—To study the incidence of eosinophilia-myalgia syndrome, the risk factors associated with the syndrome, and the clinical spectrum of illness associated with L-tryptophan use in an exposed population.
—Retrospective cohort and nested case-control studies of risk factors for eosinophilia-myalgia syndrome using inpatient and outpatient chart reviews, telephone interviews, and in-person patient interviews. Descriptive study of clinical course of L-tryptophan users.
—Office practice of one psychiatrist based in a small city (population 43467) in South Carolina.
—Eligible subjects were all patients from the practice who used L-tryptophan during the 1989 study interval. Of these, 418 (87%) were interviewed.
Main Outcome Measures.
—Clinical spectrum of illness associated with L-tryptophan use, including definite and possible cases of eosinophilia-myalgia syndrome.
—Among the 418 interviewed L-tryptophan users, we identified 47 definite cases (11%) and 68 possible cases (16%) of eosinophilia-myalgia syndrome, most of which involved patients who were using one retail brand of L-tryptophan (brand A). Among the 157 brand A users, we identified 45 definite cases (29%) and 36 possible cases (23%) of eosinophilia-myalgia syndrome, and the risk for the syndrome increased as the brand A dose increased. Fifty percent (19 of 38) of those using more than 4000 mg/day developed definite eosinophilia-myalgia syndrome, and 84% (32 of 38) developed either definite or possible eosinophilia-myalgia syndrome. On multivariate analysis, risk for definite eosinophilia-myalgia syndrome was associated with brand A dose and age of the patient; however, gender, race, and use of other medications were not associated with the syndrome.
—These results suggest that many people exposed to the agent causing eosinophilia-myalgia syndrome may develop illness, and dose of presumably contaminated L-tryptophan is the single most important predictor of eosinophilia-myalgia syndrome. The broad range of signs and symptoms reported by patients using L-tryptophan illustrates that a strict case definition may identify only about half of those affected.(JAMA. 1992;267:77-82)