Prognostic Associations of 25OH Vitamin D in NCIC CTG MA.21, a Phase III Adjuvant RCT of Three Chemotherapy Regimens (EC/T, CEF, AC/T) in High-Risk Breast Cancer (BC).

504 Background: Low vitamin D (VitD) has been associated with poor outcomes in BC. We examined in MA.21 the association of VitD blood levels with relapse-free-survival (RFS) in all patients, and in BC subtypes. Methods: Fasting blood was collected pre-chemotherapy in 935/2,104 ((44.4%) of subjects (blood collection was initiated part-way through the trial); serum was frozen at -80C and shipped on dry ice to Mount Sinai Hospital (Toronto); it was analyzed for Vit D (radioimmunoassay, Diasorin) 25(OH)D (nmol/L). VitD underwent a λ=0.5 Box-Cox transformation for analysis to reduce asymmetry; data were back-transformed for reporting. VitD was assessed as a transformed continuous factor, and by quartiles and IOM categories (<40, [40-50), [50-125), >125 nmol/L). Univariate and multivariate (mv) assessments were with Cox models, adjusted for treatment, stratification factors, and imbalances in who had VitD assessed. Results: Median age was 47.8 years; most patients were white (91.6%), premenopausal (69.4%) with good performance status (PS) 0-1 (99.9%). The majority presented with grade III (52%), HER2 neg (66.6%), HER2 missing (22.9%), ER pos (61.9%), T1-2 (89.4%), N+ (72.7%) BC. 52.1% underwent mastectomy and 74.3% received adjuvant radiotherapy. Compared to the full MA.21 population, those with VitD levels were marginally (but significantly) more likely to be white, PS 1 or 2, to have undergone mastectomy and to have ER+, HER2 neg tumors. Mean Vit D was 69.7 nmol/L [95% CI (68.1-71.3)] nmol/L. The majority (752/935 = 80.5%) had levels > 50nmol/L (20ng/ml), considered adequate by the Institute of Medicine (IOM). In adjusted mv analyses, (continuous) VitD was not associated with RFS (HR 0.98, 95% CI (0.93-1.03); p=0.36) or categorical variable (between quartiles, p=0.20-0.43; between IOM categories, p=0.33-0.78). There were no significant differences in mean VitD levels or association of VitD with RFS in tumor subtypes (ER+, HER2- / any ER, HER2+ / ER-, HER2-; p=0.14-0.50). Conclusions: There is no evidence that VitD was associated with RFS in MA.21; the majority of subjects had adequate VitD levels at study entry. Clinical trial information: NCT00014222.