Zhishi Xiebai Guizhi Decoction for coronary heart disease: A systematic review and meta-analysis

Abstract

Background:

Coronary heart disease (CHD) is a type of cardiovascular disease (CVD) caused by coronary atherosclerosis. It is a main cause of medical burden and cardiovascular related death. Zhishi Xiebai Guizhi Decoction (ZXGD) is a representative prescription of traditional Chinese medicine (TCM) in the treatment of CHD, but there is poor systemically evidence-based appraisal.

Objective:

To evaluate the efficacy and safety of ZXGD for CHD.

Methods:

Eight databases were retrieved for randomized controlled trials (RCTs). Data was extracted independently by 2 reviewers. The quality of the included studies was assessed by Cochrane Collaboration risk of bias tool. Clinical efficacy, blood lipid, vascular endothelial function, inflammatory factor and homocysteine (Hcy) were prespecified outcome measures.

Results:

Twenty-four studies (2272 patients) were included. Meta-analysis showed that compared with conventional western medicine (WM) alone, ZXGD was associated with a greater symptom improvement rate with a relative risk (RR) of 1.21 [95% CI (1.16, 1.26), P < .00001] and a greater electrocardiogram (ECG) improvement rate with a RR of 1.27 [95% CI (1.16, 1.40), P < .00001]. In terms of blood lipid, ZXGD reduced total cholesterol (TC) with a mean difference (MD) of −1.15 [95%CI (−1.75, −0.55), P = .0002] and triglyceride (TG) [MD = −0.72, 95%CI (−0.99, −0.45), P < .00001], reduced low-density lipoprotein cholesterol (LDL-C) [MD = −0.93, 95% CI (−1.17, −0.69), P < .00001], and increased high-density lipoprotein cholesterol (HDL-C) [MD = 0.31, 95%CI (0.20, 0.42), P < .00001]. In terms of vascular endothelial function, ZXGD decreased the level of endothelin-1 (ET-1) [MD = −7.81, 95%CI (−9.51, −6.10), P < .00001], and increased nitric oxide (NO) [MD = 8.90, 95%CI (7.86, 9.93), P < .00001]. ZXGD also reduced high-sensitivity C-reactive protein (hs-CRP) [MD = −1.73, 95% CI (−2.63, −0.83), P < .00001] and Hcy [MD = −2.03, 95%CI (−2.78, −1.28), P < .00001]. No significant differences were found in adverse event rate between the 2 groups with a RR of 0.77 [95% CI (0.44, 1.34), P = .36].

Conclusion:

ZXGD is effective and safe in the treatment of CHD. However, more rigorous and high-quality RCTs are needed to verify the conclusion.

1. Introduction

CVD is one of the leading causes of death and disability in human beings. Since 1990, the number of CVD deaths has increased steadily from 12.1 million to 18.6 million in 2019.^[1] CHD makes up almost one-third of the global CVD burden.^[2] CHD is caused by coronary artery stenosis or blockage due to coronary atherosclerosis, resulting in myocardial ischemia,hypoxia, and even necrosis. Patients can present with varying degrees of nausea, chest tightness, chest pain and dyspnea, and even death due to cardiac arrest.^[3] As a result of the increasing morbidity and mortality, CHD has become an important global public health problem. In recent years, the COVID-19 pandemic has directly or indirectly increased the burden of CVD, putting those who suffer from CHD at a higher risk of dying.^[4]

Treatment of CHD includes lifestyle modification, risk factor management, medications, and invasive procedures (percutaneous or surgical) with the primary goals of relieving symptoms, combating ischemia, restoring cardiac function, and improving prognosis.^[5] Currently, available clinical medicine includes antiplatelet agent, β-blocker, calcium channel blocker, nitrate, angiotensin-converting enzyme inhibitor and statin. In addition to the conventional medicine used clinically, new drugs are in research and development. Percutaneous coronary intervention (PCI), as the most active and fastest developing cardiac invasive treatment for nearly 20 years, acts on directly relieving mechanical stenosis of the coronary arteries, reducing the burden of treating complex CHD.^[6] In the wake of developments in biotechnology and tissue engineering, new therapeutic strategies for CHD such as nanotechnology, and robotic surgery, as well as 3D printing technology are in the ascendant.^[7–9]

However, despite the gratifying improvements in CHD treatment over the past few decades, the global burden of ischemic heart disease remains rising. Besides, adverse reactions of western medicine, and thrombolytic complications and restenosis after PCI cannot be ignored. It is surely an imperative topic that we need new strategies to reduce cardiovascular events.

