PURPOSE To examine the association between CYP2D6 genotype, discontinuation of tamoxifen therapy, and prognosis for breast cancer. PATIENTS AND METHODS We conducted a prospective-retrospective study linking data from a clinical breast cancer register, the Swedish Prescribed Drug Register, and self-reported questionnaires. We genotyped CYP2D6 in 1,309 patients with breast cancer who were treated with tamoxifen and were diagnosed from 2005 to 2012; they were categorized as poor, intermediate, normal, or ultrarapid CYP2D6 metabolizers. We investigated whether metabolizer status was associated with tamoxifen discontinuation and prognosis for breast cancer using Cox regression analysis. RESULTS The 6-month discontinuation rates of tamoxifen among poor, intermediate, normal, and ultrarapid CYP2D6 metabolizers were 7.1%, 7.6%, 6.7%, and 18.8%, respectively. A U-shaped association was found between CYP2D6 metabolizer status and breast cancer–specific mortality, with adjusted hazard ratios of 2.59 (95% CI, 1.01 to 6.67) for poor, 1.48 (95% CI, 0.72 to 3.05) for intermediate, 1 (reference) for normal, and 4.52 (95% CI, 1.42 to 14.37) for ultrarapid CYP2D6 metabolizers. CONCLUSION Both poor and ultrarapid CYP2D6 metabolizers of tamoxifen have a worse prognosis for breast cancer compared with normal metabolizers after receiving a standard dose of tamoxifen. This U-shaped association might call for individualized tamoxifen dosage.