Background
Stroke occurs when the blood supply to the brain is interrupted or reduced, which may be caused by blocked arteries or leaking and bursting of blood vessels [
1]. This deprives the brain of the supply of oxygen and nutrients, causing damage to the brain tissue [
2]. Being one of the most severe neurological diseases, stroke often leads to death or gross physical impairment or disability [
2], and the morbidity of stroke continues to rise with an annual increase of 8.1 %. Although the incidence of stroke is declining in many developed countries, the absolute number of stroke patients continues to enlarge because of the ageing population [
3]. Of those, acute ischemic stroke as the dominant type of stroke, account for 85 % of all strokes [
4].
Currently, rapid assessment and early intervention for patients suffering an acute ischemic stroke is of top priority [
5,
6]. There is as yet no routine, effective, generally accepted, specific treatment for AIS, except for aspirin [
7] and anti-thrombolysis with tissue plasminogen activator approved by FDA of US is administered in the setting of AIS within 4.5 h from symptom-onset [
8‐
10]. Therefore, it is necessary to test other promising therapeutic approaches to acute ischemic stroke [
11,
12]. The lack of effective and widely applicable pharmacological treatments for acute ischemic stroke patients leads to a growing interest in Chinese traditional medicine [
13,
14], for which extensive observational and anecdotal experience has accumulated over the past one thousand years [
14,
15]. Of those, Chinese herbal medicine is the most addressed in China, US and European Union [
12,
13]. In China, a series of optional treatments for treating AIS are available. Traditional Chinese medicine have proven to be effective complementary intervention for AIS, of which Chinese herb products are still prevailing well and make a significant difference [
12,
14,
15]. The overall treatment concept for Chinese traditional medicine is different from Western medicine, and herb medicines were always used to treat acute ischemic stroke. ShenXiong glucose injection is one of the most widely used Chinese herb medicine.
ShenXiong glucose injection (also named as Danshen-Chuanxiongqin injection) has properties that might be effective in acute ischemic stroke [
16], it is on the national essential drug list of China as a new drug for cardio-cerebral vascular diseases. ShenXiong glucose injection is a compound preparation, made up of Salvia miltiorrhiza and Ligustrazine [
16]. Of those, fifteen kinds of biologically active substances have been extracted from the dried roots of Salvia miltiorrhiza, a traditional medical herb known as Danshen, among which tanshinone, isotanshinone and hydroxy-tanshinone are the important ingredients [
16,
17]. Various studies suggested that Danshen could decrease intracranial pressure by decreasing overall water content of the brain and cerebrospinal fluid volume, and by reducing blood volume due to vasoconstriction [
18,
19]. DanShen may also improve cerebral perfusion by decreasing viscosity. It is reported to increase coronary blood flow, suppress thromboxane formation, decrease cerebral oedema, infarct size and neurological deficit [
19]. It is always used either alone or in combination with other herbal ingredients for patients with coronary heart diseases, hyperlipidemia, cerebrovascular diseases. In both China and other countries like US, such findings come either from biomedical studies [
19] or from clinical trials [
19,
20]. Ligustrazine, a bioactive ingredient extracted from Chuanxiong (Rhizoma Chuanxiong), was also proved to have beneficial effect on cerebrovascular diseases in pharmacological studies [
21]. ChuanXiong could improve brain microcirculation by preventing thrombus formation and platelet aggregation and blood viscosity in order that it might have significant beneficial effects in the treatment of acute ischemic stroke [
21]. As the compound of Salvia miltiorrhiza and ligustrazine, ShenXiong glucose injection has been widely accepted as a standard treatment for acute ischemic stroke in China for over 15 years [
16]. Its mechanism is associated with exciting histamine receptors, improving neuron function and reducing serum total cholesterol-triglyceride levels, increasing serum high density lipoprotein levels, and enhancing tissue plasminogen activator activity. Moreover, ShenXiong glucose injection can prevent free radical injury from ischemic brain tissue by reducing lipid peroxide [
16].
