Skip to main content
Premium Trial:

Request an Annual Quote

NEJM Studies Find New Carbamazepine Response Marker, Support Value of Genotyping in Han Chinese


By Valerie Ross

Two studies appearing this week in the New England Journal of Medicine have shed further light on the role of genetic variation in serious reactions to the seizure drug carbamazepine.

One study, led by researchers Taiwan's Academia Sinica, showed that screening Han Chinese populations for a previously identified allele, HLA-B*1502, can reduce incidence of dangerous reactions; while the other, led by researchers at the University of Liverpool, described a new allele, HLA-A*3101, that predicts risk of carbamazepine-induced reactions in a European population.

Based on earlier pharmacogenomics studies, the US Food and Drug Administration in 2007 updated the label for carbamazepine to recommend that patients of Asian ancestry taking the drug be tested for the HLA-B*1502 marker (PGx Reporter 12/19/07). While the first NEJM study supports this practice, the second suggests that the boxed warning be expanded to include the new allele for the European population.

The authors of the second study "propose that HLA-A*3101 is clinically relevant as a marker to predict hypersensitivity reactions" and recommend that "consideration be given to adding the association with HLA-A*3101 to the drug label for carbamazepine."

It is still too early to tell what impact these findings may have on the drug's label, however. An FDA spokesperson told Pharmacogenomics Reporter via e-mail that the agency "will review the matter."

Carbamazepine — used to treat epilepsy, bipolar disorder, and trigeminal neuralgia — can cause a number of hypersensitivity reactions in a small proportion of patients. The most severe of these are Stevens-Johnson Syndrome and toxic epidermal necrolysis, which cause dangerous and sometimes fatal skin and internal organ conditions. Together, the two syndromes are often denoted with the shared abbreviation SJS-TEN.

In 2004, researchers at Academia Sinica found an association between the genetic marker HLA-B*1502 and SJS-TEN in patients of Han Chinese ancestry. Nor is carbamazepine the only drug for which a particular HLA allele is associated with hypersensitivity reactions — HLA-B*5701 is associated with increased risk of such reactions in patients taking the HIV drug abacavir, and the FDA updated abacavir's label in 2008 to recommend testing for the allele.

The same researchers who found the association between HLA-B*1502 and SJS-TEN reported this week in NEJM that screening patients for the allele and treating them accordingly can reduce carbamazepine-induced SJS-TEN in patients of Han Chinese ancestry.

At 23 hospitals across Taiwan, the researchers obtained genetic material from 4,855 patients, none of whom had previously taken carbamazepine but all of whom, under normal circumstances, would have been prescribed the drug at the time of screening. Using DNA isolated from a whole blood sample, the research team determined whether each of those subjects had the HLA-B*1502 allele using Pharmigene's PG1502 DNA detection kit.

Of the subjects tested, 372 of them (7.7 percent) had the HLA-B*1502 allele and were advised to take an alternative medication or the medication they were on before the study began, while 4,120 patients who tested negative for the allele took carbamazepine and were monitored over the course of the two-month study.

Using the data from three earlier studies, the researchers calculated a historical incidence of SJS-TEN among Han Chinese patients taking carbamazepine of 0.23 percent, which would mean that about 10 of the 4,120 patients who took carbamazepine in the study could be expected to develop SJS-TEN. Over the course of the study, however, no patients developed SJS-TEN — a value significantly lower than the historical incidence, with a p-value of less than 0.001.

[ pagebreak ]

Patients did experience milder side effects — 4.3 percent had a mild rash and 0.1 percent had a rash severe enough to require hospitalization — but none developed the hypersensitivity reactions known to be linked to the genetic marker.

"The study thus illustrates the remarkable capability of HLA-B*1502 screening to prevent SJS/TEN among carbamazepine users," wrote Chen-Yang Shen, a research fellow at Academia Sinica and an author of the study, in an e-mail to PGx Reporter, "and also suggests that personalized medicine and pharmacogenomics could be extremely useful in the right clinical settings."

The researchers plan to test similar approaches with other medications. "Based on our previous finding of an association between the development of allopurinol (a very common drug used for gout treatment)-induced severe cutaneous adverse reactions and the HLA-B*5801 genotype, we have conducted a study, similar to this [study], to see whether prescreening of *5801 … can prevent allopurinol-induced serve side effects," Shen wrote.

