Introduction

Drug absorption, distribution, metabolism and elimination (ADME)-related genes are genes related to the absorption, distribution, metabolism and elimination of drugs. They can be classified into four groups depending on their roles in the ADME process. The first group is the phase I enzymes or the cytochrome P450 enzyme superfamily responsible for catalyzing the oxidation, reduction, hydrolysis, cyclization and decyclization reactions. The second group is the phase II metabolizing enzymes responsible for conjugation reactions by attaching an ionized group to the drug resulting in water-soluble metabolites. The third group is membrane transporters responsible for pumping the drugs across cellular barriers, thus having impact on a drug’s therapeutic efficacy by influencing its absorption, distribution and elimination.1 The last group is other genes that could not be classified in the previous three groups, but they are related in the ADME process, such as VKORC1 gene. However, the efficacy of any drug treatment is influenced by factors such as age, weight, gender, dosage, parallel drug use, as well as genetic variation. Genetic variations in drug ADME-related genes have been proven to be polymorphic and consequently influence the success of individual drug response owing to the fact that the activities of drug ADME-related gene variations may differ depending upon an individual genotype. In fact, genetic differences are heavily related to ethnicity,2 and therefore, large population-based studies to investigate allele frequencies in drug ADME-related genes from non-reference populations are definitely important to understand the genetic diversity of drug ADME-related genes across ethnic groups.

Currently pharmacogenomics helps facilitate the identification of biomarkers that informed optimization of drug selection, dosage of the desired drug, and prevention of adverse drug reactions.3 Lacking large-scale of pharmacogenomics information in less developed countries is definitely a bottleneck for patient care improvement. Therefore, the objectives of this study are to investigate allele frequencies of drug ADME-related genes in the Thai population and to compare drug ADME-related genes with HapMap populations including Caucasians, Africans and Asians. The knowledge obtained from this study could be used to explain the roles or the influence of genetic variations in drug responsiveness. This information will subsequently contribute to future strategies for patient-tailored drug therapy. In addition, these invaluable data might be unexpectedly useful if population drug metabolizing enzymes and transporters (DMET) information could be incorporated to improve or revise the Thai National List of Essential Medicines (NLEMs) i.e., the reimbursement drug list for patients enrolled under three schemes which are Civil Servant Medical Benefit Scheme (CSMBS) supporting the patients who are government employees and their dependents, the Social Security Scheme (SSS) supporting the patients who are employees and the Universal Coverage scheme supporting the patients who are not under CSMBS or SSS group. It is expected that population DMET information or genetic information could be used by policy makers to make decision in the drug selection process of the NLEMs, and this could yield more advantage than using traditional way of policy decision-making based on only drug efficacy and economic evaluation information.

Materials and Methods

Population samples

The 190 samples from the healthy Thai population were selected from the 3rd National Health Examination Survey (NHES) which was carried out in 2004.4 We randomly selected the samples based on the four major geographic regions and the availability of the buffy coat samples in the samples bank in the Department of Medical Sciences, Ministry of Public Health, Thailand. In summary, 47 samples from the central, 47 samples from the North, 51 samples from the Northeast and 45 samples from the South were selected to perform DMETs genotyping. For genotyping from HapMap population, all genomic dideoxy nucleic acid (DNA) samples were purchased from Coriell Institute (Camden, NJ, USA). The samples that were successfully genotyped included 58 samples of Caucasians (CEU—US Utah residents with ancestry from northern and western Europe), 119 samples of Africans (YRI—Yoruba people in Ibadan, Nigeria), 90 samples of Chinese (CHB—Han Chinese in Beijing, China) and 91 samples of Japanese (JPT—Japanese in Tokyo, Japan) (the list of sample ID that were genotyped in this study were provided in Supplementary data).

