Oral Panax notoginseng Preparation for Coronary Heart Disease: A Systematic Review of Randomized Controlled Trials

This systematic review aims to evaluate current evidence for the benefit and side effect of oral Panax notoginseng preparation for coronary heart disease (CHD). We included 17 randomized clinical trials (17 papers and 1747 participants). Comparing with no intervention on the basis of conventional therapy, oral Panax notoginseng did not show significant effect on reducing cardiovascular events, but it could alleviate angina pectoris (including improving the symptoms of angina pectoris [RR 1.20; 95% CI 1.12 to 1.28; 7 trials, n = 791], improving electrocardiogram [RR 1.35; 95% CI 1.19 to 1.53; 8 trials, n = 727], decreasing the recurrence of angina pectoris [RR 0.38; 95% CI 0.16 to 0.94; 1 trials, n = 60], duration of angina pectoris [RR −1.88; 95% CI −2.08 to −1.69; 2 trials, n = 292], and dosage of nitroglycerin [MD −1.13; 95% CI −1.70 to −0.56; 2 trials, n = 212]); oral Panax notoginseng had no significant difference compared with isosorbide dinitrate on immediate effect for angina pectoris [RR 0.96; 95% CI 0.81 to 1.15; 1 trial, n = 80]. In conclusion, oral Panax notoginseng preparation could relieve angina pectoris related symptoms. However, the small sample size and potential bias of most trials influence the convincingness of this conclusion. More rigorous trials with high quality are needed to give high level of evidence, especially for the potential benefit of cardiovascular events.


Introduction
Coronary heart disease (CHD) is one of the most leading causes of morbidity and mortality in many countries with large economic and human burdens, and it accounts for 20% of overall mortality in the United State [1]. It is reported that Ischaemic heart disease is the second leading cause for males and the third leading cause of global burden of disease for females, accounting for 6.8% and 5.3% respectively [2]. Although the benefit of some conventional drugs, such as aspirin and statin, have been demonstrated in reducing CHD mortality, annually 17.3 million people die from cardiovascular disease (CVD) worldwide (WHO 2008), and over 80% of CVD deaths take place in low and middle income countries, it is reported that by 2030 more than 23 million people will die annually from CVDs [3].
In recent years, traditional medicines have been playing more and more important roles in the maintenance of health, the prevention and treatment of diseases, and plant-based drug discovery [4][5][6][7][8]. Chinese herbal medicine or its products have been administered widely for treating CHD in China. There are more than one hundred kinds of patent herbal medicine for CHD available at present. Puerarin injection [9], Danshen preparations [10], Tongxinluo [11], compound salvia pellet [12], Suxiao jiuxin wan [13] or traditional Chinese herbal products [14] have been shown as potential benefits recently by systematic reviews. Sanqi is one of the most widely used herbal medicines in China, with function of invigorating the blood circulation according to TCM theory. Panax notoginseng was the active and effective component purified from sanqi. Oral Panax notoginseng products included xuesaitong capsule, xuesaitong dripping pills, xuesaitong pill, xuesaitong effervescent tablet, xuesaitong granule, xuesaitong dispersible tablet, sanqishutong capsule, Panax notoginseng saponins (PNS) tablet and PNS capsule. The content of Panax notoginseng varies in different agents. All of the agents have been used in clinic for patients with CHD for decades of years. Recent researches found its antioxidative [15], antiatherogenic, lipid-lowering, and anti-inflammatory [16] effects and angiogenic effect [17]. A Cochrane systematic review indicated that Panax notoginseng was effective in preventing stroke [18]. Some recent clinical trials also proved that it could benefit CHD patients [19,20]. Therefore, this systematic review aims to evaluate the safety and effectiveness of oral Panax notoginseng preparations for CHD patients.

