Patch testing for the diagnosis of anticonvulsant hypersensitivity syndrome: a systematic review.

Anticonvulsant hypersensitivity syndrome (AHS), also known by the other names drug rash (reaction) with eosinophilia and systemic symptoms (DRESS) and drug-induced hypersensitivity syndrome (DIHS), is a rare and potentially fatal reaction that occurs in susceptible patients after exposure to certain drugs, including aromatic anticonvulsants. Because of its ill-defined clinical picture and resemblance to other diseases, the diagnosis of AHS is often difficult and requires a safe and reliable diagnostic test. The skin patch test has been proven to be very useful for prediction and diagnosis of some types of hypersensitivity reactions such as delayed drug eruptions to beta-lactam antibacterials. However, the diagnostic value of patch testing for AHS is yet to be determined and its negative predictive values (NPVs) and positive predictive values (PPVs) are still unknown. This systematic review attempts to evaluate the usefulness of patch tests in the diagnosis of AHS and to examine different technical aspects of patch testing that may contribute to its performance. We included studies in which aromatic anticonvulsant drugs are the likely causes of the hypersensitivity reaction. Analysis of original publications from 1950 to August 2008 and cited in PubMed, MEDLINE and EMBASE has revealed contradictory findings, possibly due mainly to the use of unstandardized methods. Numerous factors have been suggested to affect the final result of the test, including the following: type of drug tested; concentration of drug and vehicle used; timing of the test after exposure; and the clinical picture of the reaction. The PPV of the test in optimal conditions was as high as 80-90% depending on the drug tested. On the other hand, this value is around 10-20% in many other published studies. Although patch testing may be a useful diagnostic test for AHS, accurate determination of its sensitivity and specificity is yet to be achievable due to the lack of a gold standard test against which the performance of patch testing can be measured. Its PPV appears to be higher than its NPV, a matter that necessitates the use of other confirmatory tests in case of negative patch tests (e.g. careful systemic rechallenge). The benefit of testing appears to be maximal with certain drugs (i.e. carbamazepine and phenytoin) and for specific clinical manifestations (strong reactions). It should be performed 2-6 months after recovery from the date of the ADR for best results, with adequate vehicle control.