Table 2.

Rates of distant recurrence and any recurrence according to CYP2D6 phenotype and concomitant medication usage^*

CYP2D6 phenotype	Distant recurrence		P†	Any recurrence		P†
	Tamoxifen group, No. (%)	HR (95% CI)		Tamoxifen group, No. (%)	HR (95% CI)
PM (phenotypic score = 0)	19 (18.5)	1.00 (referent)		24 (23.3)	1.00 (referent)
IM (phenotypic score = 1.0)	22 (14.3)	0.74 (0.40 to 1.37)	.34	27 (17.5)	0.67 (0.39 to 1.17)	.20
IM (phenotypic score = 1.5)	4 (8.3)	0.49 (0.16 to 1.44)	.19	6 (12.5)	0.58 (0.23 to 1.41)	.21
EM (phenotypic score = 2.0)	44 (15.6)	0.78 (0.45 to 1.35)	.38	58 (20.5)	0.82 (0.51 to 1.32)	.44

^*The relative catalytic activity of CYP2D6 for substrate O-demethylated dextromethorphan was used to calculate an “activity score”; all ATAC UK genetic substudy patients were assigned a phenotypic score based on their CYP2D6 genotype. The CYP2D6 activity score is an approach to predict most accurately a patient's CYP2D6 metabolic phenotype based on their genotype (31). The scores were adjusted for concomitant CYP2D6 inhibitors during active treatment. Cox proportional hazard model was used for multivariable analysis adjusted for tumor size (<20 mm, 20–50 mm, >50 mm, unknown), grade (well differentiated, moderately differentiated, poorly differentiated, unknown), nodal status (positive, negative, unknown), and age. CI = confidence interval; EM = extensive metabolizer; HR = hazard ratio; IM = intermediate metabolizer; PM = poor metabolizer.

^†Two-sided P values were calculated using an adjusted Cox proportional hazard model.