Background: Tamoxifen is extensively used as endocrine therapy for breast cancer. It is a prodrug that is primarily metabolized by CYP2D6 and CYP3A4, particular into endoxifen. In daily practice, the herb curcumin is widely used among patients (pts) because of its presumed anti-tumor effects. Preclinical studies show effects of curcumin on phase I and II drug metabolism, leading to altered plasma levels. We hypothesized that curcumin increases endoxifen exposure by affecting phase II metabolism. Therefore, we performed a randomized, 3-phase, cross-over study to compare tamoxifen exposure in breast cancer pts +/- curcumin, and with the addition of the bioenhancer piperine. Methods: Pharmacokinetic sampling (PK) was performed in 15 pts at the 28^th, 56^th and 84^th day of the trial. In the 28 days prior to PK, tamoxifen (20 mg qd) was either given alone, or combined with curcumin (1,200 mg TID), or with curcumin + piperine (10mg TID) in this order or vice versa. Genotyping was performed to determine CYP2D6 and CYP3A4 phenotypes. Primary endpoint was difference in geometric means for AUC_0-24h. A linear mixed model was used to analyze log-transformed area under the curve (AUC). For multiple testing a Bonferroni correction was applied. Results: Tamoxifen AUC_0-24h decreased with 7.3 % (95% CI: -0.5,-13.6%; p = .04) with curcumin and 13.0% (95% CI: -5.6,-19.9%; p = .002) with curcumin and piperine, compared to tamoxifen alone. Endoxifen AUC_0-24h decreased with 8.9% (95% CI: -0.9,-16.3%; p = .029) and 13.5% (95% CI: -2.8,-23.0%; p = .015), respectively. CYP2D6-screening resulted in 6 intermediate metabolizers (IM), and 6 extensive metabolizers (EM). Interestingly, for pts with a EM phenotype, effects of curcumin with piperine on drug exposure (AUC_0-24h) seem to be higher than for IM phenotype pts (a decrease of 21.8% vs 4.8% for tamoxifen, and 22.6% vs 10.2% for endoxifen, respectively). No severe toxicity resulting from co-treatment was observed. Conclusions: In contrast to our hypothesis, the exposure of tamoxifen and endoxifen significantly decreased by concomitant use of curcumin +/- piperine. Although limited effects in most pts, co-treatment with these herbs could drop endoxifen levels below the threshold for efficacy, especially for EM phenotype pts. Clinical trial information: NTR6149.