TCM has a long history of treating CHD. By the method of holistic review, taking specimens into consideration, combination of differentiating disease and syndrome, TCM has a significant therapeutic effect on CHD. Zhang Zhongjing first named CHD as thoracic obstruction (meaning chest pain) and created Zhishi Xiebai Guizhi Decoction for phlegm and blood stasis. After more than 1800 years, this prescription is still a commonly used and fundamental prescription in the treatment of CVD.^[10] In recent years, many RCTs of ZXGD for CHD have been conducted, discussing the effect of this classic prescription from different angles. However, most of them were single center, small sample, short-term clinical studies. Due to individual studies’ inadequate power, there is poor evidence-based medicine proof of the efficacy and safety of ZXGD in the treatment of CHD. We organized this systematic review and meta-analysis to give available evidence on ZXGD for CHD in hope of making contributions to the clinical practice of TCM.

2. Methods

2.1. Protocol and registration

A priori protocol was developed and drafted conforming to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.^[11] A clear research question was formulated based on the principle of PICO as “In patients with CHD (P), how about the efficacy and safety of ZXGD (I), when compared with WM alone (C), in improving symptoms and ECG, blood lipid (TC, TG, LDL-C and HDL-C), vascular endothelial function (ET-1 and NO), inflammatory factor (hs-CRP) and Hcy (O)?” This systematic review was protocol driven and registered on PROSPERO in the first place (registration number: CRD42023435039).

2.2. Inclusion criteria

2.2.1. Types of studies design.

This systematic review was confined to RCTs, regardless of the method of blinding, without limitation on the country where the studies were implemented and published.

2.2.2. Types of participants.

Participants clinically diagnosed with CHD were enrolled, regardless of acute or chronic coronary syndromes, according to any internationally recognized guidelines. There was no restriction on the course of CHD, and patients who undergone PCI stents implant should not be excluded. The diagnosis of traditional Chinese medicine was mainly based on Guiding Principles for Clinical Study of New Chinese Medicines (2002).^[12] The clinical manifestations of the participants were consistent with thoracic obstruction, and the syndrome types were phlegm and (or) blood stasis.

2.2.3. Types of interventions.

Intervention in the treatment group was ZXGD or combined with WM. The components of ZXGD included Zhishi (Aurantii Fructus Immaturus), Xiebai (Allii Macrostemonis Bulbus), Guizhi (Cinnamomi Ramulus), Houpo (Magnoliae Officinalis Cortex) and Gualou (Trichosanthis Fructus). It was legitimate to add or subtract herb-medicine from the original prescription based on syndrome differentiation. There was no restriction on the dosage of each herb. The oral dosage of the decoction was 100mL to 200mL each time, 2 to 3 times a day. Intervention in the control group was WM treatment alone or combined with placebo, including anticoagulation, antiplatelet aggregation, coronary artery expansion and lipid-lowering, which was the same as the treatment group in 1 trial.

2.2.4. Types of outcomes.

The primary outcome measure was clinical improvement rate, which was described respectively from 2 aspects: symptom improvement and ECG improvement. The symptom improvement was defined as the reduction in seizure frequency or duration, or the resolution of CHD and its associated symptoms. Markedly effective: symptoms almost disappeared or significantly improved, or attack frequency and duration decreased by more than 80%; Effective: symptoms were relieved, or attack frequency and duration reduced by 50% to 80%; And ineffective: no changes or aggravation of symptoms, or attack frequency and duration reduced by <50%. ECG improvement was graded according to changes in ST-segment and the height of T-wave. Markedly effective: ECG returned to normal range; Effective: ST-segment decreased and recovered more than 0.5mV after treatment, but did not reach the normal level. The inverted T-wave on the main lead became shallow or T-wave changed from flat to upright. And ineffective: no change of ECG or even deterioration, which meant depressed ST-segment further dropped by more than 0.05mV, the inverted T-wave became deeper by more than 1/4, or the vertical T-wave became flat, the flat T-wave became inverted, and the new onset of atrioventricular block or ventricular block occurred. Marked and effective were considered to be improvement.

The secondary outcome measures were: blood lipid index including TC, TG, LDL-C and HDL-C; vascular endothelial function including ET-1 and NO; inflammatory factor hs-CRP; Hcy. Any adverse events data was also collected.

