Shengmai Injection, a Traditional Chinese Patent Medicine, for Intradialytic Hypotension: A Systematic Review and Meta-Analysis

Intradialytic hypotension (IDH) is a global public health problem. A rising number of IDH sufferers resort to Chinese patent medicine, Shengmai Injection (SMI) in China. The objectives of present study are to assess the effectiveness and safety of SMI as an adjunct therapy for IDH. A systematic search of 6 medical databases was performed up to December 2011. Randomized trials involving SMI adjuvant therapy versus conventional therapy were identified. RevMan 5.0 was used for data analysis. Ten randomized clinical trials with 437 participants were identified. Methodological quality was considered inadequate in all trials. Compared with conventional therapy, SMI adjunct therapy showed significant effects in improving the clinic effective rate (P < 0.01), decreasing the incidence of IDH episode (P < 0.01), decreasing the frequency of nursing interventions (P < 0.01), and increasing diastolic blood pressure (P < 0.01). There was no statistical significance in the improvement of mean arterial pressure (P = 0.22) and systolic blood pressure (P = 0.08) between two groups. Four studies had mentioned adverse events, but no serious adverse effects were reported in any of the included trials. In conclusion, SMI adjunct therapy appears to be potentially effective in treatment of IDH and is generally safe. However, further rigorous designed trials are needed.


