Email Icon Facebook icon Twitter Icon GooglePlus Icon Contact

User Top Menu

New Study Complicates Infection Link to Psychosis

1 Dec 2015

December 2, 2015. The previously reported link between a mother's infection during pregnancy and risk for psychosis in her children may not be so straightforward, according to a study of nearly two million people published online August 24 in Schizophrenia Bulletin. Led by Asa Blomstrom and Christina Dalman at the Karolinska Institute in Stockholm, Sweden, the study did not find an association between maternal infection and non-affective psychoses such as schizophrenia in offspring. They did find a link, however, when looking among mothers with a psychiatric illness themselves. This suggests that a "double hit" of a mother's vulnerability and infection work together to increase risk.

"The most interesting finding is this evidence of a synergistic effect between infection and the mother's history of psychiatric illness," said Alan Brown of Columbia University in New York City, who was not involved in the study and who noted that this history may reflect genetic as well as environmental risk factors.

Maternal infections during pregnancy are a long-standing risk factor for psychosis, but the exact nature of the link remains unclear. Brown and colleagues first pegged maternal rubella infections as increasing risk for schizophrenia (Brown et al., 2000), and since then other types of infections have been implicated, too. This has raised the possibility that risk stems from a mother's ongoing immune activation during pregnancy, which then affects the fetus in utero.

However, the lack of an association between maternal infection and psychosis risk in this well-powered study may prompt calls to re-evaluate the role of maternal immune activation in psychosis risk.

"This is a really good paper; it's a clever paper. They've done some new analyses, and they wind up challenging the accepted dogma," said John McGrath of the University of Queensland in Brisbane, Australia. McGrath was not involved in the study.

Brown, on the other hand, told SRF, "I would say it provides further support for a hypothesis that has been around for a long time." He commented that the study only counted hospitalizations for infection, which are themselves rare.

Blomstrom didn't think her results warranted chucking out the maternal immune hypothesis of psychosis and noted the need to replicate the results first, and the finding among mothers with psychiatric illnesses.

"We got results indicating a direct intra-uterine effect of something affecting the child if the mother had an infection during pregnancy—but only among the vulnerable mothers," Blomstrom told SRF.

Unclear link

Previous studies of maternal infection during pregnancy and risk for psychosis have given mixed results. "Most have been rather small, and the results have been pointing in different directions," Blomstrom said.

Even among those studies that find a link, the interpretations are unclear: Do the infections themselves somehow derail fetal brain development in a way that paves the way to psychosis? This pregnancy-specific scenario is supported by blood samples taken from pregnant mothers: Brown has reported increased odds of schizophrenia in the children of mothers who were infected during pregnancy with influenza or Toxoplasma gondii (Brown et al., 2004; Brown et al., 2005), and a recent study found a link with a pregnant mother's level of inflammation (see SRF related news report).

Alternatively, the link may not have to do with an infection per se, but rather a parent's vulnerability to landing in the hospital for an infection. Consistent with this, a recent Danish study found that the risk for psychosis ratcheted up even when the father of the affected offspring had an infection, or when the infections occurred in the mother before or after pregnancy (Nielsen et al., 2013).

The new study aimed to clarify the nature of the link by turning to the Swedish national registry databases, which keep detailed, though anonymous, tabs on health, socioeconomic circumstances, education, and other demographic information on their residents. These registries offer hope of disentangling true risk factors from the many things correlated with them.

"We have so much information available in these data; thus, we can adjust for many of the possible confounders, which other studies have not been able to do," Blomstrom said.

Adjusting associations

The study considered all children born in Sweden between 1978 and 1997, who numbered 1,971,623. Of these, 8,330 people were eventually diagnosed with non-affective psychosis, half of which had schizophrenia. Records of both the mother and father were examined for episodes of hospitalization for infection, psychiatric diagnoses, and other variables.

It initially appeared that mothers hospitalized with an infection during pregnancy were 1.26 times more likely to have a child develop psychosis than those without an infection. But this increase in risk shrank to a non-significant 1.06 when the researchers took into account psychiatric illnesses in the parents, socioeconomic status, and the tendency for parents to seek care at a hospital. Even when the researchers analyzed infections by type (viral or bacterial), or stage of pregnancy, no sign of an association emerged.

Among the subset of mothers diagnosed with a psychiatric illness, however, infection during pregnancy nearly quadrupled psychosis risk in children. This increase in risk was more than the sum of the individual effects of psychiatric illness and infection.

Infection exerted its effect strictly through the mother, as a father's infection did not contribute to risk. This argues against general vulnerability to infection as driving psychosis risk.

Infections during pregnancy carried the most risk in this subset of mothers, though a weak yet significant increase in risk was detected for maternal infections that occurred outside of pregnancy. This suggests that these individuals have more infections than others in general, thus lending support to the idea that proneness to infection may contribute to risk.

Missed infections

The study captures only the most severe infections, thus missing the majority who are managed without needing to go to the hospital. Thus, the study does not detract from the blood sample studies that detect specific maternal infections, Blomstrom said.

The study also averaged across effects of different kinds of infections, some of which might impact risk for psychosis more than others.

"I don't think you're going to see large effects because there will be this averaging out," Brown said. "I think this is probably the best you could expect using this kind of registry-based method."

Just how the mother's psychiatric illness contributes to risk remains an open question. Though genetic factors are an obvious possibility, environmental factors associated with mental illness, such as smoking, substance abuse, or other behaviors, could also figure in to increase risk for psychosis.

"The challenge of the research community is to build on these new findings and to pull apart what they may be," McGrath said.—Michele Solis.


Blomstrom A, Karlsson H, Gardner R, Jorgensen L, Magnusson C, Dalman C. Associations Between Maternal Infection During Pregnancy, Childhood Infections and the Risk of Subsequent Psychotic Disorder—A Swedish Cohort Study of Nearly 2 Million Individuals. Schizophr Bull. 2015 Aug 24. Abstract