Inclusion Criteria.
We included randomized controlled trials (RCTs) or cross-over trials in English and Chinese regardless of publication type in this review. Quasirandomized trials were excluded and the first stage of data was used if it was cross-over trial. Any adult participant with CHD meeting with at least one of the current or past definitions or guidelines of CHD (including acute coronary syndrome (ACS) and X syndrome) was considered. Those who did not introduce diagnostic criteria in the text but stated patients with definite CHD were also included. The trial was included if oral Panax notoginseng preparation was in intervention group regardless of dosage, treatment course, and agents; trials should be excluded if there were other Chinese herbal medicines in intervention group; trials also should be excluded if there was a combination of Panax notoginseng preparation and a kind of western medicine on the basis of control group. Chinese herbal injection should be excluded in this review. Placebo, no intervention, or nitrate was considered in control group, Chinese herbal medicine in control group should be excluded. Oral Panax notoginseng preparations versus conventional therapy (except for nitroglycerin) were excluded for limited extension.
Outcome measures include primary outcomes: all cause mortality, cardiovascular events (e.g., CHD mortality, incidence of myocardial infarction (MI), revascularization, and rehospitalization for unstable angina); secondary outcomes: quality of life, attack of angina pectoris (measuring by recurrence of angina pectoris, frequency of angina pectoris, duration of angina pectoris, dosage of nitroglycerin, decrement of nitroglycerin, efficacy of angina pectoris, and others), electrocardiogram (ECG), and adverse events. We defined the efficacy of angina pectoris as improvement was more than 50%; the efficacy of ECG as elevation of ST segment was more than 0.05 mv.

Search Strategy.
Two review authors (Qinghua Shang, Hao Xu) searched the following databases up to January 2013 independently for the identifications of trials (publication or nonpublication): the Cochrane Library, Pubmed, Chinese Biomedical database (CBM), China National Knowledge Infrastructure (CNKI), Chinese VIP Information (VIP), and Wanfang databases. We used the terms as follows: coronary heart disease, CHD, coronary artery disease, angina pectoris, myocardial infarction, acute coronary syndrome, cardi * , sanqi, sanchi, jinbuhuan, tiansanqi, tianqi, panlongqi, tongpitiegu, xueshancao, liuyuelin, xuesaitong, xueshuantong, notoginseng, pseudoginseng, Panax notoginseng, ginsenosides Panax, sanchinoside, and so forth. Because of different characteristics of various databases, MeSH terms and free text terms were used regardless of the report types in full text, title, keyword, subject terms, or abstract.

Data Extraction and Quality
Assessment. Two review authors (Qinghua Shang, Hao Xu) independently extracted data according to a data extraction form made by the authors. Disagreements were resolved by consensus or consultation from a third reviewer (Jianping Liu or Zhaolan Liu). The methodological quality of trials was assessed independently using criteria from the Cochrane Handbook for Systematic Review of Interventions, Version 5.0.1 (Qinghua Shang, Hao Xu) [15]. The items included random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other biases. We judged each item from three levels ("Yes" for a low of bias, "No" for a high risk of bias, and "Unclear" otherwise), and then we assessed the trials and categorized them into three levels: low risk of bias (all the items were in low risk of bias), high risk of bias (at least one item was in high risk of bias), and unclear risk of bias (otherwise).

Data Synthesis.
We used Revman 5.1 software provided by the Cochrane Collaboration for data analyses. Studies were stratified by the types of comparisons. We will express dichotomous data as risk ratio (RR) and its 95% confidence intervals (CI). Continuous outcome will be presented as mean difference (MD) and its 95% CI. Heterogeneity was recognized significant when 2 ≥ 50%. Fixed effects model was used if there is no significant heterogeneity of the data; random effects model was used if significant heterogeneity existed (50% < 2 < 85%). Sensitive analysis would be used if there was any heterogeneity (including differences of clinical characteristics among trials and the statistical heterogeneity); subgroup analysis would be used in patients prescribed Xuesaitong softy capsule. Publication bias was explored using a funnel plot.
There were 2 comparisons in the review according to various control groups: (1) Panax notoginseng preparations and conventional therapy versus conventional therapy (15 trials) [19, 21-24, 26-29, 31-36]; (2) Panax notoginseng preparations and conventional therapy versus nitrates and conventional therapy (2 trials) [25,30]. Two trials [24,28] were designed as three groups and four groups, respectively. Wang et al. [28] designed three groups with 2 comparisons: Panax notoginseng preparations and conventional therapy versus conventional therapy; Panax notoginseng preparations and trimethazine and conventional therapy versus conventional therapy; however, we extracted the data of first comparison for inclusion criteria. Ji and Zhang [24] designed four groups with 3 comparisons: Panax notoginseng coarse power and conventional therapy versus conventional therapy; Panax notoginseng semi-micron power and conventional therapy versus conventional therapy; Panax notoginseng micron power and conventional therapy versus conventional therapy; however, we summed up the three groups included Panax notoginseng as intervention group and conventional therapy as control group for data analysis.

