Efficacy of acupuncture for lifestyle risk factors for stroke: A systematic review

Background Modifications to lifestyle risk factors for stroke may help prevent stroke events. This systematic review aimed to identify and summarise the evidence of acupuncture interventions for those people with lifestyle risk factors for stroke, including alcohol-dependence, smoking-dependence, hypertension, and obesity. Methods MEDLINE, CINAHL/EBSCO, SCOPUS, and Cochrane Database were searched from January 1996 to December 2016. Only randomised controlled trials (RCTs) with empirical research findings were included. PRISMA guidelines were followed and risk of bias was assessed via the Cochrane Collaboration risk of bias assessment tool. The systematic review reported in this paper has been registered on the PROSPERO (#CRD42017060490). Results A total of 59 RCTs (5,650 participants) examining the use of acupuncture in treating lifestyle risk factors for stroke met the inclusion criteria. The seven RCTs focusing on alcohol-dependence showed substantial heterogeneity regarding intervention details. No evidence from meta-analysis has been found regarding post-intervention or long-term effect on blood pressure control for acupuncture compared to sham intervention. Relative to sham acupuncture, individuals receiving auricular acupressure for smoking-dependence reported lower numbers of consumed cigarettes per day (two RCTs, mean difference (MD) = -2.75 cigarettes/day; 95% confidence interval (CI) = -5.33, -0.17; p = 0.04). Compared to sham acupuncture those receiving acupuncture for obesity reported lower waist circumference (five RCTs, MD = -2.79 cm; 95% CI: -4.13, -1.46; p<0.001). Overall, only few trials were considered of low risk of bias for smoking-dependence and obesity, and as such none of the significant effects in favour of acupuncture interventions were robust against potential selection, performance, and detection bias. Conclusions This review found no convincing evidence for effects of acupuncture interventions for improving lifestyle risk factors for stroke.


Introduction
Stroke is a major health issue with a significant burden upon quality of life and disability [1].The control of stroke risk factors plays a vital role in reducing the risk of new or subsequent strokes of all types [2].Three types of risk factors have been identified for stroke, including non-modifiable risk factors, medical risk factors, and lifestyle risk factors [2,3].Lifestyle risk factors for stroke-hypertension, high cholesterol, smoking-dependence, alcohol-dependence, obesity, poor diet/physical inactivity-approximately accounted for 80% of the global risk of stroke [3].Therefore, lifestyle risk factors for stroke are an ideal target for stroke prevention in comparison with other risk factors [4].A growing stroke burden throughout the world suggests contemporary stroke prevention strategies for modifiable lifestyle risk factors may be insufficient and new effective approaches are needed [5].However, the evidence for modification of lifestyle risk factors which are recommended by clinical guidelines for stroke management are not satisfactory [5,6].
Acupuncture is a traditional Chinese therapeutic intervention characterised by the insertion of fine metallic needles through the skin at specific sites (acupoints), with body and ears being the most common locations of acupoints [7].Needles may be stimulated manually or by applying electric current [8].There are various types of acupuncture treatments, such as needle acupuncture, electroacupuncture, acupressure, laser therapy, and transcutaneous electric acupoint stimulation (TEAS) [9].Acupuncture has long been used for chronic diseases including musculoskeletal pain and hypertension [7].The biological effects of acupuncture treatments, such as local inflammatory responses, anti-analgesia effects, and increase of opioid peptides, play an important role in the therapeutic effects of such therapy [10].Nevertheless, the challenges inherent in designing and implementing rigorous acupuncture research may limit the understanding of the effectiveness of acupuncture, such as those relating to acupuncturists' use of distinct syndrome classifications identified among people with the same condition and use of different skills when selecting and manipulating acupoints [11].
Using acupuncture to manage each lifestyle risk factor for stroke has attracted substantial and growing research interest over many decades.Previous reviews reported promising results of acupuncture use in controlling hypertension-associated symptoms [12], attaining weight loss [13], and reducing nicotine withdrawal symptoms [9].In addition, WHO has indicated the effect of acupuncture for alcohol-dependence, in particular auricular acupuncture [14].Nonetheless, a comprehensive systematic review assessing the effect of all forms of acupuncture for all identified lifestyle risk factors for stroke has not been conducted.As such, the aim of this paper is to identify and summarise the contemporary evidence of acupuncture interventions for lifestyle risk factors for stroke.

Methods
The systematic review reported in this paper has been registered with PROSPERO (International prospective register of systematic reviews, #CRD42017060490).

