Abstract

Objective . To evaluate the evidence for the effectiveness of acupuncture in peripheral joint osteoarthritis (OA).

Methods . Systematic searches were conducted on Medline, Embase, AMED, Cochrane Library, CINAHL, British Nursing Index, PsychINFO and CAMPAIN until July 2005. Hand-searches included conference proceedings and our own files. There were no restrictions regarding the language of publication. All randomized controlled trials (RCTs) of acupuncture for patients with peripheral joint OA were considered for inclusion. Trials assessing needle acupuncture with or without electrical stimulation were considered if sham- or placebo-controlled or controlled against a comparator intervention. Trials testing other forms of acupuncture were excluded. Methodological quality was assessed and, where possible, meta-analyses were performed.

Results . Thirty-one possibly relevant studies were identified and 18 RCTs were included. Ten trials tested manual acupuncture and eight trials tested electro-acupuncture. Overall, ten studies demonstrated greater pain reduction in acupuncture groups compared with controls. The meta-analysis of homogeneous data showed a significant effect of manual acupuncture compared with sham acupuncture (standardized mean difference 0.24, 95% confidence interval 0.01–0.47, P = 0.04, n = 329), which is supported by data for knee OA. The extent of heterogeneity in trials of electro-acupuncture prevented a meaningful meta-analysis.

Conclusions . Sham-controlled RCTs suggest specific effects of acupuncture for pain control in patients with peripheral joint OA. Considering its favourable safety profile acupuncture seems an option worthy of consideration particularly for knee OA. Further studies are required particularly for manual or electro-acupuncture in hip OA.

Introduction

Osteoarthritis (OA) is the most common form of arthritis, and the most common reason for total hip and total knee replacement [ 1 ]. The underlying disease processes of OA involve cartilage degeneration, proliferation and remodelling of subchondral bone structure. Weight-bearing peripheral and axial joints are most often affected [ 2 ]. OA is associated with symptoms of pain and functional disability. Physical disability arising from pain and loss of functional capacity reduces the quality of life and increases the risk of further morbidity and mortality [ 3 ]. Among adults aged ≥30 yrs, symptomatic knee OA occurs in ∼6% and symptomatic hip OA in about 3% [ 1 ]. Before the age of 50 yrs, the prevalence of OA in most joints is higher in men than in women, whereas in later life women are more often affected than men in hands, feet and knees [ 4 ].

The treatment of OA is largely symptomatic and includes analgesics, NSAIDS, glucosamine, topical analgesics such as capsaicin cream as well as exercise, behavioural interventions and surgical treatment [ 5 ]. Most drug treatments are associated with well-documented risks such as gastrointestinal irritation and bleeding, renal and hepatic toxicity, as well as an increased risk of hypertension. Some of these adverse events are most prominent in the elderly—the very group most commonly affected by OA [ 6 ].

Non-pharmacological treatments such as acupuncture are therefore attractive. Acupuncture is often used for OA and chronic pain relief [ 2 , 7 ]. In the US, over 2 million people use acupuncture annually [ 8 ]. To evaluate the evidence for the effectiveness of acupuncture in peripheral joint OA, we conducted a systematic review and meta-analysis of randomized controlled trials (RCTs).

Methods

Data sources

Searches were performed in July 2005 using Medline, Embase, AMED, CINAHL, British Nursing Index, PsychINFO, CAMPAIN and Cochrane Library. Search terms used were OA, degenerative arthritis, osteoarthrosis, joint pain, knee pain, hip pain, arthritis, acupuncture, ear acupuncture and electro-acupuncture. In addition, our own files were manually searched and authors were contacted. Original articles were obtained, and all reference lists were scanned for further relevant articles.

Study selection

All articles were included that reported an RCT in which patients with peripheral joint OA were treated with needle acupuncture with or without electrical stimulation. Trials testing other forms of acupuncture, such as laser acupuncture or acupuncture using moxibustion were excluded. Studies comparing two different forms of acupuncture and those in which no data or statistical comparisons were reported were also excluded. No language restrictions were applied.

Data extraction

Data were extracted independently by two of the authors (Y.D.K. and M.H.P) using a specifically designed data extraction form. For each study, trial design, randomization, blinding and handling of drop outs, inclusion and exclusion criteria, details of treatment and control procedures, main outcomes measure and main results were recorded. Differences during this process were settled by discussion.

Data synthesis

The mean change of pain scores assessed on 100 mm visual analogue scale (VAS) and on the Western Ontario MacMaster Osteoarthritis Index (WOMAC) pain scale compared with baseline were defined as primary outcome measures. They were used to assess the differences between the intervention groups and the control groups. Datapoints chosen were those at the end of the treatment period. Means and 95% confidence intervals (CIs) were calculated using the Cochrane Collaboration's standard meta-analysis software (RevMan 4.27, Update Software Ltd., Oxford, UK). For some studies, the information was insufficient. In all of these cases, we contacted the original authors requesting further information. The variance of the change was imputed using a correlation factor of 0.4 suggested by the Cochrane Collaboration. The chi-square test for heterogeneity was used to assess whether the distribution of the results was compatible with the assumption that inter-trial differences were attributable to chance variation alone. Homogeneous datasets were meta-analysed using a fixed effects model. Sensitivity analyses were planned, but abandoned due to the small number of trials that provided adequate data. Data which were not suitable for meta-analysis were assessed and weighed in a subjective manner according to important design features relating to study quality (Jadad score), sample size, analyses (e.g. baseline/inter-group comparisons), characteristics of treatment and control groups (e.g. waiting list, sham/placebo control).

