Menu Close

Acupuncture for Chronic Pain Individ. Patient Data Meta-analysis

Andrew J. Vickers, DPhil; Angel M. Cronin, MS; Alexandra C. Maschino, BS; et al                        


Background Although acupuncture is widely used for  chronic pain, there remains considerable controversy as to its value. We  aimed to determine the effect size of acupuncture for 4 chronic pain  conditions: back and neck pain, osteoarthritis, chronic headache, and  shoulder pain.

Methods We conducted a systematic  review to identify randomized controlled trials (RCTs) of acupuncture  for chronic pain in which allocation concealment was determined  unambiguously to be adequate. Individual patient data meta-analyses were  conducted using data from 29 of 31 eligible RCTs, with a total of  17 922 patients analyzed.

Results In the primary  analysis, including all eligible RCTs, acupuncture was superior to both  sham and no-acupuncture control for each pain condition (P < .001  for all comparisons). After exclusion of an outlying set of RCTs that  strongly favored acupuncture, the effect sizes were similar across pain  conditions. Patients receiving acupuncture had less pain, with scores  that were 0.23 (95% CI, 0.13-0.33), 0.16 (95% CI, 0.07-0.25), and 0.15  (95% CI, 0.07-0.24) SDs lower than sham controls for back and neck pain,  osteoarthritis, and chronic headache, respectively; the effect sizes in  comparison to no-acupuncture controls were 0.55 (95% CI, 0.51-0.58),  0.57 (95% CI, 0.50-0.64), and 0.42 (95% CI, 0.37-0.46) SDs. These  results were robust to a variety of sensitivity analyses, including  those related to publication bias.

Conclusions Acupuncture  is effective for the treatment of chronic pain and is therefore a  reasonable referral option. Significant differences between true and  sham acupuncture indicate that acupuncture is more than a placebo.  However, these differences are relatively modest, suggesting that  factors in addition to the specific effects of needling are important  contributors to the therapeutic effects of acupuncture.

Acupuncture is the insertion and stimulation of needles  at specific points on the body to facilitate recovery of health.  Although initially developed as part of traditional Chinese medicine,  some contemporary acupuncturists, particularly those with medical  qualifications, understand acupuncture in physiologic terms, without  reference to premodern concepts.1

An estimated 3 million American adults receive acupuncture treatment each year,2 and chronic pain is the most common presentation.3 Acupuncture is known to have physiologic effects relevant to analgesia,4,5 but there is no accepted mechanism by which it could have persisting  effects on chronic pain. This lack of biological plausibility, and its  provenance in theories lying outside of biomedicine, makes acupuncture a  highly controversial therapy.

A large number of randomized controlled trials (RCTs) of  acupuncture for chronic pain have been conducted. Most have been of low  methodologic quality, and, accordingly, meta-analyses based on these  RCTs are of questionable interpretability and value.6 Herein, we present an individual patient data meta-analysis of RCTs of  acupuncture for chronic pain, in which only high-quality RCTs were  eligible for inclusion. Individual patient data meta-analysis are  superior to the use of summary data in meta-analysis because they  enhance data quality, enable different forms of outcome to be combined,  and allow use of statistical techniques of increased precision.


The full protocol of the meta-analysis has been published.6 In brief, the study was conducted in 3 phases: identification of  eligible RCTs; collection, checking, and harmonization of raw data; and  individual patient data meta-analysis.

                                Data sources and searches                                

To identify articles, we searched MEDLINE, the Cochrane  Collaboration Central Register of Controlled Trials, and the citation  lists of systematic reviews (the full search strategy is shown in the eAppendix).  There were no language restrictions. The initial search, current to  November 2008, was used to identify studies for the individual patient  data meta-analysis; a second search was conducted in December 2010 for  summary data to use in a sensitivity analysis.

                                Study selection                                

Two reviewers applied inclusion criteria for potentially  eligible articles separately, with disagreements about study inclusion  resolved by consensus. Randomized controlled trials were eligible for  analysis if they included at least 1 group receiving acupuncture  needling and 1 group receiving either sham (placebo) acupuncture or  no-acupuncture control. The RCTs must have accrued patients with 1 of 4  indications—nonspecific back or neck pain, shoulder pain, chronic  headache, or osteoarthritis—with the additional criterion that the  current episode of pain must be of at least 4 weeks duration for  musculoskeletal disorders. There was no restriction on the type of  outcome measure, although we specified that the primary end point must  be measured more than 4 weeks after the initial acupuncture treatment.

It has been demonstrated that unconcealed allocation is the most important source of bias in RCTs,7 and, as such, we included only those RCTs in which allocation  concealment was determined unambiguously to be adequate (further details  are in the review protocol6).  Where necessary, we contacted authors for further information  concerning the exact logistics of the randomization process. We excluded  RCTs if there was any ambiguity about allocation concealment.

