Clinical Pharmacogenetics Implementation Consortium (CPIC) Guideline for HLA Genotype and Use of Carbamazepine and Oxcarbazepine: 2017 Update

Abstract

The variant allele HLA-B*15:02 is strongly associated with greater risk of Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) in patients treated with carbamazepine or oxcarbazepine. The variant allele HLA-A*31:01 is associated with greater risk of maculopapular exanthema, drug reaction with eosinophilia and systemic symptoms, and SJS/TEN in patients treated with carbamazepine. We summarize evidence from the published literature supporting these associations and provide recommendations for carbamazepine and oxcarbazepine use based on HLA genotypes.

INTRODUCTION

Human leukocyte antigen (HLA) genetic variation is implicated in the development of specific cutaneous adverse reactions to aromatic anticonvulsants. The purpose of this guideline is to interpret HLA-B*15:02 and HLA-A*31:01 genotyping results to guide the use of carbamazepine and oxcarbazepine. Detailed guidelines regarding the selection of alternative therapies, when to conduct genotype testing, and cost-effectiveness analyses are beyond the scope of this document. Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines are periodically updated at https://cpicpgx.org/guidelines and http://www.pharmgkb.org.

Focused Literature Review AND UPDATE

A systematic literature review focused on HLA-B*15:02 and HLA-A*31:01 genotypes and carbamazepine- and oxcarbazepine-induced cutaneous adverse reactions was conducted (details in Supplemental Material online).

This guideline is an update to the 2013 CPIC guideline for HLA-B*15:02 and carbamazepine use (1). The recommendations provided in the original guideline have not changed and are included here. However, the scope of the existing recommendations has now expanded to include the use of carbamazepine and oxcarbazepine based on HLA-A*31:01 and HLA-B*15:02 genotypes, respectively. Furthermore, the accompanying supplemental material now includes resources to facilitate the incorporation of HLA genotype results into electronic health records with clinical decision support (https://cpicpgx.org/guidelines/guideline-for-carbamazepine-and-hla-b/).

GENES: HLA-B AND HLA-A

Background

HLA-B and HLA-A are part of a large cluster of genes known as the human major histocompatibility complex (MHC). The cluster contains three subgroups: class I, II and III. The HLA-B and HLA-A genes are part of the class I complex, along with HLA-C. These genes encode cell surface proteins that present intracellular antigens to the immune system. Intracellular antigens are usually the normal breakdown products of intracellular proteins and are recognized as “self.” However, if the antigen presented derives from a pathogen or, in some cases, a transplanted tissue, it may be recognized as “non-self” and trigger an immune response. HLA is inherited in a co-dominant fashion with one set of class I and II alleles being inherited from each parent where both have full phenotypic expression.

Because HLA proteins present a wide variety of peptides for immune recognition, the HLA genes are among the most highly polymorphic genes in the human genome. HLA polymorphisms were previously ascertained serologically, but standard molecular approaches that now use DNA sequence-based typing methods either by standard Sanger or next-generation sequencing have revealed much greater complexity of genetic variation within this locus. For example, according to the World Health Organization (WHO) Nomenclature Committee for Factors of the HLA System (http://hla.alleles.org), there are more than 4,000 identified HLA-B alleles and more than 3,000 identified HLA-A alleles, many of which differ by more than one nucleotide from one another. Each allele is designated by the gene name followed by an asterisk and a four- or six-digit identifier giving information about the allele type (designated by the first two digits) and specific protein subtype (second set of digits). The details of HLA nomenclature have been described in a previous CPIC guideline (2).

The guideline presented here specifically discusses the class I HLA alleles HLA-B*15:02 and HLA-A*31:01 as they relate to carbamazepine- and oxcarbazepine-induced cutaneous adverse reactions, including Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN), drug reaction with eosinophilia and systemic symptoms (DRESS), and maculopapular exanthema (MPE).