Quality assessment

Study quality was assessed using the Jadad score [ 9 ]. Taking into account the difficulties in blinding, we used a modification of this scale [ 10 ]. For a total of 5 points, we awarded 1 point each: if the study was described as randomized; for appropriate method; if subjects were blinded to intervention; if evaluator was blinded to intervention; for description of withdrawals and dropouts. Subject blinding was assumed where the control intervention was indistinguishable from acupuncture, even if the word ‘blinding’ did not occur in the report. Trials with 4 or 5 points were considered to be of high quality. For knee and hip OA, acupuncture treatment was considered adequate if a minimum of four points were needled for a minimum of 20 min, at least six treatments were given with a frequency of at least one session weekly, sufficient electrical stimulation for a minimal sensation was given or deqi (needle sensation) was achieved.

Results

Study description

The literature searches revealed 31 possibly relevant studies [ 11–41 ] ( Fig. 1 ). Thirteen studies were excluded [ 11–22 , 41 ]. Key data of the remaining 18 that included RCTs are summarized in Table 1 . Ten of these trials employed manual acupuncture ( Table 1 ). Most trials of electro-acupuncture were performed using frequencies of 2–8 Hz [ 29 , 31 , 34 , 35 , 37–39 ]. The sensation of deqi or numbness during needle stimulation was reported in 13 of the reviewed trials [ 23 , 25 , 27 , 28 , 30 , 31 , 33–36 , 38–40 ]. The treatment duration was 20–30 min in most trials [ 23–25 , 27–39 ].

F ig . 1.

Flowchart of trial selection process.

F ig . 1.

Flowchart of trial selection process.

T able 1.

RCTs of acupuncture for osteoarthritis

First author (date) Study design, quality score Sample size, OA site Experimental treatment Control intervention Main outcome measures Results 
Christensen [ 27 ] (1992)  2 parallel groups, assessor blind 4 32, knee A: formula AT, ST-34, 35, 36, SP-10, XL-2, LI-4, B: waiting list (1) Pain VAS  (1) Intergroup difference ( P < 0.01)  
   2 sessions weekly for 3 weeks  (2) Time to walk 50 m  (2) Intergroup difference ( P < 0.01)  
     (3) Time to climb 20 steps  (3) Intergroup difference ( P < 0.01)  
     (4) HSS knee function scale  (4) Intergroup difference ( P < 0.01)  
Witt [ 36 ] (2005)  3 parallel groups, patient blind 4 300, knee A: Chinese AT, ST-34, 35, 36, SP-9, 10, UB-40, Kid-10, GB-33, 34, Li-8, Heding, Xiyan, and selective additional points (SP-4, 5, 6, ST-6, UB-20, 57, 58, 60,62, Kid-3), B: sham (superficial insertion on non-AT points) (1) WOMAC (pain, stiffness and function)  (1) Differences compared with B, C ( P < 0.01)  
   1–2 sessions weekly for 8 weeks, C: waiting list   
   12 sessions in total    
Petrou [ 25 ] (1988)  2 parallel groups, assessor and patient blind 3 31, knee A: formula AT, ST-35, 36, 43, EX-31, 32, UB-40, LI-4, B: sham (superficial insertion into near non-classical, non-tender points) (1) Starting pain (4-point scale)  (1) Intergroup difference ( P < 0.05)  
   3 sessions weekly for 3 weeks, 8 sessions in total  (2) Night pain (4-point scale)  (2) Intergroup difference ( P < 0.001)  
     (3) Walking pain, pain descending stairs and walking time (3) No intergroup differences 
Takeda [ 28 ] (1994)  2 parallel groups, assessor and patient blind 4 40, knee A: formula AT, ST-35, GB-34, SP-9, Extra–31, 32, B: sham (superficial insertion about 1 inch from the AT points) (1) McGill Pain Questionnaire (1)–(3) No intergroup differences 
   3 sessions weekly for 3 weeks  (2) WOMAC (pain, stiffness and function)  
     (3) Pain threshold (dolorimeter)  
Molsberger [ 30 ] (1994)  2 parallel groups, assessor and patient blind 4 97, knee A: formula AT, ST-34, 35, 36, UB-9, 10, GB-34, Extra-31, 32, B: sham (superficial insertion at non-AT points on homolateral leg) (1) Pain VAS  (1) Intergroup difference ( P < 0.05)  
   2 sessions weekly for 5 weeks  (2) Function (Lysholm score) (2) No intergroup difference 
Fink [ 33 ] (2001)  2 parallel groups, assessor and patient blind 5 67, hip A: formula AT, GB-30, 31, 34, BL-37, ST-40, BL-54, six ‘ah shi’ points, B: sham (insertion at same depth at least 5 cm away from AT point) (1) Pain VAS (1) No intergroup difference 
   10 sessions in total within 3 weeks  (2) Functional impairment (hip function index) (2) No intergroup difference 
Berman [ 35 ] (2004)  3 parallel groups, assessor and patient blind 5 570, knee A: Chinese EA, GB-34, SP-9, ST-35, 36, Xiyan, UB-60, GB-39, SP-6, Kid-3, B: sham (combined insertion on abdomen and noninsertion procedure on same AT points as A) (1) WOMAC (pain)  (1)(2) Differences compared with B ( P < 0.01) and C  
   23 sessions in total within 26 weeks C: education (2) WOMAC (function) (3) No intergroup difference 
     (3) Patient global assessment  
Vas [ 40 ] (2004)  2 parallel groups, assessor blind 4 97, knee A: formula EA, GB-34, SP-9, EX-LE5, ST-36, KI-3, SP-6, LI-4, ST-40, B: placebo EA (placebo needles, no perforation of the skin, electrical stimulation) plus diclofenac (1) Pain VAS  (1)–(4) Intergroup differences ( P < 0.001)  
   1 session weekly for 12 weeks plus diclofenac  (2) WOMAC pain  
     (3) WOMAC stiffness  
     (4) WOMAC function 
Dickens [ 26 ] (1989)  2 parallel groups, assessor blind 4 12, thumb A: formula AT, point not described, B: mock TENS (1) Pain VAS  (1) Changes compared with baseline for AT ( P = 0.02)  
   6 sessions within 2 weeks    
Yurtkuran [ 29 ] (1999)  4 parallel groups, assessor blind 3 100, knee A: formula EA, SP-9, GB-34, ST-34, 35, B: TENS (1) Present pain intensity (5-point scale)  (1)–(5) Changes compared with baseline for EA ( P < 0.05)  
   5 sessions weekly for 2 weeks C: ice massage (2) Stiffness (1)–(5) No differences between treatment groups 
    D: placebo TENS (electrodes disconnected) (3) 50-foot walking time  
     (4) Quadriceps muscle strength  
     (5) Active knee flexion  
Ng [ 34 ] (2003)  3 parallel groups, assessor blind 3 24, knee A: formula EA, ST-35, EX-LE-4, B: TENS (1) Numerical pain rating scale (1) Change in A, B compared with baseline 
   8 sessions in total within 2 weeks C: general education on knee care only (2) Passive range of movement (2) No changes in either group 
     (3) The timed up and go test (TUGT) (3) Change in A compared with baseline 
Ammer [ 24 ] (1988)  2 parallel groups, open 1 28, knee A: formula AT, UB-54, Liv-9, GB-34, 30, 32, ST-36, B: physical therapy (1) Starting pain, walking pain (4-point scale) (1) No intergroup differences 
   2 sessions weekly for 4 weeks, 8 sessions in total  (2) Total pain (4-point scale)  (2) Intergroup difference in favour of control ( P < 0.01)  
Stener-Victorin [ 39 ] (2004)  3 parallel groups, open 3 45, hip A: formula EA, BL-54, 36, GB-29, 30, 31, ST-31, GB-34, BL-60, B: hydrotherapy plus education (1) Pain VAS (related to motion and load) (1) No intergroup differences 
   2 sessions weekly for 5 weeks, C: education alone (2) Daytime pain VAS  (2) Change in A, B compared with baseline ( P <0.05)  
   10 sessions in total plus education  (3) Nighttime pain VAS  (3) Change in A, B compared with baseline ( P < 0.01)  
Haslam [ 32 ] (2001)  2 parallel groups, open 3 32, hip A: formula AT, GB-29, 30, 34, 43, ST-44, LI-4, four ‘ah shi’ points, B: advice plus a set of five exercises (1) modified version of WOMAC (pain, stiffness and physical function)  (1) Intergroup difference ( P = 0.02)  
   1 session weekly for 6 weeks,    
   6 sessions in total    
 
