This work was done to determine the reasons for variation in the reported rates--ranging from less than 1% to greater than 50%--of sensitivity to aspirin and cross-reactivity to acetaminophen and ibuprofen. In 47 studies that reported rates of sensitivity and in 23 reports that contained series of sensitive patients, we examined the research setting, source of patients, clinical attributes of the study group, admission process, and selection, operational definition, and method of determining sensitivity reactions. In five studies with reasonably well-specified methods, the reported sensitivity rates to aspirin were lowest (0.3% to 0.9%) for patients without allergic tendencies, higher in asthmatics, and highest if patients had nasal polyps or severe atopy. Although not determined in any of these studies, the rate of sensitivity in a general (nonclinical) population would doubtlessly be substantially lower than the rate of three per 1000 reported for nonallergic patients. The admixture of different clinical groups, varying definitions, and ascertainment of a sensitivity reaction seem to be responsible for the variations in the reported rates of sensitivity and cross-reactivity.