Scholarly article on topic 'Safflower yellow injection combined with conventional therapy in treating unstable angina pectoris: a meta-analysis'

Safflower yellow injection combined with conventional therapy in treating unstable angina pectoris: a meta-analysis Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"Safflower yellow injection" / "Angina / unstable" / Meta-analysis}

Abstract of research paper on Clinical medicine, author of scientific article — Dezhao Kong, Wei Xia, Zhe Zhang, Lei Xiao, Dongchao Yuan, et al.

Abstract Objective To evaluate the clinical efficacy of safflower yellow injection combined with conventional therapy in treating unstable angina pectoris. Methods We searched online databases: Chinese journal full-text database, China National Knowledge Infrastructure, Wanfang database, Chinese journal full-text database, Pubmed, ScienceDirect, Embase, and the Cochrane Library with manual-screening of relevant literature. Eligible randomized controlled trials (RCT) on angina pectoris were included. We conducted meta-analysis using the RevMan 5.1 software from The Cochrane Collaboration. We treated the relief rate of angina symptoms and electrocardiograph (ECG) as evaluation. Results Seven articles, including in 1134 patients, were enrolled after the evaluation. There was no significant heterogeneity among the studies (χ2=1.08, df=6, P=0.98, I2 =0%). The safflower yellow injection with conventional therapy has a higher effective rate than the control group in relieving the symptoms of angina pectoris [odds ratio (OR) = 2.95, 95% (CI) (1.81, 4.81)] and improving ischemic ECG [OR=2.85, 95% CI (1.67, 4.86)]. The difference was statistically significant in the “80 mg dosage” and “100 mg dosage” subgroups (P<0.05) for improving clinical symptoms and ECG. The funnel graphic was nearly symmetrical. Sensitivity analysis suggested that the results were stable. Conclusion Safflower yellow injection as an adjunct therapy with conventional drugs shows advantages in easing the clinical symptoms of unstable angina and improving ECG over basic therapy alone. However, the conclusions should be interpreted with care until more high-quality RCTs are reported.

Academic research paper on topic "Safflower yellow injection combined with conventional therapy in treating unstable angina pectoris: a meta-analysis"

JTCM

Online Submissions: http://www.journaltcm.com info@journaltcm.com

JTradit Chin Med 2013 October 15; 33(5): 553-561 ISSN 0255-2922 © 2013 JTCM. All rights reserved.

SYSTEMATIC REVIEW

Safflower yellow injection combined with conventional therapy in treating unstable angina pectoris: a meta-analysis

Dezhao Kong, Wei Xia, Zhe Zhang, Lei Xiao, Dongchao Yuan,Yue Liu, Guanlin Yang

Dezhao Kong, Wei Xia, Dongchao Yuan, Yue Liu, Guanlin Yang, Department of Cardiology, Liaoning University Of Traditional Chinese Medicine, Liaoning 110032, China Zhe Zhang, Lei Xiao, Guanlin Yang, Department of Cardiology, the Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Liaoning 110032, China Supported by Liaoning Province Science and Technology Plan Projects, Traditional Chinese Medicine Efficacy Evaluation Key Technology Research (No. 2010225034) and (No. 2010ZX09401 -304)

Correspondence to: Prof. Guanlin Yang, Department of Cardiology, the Affiliated Hospital of Liaoning University of Traditional Chinese Medicine; Liaoning University of Traditional Chinese Medicine, Shenyang 110032, China. yang-guanlin945@163.com Telephone: +86-15998854687 Accepted: March 21,2013

Abstract

OBJECTIVE: To evaluate the clinical efficacy of safflower yellow injection combined with conventional therapy in treating unstable angina pectoris.

METHODS: We searched online databases: Chinese journal full-text database, China National Knowledge Infrastructure, Wanfang database, Chinese journal full-text database, Pubmed, ScienceDirect, Embase, and the Cochrane Library with manual-screening of relevant literature. Eligible randomized controlled trials (RCT) on angina pectoris were included. We conducted meta-analysis using the RevMan 5.1 software from The Cochrane Collaboration. We treated the relief rate of angina symptoms and electrocardiograph (ECG) as evaluation.

RESULTS: Seven articles, including in 1134 patients, were enrolled after the evaluation. There

was no significant heterogeneity among the studies (x2=1.08, df=6, P=0.98, /2=0%). The safflower yellow injection with conventional therapy has a higher effective rate than the control group in relieving the symptoms of angina pectoris [odds ratio (OR) = 2.95, 95% (C/) (1.81, 4.81)] and improving ischemic ECG [OR=2.85, 95% C/ (1.67, 4.86)]. The difference was statistically significant in the "80 mg dosage" and "100 mg dosage" subgroups (P<0.05) for improving clinical symptoms and ECG. The funnel graphic was nearly symmetrical. Sensitivity analysis suggested that the results were stable.