Currently there are a large number of studies of the clinical efficacy of ShenXiong glucose injection in acute ischemic stroke published in the global, especially in China. However, those clinical trials have not been reviewed systematically at present, whether the existing evidence is of sufficient scientifically rigorous and whether ShenXiong glucose injection can be recommended for routine use based on current evidence is still uncertain. To make specific recommendations for further research into the effectiveness and safety of ShenXiong glucose injection for patients with acute ischemic stroke, we carry out a systematic review and GRADE approach in twofold: (i) to systematically review all the randomized and quasi-randomized controlled trials of ShenXiong glucose injection for acute ischemic stroke; (ii) in addition, to evaluate the quality of current evidence by the use of GRADE system, and to provide the best available evidence for clinical practice.
Methods
Criteria for considering studies for this review and GRADE approach
Type of studies
We included all randomized controlled trials (RCTs), and quasi-randomized controlled trials (trials that used a non-random method of treatment allocation, e.g. hospital number, date of birth or day of the week), either published or unpublished.
Types of participants
Trials involving patients of any age or sex with apparent acute ischemic stroke (excluding intracerebral hemorrhage) were eligible. Stroke must be diagnosed in accordance with the diagnostic criteria of neuro-imaging verification of pathos-logical alterations in the brain, with computed tomography (CT) or Magnetic Resonance Imaging (MRI). Of these criteria, “Key Points for Diagnosing Cerebrovascular Diseases” and “Stroke diagnosis curative standard” modified in the 4th National Cerebrovascular Disease Seminar by the China Medical Society in 1995 [
22] and “Guidelines of diagnosis and treatment of Chinese acute ischemic stroke” in 2010 [
23], were adapted.
Type of interventions
Regardless of frequency, intensity or the duration of treatment, either ShenXiong glucose injection versus placebo or ShenXiong injection plus conventional treatments versus conventional treatments alone were included in this review and GRADE approach.
Any other routine treatments except ShenXiong glucose injection, such as antihypertensive, anti-platelet aggregation and other symptomatic treatments if necessary, were defined as conventional treatment, and all participants either in experimental group or in controlled group would receive same conventional treatments during whole trial period.
Type of outcome measures
The primary outcome is death or dependency at the end of scheduled follow-up. Dependency is defined as severely dependent on others in activity of daily living, based on the definition of the Barthel Index scores 60 or less, or an Oxford handicap grade 3 to 6.
Secondary outcomes were: (1) Changes of neurological deficit after ShenXiong glucose injection treatment and at the end of scheduled follow up. The measures could concentrate on specific impairment (for instance, Motricity Index, Motor Assessment Scale) or global neurological deficit (for instance, the National Institute of Health Stroke Scale or the Neurological Function Deficit Score). The clinical neurological impairment score of stroke patients is measured by the National Institute of Health Stroke Scale (NIHSS) or the Neurological Function Deficit Score (NFDS) . The neurological function deficit score changes, such as NFDS reduction rate over 18 % was taken as curative effect standard [
24]; (2) Quality of life (QOL) at the end of follow up, which is measured by the Nottingham Health Profiles or Spiter Quality of Life Index; (3) Adverse events including bleeding, dizziness, vomiting, allergic reaction, or other adverse events caused by ShenXiong glucose injection. The number of patients developing at least one adverse event listed above was estimated.
Search strategies
Electronic Search
We searched the Cochrane Central Register of Controlled Trials (CENTRAL, Ovid, from 1991 to January 2015), MEDLINE (PubMed, from 1966 to January 2015), EMBASE (Ovid, from 1966 to January 2015), CBM (Chinese Biomedicine Database, from 1978 to January 2015), Chinese TCM Database (from 1949 to January 2015), CNKI (China National Knowledge Infrastructure, until January 2015), VIP (Chinese Scientific Journal Database from 1989 to January 2015), WanFang Database (from 1998 to January 2015). The key words comprise DanshenChuanxiongqin, Danshen-Chuanxiongqin, ShenXiong glucose injection, Dan Xiong and ShenXiong injection were used as English and corresponding Chinese search terms to identify studies from aforementioned databases. Reference lists of all included studies were also searched for publications that meet the inclusion criteria.
Searching other resources
In order to search all the related studies, we also hand-searched the following relevant journals of TCM (the last issues were set at January 2015): World Journal of Integrated Traditional and Western Medicine, Chinese Journal of Integrative Medicine, Journal of Chinese Integrative Medicine, Journal of Traditional Chinese Medicine and Journal of Beijing University of Traditional Chinese Medicine.
