Orally administered progesterone may have advantages over other routes of administration in the treatment of premenstrual syndrome (PMS) because of substantially higher levels of the anxiolytic metabolites 5 alpha and 5 beta pregnanolone. The only previous placebo-controlled trial which used oral progesterone reported beneficial effects in the treatment of PMS. The present study, a double-blind crossover trial, compared the administration of 300 mg daily oral progesterone with 200 mg daily vaginal progesterone and matched placebos for 10 days premenstrually. Although there was a significant treatment effect on symptoms, no difference between active treatments and placebo was found. The trial was terminated with 25 women completing treatment as it was evident that no clinically significant effect of either form of progesterone was likely to be detected even with twice the sample size. Serum levels of progesterone and metabolites showed that oral administration resulted in supraphysiological levels of 5 alpha and 5 beta metabolites and there was a negative correlation between 5 alpha pregnanolone levels and anxiety. However, this did not translate to overall reduction in premenstrual distress or anxiety beyond that achieved by placebo, as measured by validated questionnaires.