DNA isolation and genotyping

All genomic DNA samples were isolated from blood samples using the Qiagen DNA isolation kit (QIAGEN GmbH, Hilden, Germany). The concentration of DNA samples were measured by PicoGreen dsDNA assay kit from Invitrogen (Molecular Probes, Eugene, OR, USA) as recommended. The DNA samples with the concentration greater than 60 ng ul−1 were genotyped by DMET plus assay subsequently.5 Genomic DNA samples that passed the quality control were combined for the annealing and amplification steps, in which molecular inversion probes were exploited to genotype all the genomic sites of interest in a single reaction. The DNA samples were subsequently purified, fragmented, labeled and hybridized to the array to be scanned with the Gene Chip Scanner 3000 (Affymetrix Inc, Santa Clara, CA, USA). The genotype calls were performed by standard parameters using DMET Console version 1.1 (Affymetrix Inc). Before proceeding to the data analysis, some data quality controls were performed to check the data. Firstly, the concordance between the genotyped and reported gender were tested to check for errors in sample labeling. Secondly, all subjects illustrating a genotype call rate <98% would have been removed. The samples passing these criteria proceeded to the next step of data analysis.

Data analysis

Hardy–Weinberg disequilibrium was also carried out as part of the routine quality control with P-value cutoff at 0.05. The genotypic distribution of the Thai populations were calculated based on all samples available after the quality control processes as previously described (91 male, 99 females). The genotypic distributions were compared among the five populations, which are Caucasians (CEU), Africans (YRI), Chinese (CHB), Japanese (JPT) and Thai, by the Fisher’s Exact tests. Pairwise comparisons between pairs of five populations were also carried out by the Fisher’s Exact test. The statistical analyses were completed by standard statistical package in R version 10. As the numbers of samples from each population are not equal, it is not possible to compare statistical significance across analyses, but this information provided reliable ranking for each analyses.

Ethics statement

The protocol was ethically approved by the Faculty of Medicine Ramathibodi Hospital, Mahidol University and Institute for development of human research protection endorsed by Ministry of Public Health in Thailand. Within NHES sampling, samples from Bangkok, the capital city of Thailand, were excluded owing to the heterogeneity of people living in this city.

Results

Quality control of Thai samples

In this study, the 222 samples of the healthy Thai population obtained from the 3rd National Health Examination Survey (NHES) were sent to IPIT, the University of North Carolina, USA, for the DMET plus genotyping to be performed. All samples were randomly selected based on the four major geographic regions which are the central, the northern, the northeastern and the southern parts of Thailand. By quantifying DNA concentration using the PicoGreen ds DNA assay kit, there were only 192 samples meeting the quality and quantity requirements as recommended by the manufacturer's instruction. After performing the DMET genotyping, all samples were monitored for their qualities by investigating the concordance between the genetic and reported gender. In this step, one sample failed owing to the lack of concordance between the genetic and reported gender. Therefore, 191 samples were passed to the other step of quality control by monitoring the call rates per sample. A sample with call rate per sample below 98% was excluded in this step. In conclusion, 190 samples passed quality controls, and these samples composed of 47 samples from central area (male=24; female=23), 47 samples from the northern area (male=22; female=25), 51 samples from the northeastern area (male=23; female=28) and 45 samples from the southern area (male=22; female=23). These samples could represent the Thai population because all 190 samples were obtained from four regions that covered all the regions of Thailand. In addition, the genders of individual region were closely equal. These data indicated that all samples were capable for data and statistical analyses. All samples were examined 1936 single nucleotide polymorphisms (SNPs) according to drug ADME markers in 225 genes and compared with those of HapMap populations.

Principal component analysis (PCA) was used to illustrate the similarities and discrepancies of DMET SNPs among Thai, CEU, YRI, CHB and JPT. The data demonstrated three clearly separated clusters (Figure 1), which are the clusters of Caucasians, Africans and Asians. As expected, all Thai samples were clustered in the Asian group. Interestingly, Thai samples were strongly correlated with other two Asian HapMap populations, CHB (Chinese) and JPT (Japanese).

