Acupuncture Therapy for Functional Effects and Quality of Life in COPD Patients: A Systematic Review and Meta-Analysis

Objective This study aimed to evaluate the efficacy and safety of acupuncture therapy (AT) for improving functional effects and quality of life in COPD patients. Methods PubMed, Embase, Cochrane Library, Web of Science, Chinese Biomedical Literature Database (CBM), China National Knowledge Infrastructure (CNKI), Chongqing VIP (CQVIP), and Wanfang Data were searched. The randomized controlled trials (RCTs) evaluating the effect of AT on COPD patients were included. Primary outcome measures included six-minute walk distance (6MWD) and St. George's Respiratory Questionnaire (SGRQ). Study selection, data extraction, and risk of bias assessment were independently conducted, respectively. Statistical analysis was conducted by RevMan software (version 5.3) and Stata software (version 12.0). Results Nineteen studies (1298 participants) were included. 6MWD improved more (MD: 47.84; 95% CI: 23.33 to 72.35; Z = 3.83, P = 0.0001) and effective rate was higher (OR: 2.26; 95% CI: 1.43 to 3.58; Z = 3.48, P = 0.0005) in the experimental group compared to the control group. Symptom domain scores (MD: −24.86; 95% CI: −32.17 to −17.55; Z = 6.66, P < 0.00001), activity domain scores (MD: −16.52; 95% CI: −22.57 to −10.47; Z = 5.36, P < 0.00001) and impact domain scores (MD: −13.07; 95% CI: −17.23 to −8.92; Z = 6.16, P < 0.00001) of SGRQ in the experimental group improved more compared to the control group. There was no significant improvement in SGRQ total scores between two groups. The improvement of FEV1 was not significant between two groups, yet subgroup analysis showed that patients treated with AT adjunctive to other treatments improved more in FEV1 (MD: 0.41; 95% CI: 0.28 to 0.54; Z = 6.01, P < 0.00001) compared to those treated with other treatments alone. Conclusion AT may be effective in improving functional effects and quality of life in COPD patients. Besides, AT may also improve pulmonary function of patients with COPD. However, further high-quality RCTs are needed to confirm the efficacy and safety of AT for COPD patients.


Introduction
Chronic obstructive pulmonary disease (COPD), a leading cause of morbidity and mortality, is characterized by progressive airflow obstruction, airway inflammation, and systemic effects or comorbidities and is projected to be the third leading cause of death worldwide by 2030 [1,2]. Since breathlessness, exercise limitation, and health status impairment broadly exist in patients with COPD, effective measures should be taken to improve symptoms, exercise tolerance, and health status based on an individualized assessment of disease [1]. Although appropriate pharmacologic therapy has effect in reducing COPD symptoms and the frequency and severity of exacerbations and improving health status and exercise tolerance [1], its cost and adverse effects can never be ignored.
Acupuncture therapy (AT), one of the most popular treatments in alternative medicine, has been proven to be cost-effective and safe in many conditions [3][4][5]. However, 2 BioMed Research International there is limited evidence concerning its efficacy and safety. One previous review showed that AT might result in clinically important improvements in quality of life and dyspnea of COPD patients, but it is outdated [6]. Moreover, the interventions of included studies involved point application therapy, acupressure, and transcutaneous electrical stimulation over acupuncture points (Acu-TENS), and these techniques may not genuinely reflect the efficacy of AT based on theories of traditional Chinese medicine. Therefore, the current review aims to evaluate the efficacy and safety of AT for improving functional effects and quality of life in COPD patients.

Inclusion and Exclusion Criteria.
We included randomized controlled trials (RCTs) in which the effects of AT on COPD patients were evaluated.
Participants had COPD defined as a clinical diagnosis of COPD, with a postbronchodilator fixed ratio of forced expiratory volume in 1 second (FEV 1 )/forced vital capacity (FVC) < 0.70 measured by spirometry, and those who had an acute exacerbation within four weeks before the study were excluded.
The intervention included AT, such as manual acupuncture, electroacupuncture, auricular acupuncture, and warm acupuncture, yet noninvasive techniques, such as single moxibustion, acupressure, point application, laser acupuncture, or Acu-TENS, were excluded.

Literature
Search. PubMed, Embase, Cochrane Library, Web of Science, Chinese Biomedical Literature Database (CBM), China National Knowledge Infrastructure (CNKI), Chongqing VIP (CQVIP), and Wanfang Data were searched from their inception to 31 December 2017. We developed detailed search strategies for each electronic database without language restrictions. Reference lists of eligible studies and previous systematic reviews were also reviewed to identify further eligible studies.

