Clinical effects of a standardized Chinese herbal remedy, Qili Qiangxin, as an adjuvant treatment in heart failure: systematic review and meta-analysis

The following electronic databases were searched from date of inception to March 2015: PubMed, the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (Issue January, 2015), China National Knowledge Infrastructure (CNKI), VIP Database, Sino-Med Database, and Wanfang Database. We used the following search terms: (“Qili Qiangxin” OR “qiliqiangxin” OR “qiangxinli”) and (“heart failure” OR “cardiac failure” OR “heart decompensation”). We searched for trials from mainstream registries including Current Controlled Trials (http://​www.​controlled-trials.​com), the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP; http://​apps.​who.​int/​trialsearch/​), ClinicalTrials.gov trials registry (http://​www.​clinicaltrials.​gov), the Australian New Zealand Clinical Trials Registry (http://​www.​anzctr.​org.​au), and Centre Watch (http://​www.​centerwatch.​com). We also hand-searched the reference lists of all full text papers for additional relevant reports. No language restrictions were imposed.

We accepted RCTs regardless of blinding procedures and included only parallel design studies. Only human studies were included in this review and we required the use of internationally-accepted criteria for diagnosis of HF. We placed no other requirements of the participant population in terms of gender, age, etiology, ethnic group, severity or course of disease. Only Qili Qiangxin capsules, composed of the aforementioned 11 Chinese herbs, were accepted as the intervention. This could be used alone or alongside appropriate control treatments such as placebo, conventional treatment or no treatment, and trials with any other Chinese herbal medicine in control group will be excluded. Our primary outcome measures were all-cause mortality or cardiovascular mortality due to HF, major cardiovascular events such as MI, outpatient visits, hospitalizations or re-admission for HF. Secondary outcomes measures were quality of life (QOL) measured by Minnesota Living with Heart Failure Questionnaire (MLHFQ), New York Heart Association (NYHA) functional classification, echocardiography measurements, six-minutes walking distance (6MWD), plasma amino-terminal pro-brain natriuretic peptide (NT-pro-BNP). We also collected safety and adverse events data. We included only RCTs reporting one or more of these outcomes.

Two authors (from J Sun, K Zhang, WJ Xiong, and YJ Zhang) independently identified articles for eligibility with any disagreements resolved through discussion with a third party (JP Liu). Two authors independently extracted data and which included patient characteristics, details of the intervention and control groups, outcome measures and main results. The same process was used to assess the methodological quality of included RCTs using the risk of bias tool as described in the Cochrane Handbook for Systematic Reviews of Interventions [17]. This process requires seven criteria to be assessed: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, incomplete outcome data, free of selective reporting, and other bias.

Statistical analyses were performed by using RevMan 5.2 software (The Cochrane Collaboration). Pooled risk ratio (RR) with 95 % confidence interval (CI) of dichotomous outcomes was used to estimate report effect. Continuous data was presented as mean difference (MD) with 95 % CI. We used a fixed effect model unless there was evidence of heterogeneity. Heterogeneity was assessed using both the Chi-squared test and the I-squared statistic, and we considered an I-squared value greater than 50 % to be indicative of substantial heterogeneity. If missing data from the original trials are available, intention-to-treat analysis will be applied for primary outcome.

Our search in March 2015 from six databases and other sources identified 1,390 potentially eligible articles. After removing duplicates and reviewing full text articles, we eventually included 129 randomized clinical trials published as 131 articles (Fig. 1). All 129 studies were included in our qualitative synthesis and of these, 117 studies were eligible for our meta-analysis.

Included trials were published between 2005 and 2015 with a larger proportion of trials published in 2013 (n = 41 trials, 31.78 %) and 2012 (n = 23 trials, 17.83 %). The 129 trials involved 11,547 patients diagnosed with HF with an age range of 18 to 98 years. Sixteen trials enrolled only elderly patients over the age of 60 (n = 16, 12.40 %). The majority of trials included both male and female patients (n = 127 trials, 98.45 %). Two trials included only male patients (Wei XB 2013, Miao S 2013); two trials provided gender details only for participants who completed the trial (Li XL 2013, Kuang JB 2008). 58.23 % of total populations were male except one trial failed to provide gender detail (Tao X 2011).