Genetic Test Interpretation

Clinical genotyping tests exist for identifying HLA-B and HLA-A alleles, including HLA-B*15:02 and HLA-A*31:01. Genotyping results are presented as “positive” if one or two copies of the variant allele are present or “negative” if no copies of the variant allele are present. There is no intermediate genotype. Genotype definitions for HLA-B*15:02 and HLA-A*31:01 are summarized in Table 1. Nucleotide and amino acid sequence alignments for HLA-B*15:02 and HLA-A*31:01 and the corresponding reference sequences are available in Supplemental Figures S1-S4.

TABLE 1.

ASSIGNMENT OF HLA-B AND HLA-A GENOTYPES

Genotype	Definition	Examples of Diplotypes
HLA-B
HLA-B*15:02 negative	Homozygous for an allele other than HLA-B*15:02	*Xa/*Xa
HLA-B*15:02 positive	Heterozygous or homozygous variant	*15:02/*Xa, *15:02/*15:02
HLA-A
HLA-A*31:01 negative	Homozygous for an allele other than HLA-A*31:01	*Yb/*Yb
HLA-A*31:01 positive	Heterozygous or homozygous variant	*31:01/*Yb, *31:01/*31:01

^aWhere *X = any HLA-B allele other than HLA-B*15:02

^bWhere *Y = any HLA-A allele other than HLA-A*31:01

Available Genetic Test Options

Commercially available genetic testing options change over time. Additional information about pharmacogenetic testing can be found at the Genetic Testing Registry website (http://www.ncbi.nlm.nih.gov/gtr/).

Incidental Findings

Although HLA alleles have been studied in the context of specific responses to human immunodeficiency virus and other pathogens, there are currently no specific diseases or conditions that have been strongly linked to HLA-B*15:02 or HLA-A*31:01 independent of drug use (3–5). However, HLA-B*15:02 has also been associated with SJS/TEN from phenytoin use, and other HLA-B alleles have been strongly associated with adverse drug reactions. For example, HLA-B*57:01 is associated with abacavir-induced hypersensitivity reaction, and HLA-B*58:01 is associated with allopurinol-induced severe cutaneous adverse reactions (including SJS/TEN and DRESS). CPIC guidelines are available to guide prescribing of phenytoin (6), abacavir (7), and allopurinol (8) based on HLA-B genotype.

Other Considerations

HLA-B*15:02 and HLA-A*31:01 have distinct ethnic and geographical distributions that are important for evaluating population risk (see HLA-A and HLA-B Allele Frequency Table online). The frequency of HLA-B*15:02 is highest in East Asian (6.9%), Oceanian (5.4%) and South/Central Asian (4.6%) populations. However, not all East Asian sub-populations carry this allele in such high frequencies. HLA-B*15:02 frequency is much lower in Japanese (<1%) and Korean (<2.5%) populations. The allele is also quite rare in African populations (not observed), African Americans, Middle Easterners, Caucasians and Hispanics/South Americans (<1%). In contrast, the frequency of the HLA-A*31:01 allele is higher than the HLA-B*15:02 allele in Caucasians (3%) and Hispanic/South Americans (6%). However, it is also found in high frequencies in some East Asians, specifically Japanese (8%) and South Koreans (5%), and South/Central Asians (2%). While these frequencies are helpful in determining broad population risks, they cannot replace genotypes on an individual basis.

DRUGS: CARBAMAZEPINE AND OXCARBAZEPINE

Background

Carbamazepine

Carbamazepine, an aromatic anticonvulsant related to the tricyclic antidepressants, is U.S. Food and Drug Administration (FDA) approved for the treatment of epilepsy, trigeminal neuralgia, and bipolar disorder. Carbamazepine reduces the propagation of abnormal impulses in the brain by producing a frequency- and voltage-dependent block of sodium channels, thereby inhibiting the generation of repetitive action potentials in the epileptic focus (9). Carbamazepine-induced adverse effects that may have known dose- or concentration- dependency include dizziness, ataxia and nystagmus. Other adverse effects such as aplastic anemia, hyponatremia, leucopenia, osteoporosis, liver injury and hypersensitivity reactions such as MPE, DRESS, SJS/TEN have a complex dose-response relationship such that it is difficult to delineate a clear linear dose-response relationship. For additional information regarding the pharmacokinetics and pharmacogenomics of carbamazepine, please refer to the PharmGKB website: http://www.pharmgkb.org/pathway/PA165817070 (10).

