Intervention Review

You have free access to this content

Acupuncture for treating fibromyalgia

  1. John C Deare1,
  2. Zhen Zheng2,*,
  3. Charlie CL Xue2,
  4. Jian Ping Liu3,
  5. Jingsheng Shang4,
  6. Sean W Scott5,
  7. Geoff Littlejohn6

Editorial Group: Cochrane Musculoskeletal Group

Published Online: 31 MAY 2013

Assessed as up-to-date: 15 JAN 2012

DOI: 10.1002/14651858.CD007070.pub2


How to Cite

Deare JC, Zheng Z, Xue CCL, Liu JP, Shang J, Scott SW, Littlejohn G. Acupuncture for treating fibromyalgia. Cochrane Database of Systematic Reviews 2013, Issue 5. Art. No.: CD007070. DOI: 10.1002/14651858.CD007070.pub2.

Author Information

  1. 1

    Compmed Health Institute, Southport, Queensland, Australia; and Traditional & Complementary Medicine Program, Health Innovations Research Institute, Discipline of Chinese Medicine, School of Health Sciences, RMIT University, Bundoora, Victoria, Australia

  2. 2

    RMIT University, Traditional & Complementary Medicine Research Program, Health Innovations Research Institute and Discipline of Chinese Medicine, School of Health Sciences, Bundoora, Victoria, Australia

  3. 3

    Beijing University of Chinese Medicine, Centre for Evidence-Based Chinese Medicine, Beijing, China

  4. 4

    Akupunktoren Bruno Aamo AS, Tromsoe, Norway

  5. 5

    Royal North Shore Hospital, Department of Emergency Medicine, St Leonards, New South Wales, Australia

  6. 6

    Monash Medical Centre, Department of Rheumatology, Clayton, VIC, Australia

*Zhen Zheng, Traditional & Complementary Medicine Research Program, Health Innovations Research Institute and Discipline of Chinese Medicine, School of Health Sciences, RMIT University, PO Box 71, Bundoora, Victoria, 3083, Australia. zhen.zheng@rmit.edu.au.

Publication History

  1. Publication Status: New
  2. Published Online: 31 MAY 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Assefi 2005

MethodsRandomised: computer-generated blocked random allocation sequence with block size of 4. Researcher not involved in the study conducted randomisation (used academic research centre).

Blinding: participants blinded during treatment. Staff who collected and analysed the data were blinded to treatment group. Care givers were not blinded to group allocation.

Setting: individual private offices, does not state where these are

Was study aim clear: yes

Informed consent: verbal and written

Ethics approval: institutional review boards at participating institution

WHO clinical trial register: listed and outcomes as per listing

Intention-to-treat used: yes

Follow-up: 3 and 6 months


ParticipantsTotal number of participants: 100 participants, mean duration of illness (years): directed acupuncture: 6 years SD 5; sham control groups: acupuncture for unrelated condition 5 years SD 3; sham needling 7 years SD 6; simulated acupuncture 7 years SD 4

2 male and 94 female

Mean age: directed acupuncture; 46 years SD 11; sham control groups: acupuncture for unrelated condition 46 years SD 11; sham needling 49 years SD 14; simulated acupuncture 48 years SD 10

Diagnosis: ACR

Acupuncture-naive participants: yes

Excluded: other pain conditions, contraindicated for acupuncture (bleeding disorders, severe needle phobia), pregnant or breastfeeding, use of narcotics, litigation and previous acupuncture treatments

Recruitment source: the Greater Seattle, Washington State metropolitan area using newspaper, television, university-affiliated hospitals, local fibromyalgia support groups and health care providers

Previous treatments: manual (physical, ergonometric, chiropractic, massage), mental health therapies (psychotherapy, cognitive behavioural therapy), dietary changes or other (nerve blocks, hypnosis or biofeedback)


Interventions1) Real: directed acupuncture

Randomised to this group: 25 (analysed 25)

2) Control: acupuncture for unrelated condition, treating for irregular menses or early menses due to Blood Heat

Randomised to this group: 25 (analysed 25)

3) Control: sham needling, using body points not recognised as true acupuncture points

Randomised to this group 24 (1 did not complete baseline questionnaire) (analysed 24)

4) Control: simulated acupuncture; same acupuncture points as directed acupuncture but with toothpick inside a needle guide tube to mimic needle insertion/withdrawal

Randomised to this group 25 (analysed 25)

Minimum number of treatments needed: possible 24 treatments, required to attend 80% (19/24)

Co-interventions: maintain current use of pharmacological and non-pharmacological therapies through out the study

Acupuncturists: 8 US trained and licensed with median of 10 years experience (range 4 to 18 years)

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcomes:

1) Pain; visual analogue scale (VAS) (0 = no pain, 10 = worst ever)

2) Function: Short-Form 36 health survey, mean of 50 and standard deviation of 10, with higher scores indicating better functioning

Secondary outcomes:

3) Fatigue: VAS (0 = none, 10 = worst ever)

4) Sleep: VAS (0 = worst ever, 10 = best ever)

5) Over well-being: VAS (0 = worst ever, 10 = best ever)

Outcome measures primary and secondary: 1), 2), 3), 4), 5) taken at weeks 1, 4, 8, 12 and 3 and 6 months

6) Blinding: participants rated how certain they were that they had received directed acupuncture or stimulated acupuncture on a 7-point scale (1 = very sure, 7 = very uncertain) measured at 12 weeks

7) Acupuncturist: participants rated acupuncturist skill level (1 = high, 7 = low) and adverse events measured at weeks 1, 4, 8 and 12

8) Other co-interventions: medication use measured at week 1 and week 12

Outcome measure results:

"No significant differences were detected between the directed acupuncture and the pooled control group for any of the study outcomes".

For the blinding procedures, "32% believed they were receiving acupuncture specifically designed for FM"; no significant difference between the groups (P > 0.2). 4% believed they were receiving simulated acupuncture; no difference between the groups (P > 0.2).

Combined groups had no significant difference in the skill of the acupuncturist (P > 0.2). 77% rated their skill as high, 5% as medium, 17% did not know

84% in the directed group and 79% of the pooled sham groups completed the full course of treatment. On average 21/24 (P > 0.2).

Total medication use showed no significant difference between groups (P > 0.2). Most commonly used medications were ibuprofen, acetaminophen and naproxen.

Withdrawals/drop-outs: 4 in at the randomised stage, 10 at the allocated intervention stage, directed group 2, unrelated acupuncture treatment 2, sham acupuncture 2 and simulated acupuncture 4

Complications/adverse events: 89 participants reported adverse events. 37% reported discomfort at site of needle insertion or simulation of needles, 3% reported nausea, 0.3% felt faint. Participants in simulated acupuncture (39%) had less discomfort than directed acupuncture (61%), while unrelated acupuncture (70%) and sham acupuncture (64%) were similar to directed. Bruising was reported less in the simulated acupuncture group (10%), while directed acupuncture (52%) it was reported more, the unrelated treatments (74%) was the worse, with sham acupuncture (68%) being similar to the last 2 groups. 

Data extraction methods: data were extracted from the published paper. We selected the 'directed acupuncture' group as the real acupuncture treatment and we combined all the control arms as the sham acupuncture control as per the Cochrane Handbook and confirmed with the Cochrane editors. Mean data were measured from Figure 2 in the published data and SD was taken from baseline as this was not provided in the figure and was not published anywhere else. For the comparison of invasive and non-invasive sham controls we used the 'simulated acupuncture' as it was the same tool (Sherman 2002) as used in both the Harris 2008 and Harris 2009 studies. We could not extract data from the SF-36 graphs (Figure 3) as both graphs were labelled as SF-36 Physical Component.


NotesOther info: 1 author was contacted by e-mail and confirmed intention-to-treat, however no further information was given about data or in response to other questions we had

Refunded costs to participants: not reported

Funding: NCCAM

Language: English

Publication: full paper


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated random number

Allocation concealment (selection bias)Low riskBlocked random allocation sequence with block size of 4 with an independent researcher advising the acupuncture clinic of treatment assignment

Incomplete outcome data (attrition bias)
All outcomes
Low riskData from 7 participants were not included in the analysis. However, the missing outcome data were balanced across the groups and less likely to have impacted on the outcome.

Selective reporting (reporting bias)Low riskAll outcomes reported, as per WHO clinical trials register

Blinding of participants and personnel (performance bias)
All outcomes
Low riskRestricted conversation during treatment, participants blindfolded and used acupuncture-naive participants. Care giver not blinded to group allocation. Participants tested for blinding could not detect which group they belonged to at the end of treatment.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskData collection staff and data analysts were blinded to treatment group

Deluze 1992

MethodsRandomised: electronic number generator, closed envelopes, numbered 1 to 70, prepared before study and opened in numerical order after recruitment

Blinding: participants and outcomes assessors were blinded. Care giver was NOT blinded to group allocation.