2.3. Exclusion criteria

This systematic review excluded literatures with repeated publications, incorrect or incomplete data, mistaken statistical methods, and unexpected outcome measures. Participants should be excluded if they had heart failure, depression, or anxiety as complication. In circumstances where there were other TCM prescriptions or Chinese patent drug serving as comparator, the studies should be excluded. Any other concomitant interventions such as other TCM prescriptions or Chinese patent drug, external therapies as acupuncture, and exercise therapies as Taijiquan was not allowed.

2.4. Search strategy

Two researchers (Li Ming and Song SQ) conducted the literature search as extensive as possible independently. In case of disputes, the third party (Wu Di) shall participate in the adjudication. The process involved 8 bibliographic databases including PubMed, Embase, The Cochrane Library, Web of Science, China National Knowledge Infrastructure, Wanfang, Chinese Scientific Journal Database (VIP) and Chinese BioMedical Literature Database. The retrieval date was from establishment of the databases to June 2023. An advanced search strategy was designed in order to identify relevant research reports efficiently and both subject headings and free-text words were utilized. The search concepts included: “coronary heart disease,” “acute coronary syndrome,” “chronic coronary syndrome,” “angina pectoris,” “myocardial infarction” and “Zhishi Xiebai Guizhi decoction.” The strategy was transformed according to the characteristics of each database. Besides, references of identified studies and previous systematic review on the topic of ZXGD and CHD were searched manually for extensive literature. Table 1 specifically documents the search strategy.

Table 1.

Search strategy.

Databases	Search strategy
PubMed	((coronary heart disease[MeSH Major Topic]) OR ((((((((((coronary heart disease[Title/Abstract])) OR (CHD[Title/Abstract])) OR (acute coronary syndrome[Title/Abstract])) OR (chronic coronary syndrome[Title/Abstract])) OR (angina pectoris[Title/Abstract])) OR (myocardial infarction[Title/Abstract]))) AND ((Zhishi Xiebai Guizhi decoction[MeSH Major Topic]) OR ((Zhishi Xiebai Guizhi decoction[Title/Abstract]) OR (Zhishi Xiebai Guizhi*[Title/Abstract])))
Embase	#1 = “ischemic heart disease”/exp OR “ischemic heart disease” #2 = “coronary heart disease” OR “CHD” OR “acute coronary syndrome” OR “chronic coronary syndrome” OR “angina pectoris” OR “myocardial infarction” #3 = #1 OR #2 #4 = “zhishi xiebai guizhi decoction”/exp OR “zhishi xiebai guizhi decoction” #5 = #3 AND #4
The Cochrane Library	#1 = MeSH descriptor: [Coronary Disease] explode all trees #2 = (coronary disease):ti,ab,kw OR (coronary heart disease):ti,ab,kw OR (acute coronary syndrome):ti,ab,kw OR (chronic coronary syndrome):ti,ab,kw OR (angina pectoris):ti,ab,kw OR (myocardial infarction):ti,ab,kw #3 = #1 OR #2 #4 = (Zhishi Xiebai Guizhi decoction):ti,ab,kw OR (Zhishi Xiebai Guizhi*):ti,ab,kw #5 = #3 AND #4
Web of Science	(TS = (coronary heart disease OR CHD OR acute coronary syndrome OR chronic coronary syndrome OR angina pectoris OR myocardial infarction)) AND TS = (Zhishi Xiebai Guizhi decoction OR Zhishi Xiebai Guizhi*)
Wanfang	主题:(枳实薤白桂枝汤 or枳实薤白桂枝汤加味 or枳实薤白桂枝汤加减) and 主题:(冠心病 or 冠状动脉心脏病 or 缺血性心肌病 or 心肌缺血 or 急性冠脉综合征 or 慢性冠脉综合征 or 心绞痛)
CNKI	(主题: 枳实薤白桂枝汤+ 枳实薤白桂枝汤加味 + 枳实薤白桂枝汤加减) AND(主题: 冠心病 + 冠状动脉心脏病 + 缺血性心肌病 + 心肌缺血 +急性冠脉综合征 + 慢性冠脉综合征 + 心绞痛(精确))
VIP	题名或关键词 = 枳实薤白桂枝汤 or枳实薤白桂枝汤加味 or枳实薤白桂枝汤加减 AND 题名或关键词 = 冠心病 or 冠状动脉心脏病 or 缺血性心肌病 or 心肌缺血 or急性冠脉综合征 or 慢性冠脉综合征 or 心绞痛
CBM	(“枳实薤白桂枝汤”[常用字段:智能] OR “枳实薤白桂枝汤加味”[常用字段:智能] OR “枳实薤白桂枝汤加减”[常用字段:智能]) AND(“冠心病”[常用字段:智能] OR “冠状动脉心脏病”[常用字段:智能] OR “缺血性心肌病”[常用字段:智能] OR “心肌缺血”[常用字段:智能] OR “急性冠脉综合征”[常用字段:智能] OR”慢性冠脉综合征”[常用字段:智能] OR “心绞痛”[常用字段:智能])

