Recreational physical activity and overall survival for women diagnosed with breast cancer in the Prospective Family Study Cohort (ProF-SC)

Authors: Kehm RD, Liao Y, Zeinomar N, Phillips KA, Daly MB, John EM, Andrulis IL, Buys SS, Hopper JL, Terry MB

Category: Lifestyles Behavior, Energy Balance & Chemoprevention
Conference Year: 2019

Abstract Body:
Purpose: To examine the association of recreational physical activity (RPA) with overallsurvival (OS) after breast cancer (BC) diagnosis in a large prospective family cohort. Methods: We studied 4,709 women from the Prospective Family Study Cohort (ProF-SC) whowere enrolled within 2 years of first primary BC diagnosis. Women were followed for up to21 years (median=11.2 years); 1,176 deaths were ascertained by active follow-up,relative-report, or death records. At enrollment, women self-reported average hours perweek of RPA during the 3 years prior to diagnosis. We categorized women into quintiles oftotal metabolic equivalents (METs) per week and tested the association with OS usingmultivariable Cox proportional hazards regression. Models were stratified by age group atdiagnosis (10-year intervals), and a robust variance estimator accounted for family-levelclustering. We examined whether the association was confounded or modified by age atdiagnosis, education, lifestyle factors, body mass index (BMI), tumor characteristics, orabsolute predicted 1-year BC risk estimated from pedigree models. Results: Overall, women in the highest RPA quintile (≥35 METs/week) had a 15% reduced riskof death compared to women in the lowest RPA quintile (0-4 METs/week) after first primaryBC diagnosis (Hazard Ratio (HR)=0.85, 95% CI=0.70-1.03); no significant linear trend wasfound across quintiles of RPA (p=0.21). Consistent patterns were found across strata ofeducation, smoking status, age at diagnosis, estrogen receptor status, and tumor stage.Across strata of BMI, an association was only found for women ≤ 25 kg/m2 (interactionp=0.03). The RPA-OS association was positively associated with absolute predicted 1-yearBC risk (interaction p=0.03). Compared to the 15% reduced risk of death associated withhigh versus low RPA in the overall sample, high RPA was associated with a 2% (HR=0.98, 95%CI=0.79-1.22), 8% (HR=0.92, 95% CI=0.75-1.12) and 24% (HR=0.76, 95% CI=0.62-0.94) reducedrisk of death for women at the 10th, 50th, and 90th percentile of absolute predicted BCrisk, respectively.Conclusion: RPA may improve OS for women diagnosed with BC, particularly for women who areat higher risk for developing BC based on a family history of disease.

Keywords: recreationalphysical activity,breast cancer,overall survival