In this study, researchers assessed the women using a variety of measures: the depressive symptom scale, a perceived stress scale measuring experiences of stress and coping with stress in the past month; a questionnaire gauging how much social support the women had; tests for frequency of stressful social interactions; and a short survey of how happy the women and their partners were about the pregnancy.

Blood samples were taken to measure levels of two proteins, interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-a). Both are proinflammatory cytokines, chemical messengers that are mobilized when the body is injured or has an infection, and they cause inflammation in their effort to make repairs in the body.

When these proteins circulate without an infection to fight, the body experiences excess inflammation, which is associated with a variety of diseases depending on which cells are producing the proteins. Previous studies have shown that such inflammation during pregnancy can increase the risk of preterm birth and preeclampsia, a high blood pressure condition that can occur during the last half of pregnancy.

Depressive symptoms were associated with perceived stress, and women who were unhappy about their pregnancies had significantly more depressive symptoms than women who reported being happy that they were pregnant. In addition, women with less social support and more frequent hostile social interactions also had more depressive symptoms than did women with better support and more positive social interactions.

Overall, the women reporting more depressive symptoms had significantly higher levels of IL-6 in their blood than did women with fewer symptoms. The association between depressive symptoms and TNF-a was not as strong, but was still considered significant.

In both studies, the researchers assessed a variety of health behaviors and measures, such as body mass index, cigarette smoking, prenatal vitamin use and physical activity, to gauge whether these factors might affect the presence of inflammation markers. None of the measures had a significant effect, Christian said.

"This way, we took into account the potential for two different pathways. Stress can certainly affect health behaviors, which can affect immune function," she said. "That's why we assessed different health behaviors, to be sure the effects we see aren't better explained by something else. But here, we are seeing a physiological effect of stress and depression."

The next step will be to follow more women, for longer, to see if psychological factors during pregnancy can be linked directly to birth outcomes.

Source: Ohio State University