Title: Ultra-processed food intake and postmenopausal breast cancer risk in the NIH-AARP diet and health study

BACKGROUND: UPF intake is associated with obesity. Despite obesity being a risk factor for postmenopausal breast cancer, evidence for an association between UPF and breast cancer is limited. Our objective was to assess the association between UPF intake and postmenopausal breast cancer risk in the NIH-AARP Diet and Health Study. METHODS: Participants reported dietary intake via a food frequency questionnaire at baseline in 1995 to 1996 and were followed through 2018. Food items were disaggregated into food codes and assigned Nova classification via database linkage. Multivariable Cox proportional hazards regression models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI) for quintiles of UPF (g/1000 kcal/day) and postmenopausal breast cancer risk overall, by estrogen receptor (ER) status, and type (invasive or ductal carcinoma in situ (DCIS)), with and without adjustment for body mass index (BMI). RESULTS: Among 181,460 postmenopausal women, 14,484 were diagnosed with breast cancer over a median 21 years of follow-up. Median (IQR) UPF intake was 284.6 (191.1 to 466.6) g/1000 kcal/day. No associations between UPF intake and postmenopausal breast cancer risk overall (HRQ5vs.Q1=0.98, 95% CI = 0.93-1.03; Ptrend=0.32), by ER status (ER+ HRQ5vs.Q1=1.01, 95% CI = 0.94-1.09; P trend=0.95; ER-: HRQ5 vs Q1=1.02, 95% CI = 0.87-1.20; P trend=0.58), or by type (invasive: HRQ5vs.Q1=0.98, 95% CI = 0.92-1.04; P trend=0.14; DCIS HRQ5vs.Q1=1.00, 95% CI = 0.87-1.13; P trend=0.26) were observed. Following BMI adjustment, inverse trends for overall (P trend=0.03) and invasive (P trend=0.01) breast cancer were observed. CONCLUSION: UPF intake was not associated with postmenopausal breast cancer in the NIH-AARP cohort. Inverse trends observed after adjusting for BMI are likely spurious, owing to over-adjustment.