Association between the CYP2B6 polymorphisms and nonnucleoside reverse transcriptase inhibitors drug-induced liver injury: a systematic review and meta-analysis Authors: Noppadol Chanhom, Janjira Sonjan, Jarupat Inchai, Wanvisa Udomsinprasert, Usa Chaikledkaew, Supharat Suvichapanich, Surakameth Mahasirimongkol, Jiraphun Jittikoon Published: 2024 ABSTRACT AbstractNevirapine (NVP) and Efavirenz (EFV) can cause antiretroviral drug-induced liver injury (ARVDILI). The objectives of this study were to summarize and analyze existing data on pharmacogenomics associated with nonnucleoside reverse transcriptase inhibitors drug-induced liver injury using systematic review and meta-analysis. This study systematically searched the relevant studies regarding pharmacogenes related to ARVDILI from online databases. Genes-encoding proteins were further analyzed using the STRING program to determine the protein-protein interactions (PPI). CYP2B6 polymorphisms were further meta-analyzed. Seventeen genes have been shown to be significantly associated with ARVDILI. Illustration from STRING analysis, CYP2B6, CYP1A1, and CYP2D6 enzymes have been recognized as central proteins linked to all other analyzed proteins. Meta-analysis illustrated that CYP2B6 *1/*6 (OR = 1.83; 95% CI: 1.15–2.90; P = 0.01), *6/*6 (OR = 2.48; 95% CI: 1.28–4.79; P = 0.007), and *1/*6 plus *6/*6 (OR = 1.94; 95% CI: 1.24–3.01; P = 0.003) were associated with risks of EFV-induced liver injury. Moreover, CYP2B6 *1/*6 (OR = 0.44; 95% CI: 0.22–0.91; P = 0.03) and a group combining individuals with either *1/*6 or *6/*6 (OR = 0.42; 95% CI: 0.21–0.84; P = 0.01) were associated with reduced risks of NVP-induced liver injury. This meta-analysis revealed an association between CYP2B6 genetic polymorphism and susceptibility to ARVDILI.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-024-79965-0.Keywords: Drug-induced liver injury, CYP2B6, Human immunodeficiency virus, Hepatotoxicity, Systematic review, Meta-analysis, Adverse effect, Genetic polymorphismsSubject terms: Clinical genetics, Predictive markers, Predictive markers, Genetics research Introduction Since its discovery in 1996, nonnucleoside reverse transcriptase inhibitors (NNRTIs) have been used in combination with other antiretroviral (ARV) medications to help reduce overall morbidity and mortality associated with human immunodeficiency virus infection1. Despite these beneficial features of NNRTIs, which significantly improves the prognosis and quality of life of people living with Human Immunodeficiency Virus (PLHIV), some PLHIV with ARV develop severe adverse drug reactions (ADRs), especially drug-induced liver injury (DILI). The reported incidence of ARVDILI can range from 8 to 23% of PLHIV; nevertheless, up to 30% of these PLHIV require a change of regimen or treatment discontinuation, which subsequently causes prolonged hospitalization, in which severe cases will be fatal2–5, If not detected and treated promptly, it may potentially cause severe hepatitis and even death in some PLHIV6,7. Nevirapine (NVP) is the NNRTI used in countless PLHIV due to its proven efficacy, advantageous metabolic profile, cost-effectiveness, and particularly its significance in strategies aimed at simplifying combination antiretroviral therapy (cART)8–10. Unfortunately, approximately 6–17% of NVP recipients experienced ARVDILI, with up to 12% developing severe ARVDILI11–14. In terms of its biological metabolism, cytochrome P450 primarily facilitates the biological metabolism of NVP, resulting in the formation of four mono-oxygenated metabolites. 12-OH-NVP, a byproduct of CYP, has reportedly been associated with the development of rash, which may play a role in hepatic toxicity15. In addition to NVP, Efavirenz (EFV), one of the NNRTIs, is used in numerous successful first-line HIV regimens. However, efficacy, tolerability, and availability made it an exception for certain PLHIV16. Unfortunately, EFV administration was associated with ARVDILI. Drug accumulation in the liver was considered the perpetrator17. To increase their safety, it is essential to fill the gap in knowledge about NNRTIs’ effects on the liver and their role in DILI development18. Interestingly, genetics has been hypothesized as one of the critical contributors to ARVDILI pathogenesis in several