Variability in the clinical pattern of cutaneous side-effects of drugs with systemic symptoms: does a DRESS syndrome really exist?

Objective: To improve the definition of the various clinical patterns of patients with drug-induced cutaneous side-effects with systemic symptoms, and their possible relationships with the triggering medication, with the ultimate goal of helping in the identification of the causal drug in difficult situations when the patient is taking several drugs.

Methods: Cases of drug-induced cutaneous side-effects associated with various systemic syndromes related to anticonvulsants (carbamazepine, phenytoin and phenobarbitone), minocycline, allopurinol, abacavir and nevirapine were collected retrospectively from the French Pharmacovigilance database (FPD) over a period of 15 years (1985-2000). The clinical patterns typical of the causative drugs were described and compared with data from the literature.

Results: Two hundred and sixteen patients with symptoms and signs consistent with cutaneous drug reactions with systemic symptoms were reported to the FPD during this period of time. Their pattern was similar to published data for these drugs, with fever, cutaneous eruption, hepatic abnormalities and eosinophilia being the most prominent but inconstant symptoms. There are clues suggesting that some particular lesional patterns may exist for some drugs.

Conclusions: Although some trends emerge from these retrospective data, they suggest that no clear, unified outline can currently be defined for these multi-organ drug-induced reactions. Instead, a constellation of various symptoms and signs were recorded, that might be sorted in different patterns according to the causal drug, a finding that might indeed improve accurate identification of the causative drug in patients receiving several principal medications at a time. A national prospective study systematically collecting standardized data is required better to define the outlines of these severe adverse drug reactions and to evaluate prognostic data.