To determine whether stress is associated with risk of bacterial vaginosis (BV) in pregnant women.
Prospective cohort study.
The prenatal care clinics at the University of North Carolina. The residents' clinic sees mostly government-insured and uninsured women, and the physicians' clinic sees mostly those with private health insurance.
A total of 897 women gave samples for BV analysis. Study participants were 22% African-American, 68% white; 24% unmarried and 44% nulliparous. More than half had completed college.
Women completed two questionnaires and two interviews reporting stress and psychological aspects of their lives. Measurement scales included the Sarason life events questionnaire, the Cohen perceived stress scale, Spielberger state-trait anxiety, the John Henryism coping style and the Medical Outcomes Study social support inventory. Two stress hormones, corticotrophin-releasing hormone and cortisol, were also measured.
Main outcome measures
BV at 15-19 and 24-29 weeks of gestation was diagnosed by Gram's stain.
Women in the highest quartile of stress measures, particularly state anxiety (OR=2.0, 95% CI 1.2-3.3), perceived stress (OR=2.4, 95% CI 1.5-3.9) and total life events (OR=2.0, 95% CI 1.3-3.2), had the highest risk of BV. Adjustment for confounders, especially age, race, and income, reduced these associations (state anxiety: OR=1.3, 95% CI 0.7-2.4; perceived stress: OR=1.4, 95% CI 0.8-2.5; total life events: OR=1.3, 95% CI 0.7-2.4). No clear pattern of association was seen between stress hormones and BV.
Few associations between stress and BV were seen after adjustment for confounders.