Introduction
Intradialytic hypotension (IDH) remains a common and intractable complication for end-stage renal disease (ESRD) patients undergoing hemodialysis [1]. It is de�ned as a decrease in systolic blood pressure (SBP) by ≥20 mm Hg or a decrease in mean arterial pressure (MAP) by ≥10 mm Hg associated with clinical symptoms (dizziness, blurred vision, cramps, and fatigue), affecting approximately 20% to 30% of dialysis sessions [2,3]. Frequent hypotension episodes during dialysis not only lead to a discomfortable feeling, limitation of rehabilitation, and consumption of a disproportionate amount of health care resources, but also contribute to high mortality in hemodialysis patients [4,5]. e etiology leading to IDH is still complex and incompletely understood, but the decline in blood volume, poor cardiac function, and an inadequate cardiovascular response were the main factors [6].
On the basis of the fundamental physiology of blood pressure, the predisposing factors for IDH can be divided into two categories [7]: (1) factors affecting cardiac output such as the decline of cardiac function, blood volume changes during ultra�ltration, and electrolyte changes. e combination of le ventricular hypertrophy, recurrent cardiac ischemic injury, and abnormalities of vascular structure and function may lead to myocardial �brosis with worsening diastolic function, chamber remodeling, and an increase in electrical excitability and arrhythmias. If the ultra�ltration rate exceeds plasma re�lling rates, the plasma volume, preload, and cardiac output will eventually fall. Electrolyte changes can impair myocardial electrical stability and contractility. (2) Factors affecting total peripheral resistance such as autonomic dysfunction (impaired sympathetic response, reduced barore�ex sensitivity, and Bezold-Jarisch re�ex), imbalance of vasoactive agents (impaired vasopressin response, elevated adenosine, and increased nitric oxide activity), temperature (thermogenesis and warm dialysate), and immune response to dialysis. Currently, there is no speci�c consensus on the medical therapy for the prevention and treatment of IDH. Several common therapies were utilized in the past decade including the Trendelenburg position [3], using of cool dialysate, sodium and ultra�ltration pro�ling, high dialysate calcium, blood volume control, avoidance of food during dialysis, correction of anemia, and pressor agents midodrine [8]. However, it remains necessary to seek novel effective and safe inventions for IDH.
Shengmai San is a well-known traditional Chinese herbal prescription, recorded in Yixueqiyuan (Origins of Medicine) by Zhang Yuansu at the beginning of 1186 [10], and has been applied for cardiovascular diseases routinely and prophylactically for thousands of years in China [11]. Shengmai San consists of 3 Chinese herbal medicines (CHMs): Renshen (Radix Ginseng; Ginseng), Maidong (Radix Ophiopogonis; Dwarf Lilyturf Tuber), and Wuweizi (Fructus Schisandrae Chinensis; Chinese Magnoliavine Fruit). All three herbs of SMI are included in the Chinese Pharmacopoeia (version 2010). eory of traditional Chinese medicine believes that Shengmai San has the effect of supplementing Qi and nourishing Yin, recovering pulse, and stopping abnormal sweating. Shengmai injection (SMI), which is developed on the basis of Shengmai San, is a popular modern Chinese patent herbal preparation. SMI is widely used in various cardiovascular diseases, and at least three systematic reviews to date have been conducted to evaluate the effectiveness of SMI on heart failure [12], fatality rate of acute myocardial infarction [13], and hypotension aer acute myocardial infarction [14].
Evidences have accumulated from former experiments to con�rm the effect of SMI on regulating blood pressure [15]. Especially, the widespread use of SMI on hypotension due to a variety of causes is noteworthy [14]. SMI can signi�cantly elevate blood pressure in hypotensive patients no matter if it is essential hypotension or with secondary reasons [14,16,17]. However, SMI has no signi�cant effect on blood pressure in healthy subjects [18].
Pharmacological studies have revealed the effects of SMI on multiaspects of the pathophysiology of IDH [7]. e related pharmacological mechanisms of SMI were as follows: (1) SMI can improve cardiac function through the protection of myocardial cells, reduction of ischemiareperfusion injury, reduction of myocardial apoptosis, prevention of myocardial calcium overload and alleviation of myocardial hypertrophy, enhancement of myocardial contractility, and protection of endothelial function [19]; (2) SMI can inhibit local angiotensin II activity so as to alleviate le ventricular hypertrophy [20]; (3) SMI had protective effects against oxidative damage in mitochondria, cells, and tissues [21,22]; (4) SMI had protective effects against experimental acute cardiogenic shock by improving the hemodynamics parameter [23]; (5) SMI can inhibit high sensitive C-reactive protein (hs-CRP) and in�ammatory cytokines such as tumor necrosis factor-and interleukin-8 and reduce the systemic in�ammatory reaction [24,25]; (6) SMI can enhance humoral immunity and inhibit the cellular immunity aer cardiopulmonary bypass [26]; (7) SMI can increase sympathetic tone, enhance sinus node function, and improve conduction [27]. (8) Impressively, Shengmai San can signi�cantly attenuate heat strokeinduced arterial hypotension and cerebral ischemia through inhibition of inducible nitric oxide synthase-(iNOS-) dependent nitric oxide (NO) overproduction in the brain and excessive accumulation of in�ammatory cytokines like interleukin-1 beta, interleukin-6, and tumor necrosis factor-alpha in the peripheral blood stream [28]. In addition, Ginseng, as the principal drug in the SMI, showed the effect of improving blood pressure stability in IDH patients. Chewing �orean red ginseng could signi�cantly reduce the degree of blood pressure drop during hemodialysis and the frequency of symptomatic IDH, and this bene�cial effects may be partially due to decreased NO production and more activation of vasoconstrictors including endothelin-1, renin activity (PRA), and angiotensin II (Ang II) [29].
However, the exact active ingredients of SMI for IDH treatment are still unclear. For the chemical composition of the individual Chinese herb of SMI, ginsenoside, ophiopogonin and ophiopogonone, and lignan have been proposed as the active components of Radix Ginseng, Radix Ophiopogonis, and Fructus Schisandrae Chinensis, respectively [30]. ere are a number of reports about the effective chemical constitutes and different analytical methods for analyzing constituents in SMI. High performance liquid chromatography (HPLC) have even been widely employed for content determination of Shengmai preparations [31]. Recently, by the use of the liquid chromatographyelectrospray ionization source in combination with hybrid ion trap and high-resolution time-off light mass spectrometry (LC-IT-TOF/MS), more than 30 ginsenosides and 20 lignans were readily detected and structurally characterized from SMI [30]. Interestingly, by using the on-line high performance liquid chromatography-diode array detectionchemiluminescence (HPLC-DAD-CL) method and liquid chromatography coupled with tandem mass spectrometry (LC/MS/MS) analysis, the scavenging activities of main components detected in the individual herb were different from those in whole Shengmai San, suggesting that drug interactions in complex multiherbal formula could change the activity of the constituents [32].
Over the past decades, a number of trials have indicated that SMI could have therapeutic potential in people with IDH in China. However, the evidences for the effects of SMI have not been systematically assessed. e objective of the present study is thus to assess the clinical effectiveness and safety of SMI adjunct therapy for IDH patients.

