Carbamazepine (CBZ) is a first-generation antiepileptic drugs (AEDs) used to treat a variety of neurological and psychiatric problems, including epilepsy, trigeminal neuralgia, and bipolar disorders. In Thais, there has been a well-documented association between HLA-B*15:02 and CBZ-induced Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) (Hung et al., 2006; Locharernkul et al., 2008; Tassaneeyakul et al., 2010; Tangamornsuksan et al., 2013; Sukasem et al., 2018; Sukasem et al., 2021b). In 2013, the government launched a policy requiring HLA-B*15:02 pharmacogenetic (PGx) testing before the start of CBZ in Bangkok as a pilot study (Department of Medical Sciences, 2013). Subsequently, national policy screening for HLA-B*15:02 was reinforced by the National Health Security Office (NHSO) in 2018 throughout Thailand (Ang et al., 2017; Chang et al., 2020; Sukasem et al., 2021a; Jantararoungtong et al., 2021). Besides HLA-B*15:02, some other variants such as HLA-B*15:08, HLA-B*15:11 and HLA-B*15:21 may also affect the safety of CBZ and found to be other HLA-B risk alleles associated with CBZ-induced cutaneous adverse drug reactions (CADRs) (Jaruthamsophon et al., 2017; Volpi et al., 2018; Kloypan et al., 2021).

In addition, Chen et al. reported a strong association between the HLA-B*15:02 allele and oxcarbazepine (OXC)-induced SJS/TEN in Chinese and Thai populations (Chen et al., 2017). Consequently, the Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline recommended to avoid the prescription of CBZ and OXC in HLA-B*15:02 carrier (Phillips et al., 2018).

Although knowledge and research on PGx testing in Thailand have been rapidly expanding in recent years (Jaruthamsophon et al., 2017), there has been no longitudinal research on the application of HLA-B*15:02 genotype results in daily clinical practice, including documentation of severe cutaneous adverse drug reactions (SCARs) and new adverse drug reactions (ADRs) in electronic medical records, patterns of AEDs prescriptions based on genotype results, and assessment of the CADRs. This is the first research of its kind in Thailand, and it focuses on these fascinating topics.

The goal of this study was to investigate how PGx testing (HLA-B*15:02) affected the CBZ/OXC prescriptions and the reduction of CBZ/OXC-induced SCARs in Thais. The findings of this study have the potential to help physicians regarding the prescription of appropriate AEDs in clinical practice for those patients carrying HLA-B*15:02 CBZ/OXC-risk alleles and to support NHSO Thailand’s HLA-B*15:02 screening as the national policy.

This is the first retrospective cohort study in Thailand to determine the clinical impact of pharmacogenomics (PGx) testing on carbamazepine/oxcarbazepine (CBZ/OXC) prescribing. The female preponderance was seen in the PGx testing population in this study, and it may be because of the higher incidence of hypersensitivity reactions among females (Saksit et al., 2017). Many earlier research studies support this fact. For example, a study conducted by Alvestad et al. (2007) reported that a striking difference in the incidence of hypersensitivity was seen between males (8%) and females (19%) with p < 0.001. This association is highly dependent on the age group; reproductive years 20–50 are the most significant age group. The explanation for this is that sex hormones alter the immunological process of rash. Androgens limit the inflammatory response more than endogenous glucocorticoids, whereas female sex hormones enhance the immunological response both pathologically and physiologically (Talal, 1987; Da Silva, 1999; Osman, 2003).

Pharmacogenomic testing for HLA-B*15:02 is now standard practice in Thailand. In this study, physicians requested the PGx testing for CBZ/OXC and hold the prescription until received the PGx report for 85.67% of 384 patients. This was complying with Clinical Practice Guidelines for Epilepsy (2018) of Epilepsy Society of Thailand and the Clinical Pharmacogenetics Implementation Consortium Guideline for HLA Genotype and Use of Carbamazepine and Oxcarbazepine: 2017 Update (Epilepsy Society of Thailand, 2018; Phillips et al., 2018). Do not use CBZ in naïve-patients that are positive for HLA-B*15:02. If patient used CBZ for longer than 3 months without incidence of adverse reactions, consider use with caution (Phillips et al., 2018). However, seven patients were found to receive CBZ/OXC while requesting the HLA-B*15:02 PGx testing. Two of them were CBZ naïve-patients and prescribed with CBZ due to clinical urgency and five were identified as CBZ/OXC-treated patients (used CBZ/OXC for longer than 3 months). Therefore, alternative drugs were prescribed in 65 (92.86%) patients who carried HLA-B*15:02 (positive group).

