Introduction: Previous studies have reported that physical activity is associated with improved survival after a breast cancer diagnosis. There are few data, however, for the association of physical activity with survival in Hispanic women. The purpose of this study was to investigate the association of increased physical activity in the year prior to breast cancer diagnosis with all-cause and breast cancer mortality in Hispanic and non-Hispanic white women using data from the New Mexico Women's Health Study (NMWHS).

Study Population and Methods: The NMWHS was a population-based case-control study that included women diagnosed in New Mexico with invasive or in-situ breast cancers in 1992–1994. A total of 722 cases (339 Hispanic and 383 non-Hispanic white) participated in the study. Vital status and specific causes of death were ascertained through December 2008 via linkage with the New Mexico Tumor Registry (NMTR) and the National Death Index (NDI). Usual physical activity was assessed using a standardized questionnaire for the year prior to interview. After excluding women with a previous history of breast cancer or missing covariate data, 641 cases remained for analyses. Relative risks (RR) and 95% confidence intervals (CI) were estimated using Cox Proportional Hazards models, with time since interview as the underlying time metric, for quartiles of physical activity (MET-h/wk), adjusting for age, breast cancer treatment, menopausal status at time of diagnosis, stage of disease, number of years of supplementary estrogen use, body-mass index (BMI) at time of diagnosis, and energy intake levels in the year preceding diagnosis.

Results: During an average 12.0 years of follow-up, 216 deaths were ascertained (123 due to breast cancer). In Hispanic women, compared to 0–25.00 Metabolic Equivalent hours per week (MET-h/wk), the RRs (95% CI) were 0.75(0.42, 1.33) for 25.00–40.15 MET-h/wk, 0.46(0.24, 0.87) for 40.15–62.85 MET-h/wk and 0.41(0.22, 0.76) for >=62.85 MET-h/wk, with a p-trend value of 0.002 for all-cause mortality, and 0.70(0.32, 1.52), 0.47(0.18, 1.19) and 0.42(0.19, 0.93) with a p-trend value of 0.028 for breast cancer mortality. In non-Hispanic white women, the corresponding RRs (95% CI) were 0.60(0.33, 1.10), 1.09(0.63, 1.87) and 0.90(0.51, 1.61) with a p-trend value of 0.763 for all-cause mortality, and 0.65(0.28, 1.53), 0.97(0.44, 2.12) and 0.70(0.31, 1.60) with a p-trend value of 0.609 for breast cancer mortality. Stratified analysis suggests an added benefit for Hispanic women with BMI>=25 when their MET-hr/wk exceeds the median. The RR (95% CI) is 0.44(0.25, 0.78). The same benefit is not seen with non-Hispanic white women or in Hispanic women with BMI<25.

These results suggest that the effects of physical activity level prior to breast cancer diagnosis decrease the risk of all-cause and breast-cancer mortality in Hispanic women but not non-Hispanic women.

Citation Information: Cancer Epidemiol Biomarkers Prev 2011;20(10 Suppl):A81.