Objective: To evaluate the effectiveness of acupuncture for pain relief and function improvement in patients with knee osteoarthritis and to determine the feasibility of an eight-week acupuncture intervention. Design: Pilot randomized controlled trial. Setting: Three teaching hospitals in China. Subjects: Patients with knee osteoarthritis (Kellgren grade II or III). Interventions: Patients were randomly assigned to an eight-week (three sessions per week) intervention of either traditional Chinese acupuncture or sham acupuncture. Main measures: The primary outcome was response rate-the proportion of patients achieving score > 36% decrease in Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain and function at week 8 compared with baseline. Secondary outcomes included pain, function and quality of life. Results: Of 42 patients randomized, 36 (85.7%) completed the study. There was no significant difference in response rate between the traditional Chinese acupuncture and control groups: 61.9% (13 of 21) versus 42.9% (9 of 21) achieved score > 36% decrease in WOMAC pain and function at week 8 (P = 0.217). The sum of WOMAC pain and function scores at week 8 was 11.6 (9.1) in the traditional Chinese acupuncture group compared with 16.3 (10.9) in the control group (P = 0.183). There was no significant difference between groups. Three adverse events were recorded and were classified as mild. Conclusion: It showed that three sessions per week acupuncture intervention of knee osteoarthritis was feasible and safe. No difference was observed between groups due to small sample size. Larger (sample size > 296) randomized controlled trials of this intervention appear justified.