Jia [ 23 ] (2005)  3 parallel groups, open 2 120, knee A: Chinese AT, ST-34, 36, SP-9, 10, GB-34, medial Xiyan, lateral Xiyan, ‘ah shi’ points, B: Chinese AT plus functional exercise (1) effective rate (4-point scale)  (1)(2) No intergroup difference A vs C ( P > 0.05)  
   1 session daily, 45 sessions within 2 months C: functional exercise (2) recurrence rate  (1) Difference in favour of B compared with A and C ( P < 0.01)  
       (2) Difference in favour of B compared with A and C ( P < 0.05)  
Berman [ 31 ] (1999)  Partial crossover, assessor blind 4 73, knee A: Chinese EA, GB-34, SP-9, ST-35, 36, Xiyan, UB-60, GB-39, SP-6, Kid -3, B: conventional therapy alone (oral) (1) WOMAC (total, pain and disability)  (1)–(2) Intergroup differences ( P < 0.001)  
   2 sessions weekly for 8 weeks, 16 sessions in total  (2) Lequesne indices  
Sangdee [ 37 ] (2002)  4 parallel groups, assessor blind 3 193, knee A: formula EA, ST-35, Liv-8, medial Xiyan, trigger point, B: formula EA plus diclofenac, (1) Pain VAS  (1) Intergroup difference ( P < 0.05), A vs C and A vs D  
   3 sessions weekly for 4 weeks, C: diclofenac plus placebo EA (patch electrodes to the selected AT points) (2) WOMAC pain  (5) Intergroup difference ( P < 0.05), A vs D  
   12 sessions in total D: placebo EA plus placebo tablet (3) WOMAC stiffness (3)(4)(6) No intergroup differences 
     (4) WOMAC disability  
     (5) Lequesne's functional index  
     (6) 50 ft walk time  
Tukmachi [ 38 ] (2004)  3 parallel groups, assessor blind 4 30, knee A: Chinese EA, LI-4, SP-9, 10, Xiyan, GB-34, ST-36, LR-3, BL-40, 57, B: Chinese EA plus symptomatic medication (1) Pain VAS  (1)–(3) Intergroup difference ( P < 0.05), A vs C and B vs C at 5 weeks  
   2 sessions weekly for 5 weeks, C: symptomatic medication (2) WOMAC (pain)  
   10 sessions in total  (3) WOMAC (stiffness)  
First author (date) Study design, quality score Sample size, OA site Experimental treatment Control intervention Main outcome measures Results 
Christensen [ 27 ] (1992)  2 parallel groups, assessor blind 4 32, knee A: formula AT, ST-34, 35, 36, SP-10, XL-2, LI-4, B: waiting list (1) Pain VAS  (1) Intergroup difference ( P < 0.01)  
   2 sessions weekly for 3 weeks  (2) Time to walk 50 m  (2) Intergroup difference ( P < 0.01)  
     (3) Time to climb 20 steps  (3) Intergroup difference ( P < 0.01)  
     (4) HSS knee function scale  (4) Intergroup difference ( P < 0.01)  
Witt [ 36 ] (2005)  3 parallel groups, patient blind 4 300, knee A: Chinese AT, ST-34, 35, 36, SP-9, 10, UB-40, Kid-10, GB-33, 34, Li-8, Heding, Xiyan, and selective additional points (SP-4, 5, 6, ST-6, UB-20, 57, 58, 60,62, Kid-3), B: sham (superficial insertion on non-AT points) (1) WOMAC (pain, stiffness and function)  (1) Differences compared with B, C ( P < 0.01)  
   1–2 sessions weekly for 8 weeks, C: waiting list   
   12 sessions in total    
Petrou [ 25 ] (1988)  2 parallel groups, assessor and patient blind 3 31, knee A: formula AT, ST-35, 36, 43, EX-31, 32, UB-40, LI-4, B: sham (superficial insertion into near non-classical, non-tender points) (1) Starting pain (4-point scale)  (1) Intergroup difference ( P < 0.05)  
   3 sessions weekly for 3 weeks, 8 sessions in total  (2) Night pain (4-point scale)  (2) Intergroup difference ( P < 0.001)  
     (3) Walking pain, pain descending stairs and walking time (3) No intergroup differences 
Takeda [ 28 ] (1994)  2 parallel groups, assessor and patient blind 4 40, knee A: formula AT, ST-35, GB-34, SP-9, Extra–31, 32, B: sham (superficial insertion about 1 inch from the AT points) (1) McGill Pain Questionnaire (1)–(3) No intergroup differences 
   3 sessions weekly for 3 weeks  (2) WOMAC (pain, stiffness and function)  
     (3) Pain threshold (dolorimeter)  
Molsberger [ 30 ] (1994)  2 parallel groups, assessor and patient blind 4 97, knee A: formula AT, ST-34, 35, 36, UB-9, 10, GB-34, Extra-31, 32, B: sham (superficial insertion at non-AT points on homolateral leg) (1) Pain VAS  (1) Intergroup difference ( P < 0.05)  
   2 sessions weekly for 5 weeks  (2) Function (Lysholm score) (2) No intergroup difference 
Fink [ 33 ] (2001)  2 parallel groups, assessor and patient blind 5 67, hip A: formula AT, GB-30, 31, 34, BL-37, ST-40, BL-54, six ‘ah shi’ points, B: sham (insertion at same depth at least 5 cm away from AT point) (1) Pain VAS (1) No intergroup difference 
   10 sessions in total within 3 weeks  (2) Functional impairment (hip function index) (2) No intergroup difference 
Berman [ 35 ] (2004)  3 parallel groups, assessor and patient blind 5 570, knee A: Chinese EA, GB-34, SP-9, ST-35, 36, Xiyan, UB-60, GB-39, SP-6, Kid-3, B: sham (combined insertion on abdomen and noninsertion procedure on same AT points as A) (1) WOMAC (pain)  (1)(2) Differences compared with B ( P < 0.01) and C  
   23 sessions in total within 26 weeks C: education (2) WOMAC (function) (3) No intergroup difference 
     (3) Patient global assessment  
Vas [ 40 ] (2004)  2 parallel groups, assessor blind 4 97, knee A: formula EA, GB-34, SP-9, EX-LE5, ST-36, KI-3, SP-6, LI-4, ST-40, B: placebo EA (placebo needles, no perforation of the skin, electrical stimulation) plus diclofenac (1) Pain VAS  (1)–(4) Intergroup differences ( P < 0.001)  
   1 session weekly for 12 weeks plus diclofenac  (2) WOMAC pain  
     (3) WOMAC stiffness  
     (4) WOMAC function 
Dickens [ 26 ] (1989)  2 parallel groups, assessor blind 4 12, thumb A: formula AT, point not described, B: mock TENS (1) Pain VAS  (1) Changes compared with baseline for AT ( P = 0.02)  
   6 sessions within 2 weeks    
Yurtkuran [ 29 ] (1999)  4 parallel groups, assessor blind 3 100, knee A: formula EA, SP-9, GB-34, ST-34, 35, B: TENS (1) Present pain intensity (5-point scale)  (1)–(5) Changes compared with baseline for EA ( P < 0.05)  
   5 sessions weekly for 2 weeks C: ice massage (2) Stiffness (1)–(5) No differences between treatment groups 
    D: placebo TENS (electrodes disconnected) (3) 50-foot walking time  
     (4) Quadriceps muscle strength  
     (5) Active knee flexion  
Ng [ 34 ] (2003)  3 parallel groups, assessor blind 3 24, knee A: formula EA, ST-35, EX-LE-4, B: TENS (1) Numerical pain rating scale (1) Change in A, B compared with baseline 
   8 sessions in total within 2 weeks C: general education on knee care only (2) Passive range of movement (2) No changes in either group 
     (3) The timed up and go test (TUGT) (3) Change in A compared with baseline 
Ammer [ 24 ] (1988)  2 parallel groups, open 1 28, knee A: formula AT, UB-54, Liv-9, GB-34, 30, 32, ST-36, B: physical therapy (1) Starting pain, walking pain (4-point scale) (1) No intergroup differences 
   2 sessions weekly for 4 weeks, 8 sessions in total  (2) Total pain (4-point scale)  (2) Intergroup difference in favour of control ( P < 0.01)  
Stener-Victorin [ 39 ] (2004)  3 parallel groups, open 3 45, hip A: formula EA, BL-54, 36, GB-29, 30, 31, ST-31, GB-34, BL-60, B: hydrotherapy plus education (1) Pain VAS (related to motion and load) (1) No intergroup differences 
   2 sessions weekly for 5 weeks, C: education alone (2) Daytime pain VAS  (2) Change in A, B compared with baseline ( P <0.05)  
   10 sessions in total plus education  (3) Nighttime pain VAS  (3) Change in A, B compared with baseline ( P < 0.01)  
Haslam [ 32 ] (2001)  2 parallel groups, open 3 32, hip A: formula AT, GB-29, 30, 34, 43, ST-44, LI-4, four ‘ah shi’ points, B: advice plus a set of five exercises (1) modified version of WOMAC (pain, stiffness and physical function)  (1) Intergroup difference ( P = 0.02)  
   1 session weekly for 6 weeks,    
   6 sessions in total    
 
