|| Checking for direct PDF access through Ovid
NAT2 slow acetylator is a confirmed risk of anti-tuberculosis drug-induced liver injury (ATDILI). However, NAT2 ultra-slow acetylators, a new refinement among NAT2 slow acetylators, have been recently proposed. The patients with NAT2 genotypes of *6A/*6A, *6A/*7B and *7B/*7B are referred to in this group.We aim to prove an association of the NAT2 ultra-slow acetylators with the risk of ATDILI.Systematic review and meta-analysis were performed based on each NAT2 genotype and risk of ATDILI cases and also new classification of the ultra-slow acetylators up to 31 October 2016. Meta-analysis of 18 studies with 822 ATDILI cases and 4630 controls was carried out in the RevMan software, version 5.3 with fixed-effect (low heterogeneity) and random effect (moderate to high heterogeneity) methods.The strong associations between each NAT2 slow acetylator genotypes and ATDILI were confirmed in meta-analysis except for NAT2*5B/*5B [odds ratio (OR): 1.69; 95% confidence interval (CI): 0.96–2.95; P=0.0679]. The NAT2 ultra-slow acetylators contribute to higher risk of ATDILI (OR: 3.60; 95% CI: 2.30–5.63; P=1.76E−08) than all NAT2 slow acetylators (OR: 2.80; 95% CI: 2.20–3.57; P=5.73E−18) as well as fast acetylators. Additional in-vitro study using isoniazid as a substrate supports the existence of ultra-slow acetylator alleles (NAT2*6A and NAT2*7B).This is the first meta-analysis of NAT2 and the risk of ATDILI at the genotypic level. The result demonstrated that NAT2 ultra-slow acetylator genotypes will have the most effect on the increased risk of ATDILI.