Setting: University Hospital, Geneva, Switzerland

Was study aim clear: yes

Informed consent: verbal and written

Ethics approval: Department of Medicine ethics committee

WHO clinical trial register: not listed as had not been established when study undertaken

Follow-up: none

Intention-to-treat used: not stated, but in the results section it states that the 15 participants that withdraw were not re-evaluated. Martin 2006 in their journal article stated that analysis used intention-to-treat, yet the review by Berman 1999 using intention-to-treat, found 42% had no benefit, 39% had satisfactory benefit, while 19% had an unexpectedly large benefit. In a recent review this detail was omitted by (Mayhew 2007). Assefi 2005 and Harris 2005 also did not mention this point in their discussion about other studies.


ParticipantsTotal number of participants: 70 participants, mean duration of disease (years); real acupuncture = 14.4 years (3.7) (6.9 to 22.0), control = 6.9 years (1.3) (4.3 to 9.6)

16 male and 54 female (excess of men in the control group P = 0.015)

Mean age (years): real acupuncture = 46.8, control = 49

Diagnosis: ACR

Acupuncture-naive participants: yes

Excluded: severe concomitant disease, use of morphine-like drugs or anticoagulants, peripheral neuropathy, bleeding disorders, language difficulties and past use of acupuncture

Recruitment source: referred, but does not state where from

Previous treatments: not reported


Interventions1) Real: electro-acupuncture (visible muscle contraction)

Randomised to this group: 36 (analysed 28)

2) Control: sham electroacupuncture, similar number of needles except off the acupuncture point by 20 mm and current used on electro-stimulator was weaker than the real group. No increase in electrical volume was applied once the threshold of perception had been reached.

Randomised to this group: 34 (analysed 27)

Co-interventions: individual treatments continued, physiotherapy, anti-inflammatory agents, tricyclic antidepressants and analgesics.

See STRICTA table for treatment details (Appendix 10).


OutcomesPrimary outcomes:

1) Pain; visual analogue scale (1 to 100 mm)

Secondary outcomes:

2) Pain threshold, measured by pressure gauge over the 18 tender points as defined by ACR, before and after treatment

3) Analgesic use, tablets. Initial measurements taken in the week before the evaluation took place

4) Regional pain score, body drawing in which 21 regions are indicated. Patient assesses their pain in each region on a scale of 1 to 5, with 5 being the worst.

5) Sleep quality scale (1 to 10), with 10 being the best

6) Morning stiffness, measured in minutes

7) Patient general state (1 to 10), measured by patient, with 10 being the best

8) Evaluating physician impression (1 to 10), measured by physician as to the patient's general state, with 10 being the best

Outcome measures taken before and after treatments completed

Outcome measure results:

Overall approximately 50% improved significantly, 25% had no change with the balance showing "unexpectedly large improvement, with almost complete disappearance of symptoms"; 1 in the control group was observed to have a similar result

Real group improved significantly in 5 out of the 8 areas except morning stiffness

Pain threshold improved by 70% in the real group as opposed to 4% in the control group

Withdrawals/drop-outs: real acupuncture = 8, control group = 7

Complications/adverse events: real electroacupuncture 6 (2 = increase in symptoms, 3 = unpleasantness of needle sensation, 1 = ankle oedema). Sham electroacupuncture 5 (4 = increase in symptoms, 1 = unpleasantness of needle sensation)

Data extraction method: data were extracted from the published paper using table 2 and ZZ converted the SE data to SD. We selected the 'VAS pain scale' rather than the 'regional pain score' as it was the most used measurement tool for pain.


NotesOther info: e-mail contact was made with lead author who stated they were too busy to answer questions

Refunded costs to participants: not reported

Funding: not stated

Language: English

Publication: full paper


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskElectronic number generator

Allocation concealment (selection bias)Low riskClosed envelopes, numbered 1 to 70, prepared before study and opened in numerical order after recruitment

Incomplete outcome data (attrition bias)
All outcomes
High risk11 participants dropped out from the study, and their data were not included in the whole analysis. Number of participants dropped out from the study and reasons for drop-out were comparable in both groups (5/27; 6/28). As nearly 80% of those dropped out were due to increased symptoms, attrition bias is possible although this was comparable in both treatment groups.

Selective reporting (reporting bias)Low riskWHO clinical trials register was not established at time of publication

Blinding of participants and personnel (performance bias)
All outcomes
Low riskUsed acupuncture-naive participants, care giver not blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskIndependent evaluator, unaware of group allocations

Guo 2005

MethodsRandomised: semi-randomised according to the order of admission

Blinding: patient/care giver were not  blinded. Whether assessors were blinded is not reported. 

Setting: in and outpatients of an affiliated hospital, Helongjiang, China

Was study aim clear: no

Informed consent: not reported

Ethics approval: not reported

WHO clinical trial register: not listed

Follow-up: 6 months

Intention-to-treat used: not reported


ParticipantsTotal number of participants: 38 patients, mean duration of illness: acupuncture: 11 (2.3) months; control: 10 (3.6) months

7 male and 31 female; acupuncture group: M:F 3:16; control: M:F 4:15

Mean age: real acupuncture group = 50 (2.9) yrs (not sure if this is SD); control = 49 (3.4)

Diagnosis: ACR

Acupuncture-naive participants: not reported 

Excluded: not reported (did not mention if there were exclusion criteria)

Recruitment source: not reported.

Previous treatments: not reported


Interventions1) Real: acupuncture group

Randomised to this group: 19 (analysed 19)

2) Control: Western medication group (amitriptyline, tricyclic antidepressant, start from 10 mg, increased by 10 mg every 10 days until 30 mg, dividing dose into 2 and taking them at 2 different times of the day, 30 days 1 course)

Randomised to this group: 19 (analysed 19)

Co-interventions: not reported

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcome:

1) Pain; visual analogue scale (VAS), did not specify either 1 to 10 or 0 to 100 range

Secondary outcome:

2) Number of tender points, did not specify details

Outcome measure results:

Symptoms and signs-free, no tender points; acupuncture 9; control: 3

Significantly improved: VAS and tender points both reduced by or over 50%; most symptoms and signs resolved; 5:2

Improved: VAS and tender points both reduced by 25% to 50%; some improvement in S/S; 4:8

No effect: VAS and tender points both reduced < 25%, no changes in S/S; 1:6 

Withdrawals/drop-outs: no reported but based on data there were none

Complications/adverse events: not reported

Data extraction method: data were extracted from published paper table 2


NotesOther info: we could not contact the lead author to clarify missing information

Refunded costs to participants: not reported

Funding: not reported

Language: Chinese

Publication: full paper


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskRCT, but no other information

Allocation concealment (selection bias)High riskSemi-randomised according to the order of admission

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo drop-out reported. According to the data provided, all participants were included in the analysis.

Selective reporting (reporting bias)Unclear riskNot listed on WHO clinical trials register

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot reported

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported

Harris 2005

MethodsRandomised: computer-generated random numbers in a 4-block design. Concealed in an opaque envelope and given to the acupuncturist 1 day before treatment. 

Blinding: participants blindfolded with a non-blinded research assistant present during the treatments to monitor and ensure treatment integrity. Outcome assessors were blinded to treatment allocation. Care givers knew the allocation groups and the hypothesis.

Setting: Georgetown University, Washington DC

Was study aim clear: yes

Informed consent: verbal and written

Ethics approval: Institutional Review Board

WHO clinical trial register: listed on site and outcomes as per listing

Follow-up: none

Intention-to-treat: yes


ParticipantsTotal number of participants: 114 participants, mean duration of illness (years): T/S = 5.50 years (3.71), T/O = 5.26 years (4.83), N/S = 5.17 years (4.24), N/O = 5.77 years (4.10)

8 male and 106 female

Mean age: T/S 46.0 (10.1), T/O 44.5 (10.9), N/S 51.3 (10.0), N/O 48.1 (10.9)

Diagnosis: ACR

Acupuncture-naive participants: yes

Excluded: previous acupuncture treatments including sufficient knowledge that would prevent blinding, bleeding diathesis, autoimmune or inflammatory disease, daily narcotic analgesic use or a history of substance abuse, contraindication to use of acetaminophen or ibuprofen, in other clinical studies, pregnancy or lactation, receiving disability payment or litigation related to fibromyalgia

Recruitment source: the Washington DC metropolitan area using newspaper, periodicals and screened by telephone

Previous treatments: not reported


Interventions1) Real: traditional acupuncture (T/S) with stimulation

Randomised to this group: 29

2) Control: traditional acupuncture (T/O) without stimulation

Randomised to this group: 30

3) Control: non-traditional acupuncture (N/S) with stimulation, needles at same depth and stimulation as T/S group. Needles were placed in sites not believed to effective in Traditional Chinese Medicine based acupuncture
Randomised to this group: 28

4) Control: non-traditional acupuncture (N/O) without stimulation, in non-traditional sites

Randomised to this group: 27

(Each group received treatment once per week for 3 weeks, then twice per week for 3 weeks, then 3 times per week for 3 weeks (total 18 treatments). Between each treatment, there was a 2-week washout period)

Minimum number of treatments needed: not reported

Co-interventions: participants were allowed to continue normal treatments including antidepressants. They were not allowed to make any changes during the trial and not to seek acupuncture outside of the trial.