Figure 1
figure 1

Principal component analyses (PCA) plot illustrating the grouping of ethnic populations using known genetic variation in the metabolizing and transporter genes. PCA plot shows the segregation of Caucasian (CEU, •), Africans (YRI, ), and Asians (JPT, ♦; CHB, ; THAI, ).

Comparison of allele frequencies in drug ADME-related SNPs among Thai populations from four regions of Thailand

Comparing all Thai populations from four regions of Thailand, the results indicated significant differences in 93 SNPs with the P-value lower than 0.05 (Table 1). These 93 SNPs composed of 6 groups of phase I drug biotransformation enzymes (CYP2C9, CYP2D6, CYP4F11, CYP4F12, CYP11B1 and CYP11B2), 20 groups of phase II drug biotransformation enzymes (ALDH1A1, ALDH3A1, CHST5, CHST10, DPYD, FMO2, GSTA2, GSTA5, GSTM1, GSTM4, NNMT, SULT1A2_A3, SULT1B1, SULT1C2, SULT1E1, UGT1A1, UGT1A3, UGT2A1, UGT2B15 and UGT2B17), 20 groups of transporter (ABCB1, ABCB4, ABCB7, ABCB11, ABCC1, ABCC4, ABCC5, ABCC8, ATP7B, SLC7A7, SLC7A8, SLC13A1, SLC15A1, SLC16A1, SLC22A3, SLC22A4, SLC22A5, SLC22A8, SLC28A1 and SLC28A2) and 5 groups of other genes involved in the drug ADME process (AOX1, CBR1, POR, PPARD and VKORC1). Interestingly, all top three SNPs showing the lowest P-values were in the group of phase II drug-metabolizing enzymes and other genes, which were ALDH3A1 (rs2072330), GSTM1(rs1065411) and GSTM1 (rs737497) with the P-values 1.84 × 10−6, 7.35 × 10−6 and 2.39 × 10−5, respectively. The data of the comparison of allele frequencies in drug ADME-related SNPs among Thai populations from four regions of Thailand have never been reported before.

Table 1 Comparison of SNPs of drug ADME-related genes among Thai populations in four regions of Thailand

Comparison of allele frequencies in drug ADME-related SNPs among five populations

Comparing all five populations, the results indicated significant differences in 43 SNPs with P-value lower than 0.05 (Table 2). These 43 SNPs composed of seven groups of phase I drug biotransformation enzymes (CYP2A6, CYP3A5, CYP2B6, CYP2C8, CYP2C9, CYP2C19 and CYP2D6), four groups of phase II drug biotransformation enzymes (COMT, NAT2, TPMT and UGT1A1), one group of transporter (SLCO1B1) and two groups of other genes involving in drug ADME process (G6PD and VKORC1). Interestingly, all top five SNPs showing the lowest P-values were in the group of phase I drug-metabolizing enzymes and other genes, which were VKORC1 (rs9923231), G6PD (rs1050829), CYP3A5 (rs776746), VKORC1 (rs7294) and CYP2D6 (rs28371706) with the P-values 8.94 × 10−138, 3.23 × 10−64, 5.04 × 10−63, 6.81 × 10−35 and 9.72 × 10−35, respectively. For phase II drug-metabolizing enzymes, the lowest P-value was 1.41 × 10−28 of NAT2 (rs1801280). There was only one drug transporter demonstrating the difference among five populations which was SLCO1B1 (rs2306283) with P-value of 1.60 × 10−16. Interestingly, CYP2D6 illustrated that the highest number of SNPs differed from the other four HapMap populations because there were seven SNPs of CYP2D6 showing statistically significant P-values. These SNPs of CYP2D6 were rs28371706, rs16947, rs3892097, rs59421388, rs61736512, rs35742686 and rs5030655 with P-values 9.72 × 10−35, 4.92 × 10−32, 2.61 × 10−18, 2.41 × 10−15, 2.27 × 10−11, 1.16 × 10−3 and 7.52 × 10−3 respectively.