Jia [ 23 ] (2005)  3 parallel groups, open 2 120, knee A: Chinese AT, ST-34, 36, SP-9, 10, GB-34, medial Xiyan, lateral Xiyan, ‘ah shi’ points, B: Chinese AT plus functional exercise (1) effective rate (4-point scale)  (1)(2) No intergroup difference A vs C ( P > 0.05)  
   1 session daily, 45 sessions within 2 months C: functional exercise (2) recurrence rate  (1) Difference in favour of B compared with A and C ( P < 0.01)  
       (2) Difference in favour of B compared with A and C ( P < 0.05)  
Berman [ 31 ] (1999)  Partial crossover, assessor blind 4 73, knee A: Chinese EA, GB-34, SP-9, ST-35, 36, Xiyan, UB-60, GB-39, SP-6, Kid -3, B: conventional therapy alone (oral) (1) WOMAC (total, pain and disability)  (1)–(2) Intergroup differences ( P < 0.001)  
   2 sessions weekly for 8 weeks, 16 sessions in total  (2) Lequesne indices  
Sangdee [ 37 ] (2002)  4 parallel groups, assessor blind 3 193, knee A: formula EA, ST-35, Liv-8, medial Xiyan, trigger point, B: formula EA plus diclofenac, (1) Pain VAS  (1) Intergroup difference ( P < 0.05), A vs C and A vs D  
   3 sessions weekly for 4 weeks, C: diclofenac plus placebo EA (patch electrodes to the selected AT points) (2) WOMAC pain  (5) Intergroup difference ( P < 0.05), A vs D  
   12 sessions in total D: placebo EA plus placebo tablet (3) WOMAC stiffness (3)(4)(6) No intergroup differences 
     (4) WOMAC disability  
     (5) Lequesne's functional index  
     (6) 50 ft walk time  
Tukmachi [ 38 ] (2004)  3 parallel groups, assessor blind 4 30, knee A: Chinese EA, LI-4, SP-9, 10, Xiyan, GB-34, ST-36, LR-3, BL-40, 57, B: Chinese EA plus symptomatic medication (1) Pain VAS  (1)–(3) Intergroup difference ( P < 0.05), A vs C and B vs C at 5 weeks  
   2 sessions weekly for 5 weeks, C: symptomatic medication (2) WOMAC (pain)  
   10 sessions in total  (3) WOMAC (stiffness)  