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcomes:

1) Pain: numeric rating scale, 101-point, range from 0 to 100 points in 5-point increments, 0 = no pain to 100 = worst pain imaginable (assessed before and week 3, 4 to 5, 8, 9 to 10, 13, 14 to 15)

2) Function: Short-Form 36, score ranges from 0 to 100 with higher scores indicating better function (assessed before and week 4 to 5, 9 to 10, 14 to 15)

Secondary outcomes:

3) Fatigue: Multi-Dimensional Fatigue Inventory questionnaire, scores range from 4 to 20 with larger scores indicating more fatigue (assessed before and week 4 to 5, 9 to 10, 14 to 15)

4) Blinding: participants were asked in week 4 whether they could determine which treatment arm they were in (A = acupuncture, B = placebo and C = could not tell)

Outcome measure results:

"Clinically significant improvements in pain were observed in 25% to 35% of subjects".

"Blinding assessment indicated that participants remained blinded to treatment at week 4 (P = 0.259)".

Withdrawals/drop-outs: 38 (T/S = 7, T/O = 11, N/S = 8,  N/O = 12)

Complications/adverse events: not reported

Data extraction method: data were extracted from the published paper using Table 2. We selected T/S and N/O data to represent real and sham acupuncture. For comparing different acupuncture styles we choose T/S versus T/O.


NotesOther info: contact was made with lead author who confirmed details of drop-outs and location of trial

Refunded costs to participants: not reported

Funding: NCCAM

Language: English

Publication: full paper


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated

Allocation concealment (selection bias)Low riskRandom numbers in a 4-block design. Concealed in an opaque envelope and given to the acupuncturist 1 day before treatment. 

Incomplete outcome data (attrition bias)
All outcomes
Low riskAll participants including those who dropped out from the study were included in the data analysis. The number of and reasons for drop-out were comparable among the groups.

Selective reporting (reporting bias)Low riskAll outcomes reported, as per WHO clinical trials register

Blinding of participants and personnel (performance bias)
All outcomes
Low riskUsed acupuncture-naive participants, other participants not present at clinic during treatment, blindfolded during treatment. Care giver not blinded. Non-blinded research assistant present during all treatments to ensure treatment integrity. Blinding of allocation groups was tested at week 3 with no significant differences noted.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll evaluators blinded to treatment allocation

Harris 2008

MethodsRandomised: random number generator used (blocks of 4 with 2 acupuncture and 2 sham)

Blinding: participants blindfolded during treatments. All assessors were blinded to treatment assignments.

Setting: University of Michigan, USA

Was study aim clear: yes

Informed consent: written and informed

Follow-up: none, only for term of treatment

Ethics approval: University of Michigan Institutional Review Board

WHO clinical trial register: not listed and confirmed by author

Intention-to-treat: yes, all subjects completed trial


ParticipantsTotal number of participants: 10 participants, duration of fibromyalgia for > 1 year

0 male (acupuncture and control) and 10 female

Mean age: both acupuncture and control combined mean 48 SD 15 years

Diagnosis: ACR 1990 criteria

Acupuncture-naive participants: yes

Excluded: as per Harris 2005 study

Recruitment source: fibromyalgia subject registry at the University of Michigan Chronic Pain and Fatigue Center

Previous treatments:


Interventions1) Real: acupuncture

Randomised to this group: 6

2) Control: non skin-penetrating acupuncture (Sherman 2002)

Randomised to this group: 4

Minimum number of treatments needed: 9 out of 9

Co-intervention: none

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcomes:

1) Pain: VAS subset of Short-Form McGill Pain Questionnaire (SF-MPQ)

Assessments: at baseline and end of treatment

Outcome measure results: clinical pain improved from pre- to post-treatment according to SF-MPQ rating of the sensory dimensions of pain (mean difference in clinical pain rating 3.50 (SD 4.70); P = 0.043)

SF-MPQ sensory score baseline mean (SD) = 12.3 (4.35)

SF-MPQ sensory score end of treatment mean (SD) = 8.80 (5.61)

Withdrawals/drop-outs: no drop-outs

Complications/adverse events: no adverse events

Author stated "actually the primary outcome for this study was neuroimaging changes for TA and SA. Clinical pain was never a primary outcome, it was only used as a covariate".

Data extraction method: data for pain was provided by the author directly


NotesOther info: part of an ongoing study; results for this study have not been published anywhere else. Missing details from the study were confirmed by the lead author via e-mail. They included details of randomisation, blinding, whether acupuncture-naive, score baseline/end of treatment/drop-outs and data. The lead author confirmed that this is not a subset of the Harris 2005 study.

Funding: US Department of Army grant and NIH/National Centre for Complementary and Alternative Medicine

Refunded costs to participants: not reported

Language: English

Publication: full paper 


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated

Allocation concealment (selection bias)Low riskRandom number generator used (blocks of 4, with 2 acupuncture and 2 sham)

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo missing data

Selective reporting (reporting bias)Unclear riskNot listed on WHO clinical trials register

Blinding of participants and personnel (performance bias)
All outcomes
Low riskUsed acupuncture-naive participants, participants blindfolded, care giver was not blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll assessments were blinded

Harris 2009

MethodsRandomised: random number generator used (blocks of 4 with 2 acupuncture and 2 sham)

Blinding: participants blindfolded during treatments. All assessors were blinded to treatment assignments.

Setting: University of Michigan, USA

Was study aim clear: yes

Informed consent: written and informed

Follow-up: none, only for term of treatment

Ethics approval: University of Michigan Institutional Review Board

WHO clinical trial register: not listed and confirmed by author

Intention-to-treat: yes, all subjects completed trial


ParticipantsTotal number of participants: 20 participants, duration of fibromyalgia > 1 year

0 male (acupuncture and control) and 20 female

Mean age: both acupuncture and control combined mean 44.3 SD 13.6 years

Diagnosis: ACR 1990 criteria

Acupuncture-naive participants: yes

Excluded: as per Harris 2005 study

Recruitment source: fibromyalgia subject registry at University of Michigan

Previous treatments: not reported


Interventions1) Real: acupuncture (TA)

Randomised to this group: 10

 2) Control: non skin-penetrating acupuncture (SA) (Sherman 2002)

Randomised to this group: 10

Minimum number of treatments needed: 9 out of 9

Co-intervention: medication (agreed not to change)

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcomes:

1) Pain: McGill Pain Questionnaire (SF-MPQ)

Assessments: st baseline and end of treatment

Outcome measure results: "significant reductions in pain were observed for the entire cohort for the total score SF-MPQ"
Total: mean difference (SD) treatment - baseline: -3.45 (7.39); P = 0.05

SF-MPQ Sensory Score: mean (SD): -2.65 (5.98) P = 0.06

SF-MPQ Affective Score: mean (SD): -0.80 (2.25) P = 0.13

Both TA and SA resulted in clinically meaningful reductions in pain (SF-MPQ total score mean difference (SD): TA -4.00 (6.72); SA -2.90 (8.33)

2) Assessment of masking:

Participants had to guess which group they belonged to after the first PET scan. Overall the 2 distributions were not statistically different: Chi2 = 0.88, P = 0.65

Withdrawals/drop-outs: no drop-outs

Complications/adverse events: no adverse events

Author stated "actually the primary outcome for this study was neuroimaging changes for TA and SA. Clinical pain was never a primary outcome, it was only used as a covariate".

Data extraction method: data were provided by the author. The author noted that "actually in this manuscript (Harris 2009) we did not analyse real versus sham acupuncture groups. We combined both groups together in this analysis".


NotesOther info: part of an ongoing study. Details of study were confirmed by the author via e-mail: drop-outs/adverse events and data. The author confirmed that this is not a subset study of Harris 2005.