Table 2 Comparison of SNPs of drug ADME-related genes among five populations

Interracial pairwise comparison of allele frequencies in drug ADME-related SNPs

The objectives of this study were to examine the SNPs of drug ADME-related genes of Thai samples and also to compare and contrast the SNPs among five populations with major focus on the differences between Thai SNPs and Caucasian SNPs as well as comparing and contrasting on Thai SNPs and other Asians (CHB and JPT). Herein, we emphasized data illustrating the statistical significance between Thai versus CEU and Thai versus CHB/JPT. The reason why we focused on the data comparison between Thai versus CEU is because of the fact that the Asian’s national drug policies including Thai’s national drug policies had followed the guidelines created from the data of Caucasians from developed countries. The data comparison between Thai and other Asians (CHB/JPT) are also important because these observations will help us to better understanding in genetic variation and epidemiology of drug ADME-related genes among Asian populations. However, we also discussed some data illustrating the statistical significance between Thai and YRI to better understand the genetic variations between ethnicity.

Comparing SNPs of drug ADME-related genes between Thai and Caucasian HapMap population (CEU), the data illustrated significant differences in 23 SNPs of 13 groups of drug ADME-related genes (Table 3). For phase I enzymes, there were seven groups showing statistically significant differences which were CYP2A6 (rs1801272; P=0.000654 and rs28399433; P=0.004444), CYP3A5 (rs776746; P=3.96 × 10−12), CYP2B6 (rs2279343; P=0.046178 and rs3211371; P=0.000127), CYP2C8 (rs10509681; P=9.87 × 10−7 and rs11572080; P=9.87 × 10−7), CYP2C9 (rs1799853; P=1.81 × 10−8), CYP2C19 (rs12248560; P=1.75 × 10−7 and rs4244285; P=0.005635) and CYP2D6 (rs16947; P=5.46 × 10−5, rs35742686; P=0.012539, and rs3892097; P=2.15 × 10−15). For phase II enzymes, there were four groups COMT (rs4680; P=4.46 × 10−7), NAT2 (rs1801280; P=8.76 × 10−15 and rs1799931; P=3.49 × 10−7), TPMT (rs1800460; P=0.012539) and UGT1A1 (rs8175347; P=0.000905 and rs4148323; P=0.000804) demonstrating statistical difference. There was only one SNP of drug transporter showing statistical difference which was SLCO1B1 (rs2306283) demonstrating a P-value of 8.81 × 10−12. For other genes involved in the drug ADME process, there was only one group demonstrating statistically significant differences which was VKORC1 (rs7294; P=0.003625, rs2884737; P=9.86 × 10−17, and rs9923231; P=7.63 × 10−13).

Table 3 Comparison of allele frequency of drug-metabolizing enzymes and drug transporter genes in Thais with Caucasian (CEU), African (YRI) and Asian (CHB and JPT) HapMap populations

Comparing SNPs of drug ADME-related genes between Thai and Asian HapMap population (CHB or Han Chinese in Beijing), the data illustrated significant differences in eight SNPs of five groups of drug ADME-related genes (Table 3). For phase I enzymes, there were two groups showing statistically significant differences which were CYP2A6 (rs28399433; P=0.005512) and CYP2B6 (rs3745274; P=0.001976 and rs2279343; P=0.002954). For phase II enzymes, there were two groups NAT2 (rs1801280; P=0.001979 and rs1799930; P=0.000617) and UGT1A1 (rs4148323; P=3.74 × 10−6) demonstrating statistical difference. For other genes involved in the drug ADME process, only VKORC1 illustrated statistically significant differences (rs7294; P=1.83 × 10−8 and rs9923231; P=1.17 × 10−9). There was no SNP for drug transporter showing statistical difference between Thai samples and Asian HapMap population (CHB).

