Potential Effectiveness of Chinese Patent Medicine Tongxinluo Capsule for Secondary Prevention After Acute Myocardial Infarction: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Background: Chinese patent medicine Tongxinluo capsule (TXL) is commonly used for cardio-cerebrovascular diseases. Previous research had demonstrated that TXL exhibited great clinical effects on the treatment of acute myocardial infarction (AMI), however there is a lack of systematic review. The purpose of this study was to evaluate the potential effectiveness and safety of TXL for secondary prevention in patients with AMI. Method: We searched 6 databases to identify relevant randomized controlled trials (RCTs) from inceptions to December 30, 2017. Two review authors independently assessed the methodological quality and analyzed data by the RevMan 5.3 software. The publication bias was assessed through funnel plot and Begg's test. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was used for evaluating the quality of evidence. Results: We included 19 RCTs in this review and performed a meta-analysis based on 16 studies. There were statistical differences of TXL treatment group in reducing primary cardiovascular events (cardiac death [RR = 0.27, 95%CI: 0.08~0.95, I2 = 0%], recurrent myocardial reinfarction [RR = 0.38, 95%CI: 0.20~0.74, I2 = 0%], arrhythmia [RR = 0.44, 95%CI: 0.30~0.66, I2 = 0%], recurrent angina pectoris [RR = 0.34, 95%CI: 0.17~0.69, I2 = 0%]). TXL could improve cardiac function (LVEF [MD = 4.10, 95%CI: 3.95~4.25, I2 = 0%]), regulate blood lipid TC [MD = −0.66, 95%CI: −0.94 ~ −0.37, I2 = 74%], TG [MD = −0.38, 95%CI: −0.62 ~ −0.14, I2 = 70%], LDL-C[−0.40, 95%CI: −0.65 ~ −0.16, I2 = 88%), decrease the level of hs-CRP (4-week: MD = −0.78, 95%CI: −0.97 ~ −0.60, I2 = 20%; Over 4-week: MD = −1.36, 95%CI: −1.55 ~ −1.17, I2 = 20%). However, TXL has little effects on revascularization [RR = 0.45, 95%CI: 0.13~1.56, I2 = 0%], recurrent heart failure (RR = 0.83, 95%CI: 0.27~2.57, I2 = 0%), and HDL-C (MD = 0.14, 95%CI: 0.00 ~0.29, I2 = 73%). Furthermore, TXL treatment group was more prone to suffer gastrointestinal discomfort. Conclusion: Chinese patent medicine TXL seemed beneficial for secondary prevention after AMI. This potential benefit needs to be further assessed through more rigorous RCTs. Systematic review registration number in the PROSPERO register: CRD42017068417.