CONCLUSION: Safflower yellow injection as an adjunct therapy with conventional drugs shows advantages in easing the clinical symptoms of unstable angina and improving ECG over basic therapy alone. However, the conclusions should be interpreted with care until more high-quality RCTs are reported.

© 2013 JTCM. All rights reserved.

Key words: Safflower yellow injection; Angina, unstable; Meta-analysis

INTRODUCTION

Coronary artery disease is heart disease with myocardial ischemia and hypoxia caused by coronary atherosclerosis or stenosis. These symptoms coupled with coronary artery spasms and changes are collectively known as coronary heart disease.1 The incidence of angina pec-toris has increased in recent years with improving living standards of Chinese, dietary changes,2'3 and increased stress.4'5 Angina pectoris has become the leading cause of death in China.6 The nomenclature and di-

agnostic standards of ischemic heart disease were published by World Health Organization in 1979.7 WHO divided ischemic heart disease into five types including occult coronary heart disease, angina, myocardial infarction, ischemic cardiomyopathy and sudden-death. Angina is further divided into three types: exertional angina, spontaneous angina, and mixed angina. Exertional angina includes stable exertional angina pectoris, the first episode of angina pectoris and progressive angina pectoris. Except stable exertional angina, the rest are unstable. These subcategories are included in the "unstable angina diagnosis and treatment recommenda-tions"8 developed by the Chinese Medical Society of Cardiology in 2000.

In recent years, safflower yellow injection (SYI) has been widely used for treating unstable angina pectoris. There have been many clinical trials reported confirming its effectiveness and safety.9-15 Honghua (Flos Car-thami), a Traditional Chinese Medicine, can activate blood and dredge the meridians, removing stasis and relieving pain.16 Ben Cao Hui Yan said, "Honghua (Flos Carthami) is a medicine breaking blood stasis, activating blood circulation, mediating blood, and adjusting blood." 17 Modern pharmacological studies have shown that its main chemical ingredients "safflower yellow pigment" have efficacy for some conditions. It can improve myocardial energy metabolism, reduce ischemic damage,18 and inhibit the activation of neutrophils.19 Some researchers report that it can soften artery athero-sclerosis,20 reduce myocardial ischemia,21 and resist inflammation by inhibiting nitric oxide (NO) and prostaglandin E2 (PGE2).22 SYI was made with safflower yellow pigment as the main ingredient. The common route of administration of SYI are injection and powder. Both were used in studies analyzed in this paper. In this study, we performed an analysis of multiple independent clinical randomized controlled trials (RCTs) to estimate the odds ratio (OR) and to systematically review the efficacy of SYI combined with conventional therapy in treating unstable angina.

MATERIALS AND METHODS

Data sources

We searched the literature according to the "Cochrane Handbook for Systematic Reviews of Interventions" published by The Cochrane Collaboration. We screened the online databases: Chinese journal full-text database (1989-August 2012), China National Knowledge Infrastructure (CNKI) (1979-August 2012), Wanfang Database (1977-August 2012), Pubmed (1950-2012), Sci-enceDirect (1823-2012), Embase (2009-2012), and the Cochrane Library (1948-2012). We also performed some manual-screening of relevant literature. The last search was performed on 30 August 2012.

Search strategies

We used terms depending upon the working language of the databases. Terms we used were: "Carthamus tinc-

torius injection," "Honghua (Flos Carthami)", "SYI," and "unstable angina". The following search strategy was used in PubMed and other English databases: (a) carthamus tinctorius injection; (b) Honghua (Flos Carthami); (c) SYI; (d) 1 in all fields; (e) 2 in all fields; (f) 3 in all fields; (g) 1 'OR' 2 'OR' 3. For searching the CNKI data and other Chinese databases, the following search strategy was used: (a) SYI; (b) unstable angina; (c) 1 in all fields; (d) 2 in all fields; (e) 1 'AND' 2.

Two reviewers independently read the title and abstract, and then read the full text after ruling out literature that obviously did not meet the inclusion standards. Then, a cross-check was conducted, and any disagreement was resolved through discussion or third-party opinions.

Inclusion criteria

We developed the inclusion and exclusion criteria according to the Cochrane Collaboration Handbook criteria. (a) Trials must be a randomized, controlled design. (b) Diagnosis of unstable angina should follow the "unstable angina diagnosis and treatment recommendations" developed by the Chinese Medical Society of Cardiology in 2000,8 or standards developed by the American College ofCardiology and American Heart Associa-tion.23 (c) Selected cases should be without severe organic diseases or complications. (d) The control group was given conventional treatment for the secondary prevention of coronary heart disease such as nitrates, aspirin, and adrenergic receptor blockers, statins, and angioten-sin II receptor antagonists. However, the experimental group was given SYI based on the conventional treatment. (e) There was no use of positive drugs such as Danshen injection, safflower injection, Hongdan injection and Chuanxiong injection. (f) Powder for injection ranged from 50-150 mg, without limits on the manufacturer or dosage forms. (g) The signals should be at least one of the symptoms of angina response rate and electrocardiograph (ECG) improvement rate, and should meet the evaluation standard.

