Data from 97 cataract patients and 105 normal controls with ages less than 67 years were collected with respect to sex, age, disease status, drug intake, and blood chemistries. The statistical method of fitting log-linear models was used to determine the association between the case-control indicator variable and the other variables. The following variables were shown to be important and to associate independently with the risk of developing cataracts: age, allergy, diabetes, hypotension, hypertension, use of analgesics and coronary disease.