populations19. According to several findings from previous studies as a putative mediator for liver disease in PLHIV, CYP2B6 is the most promising of numerous intriguing genes associated with ARVDILI. It is responsible for several drugs metabolism and synthesizing cholesterol, steroids, and other lipids20. Interestingly, CYP2B6 enzymes are also the primary metabolizers of nevirapine and efavirenz. In all investigated populations, CYP2B6*6 is the most common variant allele20. CYP2B6*6 haplotype is characterized by the existence of two variants, G516T (rs3745274) and A785G (rs2279343), with or without promoter variants21. This allele has been associated with a 50–75% decrease in CYP2B6 enzyme levels but not with the protein’s active or substrate recognition sites22–24. In that context, a decrease in protein level might help explain why CYP2B6*6 was associated with an increased risk of ARVDILI from efavirenz25 and nevirapine26. In contrast to the above findings, other previous studies13,27,28 unveiled no correlation between CYP2B6*6 polymorphisms and susceptibility to ARVDILI due to efavirenz as well as nevirapine. To clarify the above associations, a meta-analysis, a valuable tool for deriving meaningful conclusions from data and for resolving inconsistencies found in research, is needed. Accordingly, the objectives of this study were to summarize and analyze existing data on pharmacogenomics associated with nonnucleoside reverse transcriptase inhibitors drug-induced liver injury using systematic review and meta-analysis approaches along with network analysis to gain insight into molecular interactions of these pharmacogenes and their genetic polymorphisms associations with ARVDILI. Results A systematic review of genetic polymorphisms associated with ARVDILI Eight hundred forty-one publications were collected from PubMed, Scopus, and Cochrane databases using the search terms specified in methodologies. After data refinement using exclusion criteria, 165 duplicated publications were removed. In addition, 377 publications were excluded due to unrelated topics, 247 were not primary sources of data, 2 were studies in vitro and in silico models, and one was published in a language other than English. In the end, 50 different papers were considered for inclusion in this study. However, 39 publications reported on associations of genetic polymorphisms with ARVDILI. The top-fifth of the highest number of reports were 11 for UGT1A1, 11 for CYP2B6, four reports of ABCB1, two for HLA-Cw, and two for HLA-DRB1 (Figure S1). The protein-protein interaction (PPI) analysis of collected ARVDILI-associated genes depicted interactions among several genes. In the pathway collected by the Kyoto Encyclopedia of Genes and Genomes (KEGG), alternative findings from a PPI network showed that the network is functionally enriched in the metabolism of xenobiotics29. The results illustrated that the ARVDILI was primarily involved in xenobiotic metabolism. Interestingly, the STRING database indicates that CYP2D6, CYP1A1, and CYP2B6 seem to be the center of the interaction among those proteins (Figure S2). In PPI analysis, CYP2B6 was selected as a candidate gene due to its frequent reports and centrality. From this, CYP2B6 genetic polymorphisms were further analyzed using the Cochran-Mantel-Haenszel analysis approach. Meta-analysis of CYP2B6 associated with efavirenz-induced liver injury Study selection and characteristics The selection process is registered in PROSPERO (ID: CRD42024593948) and illustrated in Fig. 1. Initially, 350 publications were included in this study. Publications were initially included for publication appraisal. Following the Newcastle-Ottawa scale for case-control and cohort study, three publications with 106 cases of ARVDILI from a total of 795 PLHIV were selected. The publication years ranged from 2011 to 2012, as described in Table 1. All the studies were conducted on the African population. There were two cohort studies and one case-control study. Flow diagram of this meta-analysis of CYP2B6 association with EFV-induced liver injury. A total of 350 publications were retrieved from databases up to 10 October 2024. Three studies were selected as eligible publications after the screening and quality assessment process. Table 1Characteristics of reference