Methodological Quality of Included Trials.
According to the criteria introduced above, no trial was evaluated as low risk of bias. Only one trial of the 17 trials reported the method to generate the allocation sequence (random number table) [23]. Two trials were assessed as having adequate concealment (concealed letter cover) [19,23]. No trial reported blinding method. One trial [31] reported the result of followup. No trial reported information on withdrawal/dropout. All of trials provided baseline data for the comparability among groups. The results of the assessment of risk of bias are presented in a "risk of bias summary" figure produced by Revman 5.1 automatically (Figure 2). (Tables 2 and 3   3.3.6. Reduction of Nitroglycerin. The definition of successful nitroglycerin reduction was that the patients in the trial stopped using nitroglycerin or the dosage of nitroglycerin was cut off more than 50% after the trial. Two trials [24,30] reported the condition of nitroglycerin. The results showed no significant improvement of Panax notoginseng preparation comparing with no treatment on the basis of conventional therapy (RR 1.41; 95% CI 0.89 to 2.24; 2 trials, = 250).
3.3.9. Angina Pectoris Immediate Effect. There was only one trial [25] which reported the angina pectoris immediate effect. 2 notoginsenoside pills were prescribed in this trial when angina pectoris happened. The criterion was defined as remarkably effective (angina was alleviated in 3 minutes); effective (angina was alleviated in 3-5 minutes); no effect (angina was alleviated in more than 5 minutes or need to add other medicines  [21,29]. There was 1 trial [33] which reported the frequency of angina pectoris in the unit of times/day. The result indicated that Panax notoginseng (Xuesaitong softy capsule) could reduce angina pectoris frequency compared with no treatment on the basis of conventional therapy (MD −2.76; 95% CI −3.87 to −1.65; 1 trial, = 28).

Dosage of Nitroglycerol.
There were 2 studies [21,26] reporting dosage of nitroglycerol in the unit of mg/week. Compared with no intervention on the basis of conventional therapy, oral Panax notoginseng preparation (Xuesaitong softy capsule) showed a reduction of nitroglycerol dosage (MD −1.13; 95% CI −1.70 to −0.56; 2 trials, = 212). There was 1 study [36] reporting dosage of nitroglycerol in the unit of mg/day, which showed Panax notoginseng preparation (Xuesaitong softy capsule) also reduced the nitroglycerol

Publication Bias.
A funnel plot analysis of the 7 trials in comparison of Panax notoginseng preparation and conventional therapy versus conventional therapy on angina pectoris improvement was conducted and shown in Figure 3; there might be a publication bias in this review for small sample, negative report, and low quality of the included trials.