CBM = Chinese BioMedical Literature Database, CHD = coronary heart disease, CNKI = China National Knowledge Infrastructure, VIP = Chinese Scientific Journal Database.

2.5. Study screening and data extraction

The study screening was performed independently by 2 reviewers (Li Ming and Song SQ) according to the inclusion and exclusion criteria, and any conflicts were resolved by consulting the third reviewer (Rong YH). Preliminary screening was carried out by browsing titles and abstracts. In case definite judgment cannot be made, it was marked as “unclear,” followed by full-text screening to make the final decision.

The key information and data of included studies was extracted independently by 2 reviewers (Li Ming and Song S.Q.) and a third reviewer (Rong Y.H.) was involved when there were disagreements. A structured data extraction sheet was designed to record the following information: participants and disease characteristics (such as age, gender and number of participants, course of CHD, and diagnostic criteria), information on interventions (such as medicine and treatment duration), baseline characteristics, and outcome measures (primary outcome, secondary outcomes and adverse events).

2.6. Risk of bias assessment

According to the Cochrane Collaboration ROB Tool,^[13] 7 domains including random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other bias, should be assessed for quality of RCTs. All included studies were evaluated as low-risk, high-risk or unclear-risk. In order to improve the objectivity of the evaluation results, the work was performed independently by 2 researchers (Li Ming and Wu Di) and a third-party (Song SQ) acted as a consultant when there were disagreements. Review Manager 5.3 software was used to plot a visual graph to describe the ROB for each domain in each study.

2.7. Data analysis

Review Manager 5.3 software was used for data analysis. Despite that strict inclusion and exclusion criteria was taken into account for increasing precision, variations underlying different studies existed inevitably. Therefore, a statistical test for heterogeneity was performed to decide whether quantitative evidence synthesis was feasible. The I² statistic is an estimate of the ratio of true heterogeneity to total observed variation.^[14] It is generally believed that heterogeneity is not obvious when an I² is <50%, then the data synthesis quantitatively is meaningful, and the fixed-effect model will be applied. Conversely, an I² >50% indicates the presence of statistical heterogeneity. In this scenario, the study will be stratified according to the characteristics of the participant populations (such as CHD subtypes, whether there were stents, etc), that is, subgroup analysis. In case of the stubborn heterogeneity, the random-effect model will be used. The sensitivity analysis was performed by “leave one out” approach to evaluate the stability of the effect quantity.

In this meta-analysis, RR and MD was measured for dichotomous and continuous data, respectively. The CI was set to 95% and the cutoff for statistical significance was set at P < .05. The results were graphically presented with forest plots.

Publication bias was estimated by visual examination of funnel plot in circumstance where there were at least 10 studies included.

3. Results

3.1. Literature search results

The literature search retrieved 227 records from 8 electronic databases initially, of which 108 were removed as duplicates. After title and abstract screening, 51 relevant studies were identified to await further classification. Ultimately, a total of 24 records^[15–38] met the eligibility criteria after full-text review. The main reasons for excluding studies were contraventions of pre-specified criteria. The process is graphically illustrated in OA-Figure 1.

Figure 1.

Flow diagram of the literature screening.

3.2. Characteristics of included studies

The articles were published between 2007 and 2022 and all from Chinese publications. There were 2272 patients with 1136 in the ZXGD treatment group and 1136 in the control group. Patients were randomly treated with either conventional WM alone or ZXGD combination therapy for durations between 2 weeks to 24 weeks. The main characteristics of the included studies are summarized in Table 2.

Table 2.

Characteristics of the studies.

