Association between HLA-B*1502 Allele and Antiepileptic Drug-Induced Cutaneous Reactions in Han Chinese



This article is corrected by:

  1. Errata: ERRATUM Volume 49, Issue 5, 941, Article first published online: 1 May 2008

Address correspondence and reprint requests to Dr. Patrick Kwan, Department of Medicine and Therapeutics Prince of Wales Hospital, Hong Kong, China. E-mail:


Summary:  A previous study conducted in Taiwan found a 100% association between HLA-B*1502 allele and carbamazepine-induced Stevens-Johnson syndrome (SJS) in Han Chinese subjects, with an extremely high odds ratio compared with carbamazepine-tolerant subjects (odds ratio = 2,504). We examined this association in 24 Hong Kong Han Chinese subjects who had cutaneous adverse reactions induced by different antiepileptic drugs (AEDs). They were matched with 48 AED-tolerant controls. HLA-B*1502 was associated with severe cutaneous reactions (SCR) induced by AEDs, which included carbamazepine, phenytoin, and lamotrigine (p = 0.001, odds ratio = 17.6), but was not associated with maculopapular exanthema (MPE) (p = 0.32). Further studies in larger samples of ethnically matched subjects should be conducted to confirm the findings. Identification of genetic polymorphisms predisposing to development of AED-induced SCR offers the possibility of avoiding these high-risk drugs in genetically susceptible individuals.

Antiepileptic drugs (AEDs) are one of the commonest causes of cutaneous adverse drug reactions (cADRs). The manifestation of cADRs ranges from a mild maculo-papular exanthema (MPE) to life-threatening severe cutaneous reaction (SCR), which includes Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), and drug hypersensitivity syndrome (HSS). The incidence of AED-induced cADRs varies from drug to drug. Overall, isolated rash related to AEDs occurs in up to 16% of patients (Chadwick et al., 1984), while the incidence of SCR is estimated to be 1:10,000 exposures. More than 90% of SCR occurs in the first 2 months of AED use (Tennis and Stern, 1997). AEDs associated with increased risk of SCR include carbamazepine, phenobarbital, phenytoin, and lamotrigine (Tennis and Stern, 1997). cADRs represent a significant burden on the health care system, and SCR mortality reaches up to 30% (Svensson et al., 2000). Identification of genetic polymorphisms predisposing to development of AED-induced SCR offers the possibility of avoiding these high-risk drugs in genetically susceptible individuals.

Recent studies suggested a strong association between HLA-B*1502 and carbamazepine-induced SJS in subjects of Chinese/Asian ethnicity (Chung et al., 2004; Hung et al., 2006; Lonjou et al., 2006). Chung et al. (2004) genotyped 44 Han Chinese patients living in Taiwan with carbamazepine-induced SJS (CBZ–SJS), 101 CBZ-tolerant patients, and 93 healthy controls without a history of CBZ use. HLA-B*1502 allele was present in 100% of CBZ–SJS patients but in only 3% of CBZ-tolerant patients and in 9% of the control population (CBZ–SJS vs. CBZ-tolerant: odds ratio [OR] 2,504, 95% confidence interval [CI] 126–49,522). In a follow-up study by the same group of investigators with an overlapping cohort, HLA-B*1502 was found to be strongly associated with CBZ–SJS/TEN, but not with MPE or HSS (Hung et al., 2006). In this study, we aimed to examine the association between HLA-B*1502 and cADRs induced by CBZ as well as other AEDs in another Han Chinese cohort.


This case–control study was approved by the regional hospital ethics committee. Written informed consent was obtained from all participants. All subjects were actively followed up at the Prince of Wales Hospital. They were identified from the hospital's Central Management System, which is a centralized computer database in place since 1995 linking all public hospitals in Hong Kong and containing the outpatient and inpatient notes, information on medications dispensed, history of drug allergies, and investigation results of all patients attending public hospitals. Cases were patients who developed cutaneous adverse reactions within 8 weeks after commencing an AED and for which no other causes were found. MPE was defined as skin eruption with no mucosal or systemic involvement. SCR included SJS, TEN and HSS. Their definitions were the same as in previous studies (Hung et al., 2006). SJS was defined as skin detachment of 10% of body-surface area or less, TEN was defined by skin detachment of 30% or more, whereas intermediate extent of skin detachment corresponded to SJS/TEN overlap. The criteria for HSS were skin rash, plus two of the following features: fever, lymphadenopathy, and hematological abnormalities (e.g., eosinophilia, atypical lymphocytosis) with involvement of at least one internal organ (e.g., hepatitis, pneumonitis, myocarditis, pericarditis, nephritis). Controls were patients without history of drug-induced cutaneous reactions. Cases were matched in a 1:2 ratio with controls in age and AED prescribed (taken for at least 3 months). All cases and controls were ethnic Han Chinese.