. Acute myocardial infarction (AMI) is the most severe type of CVD, causing more than 4 million deaths in Europe and northern Asia Nichols et al., 2014). In United States, there would be an American diagnosed as myocardial infarction (MI) approximately every 40 s and the overall prevalence for MI was 3.0% in adults (Benjamin et al., 2017). In China, the number of CVD events manifestly increased from 0.75 million in 1990 to 1.4 million in 2013, and one million deaths were caused by MI annually Zhou et al., 2016). According to the National Institute of Health and Care Excellence clinical guideline, the treatment of AMI mainly contains primary percutaneous coronary intervention (PPCI), fibrinolysis, coronary angiography and so forth (Carville et al., 2015). And with the development of these revascularization treatments, the short-term prognosis of AMI has been greatly improved, but mortality and recurrence rate of patients with MI remain high after the acute phase of MI (García-García et al., 2010;Andrés et al., 2012). Secondary prevention of MI has been defined as a kind of evidence-based guideline recommendation and comprehensive intervention strategy to control various cardiovascular risk factors, reduce the recurrence of MI, heart failure (HF), arrhythmia, angina and the incidence of sudden death, extend patients' life and improve their life quality via using non-medical and medical therapy (Perk et al., 2013;Piepoli et al., 2018). Thus, it is vital to conduct secondary prevention for patients who suffered the MI as early as possible.
Nowadays, to decrease risk factors and mortality, available guidelines for secondary prevention of MI worldwide uniformly recommend interventions: one is the optimal use of cardioprotective medical therapies [aspirin, statins, beta-blockers and angiotensin converting enzyme inhibitors/ angiotensin receptor blockers (ACEI/ARBs)], the other one is non-medicine strategy, like smoking cessation and the increase of physical activities (Zeymer et al., 2011;Petersen et al., 2012;Bauters et al., 2014). Traditional Chinese medicine (TCM) has been used in patients with CVDs for thousands of years and is still being commonly used in modern times in both China and elsewhere worldwide (Tachjian et al., 2010). However, to date, it is still uncertain whether TCM can be used as a kind of drug therapy for secondary prevention of CVDs.
Tongxinluo Capsule (TXL), passed appraisal by National Fund for new drug research and registered in the State Food and Drug Administration of China, is a formally classical Chinese patent medicine in treating unstable angina pectoris and acute coronary syndrome (Wang et al., 2003;Ma et al., 2009). The main ingredients of TXL consist of Ginseng Radix Et Rhizoma (Ren Shen), Paeoniae Radix Rubra (Chi Shao), Ziziphi Spinosae Semen (Suan Zao Ren), Dalberglae Odoriferae Lignum (Jiang Xiang), Santali Albi Lignum (Tan Xiang), Olibanum (Ru Xiang), Hirudo (Shui Zhi), Scorpio (Quan Xie), Scolopendra (Wu Gong), Cicadae Periostracum (Chan Tui), Eupolyphaga Steleophaga (Tu Bie Chong), and Borneolum Syntheticum (Bing Pian), which can supply qi, activate blood circulation, dredge collaterals and relieve pain according to the TCM theory (Pharmacopoeia, 2015). Besides, in clinical practice, TXL is taken orally 2-4 capsules per time, 3 times a day. The detailed production process and quality control of TXL was showed in Supplementary Material 1. Plenty of studies showed that TXL could enhance the stability of vulnerable plaques, inhibit inflammation, improve the left ventricular function and exhibit other pharmacological functions Zhang et al., 2010;Bai et al., 2013). A Cochrane systematic review indicated TXL could reduce the frequency of acute angina, improve electrocardiogram and angina symptoms (Wu et al., 2006). Currently, many studies reported TXL had good therapeutic effects on patients with MI, but there was little solid evidence about secondary prevention after AMI, thus, the purpose of this study was to perform a comprehensive review and meta-analysis of TXL for secondary prevention after AMI, especially for the primary cardiovascular events and heart functions.

METHODS
The review protocol was registered on PROSPERO (CRD42017068417), and the review was constructed following the PRISMA guidelines (Supplementary Material 2).

Types of Studies
Randomized controlled trials (RCTs) with blind method or not and no limit of publishing language were included. The course of treatment was no less than 4-week, and the number of patients in treatment/control group was more than 30. We excluded the conference papers.

Types of Participants
Patients diagnosed with AMI were included, not included those with severe liver and kidney dysfunction. Age and race were not limited.

Types of Interventions
Based on the conventional western medicine treatment, the treatment group was given TXL alone or combined with western medicine, and the control group was given nothing, placebo or western medicine. According to clinical symptoms, the conventional western medicines are mainly for secondary prevention of MI, including aspirin (100 mg/d), clopidogrel (75 mg/d), metoprolol tartrate tablet (50 ∼ 100 mg/bid), atorvastatin (20 mg/d) and so forth.

Types of Outcome Measures Primary outcomes
Primary cardiovascular events: cardiac death, recurrent myocardial infarction, revascularization, arrhythmia, heart failure and recurrent angina pectoris; Heart functions: left ventricular ejection fraction (LVEF). The included studies must contain the primary outcomes.

Data Extraction and Management
Two investigators (QZ and JYH) independently conducted the literature searching, study selection, and data extraction. The extracted data of included studies was filled in a standardized form prepared for this review. The extracted data included the first author, year of publication, disease type, sample size, randomized method, average age of participants, treatment duration, interventions in the treatment and control group, dose of medicines, outcomes and so on. Disagreements were discussed and resolved in a consensus meeting with the corresponding author (HCS). If the literature has multiple time end indicators, we chose the longest one.

Unit of Analysis Issues
Subgroups were divided according to intervention or course of treatment and analyzed individually.

