A Theory That Raises Questions

"It raises a question but does not provide an answer," Insel said of the newest research. "We are not ready to jump from any of this to the policy dimension."

The Evidence

Hints about the schizophrenia-winter connection have been around for decades.

Epidemiologists have found that children of women who were pregnant during widespread flu epidemics seemed to have higher risk for schizophrenia. But critics have said there's insufficient evidence to assert a causative relationship because such studies did not confirm that the pregnant women had the flu.

Besides, other factors seemed to trigger the same effect: One wartime study found that pregnant women whose husbands died were more likely to have children who later developed schizophrenia.

A research breakthrough came in 2004 when Alan Brown, a psychiatrist at Columbia University and the New York State Psychiatric Institute, got access to data from a study that collected blood samples between 1959 and 1966 from thousands of pregnant women at different stages of their pregnancy.

In an analysis that compared blood samples from the mothers of 64 children who went on to develop schizophrenia with blood samples of similar mothers whose offspring did not, Brown and his colleagues showed that women who had higher levels of influenza antibodies in their first or second trimester of pregnancy had offspring who were three to seven times more likely to develop schizophrenia.

Brown calculated that if the women had not had the flu during pregnancy, 14 percent of the schizophrenia cases could have been prevented, an effect he calls potentially enormous for a disease believed to have several complex genetic and environmental factors.

Brown and other researchers also began seeking evidence that mothers who contracted other infections during pregnancy also had children who were at increased risk of schizophrenia; they also began looking for common pathways between the infections. One appeared to be an immune system protein, or cytokine, called interleukin-6; cytokines are activated not only through infections but also through stress.

Brown, Insel and Patterson all hasten to point out that, while schizophrenia is a relatively rare disease, as many as 10 percent of women get the flu while they are pregnant, and many pregnant women contract other infections and experience stress. This suggests that the vast majority of women who have their immune systems triggered during pregnancy will not have children who develop schizophrenia.

Insel also cites evidence that genetics may play a more dominant role than the environment in determining who gets schizophrenia: Studies of identical twins show that when one child develops schizophrenia, the other has a 50 percent chance of developing the disorder, too.

However, some of the increased risk among identical twins may be a result of maternal infections during pregnancy -- and not genetics, Patterson argues. That's because those identical twins who share a common placenta -- and who are, therefore, more likely to receive the same maternal cytokines -- seem to have a higher risk of schizophrenia than identical twins who do not share a common placenta.

(While the science assigns no blame to mothers for these risks, the long history of blaming parents for the disorder suggests it might be useful to note that people's immune systems are beyond their control, and pregnant women cannot be held responsible for the flu-schizophrenia link.)