Author, yr	Number of participants		Age		Gender (Male/Female)		Course of disease (yr)		Comparisons		Duration of treatment (wk)	Outcome measures
	Experiment	Control	Experiment	Control	Experiment	Control	Experiment	Control	Experiment	Control
Maoyu Cai, 2016	50	50	55.5 ± 4.4	55.8 ± 4.6	27/23	29/21	4.5 ± 1.4	4.6 ± 1.5	ZXGD, WM	WM	16	①⑩
Chao Chen et al, 2019	51	51	68.54 ± 7.12	68.49 ± 7.09	28/23	27/24	5.48 ± 1.32	5.45 ± 1.29	ZXGD, WM	WM	24	①
Maiwen Du, 2022	40	40	57.34 ± 6.04	57.68 ± 6.13	22/18	23/17	5.46 ± 1.07	5.54 ± 1.05	ZXGD, WM	WM	8	①⑧⑨
Fei Jia, 2021	40	40	50.150 ± 14.624	50.400 ± 15.380	22/18	21/19			ZXGD, WM	WM	8	①③④⑩
Fengtao Li et al, 2021	34	34	58.14 ± 6.33	58.22 ± 6.47	19/15	18/16	5.65 ± 1.08	5.47 ± 1.02	ZXGD, WM	WM	8	①⑧⑨
Min Li et al, 2021	35	35	64.61 ± 8.48	65.13 ± 8.7	20/15	21/14			ZXGD, WM	WM	4	①③④⑤
Yapin Li, 2007	96	96	41–74	39–71	71/25	68/28			ZXGD, WM	WM		①②
Yuchun Li et al, 2021	60	60	64.19 ± 6.25	63.52 ± 6.33	31/29	28/32	4.19 ± 1.05	4.31 ± 1.02	ZXGD, WM	WM	12	②⑥⑦⑩
Yu Liu et al, 2018	73	72	54.9 ± 9.7	53.6 ± 8.6	40/33	39/33	4.9 ± 0.8	4.6 ± 1.5	ZXGD, WM	WM	4	①③④⑤⑥⑦
Zhimin Nie et al, 2022	40	40	59.83 ± 4.65	58.57 ± 4.77	25/15	22/18			ZXGD, WM	WM	8	⑥⑩
Lichun Shi et al, 2019	36	36	57.11 ± 3.38	56.86 ± 3.16	18/18	16/20			ZXGD, WM	WM	2	①②
Yueping Shi et al, 2009	26	30	55.09 ± 10.33	54.66 ± 10.27	20/6	22/8	0.6~28	0.3~26	ZXGD, WM	WM	4	①③④
Xiaodong Tan et al, 2019	51	51	45.13 ± 4.82	46.66 ± 5.17	32/19	34/17	1.63 ± 0.21	1.57 ± 0.17	ZXGD, WM	WM	4	①③④⑦
Xiuhai Wang, 2009	50	50	60 ± 11	59 ± 9	38/12	35/15	4.2 ± 1.6	4.3 ± 1.7	ZXGD, WM	WM	4	①②
Huiyuan Wei et al, 2011	30	30	65.52 ± 9.52	64.76 ± 8.23	15/15	14/16			ZXGD, WM	WM	4	①⑩
Shukui Wu, 2015	57	57	62.84 ± 6.76	63.91 ± 5.63	43/14	41/16	5.17 ± 2.43	6.69 ± 0.95	ZXGD, WM	WM	4	①③④⑤
Lijuan Yu et al, 2020	38	38	55.68 ± 5.41	55.23 ± 6.92	22/16	20/18	5.57 ± 1.28	5.86 ± 1.60	ZXGD, WM	WM	8	①⑧⑨
Youshou Yu, 2022	50	50	63.38 ± 5.32	63.25 ± 5.21	28/22	27/23	2.20 ± 0.81	2.17 ± 0.78	ZXGD, WM	WM	4	①③④⑧
Quanying Yuan et al, 2021	81	80	49.63 ± 5.41	51.09 ± 5.81	48/33	46/34			ZXGD, WM	WM	2	②⑥⑦⑩
Li Yue et al, 2020	38	38	45.12 ± 6.20	45.78 ± 6.31	23/15	21/17	4.98 ± 1.14	5.17 ± 1.02	ZXGD, WM	WM	5	①
Jie Zhang, 2019	68	68	51~80	56~80	35/33	38/30			ZXGD, WM	WM		①⑨
Tongwen Zhang et al, 2019	30	30	45.28 ± 4.1	49.27 ± 3.6	17/13	19/11	5.2 ± 1.4	4.9 ± 1.8	ZXGD, WM	WM	4	①③④⑤
Dejian Zhu et al, 2020	32	30	62.88 ± 7.97	62.90 ± 7.94	19/13	16/14	8.74 ± 1.62	9.02 ± 1.76	ZXGD, WM	WM	2	①②
Dejian Zhu et al, 2018	30	30	66.62 ± 10.36	67.89 ± 9.96	18/12	19/11			ZXGD, WM	WM	2	①⑧⑨

WM = Western medicine, ZXGD = Zhishi Xiebai Guizhi Decoction.