Methods
is systematic review is conducted according to the paper [9]. therapy for IDH patients were included, regardless of blinding, publication status, and language. Quasi-RCTs were not considered such as using the admission sequence for treatment allocation.

Eligibility
2.1.2. Types of Participants. Patients of any age or sex with end-stage renal disease (ESRD) who were receiving longterm regular hemodialysis and had experienced episodes of IDH were included. e diagnostic criteria were adopted in accordance with the following. (1) Diagnosis of IDH was made on the basis of "De�nition of IDH" in 2005 from the European Dialysis and Transplant Association and K/DOQI guideline, a decrease in SBP ≥20 mm Hg or a decrease in mean arterial blood pressure (MAP) ≥10 mm Hg associated with dialysis-related hypotension symptoms [2]. (2) Diagnostic criteria of IDH with comparable de�nitions was made on the basis of �lood puri�cation, second edition written by Wang in 2003, a reduction in SBP below 90 mm Hg, or a decrease in SBP ≥20 mm Hg from prehemodialysis [33]. None of them received antihypertensive drugs or any other intervention known to in�uence the blood pressure before dialysis.

Types of Interventions. SMI in any dose compared
with the conventional therapy for IDH was considered. We only included studies that compared SMI with conventional therapy. Studies comparing SMI with another CHM were excluded.

Outcome
Measures. e outcome measures included the clinical effective rate of SMI for IDH, the incidence rate of hypotension, the number of nursing interventions, blood pressure level, and adverse events. Clinical effectiveness is de�ned as the ability of SMI to improve hemodynamics and clinical symptoms related to IDH. Evaluation standards for clinical therapeutic effects were as follows [34]: (1) markedly effective: the SBP increased more than 20 mm Hg or SBP >90 mm Hg or MAP increased by ≥10 mm Hg compared with pretreatment, with no hypotension-related symptoms, and dialysis to be completed successfully; (2) effective: SBP increased by 10∼20 mm Hg or SBP >90 mm Hg or MAP increased by ≥0-10 mm Hg compared with pretreatment, with no obvious symptoms of low blood pressure, and dialysis to be completed by adjusting the dialysis program; (3) ineffective or deterioration: blood pressure did not rise or continued to decline, SBP dropped to less than 90 mm Hg, and patients showed signi�cant symptoms of low blood pressure, need vasopressors, volume expansion and other drug treatment to maintain blood pressure or were forced to interrupt dialysis. We also checked the references of published studies to identify additional trials.
e following search terms were used as medical subject headings and key words when searching electronic databases: end-stage renal disease, end-stage renal failure, end-stage kidney failure, Shengmai, Sheng-mai Injection, hemodialysis related hypotension, intradialytic hypotension, IDH, and low blood pressure. ese terms were used as Mesh and freetext terms (translated into Chinese) to search the Chinese databases.

Study Selection and Data
Extraction. Two review authors (C.-y. Chen, L.-y. Lu) independently examined the titles and abstracts of the potential references. Full articles for all potentially relevant studies were retrieved. e two reviewers then read the selected papers independently and made a �nal selection decision. Disagreements were resolved through discussion or consultation with a third author (Y. Wang). If necessary. e authors of the trials were contacted and asked to provide missing data.
e review authors extracted data on study characteristics, including patients, methods, interventions, and outcomes, into a standardized data extraction form. Reasons for the exclusion of studies were recorded. For eligible studies, two review authors (C.-y. Chen, L.-y. Lu) extracted data independently. Any disagreements were resolved by consensus or by a third reviewer (Y. Wang).

Risk of Bias in Individual
Studies. Assessment of risk of bias in included studies: two review authors (C.-y. Chen, L.-y. Lu) independently assessed risk of bias for each included article, using the twelve criteria recommended by the Cochrane Back Review Group [35]. e items were scored with "yes (+), " "no (−), " or "unsure (?). " Disagreements were resolved through discussion with or involving a third author (Y. Wang).

Data Synthesis and
Analysis. e statistical package RevMan 5.0 provided by the Cochrane Collaboration was used to analyze the data. Dichotomous data were presented as odds ratio (OR), with 95% con�dence intervals (CI). Continuous outcomes were presented as weighted mean difference, with 95% CI. Meta-analysis was only performed within comparisons where individual trials compared similar treatment and control interventions.