Oxcarbazepine

Oxcarbazepine is the keto-analog of carbamazepine. With its similar structure, oxcarbazepine shares many therapeutic indications and adverse effects with carbamazepine. Furthermore, patients who have had hypersensitivity reactions to carbamazepine may also be predisposed to hypersensitivity reactions with oxcarbazepine; these patients should only be treated with oxcarbazepine if the potential benefit justifies the potential risk.

Linking Genetic Variability to Variability in Drug-Related Phenotypes

There is evidence linking HLA-B*15:02 genotype with the risk of carbamazepine- and oxcarbazepine-induced SJS/TEN (Supplemental Table S1) and linking HLA-A*31:01 genotype with the risk of carbamazepine-induced SJS/TEN, DRESS, and MPE (Supplemental Table S2). Application of a grading system to evidence linking HLA genotypic variations to phenotypic variability with respect to cutaneous adverse reactions indicates a high quality of evidence in the majority of cases. This body of evidence provides the basis for the recommendations in Table 2 and Table 3.

TABLE 2.

RECOMMENDATIONS FOR CARBAMAZEPINE THERAPY BASED ON HLA-B AND HLA-A GENOTYPES

Genotypea	Implication	Therapeutic recommendation	Classification of recommendation	Considerations for other aromatic anticonvulsants
HLA-B*15:02 negative and HLA-A*31:01 negative	Normal risk of carbamazepine-induced SJS/TEN, DRESS, and MPE	Use carbamazepine per standard dosing guidelines.b	Strong	N/A
HLA-B*15:02 negative and HLA-A*31:01 positive	Greater risk of carbamazepine-induced SJS/TEN, DRESS, and MPE	If patient is carbamazepine-naïve and alternative agents are available, do not use carbamazepine.	Strong	Other aromatic anticonvulsantsd have very limited evidence, if any, linking SJS/TEN, DRESS, and/or MPE with the HLA-A*31:01 allele, and thus no recommendation can be made with respect to choosing another aromatic anticonvulsant as an alternative agent.
		If patient is carbamazepine-naïve and alternative agents are not available, consider the use of carbamazepine with increased frequency of clinical monitoring. Discontinue therapy at first evidence of a cutaneous adverse reaction.	Optional	N/A
		The latency period for cutaneous adverse drug reactions is variable depending on phenotype; however, all usually occur within three months of regular dosing. Therefore, if the patient has previously used carbamazepine consistently for longer than three months without incidence of cutaneous adverse reactions, cautiously consider use of carbamazepine.	Optional	Previous tolerance of carbamazepine is not indicative of tolerance to other aromatic anticonvulsants.d
HLA-B*15:02 positivec and any HLA-A*31:01 genotype (or HLA-A*31:01 genotype unknown)	Greater risk of carbamazepine-induced SJS/TEN	If patient is carbamazepine-naïve, do not use carbamazepine.	Strong	Other aromatic anticonvulsantsd have weaker evidence linking SJS/TEN with the HLA-B*15:02 allele; however, caution should still be used in choosing an alternative agent.
		The latency period for drug-induced SJS/TEN is short with continuous dosing and adherence to therapy (~4–28 days), and cases usually occur within three months of dosing; therefore, if the patient has previously used carbamazepine consistently for longer than three months without incidence of cutaneous adverse reactions, cautiously consider use of carbamazepine in the future.	Optional	Previous tolerance of carbamazepine is not indicative of tolerance to other aromatic anticonvulsants.d

DRESS = drug reaction with eosinophilia and systemic symptoms; MPE = maculopapular exanthema; N/A = not applicable; SJS = Stevens-Johnson syndrome; TEN = toxic epidermal necrolysis

^aIf only HLA-B*15:02 was tested, assume HLA-A*31:01 is negative and vice versa.

