The Effect of Chinese Herbal Medicine Combined With Western Medicine on Vascular Endothelial Function in Patients With Hypertension: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Objective Vascular endothelium plays a fundamental role in regulating endothelial dysfunction, resulting in structural changes that may lead to adverse outcomes of hypertension. The aim of this study was to systematically evaluate the effect of a combination of Chinese herbal medicine (CHM) and Western medicine on vascular endothelial function in patients with hypertension. Methods We systematically searched the literature for studies published in Chinese and English in PubMed, Embase, Cochrane Central Register of Controlled Trials, Chinese Biomedical Literature Database, China Knowledge Resource Integrated Database, Wanfang Data, and China Science and Technology Journal Database. Databases were searched using terms concerning or describing CHM, hypertension, vascular endothelium, and randomized controlled trials. RevMan 5.3.0 was used for data analysis. If the included studies were sufficiently homogeneous, quantitative synthesis was performed; if studies with different sample sizes and blind methods were used, subgroup analyses were performed. GRADEpro was selected to grade the current evidence to reduce bias in our findings. Results In this review, 30 studies with 3,235 patients were enrolled. A relatively high selection and a performance bias were noted by risk of bias assessments. Meta-analysis showed that the combination of CHM and conventional Western medicine was more efficient than conventional Western medicine alone in lowering blood pressure (risk ratio, 1.21; 95% CI, 1.16 to 1.26) and increasing nitric oxide (95% CI, 1.24 to 2.08; P < 0.00001), endothelin-1 (95% CI, −1.71 to −1.14; P < 0.00001), and flow-mediated dilation (95% CI, 0.98 to 1.31; P <0.00001). No significant difference was observed between the combination of CHM and conventional Western medicine and conventional Western medicine alone for major cardiovascular and cerebrovascular events. CHM qualified for the treatment of hypertension. The GRADEpro presented with low quality of evidence for the available data. Conclusion CHM combined with conventional Western medicine may be effective in lowering blood pressure and improving vascular endothelial function in patients with hypertension. To further confirm this, more well-designed studies with large sample sizes, strict randomization, and clear descriptions about detection and reporting processes are warranted.


INTRODUCTION
Cardiovascular disease (CVD) is the primary cause of death-and disability-adjusted life-years (DALYs) worldwide and is responsible for 12.9 million deaths and 0.3 billion DALYs each year Murray et al., 2012), for which high blood pressure (BP) is the strongest risk factor (Writing Group et al., 2016). In 2015, 1.13 billion individuals were reported to have CVDs worldwide (Collaboration, 2017). The vascular endothelium plays a fundamental role in regulating the vascular tone and structure as well as endothelial dysfunction, resulting in structural changes that may lead to adverse outcomes of hypertension (Juonala et al., 2006). Well-maintained endothelial function and integrity are of great significance in numerous conditions, including hypertension, inflammatory and cardiovascular diseases, and their risk factors (Sattar et al., 2003;Touyz and Briones, 2011).
Antihypertensive agents, including diuretics, beta-adrenergic blocking agents (b-blockers), angiotensin-converting enzyme inhibitors (ACEIs), angiotensin II receptor blockers (ARBs), and calcium channel blockers (CCBs) are mostly used in the treatment of CVDs. However, around half of the patients are incapable of effectively controlling their BP by drug therapy owing to the associated cost, adverse reactions, and complications (Shaw et al., 1995). Behavioral interventions, such as exercise, weight loss, and salt intake casn control and help lower the BP, but these are hard to comply with.
Traditional Chinese Medicine (TCM) has proven to be an important part of complementary and alternative medicine (CAM) because of efficacious clinical practice in China, although its mechanism remains unclear (Harris et al., 2012).
Many studies have shown that hypertension can be effectively managed by Chinese herbal medicine (CHM), acupuncture, and tai chi (Shih et al., 2005;Wang et al., 2007;Yin et al., 2007;Kim and Zhu, 2010). In clinics, patients with essential hypertension are commonly treated with CHM combined with antihypertensive agents. However, there is no systematic summary available of the RCTs examining the efficacy of CHM plus antihypertensive drugs (CPAD) for vascular endothelial function in patients with essential hypertension. Thus, to critically assess the efficacy of CPAD for essential hypertension, we performed this systematic review.