Description of Studies.
We identi�ed and screened 181 potentially relevant articles. Of these, 102 articles were initially excluded due to duplicate publications by reading titles and abstracts, and 53 articles were excluded because they were case reports or lack in-comparison group, or not reports of clinical trials, or effectiveness of SMI not being objective of the studies. In the identi�ed 2� potentially eligible reports, aer reading the full text, 14 articles were excluded due to comparing SMI with another CHM, and 2 more articles were Full-text articles excluded -Not real RCT (2) -Comparing SMI with another CHM (14) -E cacy of SMI not being F 1: �RISMA 2009 �ow diagram, from [9]. For more information, visit http://www.prisma-statement.org/statement.htm. excluded because they were not real RCTs with hemodialysis order used for treatment allocation [36,37]. erefore, a total of 10 studies were �nally included papers [38][39][40][41][42][43][44][45][46][47]. Flow diagram was summarized in Figure 1.

Characteristics of Included Studies.
A total of 437 participants were involved in the 10 studies included (Table 1). All studies were conducted in China and published between 1999 and 2010 on Chinese journals. Each study was performed in a single center, parallel-designed, and claimed to have applied randomization. 8 studies included 180 male and 124 female, while the other 2 studies did not mention the gender condition [44,46]. e age of the participants ranged from 15 years to 78 years. Etiology for ESRD was introduced in 184 patients in 5 studies [38,[41][42][43]47], including 107 chronic glomerulonephritis, 15 diabetic nephropathy, 35 hypertensive nephropathy, 10 obstructive nephropathy, 2 polycystic renal disease, 1 chronic pyelonephritis, 5 gouty nephropathy, and 9 other types of nephropathy. 3 studies reported the modality of dialysis, on bicarbonate dialysis for 4-5 hrs and 2-3 times a wee� with a low-�ux polysulfone hollow-�ber dialyzer [40,41,45]. 6 studies reported the duration of the dialysis from one month to 5 years [38,39,[41][42][43]45]. All of the 10 included trials were two-group parallel design studies.
In the interventions, conventional therapy referred to treatment according to the European Dialysis and Transplant Association and K/DOQI guidelines, including the use of cool dialysate, sodium and ultra�ltration pro�ling, high dialysate calcium, blood volume control, avoidance of food during dialysis, correction of anemia, and the use of pressor agents such as midodrine [2,8]. e doses of SMI used ranged from 40 mL to 60 mL. SMI was administered intravenously in all included studies. A variety of outcome measures were reported. Evaluation of the outcomes was performed at the end of the treatment.

Risk of Bias in Included
Studies. e risk of bias of each study was assessed using the twelve criteria recommended by Cochrane Back Review Group. e number of criteria met varied from 2/12 to 5/12. All of the studies included claimed randomization. No study described allocation concealment.
No trials mentioned the blinding procedures. One study described intention-to-treat analyses [47]. None of the trials mentioned drop-out data. ere was selective reporting in all the studies. All the studies described similarity of baseline except two studies [38,40]. All of the included studies appeared to have adequate and acceptable compliance and timing of outcome assessments were similar. In general, all of 10 RCTs have an unclear risk of bias. e methodological quality of each study is summarized in Table 2.