Included trials used the following diagnostic criteria: American College of Cardiology/American Heart Association (ACC/AHA) guidelines, European Society of Cardiology (ESC) guidelines, World Health Organization/International Society and Federation of Cardiology (WHO/ISFC) guidelines, World Health Organization/International Society of Hypertension (WHO/ISH) guidelines, Chinese Medical Association (CMA) guidelines, guidelines issued by Ministry of Health (MoH) in China, Framingham criteria, or textbook criteria which were consistent with internationally-used diagnostic criteria .

One hundred seventeen trials (s1-s118, Additional file 2) involving 10,170 patients compared Qili Qiangxin capsules plus conventional treatment to conventional treatment alone. Three trials (s119-s121, Additional file 2) involving 702 patients compared Qili Qiangxin capsule plus conventional treatment to placebo plus conventional treatment. Nine trials (s122-s131, Additional file 2) included 675 patients comparing Qili Qiangxin capsules plus conventional treatment to recommended pharmaceuticals (including thiazide diuretics, benazepril, captopril, metoprolol, irbesartan, hydrochlorothiazide and digoxin plus conventional treatment). Almost all patients in the included trials presented with comorbidities such as hypertension, hypotension, valvular heart disease, pulmonary heart disease, cardiomyopathy, diabetes mellitus, chronic renal insufficiency, coronary heart disease, acute MI and sinus bradycardia. The duration of treatment ranged from 1 to 12 months, and the most commonly adopted was 1 month (n = 54 trials). Nineteen trials had 6 or more months of treatment.

The main outcome in the majority of trials was reported to be left ventricular ejection fraction (LVEF) (n = 91 trials). Other primary outcomes adopted included mortality (n = 6 trials), major cardiovascular events (n = 3 trials) or re-hospitalization (n = 9 trials). Details of each study are listed in Table 1.

The methodological quality of included trials was generally low as most failed to provide protocols (Additional file 3). Nineteen trials (n = 2388 participants, 20.68 %) were judged as low risk on random sequence generation. Of these twenty, 17 trials used random number tables (n = 1774 participants, 72.29 %), one trial used statistical software (n = 512 participants, 21.44 %) and one trial used drawing lots. Only one (Li XL 2013) trial reported allocation concealment was judged as low risk; one trial (Yang F 2007) described using an envelope but failed to provide further details and was subsequently judged as presenting with unclear risk of bias.

We judged five trials on the use of blinding methods. Of these, three double-blind placebo controlled trials (Yu JH 2008, Li XL 2013, Yu JH 2008) were judged as low risk (n = 702 participants, 6.08 %). For the remaining two trials, one single-blind trial (Guo WB 2013) and one double-blind trial (Wei XB 2013) did not mention who was blinded and were judged as unclear risk. Besides, only one trial (Li XL 2013) reported the blinding of outcome assessment (n = 512 participants, 4.43 %).

Twenty-four trials failed to report all outcomes listed in the methods. The remaining trials reported all outcomes as specified in methods, although only one provided a study protocol. Insufficient information was available for us to conduct a risk of bias assessment in terms of selective reporting. In terms of other biases, only 122 trials reported comparability of baseline data and only one study (Li XL 2013) reported conducting a sample size calculation (Fig. 2).

A funnel plot analysis was generated for 50 trials comparing Qili Qiangxin plus conventional treatment to conventional treatment alone for the outcome of NYHA levels. No asymmetry was observed, suggesting no publication bias (Fig. 4).

The findings of this review suggest that in patients with HF, Qili Qiangxin used as an adjunct to conventional treatment leads to a statistically significant reduction in major cardiovascular events and re-hospitalization due to HF when compared with conventional treatment alone. However, there appeared to be no additional effects of Qili Qiangxin in terms of mortality rate. Qili Qiangxin plus conventional treatment also appears to significant improvements in cardiac function measured by levels of NHYA and showed significant beneficial effects on NT-proBNP levels, QOL, LVEF and 6MWD. Pooled data indicated that Qili Qiangxin as an adjuvant treatment have clinical significance in improving exercise capacity as well as symptomatic status. Preliminary data suggests that Qili Qiangxin appears to be safe. However, it should be noted that the majority of included trials had low methodological quality and a high risk of bias. Any conclusions drawn from this review should therefore be interpreted with caution.