Discussion
This systematic review included 17 RCTs and a total of 1747 participants. The review showed that, (1) comparing with no intervention on the basis of conventional therapy, oral Panax notoginseng showed no significant improvement for reducing the cardiovascular events, but it could relieve angina pectoris and related symptoms (including reducing the recurrence of angina pectoris, duration and frequency of angina pectoris, and dosage of nitroglycerol, as well as ECG changes); (2) oral Panax notoginseng showed similar immediate effect on angina pectoris compared with nitrate, but we could not make a significant conclusion from this equivalence due to small sample and low methodological quality trial; (3) The results also showed that oral Panax notoginseng was safe for CHD patients according to the information in hand, but it was too limited to make a conclusion for high risk bias and small sample in these trials.
Oral Panax notoginseng preparations have been used widely for treating CHD in China. Most of the researchers paid more attention to their pharmacological mechanism. Yang et al. comprehensively collected the pharmacological action of Panax notoginseng and concluded that it could provide protective effects against cardiovascular diseases through many pharmacological mechanisms including improving myocardial microcirculation, reducing arrhythmia, regulating blood lipid, preventing atherosclerosis, lowering blood pressure, and antishock [37]. Du et al. summarized the experiments on Panax notoginseng for MI and concluded that Panax notoginseng could inhibit the inflammatory reaction and improve ischemia reperfusion injury in patients with MI [38]. Chan et al. concluded that Trilinolein purified from Panax notoginseng could provide protective effects against cardiovascular disease including reducing thrombogenicity and arrhythmia and increase erythrocyte deformability. It was also an antioxidant which could counteract free radical damage associated with atherogenesis and myocardial damage [39]. All these experiments provided us laboratory evidence on protective effect of Panax notoginseng for CHD. Although many clinical trials were conducted on effect of oral Panax notoginseng preparations for CHD, there was no critical appraisal for these up to now. There was still no enough evidence for clinicians to prescribe oral Panax notoginseng preparations in CHD patients.
The impact of this review was to take a light on oral Panax notoginseng for CHD. Although it failed to prove the protective effect of Panax notoginseng on major cardiovascular events (cardiovascular mortality, MI incidence, and rehospitalization), it demonstrated that Panax notoginseng preparation might be recommended for improving symptoms of angina pectoris.
However, before translating the conclusion of this review to clinical practitioners, we have to consider the following weaknesses in this review. (1) Firstly, the "randomization" was not clear in most of the trials for insufficient reporting of generation methods of the allocation sequence, allocation concealment. Most trials stated only that patients were randomly assigned. (2) Secondly, no trial used placebo in control group, most of trials did not introduce double blind in this review, and one trial introduced blinding of outcome assessment. Therefore, in nonplacebo-controlled and nondouble blind trials, placebo effects may add to the complexity of interpreting the conclusion. (3) Thirdly, most of the trials did not introduce the study plan, and attrition bias and selective reporting bias might exist in this conclusion. (4) Fourthly, funnel plot indicated that publication bias would exist in this review. The reasons we considered were as follows: we only selected trials published in Chinese and English trials published in other languages or originated from other countries might be omitted; we only identified unpublished studies from conference paper or academic thesis, and negative trials might not be reported and induce publication bias.  [21] No abnormal changes appeared and no Ads was reported in the trial.
Ge and Zhao 2010 [22] Blood regular test, urine regular test, and blood biochemistry test had no changes compared with the previous.
Han and Chen 2008 [23] No serious Ads were reported in the trial; blood, urine, and stool routine tests, blood biochemistry test had no changes comparing with the previous.
Ji and Zhang 2003 [24] Blood, urine, and stool routine tests, and blood biochemistry test had no changes compared with the previous. No Ads was reported in the trial.  [33] No abnormal changes appeared and no Ads was reported in the trial.
Although this review suggested some benefit of Panax notoginseng preparation for CHD, the recommendation should be discreet due to poor quality and high risk bias of these trials, further rigorously designed, and well reported RCTs are still needed to prove the effectiveness and safety of Panax notoginseng preparation for CHD.

Conclusion
In this systematic review, oral Panax notoginseng preparation did not show benefit on reducing major cardiovascular events and relapse (including cardiovascular death, MI incidence, incidence of intractable angina pectoris, and rehospitalization), although it was effective in alleviating angina pectoris (including the recurrence, frequency, and duration of angina pectoris, ECG presentation, and dosage of nitroglycerin) with low adverse reaction. However, the small sample size and potential bias of most trials influence the convincingness of this conclusion. Before recommending oral Panax notoginseng preparation as an alternative herbal medicine in CHD patients, more rigorous trials with high quality are needed to prove the benefit of oral Panax notoginseng preparation and provide high level of evidence.