Exclusion criteria

The exclusion standards were: (a) unpublished literature; (b) no control group set in the documents; (c) quasi-experimental design; (d) no clearly identified result assessment; (e) repeated original literature; or (f) multiple contributions ofone manuscript.

Symptom improvement standards

We performed the standard of symptom improvement according to the "Guidelines for clinical research ofTradi-tional Chinese Drug treating angina pectoris."24 After a course of treatment, the improvement of symptoms should be at least 80%, or nitroglycerin tablet consumption should decrease by 80% (i.e., significant symptomatic improvement). The improvement of symptoms should be at least 50%, or nitroglycerin tablet consumption should decrease by 50% (i.e., basic symptomatic improvement).

ECG improvement standard

We set out the ECG improvement standard according to the "Diagnosis and treatment guidelines for Chinese common cardiovascular and cerebrovascular disease."25 If the ST segment and T wave returned to normal then there was significant ECG improvement. If the ST segment recovered at least 0.05 mv, and the T wave inversion decreased more than 50% there was basic ECG improvement.

Data synthesis and statistical analysis

We used the Revman 5.1 software provided by "The Cochrane Collaboration" to analyze the collected data. An OR with 95% confidence intervals (CI) was chose to count data. A fixed-effect model of the Man-tel-Haenszel method was used to pool these OR if there was no significant heterogeneity. Subgroup analyses were performed to explore more information. We used the Q test and I2 test to analyze the heterogeneity across studies. A condition of P<0.05 was defined as a signal of heterogeneity. A condition of I2>50% suggested the data might not be probable to be combined.26 In

addition, we created a funnel plot to analyze potential publication bias,27 and tested the stability of the results by using sensitivity analysis.

Risk of bias in individual studies

According to the "Cochrane risk of bias tool"28 and "CONSORT 2010 checklist,"29 two reviewers independently assessed the quality of included studies in items such as randomization, allocation concealment, and blinding.

RESULTS

Search results

The process of study selection is shown in Figure 1. Overall, we collected 92 articles in Chinese, 354 in English, and 5 through manual screening. We screened 28 papers for full text after we sieved out the obviously ineligible and repeated documents. We screened out 21 of these studies for reasons such as lack of control group, and multiple contributions of one manuscript. Finally, seven articles were included in this review.

Studies included in quantitative

synthesis (meta-analysis)

Figure 1 Process of searching and screening study

CNKI: China National Knowledge Infrastructure; WF: Wangfang data; VIP: Chinese journal full-text database.

Characteristics of included trials

All enrolled patients were diagnosed as having unstable angina pectoris. The control group was given the basic treatment widely recognized as valid, and the experimental group was given SYI based on the same essential treatment. SYI dosage ranged form 50-150 mg once a day intravenously, with a course of 14-15 days (only "Liu JH 2009" used 10 days). The screened results by the two reviewers were the same (Table 1).

Quality assessment

We included 1134 subjects in seven papers, with average samples of 162. There were 554 total patients in the trial group and 580 in the control group. All stud-

Table 1 Characteristics of included trials in the meta-analysis

ies showed comparability or no significant difference between the two groups (P>0.05). One study described the random methods,11 and another included adverse drug reactions.10 All studies did not drop out of the clinical trails (Table 2).

Information from the CONSORT 2010 checklist conformity is shown in Table 3. The rates (total items actually reported/total number of items in CONSORT 2010) were: 100% for title and abstract, 71.43% for introduction, 85.87% for results, 66.67% for discussion, and 29.4% for methods. The rate of abiding by overall CONSORT 2010 of all included studies was 44.02%, indicating 44.02% of all items were satisfactorily reported on average.

Study Age Included number T Cb Treatment (day) Dosage of SYI mL (mg) Positive control Concomitant medication

Jin C 20119 45-75 65 65 14 150 None Secondary prevention

Wei ZF 201010 45-78 310 348 10-14 100 None Secondary prevention

Yu XW 201111 60-78 41 39 15 100 (80) None Secondary prevention

Liu JH 200912 42-78 28 20 10 50 None Secondary prevention

Zhou SJ 201113 45-75 38 38 14 100 (80) None Secondary prevention

Tang Q201114 52-76 43 42 14 150 None Secondary prevention

Xia HZ 201015 60±8.5 31 31 14 100 None Secondary prevention

Notes: T: test group; C: control group. SYI: Safflower yellow injection.

Table 2 Characteristics of included trials in the meta-analysis

Study Angina symptom relief ECG improvement

T C T C Adverse reaction

Jin C 20119 60 53 59 52 No

Wei ZF 201010 301 322 NA NA Yes

Yu XW 201111 39 35 38 25 No

LiuJH 200912 26 15 25 15 No

Zhou SJ 201113 35 28 34 28 No

Tang Q201114 41 33 36 32 No

Xia HZ 201015 30 27 NA NA No

Notes: ECG: electrocardiograph; T: test group; C: control group; NA: no data available.