^bHLA-B*15:02 has a 100% negative predictive value for carbamazepine-induced SJS/TEN, and its use is currently recommended to guide use of carbamazepine and oxcarbazepine only. Because there is a much weaker association and less than 100% negative predictive value of HLA-B*15:02 for SJS/TEN associated with other aromatic anticonvulsants, using these drugs instead of carbamazepine or oxcarbazepine in the setting of a negative HLA-B*15:02 test in Southeast Asians will not result in prevention of anticonvulsant-associated SJS/TEN (40).

^cIn addition to HLA-B*15:02, risk for carbamazepine-induced SJS/TEN has been reported in association with the most common B75 serotype alleles in Southeast Asia, HLA-B*15:08, HLA-B*15:11, and HLA-B*15:21. Although not described, the possibility of carbamazepine-induced SJS/TEN in association with less frequently carried B75 serotype alleles, such as HLA-B*15:30 and HLA-B*15:31, should also be considered.

^dAromatic anticonvulsants include carbamazepine, oxcarbazepine, eslicarbazepine, lamotrigine, phenytoin, fosphenytoin, and phenobarbital.

TABLE 3.

RECOMMENDATIONS FOR OXCARBAZEPINE THERAPY BASED ON HLA-B GENOTYPE

Genotype	Implication	Therapeutic recommendation	Classification of recommendation	Considerations for other aromatic anticonvulsants
HLA-B*15:02 negative	Normal risk of oxcarbazepine-induced SJS/TEN	Use oxcarbazepine per standard dosing guidelines.	Strong	N/A
HLA-B*15:02 positive	Greater risk of oxcarbazepine-induced SJS/TEN	If patient is oxcarbazepine-naïve, do not use oxcarbazepine.	Strong	Other aromatic anticonvulsantsa have weaker evidence linking SJS/TEN with the HLA-B*15:02 allele; however, caution should still be used in choosing an alternative agent.
		The latency period for drug-induced SJS/TEN is short with continuous dosing and adherence to therapy (~4–28 days), and cases usually occur within three months of dosing; therefore, if the patient has previously used oxcarbazepine consistently for longer than three months without incidence of cutaneous adverse reactions, cautiously consider use of oxcarbazepine in the future.	Optional	Previous tolerance of oxcarbazepine is not indicative of tolerance to other aromatic anticonvulsants.a

N/A = not applicable; SJS = Stevens-Johnson syndrome; TEN = toxic epidermal necrolysis

^aAromatic anticonvulsants include carbamazepine, oxcarbazepine, eslicarbazepine, lamotrigine, phenytoin, fosphenytoin, and phenobarbital.

HLA-B*15:02

HLA-B*15:02 is specific for carbamazepine- and oxcarbazepine-induced SJS and TEN, though the data are strongest for carbamazepine. SJS is characterized by epidermal detachment affecting up to 10% of the body surface area (BSA) while TEN usually involves more than 30% of the BSA. Patients with between 10-30% of the BSA blistered are defined as having an SJS/TEN overlap syndrome. Mortality rates are typically below 5% for SJS and can be above 30% for TEN, with sepsis being the most frequent cause of death (11). Mortality from SJS/TEN is also related to age, the drug half-life and how early the drug is discontinued (12, 13). An immune-mediated etiology has been shown for these reactions, which is consistent with the anamnestic response often seen clinically on drug re-challenge (14). In terms of the immunopathology, cytotoxic T cells, or CD8+ T cells (lymphocytes matured in the thymus that express the CD8 protein on their surface) are involved in SJS and TEN (15, 16). Further discussion on the mechanism of carbamazepine-induced SJS/TEN is presented in the Supplemental Material.

