The objective of this study is to assess the efficacy of acupuncture as an adjunctive therapy to pharmacological treatment of chronic pain due to knee osteoarthritis, as well as for, the improvement of physical functioning, reduction of stiffness, and improvement in quality of life. This is a 3-armed single-blinded randomized sham-controlled trial, comparing acupuncture along with pharmacological treatment, sham acupuncture including pharmacological treatment, and pharmacological treatment alone. One-hundred and twenty patients with knee osteoarthritis were randomly allocated to 3 groups. Group I was treated with etoricoxib, Group II was treated with acupuncture and etoricoxib, and Group III was treated with sham acupuncture and etoricoxib.

The participants were recruited randomly from the outpatient population at the Orthopaedic Clinic of the General Hospital of Florina from September1, 2007 until February 15, 2008. The study was approved by the Scientific committee and the Ethics and Deontology committee of the Athens University Hospital "Aretaieion" and the General Hospital of Florina "Eleni Dimitriou". Of 259 patients diagnosed with osteoarthritis of the knee, 55 chose not to participate and from the remaining 201 patients, only 120 me the participation criteria and were included in the study.

To be eligible for the study, patients must have met the American College of Rheumatology criteria for diagnosis of knee osteoarthritis. Patients must also have had a Kellgren-Lawrence (radiologic criterion) score of at least 2 and chronic pain in the knee joint for longer than 3 months. After being fully informed of the study's aims, all participants provided written consent abiding with the declaration of Helsinki. Ineligible patients included those who had received intra-articular corticosteroid or hyaluronate injection during the last 3 months, corticosteroids, antiplatelet drugs (apart from acetylsalicylic acid 100mg), or immunosuppressive drugs. Patients who were pregnant, or who had suffered from a malignancy of any kind, psychiatric disease, stroke, heart attack, kidney failure, active gastric or duodenal ulcer, or gastrorrhagia were also excluded. Finally patients who had previous treatment with acupuncture, other forms of arthritis, or who had arthroplasty during the last year were also ineligible for the study.

After screening, patients meeting the participation criteria were asked for informed consent, and were randomly assigned to 3 different groups: Group I was assigned to treatment with etoricoxib alone, Group II was assigned to treatment with acupuncture and etoricoxib, and Group III was assigned to treatment with sham acupuncture and etoricoxib. Baseline Western Ontario and McMaster Universities Osteoarthritis (WOMAC) values, Visual Analogue Scale (VAS) and SF-36v2 health survey (for quality of life) parameters were recorded. Baseline body mass index (BMI), complete blood count (CBC), C-reactive protein levels, erythrocyte sedimentation rate (ESR), Rheumatoid (Rh) Factor type, kidney and liver function tests, uric acid levels, and arterial blood pressure were recorded and followed throughout the study. Finally, a unique tender point around the knee, was found for each patient and was marked and photographed. A Pain Test™ FDK 20 Algometer was used to score the pain throughout the study using this predetermined tender point. During follow-up, the doctor performing the acupuncture was the only one who was aware of the group each patient belonged to. This physician did not perform randomization or any patient evaluations or statistical analysis.

Randomization Eligible patients were randomly allocated to 3 treatment groups. In order to ensure similar treatment numbers, blocked randomization was used. Blocks of 3 patients were allocated to the 3 treatment groups and a random, computer generated sequence was assigned to each block.

Interventions

Group I (Pharmacological) All patients received 60-mg etoricoxib tablets once-daily for 60 days.

Groups II and III (Acupuncture) A doctor specializing in acupuncture (accredited by the International Council of Medical Acupuncture and Related Techniques [ICMART] and the Medical Acupuncture Society of Northern Greece) selected the acupuncture points that would be effective for osteoarthritis of the knee according to traditional treatment methods.

The protocol for acupuncture was to insert single use, sterile, 30 mm long and 30 gauge acupuncture needles into the local points ST36, SP9, SP10, GB34, Ex-LE 2, and Ex-LE5 as well as the distal points Li4, Ki3, ST40, and SP6. At each point, the patient confirmed the De Qi sensation (a feeling that indicates effective needling). This treatment was given biweekly for 8 weeks. Starting from the third session, the ES - 160 electrostimulator ITO co. (2-6Hz, 150 ms for 20 minutes) was used to stimulate the needles in pairs ST36-SP9 and GB34-SP10.

Sham acupuncture was administrated at the same duration and frequency by the same specialist who performed the non-sham acupuncture. Retractable needles were placed into small adhesive cylinders, so that the needles were supported but did not perforate the skin [30]. The acupuncturist placed the needles at the same points as the non-sham group and used the same pairs of electrodes to simulate the electrical connection.

All patients were instructed to check their arterial blood pressure daily. Patients with risk factors for upper gastrointestinal (GI) tract complications received proton pump inhibitors (PPIs).

All patients completed standard questionnaires at baseline, and after 4, 8 (final treatment week), and 12 weeks.

An examination for possible side effects of both acupuncture and pharmacological treatment was performed at each visit. These side effects included gastrointestinal tract complications, blood pressure increases, and haematomas at the acupuncture point.

The scale for the primary endpoint was the WOMAC index Version VA3.1 and its three subscales (pain, stiffness, and physical function) at the end of treatment, week 8. Additional outcome measures were the WOMAC scales at weeks 4 and 12, VAS (0-100) [26,12] at weeks 4, 8 and 12, and the Greek version of SF-36v2 Health Survey with its two components (physical-mental) for assessment, of health related quality of life at the end of week 8. The score from the direct measurement of pain at the predetermined unique trigger point was measured with the Pain Test™ FDK 20 Algometer of Warner USA at the end of weeks 4, 8, and 12. In cases of bilateral osteoarthritis of the knee, the most painful knee at baseline was assessed throughout the study.

Statistical Analyses Descriptive statistics like the mean ± standard deviation (SD), median, minimum (min) and maximum (max) are used to present continuous variables. Categorical variables were analyzed using χ2 tests and are reported as frequencies or percentages.

For multiple group comparisons, the one-way ANOVA test was used to determine statistically significant differences between groups of variables with data that were normal distributed, while the Kruskal Wallis test was used for groups with data that were not normally distributed. Post hoc tests were performed only for statistically significant group differences. Paired t-tests (or Wilcoxon tests when appropriate) were performed in order to assess differences between different time points in the same group. Data analysis was performed using the statistical package SPSS v 15.0

Inclusion Criteria:

• American College of Rheumatology criteria for diagnosis of knee osteoarthritis
• Kellgren-Lawrence (radiologic criterion) score of at least 2
• chronic pain in the knee joint for longer than 3 months

Exclusion Criteria:

• intra-articular corticosteroid or hyaluronate injection during the last 3 months
• corticosteroids
• antiplatelet drugs (apart from acetylsalicylic acid 100mg)
• immunosuppressive drugs
• malignancy of any kind
• psychiatric disease
• stroke
• heart attack
• kidney failure
• active gastric or duodenal ulcer, or gastrorrhagia
• other forms of arthritis
• arthroplasty during the last year
• previous treatment with acupuncture