Treatment with flurbiprofen (100 mg twice a day for 5 days), tranexamic acid (1.5 gm three times a day for 3 days and 1 gm twice a day for another 2 days), and an intrauterine contraceptive device releasing 20 mu-g levonorgestrel per day was compared in women with idiopathic menorrhagia. The menstrual blood loss during two control periods in 15 women subsequently treated with flurbiprofen and tranexamic acid was 295 +/- 52 ml and 203 +/- 25.2 ml in the 16 women later fitted with a levonorgestrel-releasing intrauterine contraceptive device. Menstrual blood loss was reduced by all three forms of treatment. The reduction in menstrual blood loss expressed as a percentage of the mean of two control cycles for each form of treatment was as follows: flurbiprofen, 20.7% +/- 9.9%; tranexamic acid, 44.4% +/- 8.3%; levonorgestrel-releasing intrauterine contraceptive device after 3 months, 81.6% +/- 4.5%; levonorgestrel-releasing intrauterine contraceptive device after 6 months, 88.0% +/- 3.1%; levonorgestrel-releasing intrauterine contraceptive device after 12 months, 95.8% +/- 1.2%. The reduction in menstrual blood loss achieved by the levonorgestrel-releasing intrauterine contraceptive device was greater than that recorded with flurbiprofen (p < 0.001) and tranexamic acid (p < 0.01), and was greater for tranexamic acid when compared with flurbiprofen (p < 0.05). The levonorgestrel-releasing intrauterine contraceptive device was the only form of treatment to reduce mean menstrual blood loss below 80 ml per menstruation, the upper limit of normal menstrual blood loss.