Assessment of Quality of Included Studies
According to Cochrane Reviewer's Handbook, two of the authors (ML and RJQ) individually assessed the risk of bias. Six evaluation criteria of the quality of RCTs were used, which involved the generation of random sequence, randomization concealment, blind method, integrity of outcome data, selective reporting and other bias.

Measures of Treatment Effect
Revman 5.3 software provided by the Cochrane Collaboration was used for data analyses. Continuous outcomes were presented as weighted mean difference (WMD), dichotomous data as risk ratio (RR) and 95% confidence interval (CI). Means and standard deviations were extracted for continuous outcomes, and as for the dichotomous data, we recorded the number of patients who experienced the event in each group.

Assessment of Heterogeneity
Clinical heterogeneity of included studies was analyzed with χ 2 test. If I 2 < 50%, then there is no statistical heterogeneity between studies, and fixed effect model was used for data analysis; If I 2 > 50%, then statistical heterogeneity between studies exists, random effect model was used, and the cause of heterogeneity was analyzed. Subgroup analysis was used when clinical heterogeneity exists. Subgroup was divided according to the sources of heterogeneity, such as intervention measure, course, and dosage of medicines. Descriptive analysis was used if clinical heterogeneity still exists.

Assessment of Publication Biases
The potential publication bias was assessed using funnel plots and Begg's test by Stata 14 software (Begg and Mazumdar, 1994). If P > 0.05, there is no publication bias, whereas publication bias exists.

Sensitivity Analysis
Compare the pooled statistics before and after excluding studies of low quality and great weight and those which have different results from other studies. If they have the same results (both results have differences or have no differences), then the meta-analysis result is stable and vice versa.

Statistical Analysis
The primary and second outcomes from the included studies were entered into Revman 5.3 (Higgins and Green, 2011). P < 0.05 was considered statistically significant.

Grade
The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was used to evaluate the quality of evidence by ML and XYZ, which was classified as high, moderate, low, or very low based on the judgment of the risk of bias, inconsistency, indirectness, imprecision and other considerations (Schunemann et al., 2013). Summary of Findings (SOF) tables were prepared using the software program "GRADE pro GDT."

Search Process
The initial search using the electronic search strategies yielded 1328 studies. After removing 708 duplicates from different databases, we evaluated 620 potentially relevant articles for eligibility. After screening the titles and abstracts, we excluded 507 studies. Of the 113 remaining studies, we further excluded 94 studies after screening the full-text articles. Eventually, we included 19 studies. There may exist the clinical heterogeneity between long-term intervention (≥1-year) and short-term duration (<1-year) in included studies. Thus, we conducted a descriptive analysis of 3 articles Shen, 2011; with long-term intervention and do a meta-analysis of the remaining 16 studies (You et al., 2005(You et al., , 2012Chen et al., 2007Chen et al., , 2008Huang, 2010;Liang et al., 2010;Liao et al., 2010;Tian and Li, 2011;Yang and Yu,  2011; Tian et al., 2014;Qin, 2015;Xia et al., 2016;Fan et al., 2017;Li X. F. et al., 2017;Peng et al., 2017;Wu, 2017). A flow chart (Figure 1) illustrates our search process and study selection.

Included Studies
We included 19 studies in this review. All studies were conducted in China and published in full.

Participants
Total of 1877 participants with AMI were included. The average sample size of the trials was 99 participants (ranging from 60 to 190 participants per trial). The course of treatment fluctuated between 4-week and 2-year.

Risk of Bias Assessment
All the trials provided very limited information about design and methodology. Six studies (Chen et al., 2008; Zhang and Zhang, 2009; Tian et al., 2014;Qin, 2015;Xia et al., 2016;Wu, 2017) described the random sequence generation methods, among which 5 articles (Chen et al., 2008;Tian et al., 2014;Qin, 2015;Xia et al., 2016;Wu, 2017) showed the random number table, the other one used lottery . The rest included studies only mentioned random, did not reported it in detail. In addition, allocation concealment and blind were not very clear. Only one study used a double-blind method (Chen et al., 2008). None of the trials had a pretrial estimation of sample size, which indicated the lack of statistical power to ensure appropriate estimation of the therapeutic effect, as shown in Figure 2.