studies.ReferencesStudy designCountryEthnicityControl SourceDILIdefinitionARVDILI(case)Non-ARVDILI (control)Mean age of the caseMean age of controlSex in case(male/female)Sex in control(male/female)Publications in the Efavirenz groupMugusi et al. (25)CohortTanzaniaAfricanHBCIOMS2832141.8139.7821/16182/254Yimer et al. (28)CohortEthiopiaAfricanHBCIOMS37208NANA9/3266/154Yimer et al. (27)Case-controlEthiopiaAfricanHBCIOMS41160NANA26/39140/148Publication in Nevirapine groupCarr et al. (26)Case-controlMalawiAfricanHBALT > 40 IU/L with jaundice3042837.436.278/131205/258Case-controlUgandaAfricanHB2946037.436.2Giacomelli et al. (29)Case-controlItalyCaucasianNARUCAM835439.638.52/6227/127CIOMS, Council for International Organizations of Medical Science; HB, hospital-based; NA, not available; RUCAM, Roussel Uclaf Causality Assessment Methods. Characteristics of reference studies. CIOMS, Council for International Organizations of Medical Science; HB, hospital-based; NA, not available; RUCAM, Roussel Uclaf Causality Assessment Methods. Meta-analysis results The pooled effects of CYP2B6 polymorphisms on EFV-induced liver injury occurrences in PLHIV are summarized in Table 2 and illustrated in Fig. 2. Compared with normal metabolizers, CYP2B6 *6 carriers were significantly more susceptible to EFV-induced liver injury, with an odds ratio (OR) and 95% confidence interval (CI) of 1.83 (1.15–2.90), 2.48 (1.28–4.79), and 1.94 (1.24–3.01) for heterozygous, homozygous, and combination of both alleles, respectively, as well as P value less than 0.05 (Table 2). Table 2Summary of meta-analyses between CYP2B6 genotypes and the risk of ARVDILI.CYP2B6 genotype compared with CYP2B6 *1/*1Test of heterogeneityEffects modelMutant genotype countNormal metabolizer genotype countsTotalNumber of included studiesOR95% CIP-valueI2(%)P-valuecasecontrolcasecontrolEfavirenz-induced liver injury*1/*600.53Fixed552923431970031.831.15–2.900.01*6/*600.87Fixed17783431944832.481.28–4.790.007*1/*6 + *6/*600.56Fixed723703431979531.941.24–3.010.003Nevirapine-induced liver injury*1/*600.97Fixed153831929771420.440.22–0.910.03*6/*600.85Fixed41021929753320.450.15–1.320.14*1/*6 + *6/*600.89Fixed194851929782020.420.21–0.840.01ARVDILI, antiretroviral drug-induced liver injury; CI, confidence interval; OR, odds ratio. Summary of meta-analyses between CYP2B6 genotypes and the risk of ARVDILI. *1/*6 + *6/*6 *1/*6 + *6/*6 ARVDILI, antiretroviral drug-induced liver injury; CI, confidence interval; OR, odds ratio. Forest plot comparison between CYP2B6 variance and wild-type and the susceptibility of EFV-induced liver injury from 3 publications. CI, confidential interval. Sensitivity, heterogeneity, and publication bias analyses To determine the impact that each study had on the total estimate, a sensitivity analysis was performed in which a single study was excluded at a time. The findings of the study indicated that the sensitivity test for the meta-analysis of CYP2B6 genotypes was successfully completed. Heterogeneity had not been noticed. A funnel plot was utilized in order to analyze publication bias. The condensed plot that is provided at the top of the 95% confidence interval triangle reveals publication bias for studies that had a large number of PLHIV participants, as shown by the funnel plots. But the plots were symmetrical between the sides of the triangle, which showed that there was no publication bias observed for either negative or positive results (Figure S3). Meta-analysis of CYP2B6 associated with nevirapine-induced liver injury Study selection and characteristics The selection process is registered in PROSPERO (ID: CRD42024593948) and demonstrated in Fig. 3. To begin, 356 publications were included for selection. Then, due to duplication, 49 publications were removed. Following that, 307 publications were screened, with 305 excluded from the analysis; most studies fulfilled the exclusion criteria, while some lacked eligible data for analysis. Finally, two publications were included for the appraisal (Supplementary Table S1), including 38 ARVDILI and 782 non-ARVDILI PLHIV (Table 1). Flow diagram of this meta-analysis of CYP2B6 association with NVP-induced liver injury. A total of 347 publications were retrieved from databases up to 10 October 2024. Two studies were selected as eligible publications