This study presents the first comprehensive and rigorous review of Qili Qiangxin as a supplementary treatment for patients with HF. Previous criticisms have been made regarding the use of different terms to describe types of HF potentially being confusing [7]. It is strength of this review that we included all types of HF with different comorbidities in order to maximize the impact and clinical relevance of our findings.

There are a number of limitations to this review which need to be acknowledged. The majority of included studies were assessed as unclear risk of bias. This was largely owing to the lack of details, particularly in terms of random sequence generation, allocation concealment, blinding methods and availability of a protocol. Due to limited resources and time constraints we were not able to contact trial authors to request missing data and other information preventing us from being able to make a complete assessment regarding risk of bias. However, the sensitivity analysis indicated that the missing data not change the results in this review. Furthermore, the follow-up period in the included studies was no longer than six months, with the majority reporting a follow-up of three months or less. This makes it difficult to interpret the present evidence on mortality as an outcome, and in assessing the long-term effects and safety of administering Qili Qiangxin. We excluded 40 articles on the basis of no relevant outcomes reported. These articles did not report the outcomes we listed in protocol, instead, they chose composite outcome indicator. We excluded these articles as they failed to provide data of separate components. Considering all studies excluded for this reason reported positive results for the composite outcome, we believed the exclusion was unlikely to affect the results of our review.

A recent editorial has suggested that Qili Qiangxin showed promising results. If Qili Qiangxin is shown to be safe and effective from further rigorous clinical trials research, this presents an interesting area of further work that may fundamentally challenge our current need to precisely understand the pharmacodynamics of all drug therapies [19]. Herbal medicines appear to operate through a variety of often poorly defined synergistic mechanisms involving multiple chemical components. In our study we found that Qili Qiangxin capsules have a positive effect on NT-proBNP levels. Previous studies showed that NT-proBNP level can be used as a prognostic marker for congestive heart failure as decreased NT-proBNP levels predicted reducing mortality in 10 years [20‐22]. Levels of NT-proBNP clearly differ among various congenital heart lesions, and a higher level of NT-proBNP correlates with diastolic dysfunction parameters. NT-proBNP levels are related to exercise capacity and also increase with the more dysfunctional HF stages [23]. In our study, we included patients regardless of gender, age, etiology, ethnic group, severity, and course of diseases. Consequently the patients’ in different trials had different underlying diseases. All of these factors might explain the high heterogeneity in the meta-analysis of NT-proBNP. The high heterogeneity of LVEF and 6MWD might separately due to the different population baseline of LVEF and 6MWD in the included trials.

The results from this review suggest that further research is warranted in order to provide further evidence assessing the effects and safety of Qili Xiangxin as an adjuvant to conventional treatments for HF. We have a number of recommendations for future research. Various diagnostic criteria are used internationally for HF and future research should ideally use internationally recognized diagnostic criteria such as the ACC/AHA guidelines or ESC guidelines as part of their inclusion criteria. Further studies of Qili Qiangxin should also incorporate a minimum one year follow-up period in order that clinically important data on outcomes such as mortality and cardiac events can be provided. These were rarely reported amongst the RCTs we found in this review and further data in this area would be clinically meaningful to patients and providers. Furthermore, we found the reporting of clinical trial methods such as random sequence generation and allocation concealment inadequate and we recommend researchers report in full their trial methodology in future publications. Linked to this, none of the RCTs we included in this review provided trial protocols, and some did not provide all outcomes that had been described in the methods section. For transparency, we recommend that researchers prospectively register trials, publish trial protocols and cite the protocol or registration number in subsequent publications. This will enable future researchers and guideline developers to consider the evidence presented in light of what had been planned by the research team prior to trial commencement. Finally, we found few RCTs using placebo-control design in this review. Studies in future should adopt a double-blinded placebo-controlled design in order that further information regarding specific effects of Qili Qiangxin in HF can be provided.