Using a case-control study design, patients with eosinophilia myalgia syndrome (EMS) who had used L-tryptophan (LT) were compared with LT users who did not develop EMS. Of the 113 case patients and 95 controls who had used a retail brand that could be traced to a bulk LT producer, all (100%) case-patients and 69 (73%) controls used LT brands that were traced to Showa Denko K.K. (lower 95% CL = 10.0). Among the users of LT produced by Showa Denko K.K., the risk of EMS was greater for persons who used LT produced after December 1, 1988 (OR = 25.8, [95% CL = 7.1, 101.4]). The risk of developing EMS increased with increased dosage of LT, increased age, and use of LT as a sleeping aid. These epidemiologic data support the hypothesis that the etiologic agent in EMS is a contaminant introduced into LT products during production.