There is no hypothesis shortage out on the frontier where infectious
Two more studies now pave a little more of the road to that borderland, but
they also show how much farther the science needs to travel.
Over the last few decades intriguing, if inconclusive, studies suggested
that maternal infection with influenza, genital or reproductive pathogens, or
toxoplasmosis during pregnancy may be linked to higher rates of schizophrenia
among off-spring. Studies of Swedish health registry data and U.S. military
personnel now provide further hints that risk may extend to infections
acquired in childhood.
The Swedish study examined health records of children born from 1973 to
1985 to identify patients who were hospitalized for central nervous system
(CNS) infections between birth and age 12 and were also admitted for
nonaffective psychotic illnesses after their 14th birthday. They found 8,985
out of 1.2 million children were exposed to CNS infection, including 2,435
bacterial and 6,550 viral infections; 2,269 subjects had been hospitalized for
nonaffective psychosis; and 23 subjects recorded both CNS infections and
"We found a weak association between viral CNS infections in
childhood and later development of nonaffective psychotic disorders,"
wrote Christina Dalman, M.D., Ph.D., of the Department of Psychiatric
Epidemiology at Stockholm's EPI/Karolinska Institutet, and nine colleagues in
the January American Journal of Psychiatry (AJP).
Bacterial infections occurred in only four subjects with psychosis and
accounted for no increased risk. However, the 19 subjects with viral CNS
infections had a greater risk for developing psychosis (risk ratio [RR] = 1.5,
confidence interval = 1.0-2.4). Adjustment for sex, age, urbanicity, or
parental psychosis reduced the risk slightly but brought the results below the
level of statistical significance (see chart).
Enterovirus infection (RR=1.0) was not associated with later development of
psychosis, but mumps (RR=2.7) and cyto-megalovirus (RR=16.6) were. The authors
cited prior research indicating that mumps virus can invade the brain,
infecting ependymal cells, persisting in neurons, and perturbing development
of or causing problems with memory, learning, or sensory and motor abilities.
A mumps vaccine was introduced in Sweden in 1982, before which mumps was the
most common cause of aseptic meningitis or mild encephalitis, according to
The number of cases may be small, but the fact that the study focuses on
CNS infection that caused delirium, headache, or meningitis worthy of
hospitalization is a key point, said Alan Brown, M.D., an associate professor
of clinical epidemiology at the Mailman School of Public Health and an
associate professor of clinical psychiatry at New York State Psychiatric
Institute at Columbia University. Brown has studied the connection between
prenatal infections and schizophrenia, and wrote an editorial accompanying the
"If anything, the study might have missed milder cases, since some
viral infections would not warrant hospitalization," he told
Psychiatric News. "Plus, some patients may have had such
serious neurological effects from their infections that psychiatric disorders
might be masked."
The CNS infections were spread across the entire 12-year age span studied,
which might make it hard to tie infection to specific neurodevelopmental
timing or mechanisms, said Brown. Perhaps infection causes some brain
inflammatory process, especially among glial cells, leading to fibrosis in
brain tissue, he speculated.
The second report, also in the January AJP, is a pilot study that
compared antibodies for nine common infectious agents from 180 members of the
U.S. Army, Navy, and Air Force discharged from military service between 1992
and 2001 after a diagnosis of schizophrenia. The armed forces routinely take
serum samples when a person enlists and about every two years afterward. The
180 patients were matched with 532 healthy military service members with no
Researchers led by Col. David Niebuhr, M.C., a physician and chief of the
Department of Epidemiology at Walter Reed Army Institute of Research, measured
immunoglobulin G (IgG) levels of antibodies to Toxoplasma gondii and
to several common viruses. T. gondii IgG was associated with a
significantly higher hazard ratio (HR=1.24, p<0.01) for schizophrenia.
Human herpes virus 6 also had a significant positive association
However, while 58 percent of serum samples were taken at least six months
before diagnosis, the only significant associations between T. gondii
IgG and schizophrenia occur within six months prior to diagnosis or after that
"That means the cause-and-effect mechanism could work either
way," said Brown. "T. gondii might cause schizophrenia,
but a lifestyle associated with schizophrenia might just as well increase the
risk of contracting toxoplasmosis."
Niebuhr's team found little evidence of seroconversion, indicating that
infection generally occurred before the first specimen had been drawn, he told
Psychiatric News. Given that, he said, "... from our data, we
cannot address the issue of prodromal behaviors resulting in T.
The report had a few other issues that Brown hoped will be addressed when
Niebuhr and colleagues expand their research to cover all 1,200 subjects they
have identified so far. Serial data on antibody levels are not reported by
individual subjects, so it's hard to tell when they became seropositive, said
"Most of the serum specimens were negative for both cases and
controls, but we found that cases were more likely to have higher antibody
levels than controls," explained Niebuhr. "Therefore, the sera for
all groups contain both positive and negative specimens, and both positive and
negative specimens were included in the analyses."
The expanded study will include more cases and more specimens per case,
which might help tease out the temporal relationship of infection and onset of
schizophrenia, said Niebuhr. But if new infection occurs not long before or
after schizophrenia, a clear answer may not be possible without further
investigation, he said.
Although independent replications are warranted, both studies are"
promising," said Brown, and support an infectious etiology for
schizophrenia once thought implausible.
The size of the effects observed was small, compared with the 20 percent of
the U.S. population exposed to T. gondii, for example, he said. If
elevated antibodies were shown prospectively to be a risk factor for
schizophrenia, studies like these would imply the potential for
"Given that some infections linked to risk of schizophrenia may be
treatable or prevented by vaccines, we may already be preventing cases of
schizophrenia by administering mumps vaccine during childhood," he said."
An excellent test of the toxoplasma hypothesis would be to see if
giving antitoxoplasmosis antibiotics to people seropositive for T.
gondii decreased antibodies and improved symptoms of
Whatever the outcome, such a trial might be another step toward ruling in
or out a postnatal role for the infection in the origins of schizophrenia.
"Infections in the CNS During Childhood and the Risk of
Subsequent Psychotic Illness: A Cohort Study of More Than One Million Swedish
Subjects" is posted at<http://ajp.psychiatryonline.org/cgi/content/full/165/1/59>."
Selected Infectious Agents and Risk of Schizophrenia Among U.S.
Military Personnel" is posted at<http://ajp.psychiatryonline.org/cgi/content/full/165/1/99>.▪