But for some people in situations typically considered stress-free -- perhaps even pleasant -- the levels of one hormone, cortisol, may even rise. Scientists have long believed that cortisol levels increase in times of stress and decrease as the stress is eased. The new finding is puzzling researchers but also pointing them to an entirely new area for future research.

The work was described August 4 by Janice Kiecolt-Glaser, professor of psychiatry and psychology at Ohio State University, at the annual meeting of the American Psychological Association in Washington, D.C.

The new work stems from a series of experiments done more than a decade ago and intended to look for physiologic changes caused by stressful situations. A group of about 90 newlywed couples took part in the study at Ohio State's Clinical Research Center. The couples completed questionnaires and then were asked to discuss several areas of disagreement regarding their marriage.

Blood samples were taken from each subject at the beginning of the session and at 30-minute intervals until they were discharged. When researchers analyzed the data, they found changes in hormones and other bloodstream components that could indicate a weakening of the immune status.

Today's findings, however, arise from a new analysis of data from blood samples taken immediately after those "conflict" discussions. Following a short "transition" period, the newlyweds had been asked to discuss the history of their relationships - how they met, what attracted them to each other, how did they decide to marry.

"For most people, not surprisingly," explained Kiecolt-Glaser, "this was a pretty positive interaction. And they were coming down off the earlier discussion which was usually seen as a negative emotional experience."

When Kiecolt-Glaser and colleagues Ronald Glaser, professor of molecular virology, immunology and medical genetics, and William Malarkey, professor of internal medicine, looked at the levels of cortisol in the blood however, they were surprised.

"In 75 percent of the subjects, the hormone levels had fallen just as we expected - 26 percent on average for the men and 35 percent for the women," Kiecolt-Glaser said. "But in 25 percent of them, cortisol levels stayed relatively the same or, in some cases, actually went up."

Normally, levels of cortisol drop after we wake in the morning. The experiments were done in midmornings and the newlyweds were "coming down" off the negative discussion, which should have forced cortisol levels lower.

But in one of every four people in the study, the cortisol levels failed to drop. In some people, the levels even rose. The findings were especially interesting when coupled with information from follow-up surveys of the participants. The researchers had tracked down each person to check their current marital status and to find out if the newlyweds were still together.

They found that the women whose cortisol levels rose during the discussions of their relationship history were twice as likely to have been divorced from their husbands. No similar relationship appeared among men whose levels had risen.

The study also looked at the exact word choice the couples used in describing their relationship. They used an instrument called the Linguistic Inquiry and Word Count program which lists both "positive" and "negative" words - 261 which describe optimism, energy and positive feelings and 345 words suggesting anxiety, fear, sadness, depression and anger.

Cortisol levels for three-fourths of the men in the study dropped - on average 26 percent. They used significantly more words considered "positive" than did their counterparts whose cortisol remained steady or even rose. There was no similar trend in positive word choice among women whose cortisol had dropped.

Women whose cortisol levels failed to drop, or even rose, however tended to use more negative words as they discussed their relationship history. "What I think is happening here is that in some ways, cortisol levels may serve as a bellwether of what's going to happen," whether women's marriages would survive, Kiecolt-Glaser said.

The most important finding, she says however, is that this proves that "positive" interactions, along with their health implications, deserve as much study as "negative" interactions have garnered.

The National Institutes of Health supported the study.

Note: If no author is given, the source is cited instead.

• Stress
• Gender Difference
• Racial Issues

• Stress (medicine)
• Emotional detachment
• Panic attack
• Social cognition
• Hormone
• Conflict resolution research