Among the 314, negative HLA-B*15:02 group, 60% were prescribed with CBZ/OXC, whereas the remaining 40% switched over to other AEDs overall before and after the PGx test. The retrospective follow up analysis showed that lengthy HLA-B testing results turnaround time and/or lengthy reach out time of the available PGx testing results to physicians were the two important reasons for it. This is a very common barrier while translating the PGx into the real clinical world (Krebs and Milani, 2019). Thailand has recently seen a lot of technological improvement in genetic analysis but we used samples from 2011 in our research. This could explain why clinician prescribed CBZ/OXC without waiting for the PGx results in 55 (14.32%) patients. Physicians, on the other hand, awaited for the PGx results of 329 patients before making a clinical decision. This supports our hypothesis that physician CBZ/OXC prescribing decisions were influenced by HLA-B*15:02 genotyping. Moreover, 92.85% (65/70) of HLA-B*15:02 risk allele carriers were avoided by CBZ/OXC, so it can be assumed that the risk population is separated and SCARs are prevented in this population.

However, the screening for HLA-B*15:02 is not a compulsory testing in Thailand prior to initiation of CBZ treatment. During 2013–2019, CBZ-induced SCARs were reported from 13 patients who received CBZ without PGx screening in Thai-SCARs cohort. Remarkably, HLA-B*15:02 were positive in all SJS/TEN cases (n = 10, 100%), where as negative HLA-B*15:02 in DRESS and FDE. Assumably, these patients can be prevented from SJS/TEN if the HLA-B*15:02 PGx testing was required as a mandatory testing for CBZ prescription. Additionally, the hospital policy and clinical decision support (CDS) alert system is essential to overcome the barriers associated with the utilization of PGx guidelines into clinical practice. The PGx-CDS should be integrated into EMRs and pop-up alert in the therapeutics order system recommending the clinician to order a PGx screening test prior initiation of preventable drug-induced SCARs such as CBZ/OXC (HLA-B*15:02), allopurinol (HLA-B*58:01), abacavir (HLA-B*57:01) and cotrimoxazole/dapsone (HLA-B*13:01) (Sukasem et al., 2016; Sukasem et al., 2020; Kloypan et al., 2021; Satapornpong et al., 2021).

An earlier study in Thailand reported that out of 214 SCARs-treated patients over a period of 10 years, 23 patients (10.6%) were treated with CBZ, and this shows that CBZ is the major culprit for SCARs. If MPE and other mild-moderate level reactions were included, this number could go higher than this. However, it cannot be assumed that all the risk allele carriers will have CBZ-induced CADRs. In this study, 7.14% of patients who carried HLA-B*15:02 allele did not manifest any CADRs. Because the frequency of SJS/TEN is substantially lower than the frequency of the risk allele in these groups, previous studies have shown that the test’s positive predictive value is poor (1–5%). Although having HLA-B*15:02 increases the risk of CBZ/OXC induced SJS/TEN by up to >700-fold compared to non-carriers, the vast majority of patients with HLA-B*15:02 (95%–99%) do not develop SJS/TEN from CBZ/OXC (Amstutz et al., 2014).

However, in this study if the mild reactions are included this rate goes up to 11%. The reason could be that 29% of the negative patients were reported with multiple comorbidities and most of them were immune related disorders like Systemic lupus erythematosus (SLE), hypo/hyperthyroid, cancer, rheumatoid arthritis (RA), diabetes mellitus (DM), hypertension (HTN), and central nervous system (CNS) infections. Earlier studies showed that some non-clinical risk factors may be involved in the initiation of CADRs (Campos-Fernández Mdel et al., 2005; Patel et al., 2014; John et al., 2021).

