Background: Greater levels of fear, anxiety or isolation (psychosocial stress) have been hypothesized to negatively alter the sympathetic-parasympathetic autonomic nervous system and suppress immune function, physiologic effects that in turn might impact breast tumor aggressiveness. There is a paucity of cohort data relating psychosocial stress to breast tumor aggressiveness. We examined associations between patient-reported psychosocial stress with aggressive breast cancer in a cross-sectional study of 989 recently diagnosed breast cancer patients (397 non-Hispanic white, 411 non-Hispanic black, 181 Hispanic). We examined associations of psychosocial stress with breast cancer aggressiveness and whether racial/ethnic differences in psychosocial stress might account for disparities in breast cancer aggressiveness. A major assumption of these analyses was that patients reporting greater psychosocial stress at interview (2–3 months after diagnosis) would have also reported relatively greater psychosocial stress had they been interviewed prior to diagnosis.

Methods: Psychosocial stress was assessed using the Cohen perceived stress 4 item subscale, the UCLA felt loneliness scale (3 items), and the Cockburn psychological consequences scale (12 items). These three scales loaded onto a single factor in factor analysis, and were used to create a single, standardized psychosocial stress score (Cronbach's alpha=0.56). Breast cancer aggressiveness was defined as Hormone (estrogen and progesterone) receptor negative tumors and tumors high in histologic grade. Differences in psychosocial stress scores were evaluated via t-tests, and logistic regression was used to estimate odds ratios for breast cancer aggressiveness corresponding to a one standard deviation increase in psychosocial stress score. All reported p-values are two-sided. Because aggressive breast cancer tends to lead to a later stage at diagnosis and an earlier age at diagnosis, adjustment for these variables may inappropriately attenuate the association of interest between psychosocial stress and tumor aggressiveness. Nonetheless, because psychosocial stress was measured after diagnosis, we adjusted for age and stage in some models.

Results: Patient reported psychosocial stress was one-third of a standard deviation higher for patients with receptor negative vs. positive disease (mean score difference = 0.32, p=0.0003), and higher for patients with high vs. low or intermediate grade disease (difference = 0.17, p=0.03). Compared to nH-Whites, psychosocial stress scores were higher for nh-Black patients (difference = 0.22, p=0.006), and higher for Hispanic patients (difference = 0.44, p<0.001). In logistic regression, the unadjusted association of greater psychosocial stress with receptor negative disease (OR=1.38, p<0.0005), was somewhat attenuated after adjusting for sociodemographics (OR=1.27, p=0.01), and further attenuated with additional adjustment for age and stage at diagnosis (OR=1.22, p=0.06). The association between psychosocial stress and high grade disease that was evident in unadjusted (OR=1.18, p=0.03) and partially adjusted analyses (OR=1.17, p=0.05) disappeared with additional adjustment for age and stage at diagnosis (OR=1.10, p=0.30). In path models, there were significant indirect paths from minority ethnicity to both receptor negative and high grade disease via psychosocial stress (p=0.02 and 0.02, respectively), accounting for roughly 10% of ethnic differences in tumor aggressiveness.

Conclusions: In this cross-sectional study, greater levels of fear, anxiety or isolation were associated with more aggressive breast cancer, consistent with a potential role for psychosocial stress in the etiology of breast cancer aggressiveness.

Citation Information: Cancer Epidemiol Biomarkers Prev 2011;20(10 Suppl):A91.