Polymerase chain reaction (PCR) using sequence specific primers was employed to genotype the patients, specifically for determination of HLA-B*1502 (Bunce et al., 1995). To genotype HLA-B*1502, 4 separated multiplex PCRs were performed using 4 sets of HLA-specific primers (Set 1: Forward primer [F1]–5′-CGA GAG AGC CTG CGG AC-3′, reverse primer [R1]–5′-GCC CAC TTC TGG AAG GTT CT-3′; Set 2: F2 –5′-CGC GAG TCC GAG GAT GGC-3, R2 –5′-GCA GGT TCC GCA GGC TCT-3′; Set 3: F3 –5′-ACC GGA ACA CAC AGA TCT G, R3 –5′-GAG CCA CTC CAC GCA CAG; Set 4: F4 –5′-GGA GTA TTG GGA CCG GAAC-3′, R4 –5′-GCC ATA CAT CCT CTG GAT GA-3′) and 2 sets of control primers (Set C1: FC1 –5′-TGC CAA GTG GAG CAC CCA A-3′, RC1 –5′-GCA TCT TGC TCT GTG CAG AT-3′; Set C2: FC2 –5′-ATG ATG TTG ACC TTT CCA GGG-3′, RC2 –5′-TTC TGT AAC TTT TCA TCA GTT GC-3′). Set C1 primers were used as internal PCR control in 3 multiplex PCRs using Set 2, 3 or 4 primers and Set C2 with Set 1 primers. Presence of all 4 HLA-specific PCR products (1,340 base pairs (bp) for Set 1, 124 bp for Set 2, 562 bp for Set 3, and 369 bp for Set 4) along with internal control PCR product (796 bp for Set C1 and 256 bp for Set C2) denotes HLA-B*1502.

Data are expressed as positive or negative for HLA-B*1502. Chi-square test and Fisher's exact test were used to analyze the association between AED-induced cADRs and HLA-B*1502 status. In order to reduce bias in estimating the odds ratio, whenever a zero-count cell is encountered, 0.5 was added to all cells in a 2 × 2 table (Jewell et al., 1986). P-values ≤ 0.05 (two-sided) were considered statistically significant.


HLA-B genotypes were determined in 24 patients who had AED-induced cADRs (8 with SCR, 16 with MPE) and 48 AED-tolerant controls. In the SCR group, 4 were male and 4 were female; their mean age was 36 years. In the MPE group, 9 were male and 7 were female; their mean age was 40 years. In the control group, 30 were male and 18 were female; their mean age was 40 years. Table 1 shows the AEDs taken by the various patient groups.

Table 1. Number of subjects in each patient group taking different antiepileptic drugs
  1. CBZ, carbamazepine; LTG, lamotrigine; MPE, maculopapular exanthema; PB, phenobarbital; PHT, phenytoin; SCR, severe cutaneous reaction; TPM, topiramate; VPA, valproate.

CBZ 4 416
LTG 2 411
PB 1 0 2
PHT 1 513
TPM 1 2
VPA 2 4
Total 81648

The HLA-B*1502 status and other characteristics of patients with SCR are shown in Table 2. Significantly more patients with AED-induced SCR had the HLA-B*1502 allele compared with controls (75% vs. 14.5%, p=0.001, OR 17.6, 95% CI 2.9–105.2). HLA-B*1502 was found in all patients with AED-induced SJS/TEN and in none with HSS. Thus the association with HLA-B*1502 allele was highly significant when the analysis was confined to comparing patients with AED-induced SJS/TEN alone with controls (100% vs. 14.5%, p = 1.48 × 10−4; OR 71.9, 95% CI 3.7–1,415.8). There was no difference in the frequency of subjects with the HLA-B*1502 allele between the MPE (12.5%) and control groups (14.5%; p = 0.32, OR 0.84, 95% CI 0.15–4.51). Among patients with SCR, four were induced by CBZ; of those four, all were positive for HLA-B*1502. In the other two patients who were positive for HLA-B*1502, SCR was induced by lamotrigine and phenytoin, respectively. In the former patient, lamotrigine was added to background valproate treatment.