Search strategy
In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guideline, a systematic search of the literature was conducted using the MEDLINE, CINAHL/EBSCO, Scopus, and Cochrane Database of Systematic Reviews databases for studies published from January 1996 to December 2016.The lifestyle risk factors for stroke included in this systematic review are high blood pressure (hypertension & prehypertension), high cholesterol, obesity (overweight/obesity), smoking-dependence, alcohol-dependence, and physical inactivity.The literature search employed keyword and MeSH searches for terms relevant to 'acupuncture' and each lifestyle risk factor for stroke.Search terms used for each database are available in Table 1.Relevant randomised controlled trials (RCT) listed as references of published systematic review papers on selected lifestyle risk factors for stroke were also searched via Google Scholar by title, in order to include all relevant RCTs in this field.

Types of interventions.
There was no limitation on the forms of (traditional) acupuncture and the frequency and duration of the intervention.However, contemporary acupuncture such as trigger points and dry needling was not eligible for inclusion in this review.
Types of outcome measures.Only anthropometric parameters and the widely used indicators of each lifestyle risk factor for stroke were included.The primary outcomes were a change in systolic blood pressure (SBP) and/or diastolic blood pressure (DBP) for hypertension-focused RCTs; triglycerides, LDL/HDL cholesterol for hyperlipidemia/dyslipidemiafocused RCTs; body weight (BW), body mass index (BMI), waist circumference (WC) for obesity-focused RCTs; alcohol craving, completion rate of treatment, withdrawal symptoms for RCTs focusing on alcohol-dependence; withdrawal symptoms, daily cigarette consumption, abstinence rate for RCTs focusing on smoking-dependence; physical activity minutes/day and cardiorespiratory fitness for physical inactivity-focused RCTs.

Data extraction
Title and abstracts of all citations identified in the search were imported to Endnote (Version X8) and duplicates removed.These citations were independently reviewed for eligibility by two authors (WP and RL) and the full texts of ambiguous articles were retrieved if consensus was not reached.Any disagreements were assessed by a third author.We contacted authors regarding raw data of their RCTs where necessary for meta-analysis.Where we failed to obtain such raw data, the RCT had to be excluded in the meta-analysis.According to the RCT description in the articles included, raw data were extracted from post-intervention effect and/ or follow-up (long-term) effect.
Data were extracted into a pre-determined table (Table 2) and checked for coverage and accuracy by two authors independently.Table 2 includes detailed information on sample size, inclusion criteria, participants' characteristics, intervention groups, add-on strategy, results of outcome measures, and side-effects.Both statistically significant within-group and/or between-group effect of acupuncture interventions for each lifestyle risk factor for stroke were recorded if reported.

Data syntheses
Cochrane RevMan version 5.3 software was employed to conduct meta-analysis of the outcome measures and heterogeneity was determined using I 2 statistic [15].The meta-analysis included all studies where acupuncture was employed with or without co-interventions, provided that such intervention was given to all groups.However, meta-analyses were conducted only if at least two RCTs were available exploring a specific outcome of a risk factor.Acupuncture approaches shown in the meta-analysis include needle acupuncture (body, aural region, electroacupuncture), laser acupuncture, and acupressure.Analyses were performed separately for type of experimental interventions (acupuncture, acupressure, laser acupuncture, or the combination of acupuncture and acupressure) according to the RCT design.Random effects model (Mantel-Haenszel for dichotomous/categorical variables and inverse variance for continuous variables) was used to calculate mean differences (MD), standardized mean differences (SMD), or risk ratios (RR), and 95% confidence intervals (CI) were reported.Sensitivity analyses were used to test the robustness of statistically significant results for RCTs with low risk versus high risk of bias for the domains selection bias and performance/detection bias.Effects sizes of acupuncture compared to other interventions were shown in Table 3.   ; male a Add-on strategy of all the intervention groups.

Quality assessment
Two authors (DS and WP) independently assessed the risk of bias of all included studies using the Cochrane Risk of Bias Tool for selection bias (random sequence generation and allocation concealment), performance bias (blinding of participants and personnel), detection bias (blinding of outcome assessment), attrition bias (incomplete outcome data), reporting bias (selective outcome reporting), and other bias (Table 4).Disagreements were assessed by a third author.It is worth noting that, due to methodological reasons and the uniqueness of acupuncture treatments, it is not feasible to blind the acupuncturist in acupuncture RCTs.Therefore, we adopted the domain of performance bias and only focused on adequate participant blinding.

Results
The key database searches identified 2,502 records with another six records from Google Scholar search, of which 299 duplicates were removed.After screening, the full texts of 305 papers were reviewed, of which a total of 62 full-text articles (reporting on 59 RCTs) were considered eligible and included in this systematic review.The PRISMA flowchart of literature search and article selection details has been shown in Fig 1.
There were 59 RCTs (5,650 participants) regarding the use of acupuncture interventions in treating lifestyle risk factors for stroke, of which 7 RCTs for alcohol-dependence (845 participants), 15 RCTs for smoking-dependence (1,960 participants), 12 RCTs for hypertension (927 participants), and 25 RCTs for obesity (1,918 participants).No publication reported on a trial examining the efficacy of acupuncture for the lifestyle risk factor for stroke of high cholesterol or physical inactivity as a primary outcome.