Consistent with the regional and ethnic distribution of the HLA-B*15:02 allele, studies have shown the genetic risk of carbamazepine-associated SJS/TEN to be higher in several Asian countries with increased frequency of the HLA-B*15:02 allele, including Vietnam (17), Cambodia (17), Reunion Islands (17), Thailand (18, 19), some parts of India (20), Malaysia (21) and Hong Kong (22). The HLA-B*15:02 allele has not been observed in cases of SJS/TEN in some ancestral groups, such as Japanese and Korean populations or non-Asian descendants in Europe or North America (17, 23–26), where the frequency of the allele is very low. In the Han Chinese population, the sensitivity of HLA-B*15:02 as a predictive test for SJS/TEN has been estimated at 98% and specificity at 97%; the positive predictive value is estimated at 7.7% and negative predictive value at 100% (27). However, it is important to note that in one study, in a group of individuals thought to be of European origin, four of 12 individuals with SJS/TEN carried the HLA-B*15:02 allele (24). Subsequently, they were found to have some Southeast Asian ancestry. This example underscores the importance of considering the HLA-B*15:02 allele carrier status in therapeutic decision making regardless of self-reported ethnicity.

Based on the strong evidence linking HLA-B*15:02 to carbamazepine-induced SJS/TEN, the FDA issued a Health Alert in 2007 about changes to package labeling and recommendations for genetic testing in patients treated with carbamazepine (28). The FDA label for carbamazepine carries a boxed warning about the risk of SJS/TEN with the presence of the HLA-B*15:02 allele and states that patients testing positive for the allele should not be treated with carbamazepine unless the benefit clearly outweighs the risk. The FDA label for oxcarbazepine does not carry this boxed warning, but there is mention of the association between HLA-B*15:02 and the risk of SJS/TEN in the warnings and precautions section that advises avoiding oxcarbazepine in HLA-B*15:02 positive patients unless the benefit clearly outweighs the risk. The positive predictive value of HLA-B*15:02 for oxcarbazepine-induced SJS/TEN is estimated to be 0.73%, which is much lower than that of carbamazepine-induced SJS/TEN (7.7%); however, the negative predictive value for both nears 100% in Southeast Asian populations (29).

HLA-A*31:01

Unlike HLA-B*15:02, the HLA-A*31:01 allele is associated with a wider range of carbamazepine hypersensitivity reactions, including MPE, DRESS, and SJS/TEN, in many different populations (30). DRESS is a severe hypersensitivity reaction characterized by generalized cutaneous eruptions with systemic manifestations that can be life-threatening, whereas MPE is a milder reaction with only the presence of rash without mucosal or organ involvement, or systemic features. Available evidence suggests an association between the presence of HLA-A*31:01 and carbamazepine-induced MPE, DRESS, and SJS/TEN, with the data strongest for DRESS and SJS/TEN in European and Japanese populations, where the allele frequency is higher; however, no such evidence exists for oxcarbazepine.

In Southeast Asian populations, the strong association between HLA-B*15:02 and carbamazepine-induced SJS/TEN would overwhelm any potential association between HLA-A*31:01 and carbamazepine-induced SJS/TEN. In European, African, and Japanese populations where the carriage rate of HLA-B*15:02 is less than 1%, HLA-A*31:01 appears to be the primary driver of carbamazepine-induced SJS/TEN and other hypersensitivity reactions. HLA-A*31:01 is also a risk factor for MPE and DRESS in Han Chinese populations. The positive predictive value and number needed to test to prevent one case of all carbamazepine-induced hypersensitivity reactions (most influenced by MPE ⋙ DRESS) combined are most favorable for European populations, and they are estimated at 43% and 47, respectively (31). Limited, if any, evidence exists to support an association between HLA-A*31:01 and hypersensitivity associated with other aromatic anticonvulsants including lamotrigine (32), oxcarbazepine, eslicarbazepine, phenytoin, fosphenytoin, and phenobarbital, and thus no recommendations can be given regarding the safety of these agents in HLA-A*31:01 positive patients. In light of evidence supporting clinical cross-reactivity among aromatic anticonvulsants, however, in the instance where a severe hypersensitivity reaction has occurred with one agent, avoidance of the others is recommended (33).