                                Data extraction and quality assessment                                

The principal investigators of eligible studies were  contacted and asked to provide raw data from the RCT. To ensure data  accuracy, all results reported in the RCT publication, including  baseline characteristics and outcome data, were then replicated.

Reviewers assessed the quality of blinding for eligible  RCTs with sham acupuncture control. The RCTs were graded as having a low  likelihood of bias if either the adequacy of blinding was checked by  direct questioning of patients (eg, by use of a credibility  questionnaire) and no important differences were found between groups,  or the blinding method (eg, the Streitberger and Kleinhenz sham device8)  had previously been validated as able to maintain blinding. Randomized  controlled trials with a high likelihood of bias from unblinding were  excluded from the meta-analysis of acupuncture vs sham; a sensitivity  analysis included only RCTs with a low risk of bias.

                                Data synthesis and analysis                                

Each RCT was reanalyzed by analysis of covariance with  the standardized principal end point (scores divided by pooled standard  deviation) as the dependent variable, and the baseline measure of the  principal end point and variables used to stratify randomization as  covariates. This approach has been shown to have the greatest  statistical power for RCTs with baseline and follow-up measures.9,10 The effect size for acupuncture from each RCT was then entered into a meta-analysis using the metan command in Stata software (version 11; Stata Corp): the meta-analytic  statistics were created by weighting each coefficient by the reciprocal  of the variance, summing, and dividing by the sum of the weights.  Meta-analyses were conducted separately for comparisons of acupuncture  with sham and no-acupuncture control, and within each pain type. We  prespecified that the hypothesis test would be based on the fixed  effects analysis because this constitutes a valid test of the null  hypothesis of no treatment effect.

                                Results                                                                 Systematic review                                

We identified 82 RCTs (Figure 1),1193 of which 31 were eligible (Table 1 and eAppendix). Four of the studies were organized as part of the German Acupuncture Trials (GERAC) initiative,1114 4 were part of the Acupuncture Randomized Trials (ART) group1518; 4 were Acupuncture in Routine Care (ARC) studies1922; 3 were UK National Health Service acupuncture RCTs.23,24,98 Eleven studies were sham controlled, 10 had no-acupuncture control, and  10 were 3-armed studies, including both sham and no-acupuncture  control. The second search for subsequently published studies identified  an additional 4 eligible studies,9497 with a total of 1619 patients.

An important source of clinical heterogeneity between  studies concerns the control groups. In the sham RCTs, the type of sham  included acupuncture needles inserted superficially,13 sham acupuncture devices with needles that retract into the handle rather than penetrate the skin,25 and nonneedle approaches, such as deactivated electrical stimulation26 or detuned laser.27 Moreover, cointerventions varied, with no additional treatment other than analgesics in some RCTs,15 whereas in other RCTs, both acupuncture and sham groups received a  course of additional treatment, such as exercise led by physical  therapists.24 Similarly, the no-acupuncture control groups varied among usual care,  such as an RCT in which control group patients were merely advised to  “avoid acupuncture”98; attention control, such as group education sessions28; and guidelined care, in which patients were given advice as to specific drugs and doses.13

                                Data extraction and quality assessment                                

Usable raw data were obtained from 29 of the 31 eligible  RCTs, including a total of 17 922 patients from the United States,  United Kingdom, Germany, Spain and Sweden. For 1 RCT, the study database  had become corrupted29;  in another case, the statisticians involved in the RCT failed to  respond to repeated enquiries despite approval for data sharing being  obtained from the principal investigator.30

The 29 RCTs comprised 18 comparisons with 14 597  patients of acupuncture with no-acupuncture group and 20 comparisons  with 5230 patients of acupuncture and sham acupuncture. Patients in all  RCTs had access to analgesics and other standard treatments for pain.  Four sham RCTs were determined to have an intermediate likelihood of  bias from unblinding13,27,31,32;  the 16 remaining sham RCTs were graded as having a low risk of bias  from unblinding. On average, dropout rates were low (weighted mean,  10%). Dropout rates were higher than 25% for only 4 RCTs: those by  Molsberger et al30,97 (27% and 33%, respectively, but raw data were not received and neither RCT included in main analysis); Carlsson et al32 (46%, RCT excluded in a sensitivity analysis for blinding), and Berman et al28 (31%). This RCT had a high dropout rate among no-acupuncture controls  (43%); dropout rates were close to 25% in the acupuncture and sham  groups. The RCT by Kerr et al31 had a large difference in dropout rates between groups (acupuncture,  13%; control, 33%) but was excluded in the sensitivity analysis for  blinding.


