Oral Huangqi Formulae for Stable Chronic Obstructive Pulmonary Disease: A Systematic Review and Meta-Analysis

Objective. To evaluate the efficacy and safety of oral Huangqi formulae for the treatment of stable COPD. Methods. The major databases were searched until September 2010 and supplemented with a manual search. Randomized controlled trials (RCTs) of oral Huangqi formulae that reported on lung function, St. George's Respiratory Questionnaire, symptom improvement and/or frequency of exacerbations were extracted by two reviewers. The Cochrane tool was used for the assessment of risk of bias in the included trials. Data were analyzed with RevMan 5.1.2 software. Results. 25 RCTs (1,661 participants) were included. Compared with conventional therapy (CT) alone, oral Huangqi formulae plus CT increased FEV1, and a similar result was found comparing Huangqi formulae with no treatment. Improvements in SGRQ total score, COPD-related symptoms and reduction of frequency of exacerbations were found in patients receiving Huangqi formulae plus CT compared to those receiving CT alone or CT plus placebo. No serious adverse events were reported. However, there were some methodological inadequacies in the included studies. Conclusions. The benefits of Huangqi formulae for stable COPD were promising, but its efficacy and safety have not been established due to methodological weakness and possible bias in the reported results. Further rigorously designed studies are warranted.


Introduction
Chronic obstructive pulmonary disease (COPD) is characterized by airflow limitation and manifests as progressive dyspnea accompanied by deterioration of lung function [1]. It is a major cause of morbidity, disability, and mortality. The World Health Organization (WHO) estimated that COPD ranked fifth in terms of burden of disease worldwide [2]. In China, COPD affects 8.2% of people aged 40 years or older [3]. It was the fourth leading cause of death in cities and the third in rural areas [4].
Studies have shown that pharmacotherapy cannot modify the trend of decline in lung function [5,6]. Chinese herbal medicine is commonly used for COPD, especially for the stable stage, in China and other Asian countries. Clinical studies suggest that herbal formulae that include Huangqi (Radix Astragalus membranaceus) were effective for stable COPD . It often serves as a principal medicine (i.e., the ingredient provides the principal curative action on the main syndrome or primary symptom [32]) in formulae for COPD. However, the quality of these studies had not been evaluated systematically, and some of the reports for the effects of Huangqi were conflicting. Therefore, this systematic review was conducted to evaluate the evidence for the efficacy and safety of oral Huangqi for treating stable COPD.

Search Strategy.
Comprehensive searches were performed for three English language databases and five Chinese databases, which included PubMed (from 1966), Embase (from 1985), Cochrane Central Register of Controlled Trials, Chinese Biomedical Database (CBM, from 1979), China National Knowledge Infrastructure (CNKI, from 1994), VIP medicine information system (VMIS, from 1989), Wanfang (from 1998), and TCM-Online (from 1949), from the inceptions of the databases to September 2010, without language restriction. A manual search was conducted of evidencebased medicine (EBM) reports on Chinese prescriptions by Japan Society for Oriental Medicine, EBM special committee [31]. The focus of the search was randomized controlled trials (RCTs) of oral Huangqi formula for stable COPD.
Search terms included chronic obstructive pulmonary disease, chronic bronchitis, emphysema, COPD, chronic obstructive lung disease, chronic obstructive airway disease, chronic airflow obstruction, traditional Chinese medicine, Chinese herbal drugs, complementary and alternative medicine, phytotherapy, herbs, herbal Medicine, Astragalus, Huangqi, Beiqi, Milkvetch Root, controlled clinical trial, and their synonyms. The literature was screened based on title, abstract, and full text as needed. Full details on the search strategy are described in the appendix.

