influenza as risk factor for psychosis in offspring is insufficient: critical review and meta-analysis.

Maternal influenza during pregnancy has been suggested to increase the psychosis risk for the offspring. This hypothesis has been tested using "ecological" studies, which examined the risk for individuals born after epidemics, and "serological" studies, based on serological evidence. A study of the latter type obtained an increased schizophrenia risk for individuals exposed during the first trimester. A second study found a relationship between influenza at any time during gestation and risk for bipolar disorder with psychotic features. The aims of this paper are to assess the validity of the serological studies and to evaluate the combined results of ecological and serological investigations using meta-analysis. The serological studies turned out to be of limited validity, because they utilized a single serum specimen. Since influenza antibodies can remain positive for years after infection, many mothers of cases may have been infected before pregnancy. For an adequate timing of exposure one needs an acute and a convalescent specimen, obtained 10-20days later. Meta-analysis with respect to schizophrenia: we pooled the results of the single serological investigation and 8 ecological studies related to the 1957 pandemic (with negative results) and found that the first investigation carried hardly any weight. Bipolar disorder: we pooled the results of the serological investigation and three other studies and obtained a mean, weighted odds ratio of 1.34 (95% CI 0.78-2.29) for individuals possibly exposed during prenatal life. The evidence for gestational influenza as psychosis risk factor is insufficient.