① Symptom improvement rate; ② ECG improvement rate; ③ TC; ④TG; ⑤ HDL-C; ⑥ ET-1; ⑦ NO; ⑧ hs-CRP; ⑨ Hcy; ⑩ adverse events.

3.3. Assessment of risk of bias

Overall risk of bias of included studies was assessed as moderate. The random sequences of 14 studies^{[16,18–20,22,24,25,27,31–34,37,38]} were generated using the random number table and 1 study^[15] using envelope lottery, which were assessed as low bias. Although 1 study^[23] claimed to be randomized, participants were actually assigned according to the order of admission, which was suspected of high selection bias. The methods of allocation concealment and blinding were absent in all studies, so the performance and detection bias was assessed as unclear. There were 2 patients dropping out in the control group in 1 study,^[18] while the researcher failed to report treatment measures of the missing data, so this study was considered to be at high risk of attrition bias. All studies reported the pre-set outcomes with low risk of reporting bias. Other bias was not found among the studies. The risk of bias is summarized in OA-Figure 2.

Figure 2.

Risk of bias summary for the included literatures.

3.4. Effects of intervention

3.4.1. Total clinical effective rate.

21 studies^{[15–21,23,25–32,34–38]} provided information on symptom improvement rate, and 6 studies^{[21,22,25,28,33,37]} provided information on ECG improvement rate. ZXGD was found to be superior to conventional WM alone for symptom improvement [RR = 1.21, 95% CI (1.16,1.26), I² = 0%, P < .00001] (OA-Figure 3) and ECG improvement [RR = 1.27, 95% CI (1.16, 1.40), I² = 13%, P < .00001] (OA-Figure 4).

Figure 3.

Meta-analysis forest plot of the symptom improvement rate.

Figure 4.

Meta-analysis forest plot of the ECG improvement rate. ECG = electrocardiogram.

3.4.2. Blood lipid.

8 studies^{[18,20,23,26,27,30,32,36]} provided information on TC. Since there was a substantial heterogeneity among studies (I² = 96%, P < .00001), subgroup analysis was performed according to CHD subtypes to examined the source of heterogeneity. Participants were divided into stable angina group^{[18,20,23,26,30,36]}and unstable angina group,^[27,32] while significant heterogeneity in both subgroups still existed. This suggested CHD subtypes was not a heterogeneous source. Therefore, a random-effect model was more appropriate. The meta-analysis results showed that the treatment of ZXGD had a significantly lower TC compared with control group [MD = −1.15, 95% CI (−1.75, −0.55), P = .0002] (OA-Figure 5).

Figure 5.

Meta-analysis forest plot of TC. TC = total cholesterol.

8 studies^{[18,20,23,26,27,30,32,36]} provided information on TG. Likewise, there was also a great heterogeneity (I² = 83%, P < .00001), and subgroup analysis was carried out in accordance with CHD subtypes. However, the quantitative synthesis of data neither in the stable angina group^{[18,20,23,26,30,36]} or unstable angina group^[27,32] was meaningless, because there was still a great heterogeneity in the stable angina group. The subgroup analysis failed to find the source of heterogeneity. Therefore, a random-effect model was used and data pooling showed that the treatment of ZXGD had a significantly lower TG compared with control group [MD = −0.72, 95% CI (−0.99, −0.45), P < .00001] (OA-Figure 6).

Figure 6.

Meta-analysis forest plot of TG. TG = triglyceride.

7 studies^{[18,20,23,26,27,30,32]} provided information on LDL-C. Meta analysis results showed a significant I² value (I² = 76%, P = .0003), which indicates that there was heterogeneity between studies. No source of clinical heterogeneity was found after subgroup analysis according to CHD subtypes or duration of treatment. Data of the studies was analyzed with a random-effect model and a comforting result showed that relative to WM alone, ZXGD combined with WM had a lower LDL-C [MD = −0.93, 95% CI (−1.17, −0.69), P < .00001] (OA-Figure 7).

Figure 7.

Meta-analysis forest plot of LDL-C. LDL-C = low-density lipoprotein cholesterol.