METHODS
This meta-analysis was carried out and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Moher et al., 2010). A protocol has been registered in PROSPERO for this review (registration number: CRD42019140743).

Participants
In this meta-analysis, we did not restrict based on patient's age, gender, course of disease, case source, nationality, or race. In the original literature, the definition of hypertension is consistent with past guidelines (systolic BP [SBP] ≥140 mmHg or diastolic BP [DBP] ≥90 mmHg) (Hypertension Alliance (China) et al., 2019). The exclusion criteria included (a) subjects with hypertension complicated by other serious CVDs, hepatic failure, or renal failure, (b) secondary hypertension, (c) gestational hypertension, or (d) isolated systolic hypertension. effect under certain circumstances. All types of CPADs without considering the dose, method of dosing, composition of the formula, or time of drug administration were compared with the Western medicine. The comparisons included in this study were as follows: (a) CHM combined with CCB vs CCB; (b) CHM combined with ACEI vs ACEI; (c) CHM combined with ARB vs ARB; (d) CHM combined with diuretic vs diuretic; (e) CHM combined with multiple antihypertensive drugs (CCB/ACEI/ ARB/diuretic/b-blocker) vs multiple antihypertensive drugs.

Control
As mentioned above, the control could be CCB, ACEI, ARB, diuretic, b-blocker alone, or multiple antihypertensive drugs.

Outcomes
Primary outcomes were 24 h ambulatory BP monitoring (24 h-SBP and 24 h-DBP), SBP, DBP, and therapeutic effectiveness with reference to the standards of Chinese Medicine Clinical Research of New Drugs Guiding Principles. Marked effectiveness was considered as: DBP decreased ≥20 mmHg but did not reach normal level; or 10 mmHg ≤ DBP decrease < 20 mmHg and reached normal level; BP decrease <10 mmHg with normal level; or 10 mmHg ≤ DBP decrease < 20 mmHg but did not reach normal level; or SBP decrease ≥ 30 mmHg but did not reach normal level; ineffectiveness: DBP decreased < 20 mmHg and did not reach normal levels. Secondary outcomes were nitric oxide (NO), endothelin-1 (ET-1), flow-mediated dilation (FMD), vascular endothelial growth factor (VEGF), high-sensitivity C-reactive protein (hs-CRP), angiotensin II (Ang II), von Willebrand factor (vWF), and transforming growth factor b-1 (TGFb-1). Among them, NO protects the vascular endothelial function, whereas increase in ET-1, VEGF, hs-CRP, Ang II, vWF, and TGFb-1 adversely affects vascular endothelial function. The decrease in FMD indicates an impairment of the vascular endothelial function.

Study Type
Randomized controlled trials (RCTs) that combined CHM with antihypertensive drugs to treat hypertensive patients regardless of blinding, were included. To minimize publication bias, there was no restriction on language and time.

Literature Searches
We conducted a systematic search of studies published in Chinese and English using the following databases: PubMed, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), Chinese Biomedical Literature Database (CBM), China Knowledge Resource Integrated Database (CNKI), Wanfang Data, and China Science and Technology Journal Database (VIP). These databases were searched from inception to April 2019. Terms related to CHM, hypertension, and RCTs were searched in these databases. The search was not restricted by language or publication dates. The search strategy that suited each database was as follows: ("Hypertension" OR "High Blood Pressure") AND ("Endothelium, Vascular" OR "Endothelium") AND ("Medicine, Chinese Traditional" OR "Drugs, Chinese Herbal" OR "Traditional Chinese Medicine" OR "Chinese medicine" OR "Chinese medica" OR "Chinese drug" OR "Chinese patent medicine" OR "CHM" OR "Herbal medicine" OR "Chinese Herbal Drug" OR "Chinese Plant Extract").
After removing duplicates, two investigators independently screened the titles and abstracts of the articles obtained via initial search for relevance. Abstracts that did not meet our criteria were excluded, and few abstracts with insufficient information regarding the inclusion criteria were further reviewed. The full texts of the remaining results were further determined by the same investigators but by blinding to each other's review. All disagreements were resolved by consensus. The flow chart of the study selection is shown in Figure 1.

Data Extraction
Data were extracted in duplicate by two investigators independently and were inputted to a dedicated database. The data extracted from each article included basic information (study ID, document type, author, and publication year), participant's demographic details (sample size, age, and sex), diagnostic criteria, inclusion and exclusion criteria, study drug and control treatment, outcomes, fall outs, follow-up duration, and outcomes. Disagreements regarding the extracted data were settled by a third reviewer (AAN).