Therapeutic Recommendations

The therapeutic recommendations for HLA-B*15:02 and carbamazepine remain unchanged from the original guideline (1), but in this update they are now also applicable to oxcarbazepine (Tables 2 and 3). These recommendations hold irrespective of the patient’s region of origin or ethnic group. For patients who are HLA-B*15:02 negative, carbamazepine or oxcarbazepine may be prescribed per standard guidelines. If a patient is carbamazepine-naïve or oxcarbazepine-naïve and HLA-B*15:02 positive, carbamazepine and oxcarbazepine should be avoided, respectively, due to the greater risk of SJS/TEN. Other aromatic anticonvulsants, including eslicarbazepine, lamotrigine, phenytoin, fosphenytoin, and phenobarbital, have very limited evidence, if any, linking SJS/TEN with the HLA-B*15:02 allele; however, caution should still be used when choosing an alternative agent. With regular dosing, carbamazepine- or oxcarbazepine- induced SJS/TEN usually develops within the first 4-28 days of therapy; therefore, patients who have been continuously taking carbamazepine or oxcarbazepine for longer than three months without developing cutaneous reactions are at extremely low risk (but not zero) of carbamazepine- or oxcarbazepine-induced adverse events in the future, regardless of HLA-B*15:02 status (34, 35).

For patients who are HLA-A*31:01 negative, carbamazepine may be prescribed per standard guidelines (Table 2). If a carbamazepine-naïve patient also received testing for HLA-B*15:02 and is positive for this allele, carbamazepine should be avoided regardless of the HLA-A*31:01 genotype result. If a patient is carbamazepine-naïve and HLA-A*31:01 positive, and if alternative agents are available, carbamazepine should be avoided due to the greater risk of SJS/TEN, DRESS, and MPE. Other aromatic anticonvulsants, including oxcarbazepine, have very limited evidence, if any, linking SJS/TEN, DRESS, and/or MPE with the HLA-A*31:01 allele, and thus no recommendation can be made with respect to choosing another aromatic anticonvulsant as an alternative agent. If alternative agents are not available, consider the use of carbamazepine with increased frequency of clinical monitoring. Discontinue therapy at first evidence of a cutaneous adverse reaction. As previously mentioned, since the latency period for cutaneous adverse drug reactions is known, if the patient is HLA-A*31:01 positive and has previously used carbamazepine for longer than three months without incidence of a cutaneous adverse reaction, cautiously consider use of carbamazepine.

Pediatrics

Data describing the relationship between HLA-B*15:02 and HLA-A*31:01 genotype and carbamazepine- or oxcarbazepine-induced cutaneous adverse reactions in pediatric patients are scarce (Supplemental Tables S1 and S2). In the absence of data suggesting a different relationship between these HLA alleles and drug-induced hypersensitivity in pediatric patients, the recommendations may be used to guide use of carbamazepine and oxcarbazepine in both adult and pediatric patients.

Recommendations for Incidental Findings

Aromatic anticonvulsants that are structurally similar to carbamazepine have also been associated with SJS/TEN and HLA-B*15:02. The drug-specific evidence linking HLA-B*15:02 and SJS/TEN is discussed in the Supplemental Material and may have implications for choosing alternatives to carbamazepine in those who carry the HLA-B*15:02 allele.