RCT, randomized controlled trial; ST, stomach; UB or BL, urinary bladder; LI, large intestine; SP, spleen; Kid, kidney; Liv, liver; GB, gall bladder; EX, extraordinary; AT, manual acupuncture; EA, electro-acupuncture; VAS, visual analogue scale; TENS, transcutaneous electrical nerve stimulation.

Study quality

Two sham-controlled studies reported assessor and subject blinding and scored the maximum on the modified Jadad scale [ 33 , 35 ], whereas eight studies scored 4 points [ 26–28 , 30 , 31 , 36 , 38 , 40 ]. Patient blinding was judged to have been achieved in six studies [ 25 , 28 , 30 , 33 , 35 , 36 ]. The assessor was reported as blinded in 13 studies ( Table 1 ) [ 25–31,33–35,37 , 38 , 40 ]. Acupuncture was considered adequate in all but one [ 34 ] trial.

Outcomes

Overall, 10 studies suggested greater pain reduction in acupuncture groups compared with control groups [ 25 , 27 , 30–32,35–38 , 40 ]. Of the 10 high-quality studies scoring 4–5 points [ 26–28 , 30 , 31 , 33 , 35 , 36 , 38 , 40 ], seven trials reported acupuncture to be superior to a range of control interventions [ 27 , 30 , 31 , 35 , 36 , 38 , 40 ]. The extent of heterogeneity in the dataset prevented a meaningful meta-analysis across all trials.

Manual acupuncture was compared with waiting list control groups in two studies, which reported intergroup differences for pain on VAS and WOMAC scales ( Table 1 ). When compared with superficial needling as sham control, intergroup differences on pain measures were reported in three of four trials ( Table 1 ). Three trials provided sufficient data for meta-analysis, which showed a significant effect of manual acupuncture compared with sham for treating peripheral joint OA (standardized mean difference 0.24, 95% CI 0.01–0.47, P = 0.04, n = 329, Fig. 2 ). This was confirmed by the results for knee OA ( Fig. 3 ). Compared with mock TENS, there were no intergroup differences [ 26 ] and compared with conventional therapy the results were mixed [ 23 , 24 , 32 ]. In one case of insufficient data reporting [ 30 ], additional information was requested from the original author but has not been received so far.

F ig . 2.

Meta-analysis of randomized controlled trials (specifics are mentioned at the top of each figure).

F ig . 2.

Meta-analysis of randomized controlled trials (specifics are mentioned at the top of each figure).

F ig . 3.

Meta-analysis of randomized controlled trials (specifics are mentioned at the top of each figure).

F ig . 3.

Meta-analysis of randomized controlled trials (specifics are mentioned at the top of each figure).

Electro-acupuncture was compared against sham or placebo acupuncture ( Table 1 ) in three trials [ 35 , 37 , 40 ], which reported intergroup differences for pain on VAS and WOMAC scales. Electro-acupuncture was compared with conventional therapy in seven trials [ 29 , 31 , 34 , 35 , 37–39 ]. Intergroup differences in favour of electro-acupuncture as compared with oral medication such as diclofenac on pain VAS [ 37 , 38 ], and WOMAC scales [ 31 , 38 ]. Intergroup differences on pain were not reported compared with transcutaneous electrical nerve stimulation (TENS) [ 29 , 34 ] and education [ 34 , 39 ], whereas in another study [ 35 ], contact with the original author confirmed the differences compared with education for WOMAC pain scores. The extent of heterogeneity in trials of electro-acupuncture prevented a meaningful meta-analysis of these data.