Risk of Bias Assessment
We used the Cochrane Collaboration "risk of bias" (ROB) tool for ROB assessment. Referring to the Cochrane Handbook criteria for assessing ROB in the ROB assessment tool, two investigators independently assessed the methodological quality of the included studies using RevMan 5.3.0. Following the handbook, the assessment of ROB was graded as low, unclear, or high based on the following seven aspects: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcomes assessment (detection bias), incomplete outcomes data (attrition bias), selective reporting (reporting bias), and other biases. All differences during ROB assessment were resolved by consensus.

Statistical Analyses
RevMan 5.3.0 was used to analyze the results of this study. For binary data, estimates were described as relative risk (RR) and 95% confidence interval (CI). For continuous data, the weighted mean difference (MD) or standard mean difference (SMD) and 95% CIs were calculated. Only complete case data were selected for further analysis. Heterogeneity between the studies in effect measures was assessed using both the chi-squared test and the I 2 statistic (Higgins et al., 2003) with an I 2 value >50%, indicative of substantial heterogeneity. Sufficiently homogeneous distribution allowed the use of quantitative synthesis in both statistics and clinic. When the I 2 value was lower than 50% and P >0.10, a fixed-effect model was adopted; otherwise, a random-effect model was suitable.
Because of significant heterogeneity in primary outcomes, studies with different sample sizes and bindings were subjected to separate subgroup analyses. As few results of the subgroup analysis revealed low methodological quality of the included studies and significant positive results, no further sensitivity analyses were performed.
As the least number of studies in each project was 10, a funnel plot was drawn to detect publication bias. To minimize bias in our findings, we selected the online software GRADEpro to summarize findings for outcomes to evaluate the available evidence. The assessment included biases including risk, inconsistency (heterogeneity), indirect, imprecision, and publication biases; each evidence was graded as very low, low, moderate, or high.

Literature Screening
Herein, we have described the literature retrieval process. A total of 578 potentially relevant articles from seven electronic databases were retrieved after the literature search. After removal of duplicates, 395 articles were identified. After going through the titles and abstracts, 272 articles that were case reports, case series, reviews, and animal studies irrelevant to hypertension were excluded. After reading the full text of the remaining 123 articles, 93 studies were further removed for at least one of the following reasons: no RCT (n = 11); participants failing to meet the inclusion criteria (n = 24); duplicates (n = 4); therapeutic measures failing to meet the predetermined inclusion criteria (n = 28), no Western medicine used in the control group (n = 5), and no BP data for extraction (n = 21). Thirty articles in accordance with the inclusion criteria were identified.

Characteristics of Included Studies
The enrolled 30 articles were published between 2010 and 2019, in which all studies related to comparison of CPAD group vs antihypertensive drug group (Control group). In the CPAD group, the standard, type, and dosage of antihypertensive drugs used were identical to those in the control group.

ROB Assessment
Two reviewers independently extracted the data from these included studies and conducted an ROB assessment using the Cochrane Collaboration's tool for ROB assessment. In this systematic review, all 30 trials were reported as RCTs. Of 30, 18 reported the generation of the allocation sequence. Among them, 15 used a random number table Shen et al., 2013;Ou and Li, 2014;Lu et al., 2015;Bian, 2016;Feng et al., 2016;Teng et al., 2016;Wang et al., 2016;Zheng et al., 2016;Ruan et al., 2017;Zeng Z. C. et al., 2017;Han et al., 2018;Liu et al., 2018;Luo, 2018;Shi et al., 2018), two used a randomization list generated with an SAS software package (Wu et al., 2014;Cao et al., 2019) and one used the stratified block randomization (Sun, 2019). Only one trial (Wu et al., 2014) reported allocation concealment. Two trials reported the use of double blinding (Xu et al., 2010;Wu et al., 2014). Three trials reported withdrawals (Gao et al., 2012;Teng et al., 2016;Zheng et al., 2016). No protocols for the included studies were available for us to investigate the selective reporting. We found only one study (Wu et al., 2014) that is included in this review to be at low risk for selective reporting, with the remaining 29 studies being assessed as high risk. No other potential sources of bias were detected. The ROB in the included studies is shown in Figures 2 and 3.