Results of Individual Studies.
Zhao et al. [38] conducted an RCT to test the effect of SMI on correcting IDH. 100 hemodialysis sessions were divided into two subgroups: the treatment group received SMI 40 mL intravenously, and the control group received normal saline injection. e results showed that the total clinical effective rate was 85% in treatment group and 55% in control group ( ). In the study of Liu and Su [39], 70 IDH patients were randomly divided into experimental group and control group. e experimental group received SMI 60 mL intravenously plus conventional therapy, while only conventional therapy was given for control group. e total clinical effective rate was 88.57% in experimental group and 62.86% in control group ( ). e frequency of �uid infusion treatment in experimental group was signi�cantly lower than that in the control group ( ). Zhou [40] recruited 14 patients (totally 140 hemodialysis sessions) and randomly divided into two groups. e therapy group was given SMI 60 mL intravenously and the control group was given 50% glucose 60 mL correspondingly. e result showed that the hypotension rate was 8% in therapy group and 38% in control group ( ). In the study of Zheng et al. [41], patients in treatment group were additionally given SMI 50 mL intravenously. e hypotension rate was 18.8% in treatment group and 33.1% in control group ( ). SBP, diastolic blood pressure (DBP), and MAP were all signi�cantly higher in treatment group than in control group ( ). e difference of MAP between the two groups was also statistically signi�cant ( ). Jiang et al. [42] selected 18 patients (352 hemodialysis sessions) and randomly divided them into two groups. Patients in control group were given midodrine hydrochloride tablet before and aer dialysis. Patients in therapy group were additionally given SMI 60 mL intravenously on that basis. e overall effective rate was 88.5% in therapy group and 69.7% in control group ( ). SBP, DBP, and MAP were all signi�cantly higher in therapy group compared with control group ( ). e difference of MAP aer dialysis was signi�cant between the two groups ( ). In the study of Lv and Liu [43], patients of treatment group received SMI 60 mL treatment, while patients of control group only received 0.9% saline. ere were no signi�cant differences between two groups in MAP, systolic pressure, diastolic pressure, and heart rate ( ). Clinical effective rate in experimental group was signi�cantly higher than control group ( ). Number of measures taken to rectify the dialysis-related symptoms were treatment group 2 3± 2 times and control group 4± times. e difference was statistically signi�cant ( ). In the study of Wang [44], the total effectiveness rate was 86.8% in SMI group and 62.9% in conventional therapy group ( ). Number of measures taken to rectify the dialysis-related symptoms were: SMI group 2 3 ± times, conventional therapy group ± 3 times ( ). In the study of Yu [45], control group was given 50% glucose + conventional therapy. Treatment group was given SMI 60 mL + conventional therapy. e rate of hypotension in treatment group was signi�cantly lower than that of control group ( ). e clinical effective rate in treatment group was higher than that of control group ( ). Cao et al. [46]   conventional group. e mean arterial blood pressure of SMI group was, predialysis: 96.4 ± 13.1 mm Hg; postdialysis: 97.8 ± 9.1 mm Hg; conventional group, predialysis: 99.2 ± 9.5 mm Hg; postdialysis: 99.7 ± 8.6 mm Hg. Number of measures taken to rectify the dialysis-related symptoms were SMI group: 2.4±1.1 times; conventional group: 5.4±1.8 times. ere was a signi�cant difference between the two groups ( . 5). In the study of Li [47], the therapy group was given the following treatment: SMI + 50% glucose, i.v., while the control group was given 0.9% sodium chloride injection or 20% human albumin or fresh plasma, ivgtt. Results showed that level of blood pressure and improvement of clinical symptoms were signi�cantly better in therapy group than in control group ( . 1). [38-41, 44, 45, 47] calculated the clinical effective rate with the ratio between the proportion of responders in the treatment group and in the control group. e 7 independent trials showed homogeneity in the consistency of the trial results (chi-square = 3.70, .72, 2 %). us, ��ed-effects model should be used for statistical analysis. e combined effects showed that patient with IDH receiving SMI therapy had signi�cantly improved the clinical effective rate when compared with the control group (OR 3.74, 95% CI 2.59 to 5.39; 7. 5, .

e Number of Nursing
Interventions. 4 studies recorded the number of nursing interventions for IDH episode [39,43,44,46]. Routine nursing interventions are as follows: placing the patient in the Trendelenburg position, saline and hyperoncotic albumin boluses, decreasing the transmembrane ultra�ltration pressure, and early termination of dialysis. e 4 trials showed homogeneity in the results (chi-square = 0.58, .9 , 2 %). us, ��ed effects model should be used for statistical analysis. ere was a signi�cant decrease on frequency of nursing interventions in SMI group (WMD −3.01, 95% CI −3.   be used for statistical analysis. ere was no statistical signi�cance in increasing MAP between two groups (WMD 7.83, 95% CI −4.66 to 20.33, , ), Figure  6. 3 studies reported pre-and post-SBP, and pre-and post-DBP [39,41,42]. e trials did not show homogeneity in the trial results, thus random-effects model should be used for statistical analysis. ere was no statistical signi�cance in increasing SBP when compared with control group (WMD 9.02, 95% CI −1.07 to 19.11, , ), Figure  7, but there was a signi�cant increase in DBP in SMI group (WMD 2.84, 95% CI 1.42 to 4.27, , ), Figure 8.