after the screening and quality assessment process. Meta-analysis results The association of CYP2B6 with NVP-induced liver injury sensitivity in PLHIV is detailed in Table 2. Meta-analysis showed that CYP2B6 *1/*6 allele and combination of *1/*6 + *6/*6 alleles were significantly associated with a reduction in susceptibility to NVP-induced liver injury, compared with normal metabolizer (Fig. 4), with pooled OR and 95% CI of 0.44 (0.22–0.91) and 0.42 (0.21–0.84), respectively, as well as P value less than 0.05 (Table 2). However, there was no significant association between CYP2B6 homozygous *6 with ARVDILI risk (Fig. 4). Forest plot comparison between CYP2B6 variance and wild-type and the susceptibility of NVP-induced liver injury from 2 publications. CI, confidential interval. Sensitivity, heterogeneity, and publication bias analyses According to the results of the sensitivity analysis, the substantial relationships between the CYP2B6 *1/*6 allele and a combination of *1/*6 + *6/*6 alleles with a lower risk of NVP-induced liver injury were not attributable to chance. Heterogeneity had not been noticed. An investigation on publication bias was carried out utilizing a funnel plot, which revealed that publication bias was not discovered for either a negative or positive result (Figure S4). Discussion Recently, genetic polymorphisms associated with ARVDILI have been studied, increasing the possibility of utilizing them as biomarkers to predict the chance of developing ARVDILI28. Consistent with this, our systematic review showed numerous reports on the relationship between genetic polymorphisms and ARVDILI in several populations. The fifth highest frequency of reported gene-associated ARVDILI was UGT1A1, CYP2B6, ABCB1, HLA-Cw*04, and HLA-DRB1, respectively. Additionally, the PPI analysis illustrated the functional enrichment of the protein in the network involved in the metabolism of xenobiotics according to the Kyoto Encyclopedia of Genes and Genomes (KEGG). The results suggested that the ARVDILI was mainly affected by the function of xenobiotic metabolisms. In addition, PPI network analysis of these gene-encoded proteins demonstrated links between those proteins with CYP1A1 and CYP2B6 as central to their interactions. The result demonstrated the association of CYP1A1 and CYP2B6 proteins with other proteins in the network. Due to the centrality in the PPI network and reports frequency, the CYP2B6 protein has been hypothesized to be the crucial molecule. Consequently, based on this premise, we conducted meta-analyses to verify the relationship between CYP2B6 polymorphisms and their association with ARVDILI susceptibility. Our findings demonstrated that a genotype known as CYP2B6*6, which is related with a decreased ability to metabolize drugs, was linked to an increased risk of ARVDILI in PLHIV patients who were using EFV. Conversely, the mentioned genotype is also associated with a reduced risk in nevirapine-administered PLHIV. CYP2B6 is the drug-metabolizing enzyme involved in EFV and NVP metabolisms, especially EFV. Plentiful evidence obtained from hepatocyte experiments uncovered that EFV compromised cell viability and altered mitochondrial function by diminishing mitochondrial membrane potential, decreasing mitochondrial O2 consumption, and increasing ROS production30–33. In addition to EFV, 8-hydroxyefavirenz (8-OH-EFV), the primary metabolite of EFV, generally has the capability to induce cell death and ROS production in hepatocytes34. Therefore, the EFV and its primary metabolite are the reactive forms that can cause ARVDILI in PLHIV. In addition, according to a previous study, the CYP2B6 enzyme metabolizes EFV into 8-OH-EFV, then 8,14-diOHEFV, respectively35. Therefore, it is unsurprising that the poor metabolizer form of the CYP2B6 enzyme, such as *6, has been linked to ARVDILI in several populations. On the contrary, in NVP’s circumstance, the meta-analysis results showed that CYP2B6 *1/*6 and a combination of *1/*6 and *6/*6 were dominantly found in the controls group, thus indicating that the poor metabolizer form of CYP2B6 may reduce the risk of ARVDILI. In contrast to this result, the LiverTox database and a previous study by Ciccacci et al. both depicted that the CYP2B6 inducer produced a toxic intermediate and subsequently caused DILI