The prevalence of asthma is higher in women than men of reproductive age and almost half of all hospitalisations for asthma in women occur during the perimenstrual phase of the cycle. The mechanisms of premenstrual asthma (PMA) are unknown and a definition of PMA has not been clearly presented in the literature. The objective of this study was to determine the prevalence of PMA using a variety of definitions, to investigate the cycle-to-cycle variation in PMA, the effects of oral contraceptive use and the relationship between PMA and premenstrual symptoms. Premenopausal women with asthma (n = 28) were prospectively followed for at least 12 weeks over 2-4 consecutive menstrual cycles. Asthma symptoms, beta(2)-agonist and inhaled corticosteroid use and morning and evening peak expiratory flow were recorded daily. The following types of PMA definitions were investigated: self-reported PMA, increased symptoms, increased medication use, decreased peak flow or a combination of changes in symptoms, medication and peak flow. Changes of more than 20%, for at least 2 consecutive days of the luteal phase (last 14 days of the cycle prior to menstruation) compared to the early follicular phase average (first 7 days after menstruation) were considered PMA. Using a composite definition where subjects experienced increased symptoms and medication use with or without a change in peak flow, 16 subjects were classified as having PMA (57%), while 12 did not have PMA. Only 4 subjects (25%) had PMA in every cycle examined. Fifty-five percent of subjects who used oral contraceptives had PMA, while 59% of subjects who did not use oral contraceptives had PMA. Women who were defined PMA using the composite definition were more likely than those without PMA to experience a 20% decrease in peak flow during the luteal phase. There was no relationship between asthma symptoms and premenstrual symptoms on day 1 of the menstrual cycle in women with PMA. PMA resulting in increased symptoms and medication use occurred in 57% of subjects studied for 2-4 menstrual cycles. The use of oral contraceptives is not protective and further work is required to elucidate the mechanisms of PMA.