Table 3 Total number of items compliant with the "CONSORT 2010 checklist"

Author Title and Introduction abstract Methods Results Discussion Other Total of each information study

JinC9 2 1 5 6 2 0 16

Wei ZF10 2 2 5 6 2 0 17

YuXW11 2 1 6 6 2 0 17

LiuJH12 2 1 5 6 2 0 16

Zhou SJ13 2 2 5 6 2 0 17

Tang Q14 2 1 5 6 2 0 16

Xia HZ 201015 2 2 4 5 2 0 15

Total of each field 14 10 35 41 14 0 114

Rated percentage (%) 100.00 71.43 29.41 58.57 66.67 0.00 44.02

Notes: rated percentage is total items actually reported in all studies/total number of items in CONSORT 2010 checklist.

Random sequence generation (selection bias) Allocation concealment (selection bias) Blingding of participants and personnel (performance bias) Blinding of outcome assessment (detection bias) Incomplete outcome data (attrition bias) Selective reporting (reporting bias) Other bias

I I Low risk of bias Q

Figure 2 Analysis on the risk of bias of included trails

The Cochrane risk of bias tool gave risk of bias within studies in Figure 2, 3: low bias in incomplete outcome data, unclear bias in selective outcome reporting, blind-

0% 25%

Unclear risk of bias

50% 75%

I High risk of bias

Jin C 2011

Liu JH 20091

Tang Q 201114

Wei ZF 20101

XiaHZ 20101

Yu XW 20111

Zhou SJ 20111 Figure 3 Summary of the risk of bias of trails included

Study or Subgroup

i conventional therapy conventional therapy Odds Ratio

Events Total Events Total Weiqht M-H. Fixed. 95% CI

60 65 53 65 18.3% 2.72 [0.90, 8.22]

26 28 15 20 5.6% 4.33[0.75, 25.15]

39 43 35 42 14.8% 1.95(0.53, 7.23]

301 310 322 348 39.6% 2.70(1.25,5.86]

30 31 27 31 3.9% 4.44 (0.47, 42.26]

39 41 35 39 7.9% 2.23 [0.38,12.92]

35 38 28 38 9.9% 4.17(1.05,16.61]

556 583 100.0% 2.86 [1.78,4.59]

ing of participants, personnel and outcome assessors, sequence generation (except for study 11), allocation concealment and other sources of bias.

Symptoms of angina

Meta-analysis of the seven studies showed no heterogeneity between studies (x2=1.08, f=6, P=0.98>0.05, /= 0%). Therefore, the fixed effects model (M-H: Man-tel-Haenszel formula) could be taken for analysis. In Figure 4, the summary odds ratio (OR) of the number of angina attacks decreased by >50% was 2.86, 95% CI (1.78, 4.59). The diamond in the forest plot is on the right side of the middle line. The overall effect of test Z=4.35, P<0.000 01 tells that efficacy of SYI with basic therapy on the symptoms of angina pectoris was much better than the control. Posted bias assessment of the included studies is shown in Figure 5. The funnel graphics are almost symmetrical, suggesting no publication bias.

ECG improvement

Among the five studies9,11-14 there were 419 patients, with 215 in the test group and 204 in the control group. No obvious heterogeneity existed among the studies (x2=2.94, f=4. P=0.57>0.05, /= 0%). Therefore, we used the fixed effect model analysis (MH Fixed method) to analyze. The results are shown in Figure 6. The diamond is in the right side of the middle line. The efficacy of SYI on ECG was significantly superior to that of drugs in the control group [OR=2.85, 95% CI(1.67, 4.86)] (Z=3.84. P=0.00001).

Odds Ratio r.l-H. Fixed. 95% CI

Jin C2011 Liu JH2009 Tang Q2011 WeiZF2010 Xia HZ2010 Yu XW2011 Zhou SJ2011

Total (95% CI) Total events 530

Heterogeneity ChF= 1.08, df= 6 (P = 0.98); l'= 0% Test for overall effect: Z= 4.35 (P < 0.0001)

Figure 4 Meta-analysis results of symptoms of angina SYI: Safflower Yellow Injection.

