Objective: To evaluate the efficacy of chiropractic therapy on the treatment of symptoms associated with premenstrual syndrome.Design: A prospective, randomized, placebo-controlled, crossover clinical trial.Setting: Multicenter private clinics.Subjects: Twenty-five subjects with diagnosed premenstrual syndrome (with a Moos premenstrual syndrome questionnaire plus daily symptom monitoring).Intervention: After randomization, 16 of the subjects received high-velocity, low-amplitude spinal manipulation plus soft tissue therapy 2 to 3 times in the week before menses for at least 3 cycles. The remaining 9 subjects received a placebo treatment with a spring-loaded adjusting instrument wound down for minimum force. After a 1-cycle washout, the 2 groups changed over.Outcome measure: Daily rating of symptom level, comparing total scores for premenstrual week with baseline for treatment and placebo phases.Data analysis: The data were analyzed with paired Student t tests and Wilcoxon signed rank tests, with the statistical significance set at P < .05.Results: There was a significant decrease in scores after treatment compared with baseline scores (P = .00001) and a statistically significant decrease in scores for the treatment phase compared with the placebo (P = .006). For group 1 (n = 16), there was a significant decrease in scores after treatment compared with baseline scores (P = .0001) and a statistically significant decrease in scores for the treatment phase compared with the placebo (P = .041). For group 2 (n = 9), there was a significant decrease in scores during treatment compared with the baseline (P = .01); however, there was no difference at the P = .05 level between treatment and placebo scores.Conclusions: Within the limitations of the study, the results support the hypothesis that the symptoms associated with PMS can generally be reduced by chiropractic treatment consisting of adjustments and soft-tissue therapy. However, the role of a placebo effect needs further elucidation, given that the group receiving the placebo first, although improving over the baseline, showed no further improvement when they had actual treatment.