Table 2. Characteristics of patients with severe cutaneous reactions
PatientSexAgea (years)Associating AEDbType of SCRHLA-B* 1502 allele
  1. AED, antiepileptic drug; CBZ, carbamazepine; HSS, drug hypersensitivity syndrome; LTG, lamotrigine; PB, phenobarbital; PHT, phenytoin; SCR, severe cutaneous reactions; SJS, Stevens-Johnson syndrome; TEN, toxic epidermal necrolysis.

  2. a Age at development of SCR.

  3. b An associating AED is one that was commenced within 8 weeks prior to the development of SCR and no other causes were found for the SCR.

1F 28 CBZSJS Positive
2M 23 CBZSJS Positive
3F 53 CBZSJS Positive
4M 10 CBZTEN Positive
5F 53 PHTSJS Positive
6F 41 LTGTEN Positive
7M 14 PBHSS Negative
8M 44 LTGHSS Negative


We found a strong association between HLA-B*1502 and AED-induced SCR. Consistent with the original Taiwanese report, all patients with CBZ-induced SJS/TEN in our study were positive for HLA-B*1502 (Chung et al., 2004). As the odds ratio for this allele was extremely high in the original (Chung et al., 2004) and follow-up (Hung et al., 2006) Taiwanese studies (2,504 [95% CI 126 to 49,522] and 1,357 [95% CI 193–8,838], respectively), our study was of sufficient size to detect such a strong effect. Similar to the follow-up study by the Taiwan group, HLA-B*1502 was not found in the patients with HSS, nor was it associated with MPE. SJS and TEN are cADRs characterized by bullous lesions, while MPE and HSS are regarded as nonbullous reactions. These findings lend further support to the notion that genetic association with CBZ-induced cADRs is phenotype specific (Hung et al., 2006).

Current data also support the hypothesis that genetic susceptibility to CBZ-induced SJS/TEN might be ethnicity specific. In a European study, Lonjou et al. (2006) studied 12 CBZ-induced SJS/TEN patients. HLA-B*1502 was found in only 4 patients, all of whom had a parent of Asian origin. Compared with the Chinese population (8.6%) (Chung et al., 2004), HLA*B-1502 is rare in people of European ancestry (1–2%) (Lonjou et al., 2006). It could be postulated that failure to detect an association between HLA-B*1502 and CBZ–SJS/TEN in the latter might be explained by differences in linkage disequilibrium patterns between the ethnic groups. However, fine recombinant genetic mapping among Taiwan Han Chinese on the extended HLA-B*1502 haplotype narrowed down the susceptibility region for CBZ–SJS/TEN to within 86 kb flanking the HLA-B gene, and within this region HLA-B is the only gene present, suggesting that B*1502 is the susceptibility haplotype (Hung et al., 2006).

The cascade of immune mechanisms underlying cADRs remains unclear. Infiltrating CD4+ T cells are present in the skin lesions of MPE and HSS, whereas CD8+ T-cell mediated cytotoxic responses appear to be the major event in SJS/TEN (Naisbitt et al., 2003; Nassif et al., 2004). It has been suggested that HLA-B allele may illicit immune response by presenting peptides bound to the drug and/or its metabolites to specific T cells (Hung et al., 2006), resulting in proliferation of the cytotoxic cells (Wu et al., 2006).

In addition to CBZ, we also found HLA-B*1502 allele in patients with lamotrigine and phenytoin-induced SJS/TEN. Whether this implies a similar immune response against the different AEDs will need to be confirmed in larger studies. Cross-reactivity among the aromatic AEDs (carbamazepine, phenytoin, phenobarbital) in inducing cADRs is recognized, but has not been observed between aromatic AEDs and lamotrigine (Arroyo and de la Morena, 2001). The mechanism of cross-reactivity is thought to be due to accumulation of toxic hydroxylated aromatic metabolites from the aromatic AEDs. Interestingly, in the patient developing TEN associated with lamotrigine and had the HLA-B*1502 allele, lamotrigine was added to background valproate treatment. It is well documented that the risk of SCR associated with lamotrigine is higher with concurrent valproate therapy (Guberman et al., 1999).

The current study is limited by its small sample size, resulting in a wide confidence interval. Nonetheless, further research in Chinese and other ethnic cohorts to test the association, particularly with AEDs other than CBZ, and to identify other HLA alleles predisposing to AED-related SCR is warranted. Knowledge of this association might be used to help prevent life-threatening SCR by screening patients, particularly those of Chinese ethnicity, for HLA-B*1502 before prescribing these high-risk AEDs.