Adverse Event
Of all the included studies, four studies (Gao et al., 2012;Wu et al., 2014;Teng et al., 2016;Zheng et al., 2016) reported losses to follow, and nine (Wu et al., 2014;Ou and Li, 2014;Lu et al., 2015;Weng and Lin, 2015;Bian, 2016;Wu et al., 2016;Zeng Z. C. et al., 2017;Luo, 2018;Shi et al., 2018) studies reported adverse events. On initial recruitment, there were 3,265 patients whose blood pressures were measured. Altogether, 24 cases failed to follow-up and we finally collected 3,235 cases (99.09%) with complete data. Of nine studies reporting adverse events, five reported no adverse events during the study and three described the frequency of adverse effects in detail. The adverse reactions of the CPAD and control groups included edema of the lower extremity, flushing, and headache. All the reported adverse reactions were not aggravated and they all disappeared after symptomatic treatment.

GRADE Evidence Profile
The GRADE evidence profile and summary of the findings are detailed in Table 2. Due to serious ROB in study methods, the heterogeneity and reporting bias, overall quality of evidence for 24 h-SBP, 24 h-DBP, SBP, DBP, NO, ET-1, and FMD were assessed as very low quality, low quality, indicating that these estimates were uncertain, and further studies are likely to influence our confidence when estimating CHM effects ( Table 2).

Publication Bias
Funnel plot analysis for the outcomes of SBP (A), DBP (B), NO (C), ET-1 (D), and therapeutic effects (E) was performed to explore the publication bias. The funnel plot was asymmetric, suggesting a mild publication bias in this systematic review ( Figure 6).

DISCUSSION
TCM herbal formulas have always been recommended as complementary and alternative treatments for hypertension in China and other countries (Xiong et al., 2018). With the concerns of long-term medication and adverse reactions of antihypertensive drugs, some mild to moderate hypertensive patients who are not willing to take antihypertensive drugs would prefer CHM either used alone or in combination with antihypertensive drugs. In their opinion, TCM was efficacious in improving symptoms, reducing fluctuations in BP, improving vascular endothelial function associated with hypertension, and reducing the amount of conventional Western medicine despite its bitter and slightly sweet taste (Wu et al., 2014). In addition, as CHM has been used for thousands of years, it seems to be relatively safe (Xiong et al., 2015). However, whether TCM is beneficial for hypertension on vascular endothelial function is not well recognized until now. To our knowledge, this is the first systematic review and meta-analysis of the published RCTs to sum up the effects of TCM for hypertension treatment on vascular endothelial function and provide the latest level of evidence for patients, policymakers, and clinicians.

Summary of Main Results
As compared with the conventional Western medicines, the results of this meta-analysis showed marked improvements in BP and vascular endothelial function for hypertensive patients treated with CPAD, although there was some heterogeneity among these studies. NO and ET are essential substances synthesized and secreted mainly by endothelial cells for dilatation and vascular contraction, respectively, and their levels in blood were used to evaluate vascular endothelial injury (Zhong et al., 2011). FMD is a conventional assessment of conduit artery function with great cardiovascular insight            (Broxterman et al., 2019). Meta-analysis in this study proves that CPAD could obviously protect vascular endothelial function in hypertensive patients by increasing the serum NO levels, reducing that of ET, and improving FMD. Hypertension is also related to endothelial dysfunction. Hypoxia induced by vascular endothelial injury is one of the most important factors in inducing VEGF expression. VEGF is an important mediator of the Ang II-induced inflammatory reaction of vasculature. It releases VEGF, attracts circulating neutrophils and monocytes, and increases the production of inflammatory mediators, leading to hypertension (Zhao et al., 2004). VEGF and other inflammatory markers are greatly elevated in hypertensive patients, especially in patients with uncontrolled BP, and VEGF levels are directly related to the levels of SBP and CRP (Marek-Trzonkowska et al., 2015). VWF and TGFb-1 are also vital indicators for the evaluation of vascular endothelial injury (Ryan et al., 2003;Kido et al., 2008). This meta-analysis showed that compared with conventional Western medicines, remarkable improvements were displayed in VEGF, hs-CRP, vWF, and TGFb-1 for hypertensive patients treated with CPAD. We conducted a subgroup analysis of the above results with more than 10 studies. The antihypertensive drugs were divided into CCB, ACEI, ARB, diuretics, and combination drugs. The results showed that CHM combined with CCB and combined intervention could significantly improve 24 h-SBP, 24 h-DBP, SBP, DBP, NO, ET-1, and FMD. CHM combined with ACEI could remarkably improve SBP, DBP, NO, ET-1, but failed to reduce 24 h-SBP and 24 h-DBP. CHM combined with ARB could greatly improve SBP, DBP, NO, and ET-1, but failed to increase FMD. CHM combined with diuretics could obviously increase FMD but failed to reduce 24 h-SBP and 24 h-DBP. Moreover, the completion rate of all studies was more than 99% without severe adverse events, indicating that CHM might be an effective and safe choice for hypertension by alleviating symptoms and improving the well-being of hypertensive patients.