A New Allele

In the second NEJM study published this week, researchers at the University of Liverpool and others associated with EPIGEN, an international consortium devoted to epilepsy genetics, described an HLA allele, HLA-A*3101, associated with carbamazepine-induced hypersensitivity reactions in patients with European ancestry.

In addition to looking for a predictive marker for SJS-TEN, the team was also investigating possible markers for two other related side effects of the drug: carbamazepine-hypersensitivity syndrome, a skin rash or liver symptoms accompanied by at least two other symptoms such as fever, blood abnormalities, or involvement of other organ systems; and maculopapular exanthema, a mild skin rash that doesn't involve other organ systems.

The researchers recruited patients taking carbamazepine at the University of Liverpool and Walton Centre for Neurology, in the UK, and EPIGEN-affiliated centers in Europe and the US. All of the subjects were of European ancestry, determined by self-report and/or genetic analysis. Twelve patients with SJS-TEN, 26 with hypersensitivity syndrome, and 106 with maculopapular exanthema enrolled in the study, as did 257 clinical controls, patients with epilepsy who had been taking carbamazepine for at least three months and showed no signs of hypersensitivity. For population controls, the researchers used samples from 3,987 people of European ancestry in the UK National Blood Services Collection and the 1958 British Birth Cohort.

The researchers performed genome-wide analysis for 65 subjects with carbamazepine reactions and 200 controls, using the Illumina Infinium 1.2M chip and the Illumina 610K Quad platform. "Keeping with other areas where other drugs are known to cause hypersensitivity reactions, we got a very strong signal on the short arm of chromosome six," where the HLA gene is located, Munir Pirmohamed, chair of pharmacogenetics at the University of Liverpool and an author of the study, told PGx Reporter.

The team then compared 22 of those patients with hypersensitivity reactions recruited by the Liverpool researchers to 2,767 controls of European Ancestry who had provided samples to the UK National Blood service, and imputed classic HLA alleles using "a dense reference panel of SNP data and four-digit HLA types," they wrote in the paper. They used the MACH 1.0 software to impute HLA alleles in the forty-three patients with carbamazepine reactions recruited by the EPIGEN consortium, compared to controls from the 1958 British Birth Cohort.

To validate that the imputation was accurate, high-resolution, sequence-based HLA-A genotyping was performed by Histogenetics on 22 patients with hypersensitivity syndrome and 44 clinical controls. In all cases, results from imputation and direct genotyping were the same.

The HLA-A*3101 allele was associated with all three reactions, the researchers found: maculopapular exanthema, carbamazepine-induced hypersensitivity syndrome, and SJS-TEN. The associations between the HLA-A*3101 allele and each of the three reactions generated odds ratios of 8.33, 12.41, and 25.93, respectively.

Overall, the presence of the allele increased a patient's risk of adverse reaction from 5 percent to 26 percent, while the allele's absence decreased a patient's risk to less than 4 percent. The authors note that these numbers "are similar to those underlying the test criteria calculated for the presence of the HLA-B*5701 allele and abacavir hypersensitivity."

"It's definitely in the realm of clinical relevance," Gianpiero Cavalleri, an author of the study and a geneticist at the Royal College of Surgeons in Ireland, told PGx Reporter.

"The strength of the study is that it not only shows something in one phenotype but in several phenotypes," Pirmohamed said, "and it has been validated using several groups of patients," since the Liverpool and EPIGEN researchers initially recruited patients and analyzed data independently before coming to the same result.

Another notable feature of the results, said Cavalleri, is that the allele is more predictive of risk for more severe reactions. "As you go from rash to hypersensitivity to SJS, the odds ratio gets stronger."

The regulatory implications of the finding are still unclear, however.

"It's for the FDA to decide" whether to update carbamezepine's label "on the basis of the evidence that's there, but looking at the likes of abacavir and so on, it's within the effect size that has led to FDA changes in the past," Cavalleri said.

Both regulatory agencies and clinicians often want to see the results of a prospective pharmacogenomics study before making clinical decisions, he added, and the lack of a rapid test for the allele could be an obstacle to widespread testing of patients.

Given these results, Cavalleri said, "If I personally carried this allele and was offered carbamazepine, I wouldn't take it." Pirmohamed, independently, said, "If I had to have carbamazepine, I'd want to know what my *3101 type was."

Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at tray [at] genomeweb [.] com.