0.01 0.1 1 10 Favors experimental Favors control

SE (log [OR])

0.01 0.1 1 10 100 Figure 5 Posted bias assessment of the included studies Subgroup analysis

We made the subgroup analysis according to the dosage of injection. Studies were divided into four sub-

Study or Subgroup SYI with conventional therapy Events Total conventional therapy Events Total Weight Odds Ratio M-H. Fixed. 95% CI

Jin C2011 59 65 52 65 28.6% 2.46 [0.87, 6.93]

Liu JH2009 25 28 15 20 11.2% 2.7810.58,13.32]

Tang Q2011 36 43 32 42 31.4% 1.61 [0.55,4.72]

Yu XW2011 38 41 25 39 11.2% 7.09 [1.85, 27.23]

Zhou SJ2011 34 38 28 38 17.6% 3.04 [0.86,10.73]

Total (95% CI) 215 204 100.0% 2.85 [1.67,4.86]

Total events 192 152

Heterogeneity: Chi= = 2.94, df= 4 (P = 0.57); P = 0%

Test for overall effect: Z= 3.84 (P = 0.0001)

groups: 50 mg dosage,12 80 mg,11,13 100 mg,10,15 and 150 mg.914 The results showed that SYI with conventional therapy had a higher effective rate than routine therapy alone in the control group, except the subgroup "50 mg dosage" (Z=1.63, P=0.10). Five studies were divided into three subgroups about ECG improvement. Subgroup "80 mg dosage" had a higher effective rate than that in the control group on ECG improvement (Z=3.30, P=0.001). The difference was insignificant in subgroup "50 mg dosage" and "150 mg dosage" on ECG improvement (P>0.05) (Figure 7, 8).

Sensitivity analysis

To perform the sensitivity analysis, we replaced fixed

Odds Ratio IM-H, Fixed. 95% CI_

Figure 6 Meta-analysis results of electrocardiograph improvement

Experimental Control Odds Ratio

Study or Subgroup Events Total Events Total Weight M-H. Fixed. 95% CI

0.01 0.1 1 10

Favors experimental Favors control

Odds Ratio M-H, Fixed. 95% CI

42.1.1 50mg dosage

Liu JH2009 26 28 15 20 5.6% 4.33 [0.75, 25.15]

Subtotal (95% CI) 28 20 5.6% 4.33 [0.75, 25.15]

Total events 26 15

Heterogeneity: Not applicable

Test for overall effect: Z = 1.63 (P = 0.10)

42.1.2 80mg dosage

Yu XW2011 39 41 35 39 7.9% 2.23 [0.38,12.92]

Zhou SJ2011 35 38 28 38 9.9% 4.17(1.05,16.61]

Subtotal (95% CI) 79 77 17.8% 3.31 [1.12,9.74]

Total events 74 63

Heterogeneity: Chr = 0.30, df= 1 (P = 0.58); P = 0%

Test for overall effect: Z = 2.17 (P = 0.03)

42.1.3 100mg dosage

Wei ZF2010 301 310 322 348 39.6% 2.70 [1.25, 5.86]

Xia HZ2010 30 31 27 31 3.9% 4.44 [0.47,42.26]

Subtotal (95% CI) 341 379 43.5% 2.86 [1.38,5.93]

Total events 331 349

Heterogeneity: ChP = 0.17, df= 1 (P = 0.68); P = 0%

Test for overall effect: Z = 2.82 (P = 0.005)

42.1.4 150mg dosage

Jin C2011 60 65 53 65 18.3% 2.72 (0.90,8.22]

Tang Q2011 39 43 35 42 14.8% 1.95 [0.53, 7.23]

Subtotal (95% CI) 108 107 33.1% 2.37 [1.02,5.52]

Total events 99 88

Heterogeneity: ChP = 0.14, df= 1 (P = 0.70); P = 0%

Test for overall effect: Z = 2.01 (P = 0.04)

Total (95% CI) 556 583 100.0% 2.86 [1.78,4.59]

Total events 530 515

Heterogeneity: ChP= 1.08, df= 6 (P = 0.98) ;i== 0%

Test for overall effect: Z = 4.35 (P < 0.0001)

Test for subarouo differences: ChP = 0.47. df= 3 (P= 0.92). P = 0% Figure 7 Subgroup analyses on symptoms of angina

0.01 0.1 1 10 Favors experimental Favors control

Study or Subgroup

Experimental Events Total

Control Events Total

Odds Ratio Weiqht M-H. Fixed. 95% CI

Odds Ratio M-H. Fixed. 95% CI

39 38 77

43.1.1 50mg dosage Liu JH2009 25 28 Subtotal (95% CI) 28 Total events 25 Heterogeneity: Not applicable Test for overall effect: Z= 1.28 (P = 0.20)

43.1.2 80mg dosage Yu XW2011 38 41 Zhou SJ2011 34 38 Subtotal (95% CI) 79 Total events 72 Heterogeneity: Chf = 0.81, df= 1 (P = 0.37); \'= 0% Test for overall effect: Z= 3.30 (P = 0.0010)

43.1.3 150mg dosage

Jin C2011 59 65 52 65

Tang Q2011 36 43 32 42

Subtotal (95% CI) 108 107

Total events 95 84

Heterogeneity: Chi3 = 0.31, df= 1 (P = 0.58); P= 0% Test for overall effect: Z= 1.84 (P= 0.07)

11.2% 11.2%

2.78 [0.58,13.32] 2.78 [0.58,13.32]