Study Selection. Two review authors (Yang Xie and
Xueqing Yu) independently examined titles and abstracts retrieved from the search and selected all potentially eligible studies. Then these full-text articles were obtained and the same review authors reviewed them independently against the inclusion and exclusion criteria. A third review author (Jiansheng Li) acted as an arbiter when consensus could not be reached.

Data Extraction.
Data extraction was independently conducted by two review authors (Yang Xie and Xueqing Yu) using a standardized data extraction sheet, involving information of authors, year of publication, study design, participants, intervention, comparator, and outcomes, with a third review author (Jiansheng Li) acting as an arbiter when disagreements existed between Yang Xie and Xueqing Yu.

Assessment of Risk of Bias.
Methodological quality was evaluated using the Cochrane tool for assessing risk of bias in RCTs [29]. Two review authors (Yang Xie and Xueqing Yu) independently assessed and scored each study with a third review author (Jiansheng Li) acting as an arbiter when disagreements existed.
2.6. Statistical Analysis. Statistical analysis was conducted by RevMan software (version 5.3) [30] and Stata software (version 12.0; StataCorp LP, USA). We summarized data using odds ratio (OR) with 95% confidence intervals (CI) for dichotomous outcomes and mean difference (MD) with 95% CI for continuous outcomes. If the data could not be combined into a meta-analysis, we summarized them in the text. We used a 2 test to estimate heterogeneity of both the MD and OR. Further analysis was performed using the I 2 test. A random-effect model was used to interpret the results if heterogeneity was statistically significant, whereas a fixedeffect model was used if heterogeneity was not statistically significant. We regarded heterogeneity as substantial when I 2 was greater than 50% or a low value ( < 0.10) was reported for the 2 test [31]. When more than 10 studies were included in the meta-analysis, we would investigate publication bias by funnel plots. In addition, a metaregression analysis was performed to explore potential associations between effect size and covariates of interest (publication year, region, intervention forms, sample size, and treatment period). If necessary, we conducted subgroup analysis to assess whether the treatment effects were different in different subgroups.

Data Extraction and Assessment of Risk of Bias.
A detailed description of the characteristics of included studies was outlined in Table 1. We determined the Cochrane "risk of bias" score for each study and this information was summarized in Table 2 and Figures 2 and 3. [9, 10, 16-18, 21, 23, 25] provided numerical data for 6MWD and were included in the meta-analysis. Analysis of the data indicated that there   was heterogeneity ( 2 = 65.96, < 0.00001; 2 = 89%); hence, a random-effect model was used. The pooled results showed that 6MWD in the experimental group improved more compared to the control group (MD: 47.84; 95% CI: 23.33 to 72.35; = 3.83, = 0.0001) (Figure 4).
Two studies [10,17] provided numerical data for symptom domain scores of SGRQ and were included in the metaanalysis. Analysis of the data indicated that heterogeneity was not statistically significant ( 2 = 0.05, = 0.83; 2 = 0%); hence, a fixed-effect model was used. The pooled results showed that symptom domain scores of SGRQ in the experimental group improved more compared to the control group (MD: −24.86; 95% CI: −32.17 to −17.55; = 6.66, < 0.00001) ( Figure 6).
Two studies [10,17] provided numerical data for activity domain scores of SGRQ and were included in the metaanalysis. Analysis of the data indicated that heterogeneity was not statistically significant ( 2 = 0.04, = 0.84; 2 = 0%); hence, a fixed-effect model was used. The pooled results showed that activity domain scores of SGRQ in the experimental group improved more compared to the control group (MD: −16.52; 95% CI: −22.57 to −10.47; = 5.36, < 0.00001) (Figure 7).
Two studies [10,17] provided numerical data for impact domain scores of SGRQ and were included in the metaanalysis. Analysis of the data indicated that heterogeneity was not statistically significant ( 2 = 0.00, = 0.97; 2 = 0%); hence, a fixed-effect model was used. The pooled results showed that impact domain scores of SGRQ in the experimental group improved more compared to the control group (MD: −13.07; 95% CI: −17.23 to −8.92; = 6.16, < 0.00001) (Figure 8).

mMRC.
Only one study [8] provided numerical data for mMRC scores; thus, the meta-analysis was not performed. Changes from baseline in mMRC scores in AT plus pulmonary rehabilitation (PR) group and PR group were −0.3 ± 0.5 and −0.3 ± 0.9, respectively. There was significant difference reported within AT plus PR group ( = 0.04).