Six of the 11 RCTs with sample sizes of ≥40 reported intergroup differences on VAS or WOMAC in favour of acupuncture ( Table 1 ), while four [ 25 , 27 , 32 , 38 ] of the seven studies [ 24–27 , 33 , 34 , 38 ] with sample sizes <40 also reported the same. Eight studies [ 25 , 27 , 30 , 31 , 35 , 36 , 38 , 40 ] of trials indicating deqi sensation, and nine of 14 studies on OA of the knee had intergroup differences on pain measures favouring acupuncture [ 25 , 27 , 30 , 31 , 35–38 , 40 ], whereas one [ 32 ] of three studies of OA of the hip gave a similar result [ 32 ]. Beneficial effects are reported for OA of the thumb joint [ 26 ].

Discussion

A previous review indicated favourable evidence of acupuncture for pain control when compared with sham acupuncture, and concluded that the evidence suggests that acupuncture may play a role in the treatment of knee OA [ 42 ]. Since the publication of this article, a number of further RCTs were published. Whereas the previous review [ 42 ] could not perform meta-analyses due to the numerous types of control groups and insufficient reporting of data, the newly identified trials enabled us to conduct meta-analyses. Our meta-analysis of a homogeneous dataset suggests significant pain reduction for patients with peripheral joint OA, particularly knee OA ( Figs. 2 and 3 ). A best evidence synthesis [ 43 ], which contains a degree of subjectivity, suggests that, overall, the data for manual acupuncture could be classified as a fairly strong evidence.

We opted to assess quality using a modified Jadad scale whereby a point is scored for subject blinding and assessor blinding separately. Of the 18 RCTs, only five trials were both patient-blinded and assessor-blinded [ 25 , 28 , 30 , 33 , 35 ], whereas four studies [ 23 , 24 , 32 , 39 ] did not make an attempt at either subject or assessor blinding. Trials with inadequate levels of blinding are likely to exaggerate the treatment effects and thus limit the reliability of the findings [ 44 , 45 ]. Of course, many other systems to assess methodological quality exist; some of which are more elaborate. However, the Jadad score is an accepted, frequently used and easy to apply score. Jüni et al . [ 46 ] have shown that there are no strong advantages of these instruments over the Jadad score.

Acupuncture was compared with various types of control interventions. Depending on the nature of the control intervention, different conclusions can be drawn. Manual acupuncture seems to reduce pain compared with waiting-list controls. This may suggest effectiveness against the natural course of OA but does not allow for placebo effects. Trials of manual acupuncture were also conducted against sham acupuncture, and the meta-analysis of three RCTs, which provided sufficient data, suggests a positive effect. In addition, five trials [ 30 , 35–37 , 40 ] reported a degree of blinding, suggesting beneficial effects compared with sham or placebo control for VAS or WOMAC pain scores. This indicates analgesic effects of acupuncture beyond a placebo response. There were no intergroup differences compared with placebo TENS [ 26 , 29 ]. Compared with conventional medication, electro-acupuncture was superior to NSAIDs on VAS [ 37 , 38 ] and WOMAC [ 31 , 38 ]. Further trials are required to confirm these data.

The placebo effect of acupuncture as a treatment for pain can be impressive. Several placebo or sham acupuncture methods have been tried. These range from puncturing the skin outside acupuncture points, eliciting a sensation on the skin without puncturing the skin, superficially puncturing the skin without stimulation or in cases of electro-acupuncture, using electro-stimulators with disconnected cables. To reliably account for the placebo effect, it is crucial that the sham procedure is indistinguishable from the real treatment. Therefore, the success of the blinding procedures should be assessed.

This systematic review has several limitations. We cannot be absolutely certain that our searches located all relevant RCTs, which is a limitation that indeed applies to systematic reviews in general. Although strong efforts were made to retrieve all RCTs on the subject, it is conceivable that some were not found. In this study, a large number of different databases were searched and there were no restrictions in terms of publication language. Further limitations include the paucity and the often suboptimal quality of the primary data. It must be noted that design features such as placebo or blinding are difficult to incorporate for acupuncture trials and that research funds for acupuncture studies are in short supply. Even though the total number of 18 RCTs is encouraging, it is too small considering the heterogeneity of the overall dataset.

One argument for using acupuncture in the management of painful musculoskeletal conditions is that it is safer than standard drug treatments [ 10 ]. Serious adverse effects of acupuncture have been reported, although these may be rare [ 47 ]. Relative to those of standard drug treatments these may be infrequent or even negligible [ 2 , 48 ]. Several prospective studies have shown that mild adverse effects after acupuncture occur in about 7% of all cases [ 49 ].

In conclusion, sham-controlled RCTs suggest specific effects of acupuncture for pain control in patients with peripheral joint OA. Considering its favourable safety profile, acupuncture seems an option worthy of consideration particularly for knee OA. Further studies are required particularly for manual or electro-acupuncture for hip OA.

graphic

Acknowledgement

Y.D.K. was supported by a Soongsan grant from Wonkwang University, South Korea in 2005.

The authors have declared no conflict of interest.