Funding: US Department of Army grant and NIH/National Centre for Complementary and Alternative Medicine

Refunded costs to participants: not reported

Language: English

Publication: full paper


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated

Allocation concealment (selection bias)Low riskRandom number generator used (blocks of 4, with 2 acupuncture and 2 sham)

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo missing data

Selective reporting (reporting bias)Unclear riskNot listed on WHO clinical trials register

Blinding of participants and personnel (performance bias)
All outcomes
Low riskUsed acupuncture-naive participants, participants blindfolded, care giver was not blinded. Participants had to guess which group they belonged to after first PET scan. Overall the 2 distributions were not statistically different.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll assessments were blinded

Itoh 2010

MethodsRandomised: randomly assigned with a computerised randomisation program (SAMPSIZE V2.0, Blackwell Science Ltd, UK), permutated block randomisation to either group

Blinding: outcome assessors were blinded to treatment assignments

Setting: Acupuncture and Moxibustion Center, Meiji University of Integrative Medicine, Kyoto, Japan

Was study aim clear: yes

Informed consent: written and informed

Follow-up: none, only for term of treatment

Ethics approval: Ethics Committee of Meiji University of Integrative Medicine

WHO clinical trial register: not listed and confirmed by author

Intention-to-treat: no, analysis only of participants that completed study


ParticipantsTotal number of participants: 13 participants, duration of fibromyalgia for Group A 3.9 (SD 8.4), Group B 4.4 (SD 2.3)

3 male and 13 female; does not state which group they were allocated to

Mean age: Group A 45.7 (15.2), Group B 47.3 (13.3) years

Diagnosis: ACR 1990 criteria

Acupuncture-naive participants: yes

Excluded: previous acupuncture, bleeding disorders, autoimmune or inflammatory diseases, participation in other trials, pregnancy or lactation, receiving disability payments or involved in litigation related to fibromyalgia

Recruitment source: fibromyalgia specialists at hospitals

Previous treatments: maintain current medication use


Interventions1) Real: acupuncture (TA)

Group B, electro- and trigger point acupuncture

Randomised to this group: 7

 2) Control: Group A received 5 acupuncture sessions after 5 weeks of weekly or twice weekly clinical examinations only

Randomised to this group: 6

Minimum number of treatments needed: not reported

Co-intervention: medications using amitriptyline, SSRIs and SNRIs (agreed not to change)

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcomes:

1) Pain: VAS 19 cm scale with higher score indicating negative impact

2) Function: Fibromyalgia Impact questionnaire FIQ), 20 items covering physical functioning, work status, depression, anxiety, sleep (rest/morning tiredness), pain, stiffness, fatigue and well-being. Each scored 0 to 10. The higher the combined score the worse the condition is affecting the participant. Full details on scoring can be found in Burckhardt 1991.

Assessments: before start, week 5 and at end of treatments, week 10 

Outcome measure results: VAS, Group A remained unchanged until acupuncture treatment started, while Group B decreased by week 5. No differences between groups at baseline (P = 0.566), while at week 5, significant differences in VAS between groups (P = 0.022) and at week 10 no difference (P = 0.252).

FIQ, Group A remained unchanged until acupuncture treatment started, while Group B decreased by week 5. No differences between groups at baseline (P = 0.616), while at week 5 significant differences in FIQ between groups (P = 0.026) and at week 10 no difference (P = 0.86).

Withdrawals/drop-outs: Group A 2, Group B 1, both lost due to lack of response to treatment

Complications/adverse events: none reported 

Data extraction method: data were provided by the lead author directly


NotesOther info: author was contacted by e-mail and provided data

Funding: not reported

Refunded costs to participants: not reported

Language: English

Publication: full paper

Results: see the comparisons


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated

Allocation concealment (selection bias)Low riskPermutated block randomisation to either group

Incomplete outcome data (attrition bias)
All outcomes
High riskIntention-to-treat not used, 3 drop-outs (1 acupuncture, 2 no treatment groups) due to lack of response to treatment, however drop-out rates were not significantly different between groups

Selective reporting (reporting bias)Unclear riskNot listed on WHO clinical trials register

Blinding of participants and personnel (performance bias)
All outcomes
High riskUsed acupuncture-naive participants, however participants at start of treatment would have known which group they belonged to due to study design; care giver was not blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll assessments performed by independent investigator not aware of treatment sequence or treatment received

Martin 2006

MethodsRandomised: immediately before first treatment. Opaque envelopes, labelled sequentially. Each contained a 3 x 5 index card, which was printed with the group assignment. Every 4 envelopes contained 2 control and 2 experimental assignments. This was done in blocks of 2 to prevent imbalances in treatment allocation. Envelopes were opened in order.

Blinding: participants seated in an arrangement that blocked their view to treatment. Outcome assessors blinded to group allocation. Care givers and participants maintained neutral conversion.

Setting: Mayo Fibromyalgia Treatment Program, Rochester, Minnesota, USA

Was study aim clear: yes

Informed consent: verbal and written

Ethics approval: Mayo Foundation Institutional Review Board

WHO clinical trial register: not listed and confirmed by author who stated they did not know about it

Follow-up: 1 and 7 months

Intention-to-treat: yes


ParticipantsTotal number of participants: 50 patients, mean duration of illness (years): not reported

1 male (control group) and 49 female (real and control)

Mean age: acupuncture 47.9 SD 11.2 years, control 51.7 SD 14.1 years

Diagnosis: ACR

Acupuncture-naive participants: yes

Excluded: prior acupuncture experience, bleeding diathesis, had to be able to understand consent and to be able to fill out the questionnaires

Recruitment source: referrals to programme from physician after conservative management

Previous treatments: patients had received conservative management, but this was not described; "many had already used most of the basic treatments for fibromyalgia"


Interventions1) Real: electroacupuncture, used a special table arrangement that did not allow the patient to see what was happening;

Randomised to this group: 25

2) Control: sham electroacupuncture, setting was same as real group, except needle was attached to plaster and did not break the skin;

Randomised to this group: 25

Minimum number of treatments needed: all patients completed at least 5 treatments

Co-interventions: 1.5 days of education, counselling and group discussion about symptom management (done before enrolment into study, 4-week wash-out period before start of treatments). No other co-interventions were reported.

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcomes:

1) Pain: Multidimensional Pain Inventory (MPI), 61-item questionnaire developed for chronic pain. Composed of 13 scales that measure different pain-related aspects. 4 of the questions that related to support from spouse or significant other were excluded, as it was not part of the standard treatment programme.

2) Function: Fibromyalgia Impact Questionnaire (FIQ), 20 items covering physical functioning, work status, depression, anxiety, sleep (rest/morning tiredness), pain, stiffness, fatigue and well-being. Each scored 0 to 10. The higher the combined score the worse the condition is affecting the patient. Full details on scoring can be found in the article by Burckhardt 1991.

Secondary outcomes:

3) Participants were asked their opinion regarding group assignment

Assessments: before start, immediately at end of treatment sessions and at 1 and 7 months

Outcome measure results:

FIQ showed significant improvement in the acupuncture group over control acupuncture during study period (P = 0.01), with the greatest difference at the 1 month (P = 0.007). Subscale analysis showed significant group effect for fatigue (P = 0.001) and anxiety (P = 0.003) at 1 month. Other sub-scales showed trends towards improvement but were not statistically significant.

MPI group effect showed significant improvement in pain at 1 month (P = 0.03) but effect was lost at the 7-month measure (P = 0.05).

Blinding of participants did not exceed chance

Withdrawals/drop-outs: 1 lost to follow-up

Complications/adverse events: many participants in both groups experienced feeling tired and/or relaxed after treatment. Mild bruising and soreness was common in acupuncture group. 2 patients experienced mild vasovagal symptoms (1 from each group). 1 patient experienced a pulmonary embolism (believed to be unrelated to the study). 

Data extraction method: data were extracted from published paper table 3 and we used data at the one month after treatment point as per our protocol


NotesOther info: author was contacted by e-mail and confirmed allocation concealment, WHO listing and point locations

Funding: Mayo Foundation and Mayo Anaesthesia Clinical Research Unit

Refunded costs to participants: financial compensation provided for parking

Language: English

Publication: full paper


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandom sequence not reported

Allocation concealment (selection bias)Low riskImmediately before first treatment. Opaque envelopes, labelled sequentially. Each contained a 3 x 5 index card, which was printed with the group assignment. Every 4 envelopes contained 2 control and 2 experimental assignments. This was done in blocks of 4 to prevent imbalances in treatment allocation. Envelopes were opened in order.

Incomplete outcome data (attrition bias)
All outcomes
Low riskAll participants completed outcome measures at the end of treatment; 1 from the control group did not have data for 1 and 7-month follow-ups and was excluded from follow-up analysis. Given this was only 1 participant, this exclusion would have minimal impact on the outcome.

Selective reporting (reporting bias)Unclear riskNot listed on the WHO clinical trials register; author stated did not know about it

Blinding of participants and personnel (performance bias)
All outcomes
Low riskUsed acupuncture-naive participants, blinded to actual treatment by seated arrangement of protocol, restricted conversation during treatment and use of blinded study co-ordinator for questions. Care giver was not blinded. Group allocation was tested and ability of patients to determine treatment received did not exceed chance.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll evaluations were obtained from participants by study co-ordinator who was blinded to group allocations

Targino 2008

MethodsRandomised: using "a computer-generated random sequence of numbers provided by the Hospitals Informatics Departments. The randomization was conducted by one physician who was not involved with the inclusion or exclusion process."

Blinding: assessor blind - "participants rated their pain intensity using a VAS. Blinded evaluation of PPT18 and TePsN was carried out by a single physician (HHSK), while blind evaluation of quality of life was conducted by one psychologist (LPMS). Even though the participants knew which group they were in (either the acupuncture treatment group or the standard care group), they were instructed not to communicate this information to the outcome assessors."