11.2% 17.6% 28.8%

7.09 [1.85, 27.23] 3.04 [0.86,10.73] 4.61 [1.86.11.45]

28.6% 31.4% 60.1%

2.46 [0.87, 6.93] 1.61 [0.55, 4.72] 2.01 [0.96,4.23]

Total (95% CI) Total events 192 152

Heterogeneity: Chi== 2.94, df= 4 (P = 0.57); P= 0% Test for overall effect: Z= 3.84 (P = 0.0001) Test for subarouo differences: Chi== 1.92. df= 2 (P = 0.38). ?= 0% Figure 8 Subgroup analyses on electrocardiograph improvement

effects methods with random effects methods, excluded literature which had the largest number ofsamples,10 and short treatment course.10,12 As shown in Figure 9, 10, and 11: OR=2.85, 95% CI (1.77, 4.57); OR=2.97, 95% CI (1.63, 5.39); and OR=2.83, 95% CI (1.50, 5.33), respectively. OR values did not change significantly under different interventions, suggesting the stability of this result.

Adverse reactions

Six studies described adverse reactions. Only one study had minor side-effects including rash and dizziness, which were improved by adjusting the medication drip rate.

204 100.0% 2.85 [1.67, 4.86]

0.01 0.1 1 10 Favors experimental Favors control

DISCUSSION

SYI is a product of Traditional Chinese Medicine. However, the efficacy of the combined use of SYI and basic therapy remains controversial. In China, it is common to use SYI with routine therapy to treat unstable angina. However, no systematic review has been published in English, which hinders SYI's ability to gain worldwide recognition. We reviewed current RCTs to provide evidence for the use of SYI as an adjunctive therapy for unstable angina. Each study took broadly accepted diagnostic standards to determine treatment outcomes. In this study, we took the two signals of effects in symptoms and ECG improvement for analysis, which fully reflect the effectiveness of treating unstable

angina pectoris. Meta-analysis is the best method of literature analysis in evidence-based medicine. It can save costs and time integrating the data of multiple independent studies to increase sample size and to improving test performance.30 Moreover, it can provide clinicians with evidence-based medical evidence to guide their clinical practice. This meta-analysis showed that the efficacy of SYI combined with conventional therapy was better than basic therapy alone in terms of relieving symptoms of angina and ECG. From subgroup analyses, the difference was statistically significant (P<0.05) in the subgroup "80 mg dosage" and "100 mg dosage." Therefore, we suggested that the dosage of SYI in treating unstable angina should be 80 or 100 mg. Sensitivity analysis suggested that the results were stable. However, further clinical trials need to be performed before any firm conclusions can be drawn. Evidence from these articles9-15 and our analysis suggest that conventional therapy combined with SYI has an advantage in treating unstable angina.

There are several limitations in this study. First, only one study reported the randomization method.11 A lack of descriptions of randomization methods and allocation concealment may have biased the results. Second, blinding was not clearly described in any of the trials, which may overestimate the efficacy of the treatment group. In spite of the difficulty of implementing blinding methods of TCM clinical research, it is necessary to use blinding during the measurement of outcomes. Third, there were variations in manufacturing stan-

Study or Subgroup

conventional therapy conventional therapy Odds Ratio

Events Total Events Total Weight M-H. Random. 95% CI

60 65 53 65 18.4% 2.72 (0.90, 8.22)

26 28 15 20 7.3% 4.33(0.75,25.15)

39 43 35 42 13.1% 1.95(0.53,7.23)

301 310 322 348 37.6% 2.70(1.25,5.86]

30 31 27 31 4.4% 4.44 [0.47, 42.26)

39 41 35 39 7.3% 2.23 [0.38,12.92]

35 38 28 38 11.8% 4.17(1.05,16.61]

556 583 100.0% 2.85(1.77,4.57)

Odds Ratio M-H. Random. 95% CI

Jin C2011 Liu JH2009 Tang Q2011 Wei ZF201 0 Xia HZ2010 Yu XW2011 Zhou SJ2011

Total (95% CI) Total events 530

Heterogeneity:Tau* = 0.00; Chi»= 1.08, df= 6 (P = 0.98); l2= 0% Test for overall effect: Z= 4.32 (P < 0.0001)

Figure 9 Sensitivity analysis with random-effects methods

SYI with conventional therapy conventional therapy Study or Subgroup Events Total Events Total Weight Odds Ratio Г.1-Н. Fixed. 95% CI

Jin C2011 60 65 53 65 30.3% 2.72 (0.90, 8.22]

Liu JH2009 26 28 15 20 9.3% 4.33(0.75,25.15]

Tang Q2011 39 43 35 42 24.5% 1.95 (0.53, 7.23]

Xia HZ2010 30 31 27 31 6.5% 4.44 [0.47,42.26]

Yu XW2011 39 41 35 39 13.0% 2.23 [0.38.12.92]