in PLHIV36,37. This contradictory result may be explained by the fact that NVP has several metabolizing enzymes and numerous metabolites, resulting in the production of 2-, 3-, 8-, and 12-hydroxynevirapine (-OH-NVP) and 4-carboxynevirapine. Unlike EFV, the active metabolite produced by CYP2B6, 8-OH-NVP, is not the most abundant metabolite. Moreover, the major metabolite of NVP produced by CYP3A4, 12-OH-NVP38, has been implicated in hepatotoxicity and skin rash caused by NVP15,39,40. Therefore, to draw a clear explanation of the association between CYP2B6 polymorphisms and hepatotoxicity in NVP-administered PLHIV, the genotypic information of CYP3A4 is required. This discrepancies between EFV and NVP results may cause different interpretation of CYP2B6 genotyping results between two medication users. While poor metabolizer of CYP2B6 may cause DILI in EFV using PLHIV and should decide the regimen alteration, NVP using PLHIV may continue without any interaction. However, due to the limited number of publications and sample size, another study on this topic is encouraged to identify the exact reason for conflicting results. The CYP2B6 enzyme is not the exclusive metabolizer of EFV and NVP41. UGTs, CYP3A4, CYP3A5, CYP1A2, and CYP2A6 may have minimal contributory roles in the xenometabolic phase of EFV22,42–44. CYP3A4, CYP3A5, CYP2C9, and CYP2D6 are also involved in the metabolism of NVP45–47. Consequently, the influence of several genes on ARVDILI susceptibility seems inescapable. Chanhom et al. observed the effects of various genes in ATDILI within the Thai population48. This finding also indicates the potential for various gene association studies in ARVDILI. Nevertheless, further studies are needed to investigate the specific interactions between these genes and ARVDILI susceptibility, especially in diverse populations. Understanding the role of genetic variations in drug metabolism can help improve personalized medicine and reduce the risk of adverse reactions. Overall, the complex interplay between genes and DILI underscores the importance of pharmacogenomics in optimizing patient care. Noteworthy, some inherent limitations need to be considered when evaluating the presented findings. First, we included publications published only in English; thus, language bias was unavoidable. Accordingly, we lost reports published in other languages, possibly leading to selection bias. Moreover, we performed a PPI analysis to investigate the potential of the protein/gene for further meta-analysis. However, the PPI function is quite restricted and cannot be confirmed by clinical outcomes. The authors encouraged further bioinformatic study to investigate the interplay among genes, proteins, and clinical outcomes in ARVDILI. Moreover, one of the possible confounders was DILI definition due to the change of the definition in 201649 which may affect the DILI interpretation discrepancies between studies. The author encouraged further observational studies or trials to comply with the updated RUCAM criteria for DILI causality assessment49. Along with this, some publications were excluded due to our rigid selection protocol, which might be the possible reason for sparse publications and sample sizes and could cause inconsistent results, especially in NVP-induced liver injury. Furthermore, the limited number of publications and sample size in this meta-analysis adversely affects the generalizability of its conclusions. The meta-analysis of the link between CYP2B6 genotypes and EFV-induced liver injury was only derived from the African population, as well as NVP-induced liver injury were obtained from only African and Caucasian populations. Consequently, referencing this may produce divergent outcomes in other populations with varying polymorphism frequencies. Contrastingly, this meta-analysis highlights the research deficit in different populations, and a rigid selection protocol might construct an exceptional foundation for future studies. For this reason, we might mitigate the effect of the confounding factors, which provide us with more precise analysis results. Notably, this is the first systematic review and meta-analysis on the association between CYP2B6 polymorphisms and ARVDILI in PLHIV. In conclusion, our systematic review illustrated that UGT1A1, CYP2B6, and