4 studies^{[20,23,30,36]} provided information on HDL-C. A fixed-effect model was used given that the negligible heterogeneity (I² = 9%, P = .35). The meta-analysis revealed that ZXGD combination group was superior to conventional WM alone group for improving HDL-C [MD = 0.31, 95% CI (0.20, 0.42), P < .00001] (OA-Figure 8).

Figure 8.

Meta-analysis forest plot of HDL-C. HDL-C = high-density lipoprotein cholesterol.

3.4.3. Vascular endothelial function.

4 studies^[22–24,33] provided information on ET-1. Since there was a substantial heterogeneity among studies (I² = 60%, P = .06), subgroup analysis was performed in accordance with duration of treatment. Participants received 8 to 12 weeks of treatment in 2 studies^[22,24] and 2 to 4 weeks in the other 2 studies.^[23,33] However, between-study heterogeneity in 2 subgroups was large, respectively. This indicated that the treatment duration was not a heterogeneous source. Then the sensitivity analysis was performed, and the heterogeneity decreased (I² = 0, P = .73) when 1 studies^[33] were deleted. A fixed-effect model was used, and meta-analysis results of the remaining 3 studies^[22–24] showed that the treatment of ZXGD had a significantly lower ET-1 compared with control group [MD = −7.81, 95% CI (−9.51, −6.10), P < .00001] (OA-Figure 9).

Figure 9.

Meta-analysis forest plot of ET-1. ET-1 = endothelin-1.

4 studies^{[22,23,27,33]} provided information on NO. A fixed-effect model was used due to the low heterogeneity (I² = 21%, P = .28). The meta-analysis revealed that ZXGD combination group was superior to conventional WM alone group for improving NO [MD = 8.90, 95% CI (7.86, 9.93), P < .00001] (OA-Figure 10).

Figure 10.

Meta-analysis forest plot of NO. NO = nitric oxide.

3.4.4. Inflammatory factor.

5 studies^{[17,19,31,32,38]} provided information on hs-CRP. The heterogeneity test indicated that there was a high statistical heterogeneity (I² = 90%). Then subgroup analysis was performed based on whether there were heart sents to examined the heterogeneity source. The results showed that ZXGD decreased hs-CRP in patients without heart sents^[31,32,38] [MD = −0.88, 95% CI (−1.04, −0.71), P < .00001, I² = 0] and patients with 1 or 2 heart sents^[17,19] [MD = −3.07, 95% CI (−3.71, −2.43), P < .00001, I² = 0]. The total combined effect under the random-effect model suggested the same conclusion [MD = −1.73, 95% CI (−2.63, −0.83), P < .00001] (OA-Figure 11). Heterogeneity significantly reduced in both groups, proving that whether there were heart stents was possibly the sources of heterogeneity.

Figure 11.

Meta-analysis forest plot of hs-CRP. hs-CRP = high-sensitivity C-reactive protein.

3.4.5. Hcy.

5 studies^{[17,19,31,35,38]} provided information on Hcy. A fixed-effect model was used in the presence of small heterogeneity (I² = 0, P = .55). Pooled data from the 5 RCTs suggested that ZXGD combination therapy significantly decreased Hcy in plasma [MD = −2.03, 95%CI (−2.78, −1.28), P < .00001] (OA-Figure 12).

Figure 12.

Meta-analysis forest plot of Hcy. Hcy = homocysteine.

3.4.6. Adverse events rate.

Any adverse events data was also recorded. There were 6 studies^{[15,18,22,24,29,33]} reporting occurrence of adverse events, and 3 of them^[15,22,24] declared no adverse events happened throughout the RCTs. In the other 3 studies,^[18,29,33] there were 19 cases of gastrointestinal reactions such as abdominal pain and diarrhea, 6 cases of liver and kidney function damage, and 15 cases of dizziness, palpitation, rash or fever. The meta-analysis revealed no significant difference between 2 groups in means of adverse events [RR = 0.77, 95% CI (0.44, 1.34), I² = 0, P = .36] (OA-Figure 13).

Figure 13.

Meta-analysis forest plot of the adverse events rate.

3.5. Sensitivity analysis

Sensitivity analysis was performed and the results showed that little changes were found in effect quantity each time 1 study was excluded. Consequently, the results of this meta-analysis had relatively good stability.

3.6. Publication bias

Symptom improvement rate was used to create a funnel plot to identify the publication bias. 21 individual studies showed visually centralized and symmetrical distribution, so it was reasonable to believe that publication bias of the included RCTs had little effect on the results of meta-analysis (OA-Figure 14).