Zhou SJ2011 35 38 28 38 16.4% 4.17(1.05,16.61]

Total (95% CI) 246 235 100.0% 2.97 (1.63, 5.39]

0.01 0.1 1 10

Favors experimental Favors control

Odds Ratio _M-H. Fixed. 95% CI

Total events

Heterogeneity: Chi== 1.05, df= 5 (P = 0.96); l== 0% Test for overall effect: Z= 3.57 (P = 0.0004)

Figure 10 Sensitivity analysis excluding maximum sample study

SYI with conventional therapy conventional therapy Study or Subgroup Events Total Events Total Weight Odds Ratio M-H. Fixed. 95% CI

Jin C2011 60 65 53 65 33.4% 2.72(0.90,8.22]

Tang Q2011 39 43 35 42 27.0% 1.95(0.53, 7.23]

Xia HZ2010 30 31 27 31 7.1% 4.44 (0.47, 42.26]

Yu XW2011 39 41 35 39 14.3% 2.23(0.38,12.92]

Zhou SJ2011 35 38 28 38 18.1% 4.17(1.05,16.61]

Total (95% CI) 218 215 100.0% 2.83 (1.50, 5.33]

0.01 0.1 1 10 Favors experimental Favors control

Odds Ratio M-H. Fixed. 95% CI

Total events 203

Heterogeneity: ChP= 0.84, df= 4 (P = 0.93); P = 0% Test for overall effect: Z = 3.21 (P = 0.001)

0.01 0.1 1 10 100 Favors experimental Favors control

Figure 11 Meta-analysis excluding studies with a short treatment

dards and two studies10'12 had a very short treatment course. These factors may contribute to heterogeneity among the studies. Nevertheless, the results proved to be stable after excluding these reports (Figure 11). Fourthly, the follow-up period is an important factor for the medical efficacy for patients with unstable angina pectoris. This meta-analysis included seven RCTs with short follow-up periods, which may influence the result. More studies with longer follow-up periods are needed to prove this trend. Fifthly, some studies reported the signals such as blood lipids and blood viscosity.912-14 However, methodological heterogeneity31 may be caused by different instruments. Therefore, these signals were not used in our analysis. Finally, publication bias may exist in the study. Although we thoroughly screened the English database, there were no English studies. Additionally, all the studies are positive results. Some negative results may be unreported and therefore, are not included in the review.

As most of the studies reviewed in the TCM field come from China, Chinese researchers should formally train in clinical trial design. Future trials should describe randomization methods, allocation concealment, and blinding methods during clinical research and measuring outcomes. Otherwise, it is difficult to exclude the

course

possibility of biases in the assessment of clinical efficacy, symptom scores, and other outcome indexes. Withdrawal and follow-up data are also necessary. Publishers play an important role in this effort as well. Researchers should strictly obey "The Randomized Controlled Experiment Report Specification-Consort,"32 especially in TCM clinical trials to provide quality evidence for the second evaluation and to increase the strength of the arguments.

This meta-analysis found that SYI combined with conventional therapy for unstable angina is more effective than basic therapy alone in terms of improving the symptoms of angina and ECG. However, this conclusion should be interpreted with care because of the low quality of the included trials. More rigorous design, large samples, multicenter, and high-quality RCTs need to be performed before any firm conclusions can be drawn.

ACKNOWLEDGEMENTS

We would like to thank Prof. Qinrong Zhang of Liaon-ing University of Traditional Chinese Medicine for his work on this article.

REFERENCES

1 Wang JY, Liao EY, Huang CX, Qi H. Internal medicine (for 8-year and 7-year clinical and other professional). Beijing: People's Medical Publishing House, 2010: 273.

2 Collaborative Group of the National "Ninth Five-Year" scientific and technological issues. Prevalent status and trends of cardiovascular disease risk factors of Chinese middle-aged crowd from the early 1980s to the late 1990s. Zhong HuaXinXue Guan Bing Za Zhi 2001; 29(2): 74.

3 Lu ZL. The control of dyslipidemia and cardiovascular risk factors. Zhong Hua Xin Xue Guan Bing Za Zhi 2001; 29(5): 315.

4 Yu SF, Li KR, Yang Y, et al. The relationship between occupational stress and cardiovascular disease risk factors. Zhong Hua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi 2003; 21(1): 12-15.

5 Yao SQ, Fan XY, Jin YL, et al. The impact of occupational stress on the vocational crowd cardiovascular function. Zhong Hua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi 2003; 21(1): 20-22.

6 Cardiovascular disease prevention research center of the ministry of health. Cardiovascular disease in China report 2008-2009. Beijing: China Encyclopedia Publishing House, 2009: 2-5.

7 Nomenclature and criteria for diagnosis of ischemic heart disease. Report of the joint international society and federation of cardiology/World Health Organization task force on standardization of clinical nomenclature. Circ 1979; 59 (3): 607-609.