Study Selection.
The inclusion criteria were as follows: (1) RCTs with patients diagnosed with COPD in the stable stage [1,33], which manifests as dyspnea, cough, and phlegm which remain stable or are rather mild; (2) Huangqi formula (taken orally as decoction, pill, powder, or capsule) alone or in combination with conventional therapy compared with placebo, no treatment, or conventional therapy as controls. Huangqi serves as a principal medicine, defined as follows: the properties of Huangqi are consistent with the main aims of the formula, or the dosage of Huangqi is relatively large (more than 15 g). Conventional therapy includes bronchodilators (beta2-agonists, anticholinergics, methylxanthine), corticosteroids, exercise training, smoking cessation, etc [1]. (3) Outcome measurements include spirometric parameters (forced expiratory volume in one second, FEV 1 ), St. George's Respiratory Questionnaire (SGRQ) total score, symptom improvement, and/or frequency of exacerbations. The exclusion criteria were (1) trials that included patients with asthma or other non-COPD disorders; (2) test interventions that were combined with other TCM therapies such as acupuncture, acupoint injection; (3) Chinese herbs or other TCM therapies were used in the control group. Chen) independently assessed studies based on the inclusion and exclusion criteria. If needed, two other reviewers (Lin Lin and Zehuai Wen) were consulted. Data on the details of study design, participants, interventions, control medicine, outcome measures, numbers of dropouts, and number and nature of any adverse events reported were extracted to a predefined form.

Data Extraction and Risk of Bias Assessment. Two authors (Lei Wu and Yuanbin
Assessment of symptom improvement was based on the chronic bronchitis section on the Guidance for Clinical Research on New Drugs of TCM [34], where responses were categorized into four levels (symptom control, very good, good, and no effect). The proportion of patient responses for the following symptoms (cough, sputum, dyspnea, and rale) was assessed according to the previous levels.
Two authors (Lei Wu and Yuanbin Chen) independently assessed the risk of bias of the included studies using the Cochrane tool [35]. Any discrepancies in assessment were decided by discussion. Other authors (Zehuai Wen and Xinfeng Guo) were consulted to make the final decision when needed. To verify unclear information on methodology and therapy, attempts were made to contact the authors of the original papers via phone, email, or mail. If the authors were not contactable after 3 times by phone or email, they were sent mail and given one month to reply.

Data
Analysis. Data were analyzed by RevMan 5.1.2 (Cochrane Collaboration), and Stata 12.0 software (StataCorp LP, College Station, TX, USA). Dichotomous data were presented as risk ratio (RR) and continuous outcomes as mean difference (MD), with 95% confidence intervals (95% CI). Statistical heterogeneity was assessed by Cochrane's Q test. If the analysis showed low heterogeneity ( ≥ 0.10 and 2 ≤ 50%), data were synthesized using a fixed-effects model. Otherwise, a random-effects model was applied. Subgroup analyses were performed if sufficient numbers of RCTs were available and sensitivity analyses were undertaken as required. Publication bias was assessed by funnel plot analysis if the group included more than five trials [36]. Egger's test was conducted if it was difficult to determine whether the funnel plots were symmetrical.

Description of Studies.
Of 13,254 potentially relevant citations, 8,952 were excluded as duplications, 4,012 were excluded for not meeting the inclusion criteria after reading the titles and abstracts, and further 265 studies were excluded after reading the full articles and/or contacting the authors. Finally, 25 studies (including 1,661 participants) that met all the selection criteria were included ( Figure 1).
Of these 25 studies, 24 were conducted in China  and 1 in Japan [31]. Twenty-two were published and 3 were dissertations [20,22,26]. All patients were diagnosed as stable COPD according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) or guidelines issued by Committee of Respiratory Disease, Chinese Medical Association. Disease severity in the included trials ranged from mild to very severe COPD. The duration of patients' COPD ranged from 4 to 30 years.
Nineteen studies compared Huangqi formulae plus conventional therapy with conventional therapy [7, 9-11, 14-19, 22, 24-31]. Two studies compared Huangqi formulae plus conventional therapy with placebo plus conventional therapy [12,21]. Three studies compared Huangqi formulae with no treatment [8,13,23]. One study compared a Huangqi formula with conventional therapy [20]. The duration of treatment varied from 2 weeks to 6 months. One trial reported a 1-year follow-up period [16].  Five studies indicated that Huangqi was used as the principal medicine [12,17,18,20,26]. In the twenty remaining studies, we considered Huangqi to be the principal medicinal when its properties were consistent with the main aims of the formulae or its dosage was more than 15 g, even when the authors did not indicate that it was used as a principal medicinal. The characteristics of the included studies are summarized in Tables 1 and 2. 3.2. Risk of Bias of the Included Studies. After attempts at verification by contacting the authors of the original papers through phone, e-mail, or mail, 14 of the included trials still lacked a detailed description of the method of randomization, and 19 trials did not indicate whether blinding was applied. Only two trials described the method of generation of randomization sequence, allocation concealment, double blinding, and placebo manufacturing method. Two trials reported dropouts; one of them explained the reason was loss of contact with the participants, and the other did not provide reasons. No trials reported whether they had used intention-to-treat analysis. Selective outcome reporting was judged as low risk of bias in 2 trials, uncertain risk in 19 trials, and high risk in 4 trials. No study had a registered or published protocol; so, judgment was based on the outcome measures specified in the method section of the study and was informed by discussion with the author when possible. Most of the included trials lacked a sample size calculation ( Figure 2).