Strengths and Limitations
In clinical practice, antihypertensive Western medicines have a clear curative effect, but all of them have certain side effects. They cannot be completely eliminated from clinical use, and it is impossible to avoid situations where some patients may not be willing to take Western drugs and may not fully meet the needs of clinical treatment for hypertension. TCM classic herbal formulas with fixed herbs, definite curative effects, and fewer adverse effects for certain diseases have been practiced since ancient times (Xiong et al., 2017). Meanwhile, TCM classic herbal formulas have been recommended as complementary and alternative treatments in China and other countries. Thus, in Asian countries, some hypertensive patients have turned to TCM treatment with fewer adverse effects (Xiong et al., 2013). A previous meta-analysis demonstrated that a TCM adjuvant to antihypertensive drugs might be beneficial for hypertensive patients for lowering BP, improving depression, regulating blood lipids, and inhibiting inflammatory responses (Xiong et al., 2015;Chen et al., 2018;Xiong et al., 2018;Huang et al., 2019;Xiong et al., 2019). However, these studies do not focus on vascular endothelial function injury, which is one of the most common reactions of hypertension and plays a vital role in the occurrence and development of hypertension. The combination of CHM and Western medicine for essential hypertension treatment has become a trend in East Asia. At present, there is no definite index for evaluating the function of the vascular endothelium. To comprehensively evaluate the function of the vascular endothelium, we included many blood indicators to ensure that the results are comprehensive. The evaluation of the vascular endothelial function has not been included as one of the important factors affecting the cardiovascular prognosis of patients in the latest hypertension treatment guidelines.

Ren et al. Effect of CHM on Hypertension
Moreover, there are no drugs specifically for vascular endothelial injury. The analysis in this study may provide some evidence for further improvement of the treatment guidelines. After quantitative synthesis, our review is the first to demonstrate that CPAD can lower BP and improve vascular endothelial function in patients with hypertension, suggesting that TCM as an adjuvant therapy could be used for hypertension treatment by alleviating symptoms and improving the well-being.
However, there are some limitations to our review. We only conducted a search based on Chinese and English studies, and it is possible that articles related to CPAD for hypertension may have been published in other languages. Moreover, in this systematic review, we did not consider differences in the composition and dosage of CPAD and the course of medication, which may have some influence on treatment efficacy. The methodological quality of the included trials was quite low, except that CPAD was difficult to blind. The included studies had other flaws, including poor randomization and allocation concealment. According to the GRADE system, the evidence of CPAD for hypertension treatment was assessed to be of very low, low, or moderate quality. Therefore, evidence supporting CPAD in the treatment of patients with hypertension was inconclusive.

Implications for Research
In this review, CPAD may be effective and safe for the treatment of hypertension, but the methodological quality of the included studies was poor, evidence for efficacy and safety was insufficient, and clinical heterogeneity was large; therefore, great attention should be paid when interpreting current evidence and potential findings. Further research is needed to evaluate the efficacy and safety of CPAD for treating hypertension. Rigorous RCTs with large sample sizes and high-quality methodologies are required to explore the efficacy of CPAD in clinic and provide evidencebased data for promoting the use of CPAD.

CONCLUSION
To summarize, our meta-analysis suggests that compared with conventional Western medicine alone, CPAD might be effective in reducing BP levels and improving vascular endothelial function in patients with hypertension. As there are some methodological limitations to the studies included, these findings are required to be interpreted carefully. To further strengthen this evidence, new, well-designed studies with large sample sizes, strict randomization, and detailed descriptions about the detection and reporting processes are warranted.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/ supplementary material.