Effects of Interventions
We conducted a descriptive analysis of 3 studies. A study  showed TXL could increase LVEF (P < 0.05), and the other 2 articles mainly reported the role of TXL on primary cardiovascular events. One study (Shen, 2011) exhibited TXL was beneficial to reduce the total cardiovascular events, the other ) demonstrated TXL had no effects on cardiac death, non-fatal MI, HF, recurrent angina pectoris except revascularization. Then we did a meta-analysis of the remaining 16 studies with a short-term duration and assessed the clinical functions of TXL by primary/s outcomes.

Heart function
AMI is the most severe manifestation of coronary artery disease, which has a negative effect on the myocardial function and blood supply. LVEF is closely related to the myocardial contractility, reflecting myocardial systolic function and bloodstream of the coronary arteries. Eleven studies (You   (You et al., 2012) detecting LVEF by radionuclide ventriculography, others used the doppler echocardiography. Thus, the meta-analysis we conducted from the remaining ten studies. The meta-analysis showed that TXL treatment group could increase the LVEF significantly and improve heart function (MD = 4.10, 95%CI: 3.95∼4.25, I 2 = 0%) (Figure 4).

The inflammation reaction
Inflammation reaction is one of the key links of AMI progression. The necrosis of numerous cardiomyocytes in the infarct tissue will lead to a sharp inflammation. The rapid accumulating of massive leukocytes and monocytes accompanied by the release of a great number of inflammatory factors, leading to further aggravate myocardial injury (Stirrat et al., 2017). Hs-CRP, one kind of C-reactive protein in plasma, is a non-specific marker of systemic inflammatory response in acute phase, could predict the risk of cardiovascular events. Six studies (Huang, 2010;Liao et al., 2010;Tian et al., 2014;Xia et al., 2016;Fan et al., 2017;Wu, 2017) revealed hs-CRP involving in 293/287 patients in both groups. Due to high heterogeneity, we carried out a subgroup analysis based on different treatment courses. When the intervention time was 4 weeks, TXL treatment group could effectively reduce the hs-CRP level (MD = −0.78, 95%CI: −0.97 ∼ −0.06, I 2 = 20%), and this therapeutic

Adverse Reaction
Seven studies (Chen et al., 2007;Huang, 2010;Shen, 2011;Tian and Li, 2011;Yang and Yu, 2011; Wu, 2017), involving in 362/332 patients in two groups, have reported the adverse events of the TXL. Adverse drug reactions mainly included the gastrointestinal discomfort, transaminase elevation, stroke, pulmonary infection and so forth. Compared with the control group, the total risk rate of adverse events in the treatment group was relatively high (6.91% versus 4.22%), and the gastrointestinal discomfort accounted for the main proportion, as shown in Table 2.

Publication Biases
We proceed the publication bias (Recurrent myocardial reinfarction and LVEF) to evaluate the quality of our metaanalysis by funnel plot and Begg' test using State 14 software. Funnel chart analysis showed there were no obvious publication biases and P-values were 0.536, 0.371, respectively, as shown in Figures 8A,B.

Grade
The quality of evidence was evaluated by ML and XYZ with GRADE approach, and most of cardiac events were considered as the moderate score. Some outcomes exhibited the low or very low evidence due to the large heterogeneity and/or the wide confidence interval (Table 3).

DISCUSSION
In this meta-analysis from 16 studies of TXL, we observed the results as following: There was significant difference of TXL in preventing restenosis and recurrence of cardiovascular events (cardiac death, myocardial reinfarction, arrhythmia, recurrent angina pectoris). TXL enabled to improve cardiac function (LVEF), regulate lipid metabolism (TC, TG, LDL), and inhibit inflammation reaction (hs-CRP). However, TXL has little effects on revascularization, recurrent heart failure, and HDL. Moreover, TXL treatment group was more prone to suffer gastrointestinal discomfort.

