Cancer
Temporal changes in modifiable risk factors of breast cancer in the Generations Study Isobel Jackson* Isobel Jackson Alicia Heath Michael Jones Montse Garcia-Closas Amy Berrington de González Marc Gunter
Background: Several established breast cancer risk factors are modifiable and exposure may change over the life course. Understanding temporal changes in these risk factors is crucial to inform breast cancer prevention strategies. However, most cohorts rely on single exposure measurements and longitudinal data assessing long-term exposure patterns are scarce.
Methods: Using the Generations Study, a prospective cohort of >113,700 UK women aged ≥16 years, we investigated changes in self-reported alcohol intake, smoking, body mass index (BMI) and physical activity (PA). We analyzed changes in risk factors between baseline and follow-ups at 6.2 and 14.3 years after study entry. Within-person changes were calculated comparing follow-up to baseline exposure measurements overall and stratified by birth cohort. Predictors of exposure change were analyzed using logistic and linear regression for binary and continuous changes, respectively, adjusting for baseline age and exposure.
Results: Changes in risk factors were modest overall. At each round, most women reported current drinking, non-smoking, high PA and healthy BMIs, with minimal changes in observed values across rounds. Over a median follow-up of 14.3 years, participants’ BMI and total PA increased by a median of 0.5 kg/m2 (interquartile range [IQR]:-0.9, 2.2) and 1.4 MET-h/week (IQR:-14.3, 18.5), respectively, while alcohol intake and smoking intensity decreased by a median of 2.4 g/day (IQR:-9.6, 3.3) and 9 cigarettes/day (IQR:-15.0, -3.1), respectively, among current/former drinkers or smokers at follow-up. Exposure patterns were heterogeneous across birth cohorts. Factors associated with exposure changes included socioeconomic status and education.
Conclusion: Although exposure change varied by birth cohort, modifiable risk factors of breast cancer appear relatively stable over time. For these risk factors, a single baseline measurement may be sufficient to capture medium to long term exposures for risk assessment.