Smith-Whitley and Lange raise several important points. First, the "retinoic acid syndrome" is a constellation of clinical signs (fever, weight gain, pulmonary infiltrates, and pleural effusions) that can mimic various illnesses to which patients with acute promyelocytic leukemia (APL) are susceptible, especially congestive heart failure and infectious pneumonia [1]. (Indeed, our housestaff has identified two recent patients who presented with the "syndrome" before treatment was initiated.) Second, this complex by no means accounts for all idiosyncratic drug-related effects observed in patients with APL; variations have included pericarditis and high spiking fevers without an apparent infectious source, both of which also appear to be responsive to steroids. Third, the use of high-dose corticosteroids modifies the course of this syndrome in many—but certainly not all—patients. Fourth, children are extraordinarily sensitive to the central nervous system effects of all-trans retinoic acid (TRA) and tolerate only a fraction of the dose that can be administered to adults [2]. The French experience suggests that patients with APL have responded to TRA doses of 15 mg/m² (Degos L. Personal communication). Because intracranial hypertension is dose-related rather than idiosyncratic in children, further study of their dose-response is warranted. Fifth, although the cause of TRA-induced leukocytosis is unclear, worsening of the coagulopathy is unusual.

Most importantly, their experience re-emphasizes that most patients with TRA-induced leukocytosis do not develop symptoms of the "TRA syndrome"; therefore, treatment with cytotoxic drugs specifically directed at the leukocyte count risks exacerbation of the coagulopathy. For the next year or so, we strongly suggest that practitioners eschew compassionate treatment requests for this drug and continue to refer patients to major centers until this important safety issue is resolved.