Outpatients aged 60 years or older with knee OA were recruited from the Meiji University of Oriental Medicine Hospital. The patients had been clinically and radiologically diagnosed of knee OA according to the criteria of the American College of Rheumatology. Further inclusion criteria for the present study were: (1) knee pain lasting for six months or longer; (2) no radiation of knee pain; (3) radiographic evidence of at least one osteophyte at the tibiofemoral joint (Kellgren-Lawrence grade ≥ 2); (4) normal neurological functions of lumbosacral nerve, including deep tendon reflexes, voluntary muscle action and sensory function; and (5) not receiving acupuncture treatment for more than six months. Exclusion criteria were: (1) major trauma or systemic disease; and (2) receiving conflicting or ongoing co-interventions. Patients under drug treatment were included if there had been no change in medicine and its dosage for one month or longer. This study was approved by the Ethics Committee of the Meiji University of Oriental Medicine.

All enrolled patients gave their written informed consent. According to a block randomised allocation table (generated by Sample Size, version 2.0, Int), the enrolled patients were allocated to (1) the control (CT) group, (2) the acupuncture (ACP) group, (3) the transcutaneous electrical nerve stimulation (TENS) group or (4) the acupuncture and TENS (A&T) group.

The design of this study was a randomly controlled clinical trial using a block randomised procedure. Each patient received a total of five treatments, once per week, and follow-up was measured for ten weeks after the first treatment.

Primary outcome measures were: (1) pain intensity, quantified with a 10 cm visual analogue scale (VAS, 0 – 100 mm) and (2) pain disability measured with the Western Ontario and McMasters Universities Osteoarthritis Index (WOMAC, 0 – 100 points) 14. The WOMAC consists of 24 questions each with five possible responses.

The VAS scores were measured immediately before the first treatment and subsequently at one, two, three, four, five and ten weeks after the first treatment. The WOMAC scores were measured immediately before the first treatment and subsequently five and ten weeks after the first treatment. Each VAS and WOMAC score was measured immediately before treatment of the specified week.

Repeated measures analysis of variance (ANOVA) was used to study the changes in the VAS and WOMAC scores in the three groups. Changes in the time course among groups were considered significant when the interaction was significant at a level of 0.013 (0.05/4). After detection of significant changes in the overall time course with repeated measures ANOVA, pair comparisons were detected with Bonferroni correction. StatView for Windows (version 5.0) or SYSTAT 10 (SYSTAT Inc) was used for the statistical analysis. The results with P values of less than 0.05 were considered statistically significant.

A total of 32 patients (21 women, 11 men; aged 62 – 83 years) were randomly allocated to four groups for specific treatment. Before the treatments, no significant difference was found in baseline variables including age, disease, pain duration, and pain intensity VAS scores among the four groups.

One patient in the ACP group, two patients in the TENS group, one patient in the A&T group, and two patients in the CT group dropped out as they had not responded to the respective treatment. In addition, one patient in the ACP group and one patient in the A&T group dropped out due to adverse effects (i.e. deterioration of symptoms). The total dropout rate was the same in every group. The analyses were performed on the 24 patients who completed the study and provided required information (Figure 1).

The mean pain intensity VAS scores decreased in all groups during treatment, although the exact time courses varied (Table 1). In the A&T group, the pre-treatment (0-week) VAS score and 5-week VAS score were significantly different (P = 0.039). However, differences between the pre-treatment (0-week) scores and 5-week scores of the ACP, TENS and CT groups were not statistically significant.

During the first five weeks of treatment, while the ACP, TENS and A&T groups all reported lower mean pain intensity VAS scores than the CT group, only the A&T group gave a statistically significant reduction in VAS scores.

The mean WOMAC scores decreased in all groups during treatment, although the exact time courses varied (Table 2). In the A&T group, the pre-treatment (0-week) WOMAC score and 5-week WOMAC score were significantly different (P = 0.008). By the end of the fifth week of treatment, the ACP, TENS and A&T groups reported lower WOMAC scores than the CT group. However, the differences between the treatment groups were not statistically significant.