Comparing SNPs of drug ADME-related genes between Thai and Asian HapMap population (JPT or Japanese in Tokyo), the data illustrated significant differences in 10 SNPs of six groups of drug ADME-related genes (Table 3). For phase I enzymes, there were four groups showing statistically significant differences which were CYP2A6 (rs28399433; P=0.000898), CYP3A5 (rs776746; P=0.009769), CYP2B6 (rs3745274; P=0.010841 and rs2279343; P=0.005689) and CYP2C19 (rs12248560; P=0.02956 and rs4986893; P=0.025513). For phase II enzymes, there was only one group NAT2 (rs1801280; P=0.000515 and rs1799931; P=0.001812). For other genes involved in the drug ADME process, VKORC1 demonstrated statistically significant differences (rs7294; P=0.000189 and rs9923231; P=1.65 × 10−5). There was no SNP for drug transporter showing statistical difference between Thai samples and Asian HapMap population (JPT).

As described previously, some but not all SNPs comparing in drug ADME-related genes between Thai and African HapMap population (YRI) were reported here particularly data illustrating the statistical significance between Thai versus CEU and Thai versus CHB/JPT. In these SNPs, the data illustrated significant differences in 13 SNPs of seven groups of drug ADME-related genes (Table 3). For phase I enzymes, there were four groups showing statistically significant differences which were CYP3A5 (rs776746; P=2.84 × 10−33), CYP2B6 (rs3745274; P=0.00240), CYP2C19 (rs12248560; P=8.61 × 10−13 and rs4986893; P=0.008578) and CYP2D6 (rs16947; P=5.93 × 10−29 and rs3892097; P=0.003394). For phase II enzymes, there were only two groups NAT2 (rs1801280; P=1.52 × 10−5, rs1799930; P=0.010754 and rs1799931; P=4.12 × 10−10) and UGT1A1 (rs8175347; P=1.20 × 10−22 and rs4148323; P=1.85 × 10−6) demonstrating statistical difference. For other genes involved in the drug ADME process, VKORC1 also demonstrated significant differences (rs7294; P=1.42 × 10−13 and rs9923231; P=2.21 × 10−83). There was no SNP for drug transporter showing statistical difference between the Thai samples and African HapMap population.

Discussion

Genetic variations of drug ADME-related genes have long been known to be significant factors affecting pharmacokinetics and pharmacodynamics of drugs and xenobiotics. Therefore, the investigation of allele frequencies of drug ADME-related genes including genetic variations or SNPs in phase I drug-metabolizing genes, phase II drug-metabolizing genes as well as drug transporter genes in Thais is absolutely important because these information will provide us a better understanding in predicting in metabolic behaviors of drugs or xenobiotics in Thai people. Moreover, as genetic differences are heavily related to ethnicity,2 comparisons of drug ADME-related genes with other ethnic groups are also immensely worthy because we can exploit these results to translate some unexplainable clinical evidences particularly found in Thais as well as to avoid some medicinal usages demonstrating high risk in producing adverse drug reaction. These processes will definitely lead us to develop the pharmaceutical practice in our country.

In this study, we performed an advanced method to examine the genetic variation of drug ADME-related genes in 190 healthy Thai participants using the DMET plus genotyping assay system. This method was demonstrably robust, cost-effective and practical as well as easy to use on a large scale. Using this system, we could investigate 1936 SNPs related to 225 genes that have been documented for functional significance in phase I and phase II drug-metabolism enzymes as well as drug transporters and other genes involved in ADME process. All Thai SNPs obtained from this study were compared with other HapMap populations (CEU, YRI, CHB and JPT). As described previously, we mainly focused on the drug ADME-related genes demonstrating statistical significance when pairwise comparing with Caucasians and other Asians in this study. Comparing among five populations, there were 43 SNPs illustrating statistical significance. However, data illustrating the statistical significance between Thai versus CEU and Thai versus CHB/JPT were only 26 SNPs comprising of six groups in phase I drug biotransformation enzymes (CYP2A6, CYP3A5, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6 and VKORC1), four groups in phase II drug biotransformation enzymes (COMT, NAT2, TPMT and UGT1A1) and one group in transporter (SLCO1B1) (Table 3). Interestingly, all these all 13 groups of drug ADME-related genes typically demonstrated clinical significances as previously observed in many studies. Herein, we concluded the data of these SNPs together with their functions and their substrates listed in Table 4. Importantly, we have found the potential results which could explain many previous clinical evidences based on genetic variations in drug ADME-related genes, which is also emphasized in this article.

