Background: The clinical impact of CYP2D6 genetic polymorphisms on tamoxifen (Tam) remains unclear. Active metabolite levels of Tam can be influenced by drug interaction and compliance. This study evaluated the association of steady state plasma levels of Tam and its metabolites with clinical outcomes in samples collected in MA.12. Premenopausal women (n=672) with high risk node negative and node positive breast cancer following adjuvant chemotherapy were randomized to Tam or placebo. There was an improvement in DFS (HR 0.77; p=0.056) but not OS (HR 0.78; p=0.12) favoring Tam. Methods: TAM, N-desmethyl-tamoxifen (NDM), 4-hydroxy-tamoxifen (4-OH), and endoxifen (Endo) were quantified in serum by liquid chromatography–tandem mass spectrometry. Samples collected after 120 days on Tam were analyzed, (n=121 Tam, n=121 Placebo). Exploratory analyses assessed the association of DFS and OS with each metabolite level, as a continuous variable, as well as a categorical variable using an optimal cut point. Cut points were determined using the minimum p-value. Factors associated with DFS or OS were evaluated in a multivariate Cox model adjusting for age, performance status, chemotherapy, receptor status, nodal status, histology and tumor stage. Results: Baseline characteristics of the substudy were similar to the full trial population. Descriptive metabolite levels in the Tam cohort are listed in the Table. There was no crossover at 120 days. As a continuous measure, none of the metabolite values were significantly correlated with DFS or OS. As a categorical variable, only 4-OH was associated with clinical outcomes. In a multivariate Cox model, the optimal 4-OH cutpoint (1.72 ng/mL) was significantly correlated with 5 year DFS HR 0.30 [95% CI(0.10-0.95), p=0.04)] and 5 year OS HR 0.21 [95% CI(0.05-0.93), p=0.04]. Conclusions: In this exploratory, hypothesis generating analysis of MA.12, wide ranges in serum concentrations of Tam metabolites were observed. Only 4-OH levels were associated with clinical outcomes.