Effectiveness of TXL
AMI is a common cardiovascular disease, remaining a leading cause of morbidity and mortality worldwide. As we all know, AMI is caused by atherosclerotic coronary artery stenosis, plaque rupture and thrombosis appear, engendering insufficient blood supply of coronary, further resulting in myocardial ischemia or even necrosis. If AMI does not promptly recanalize the infarctrelated coronary arteries after the onset of the disease, many patients eventually die of ventricular arrhythmias and heart failure. Despite substantial improvements in prognosis over the past decade, the progress is a result of several major trends, including improvements in risk stratification, more widespread use of an invasive strategy and medication (Kvakkestad et al., 2016). However, mortality and recurrence rate of the adverse cardiac events remain high after the acute phase of MI which will seriously affect the patients' quality of life and prognosis. Therefore, the second prevention of adverse cardiac events after MI is very important. Coronary artery in patients with MI has extensive lesions and multiple "vulnerable" plaques. While interventional therapy is often focusing on some single serious plaques, it is vital for many drugs to widely protect cardiovascular system through inhibiting vascular inflammation, improving myocardial energy metabolism, reducing myocardial remodeling, regulating blood lipids and so forth. Compared with the control group, TXL alone or combined with conventional western medicine exhibited a specific advantage in the secondary prevention of AMI, likely due to the comprehensive protection of blood vessels and myocardium. Many clinical trials and experimental evidence have just proved it.

Endothelial Protective Effect
Dysfunction of endothelial cells is directly associated with impaired vasorelaxation, increased inflammation, and increased migration and proliferation of smooth muscle cells, is the initiating factor in CVDs (Gutiérrez et al., 2013). And oxidized low-density lipoprotein-(ox-LDL-) induced dysfunction of the vascular endothelium is one of the key factors in the pathogenesis of many CVDs (Wang et al., 2009). It has been demonstrated that TXL alleviated ox-LDL-induced hyperpermeability via the activation of ERK1/2 in the vascular endothelial cell (Chang et al., 2017). Vascular injury after chronic hypoxia usually leads to endothelial injury and structural damage to tight junctions, TXL could promote endothelial cell proliferation and reverse the expression of tight junctions' proteins in vivo and in vitro (Zheng et al., 2015). Additionally, we found that TXL could regulate gene expression in hypoxia-treated human cardiac microvascular endothelial cells (CMECs) (Li Y. N. et al., 2015) and attenuate human CMEC injury via inhibiting peroxynitrite, which contributed to macrophage-mediated human CMEC injury .

Protection Against Ischemia-Reperfusion (I/R) Injury
When AMI occurs, reperfusion of the ischemic myocardium is a key approach for limiting infarct size. However, reperfusion itself paradoxically leads to further cardiovascular complications, which is called myocardial I/R injury. Nowadays, no effective therapy is available to protect the heart from this I/R injury (Eltzschig and Eckle, 2011). Previous studies have demonstrated that TXL enabled to effectively protect hearts against no-reflow and reperfusion injury in a protein kinase A-dependent manner (Li et al., 2010(Li et al., , 2013 and ameliorate myocardial I/R injury by promoting cardiac microvascular endothelial barrier function . Besides, as for cerebral I/R injury, TXL could reduce cell death via Cx43/Calpain II/Bax/Caspase-3 pathway (Cheng et al., 2017).

Anti-inflammation
The immune system and inflammation, through several cellular and soluble inflammatory mediators, play a great role in the local tissue structural changes of AMI (Swirski and Nahrendorf, 2013). TXL was able to decrease serum contents of P-selectin, intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1) and pro-inflammatory cytokines (interleukin 6, interleukin 10), regulate anti-inflammatory factor levels in the early reperfusion of AMI (Zhang H. T. et al., 2009). As the above result showed, TXL treatment group could effectively reduce the hs-CRP level and this therapeutic effect would get better along with the prolonging of treatment period. Moreover, TXL exerted its vascular protective effects by inhibiting the vascular inflammatory response and suppressing the gene expression of miR-155 (Zhang et al., 2014).

Regulation of Lipid Metabolism
The pathogenesis of AMI has close links with diabolism of lipids. The rise of blood fat will easily lead to the occurrence and development of vulnerable plaques, further resulting in AMI. The clinical trials indicated that TXL could significantly reduce the level of TC, TG, and LDL. Relevant experimental study reported that TXL enabled to enhance the stability of vulnerable plaques and prevent plaques from rupture in a dose-dependent manner.
In addition, high-dose TXL group showed a low serum level of  low-density lipoprotein cholesterol and ox-LDL. Besides, TXL was effective for decreasing serum lipid levels and inhibiting plaque inflammation in the balloon-induced abdominal aortic endothelial injury rabbits model (Chen et al., 2009).