Publication Bias.
The number of studies reporting SGRQ scores and the frequency of exacerbations was less than five; so, a funnel plot was not applicable. The characteristics of studies reporting FEV 1 were different, and the number of trials in each subgroup was less than five; so, funnel plots were also not applicable. For symptom improvement, the symmetry of the funnel plot was not clear; so, Egger's test was conducted. It indicated that the effect of publication bias was not significant ( = 1.39, 95% CI, −0.69 to 2.48, = 0.215) ( Figure 3).

Spirometric Parameters.
Subgroup analyses were performed according the type of comparison (Figure 4). Subgroup 1 compared Huangqi formulae plus conventional therapy with conventional therapy alone [11, 16-19, 25, 27, 28, 30]. trials, but there was high heterogeneity in this subgroup. So, further analyses were performed according to duration of treatment as follows: more than, equal to, or less than 3 months. The former 2 categories had low heterogeneity. There was a significant difference in FEV 1 in the trials which had durations of more than 3 months (MD 0.33, 95% CI: 0.20 to 0.46) [17,19,30] and no significant difference in those of 3 months duration (MD 0.07, 95% CI: −0.07 to 0.20) [18,28]. For trials of less than 3 months [11,16,25,27], FEV 1 increased (MD 0.36, 95% CI: 0.29 to 0.43), but there was heterogeneity. So, sensitivity analysis was conducted. It produced low heterogeneity with a significant difference (MD 0.49, 95% CI: 0.41 to 0.57) when one trial [16] was removed because the age distribution of participants differed from the other trials.

Frequency of Exacerbations.
Four trials reported frequency of exacerbations at pre-and posttreatment [16,17,21,24]. Compared with placebo plus conventional therapy or conventional therapy alone, the frequency of exacerbations in patients receiving Huangqi formulae plus conventional therapy was significantly reduced (MD −0.97, 95% CI: −1.57 to −0.37), but the heterogeneity was high (Figure 7). This Study or subgroup 11       appeared due to variation between trials in terms of the control interventions used, duration, and trial quality. Overall, the better result appeared to be of He et al., 2010 [21].

Adverse Events.
Six trials reported that no adverse events occurred [7,8,11,21,26,31]. One trial (60 participants) reported hoarseness in the control group treated by salmeterol/fluticasone [28], but there was no causality assessment for this adverse event. The other trials did not report adverse events; so, there was insufficient data to assess whether the combination of Huangqi formulae plus conventional therapy affected the adverse event rate.