Unexpectedly, 2 cases with CBZ-induced SCARs (DRESS and SJS) were reported among 189 patients with HLA-B*15:02 negative after CBZ/OXC treated patients. Likewise, 2 cases with CBZ-induced DRESS were reported after CBZ treatment without HLA-B*15:02 PGx testing. Commonly, the HLA-B*15:02 allele has been reported to be specifically associated with CBZ-induced SJS/TEN in Asian populations, and no associations have been reported for drug-induced MPE and DRESS (phenotype-specific biomarker) (Sukasem et al., 2021b). In contrast, HLA-A*31:01 was reported to be associated with CBZ-induced DRESS and MPE (Mockenhaupt et al., 2019; Ahmed et al., 2021). US-FDA recommended that HLA-B*15:02 and HLA-A*31:01 genotypes should be detected to stratify the high risk patients prior CBZ prescription. The use of CBZ should be avoided in patients who test positive for the HLA-A*31:01 or HLA-B*15:02 alleles (US Food and Drug Administration, 2009). However, conflicting evidence has been found in this study, one SJS patient was reported from negative of HLA-B*15:02 group. This suggested that other genetic and clinical factors may also influence a patient’s risk for adverse reactions.

PGx testing has been implemented in many hospitals throughout the world over the last decade to guarantee that the best treatment is chosen. For example, 27 institutions in the United States have established numerous PGx implementation programmes (Dunnenberger et al., 2015; Volpi et al., 2018), while other PGx implementation programmes in Europe and East Asia include EU-funded Ubiquitous Pharmacogenomics (U-PGx), PREemptive Pharmacogenomic testing for the prevention of Adverse Drug Reactions (PREPARE) (van der Wouden et al., 2017), and the Southeast Asian Pharmacogenomics Research Network (SEAPharm) (RIKEN, 2012; Chumnumwat et al., 2019).

We discovered that the number of HLA-B*15:02 tests per year has increased significantly since 2013 in Thailand. The reason for the high PGx testing could be because of the implementation of a pilot project for HLA-B*15:02 PGx testing before prescribing CBZ in Bangkok by the Department of Medical Sciences, Ministry of Public Health, Thailand. This phenomenon was emphasised later by NHSO announcement for implementation of national screening policy in 2018.

In the HLA-B*15:02 positive group, levetiracetam, sodium valproate, LTG, gabapentin, and phenytoin (PHT) sodium were the most prescribed as alternative AEDs in epilepsy. The reason that levetiracetam and sodium valproate were used is that they are non-aromatic AEDs with a low rate of cross-reactivity with CADRs. In 2014, Li et al. (2015) reported a significant association between HLA-B*15:02 and PHT or LTG-induced SJS/TEN. Nevertheless, our study found that LTG and PHT were prescribed in HLA-B*15:02 positive patients without any ADRs. The reason could be the strength of the association is much weaker than it was for CBZ/OXC-related SJS/TEN. A study from Thailand reported a moderate association (limited evidence) (Locharernkul et al., 2008) and a case report of Han Chinese also showed a weak association (insufficient evidence) (Shi et al., 2012).

Although the retrospective nature of this study is its limitation, this is the first large cohort study in Thailand that provides valuable information on the real clinical impact of HLA-B*15:02 PGx testing for CBZ prescription and associated CADRs in Thai patients.

The selection of appropriate AEDs was influenced by pharmacogenomic testing of HLA-B*15:02 risk alleles. HLA-B genotyping can guide the physician in selecting CBZ/OXC and other AEDs prescriptions rationally. The number of HLA-B*15:02 screening has increased significantly in Thailand. In 85% of cases, physicians waited for HLA-B*15:02 testing to make clinical decisions on CBZ/OXC prescribing. However, physicians have administered CBZ/OXC without PGx testing in the urgent cases. Though there were no CADRs in the HLA-B*15:02 positive group, CBZ-induced SJS/TEN were reported from HLA-B*15:02 carriers who received CBZ without PGx screening before prescription. The presence of mild rash in the HLA-B*15:02 negative group indicates that other genetic biomarker such as HLA-A*31:01 and/or non-genetic variables are involved in CBZ/OXC-induced CADRs, emphasizing that CBZ/OXC prescriptions necessitate CADRs monitoring even in patient without HLA-B*15:02 risk alleles.