Setting: not reported, but assume it was the Clinics Hospital

Was study aim clear: yes

Informed consent: written and informed consent

Follow-up: 3 months (after randomisation and at the end of the treatment), 6 months, 12 months and 24 months

Ethics approval: the ethics review committee of the Clinics Hospital

WHO clinical trial register: listed and outcomes as per listing

Intention-to-treat: yes, up to 12 months follow-up, but not at 24 months follow-up     


ParticipantsTotal number of participants: 58 patients, mean duration of pain (months): acupuncture group 118.8 (117.3), control 93.0 (75.25). Did not describe what they meant by pain duration, whether it was fibromyalgia diagnosis or pain. Assume it refers to fibromyalgia.

Gender: 0 male (acupuncture and control) and 58 female

Mean age: acupuncture 52.09 SD 10.97 years, control 51.17 SD 11.20 years

Diagnosis: ACR: "ACR criteria were applied by one of the physicians (HHSK) to confirm the diagnosis prior to the enrolment to the study."

Inclusion: 20 to 70 yrs old, have to have had pain VAC > 4/10; using an antidepressants at an analgesic dose (12.5 to 75 mg/kg). Author did not specify the name of medication

Acupuncture-naive participants: no, patients who had acupuncture in the previous 12 months were excluded. Presumably, this is not an important item because sham acupuncture is not used.

Excluded: patients with "severe psychiatric disease, the presence of neurological deficits, cardiac disease or glaucoma, and treatment with acupuncture within one year prior to the start of the study."

Recruitment source: "were recruited by physicians from the Clinics Hospital in Sao Paulo. They included doctors practising in the Pain Clinic of the Department of Neurology, those in the Rheumatology Clinic and those in the Division of Physical Medicine of the Institute of Orthopedics and Traumatology."

Previous treatments: none reported except for current medication


Interventions1) Real: acupuncture + standard care (12.5 to 75 mg of tricyclic antidepressants per day), individualised plus exercise, including "oral instruction to walk for 30 min twice a week at their own pace, to breathe deeply and to perform mental relaxation exercises for another 30 min. They were also told to perform twice-weekly stretching exercises involving the para-spinalis muscles, glutei, hamstrings, ankle plantar flexors and hip flexors." Patients in the acupuncture group always had their sessions performed by the same physician (RAT).

Randomised to this group: 34

2) Control: standard care, tricyclic antidepressant (individualised, ranging from 12.5 to 75 mg/day, most (84.5%) received 50 mg/day) plus exercise. Participants "were seen by a physician at the beginning of the study and during the follow-up visits. No additional visits were scheduled for the controls to compensate for the extra attention being received by patients in the acupuncture treatment group. Compliance with the use of either exercise or antidepressant drugs was based on participants reports during the outcome evaluation interviews."

Randomised to this group: 24

Compliance: 97.1% (33 participants) completed all 20 sessions, with one leaving after 17 sessions due to complete relief from pain

Minimum number of treatments needed: not reported

Co-intervention: not reported

See STRICTA table for treatment details (Appendix 10)


OutcomesPrimary outcomes:

1) Visual analogue scale (VAS) with 0 = no pain and 10 = worst pain experienced

2) Quality of life: SF-36 form. Portuguese version of MOS 36-item short-form health survey (8-items), higher scores indicate better quality of life.

Secondary outcomes:

3) Number of tender points below 4 kg/cm2 (TePsN). The lower the number, the less the severity of symptoms.

4) Mean pressure pain threshold value, over the 18 fibromyalgia points (PPT18). The higher the values the less severe the symptoms, measured with algometry (not sure if electronic or manual).

Assessments: at baseline, 3 months (after randomisation and at the end of the treatment), 6 months, 12 months and 24 months

Outcome measure results:

The 2 groups were comparable at baseline. Also mentioned in the discussion, the usage of medication was not different between the 2 groups.

VAS showed statistically significant improvement in the acupuncture group at 3 months (P < 0.001, however at 6, 12 and 24 months follow-up, it was not statistically different between the groups (P > 0.05)

SF-36 showed improvement in only 5 sub-scales of the acupuncture group at 3 months (PF, BP, VT, RE, MH). At 6 months the acupuncture group benefit was for only 1 sub-scale (GH) and at 12-month follow-up only 1 sub-scale showed improvement (RP).

TePsN and PPT18 showed improvement in the acupuncture group at 3 and 6 months

There was no statistical difference in the scores of the standard care group at any time (P > 0.05)

Withdrawals/drop-outs: at 24 months, 2 were lost to follow-up in the acupuncture group (follow-up rate 94.1%), 1 in the control group (follow-up rate 95.8%)

Complications/adverse events: 2 patients in the acupuncture group reported temporary oedema of the left hand at LI4. There were no reported incidences of discomfort, soreness, vasovagal symptoms, bruising or haematoma at time of treatment or the during the follow-up period of 24 months.

Data extraction method: data were extracted from the published paper using table 2 and medians/ranges were converted by the CMSG statistician. We could not use SF-36 data as there were no available data for converting median/range to mean. We were waiting for details from the author at the time of publication.


NotesOther info: 1 author of our review was in contact with the study author who that advised she had a paper awaiting publication which she provided direct. This was outside of our search at the start of review and provided further data.

Funding: no funding provided

Refunded costs to participants: not reported

Language: English

Publication: full paper


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated random sequence of numbers

Allocation concealment (selection bias)High riskDid not use, as study objective was to evaluate the benefit of the addition of acupuncture to a standard course of tricyclic antidepressants and exercise

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo drop-out was reported during the treatment or in the first 12 months of follow-up. 3 participants were not contacted at the 24-month follow-up and their data were excluded from analysis at that time. This exclusion will not impact on the outcome at the end of treatment or 12-month follow-up. Furthermore, the number of drop-outs at 24 months was comparable between the 2 groups.

Selective reporting (reporting bias)Low riskAll outcomes reported as per WHO clinical trials register

Blinding of participants and personnel (performance bias)
All outcomes
High riskParticipants were restricted to those who not had acupuncture in the last 12 months; care giver was not blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskOutcome assessors were blinded and participants were told not to inform them of their group allocation

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Cao 2003Article in Chinese. "Randomised" stated but methods not described. Excluded as study had an extra therapy (mobile cupping) in the acupuncture and Western medicine arm that was not included in the control arm of Western medicine (i.e. acupuncture + Western medicine + mobile cupping versus Western medicine).

Cassisi 1994Article in Italian. "Patients were randomly chosen and divided into three therapeutic groups." Translation by Italian Cochrane Centre confirms article same as Cassisi 1995 and is a case series.

Cassisi 1995Article in Italian. "Patients were randomly chosen and divided into three therapeutic groups." Translation by Italian Cochrane Centre confirms article is a case series. This study has been quoted in a number of reviews and articles as a RCT but this is incorrect. Contact with one of the authors: the data in this reprint are incorrect and they advise to use the 1994 article.

Chen 2009Article in Chinese; case series

Collazo Chao 2010Article in Spanish; not a RCT

Dai 2009Article in Chinese; case series

Feldman 2001Conference report. RCT. The publisher and conference organisers were contacted and could not provide details of the whereabouts of the authors. Internet searches were undertaken to try and locate either author without success.

Gong 2010Article in Chinese. Extra therapy not included in both arms (acupuncture + mind focus versus Western medicine).

Gou 2010Article in Chinese. Extra therapy not included in both arms (acupuncture + infrared lamp versus Western medicine).

Guan 2005Article in Chinese. Data unusable as responder only.

Guevara 2007Conference report. No primary clinical outcomes published.

Guo 2003Article in Chinese. Data unusable as responder only.

Guo 2005aArticle in Chinese. Data unusable as responder only (although number of tender points reported for baseline).

Harris 2006Secondary analysis of original article (Harris 2005)

Harris 2007Data were from before acupuncture treatment (cross-sectional study)

Harris 2007aConference report, RCT. Author was contacted and has advised awaiting full journal publication.

Harris 2007bConference report, RCT. Author was contacted and has advised awaiting full journal publication.

Jiang 2010Article in Chinese. Invalid control (acupuncture + cupping + Western medicine versus acupuncture + cupping versus+ Western medicine).

Lautenschlager 1989Article in German. ACR criteria not met.

Li 2005Article in Chinese. Data unusable as responder only.

Li 2005aArticle in Chinese. Data unusable as responder only.

Li 2006Article in Chinese. 'Randomised' stated but methods not described. Excluded as study had an extra therapy (mobile cupping) in the acupuncture and Western medicine arm that was not included in the control arm of Western medicine (acupuncture + Western medicine + mobile cupping versus Western medicine).

Li 2008Article in Chinese. Did not report any of our primary outcomes measures.

Li 2010Article in Chinese. Invalid control (acupuncture + moxa versus acupuncture + moxa + Western medicine).

Lui 2002Article in Chinese. Did not meet ACR criteria.

Sandberg 1999Swedish study. Not a RCT.

Sandberg 2004Comparison study. None of the review's primary outcome measures were used.

Sprott 1995Article in German. Conference report. Author asked that the 1998 study be considered, although this report had more data than the 1998 study.

Sprott 1998"Randomly subdivided into 3 groups". Data unusable as shows only 'mean' results.