Data collection and analysis
Selection of studies
Clinical trials were identified by screening the titles and abstracts of targets by two independent reviewers (XL, LP), and any disagreements were resolved through discussion and consultation with a third reviewer (DK).
Data extraction and management
Two reviewers (XL, LP) independently extracted data from studies that met the inclusion criteria using a pre-designed form, and disagreements were resolved by discussing with the third reviewer (DK). We planned to retrieve the following data: methods (study design and execution, method of randomization), participant characteristics (number of patients, age, sex, inclusion criteria, method of diagnosis), interventions (description of interventions given to each treatment group including, with or without other combined treatment), outcomes, methods of analysis (intention-to-treatment analysis or per-protocol analysis, or both) and statistical methods used.
Assessment of risk of bias in included studies
Two reviewers (XL, LP) independently assessed the methodological quality of each included studies using the “risk of bias” assessment tool outlined in the Cochrane Handbook for Systematic Reviews of interventions (Version 5.1.0) [
25]. Any disagreements were resolved by discussion with the third reviewer (DK).
Statistical analysis
The Cochrane Review Manager software, RevMan 5.1 [
26], was used to calculate treatment effects across trials. For discontinuous data, we used relative risk (RR) with the corresponding 95 % confidence interval (CI) as the measure of effect. For continuous data, the measure of the treatment difference for any outcome would be the mean difference (MD) when the pooled trials using the same rating scale or test, otherwise, the standardized mean difference (SMD) with 95 % confidence interval (CI) would be used. A Chi-square based test of homogeneity was performed using Cochrane Q statistic [
27] and I
2. If
P value less than 0.10 or I
2 value exceed 50 % substantial, indicated heterogeneity among trial results significantly, we would carry out a subgroup analysis to explore the sources of heterogeneity, including clinical heterogeneity, such as the route of administration, time and dose interval of start of treatment, type of intervention, as well as the methodological heterogeneity, such as design type of the trials, the risk bias of happening and so on. If I
2 was less than or equal to 50 % we would use a fixed-effect meta-analysis, otherwise, if the underline true effect being assumed to follow up a normal distribution, we would use a random-effects model (in which case the confidence intervals will be broader than those of a fixed-effect model). Sensitivity analysis was carried out by using the leave-one-out approach to explore the influence of heterogeneity, to determine how robust the results of analysis were. Funnel plots and Egger analysis were used to assess the existence of publication bias, if possible.
Besides, in the process of interpreting results and coming to conclusions, we planned to use GRADE approach (grading of recommendation assessment, development, and evaluation, GRADE) to evaluate the quality of evidence. This approach is a systematic method for evaluating quality of evidence and formulating strength of recommendation if sufficient information is available [
28]. Firstly, we subsequently classified the importance of each outcome. Three categories of outcome were specified based on their importance (median score of 1 to 9). Then the quality of evidence is divided into four categories: high, moderate, low, and very low. The GRADE approach defines clearly the methodological criteria by which evidence can be upgrade or downgrade and the justification for the strength of recommendation being formulated as “strong” or “weak”. Critical and important outcomes were used for decision making and were included in the evidence profile and summary of Findings Tables [
29].
Discussion
Summary of main findings
This review was done in accordance with the methods of a Cochrane systematic review. We identified and included 12 trials involving a total of 974 participants. This systematic review has shown that ShenXiong glucose injection given as a treatment in the acute stage of ischemic stroke significantly effect on improving activities of daily living function at 7 days, 14 days and 28 days whilst on treatment, but a meta-analysis of this outcome using data from the five studies was not performed due to obvious heterogeneity among the study estimates. Significant improvements on activities of daily living function were observed from individual trials’ results. We performed subgroup analysis based on the type of treatment (ShenXiong glucose injection and Danshen ligustrazine injection). No clear difference in Barthel index score at 28 days was present between ShenXiong glucose injection trial and Danshen ligustrazine injection trial. None of the included studies reported the pre-specified primary outcome of death during the follow-up period, it may be explained that authors attempted to emphasize short-term effect of ShenXiong glucose injection, or that there were no deaths during the short-term follow up period. Furthermore, pooled results showed ShenXiong glucose injection had significant effect on the improvement of neurological function deficit at 14-day and 28-day whilst on treatment. But this results should be interpreted with caution because of severe heterogeneity was seen among the review. We performed subgroup analysis based on the type of treatment (ShenXiong glucose injection and Danshen ligustrazine injection) to explore the reasons for large heterogeneity. It showed no evidence of difference in efficacy between ShenXiong glucose injection and Danshen ligustrazine injection. There is evidence that ShenXiong glucose injection has clinical effects, such as higher response rate, this result is consistent with the improvement of neurological function deficit. Lastly, only six trials reported that no adverse events were observed during the trials period, while the left studies failed to report whether adverse events happening or not. We cannot draw a conclusion on the safety of ShenXiong glucose injection. There is currently not enough evidence to determine whether the use of ShenXiong glucose injection in acute ischemic stroke would result in any harmful effects.
