Table 1.

Multivariable Analyses for Genetic Determinants of MP Dose Intensity

Variable	Model One		Model Two		Model Three
	Regression Coefficient Estimate*	P†	Regression Coefficient Estimate*	P†	Regression Coefficient Estimate*	P†
African ancestry	−0.01	.71	−0.01	.66	−0.009	.78
East Asian ancestry	−0.14	.0003	−0.08	.019	−0.102	.005
Native American ancestry	−0.03	.43	0.0006	.99	0.013	.75
NUDT15 variant (rs116855232)			−0.22	8.8 × 10−9	−0.22	2.7 × 10−8
TPMT variants‡					−0.20	4.6 × 10−9

NOTE. Three multivariable models illustrate contributions of NUDT15 variants, TPMT variants, and genetic ancestry to interpatient variability in MP dose intensity.

Abbreviations: MP, mercaptopurine; SNP, single-nucleotide polymorphism.

^*Regression coefficient estimate represents increase (positive value) or decrease (negative value) in percent MP dose intensity for those patients with certain levels of genetic ancestry (model one), those carrying risk allele at NUDT15 SNP (model two), or those carrying risk alleles at TPMT SNPs. For instance, mean MP dose intensity for patients with 100% East Asian genetic ancestry is estimated 14 percentage points lower than that of patients with 100% European genetic ancestry, as in model one. For each additional T allele at rs116855232, mean MP dose intensity is estimated to decrease 22 percentage points while holding ancestry factors at any fixed level, as in model two. In model three, for each additional T allele at rs116855232, mean MP dose intensity is estimated to decrease 22 percentage points while holding ancestry factors and TPMT variation at any fixed levels.

^†P values were estimated using mixed linear-effects model including independent variables as described.

^‡Genetic variation in TPMT represents composite genotype at rs1142345, rs1800462, and rs1800460 (contributing *2, *3A, *3B, and *3C variants).