Table 4 List of the important drug-metabolizing enzymes and drug transporter genes including their phenotype and example of drug substrates

CYP2C19 variations

CYP2C19 is an isoenzyme that metabolizes and eliminates a wide variety of drugs, including anticonvulsants, antidepressants, antimalarial drugs, antiulcer drugs, such as omeprazole, and antithrombotic drugs, such as clopidogrel. Owing to the fact that CYP2C19 variation is dependent on ethnicity, not surprisingly, the major allele frequencies of Thai population that have been found in this study were CYP2C19*2 (rs4244285), CYP2C19*3 (rs4986893) and CYP2C19*17 (rs12248560) as demonstrated in a previous study.6 The CYP2C19*2 and CYP2C19*3 alleles are considered to be poor metabolizers, whereas CYP2C19*17 is considered to be an ultra-rapid metabolizer. One of the famous clinical impacts of CYP2C19 is the efficacy of treatment by clopidogrel in cardiovascular disease.7 More importantly, the most prevalent CYP2C19 in Thai and other Asians is CYP2C19*2, the poor metabolizer, which could be a major impact in treatment by clopidogrel suggesting to monitor and consider to adjust dosing in Thai and other Asian populations.

CYP2D6 variations

Many drugs such as antipsychotics, antidepressants and analgesics are specific substrates and/or inhibitors of the CYP2D6.8 As investigated in previous studies, allele frequencies of the poor metabolizers, CYP2D6*3 (rs35742686) and CYP2D6*4 (rs3892097), in Thai and other Asians are lower than those of Caucasian.9, 10, 11 The clinical impact of CYP2D6 in Thai is also important even though it was found the low allele frequencies of these two forms of poor metabolizer. The crucial reason of clinical significance is not only focusing on clinical pharmacology, but also emphasizing on drug–drug interaction owing to its capabilities to metabolize many drugs and to be inhibited by many drugs. This occurrence could lead to adverse of drug reaction as discovered in previous researches. For example, the unfavorable drug–drug interactions of selective serotonin reuptake inhibitor antidepressants with CYP2D6 drugs (metoprolol, donepezil, galantamine) demonstrated the reduction of drug elimination by CYP2D6 leading to symptomatic adverse drug reactions.12 To avoid adverse drug reactions, the CYP2D6 biomarker should be suggested to indicate on the drug label for all populations including Thai.

CYP2C9 and VKORC1 variations

CYP2C9 is responsible for the hydroxylation of tolbutamide, an oral sulfonylurea hypoglycemic agent used in the treatment of type II diabetes mellitus.13 Several reports indicate that the most common allelic variants are CYP2C9*2 (Arg144Cys/ rs1799853) and CYP2C9*3 (Ile358Leu), which encode enzymes with decreased substrate turnover. Our results indicated the concordance of allele frequency for CYP2C9*2 in Thais investigated in this study with other previous findings in Thai and other Asians with lower occurrence in this variants than that of Caucasian.14, 15 An outstanding clinical significance of CYP2C9 was reported for warfarin metabolism and its clinical relevance. Although, CYP2C9*2, a low metabolizer, was found at a higher frequency in Caucasians than in Asians including Thai, the administrative dose of warfarin for Thai and other Asians were lower than the dose for Caucasians as warfarin metabolism not only associates by CYP2C9 but also by VKORC1. VKORC1 has a major role in the vitamin K pathway and is the target protein of warfarin. Interestingly, our data of VKORC1 allele frequency in Caucasian, Asians including Thai agree with the previous reports.16, 17 The major prevalence in Thai and other Asians is VKORC1_c.-1639G>A (rs9923231) with 75% illustrating a low function enzyme. Unlike Asians, the occurrence of this allele in Caucasians is only 38%. This is the key reason for Thai and other Asians to reduce warfarin dosage even patients having CYP2C9*2. Our results accentuate the necessity of CYP2C9 and VKORC1 genetic identifications before prescribing to avoid warfarin toxicity.