Other Considerations

HLA-B75 serotypes

HLA-B*15:02 is the most common HLA-B75 serotype allele in Southeast Asia. Other less frequently carried members of the HLA-B75 serotype include HLA-B*15:08, HLA-B*15:11, and HLA-B*15:21. The HLA proteins coded by these alleles share structural similarity and peptide binding grooves, and hence peptide binding specificities, with HLA-B*15:02 and have also been reported in association with carbamazepine-induced SJS/TEN (26, 36–38). Currently the majority of available data focuses on the risk of carbamazepine-induced SJS/TEN conferred by the presence of HLA-B*15:02 and is the basis for the design of efficient single allele molecular typing assays. However, some labs may provide high-resolution HLA-B typing and the possibility of carbamazepine-induced SJS/TEN with HLA-B*15:08, HLA-B*15:11, HLA-B*15:21, and even less common HLA-B75 serotype alleles such as HLA-B*15:30 and HLA-B*15:31 where carbamazepine-induced SJS/TEN has yet to be described, needs to be considered a potential risk if this information is available.

Implementation of this guideline

The guideline supplement and CPIC website (https://cpicpgx.org/guidelines/guideline-for-carbamazepine-and-hla-b/) contains resources that can be used within electronic health records (EHRs) to assist clinicians in applying genetic information to patient care for the purpose of drug therapy optimization (see Resources to incorporate pharmacogenetics into an electronic health record with clinical decision support in the Supplemental Material).

POTENTIAL BENEFITS AND RISKS FOR THE PATIENT

A potential benefit of HLA-B*15:02 and HLA-A*31:01 testing is a reduction in the incidence of serious, and sometimes fatal, cutaneous adverse reactions to carbamazepine and oxcarbazepine by identifying those who are at significant risk and using alternative therapy. The success of HLA-B*15:02 prospective screening in reducing the rate of SJS/TEN has been demonstrated clinically in a Chinese population (39).

A potential risk of HLA-B*15:02 or HLA-A*31:01 testing is ruling out the use of carbamazepine or oxcarbazepine in patients who may not ever develop a hypersensitivity reaction to the drug. This risk is mitigated by the fact that there are often alternatives to carbamazepine or oxcarbazepine with comparable effectiveness; however, consideration must be given to the risk of cutaneous adverse reactions with other anticonvulsants. For example, it has been demonstrated in an Asian population that a HLA-B*15:02 screening policy for carbamazepine will not decrease the overall rate of SJS/TEN if other anticonvulsants associated with SJS/TEN (e.g., phenytoin) are used instead of carbamazepine (40). The risk of phenytoin-associated SJS/TEN is described in more detail in the CPIC guideline for CYP2C9 and HLA-B genotypes and phenytoin dosing (6). Furthermore, other anticonvulsants may be associated with more unfavorable adverse effect profiles compared to carbamazepine or oxcarbazepine.

Although genotyping is considered reliable when performed in qualified clinical laboratories, laboratory error and sample mix-up is always a distinct possibility. If a HLA-B*15:02 negative, Southeast Asian individual who does not carry another B75 serotype of HLA develops carbamazepine-induced SJS/TEN, for instance, the HLA typing should be repeated to rule out sample or typing error. Genotype results are associated with a patient for a lifetime; as such, a genotyping error could have a broader impact on health care should other HLA-B*15:02 or HLA-A*31:01 associations be identified in the future.

CAVEATS: APPROPRIATE USE AND/OR POTENTIAL MISUSE OF GENETIC TESTS

If a patient has taken carbamazepine or oxcarbazepine consistently for more than three months, it is highly unlikely that a severe cutaneous adverse reaction will occur after that time. As a result, known HLA-B*15:02 or HLA-A*31:01 genotypes will be less helpful for treatment-experienced patients compared to treatment-naïve patients. Furthermore, because extensive ethnic admixture has occurred globally and not all carbamazepine- and oxcarbazepine-induced cutaneous adverse reactions can be attributed to HLA-B*15:02 or HLA-A*31:01, clinicians should carefully monitor all patients as standard practice.