25-hydroxy vitamin D (VitD) and associated variables as predictors of breast cancer (BC) risk and tamoxifen benefit in NSABP-P1.

Abstract

1528 Background: Observational studies suggest that host lifestyle factors are associated with BC risk. Evidence is strong for obesity after menopause, but less convincing for VitD. Obesity and low VitD are associated with insulin resistance, inflammation and elevated adipocytokines such as leptin. We have examined associations of obesity, VitD and these related factors with BC risk. Methods: A nested case control study of the NSABP-P1 trial was conducted. Cases were participants who developed invasive BC. Control participants (≤4 per case) did not develop invasive or non-invasive BC during follow-up and were matched for age, race, 5-year Gail score and geographic location of clinical center as a surrogate for latitude. Fasting serum banked at trial enrolment was assayed for 25-hydroxy VitD, insulin, leptin and C-reactive protein (CRP), a marker of inflammation. Associations between study variables and the risk of invasive BC were evaluated using conditional logistic regression with adjustment for baseline factors including tamoxifen treatment and body mass index (BMI). Results: 231 cases were matched with 856 controls. Mean age was 54 and 45% were postmenopausal. There were significant (p<0.05) negative correlations for VitD (mean 57.9 nmol/L + 25.3)with insulin, CRP and leptin (Spearman’s rho =−0.23,−0.15,−0.22). There were positive correlations for BMI (mean 27.2 kg/m2 + 5.7) and these variables (rho = 0.58, 0.54, 0.79). Higher BMI was associated with higher BC risk [odds ratio (OR) 1.45 (≥25 vs. <25 kg/m2), 95% confidence interval (CI) 1.06-2.00, p=0.02]. Tamoxifen treatment (vs. placebo) was protective (OR 0.44, 95% CI 0.32-0.61 p<0.001). VitD deficiency (<72 nmol/L or 30 ng/mL) did not influence BC risk (OR 1.05 95% CI 0.73-1.53, p=0.76). When evaluated as continuous variables, insulin, CRP and leptin levels were not associated with BC risk (OR 0.88 [p=0.34], OR 0.97 [p=0.68] and OR 1.05 [p=0.70]). There were no significant interactions between treatment and BMI or serum variables. Conclusions: Higher BMI is associated with a greater BC risk. Serum levels of VitD, insulin, CRP and leptin are not independent predictors of BC risk or tamoxifen benefit in this high risk population.