Adverse Events.
Four studies reported nonserious adverse events [38,41,42,47]. e other 6 studies did not report adverse events [39,40,[43][44][45][46]. Zhao et al. [38] indicated no statistically signi�cant difference in serum creatinine, blood urea nitrogen, serum electrolytes, and electrocardiogram before and aer hemodialysis in treatment group and control group ( ). ere was no case report of toxic side effects or allergy in treatment group. Zheng et al. [41] found no signi�cant change in heart rate before and aer dialysis. ere were no adverse reactions in the two groups during dialysis. Jiang et al. [42] reported that no signi�cant difference in heart rate before and aer dialysis in the two groups. ere was no signi�cant difference in routine blood test, blood urea nitrogen, creatinine, alanine aminotransferase, albumin, urea clearance index (Kt/V) in the two groups before and aer treatment ( ). ere were no adverse reactions in patients of the two groups, and the treatment was well tolerated. Li [47] demonstrated that the side effects in SMI group were lower than that of control group. In the control group, allergic reactions and transfusion reactions occurred in 4 cases, heart failure in 2 cases, dialyzer clotting in 8 cases, and early termination of dialysis was 6 cases due to no improvement of clinical symptoms and blood pressure. In the SMI group, dialyzer clotting occurred in 1 case and could continue hemodialysis aer replacing the dialyzer. All patients completed the expected dialysis and no adverse reactions such as allergic reactions, abdominal distension, tachycardia, and hypotension happened. clinical trials must be registered in order to be considered for publication [48]. However, none of the included studies in this paper had been formally registered in WHO International Clinical Trials Registry Platform. us, protocols were not available to con�rm free of selective reporting. ere are also a number of methodological limitations in this meta-analysis. Firstly, the data were all collected from the published articles without directly contacting the authors for obtaining additional information about the included studies. erefore, the twelve criteria of the "risk of bias" assessment tool could only be classi�ed as "unclear. " Secondly, all studies included in this paper used an "A+B versus B" design where patients were randomized to receive SMI plus conventional therapy versus conventional therapy, without a rigorous control for placebo effect. is kind of design is likely to generate false positive results [49]. irdly, all 10 studies claimed to be RCTs, but they all failed to give adequate and convincing information on how the random allocation was generated and concealed, which is necessary to avoid selection bias. ey also did not mention blinding method, and thus could produce performance bias and detection bias. erefore, outcome assessment was prone to signi�cant systemic errors. Intention-to-treat analysis was mentioned only in one study [47], and no dropouts were reported. us, the results generated from these studies should be interpreted with caution. Fourthly, the included studies were of relatively small sample size and without formal sample size calculation. Trials that lacked proper sample size estimation placed their statistical analysis's validity in doubt. Baseline information on ESRD patients was insufficient, with 6 trials provided information on chronic hemodialysis duration [38,39,[41][42][43]45] and 5 studies reported the etiology of ESRD [38,[41][42][43]47]. Varying dialyser, dialysis, membrane and dialysate were used in different studies. e lack of baseline information may lead to selection bias and not to comparable baseline.

Discussion
No study found severe adverse effects of SMI. Due to the small sample size, safety still needs to be assessed. Publication bias may also exist because only Chinese language publications were found and included.

Implications for
Practice. is is the �rst meta-analysis of randomized, controlled trials to assess the effectiveness and safety of SMI adjuvant therapy in patients with IDH. However, the evidences available from this systematic review is insufficient to recommend the routine use of SMI as adjuvant therapy for IDH, because the strength of the evidences is compromised by methodological �aws and lack of replicable validation. e effectiveness and safety of SMI therapy for IDH remain to be further determined.

Implications for Research.
First, improvement in the methodological quality of randomized controlled trials is critical for future research and more methodologically rigorous studies are justi�ed to con�rm or refute the effects reported here. Second, the included trials were generally of small sample size. All the trials were in lack of sample size estimation, so sample size calculation should be conducted before enrollment. Relevant clinical events such as death, dependency, and activities of daily living at the longer followup period should be included in outcome assessment. ird, well-designed, randomized, double-blind, placebocontrolled trials need to be carried out and reported in detail according to CONSORT [50] or CONSORT for TCM [51,52].

Con�ic� of �n�eres�s
e authors do not have any con�ict of interests.