CYP3A5 variations

The CYP3A5 gene encodes an enzyme catalyzing many reactions involved in drug metabolism such as olanzapine, tacrolimus, nifedipine and cyclosporine. The clinical impact of CYP3A5 was the adverse event of tacrolimus in patient having CYP3A5*3 which represent as a nonexpressor group as the given mutation causing the non-expressing enzyme. Our results demonstrated allele frequencies of Caucasians, Africans and other Asians with almost equal values as reported previously with the highest prevalence in Africans and the lowest prevalence in Caucasians.18 For Asians, the prevalence of CYP3A5*3 occurred at around 40–50%. The clinical importance of CYP3A5 involved in tacrolimus metabolism was because the patient with CYP3A5*3 was required to reduce the dosage of tacrolimus to avoid the toxicity of this drug.19 Although, there is no regulation announced by US-FDA/EMEA to test this marker before tacrolimus administration, our data now strongly support the idea that there is a high possibility of Thai and other Asians facing tacrolimus toxicity owing to the high prevalence of this gene in these populations. Therefore, investigating the polymorphism of CYP3A5 before using this drug is highly recommended.

NAT2 variations

NAT2 is responsible for N-acetylation of arylamine drugs and displays genetic variation to be a slow acetylation or rapid acetylation enzyme depending on its polymorphism. NAT2*5, NAT2*6 and NAT2*7 are the major groups of alleles that are associated with decreased enzyme activity (owing to amino acid changes) and demonstrating the slow acetylator phenotype. Not surprisingly, our data also demonstrated the concordance of allele frequency of this gene in Thai when compared with the previous report.20 The outstanding clinical impact of NAT2 was the adverse drug reaction of isoniazid (INH) as the slow acetylator phenotype has been postulated to be responsible for the development of INH-induced hepatitis. INH is an anti-tuberculosis drug that is principally metabolized by NAT2. Ethnicity is the risk factor for hepatitis from anti-tuberculosis,21 which might partly be explained by their higher proportion of slow acetylator genotype and phenotype. There are many reports illustrating the correlation between the slow acetylator phenotype and the INH-induced hepatotoxicity, therefore, suggesting to monitor and consider adjusting the INH dosing in Thai and other Asian populations must be highly supported.

SLCO1B1 variations

SLCO1B1 gene encodes an organic anion transporter 1B1, which is responsible for the drug clearance. Several SNPs have been identified in the SLCO1B1 gene encoding for OATP1B1. One common SNP of the SLCO1B1 gene, SLCO1B1*1B, has been associated with increased OATP1B1 transport activity. In our current study, the allele frequency of SLCO1B1*1B gene in Thai populations demonstrated approximately twice higher than that in Caucasians, therefore, the effect of this transporter could be found highly impact for Asians including Thai populations. The clinical importance of the SLCO1B1*1B involved in repaglinide as previously reported in Chinese.22 In addition, SLCO1B1*1B has been proved that it had strong association with statin-induced myopathy.23 Although, there is no regulation to investigate the variation of this transporter gene, our data recommended to examine the polymorphism of this gene before taking repaglinide to gain the efficacy of the treatment as this issue seems to be more important for Asians rather than Caucasians owing to the difference in allele frequency among ethnicities.

In conclusion, our data not only demonstrated the allele frequency of drug ADME-related genes in Thai, but also indicated the clinical impact of these SNPs of drug ADME-related genes. The information of genetic variations in phase I enzyme, phase II enzyme and drug transporters together with the drugs substrates/inhibitors and their clinical impact listed in Table 4 must be carefully reviewed as our findings could be a pinpoint to develop the system in improving the national drug formulary, which could be useful in many levels such as developing the biomarker for drug label and using our data to support for revising Thai National List of Essential Medicines (NLEMs) in the future.