Figure 14.

Funnel plots of the symptom improvement rate.

4. Discussion

Recently, a large number of clinical and basic studies are investigating the regulatory effect of ZXGD in the treatment of CHD. However, the outcomes have not been evaluated systematically. This systematic review and meta-analysis aim to appraise the efficacy and safety of ZXGD for CHD. A total of 24 RCTs involving 2272 patients were included and the results showed that on the basis of traditional WM, ZXGD combination therapy had advantages on improving clinical efficacy (both symptom and ECG) for the treatment of CHD. In terms of blood lipid, ZXGD combined with WM is superior to WM alone to decrease TC and TG, meanwhile, increase HDL-C. It has been universally known that dyslipidemia is one of the risk factors for CHD. TC and TG are positively correlated with the incidence of CHD while HDL-C is negatively correlated. ET-1 and NO are vasoactive substances that synthesize endothelial cells, whose synthesis, release and imbalance directly affect the basic tension of blood vessels, and are closely related to the occurrence and development of CHD.^[39] Our meta-analysis suggested that ZXGD significantly decreased ET-1 and increased NO. Studies have shown that inflammatory response actives an important part in changing coronary plaques.^[40] hs-CRP is an important inflammatory factor and elevated hs-CRP can predict the increasing risk of CHD. Our meta-analysis revealed that ZXGD could decrease hs-CRP. As a new independent risk factor for CHD, excess of Hcy can damage vascular endothelial cells, promote the proliferation of smooth muscle cells, and activate inflammatory reaction and oxidative stress reaction, thus lead to atherosclerotic diseases.^[41] Our meta-analysis presented that ZXGD could decrease Hcy. As for the adverse reactions, there was no statistical difference between ZXGD group and the control group, indicating that ZXGD was safe and reliable for CHD.

Nevertheless, the high heterogeneity of the results should not be ignored. Subgroup analysis was employed in TC, TG and ET-1 according to CHD subtypes and duration of treatment, but the heterogeneity was still considerable. Blood lipid and vascular endothelial function can be affected by age, BMI, lifestyle (such as smoking, drinking, etc) and exercise, which may lead to clinical heterogeneity in the present study. Subgroup analysis of hs-CRP found the source of heterogeneity, that is, whether there were heart sents.

TCM has a long history in treating CVD and exhibits promising preventive and curative effects on CHD. According to the theory of TCM, the pathogenesis of CHD is summarized as the decline of Yang in the upper energizer and the invasion of pathogen. As a well-known traditional herbal formula, ZXGD has the function of activating Yang and regulating Qi, reducing phlegm and eliminating nodes. Thousands of years of clinical practice has proved its definite curative effect on CHD. Besides, experimental studies have also revealed the mechanism of ZXGD. It has been shown that adenosine, guanosine and quercetin are the effective components of ZXGD in the treatment of CHD.^[42] The cell metabolomics has demonstrated that ZXGD can regulate oxidative stress, sustain mitochondrial homeostasis, ameliorate the disordered energy metabolism and improve amino acid metabolism, thus to protect myocardium.^[43] Animal experiments have shown that ZXGD may alleviate myocardial injury in rats through inhibiting TNF/NF-κB signaling pathway.^[44]

There are some restrictions to the study. The primary shortness is the small sample size with 13 trials having fewer than 50 participants in ZXGD group. In addition, short treatment and observation period, and a lack of follow-up hinders the way to explore the long-term efficacy and safety of ZXGD for CHD. A further limitation is the general low quality of included studies, which can reduce the power of the meta-analysis. Therefore, the analysis results of this outcome should be taken a cautious attitude.

5. Conclusion

The present study shows that ZXGD combined with traditional WM seems to be safe and more effective than WM alone for CHD patients. However, due to the limitation of the quality of evidence, more large sample, rigorously designed and conducted RCTs are still needed to supplement the evidence of this conclusion.

Author contributions

Conceptualization: Yongtian Yin.

Data curation: Shengqiang Song.

Funding acquisition: Ming Li.

Investigation: Ming Li, Shengqiang Song.

Methodology: Yuanhang Rong, Di Wu.

Project administration: Yongtian Yin.

Resources: Shengqiang Song.

Software: Shengqiang Song.

Supervision: Yuanhang Rong, Di Wu, Yongtian Yin.

Visualization: Ming Li.

Writing – original draft: Ming Li.

Writing – review & editing: Yongtian Yin.