It has been hypothesized that hyperestrogenemia may underlie myocardial infarction. As a test of this hypothesis, the serum estradiol and testosterone levels were estimated in samples collected prospectively from 96 male patients aged 52-74 years [mean age 60.8 +/- 6.3 (SD)] who had had a myocardial infarction and from 96 matched control subjects in the Honolulu Heart Program. Established risk factors for myocardial infarction were measured prospectively. No significant difference between patients and control subjects in mean estradiol or testosterone level was observed. The only established risk factor that was significantly different was blood pressure, which was higher in the patients. Thus, the data did not confirm the hypothesis. However, two major pitfalls for prospective studies of estradiol in myocardial infarction, which might have affected the validity of the results, were observed, namely, deterioration of estradiol values with prolonged storage (8.5-12 years in this study) and intervention. The value of this study, therefore, may lie in pointing out difficulties in carrying out prospective studies on sex hormones.