The DRESS syndrome: the great clinical mimicker.

The life-threatening DRESS (drug rash with eosinophilia and systemic symptoms) syndrome is characterized by the presence of at least three of the following findings: fever, exanthema, eosinophilia, atypical circulating lymphocytes, lymphadenopathy, and hepatitis. This syndrome is difficult to diagnose, as many of its clinical features mimic those found with other serious systemic disorders. This idiosyncratic reaction occurs most commonly after exposure to drugs such as allopurinol, sulfonamides, and aromatic anticonvulsants such as phenytoin, phenobarbital, and carbamazepine. We describe a 44-year-old woman who was brought to the emergency department with new-onset hemorrhagic stroke. She was admitted to the intensive care unit where she received supportive care that included clonidine and hydralazine for blood pressure control and phenytoin for seizure prophylaxis. On hospital day 21, the patient developed signs and symptoms of severe sepsis. Despite receipt of broad-spectrum antibiotics (vancomycin and piperacillin-tazobactam) and supportive care, the patient's clinical condition worsened with progressive jaundice, severe oliguria, and labile blood pressures. All cultures revealed no growth, and her chest radiograph remained clear. Several days after the onset of her fever, the patient developed several hematologic abnormalities including thrombocytopenia, with schistocytes present on a peripheral smear. She also had an elevated lactate dehydrogenase level. A provisional diagnosis of thrombotic thrombocytopenic purpura was made; however, the patient then developed severe facial edema, nearly global erythroderma, and severe exfoliative dermatitis. A punch biopsy of the skin was compatible with the DRESS syndrome. Phenytoin, vancomycin, and piperacillin-tazobactam were discontinued, and the patient was started on systemic corticosteroids, with rapid resolution of her fever and eosinophilia and progressive improvement in her skin rash and multiorgan system dysfunction. Use of the Naranjo adverse drug reaction probability scale indicated a probable relationship (score of 5) between the patient's development of DRESS syndrome and treatment with phenytoin. Clinicians should have a high index of suspicion for the DRESS syndrome in patients being treated with aromatic anticonvulsants who develop a sepsis-like syndrome. Furthermore, considering the potential severe effects associated with phenytoin, the risks and benefits should be carefully evaluated before using this agent for seizure prophylaxis.