Effective Rate.
Seven studies [7, 13-16, 19, 25] provided categorical data for effective rate and were included in the meta-analysis. Analysis of the data indicated that heterogeneity was not statistically significant ( 2 = 2.38, = 0.88; 2 = 0%); hence, a fixed-effect model was used. The pooled results showed that effective rate in the experimental group was higher compared to the control group (OR: 2.84; 95% CI: 1.59 to 5.06; = 3.53, = 0.0004) (Figure 15).

Reporting Biases.
We did not investigate publication biases by funnel plot because each comparison included not more than 10 studies.

Discussion
This systematic review provided a detailed summary of the current evidences related to the efficacy and safety of AT for functional effects and quality of life in COPD patients.
6MWD is an important measure of functional exercise capacity of patients with COPD. The distance walked is associated with clinical outcomes such as hospitalization and mortality, and its changes are used to evaluate the efficacy of therapeutic interventions such as pulmonary rehabilitation, surgery, and pharmaceutical management [33,34]. In this review, 6MWD in the experimental group improved more compared to the control group, and the MD was 47.84 Heterogeneity:  2 = 0.00, ＞＠ = 1 (P = 1.00); I 2 = 0% Test for overall effect: Z = 6.01 (P < 0.00001) Figure 14: AT adjunctive to other treatments versus other treatments alone, FEV 1 . Heterogeneity:  2 = 2.38, ＞＠ = 6 (P = 0.88); I 2 = 0% Figure 15: AT adjunctive to other treatments versus other treatments alone, effective rate. meters, which was greater than 25 meters, the minimal clinically important difference (MCID) of 6MWD for COPD patients [34]. This result might indicate the potential of AT in improving exercise capacity of COPD patients. Two subgroup analyses supported this result as well, and the MD of 6MWD change was 63.05 meters and 35.15 meters, respectively. SGRQ, another primary outcome measure in this review, is a well-established disease-specific instrument to measure quality of life for asthma and COPD. In this review, there was no statistically significant improvement in SGRQ total scores between two groups. However, MD of symptom domain scores, activity domain scores, and impact domain scores of SGRQ was 24.86 units, 16.52 units, and 13.07 units, respectively. Although there was no MCID available for each domain, each MD was at least three times greater than 4 units, the MCID for SGRQ total scores in COPD patients [35], and this might suggest the effect of AT on different aspects of health status in COPD patients. Subgroup analysis (AT adjunctive to other treatments versus placebo or sham acupuncture adjunctive to other treatments) supported these above results as well.

Experimental
FEV 1 is widely used by physicians in the diagnosis, classification, treatment, monitoring, and establishing prognosis for COPD patients. In this review, there was no statistically significant improvement in FEV 1 between two groups. However, subgroup analysis (AT adjunctive to other treatments versus other treatments alone) showed MD of FEV 1 change was 410 mL, which was four times greater than 100 mL, the MCID of FEV 1 for COPD patients [36]. And this result might suggest the potential of AT in improving pulmonary function in COPD patients.
mMRC is a major instrument to measure breathlessness. In this review, since mMRC scores were only available in one study, the meta-analysis was not performed. According to this study, change from baseline in mMRC scores in AT plus PR group and PR group was 0.3 units in both, and it was reported that there was significant difference within AT plus PR group. However, it was limited to support the effect of AT in improving breathlessness in COPD patients.
Effective rate, an important outcome measure in clinical studies of Chinese medicine, was also evaluated. In this review, effective rate in the experimental group was higher compared to the control group; to some extent, this might suggest that AT was a more effective treatment compared to other treatments. Importantly, subgroup analysis (AT adjunctive to other treatments versus other treatments alone) also supported this result with OR of 2.84.
Adverse effects were poorly reported in included studies. One study reported some minor adverse effects, and 5 studies reported no adverse effects. This might indicate the safety of AT for COPD patients.
There were some limitations in this study. Firstly, methodological quality of the included studies was generally low. For example, most of the included studies had high risk of performance bias. Secondly, most analysis of the data in the meta-analysis indicated that there was heterogeneity. Thirdly, there were various intervention forms of AT, which might make it difficult to evaluate the efficacy of AT alone.