References

1
Felson
DT
Zhang
Y
An update on the epidemiology of knee and hip osteoarthritis with a view to prevention
Arthritis Rheum
 , 
1998
, vol. 
41
 (pg. 
1343
-
55
)
2
Ernst
E
Acupuncture as a symptomatic treatment of osteoarthritis: a systematic review
Scand J Rheumatol
 , 
1997
, vol. 
26
 (pg. 
444
-
7
)
3
Jordan
KM
Arden
NK
Doherty
M
, et al.  . 
EULAR Recommendations 2003: an evidence based approach to the management of knee osteoarthritis: report of a Task Force of the Standing Committee for International Clinical Studies Including Therapeutic Trials (ESCISIT)
Ann Rheum Dis
 , 
2003
, vol. 
62
 (pg. 
1145
-
55
)
4
Felson
DT
Lawrence
RC
Dieppe
PA
, et al.  . 
Osteoarthritis: new insights. Part 1: the disease and its risk factors
Ann Intern Med
 , 
2000
, vol. 
133
 (pg. 
635
-
46
)
5
Felson
DT
Lawrence
RC
Hochberg
MC
, et al.  . 
Osteoarthritis: new insights: Part 2: treatment approaches
Ann Intern Med
 , 
2000
, vol. 
133
 (pg. 
726
-
37
)
6
Griffin
MR
Ray
WA
Schaffner
W
Nonsteroidal anti-inflammatory drug use and death from peptic ulcer in elderly persons
Ann Intern Med
 , 
1988
, vol. 
109
 (pg. 
359
-
63
)
7
Eisenberg
DM
Davis
RB
Ettner
SL
, et al.  . 
Trends in alternative medicine use in the United States, 1990–1997: results of a follow-up national survey
JAMA
 , 
1998
, vol. 
280
 (pg. 
1569
-
75
)
8
Tindle
HA
Davis
RB
Phillips
RS
Eisenberg
DM
Trends in use of complementary and alternative medicine by US adults: 1997–2002
Altern Ther Health Med
 , 
2005
, vol. 
11
 (pg. 
42
-
9
)
9
Jadad
AR
Moore
RA
Carroll
D
, et al.  . 
Assessing the quality of reports of randomized clinical trials: is blinding necessary?
Control Clin Trials
 , 
1996
, vol. 
17
 (pg. 
1
-
12
)
10
White
AR
Ernst
E
A systematic review of randomized controlled trials of acupuncture for neck pain
Rheumatology
 , 
1999
, vol. 
38
 (pg. 
143
-
7
)
11
Gaw
AC
Chang
LW
Shaw
LC
Efficacy of acupuncture on osteoarthritic pain. A controlled, double-blind study
N Engl J Med
 , 
1975
, vol. 
293
 (pg. 
375
-
8
)
12
Junnila
SY
Acupuncture superior to piroxicam in the treatment of osteoarthrosis
Am J Acupunct
 , 
1982
, vol. 
10
 (pg. 
341
-
6
)
13
Lundeberg
T
Eriksson
SV
Lundeberg
S
Thomas
M
Effect of acupuncture and naloxone in patients with osteoarthritis pain. A sham acupuncture controlled study
The Pain Clinic
 , 
1991
, vol. 
4
 (pg. 
155
-
61
)
14
Thomas
M
Eriksson
SV
Lundeberg
T
A comparative study of diazepam and acupuncture in patients with osteoarthritis pain: a placebo controlled study
Am J Chin Med
 , 
1991
, vol. 
19
 (pg. 
95
-
100
)
15
McIndoe
AK
Young
K
Bone
ME
A comparison of acupuncture with intra-articular steroid injection as analgesia for osteoarthritis of the hip
Acupunct Med
 , 
1995
, vol. 
13
 (pg. 
67
-
70
)
16
Fargas Babjak
AM
Pomeranz
B
Rooney
PJ
Acupuncture-like stimulation with codetron for rehabilitation of patients with chronic pain syndrome and osteoarthritis
Acupunct Electrother Res
 , 
1992
, vol. 
17
 (pg. 
95
-
105
)
17
Berman
BM
Lao
L
Greene
M
, et al.  . 
Efficacy of traditional Chinese acupuncture in the treatment of symptomatic knee osteoarthritis: a pilot study
Osteoarthritis Cartilage
 , 
1995
, vol. 
3
 (pg. 
139
-
42
)
18
Fink
MG
Kunsebeck
HW
Wippermann
B
Effects of acupuncture on pain and impairement in patients with osteoarthritis of the hip
Z Rheumatol
 , 
2000
, vol. 
59
 (pg. 
191
-
9
)
19
Singh
BB
Berman
BM
Hadhazy
V
, et al.  . 
Clinical decisions in the use of acupuncture as an adjunctive therapy for osteoarthritis of the knee
Altern Ther Health Med
 , 
2001
, vol. 
7
 (pg. 
58
-
65
)
20
Fink
M
Karst
M
Wippermann
B
Gehrke
A
Acupuncture as a complementary treatment in physical medicine: its use in osteoarthritis of the hip
Physikalische Medizin Rehabilitationsmedizin Kurortmedizin
 , 
2001
, vol. 
11
 (pg. 
123
-
8
)
21
Tillu
A
Roberts
C
Tillu
S
Unilateral versus bilateral acupuncture on knee function in advanced osteoarthritis of the knee – a prospective randomised trial
Acupunct Med
 , 
2001
, vol. 
19
 (pg. 
15
-
8
)
22
Vas
J
Perea Milla
E
Méndez
C
Acupuncture and moxibustion as an adjunctive treatment for osteoarthritis of the knee – a large case series
Acupunct Med
 , 
2004
, vol. 
22
 (pg. 
23
-
8
)
23
Jia
J
Mao
GL
Hu
SH
Dong
XC
Acupuncture combined with function exercise for the elder patients with knee osteoarthritis
Chin J Clin Rehab
 , 
2005
, vol. 
9
 (pg. 
18
-
9
)
24
Ammer
K
Petschnig
R
Vergleich der Wirksamkeit von Akupunktur und Physikotherapie bei ambulanten Gonarthrosepatienten
Wien Med Wochenschr
 , 
1988
, vol. 
138
 (pg. 
566
-
9
)
25
Petrou
P
Winkler
V
Genti
G
Balint
G
Double-blind trial to evaluate the effect of acupuncture treatment on knee osteoarthrosis
Scand J Acupunct
 , 
1988
, vol. 
3
 (pg. 
112
-
5
)
26
Dickens
W
Lewith
GT
A single-blind, controlled and randomised clinical trial to evaluate the effect of acupuncture in the treatment of trapezio-metacarpal osteoarthritis
Comp Med Res
 , 
1989
, vol. 
3
 (pg. 
5
-
8
)
27
Christensen
BV
Iuhl
IU
Vilbek
H
Bulow
HH
Dreijer
NC
Rasmussen
HF
Acupuncture treatment of severe knee osteoarthritis. A long-term study
Acta Anaesthesiol Scand
 , 
1992
, vol. 
36
 (pg. 
519
-
25
)
28
Takeda
W
Wessel
J
Acupuncture for the treatment of pain of osteoarthritic knees
Arthritis Care Res
 , 
1994
, vol. 
7
 (pg. 
118
-
22
)
29
Yurtkuran
M
Kocagil
T
TENS, electroacupuncture and ice massage: comparison of treatment for osteoarthritis of the Knee
Am J Acupunct
 , 
1999
, vol. 
27
 (pg. 
133
-
40
)
30
Molsberger
A
Böwing
G
Jensen
KU
Lorek
M
Schmerztherapie mit Akupunktur bei Gonarthrose
Der Schmerz
 , 
1994
, vol. 
8
 (pg. 
37
-
42
)
31
Berman
BM
Singh
BB
Lao
L
, et al.  . 
A randomized trial of acupuncture as an adjunctive therapy in osteoarthritis of the knee
Rheumatology
 , 
1999
, vol. 
38
 (pg. 
346
-
54
)
32
Haslam
R
A comparison of acupuncture with advice and exercises on the symptomatic treatment of osteoarthritis of the hip – a randomised controlled trial
Acupunct Med
 , 
2001
, vol. 
19
 (pg. 
19
-
26
)
33
Fink
MG
Kunsebeck
H
Wipperman
B
Gehrke
A
Non-specific effects of traditional Chinese acupuncture in osteoarthritis of the hip
Complement Ther Med
 , 
2001
, vol. 
9
 (pg. 
82
-
9
)
34
Ng
MML
Leung
MCP
Poon
DMY
The effects of electro-acupuncture and transcutaneous electrical nerve stimulation on patients with painful osteoarthritic knees: a randomized controlled trial with follow-up evaluation
J Altern Complement Med
 , 
2003
, vol. 
9
 (pg. 
641
-
9
)
35
Berman
BM
Lao
L
Langenberg
P
Lee
WL
Gilpin
AMK
Hochberg
MC
Effectiveness of acupuncture as adjunctive therapy in osteoarthritis of the knee. A randomized, controlled trial
Ann Intern Med
 , 
2004
, vol. 
141
 (pg. 
901
-
10
)
36
Witt
C
Brinkhaus
B
Jena
S
, et al.  . 
Acupuncture in patients with osteoarthritis of the knee: a randomised trial
Lancet
 , 
2005
, vol. 
366
 (pg. 
136
-
43
)
37
Sangdee
C
Teekachunhatean
S
Sananpanich
K
, et al.  . 
Electroacupuncture versus Diclofenac in symptomatic treatment of osteoarthritis of the knee: a randomized controlled trial
BMC Complement Altern Med
 , 
2002
, vol. 
2
 
