To determine the efficacy of acupuncture in the treatment of patients with asthma.

MEDLINE, Biological Abstracts and Dissertation Abstracts, and references from published papers and narrative reviews, were searched from 1970 to 2000. The keywords for the search were reported. In addition, the authors of eligible reports were contacted for further published and unpublished articles.

Study designs of evaluations included in the review

Randomised clinical trials (parallel and crossover trials) were eligible for inclusion.

Specific interventions included in the review

Studies comparing real and placebo acupuncture (i.e. acupuncture at real and placebo points) were eligible for inclusion. No further details were reported.

Participants included in the review

People with asthma were eligible for inclusion in the review. The mean age of the participants in the included studies ranged from 11 to 43 years and the mean duration of asthma, where specified, ranged from 16 to 23 years.

Outcomes assessed in the review

Studies measuring either the peak expiratory flow rate (PEFR), forced expiratory volume in one second (FEV1) or forced vital capacity (FVC) were eligible for inclusion.

How were decisions on the relevance of primary studies made?

The authors did not state how the papers were selected for the review, or how many reviewers performed the selection.

A questionnaire addressing study design, statistical analysis and presentation of results was designed to assess the quality of the included studies. A score was assigned to each section: 27 for study design, 10 for statistical analysis and 2 for presentation of the results. The maximum score possible was 39 points. Four biostatisticians, blinded to the author and source of publication, independently assessed the validity of the eligible studies.

The authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction. The extracted data included the number of participants, gender, age, duration of asthma, use of blinding, duration of treatment, the number of excluded patients, if other medication was being taken, study design and the method of outcome assessment. Data for these categories were not available for all of the included studies.

Individual patient data were available for 3 trials, with summary data used from the remaining 8 trials. Standardised mean differences (SMDs) were calculated for the individual studies. In crossover studies the correlation coefficient was used to compute the variance of the mean difference. A weighted combination of correlation coefficients was used where there were insufficient data to calculate a correlation coefficient. A scanner and technical drawing programme were used to estimate the means and standard errors (SEs) from graphs. Where a series of repeated measurements were reported, the maximum mean change was used.

How were the studies combined?

The SMDs were combined using a fixed-effect model, weighted by the inverse of the variance. When heterogeneity was not explained by an identifiable cause, the treatment effect was assessed using a random-effects model (a more conservative analysis giving wider confidence intervals). The authors reported carrying out two analyses, one on PEFR (using FEV1 where PEFR was unavailable) and the other on FEV1 (using PEFR where FEV1 was unavailable).

How were differences between studies investigated?

Statistical heterogeneity was determined using the Q test. The sums of squares were partitioned into two parts, one related to between-subgroup differences and the other to within-subgroup differences.

Twelve studies met the inclusion criteria, of which 11 (n=202) were included in the review: 4 parallel (n=85) and 7 crossover (n=117) trials. A paper published in Chinese was excluded from the analysis due to translation problems.

The validity of the included trials was reported to be limited in terms of sample size, missing information, adjustment of baseline characteristics, method of placebo acupuncture and possible publication bias.

FEV1.

Based on 5 of the 6 studies reporting FEV1, acupuncture resulted in a mean difference of 3.53 (95% confidence interval, CI: 0.44, 6.62) and a SMD of 0.17 (95% CI: -0.05, 0.39) in FEV1. There was a large range of effect when the mean difference was used, and the pooled estimate crossed the line of no effect when the SMD was used. There was no statistically-significant heterogeneity.

PEFR.

The effect of acupuncture on PEFR was variable, being beneficial in 4 studies and detrimental in one; the effect size for SMD ranged from d=-1.293 (SE=0.536) to d=0.560 (SE=0.495). The effect of acupuncture on PEFR was not reported as a meta-analysis.

FVC.

No results for the effect of acupuncture on FVC were presented.

The overall treatment effect was d=0.12 (95% CI: -0.07, 0.31), which corresponds to an approximate difference in FEV1 means of 1.7 (95% CI: -1.3, 4.7).

Induced bronchoconstriction.

The authors reported a beneficial effect of acupuncture in people where bronchoconstriction was induced (estimated effect 0.3, 95% CI: 0.04, 0.56), which was not seen in patients with uninduced bronchoconstriction (estimated effect -0.08, 95% CI: -0.28, 0.20).

Overall, the review did not provide evidence of the efficacy of acupuncture in the treatment of patients with asthma. However, a small effect may have been observed in those where bronchoconstriction was induced.

The authors made attempts to identify both published and unpublished data. The criteria for selection and the validity assessment were stated clearly. The funnel plot showed a bias towards small studies with positive results. Of the three outcomes the authors stated to have been measured (PEFR, FEV1 and FVC), the authors concentrated on FEV1; PEFR was only addressed in a meta-analysis combined with FEV1, while FVC was not addressed at all, even though 4 studies reported FVC and another VC. The extraction of the largest difference from trials with repeated measurements, and the estimation of variances, may also have introduced bias into the results. In addition, the exclusion of a Chinese trial may have had an impact on the findings of the review. The authors' conclusion was appropriate given the small sample size, the presence of publication bias and heterogeneity between the trials, which resulted in the analyses giving either large CIs or the CIs crossing the line of no effect.

Practice: The authors did not state any implications for practice.

Research: The authors stated the need for a large, randomised clinical trial.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.