8 Chinese Medical Society Of Cardiology, Medical Society Of Cardiology Society Of Cardiology Editorial Committee. Recommendations for diagnosis and treatment of unstable angina. Zhong Hua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi 2000; 28(6): 409-412.

9 Jin C. Safflower yellow pigment treatment of unstable angina pectoris 65 cases. Zhong Guo Zhong Yi Za Zhi 2010; 19(11): 1952.

10 Wei ZF, Wang XY, Wu J. Safflower yellow pigment treatment of unstable angina clinical observation. Jilin Yi Xue 2010; 31(20): 3208-3209.

11 Yu XW, Zhou YB. 41 cases with Safflower yellow pigment in treating elderly patients with unstable angina pectoris. Zhong Guo Zhong Yi Ji Zheng 2011; 20(3): 467.

12 Liu JH, Hou XJ. Safflower Yellow injection in treating unstable angina: an observation report. Zhong Xi Yi Jie He Xin Nao Xue Guan Bing Za Zhi 2009; 7(6): 374-375.

13 Zhou SJ, Li YL, Huang DM. 38 cases of Safflower Yellow injection for treatment of unstable angina. Hainan Yi Xue 2011; 22(20): 63-64.

14 Tang Q. Safflower Yellow injection in treating coronary heart disease with unstable angina. Shi Yong Zhong Yi Nei Ke ZaZhi, 2011; 25(11): 39-40.

15 Xia HZ. Influence of Safflower Yellow injections on platelet activation and fibrinolysis activity in patients with unstable angina. Yi Yao Lun Tan ZaZhi 2010; 31(19): 93-95.

16 Gao XM, Wang YY, Yan ZH. Chinese materia medica. Beijing: China Press of Traditional Chinese Medicine, 2007: 323.

17 Ni ZM. Ben Cao Hui Yan. Beijing: TCM Ancient Books Publishing House, 2005: 130.

18 Piao YZ, Jin M, Zang BX. A study of safflower yellow pigment on improving hypoxia myocardial energy metabolism. Zhong Cao Yao 2003; 34(05): 436-439.

19 Wu W Jin M, Tong J. Inhibitory effect of hydroxyl safflower yellow a against PMN activation induced by LPS. Act Pharmaceutical Sonica 2011; 46(2): 153-157.

20 Kasey, Shigekazu T, Masami F. Effects of safflower seed extract on arterial stiffness. Vascular Health and Risk Management 2010 NOV 3; (6): 1007-1014.

21 Wan LH, Chen J, Li L. Protective effects of Carthamus tinctorius injection on isoprenaline-induced myocardial injury in rats. Pharm Biol 2011; 49(11): 1204-1209.

22 Wang CC, Choy CS, Liu YH. Protective effect of dried safflower petal aqueous extract and its main constituent, carthamus yellow, against lip polysaccharide-induced inflammation in RAW264.7 macrophages. Journal of the science of food and agriculture 2011; 91(2): 218-225.

23 ACC/AHA. Guidelines for the management of patients with unstable angina and non-st-segment elevation myo-cardial infarction. A report of American college of cardiology/American heart association task force or practice guidelines. J Am Coll Cardiol 2000; 36(5): 970.

24 Ministry of Health of the People's Republic of China. Guidelines for clinical research of Traditional Chinese New Drug. Beijing: China Medical Science and Technology Press, 2002: 41.

25 Medical Administrative Department of the Ministry of Health of the People's Republic of China. Diagnosis and treatment guidelines for Chinese common cardiovascular and cerebrovascular disease. Beijing: Beijing Science and Technology Press, 2000: 59.

26 Wang D, Zhai JX, Mu ZY. The heterogeneity in Me-ta-analysis and its treatment. Zhong Guo Xun Zheng Yi Xue ZaZhi 2009; 9(10): 1115-1118.

27 Kang DY, Hong Q, Liu K. Analysis, identifying and treating the publication bias in meta analysis. Zhong Guo Xun Zheng Yi Xue Za Zhi 2003; 3(1): 45-49.

28 Cochrane handbook for systematic reviews of interventions 4.2.6 updated September 2006. The Cochrane Collaboration, 2006, 136. Cited 2012-08. Available from: URL: http://www.cochrane.org/training/cochrane-hand-book.

29 Kenneth FS, Douglas GA, David M, the CONSORT Group. CONSORT 2010 Statement: Updated guidelines for reporting parallel group randomised trials. J Clin Epidemiol 2010; (63): 834-840.

30 Liu MZ, Zhou RY. Statistical software applications in Traditional Chinese Medicine. Beijing: China Press of Traditional Chinese Medicine, 2006: 330.

31 Zhou XY, Fang JQ. The common bias in meta-analysis. Zhong Guo Xun Zheng Yi Xue Za Zhi 2002; 2(4): 216-220.

32 David M, Sally H, Kenneth FS, et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. J Clin Epidemiol 2010, (63): 1-37.