In response to case reports of toxic shock syndrome, six case-comparison studies were done in 1980. Early cases were predominantly in menstruating women, and the use of tampons was strongly associated with the onset of illness. Because of the widespread publication of this finding, the case-comparison studies had problems due to differential ascertainment and recall bias. However, the number of cases among women was so great and the relation with tampon use so marked that unreasonable assumptions are necessary if the results are to be attributed to these biases. The studies show the power of epidemiologic methods, even given the unfavorable circumstance of an uncommon condition, associated with a common practice. Although the absolute risk of toxic shock syndrome is small, tampon users have more than a tenfold excess risk of the condition over women not using tampons, and the use of tampons is an adequate explanation for the excess risk of women over men.