Sprott 2000Article in German. Unsure how randomised. None of the review's primary outcome measures were used.

Sun 2008Article in Chinese; case series

Targino 2002RCT; no quantitative data for analyses

Uhlemann 2001Article in German. Randomised. Conference report. Author could not be contacted; no quantitative data for analysis.

Wang 2002Article in Chinese. Data unusable as no SD.

Wang 2004Article in Chinese. Data unusable as responder only.

Wei 2006Article in Chinese. Data unusable as responder only.

Wu 2003Article in Chinese. Data unusable as responder only.

Yao 2006Article in Chinese. Data unusable as responder only.

Zhang 2001Article in Chinese. Data unusable as responder only.

Zhou 2003Article in Chinese. Data unusable as responder only.

 
Characteristics of ongoing studies [ordered by study ID]
Vas 2011

Trial name or titleEffects of acupuncture on patients with fibromyalgia: study protocol of a multi-centre randomised controlled trial

MethodsRCT multi-centre study

Participants156 participants, aged over 17, ACR diagnosis

InterventionsTrue or sham acupuncture, 9 treatments, once per week

OutcomesFIQ, Hamilton rating scale for depression, medication use. Follow-up 6 and 12 months

Starting dateOctober 2010 to December 2013

Contact informationjorgef.vas.sspa@juntadeandalucia.es

Notes

 
Comparison 1. Acupuncture versus non-acupuncture treatment

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Pain up to 1 month after treatment113Mean Difference (IV, Random, 95% CI)-22.4 [-40.98, -3.82]

 2 Global well-being: rated by participants up to 1 month after treatment113Mean Difference (IV, Random, 95% CI)-15.40 [-25.62, -5.18]

 3 Sleep up to 1 month after treatment113Mean Difference (IV, Random, 95% CI)-0.40 [-1.01, 0.21]

 4 Fatigue up to 1 month after treatment113Mean Difference (IV, Random, 95% CI)-1.1 [-1.98, -0.22]

 5 Stiffness up to 1 month after treatment113Mean Difference (IV, Random, 95% CI)-0.90 [-1.66, -0.14]

 6 Adverse events113Risk Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

 7 Mental well-being up to 1 month after treatment113Mean Difference (IV, Random, 95% CI)-0.5 [-1.10, 0.10]

 
Comparison 2. Acupuncture versus placebo or sham acupuncture

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Pain (subgroup EA & MA) up to 1 month after treatment6286Std. Mean Difference (IV, Random, 95% CI)-0.14 [-0.53, 0.25]

    1.1 Electro-acupuncture
2104Std. Mean Difference (IV, Random, 95% CI)-0.63 [-1.02, -0.23]

    1.2 Manual acupuncture
4182Std. Mean Difference (IV, Random, 95% CI)0.14 [-0.17, 0.45]

 2 Pain follow-up to 7 months after treatment (subgroup EA vs MA)2145Std. Mean Difference (IV, Random, 95% CI)-0.12 [-0.52, 0.28]

    2.1 Electro-acupuncture
149Std. Mean Difference (IV, Random, 95% CI)-0.37 [-0.93, 0.20]

    2.2 Manual acupuncture
196Std. Mean Difference (IV, Random, 95% CI)0.05 [-0.41, 0.51]

 3 Pain: sham non-invasive acupuncture (not breaking skin) vs sham invasive acupuncture (breaking skin)6Std. Mean Difference (IV, Random, 95% CI)Subtotals only

    3.1 Sham breaking skin
3170Std. Mean Difference (IV, Random, 95% CI)-0.21 [-0.57, 0.15]

    3.2 Sham not breaking skin
4116Std. Mean Difference (IV, Random, 95% CI)0.04 [-0.64, 0.71]

 4 Physical function (SF-36)156Mean Difference (IV, Random, 95% CI)-5.80 [-10.91, -0.69]

 5 Global well-being: rated by participants (subgroup EA vs MA) up to 1 month after treatment3200Std. Mean Difference (IV, Random, 95% CI)0.29 [-0.44, 1.01]

    5.1 Electro-acupuncture
2104Std. Mean Difference (IV, Random, 95% CI)0.65 [0.26, 1.05]

    5.2 Manual acupuncture
196Std. Mean Difference (IV, Random, 95% CI)-0.40 [-0.86, 0.06]

 6 Global well-being follow-up to 7 months after treatment (subgroup EA & MA)2145Std. Mean Difference (IV, Random, 95% CI)-0.03 [-0.87, 0.81]

    6.1 Electro-acupuncture
149Std. Mean Difference (IV, Random, 95% CI)0.41 [-0.15, 0.98]

    6.2 Manual acupuncture
196Std. Mean Difference (IV, Random, 95% CI)-0.45 [-0.91, 0.01]

 7 Sleep (subgroup EA & MA) up to 1 month after treatment3200Std. Mean Difference (IV, Random, 95% CI)0.16 [-0.29, 0.61]

    7.1 Electro-acupuncture
2104Std. Mean Difference (IV, Random, 95% CI)0.40 [0.01, 0.79]

    7.2 Manual acupuncture
196Std. Mean Difference (IV, Random, 95% CI)-0.25 [-0.71, 0.21]

 8 Sleep follow-up to 7 months after treatment (subgroup EA & MA)2145Std. Mean Difference (IV, Random, 95% CI)-0.09 [-0.44, 0.26]

    8.1 Electro-acupuncture
149Std. Mean Difference (IV, Random, 95% CI)0.07 [-0.49, 0.63]

    8.2 Manual acupuncture
196Std. Mean Difference (IV, Random, 95% CI)-0.20 [-0.66, 0.26]

 9 Fatigue (subgroup EA vs MA) up to 1 month after treatment3201Std. Mean Difference (IV, Random, 95% CI)-0.10 [-0.81, 0.61]

    9.1 Electro-acupuncture
149Std. Mean Difference (IV, Random, 95% CI)-0.85 [-1.44, -0.27]

    9.2 Manual acupuncture
2152Std. Mean Difference (IV, Random, 95% CI)0.26 [-0.08, 0.61]

 10 Fatigue follow-up to 7 months after treatment (subgroup EA vs MA)2145Std. Mean Difference (IV, Random, 95% CI)0.04 [-0.52, 0.59]

    10.1 Electro-acupuncture
149Std. Mean Difference (IV, Random, 95% CI)-0.27 [-0.84, 0.29]

    10.2 Manual acupuncture
196Std. Mean Difference (IV, Random, 95% CI)0.30 [-0.16, 0.76]

 11 Stiffness up to 1 month after treatment2104Std. Mean Difference (IV, Random, 95% CI)-0.45 [-0.84, -0.06]

 12 Stiffness follow-up to 7 months after treatment149Mean Difference (IV, Random, 95% CI)-0.30 [-1.60, 1.00]

 13 Adverse events6289Risk Ratio (M-H, Random, 95% CI)0.44 [0.12, 1.63]

 14 Mental well-being up to 1 month after treatment149Mean Difference (IV, Random, 95% CI)-1.70 [-3.13, -0.27]

 15 Mental well-being follow-up to 7 months149Mean Difference (IV, Random, 95% CI)-1.40 [-3.01, 0.21]

 16 Analgesic use (number of tablets per week) up to 1 month after treatment155Mean Difference (IV, Random, 95% CI)-3.20 [-10.20, 3.80]

 17 Analgesic use (number of participants taking analgesics up to 1 month after treatment)180Risk Ratio (M-H, Random, 95% CI)0.94 [0.66, 1.32]

 18 Tenderness up to 1 month after treatment155Mean Difference (IV, Random, 95% CI)0.80 [0.02, 1.58]

    18.1 Mean pressure pain threshold (kg/cm2)
155Mean Difference (IV, Random, 95% CI)0.80 [0.02, 1.58]

 19 Overall well-being: rated by care giver at end of treatment155Mean Difference (IV, Random, 95% CI)2.0 [0.81, 3.19]

 
Comparison 3. Acupuncture versus medication

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Pain up to 1 month after treatment138Mean Difference (IV, Random, 95% CI)-17.3 [-24.13, -10.47]

 2 Adverse events138Risk Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

 3 Tenderness up to 1 month after treatment138Mean Difference (IV, Random, 95% CI)-4.00 [-6.73, -1.27]

    3.1 Number of tender points
138Mean Difference (IV, Random, 95% CI)-4.00 [-6.73, -1.27]

 
Comparison 4. Acupuncture as an adjunct therapy

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Pain up to 1 month after treatment158Mean Difference (IV, Random, 95% CI)-3.0 [-3.90, -2.10]

 2 Pain up to 7 months after treatment158Mean Difference (IV, Random, 95% CI)-0.5 [-1.49, 0.49]

 3 Adverse events158Risk Ratio (M-H, Fixed, 95% CI)3.57 [0.18, 71.21]

 4 Tenderness - number of tender points below kg/cm2 up to 1 month after treatment158Mean Difference (IV, Random, 95% CI)-4.5 [-6.20, -2.80]