Others
A large number of studies have reported that TXL presented various regulatory influences on CVDs. The cardioprotective benefits of TXL against MI were closely related to the inhibition of apoptosis and promotion of autophagy in rat hearts after AMI by activating AMPK signaling pathway . TXL also exhibited anti-platelet aggregation effects . TXL may probably exhibit a specific advantage in the prevention and treatment of myocardial fibrosis via mediating the expressions of TGF-beta1, Smad3, and Smad7 in diabetic rats (Wang X. et al., 2016). Furthermore, TXL could inhibit OX-LDL-induced maturation of dendritic cells by regulating PPAR gamma pathway . So, in summary, TXL exerted its comprehensive protective effects on myocardium and blood vessels not only in clinical trials but also in basic experiments (Figure 9). Nevertheless, there was no statistical difference between TXL treatment group and control group in some fields, such as revascularization, recurrent heart failure and HDL.
The reasons may include the following two aspects. Firstly, TXL, a compound preparation consisting of many kinds of herbs and insects, presented limited therapeutic effects on revascularization, recurrent heart failure and HDL from the current literature.
Secondly, and most importantly, the intervention time of included studies to conduct meta-analysis was less than one year, it may be difficult to observe the occurrence of revascularization and heart failure as well as the increase of HDL during the short period. Thus, long-term studies need to be further implemented.

The Safety of TXL
Seven studies reported adverse events in this review. Compared with the control group, the total risk rate of adverse events in the treatment group was relatively high, and the gastrointestinal discomfort accounted for the major proportion. We observed the reasons as following: TXL contains many animal medicines, such as Leech (Hirudo, Shui Zhi), Scorpion (Scorpio, Quan Xie), Centipede (Scolopendra, Wu Gong), and other insect drugs, which could dredge the channel and promote blood circulation, while on the other hand, this kind of function maybe partly cause stomach irritation. There were a few articles reporting adverse drug reactions which was not able to comprehensively reflect the safety of TXL. Many studies did not show adverse events of control group. Additionally, there was not statistical analysis of untoward effects in original literature, all of such reasons contributed to the large adverse events in the experimental group.

Study Limitations
With the limitation of small sample size, short-term duration and poor quality of methodology (unclear of random and blind methods), more large-scale perspective randomized double-blind long-term RCTs are needed to confirm the efficacy and safety of TXL in treating AMI. Furthermore, the included patients in our systematic review were all from China, so it is difficult to define the influence of race and region. Besides, there were not enough studies involving in dose-effect relationship of TXL.

Clinical Practice and Further Research of This Review
Given the available evidence by our work, the results of this review suggested that TXL alone or combined with conventional western medicine exhibited a specific advantage in the secondary prevention of AMI, especially in the decrease of primary cardiac events. Clinical practice could consider adding oral TXL to routine pharmacotherapy management, it may improve the living quality of the patients and clinical efficacy. Taking TXL for example, it also proved that traditional Chinese medicine, a kind of complementary and alternative medicine, has its unique advantages in the treatment of some diseases. Besides, some key points need to pay more attention. Firstly, lack of detailed description of randomization and blinding was the major methodological flaw of most included studies in this review, thus, clinical experts and statistical professors should have a better cooperation to design the high-quality clinical trials and make the blinding of participants and outcome assessors. Secondly, due to the limited evidence of safety of TXL, more research is worth further exploring. In addition, the researchers should conduct more trials in various countries to explore the influence of different regions and races. More importantly, the biological mechanism of TXL and its active compounds need to further explore, which will benefit to providing more evidence for clinical treatment in CVDs.

CONCLUSION
Chinese patent medicine TXL seemed beneficial for secondary prevention after AMI. This potential benefit needs to be further assessed through more rigorous RCTs. Furthermore, TXL may be safe for treating AMI due to the milder adverse events in clinic.

AUTHOR CONTRIBUTIONS
HS and GT provided guidelines for this systematic review and meta-analysis. ML, CL, SC, and YS wrote the main manuscript and prepared the Figures 3-7, 9 and Table 2. QZ and JH conducted the literature searching, study selection and data extraction. JH provided Table

FUNDING
The present work was supported by grants from the National Natural Science Foundation of China (No. 81430098).