ABCB1 were the most frequently studied genes. This meta-analysis illustrated that CYP2B6 *6 and *1 was significantly associated with EFV-induced liver injury and NVP-induced liver injury in PLHIV, respectively. Additionally, the results of this study may use as a fundamental knowledge to develop decision support guideline for drug selection and personalized medicine for PLHIV in the near future. Further observational studies with larger sample size, well-characterized subjects, and various ethnicities are warranted to gain a precise conclusion and develop those polymorphisms as genetic markers for predicting PLHIV with a high ARVDILI risk induced by EFV and NVP. Materials and methods Gene associated with ARVDILI exploration Scopus and PubMed were combed using the following search terminology. ((“ARV” OR “Antiretroviral treatment” OR “antiretroviral agent*” OR “Antiretroviral drug*"OR “Anti-HIV Agent*” OR “HIV therapy” OR “HIV treatment” OR “HAART” OR “NNRTI*” OR “nonnucleoside reverse transcriptase inhibitors”) AND (“genetic polymorphism*” OR “polymorphism*” OR “Pharmacogenetic*” OR “Immunogenetic*” OR “Pharmacogenomic*” OR “genome wide association study” OR “variant allele*” OR “genetic predisposition” OR “Gene Variant*”)) AND (“ARV-associated hepatotoxicity” OR “drug-induced hepatotoxic” OR “Drug-induced hepatotoxicity” OR “Hepatotoxic” OR “Hepatotoxicity” OR “Liver toxicity” OR “Liver toxic” OR “drug-induced liver injury” OR “DILI” OR “liver injury” OR “Drug induced hepatitis” OR “Drug-induced hepatitis” OR “Hepatitis” OR “hypersensitivity reaction” OR “hypersensitivity*” OR “hyperbilirubinemia”). From inception to 1 August 2023, original research articles on the association between genetic polymorphisms and ARVDILI were gathered. The studies with negative results were exclusively excluded from the analysis. Frequencies of association reports of each gene-ATDILI association were reported. Proteins associated with ARVs were analyzed using STRING online software version 11.5 (https://string-db.org/)50. The STRING integrated and classified protein/gene associations by benchmarking them in experimental and predicted physical/theoretical protein-protein interactions data. By the criteria of the STRING standard, the interaction network was deemed to have a confidence score of more than 0.70, and all interaction sources were selected, including text-mining, experiment, databases, co-expression, neighborhood, gene fusion, and co-occurrence. Each protein was represented as a node in the protein-protein interaction (PPI) network, and an interaction between two proteins in the network was denoted by an edge. These interactions might either be physical or theoretical. Meta-analysis of CYP2B6 associated with efavirenz/nevirapine-induced liver injury Search strategies of the literature for meta-analysis On 10 October 2024, the original publication search was performed using electronic databases of PubMed, Scopus, and Web of Science. The PICO strategy was used throughout the search process51. Following is the definition of the PICO: Population (P): PLHIV who receive efavirenz or nevirapine medication, Intervention (I): CYP2B6 genotype associated with an increased risk of ARVDILI in PLHIV, Comparator (C): CYP2B6 normal metabolizer, and Outcomes (O): hepatotoxicity. The search techniques were developed by combining several search phrases with various Boolean operators52. The process of selecting studies for meta-analyses The inclusion criteria were as follows: (i) the study subjects were PLHIV receiving EFV or NVP; (ii) the study investigated the association between CYP2B6 polymorphisms and ARVDILI susceptibility; and (iii) the study was a randomized controlled trial, cohort, or case-control design. The exclusion criteria were as follows: (i) the study was either a case report secondary or tertiary publication such as a review, systematic review, and meta-analysis, (ii) the study was conducted in an animal model, in vitro, or in silico, iv) full article cannot be accessed, (iii) the study was not published in English, (iv) the study did not observe the liver function test, (v) the study included confounding factors such as viral hepatitis infection, alcoholism, or concomitant hepatotoxic medication, (vi) the study included vulnerable subjects such as