3
38
Tukmachi
E
Jubb
R
Dempsey
E
Jones
P
The effect of acupuncture on the symptoms of knee osteoarthritis – an open randomised controlled study
Acupunct Med
 , 
2004
, vol. 
22
 (pg. 
14
-
22
)
39
Stener-Victorin
E
Kruse-Smidje
C
Jung
K
Comparison between electro-acupuncture and hydrotherapy, both in combination with patient education and patient education alone, on the symptomatic treatment of osteoarthritis of the hip
Clin J Pain
 , 
2004
, vol. 
20
 (pg. 
179
-
85
)
40
Vas
J
Méndez
C
Perea Milla
E
, et al.  . 
Acupuncture as a complementary therapy to the pharmacological treatment of osteoarthritis of the knee: randomised controlled trial
Br Med J
 , 
2004
, vol. 
329
 (pg. 
1216
-
20
)
41
Tillu
A
Tillu
S
Vowler
S
Effect of acupuncture on knee function in advanced osteoarthritis of the knee: a prospective, non-randomised controlled study
Acupunct Med
 , 
2002
, vol. 
20
 (pg. 
19
-
21
)
42
Ezzo
J
Hadhazy
V
Birch
S
Lao
L
Kaplan
G
Hochberg
M
Berman
B
Acupuncture for osteoarthritis of the knee: a systematic review
Arthritis Rheum
 , 
2001
, vol. 
44
 (pg. 
819
-
25
)
43
van Tulder
MW
Cherkin
DC
Berman
B
Lao
L
Koes
BW
The effectiveness of acupuncture in the management of acute and chronic low back pain. A systematic review within the framework of the Cochrane Collaboration Back Review Group
Spine
 , 
1999
, vol. 
24
 (pg. 
1113
-
23
)
44
Schulz
KF
Chalmers
I
Hayes
RJ
Altman
DG
Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials
JAMA
 , 
1995
, vol. 
273
 (pg. 
408
-
12
)
45
Day
SJ
Altman
DG
Statistics notes: blinding in clinical trials and other studies
Br Med J
 , 
2000
, vol. 
321
 (pg. 
504
-
44
)
46
Jüni
P
Witschi
A
Bloch
R
Egger
M
The hazards of scoring the quality of clinical trials for meta-analysis
JAMA
 , 
1999
, vol. 
282
 (pg. 
1054
-
60
)
47
Ernst
E
White
A
Acupuncture: safety first
Br Med J
 , 
1997
, vol. 
314
 pg. 
1362
 
48
Garcia Rodriguez
L
Hernandes-Diaz
S
The risk of upper gastrointestinal complications associated with nonsteroidal anti-inflammatory drugs, glucocorticoids, acetaminophen, and combinations of these agents
Arthritis Res
 , 
2001
, vol. 
3
 (pg. 
98
-
101
)
49
Melchart
D
Weidenhammer
W
Streng
A
, et al.  . 
Prospective investigation of adverse effects of acupuncture in 97733 patients
Arch Intern Med
 , 
2004
, vol. 
164
 (pg. 
104
-
5
)

Comments

0 Comments