Alcohol-dependence
Seven RCTs [16][17][18][19][20][21][22] focused on acupuncture treatments for alcohol-dependence using outcomes of alcohol craving (four RCTs), alcohol withdrawal symptoms (four RCTs), and drinking days (one RCT).Table 2 shows details of such RCTs' characteristics and safety-related information.Most of the included studies defined alcohol-dependence according to the 3 rd version (revised)/4 th version of the Diagnostic and Statistical Manual of Mental Disorders (DSM) or the 10 th version of the International Statistical Classification of Diseases and Related Health Problems (ICD) [16][17][18][19][20][21].The sample size of RCTs focusing on alcohol-dependence ranged from 20 to 503 participants with only two studies recruiting more than 100 participants.
Psychiatrists/nurses [17,20], acupuncturists [18,22], and oriental medical doctors [16] were reported as administering the acupuncture interventions.The modes of acupuncture delivered within the interventions included both specific and nonspecific/symptom-based auricular acupuncture (five studies), body acupuncture (one study), and combined auricular and body acupuncture (one study).Acupuncture treatment sessions ranged from 30-minutes to 45-minutes.Only one RCT employed needle stimulation technique for the acupuncture treatment of alcohol-dependence [17].
Non-significant differences between acupuncture and control groups for alcohol craving were reported in three RCTs [16,17,20], alcohol withdrawal symptoms in two RCTs [17,18], and drinking days in one RCT [20].Statistically significant within-intervention group effects were reported for alcohol craving with specific auricular electroacupuncture [21] and alcohol withdrawal symptoms with combined use of auricular and body acupuncture [19], while statistically significant between-group effects were reported for alcohol withdrawal symptoms with symptom-based auricular acupuncture (VS specific auricular acupuncture) [22].Risk of bias assessment indicated that three RCTs did not report information on random sequence generation, four RCTs failed to apply blinding to participants and personnel, one did not report adequate blinding of outcome assessors, and three failed to report complete outcome data (Table 4).Due to the great heterogeneity regarding intervention details and outcomes applied in the RCTs focusing on alcohol-dependence, no meta-analysis could be conducted.
Both statistically significant within-intervention group and between-group effects were reported in five RCTs for (a) SBP as well as DBP levels with body acupuncture (VS non-specific acupuncture) [48], combined body acupuncture and exercise (VS sham acupuncture plus exercise) [44], combined laser body acupuncture with/without music treatment (VS starch tablets) [47], body acupressure (VS sham acupuncture) [50], (b) nighttime DBP level with body acupuncture (VS sham acupuncture) [40].In addition, study results reported statistically significant within-intervention group effects for (a) SBP as well as DBP levels with laser acupuncture [41], (b) SBP level with body electroacupuncture [42], (c) DBP level with combined body and auricular acupuncture [49], and statistically significant between-group effect for SBP level with body electroacupuncture (VS sham acupuncture) [43].
Meta-analyses did not show evidence for neither post-intervention nor long-term effect of acupuncture interventions on SBP control (two RCTs on acupuncture, MD = -0.54mmHg; 95%CI: -10.69, 9.60; p = 0.92) and DBP control (two RCTs on acupuncture, MD = -1.38 mmHg; 95%CI: -4.06, 1.31; p = 0.32) compared to sham acupuncture (Table 3).Risk of bias assessment indicated only six hypertension-focused RCTs blinded participants and personnel appropriately and seven RCTs did not report information on blinding of outcome assessment (Table 4).
Relative to sham acupuncture, meta-analyses only found those receiving acupuncture interventions for obesity reported lower waist circumference (five RCTs, MD = -2.79cm; 95% CI: -4.13, -1.46; p<0.001; heterogeneity: I 2 = 0%; Chi 2 = 1.61; p = 0.81).However, after excluding RCTs with other than low risks of selection and performance/detection bias, none of the effect remained statistically significant.In comparison with no treatment intervention, meta-analyses did not show evidence for post-intervention effect of acupuncture interventions on BW (two RCTs on acupuncture, MD = -1.12kg; 95%CI: -5.51, 3.27; p = 0.62; two RCTs on auricular acupressure, MD = -2.87Kg; 95%CI: -6.47, 0.74; p = 0.12).Meta-analyses also did not show evidence for post-intervention effect of auricular acupressure interventions on BMI (two RCTs, MD = -0.41kg/m 2 ; 95%CI: -1.56, 0.73; p = 0.48) compared to no treatment (Table 3).Risk of bias assessment was unclear in numerous obesity-focused RCTs due to a lack of detail in the publications.Specifically, nine RCTs did not report random sequence generation and allocation concealment information.Twelve RCTs failed to report complete outcome data.Fifteen RCTs did not blind participants and personnel and 20 RCTs did not provide information on blinding of outcome assessment (Table 4).