 5 Tenderness - number of tender points below kg/cm2 up to 7 month after treatment158Mean Difference (IV, Random, 95% CI)-2.0 [-3.51, -0.49]

 6 Tenderness - mean pressure threshold by pressure algometry up to 1 month after treatment168Mean Difference (IV, Fixed, 95% CI)0.70 [0.41, 0.99]

 7 Tenderness - mean pressure threshold by pressure algometry, follow up to 7 months after treatment158Mean Difference (IV, Random, 95% CI)0.60 [0.26, 0.94]

 
Comparison 5. Deep needling with stimulation (T/S) versus deep needling without stimulation (T/O)

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Pain up to 1 month after treatment1Mean Difference (IV, Random, 95% CI)Subtotals only

    1.1 NRS
141Mean Difference (IV, Random, 95% CI)0.30 [-18.34, 18.94]

 2 Physical function (SF-36) up to 1 month after treatment141Mean Difference (IV, Random, 95% CI)-5.50 [-11.43, 0.43]

 3 Fatigue up to 1 month after treatment141Mean Difference (IV, Fixed, 95% CI)1.10 [-1.41, 3.61]

 
Summary of findings for the main comparison. Acupuncture versus non-acupuncture for treating fibromyalgia

Acupuncture versus non-acupuncture for treating fibromyalgia

Patient or population: patients with fibromyalgia

Settings: Japan1

Intervention: acupuncture versus non-acupuncture

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

Non-acupunctureAcupuncture (EA)

Pain up to 1 month after treatment
VAS2
No treatment
69.8 points

Scale (0 to 100)
EA

47.4 points

(Lower score indicates less pain)

22.4 MD lower
(40.98 lower to 3.82 lower)
13
(1 study)
⊕⊕⊝⊝
low4
AR % -22.40% (-40.98% to -3.82%)

RR % -30.19% (-55.23% to -5.15%)

NNT 4 (1 to 161)

Physical functionNot reportedNot reported

Global well-being up to 1 month after treatment

FIQ3
No treatment
66.5 points

Scale (0 to 100)
EA

51.1 points

(Lower score indicates better well-being)

15.4 MD lower
(25.62 lower to 5.18 lower)
13
(1 study)
⊕⊕⊝⊝
low4
AR % -15.40% (-25.62% to -5.18%)

RR % -23.88% (-39.72% to -8.03%)

NNT 4 (1 to 52)

Sleep up to 1 month after treatment

Subset (rest) FIQ3
No treatment
4.0 points

Scale (0 to 10)
EA

3.6 points

(Lower score indicates better sleep)

0.4 MD lower
(1.01 lower to 0.21

higher)
13
(1 study)
⊕⊕⊝⊝
low4
AR % -4.00% (-10.10% to 2.10%)

RR % -10.53% (-26.58% to 5.53%)

NNT N/A

Fatigue up to 1 month after treatment

Subset (fatigue) FIQ3
No treatment
4.5 points

Scale (0 to 10)
EA

3.4 points

(Lower score indicates less fatigue)

1.1 MD lower
(1.98 lower to 0.22 lower)
13
(1 study)
⊕⊕⊝⊝
low4
AR % -11.00% (-19.80% to -2.20%)

RR % -26.19% (-47.14% to -5.24%)

NNT 4 (1 to 52)

Stiffness up to 1 month after treatment

Subset (stiffness) FIQ3
No treatment
4.8 points

Scale (0 to 10)
EA

3.9 points

(Lower score indicates less stiffness)

0.9 MD lower
(1.66 lower to 0.14 lower)
13
(1 study)
⊕⊕⊝⊝
low4
AR % -9.00% (-16.60% to -1.40%)

RR % -21.95% (-40.49% to -3.41%)

NNT 3 (1 to 128)

Adverse eventsNot estimable13
(1 study)
low5No adverse events only withdrawals (3) due to non-improvement in condition

 *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
AR: absolute risk; CI: confidence interval; EA: electroacupuncture; FIQ: Fibromyalgia Impact Questionnaire; MD : mean difference; NNT: number needed to treat; RR: risk ratio; VAS: visual analogue scale
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1Itoh 2010, no follow-up.
2VAS (0 = no pain, 10 = worse pain ever).
3FIQ (20-item questionnaire, higher scores indicate participant is more affected by fibromyalgia).
4Intention-to-treat not used, single study with small sample size.
5Small sample size.
 
Summary of findings 2. Acupuncture versus placebo or sham acupuncture for treating fibromyalgia

Acupuncture versus placebo or sham acupuncture for treating fibromyalgia

Patient or population: patients with fibromyalgia
Settings: USA, Switzerland1
Intervention: acupuncture versus placebo or sham acupuncture

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

Control (sham or placebo)Acupuncture (EA or MA or combined)

Pain up to 1 month after treatment
(VAS, NRS, SF-MPQ, MPI)2
Combined EA and MA
70 points
Scale (0 to 100), lower score means less pain)2
Combined EA and MA

0.14 SMD lower
(0.53 lower to 0.25 higher)

EA: 57 points

SMD -0.63 (-1.02 to -0.23)
286
(6 studies)
⊕⊕⊕⊝
moderate8
EA

AR %: -13% (-22% to -5%)

RR %: 22% (35% to 8%)

NNT 3 (2 to 9)

MA

AR % 0.28% (-0.34% to 0.90%)

RR % 4.00% (-4.86% to 12.86%)

NNT N/A

Physical function up to 1 month after treatment
(SF-36)3
MA

28 points
Scale (0 to 100)4
MA

22.2 points

(Higher score indicates
better physical function)

5.8 MD lower
(10.91 lower to 0.69 lower)
56
(1 studies)
⊕⊕⊕⊝
moderate9
AR % -5.80% (-10.91% to -0.69%)

RR % -15.21% (-28.61% to -1.81%)

NNT 4 (2 to 53)

Global well-being: rated by participants up to 1 month after treatment
(VAS, FIQ)4
Combined EA and MA

4.1 points
Scale (0 to 10)6
Combined EA and MA

0.29 SMD higher
(0.44 lower to 1.01 higher)

EA

SMD 0.65 higher (0.26 to 1.05)

(Higher score indicates better well-being)
200
(3 studies)
⊕⊕⊕⊝
moderate8
EA

AR % 11% (4% to 17%)

RR % 23% (9% to 38%

NNT 3 (2 to 9)

MA

AR % -8.00% (-17.20% to 1.20%)

RR % -20.00% (-43.00% to 3.00%

NNT N/A

Sleep up to 1 month after treatment
(VAS, subset FIQ (rest))5
Combined EA and MA
3.03 points:

Scale (0 to 10)8
Combined EA and MA

0.16 SMD higher
(0.29 lower to 0.61 higher)

EA
3.82 points

SMD 0.40 higher (0.01 to 0.79)

(Higher score indicates
better sleep)
200
(3 studies)
⊕⊕⊕⊝
moderate8
EA

AR % 8% (0.20% to 16%)

RR % 9% (0.21% to 17%)

NNT 5 (3 to 206)

MA

AR % -5.00% (-14.20% to 4.20%)

RR % -16.50% (-46.86% to 13.86%)

NNT N/A

Fatigue up to 1 month after treatment
(VAS, MFI, subset FIQ (fatigue))6
Combined EA and MA
7.77 points:

Scale (0 to 10)9
Combined EA and MA

0.1 SMD lower (0.81 lower to 0.61 higher)

(Lower score indicates
less fatigue)

EA

6.24 points

SMD -0.85 (-1.44 to -0.27)
201
(3 studies)
⊕⊕⊕⊝
moderate8
EA

AR % -15.30% (-25.92% to -4.86%)

RR % -20.13% (-34.11% to -6.39%

NNT 3 (2 to 8)

MA

AR % 4.34% (-1.34% to 10.1%)

RR % 5.59% (-1.72% to 13.11%)

NNT N/A

Stiffness up to 1 month after treatment
(Minutes, subset FIQ (stiffness))7
EA
6.6 points:

Scale (0 to 10)10
EA

5.7 points

(Lower score indicates
less stiffness)

0.45 SMD lower
(0.84 lower to 0.06 lower)
104
(2 studies)
⊕⊕⊕⊝
moderate8
AR % -9.00% (-16.80% to -1.20%)

RR % -13.24% (-24.71% to -1.76%)

NNT 5 (3 to 35)

Adverse events

Adverse events includes withdrawal and drop-outs that could be clearly identified as due to an adverse event
Study populationRR 0.44
(0.12 to 1.63)
289
(5 studies1)
⊕⊕⊕⊝
moderate10
AR % 44.00% (12.00% to 163.00%)

RR % 56.00% (88.00% to 63.00%)

1/6 people who had acupuncture reported adverse events

1/3 people who had fake treatments reported adverse events

Overall, such events were minor and lasted less than 1 day

372 per 1000164 per 1000
(45 to 606)

Moderate

83 per 100037 per 1000
(10 to 135)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