After applying GRADE approach for this systematic review, although rating down for the risk of bias was not required, Unexplained heterogeneity was found within included studies, while the available data did not permit an interpretation as to the reason for some of the inconsistencies we found, such inconsistency is also a basis for downgrading the level of evidence to support ShenXiong injection for AIS patients. In addition, regarding the two important outcomes, the total number of events or patients does not exceed the OIS (Optimal Information Size), therefore, rating down for imprecision was warranted [
38]. By the way, the quality of evidence was not reduced in concerning indirectness. And finally, despite the fact that none of the included studies were found by pharmaceutical industry, the funnel plot present asymmetric trend on clinical efficacy, and the risk of publication bias may exist in somewhat.
Therefore, our GRADE approach leads to the decision that ShenXiong glucose injection for a individual patient with acute ischemic stroke should be approached cautiously, patients‘values and preferences should be considered and allow them to participate in the medical decision making process.
The limitations of the twelve trials assessed to be of inferior methodological quality were as follows
1. Only two trials using a random method to divide the groups, the remaining 10 trials reported ‘randomly allocating’ participants but the method of randomisation was not described. 2. None of the included trials reported allocation concealment, and whether a blinding method used or not within twelve trials is unclear, increasing the risk of selection or performance bias happening. 3. In addition, all included studies were conducted in China and were published in Chinese journals. A funnel plot analysis also indicated asymmetry trend for the outcome of response rate, publication bias maybe exist, although we attempt to perform comprehensive search as well as supplementary search for both published and unpublished literature. 4. Another limitation of the review relate to the over-proportion of small trials to be included, may cause imprecision and inconsistency of effect size among those trials. Although the estimate of effect on the outcomes is significant statistically, the observed effects could also simply be due to bias rather than to a biological effect of ShenXiong glucose injection.
In addition, the usefulness of evidence was limited due to the following reasons,. Firstly, the follow-up time was insufficient to evaluate the long term effects of ShenXiong glucose injection, in the 12 trials, follow-up time ranged from 7 days (1 week) to 28 days (4 week) after the start of treatment. Secondly, the reporting on adverse effect was poor in most of included trials, it is highly implausible that no adverse event happen among 494 patients, the safety of ShenXiong injection is still inconclusive. And additional limitations in this review include the diversity of outcome measure used by the included studies, and the limited duration of follow up (generally no more than 2 months).
In summary, the overall quality of evidence was judged to be low to moderate due to important limitations exist in study design, inconsistency, lack of directness, imprecision and possible publication bias. The result of published intervention trials are therefore of insufficient quality to enable evidence-based recommendations to be developed for clinical practice in acute ischemic stroke. This review systematically describes current evidence from latitudinal and longitudinal studies on ShenXiong glucose injection and identifies methodological aspects to be improved in further research. This methodological improvement is pivotal for the assessment of the body of evidence on ShenXiong glucose injection treatment. High-quality evidence from methodologically via double-blinded randomized controlled trials are required to provide further support for administration of ShenXiong glucose injection for patients with acute ischemic stroke.
Competing interests
This work was supported by funds from State Administration of Traditional Chinese Medicine.
Authors’ contribution
DK, PR and QH conceived and designed the study, XL and LP developed search strategy, searched data, trial selected, data extraction, as well as contacting authors for additional data. XL, LT, WY and LP performed data analyses, presented results, and revised the manuscript. XL and SW interpreted results and draw conclusions. All authors declared that there are no conflicts of interest. All authors read and approved the final manuscript.