Background and objectivesN-acetyltransferase 2 (NAT2) metabolize several drugs including isoniazid. We investigated the effect of genotype, geographical difference, and smoking habit on NAT2 phenotype in Ethiopians.MethodsGenotyping for NAT2 191G > A, 341 T > C, 590G > A, and 857G > A was performed in 163 unrelated healthy Ethiopians (85 living in Ethiopia and 78 living in Sweden). The NAT2 phenotype was determined using caffeine as a probe and log AFMU/(AFMU + 1X + 1 U) urinary metabolic ratio (MR) as an index.ResultsThe frequencies of NAT2*4, *5, *6, *7, and *14 haplotypes were 14.1, 48.5, 30.1, 5.5, and 1.8%, respectively. The frequencies of rapid (NAT2*4/*4), intermediate (heterozygous *4), and slow (no *4 allele) acetylator genotypes were 1.8, 24.6, and 73.6%, respectively. The distribution NAT2 MR was bimodal with 70% being phenotypically slow acetylators. NAT2 genotype (p < 0.0001) and country of residence (p = 0.004) independently predicted NAT2 phenotype. Controlling for the effect of genotype, ethnic Ethiopians living in Ethiopia had significantly higher NAT2 MR than those living in Sweden (p = 0.006). NAT2 genotype-phenotype concordance rate was 75%. Distinct country-of-residence-based genotype-phenotype discordance was observed. The proportion of phenotypically determined rapid acetylators was significantly higher and slow acetylators was lower in Ethiopians living in Ethiopia (39% rapid, 61% slow) than in Sweden (20% rapid, 80% slow). Sex and smoking had no significant effect on NAT2 MR.ConclusionsWe report a high prevalence of NAT 2 slow acetylators in Ethiopians and a conditional NAT2 genotype-phenotype discordance implicating a partial phenotype conversion and metabolic adaptation. Gene-environment interactions regulate NAT2 phenotype.