The analgesic efficacy and tolerance of lysine clonixinate (LC) as well as LC-induced changes in menstrual prostaglandin levels were studied according to a prospective double-blind randomized crossover design, controlled with ibuprofen (I) and placebo (P). Treatment consisted in 4 consecutive phases: in the first phase, patients refrained from taking medication and during the remaining three phases, they received double-blind fixed doses of 1 tablet of lysine clonixinate 125 mg, I 400 mg or P, q.6 h. at random, three days before onset of menses and during 8 days thereafter. Controls were carried out at each menstrual cycle, assessing pain according to a scale from 0 to 4, onset of premenstrual and intramenstrual symptoms, relief of pain and occurrence of side-effects. During menstruation, patients recorded their assessments of pain in a diary and collected the whole menstrual bleeding during the first three days. The intensity of menstrual pain remained unchanged in controls upon admission (3.16) and during the phase with no treatment (3.04), but was significantly reduced with P (2.4), LC (1.79) and I (1.54). Significantly lower pain intensities compared with placebo were seen with active treatment phases. Forty-two percent of patients treated with P reported premenstrual pain which was significantly reduced to 17% with LC and to 12.5% with I. Active treatment phases revealed 21% of asymptomatic patients during premenstrual and menstrual periods and 71% (LC) and 75% (I) of cases with partial relief of pain. Patients' diaries showed significant pain reductions with LC and I, during the 1st and 2nd days compared with P; such differences were gradually reduced to nil by the 4th day. Levels of menstrual PGs changed according to pain intensity reductions from baseline (P: 29%, (NS); LC: 58% and I: 61%; both were statistically significant, p < 0.01).