Forest plots for acupuncture against sham acupuncture  and against no-acupuncture control are shown separately for each of the 4  pain conditions in Figure 2 and Figure 3. Meta-analytic statistics are shown in Table 2. Acupuncture was statistically superior to control for allanalyses (P < .001).  Effect sizes are larger for the comparison between acupuncture and  no-acupuncture control than for the comparison between acupuncture and  sham: 0.37, 0.26, and 0.15 in comparison with sham vs 0.55, 0.57, and  0.42 in comparison with no-acupuncture control for musculoskeletal pain,  osteoarthritis, and chronic headache, respectively.

For 5 of the 7 analyses, the test for heterogeneity was  statistically significant. In the case of comparisons with sham  acupuncture, the RCTs by Vas et al37,38,41 are clear outliers. For example, the effect size of the RCTs by Vas et  al for neck pain is about 5 times greater than meta-analytic estimate.  One effect of excluding these RCTs in a sensitivity analysis (Table 3 and Table 4)  is that there is no significant heterogeneity in the comparisons  between acupuncture and sham. Moreover, the effect size for acupuncture  becomes relatively similar for the different pain conditions: 0.23,  0.16, and 0.15 against sham, and 0.55, 0.57, and 0.42 against  no-acupuncture control for back and neck pain, osteoarthritis, and  chronic headache, respectively (fixed effects; results similar for the  random effects analysis).

To give an example of what these effect sizes mean in  real terms, a baseline pain score on a 0 to 100 scale for a typical RCT  might be 60. Given a standard deviation of 25, follow-up scores might be  43 in a no-acupuncture group, 35 in a sham acupuncture group, and 30 in  patients receiving true acupuncture. If response were defined in terms  of a pain reduction of 50% or more, response rates would be  approximately 30%, 42.5%, and 50%, respectively.

The comparisons with no-acupuncture control show  evidence of heterogeneity. This seems largely explicable in terms of  differences between the control groups used. In the case of  osteoarthritis, the largest effect was in the study by Witt et al,17 in which patients in the waiting list control received only rescue pain  medication, and the smallest was in the study by Foster et al,24 which involved a program of exercise and advice led by physical  therapists. For the musculoskeletal analyses, heterogeneity is driven by  2 very large RCTs19,20 (n = 2565 patients and n = 3118 patients, respectively) for back and neck pain. If only back pain is considered (Table 3 and Table 4), heterogeneity is dramatically reduced and is again driven by one RCT, by Brinkhaus et al,15 with waiting list control. In the headache meta-analysis, Diener et al13 had much smaller differences between groups. This RCT involved  providing drug therapy according to national guidelines in the  no-acupuncture group, including initiation of β-blockers as migraine  prophylaxis. There was disagreement within the collaboration about  whether this constituted active control. Excluding this RCT reduced  evidence of heterogeneity (P = .04) but had little effect on the effect size (0.42-0.45).

Table 3 and Table 4 show several prespecified sensitivity analyses. Neither restricting the  sham RCTs to those with low likelihood of unblinding nor adjustment for  missing data had any substantive effect on our main estimates.  Inclusion of summary data from RCTs for which raw data were not obtained  (2 RCTs) or which were published recently (4 RCTs) also had little  impact on either the primary analysis (Table 3 and Table 4) or the analysis with the outlying RCTs by Vas et al37,38,41 excluded (data not shown).

To estimate the potential impact of publication bias, we  entered all RCTs into a single analysis and compared the effect sizes  from small and large studies.99 We saw some evidence that small studies had larger effect sizes for the comparison with sham (P = .02) but not no-acupuncture control (P = .72). However, these analyses are influenced by the outlying RCTs by Vas et al,37,38,41 which were smaller than average, and by indication, because the  shoulder pain RCTs were small and had large effect sizes. Tests for  asymmetry were nonsignificant when we excluded the RCTs by Vas et al37,38,41 and shoulder pain studies (n = 15; P = .07) and when small studies were also excluded (n < 100 and n = 12, respectively; P = .30).  Nonetheless, we repeated our meta-analyses excluding RCTs with a sample  size of less than 100. This had essentially no effect on our results.  As a further test of publication bias, we considered the possible effect  on our analysis if we had failed to include high-quality, unpublished  studies. Only if there were 47 unpublished RCTs with n = 100 patients  showing an advantage to sham of 0.25 SD would the difference between  acupuncture and sham lose significance.

A final sensitivity analysis examined the effect of  pooling different end points measured at different periods of follow-up.  We repeated our analyses including only pain end points measured at 2  to 3 months after randomization. There was no material effect on  results: effect sizes increased by 0.05 to 0.09 SD for musculoskeletal  and osteoarthritis RCTs and were stable otherwise.

As an exploratory analysis, we compared sham control with no-acupuncture control. In a meta-analysis of 9 RCTs,1113,1518,24,28 the effect size for sham was 0.33 (95% CI, 0.27-0.40) and 0.38 (95% CI,  0.20-0.56) for fixed and random effects models, respectively (P < .001 for tests of both effect and heterogeneity).

Leave a Reply