AR: absolute risk; CI: confidence interval; EA: electro-acupuncture; FIQ: Fibromyalgia Impact Questionnaire; MA: manual acupuncture; MD: mean difference; MFI: Multidimensional Fatigue Inventory; MPI: Multidimensional Pain Inventory; NNT: number needed to treat; NRS: numerical pain rating scale; RR: risk ratio; SF-MPQ: Short-Form McGill Pain Questionnaire; SMD: standardised mean difference; VAS: visual analogue scale

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1USA: Assefi 2005; Harris 2005; Harris 2008; Harris 2009; Martin 2006; Switzerland: Deluze 1992. Follow-up only Assefi 2005 (3 and 6 months) and Martin 2006 (1 and 7 months).
2Pain: Assefi 2005: VAS (10 cm, 0 = no pain, 10 = worse pain ever), Deluze 1992: VAS (1 to 100 mm, does say which is worse 1 or 100), Harris 2005: NRS (101 points, 0 to 100, 5-point increments, 0 = no pain, 100 = worse pain imaginable), Harris 2008 and Harris 2009 SF-MPQ (subset VAS), Martin 2006; MPI (generalised measure of pain, 61-item questionnaire, higher score means more pain). At up to 7 months follow-up (Assefi 2005 and Martin 2006), pain: 2.4 points lower on a 100-point scale (SMD -0.12; 95% CI -0.52 to 0.28, P = 0.55).
3Harris 2005: SF-36 questionnaire (score 0 to 100 with higher scores indicating better function).
4Global well-being: Assefi 2005: VAS (0 to 10, 0 = worse ever, 10=best ever), Deluze 1992: VAS (1 to 10, 10 = best), Martin 2006: FIQ (20-item questionnaire, higher scores indicate participant is more affected by fibromyalgia). At up to 7 months follow-up (Assefi 2005 and Martin 2006), global well-being: 6.7 points lower on a 100-point scale (SMD -0.03; 95% CI -0.87 to 0.81, P = 0.94).
5Sleep: Assefi 2005: VAS (0 to 10, 0 = worse ever,10 = best ever), Deluze 1992: VAS (1 to 10, 10 = best), Martin 2006: subset FIQ (rest). At up to 7 months follow-up (Assefi 2005 and Martin 2006), sleep: 1.8 points lower on a 100-point scale (SMD -0.09; 95% CI -0.44 to 0.26, P = 0.62).
6Fatigue: Assefi 2005: VAS (0 to 10, 0 = worse ever, 10 = best ever, Harris 2005: MFI (calculated using Reliability of Change Index, scores range from 4 to 20 with larger scores indicating more fatigue), Martin 2006: subset FIQ (fatigue). At up to 7 months follow-up (Assefi 2005 and Martin 2006), fatigue: 1 point lower on a 100-point scale (SMD -0.04; 95% CI -0.52 to 0.59, P = 0.90).
7Stiffness: Deluze 1992: morning stiffness (minutes), Martin 2006: subset FIQ (stiffness). At up to 7 months follow-up (Martin 2006), fatigue: 3 points lower on a 100-point scale (95% CI -1.60 to 1.00, P = 0.65).
8Deluze 1992: intention-to-treat not used.
9People who used acupuncture rated their physical function to be 4 points lower (absolute deterioration), small sample size.
10Small sample size (however some studies reported no adverse events).
 
Summary of findings 3. Acupuncture versus medication for treating fibromyalgia

Acupuncture versus medication for treating fibromyalgia

Patient or population: patients with fibromyalgia
Settings: China1
Intervention: acupuncture versus medication

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

Medication (antidepressant)Acupuncture

Pain up to 1 month after treatment

VAS2
MA

28.8 points

Scale

(0 to 100)
MA

11.5 points

(Lower score indicates less pain)

17.3 MD lower
(24.13 lower to 10.47 lower)
38
(1 study)
⊕⊕⊝⊝
low3
AR % -17.30% (-24.13% to -10.47%)

RR % -23.32% (-32.52% to -14.11%)

NNT 2 (lower 1, upper 3)

Physical functionNot reported

Global well-being: rated by participantsNot reported

SleepNot reported

FatigueNot reported

StiffnessNot reported

Adverse eventsStudy populationNot estimable38
(1 study)
⊕⊕⊝⊝
low3
No details were reported about adverse events, however from the data it would appear there were no drop-outs or withdrawals

See commentSee comment

Moderate

0 per 10000 per 1000
(0 to 0)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
AR: Absolute risk; CI: Confidence interval; MA: Manual acupuncture; MD: mean difference; NNT: number needed to treat; RR: risk ratio; VAS: visual analogue scale

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1Guo 2005, paper stated follow-up at 6 months but no data provided.
2VAS (0 = no pain, 10 = worse pain ever).
3Poorly reported paper (see 'Risk of bias' table).
 
Summary of findings 4. Acupuncture as an adjunct therapy for treating fibromyalgia

Acupuncture as an adjunct therapy for treating fibromyalgia

Patient or population: patients with fibromyalgia
Settings: Brazil1
Intervention: acupuncture as an adjunct therapy

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

Medication and exerciseAcupuncture plus medication and exercise

Pain up to 1 month after treatment

VAS2
MA

8 points

Scale (0 to 10)
MA

5 points

(Lower score indicates
less pain)

3.0 MD lower
(3.9 lower to 2.1 lower)
58
(1 study)
⊕⊕⊕⊝
moderate4
AR % -30.00% (-39.00% to -21.00%)

RR% -37.50% (-48.75% to -26.25%)

NNT 3 (lower 2, upper 10)

Physical function

MOS SF-363
Confirmed data not available at time of publication

Global well-being: rated by participantsNot reported

SleepNot reported

FatigueNot reported

StiffnessNot reported

Adverse eventsStudy populationRR 3.57
(0.18 to 71.21)
58
(1 study)
⊕⊕⊕⊝
moderate4
2 adverse events in acupuncture group (hand oedema at LI4 site)

0 per 10002 per 1000
(0 to 0)

Moderate

0 per 10000 per 1000
(0 to 0)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
AR: absolute risk; CI: confidence interval; MA: manual acupuncture; NNT: number needed to treat; RR: risk ratio

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1Targino 2008, follow-up 6, 12 and 24 months. At up to 7 months follow-up, pain: 5 points lower on a 100-point scale (95% CI -1.49 to 0.4, P = 0.32).
2VAS (10 cm, 0 = no pain, 10 = worst pain experienced).
3MOS SF-36 (Portuguese version, 8 multi-item scale measuring quality of life, higher values indicate better life).
4No control and single study with small sample size.
 
Summary of findings 5. Deep invasive acupuncture stimulation versus non-stimulated acupuncture for treating fibromyalgia

Deep invasive acupuncture stimulation versus non-stimulated acupuncture for treating fibromyalgia

Patient or population: patients with fibromyalgia
Settings: USA1
Intervention: deep needling with stimulation versus deep needling without stimulation

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

Control acupuncture: deep needling without stimulationAcupuncture: deep needling with stimulation

Pain up to 1 month after treatment
NRS2
MA

53.9 points

Scale (0 to 100)
MA

54.2 points

(Lower score indicates less pain)

0.3 MD higher
(18.34 lower to 18.94 higher)
41
(1 study)
⊕⊕⊕⊝
moderate5
AR % 0.30% (-18.34% to 18.94%)

RR % 0.57% (-34.55% to 35.68%)

NNT N/A

Physical function up to 1 month after treatment
SF-363
MA

40.2 points

Scale (0 to 100)
MA

34.7 points

(Higher score indicates
better physical function)

5.5 MD higher
(11.43 lower to 0.43 higher)
41
(1 study)
⊕⊕⊕⊝
moderate5
AR % 5.50% (-11.43% to 0.43%)

RR % 14.63% (-30.40% to 1.14%)

NNT N/A

Global well-being: rated by participantsNot reported

SleepNot reported

Fatigue up to 1 month after treatment

MFI4
MA

14.6 points

Scale (0 to 20)
MA

15.7 points

(Lower score indicates less fatigue)

1.1 MD higher
(1.41 lower to 3.61 higher)
41
(1 study)
⊕⊕⊕⊝
moderate5
AR % 5.50% (-7.05% to 18.05%)

RR % 6.74% (-8.63% to 22.11%)

NNT N/A

StiffnessNot reported

Adverse eventsContained in Table 2

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
AR: absolute risk; CI: confidence interval; MA: manual acupuncture; MD: mean difference; MFI: Multidimensional Fatigue Inventory; NNT: number needed to treat; NRS: numerical pain rating scale; RR: risk ratio

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1Harris 2005 compared one type of acupuncture (needles placed in traditional site with manual stimulation) with another type (needles placed in traditional site without stimulation). No follow-up.
2NRS rating scale (0 to 100 points, 5-point increments, 0 = no pain, 100 worse pain imaginable).
3SF-36 questionnaire (score 0 to 100, with higher scores indicating better function).
4MFI (calculated using Reliability of Change Index, scores range from 4 to 20 with larger scores indicating more fatigue).
5Single study with small sample size.