Discussion
This article reports the first systematic review of the effect of acupuncture interventions for lifestyle risk factors for stroke.A number of acupuncture techniques have been used for the management of these lifestyle risk factors and have yielded limited improvements in outcomes.No analysis can be conducted on RCTs focusing on alcohol-dependence and no evidence of the effect of acupuncture treatments on high blood pressure was shown based on meta-analysis.The meta-analysis showed individuals receiving auricular acupressure reported better outcomes in daily cigarette consumption than sham acupressure.Furthermore, acupuncture users have reported better outcomes in reducing waist circumference compared to sham acupuncture.No serious side effects occurred when using acupuncture on these four lifestyle risk factors.However, approximately half of the RCTs focusing on hypertension and obesity did not report safety information of acupuncture users.As such, acupuncture appears to be a relative safe treatment for the management of lifestyle risk factors for stroke.Some evidence of the benefits of acupuncture and/or auricular acupressure was revealed for RCTs of lifestyle risk factors for stroke-smoking-dependence and obesity-in our review.However, a total of eight and 14 types of acupuncture-related interventions have been examined in RCTs focusing on smoking-dependence and obesity, respectively.The findings reported here highlighted the gaps in the evidence of clinical acupuncture use in the specific field of lifestyle risk factors for stroke and generally.Consistent with findings of prior systematic reviews [9,78], acupuncture involves a range of techniques.Both acupuncture-associated clinical trials and observational studies are required to determine methodology issues such as the use of acupuncture only, acupressure only, or the combination of acupuncture and acupressure, and the further choices of acupuncture like needle acupuncture, electroacupuncture and laser acupuncture.Therefore, future high-quality research is warranted to confirm our preliminary findings and provide robust effect estimates of acupuncture interventions for lifestyle risk factors for stroke.
In our review, approximately half of the RCTs focusing on smoking-dependence and obesity employed auricular acupressure alone or in combination with other acupuncture intervention(s).Acupressure is considered more practical (ease of application by patients themselves) with low cost, compared to other acupuncture treatments [79].However, no consistent and convincing evidence has been found in this review on whether acupressure is effective for the management of overall lifestyle risk factors for stroke.As a result, there is insufficient evidence to conclude that the use of acupressure could improve the lifestyle risk factors for stroke and more studies are required.
Sham acupuncture is the most frequently employed comparison for acupuncture treatments in general [80] and among people with lifestyle risk factors for stroke which has been shown in our review.Although meta-analysis presented here reported statistically significant benefits of real acupuncture interventions regarding the management of the lifestyle risk factors of smoking-dependence and obesity than sham interventions, none of the effects of the RCTs included in the analyses was robust against potential selection, performance, and detection bias.In addition to the identified design challenges of acupuncture-associated RCTs regarding the choice of control group with the fact that sham acupuncture may also trigger physiological effect [81], future acupuncture-associated RCTs should avoid high risk of bias from lack of allocation concealment and missing outcome data, persuade original investigators to provide sufficient information on blinding of outcome ascertainment and if necessary, choose an appropriate comparable control intervention for clinical acupuncture research.Some limitations of our systematic review are worth noting.The acupuncture interventions varied greatly across the RCTs of each lifestyle risk factor for stroke included in this review in terms of inclusion criteria of participants, acupuncture forms, acupoint selection, manipulation methods, and frequency/duration of the treatments.Also, this systematic review was restricted to RCTs published in English-language peer-reviewed journals.Furthermore, a proportion of included studies were not registered before they were published, we therefore cannot rule out the possibility of reporting or publication bias.The findings in this systematic review regarding the effect of acupuncture for lifestyle risk factors for stroke should be interpreted with caution.However, compared to previous Cochrane and systematic reviews [9,12,13,82], based on the risk of bias evaluation (Table 4), the methodological quality of RCTs on acupuncture treatments identified in our review has improved over recent years, including regards to random sequence generation application, the reporting of acupuncture treatments, and use of long-term follow-ups.

Conclusion
This review shows no convincing evidence regarding the effect of acupuncture, acupressure, laser acupuncture or their combination use for lifestyle risk factors for stroke.However, the translation of findings of this systematic review may contribute to the evidence-base of potential clinical practice guideline recommendations for stroke prevention.