Objective

Asians who carry the HLA-B*1502 allele have an elevated risk of developing Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) when treated with the antiepileptic drugs (AEDs) carbamazepine (CBZ) and phenytoin (PHT). With a focus on Singapore, this analysis identifies circumstances in which genotyping and targeted treatment with alternative AEDs that do not induce SJS/TEN is likely to be more cost-effective than 1) treatment with CBZ or PHT without genotyping or 2) providing a more expensive drug that does not induce SJS/TEN to all patients without genotyping.

Methods

A decision tree model was developed in TreeAge. The model takes into account costs of epilepsy treatments and genotyping, reductions in quality of life and increased costs resulting from SJS/TEN complications, the prevalence of the risk allele, the positive predictive value (PPV) of genotyping, life expectancy, and other factors.

Results

Compared with no genotyping and providing CBZ to all, genotyping results in an incremental cost-effectiveness ratio of $37,030/quality-adjusted life year (QALY) for Chinese patients, $7,930/QALY for Malays, and $136,630/QALY for Indians in Singapore.

Conclusions

Because of the different population allele frequencies of HLA-B*1502 among different ethnic groups, genotyping for HLA-B*1502 and providing alternate AEDs to those who test positive is cost-effective for Singaporean Chinese and Malays, but not for Singaporean Indians. Population frequency of HLA-B*1502, PPV, duration of treatment relative to life expectancy, and costs of alternative drugs are the key drivers influencing cost-effectiveness.