Discussion
This systematic review includes 24 RCTs conducted in China and 1 RCT in Japan. All RCTs employed the herb Huangqi as a principal herb in the herbal formulae used in the test arms,  and all studies only included patients assessed as suffering stable COPD. The comparators were mostly conventional therapy, but this was variable and was not clearly specified in some studies. This is likely to have been a source of heterogeneity in the meta-analyses, but it also reflects usual care since patients with stable COPD would typically receive a variety of conventional therapies which may be varied according to response and individual need. Eight studies were of six months, while the others were of shorter treatment duration and only one had a followup at one year; so, the results can only be considered relevant to the relatively shortterm management of COPD.
The meta-analysis results indicate that the use of Huangqi formulae could significantly improve lung function measured as FEV 1 when compared with no treatment based on two studies, and it produced an additional improvement when combined with conventional therapies based on 9 studies, four of which were of over three months duration. The incidence of exacerbations also appeared to decline when Huangqi formulae were combined with conventional therapies. Improvements in quality of life based on SGRQ were evident, but this was based on only three studies. Also, these Huangqi-containing formulae appeared to be well tolerated, even when combined with conventional medications, since seven studies reported that no adverse events were noted.
All the included trials demonstrated at least some methodological deficiencies leading to potential risks of bias. Only eleven provided evidence of adequate randomization procedures, and only three were effectively blinded to participants and investigators. Consequently, the results should be interpreted with caution. Therefore, the potential benefits of oral Huangqi formulae for stable COPD need to be further appraised through trials that employ rigorous methodology and include adequate assessment of the safety profiles of the interventions. In addition, we found that the reporting of trial methods and procedures was frequently unclear and insufficient. Therefore, we suggest that all reports of RCTs published in China should be required to comply with the CONSORT statement [37] and the publication of protocols should be encouraged.
There has been increasing interest in complementary and alternative medicine (CAM) for the treatment of COPD, especially the use of Chinese herbal medicines [38,39]. A recent cross-sectional study in Australia suggested that nearly one in five (17.3%) individuals with moderate to severe COPD had used some form of herbal preparation [39]. Therefore, reviews of the state of the evidence base are essential.
From viewpoint of TCM, patients with stable COPD usually manifest with Qi-deficiency syndrome [40]. One of the characteristics of Qi-deficiency is that the patient easily suffers from colds which commonly lead to acute exacerbations of COPD. Huangqi is one of the principal herbs used for reinforcing Qi. It has been widely used for preventing and alleviating common colds; so, its clinical use in COPD is predicated on a putative benefit in preventing colds and reducing COPD exacerbations.
From the experimental perspective, one line of research into Huangqi has focused on its effects on inflammation. The main pathological characteristic of COPD is chronic airway inflammation involving a number of proinflammatory mediators and cytokines. Also, oxidative stress is increased in COPD, which amplifies inflammation and may result in corticosteroid resistance. An invivo study suggests that Huangqi may reduce inflammatory infiltration and inhibit the inflammatory response in the airway through downregulating the expression of TNF-and IL-8 [41]. Flavonoids in Huangqi may protect the erythrocyte membrane from attack by free radicals and appear to eliminate free radicals [42]. One study has shown that airway inflammation induced by cigarette smoke was reversed by astragaloside IV (AST IV), another active constituent of Huangqi, in a dose-dependent fashion. This effect appeared due to its anti-inflammatory and antioxidant properties, including NF-B inactivation [43]. Other studies suggested that AST IV possesses antiinflammatory and immune regulation activity and can be used for preventing asthma attacks [44]. The previous pharmacological properties of Huangqi may at least partially explain the clinical benefits reported by the studies included in this paper.
An earlier paper [45] evaluated the effects of a diverse range of herbal medicines in COPD, and a subsequent paper narrowed the focus to ginseng and ginseng-containing formulae [46]. The strengths of this paper are it focuses only on formulae that share the same principal ingredient and there is experimental evidence that supports the application of this herb as a modulator of inflammation.
The main limitations to this paper are the potential sources of bias due to methodological defects and inadequacies in reporting. Although publication bias was not a major issue, the previous issues potentially lead to overreporting of positive results, selective reporting of outcome measures, and underreporting of adverse events. Also, the use of a diversity of conventional therapies as comparators makes it difficult to assess the magnitude of any effects and to interpret their clinical significance.
Therefore, the potential benefits of oral Huangqi formulae for stable COPD evident in this paper need to be further appraised through suitably powered clinical trials that employ standardized conventional therapies as comparators over a sufficient period to determine whether any effects are of sufficient clinical relevance to warrant modification to current best practice for the management of stable COPD.

Conclusions
Oral Huangqi formulae appear beneficial in terms of improving lung function, quality of life, and symptoms and in reducing the incidence of exacerbations for patients with stable COPD, but these apparent benefits require further appraisal through higher quality trials that strictly adhere to methodological principles and procedures.