children, pregnant/breastfeeding women, or elderly, (vii) the study is unrelated to the association between at CYP2B6 polymorphisms and ARVDILI susceptibility or does not have sufficient data for analysis. Assessment of selected study quality The Newcastle-Ottawa scale was utilized by two reviewers, JS and JI, in order to conduct independent evaluations of the articles’ overall quality53. The issue that existed between the two reviewers was finally settled by conversation and agreement with the third reviewer (NC). The scores on the Newcastle-Ottawa scale range from 0 to 9, with 0 being the lowest and 9 being the highest. If the publication’s overall evaluation score was higher than 6, then it was determined that the publication successfully passed the evaluation (Supplementary Table 1). The studies were grouped according to the medication involved in the study, i.e., either the NVP group or the EFV group. Data extraction for meta-analysis The information from each of the research that were considered to be relevant was compiled into a single sheet of data. The following information was taken from the article: first author, publication year, study design, ethnicity, sample size, gender, age, BMI, observed medication, DILI definition and the number of ARVDILI/non-ARVDILI for each CYP2B6 genotype along with odds ratios and 95% confidence intervals. In the event where two of the reviewers, JS and JI, could not come to an agreement regarding the findings of the data extraction, the issue was settled through discussion and by reaching a consensus with the third reviewer (NC). Statistical analysis Meta-analysis of CYP2B6 gene effect sizes was conducted using Computer program Review Manager (RevMan®). Version 5.3. Copenhagen: The Nordic Cochrane Centre, the Cochrane Collaboration, 2014. In accordance with low heterogeneity among studies and similarity of population and study design, The Mantel-Haenszel fixed-effects model was used to calculate pooled OR, 95% CI, and P-value54. In order to be declared statistically significant, the P-value has to be lower than 0.05. The Chi-square test or an I2 value of more than 50% were utilized to make an approximation of the heterogeneity that existed among the studies55. The results of the individual research and the synthesis were plotted out in the forest plot so that they could be viewed visually. A funnel plot was then used to establish whether or not there was any publishing bias. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements This research project has been funded by Mahidol University (Fundamental Fund: fiscal year 2024 by National Science Research and Innovation Fund (NSRF))The authors want to thank the Faculty of Pharmacy, Mahidol University, for providing access to the databases and the authors of the original articles for providing useful information. In addition, we want to thank Genomic Medicine Centre, Division of Genomic Medicine and Innovation Support, Department of Medical Sciences, Ministry of Public Health for providing useful suggestions and comments throughout our research. Abbreviations Antiretroviral drug-induced liver injury Combination antiretroviral therapy Confidence interval Drug-induced liver injury Efavirenz Nevirapine Odds ratio People living with HIV Protein-protein interaction Author contributions Conceptualization, J.J., W.U., U.C., S.M., and S.S.; Data curation, N.C., J.S. and J.I.; Formal analysis, N.C., J.S., J.I. and J.J.; Funding acquisition, J.J., W.U., U.C., and S.S.; Investigation, J.J., W.U., S.S., N.C., J.S. and J.I.; Methodology, J.J., W.U., U.C., S.M., and S.S.; Project administration, J.J.; Supervision, J.J.; Validation, J.J., W.U., and S.S.; Visualization, N.C., J.S. and J.I.; Writing – original draft, N.C.; Writing – review & editing, J.J., W.U., and N.C. Funding This research project has been funded by Mahidol University (Fundamental Fund: fiscal year 2024 by National Science Research and Innovation Fund (NSRF)). Data availability Data is provided within the manuscript or supplementary information files. Declarations Ethics declarations This study is a systematic review and meta-analysis. Therefore, the ethics approval is not a requirement. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement Data is provided within the manuscript or supplementary information files.