Thirty patients with tension-type headache were randomly chosen to undergo a trial of traditional Chinese acupuncture and sham acupuncture. Five measures were used to assess symptom severity and treatment response: intensity, duration and frequency of headache pain episodes, headache index and analgesic intake. The five measures were assessed during a 4 week baseline period, after 4 and 8 weeks of treatment, and 1, 6 and 12 months thereafter. Before the start of the study, each patient was administered the MMPI. Split-plot ANOVAs showed that, compared to baseline, at 1 month after the end of treatment and for the 12 month follow-up, the frequency of headache episodes, analgesic consumption and the headache index (but not the duration or intensity of headache episodes) significantly decreased over time; however, no difference between acupuncture and placebo treatment was found. No single MMPI scale predicted the response to treatment, but the mean MMPI profile of acupuncture non-responders showed the presence of 'Conversion V'.