Acupuncture Efficacy on Ischemic Stroke Recovery
Multicenter Randomized Controlled Trial in China
Background and Purpose—Acupuncture is a frequently used complementary treatment for ischemic stroke in China but the evidence available from previous randomized trials is inconclusive. The objective of this study was to assess the efficacy and safety of acupuncture in a more robustly designed larger scale trial.
Methods—This is a multicenter, single-blinded, randomized controlled trial. Eight hundred sixty-two hospitalized patients with limb paralysis between 3 to 10 days after ischemic stroke onset were allocated acupuncture plus standard care or standard care alone. The acupuncture was applied 5 times per week for 3 to 4 weeks. The primary outcomes were defined as follows: (1) death/disability according to Barthel index and (2) death/institutional care at 6 months.
Results—There was a tendency of fewer patients being dead or dependent in acupuncture group (80/385, 20.7%) than in control group (102/396, 25.8%) at 6 months (odds ratio, 0.75; 95% confidence interval, 0.54–1.05). The benefit was noted in subgroup receiving ≥10 sessions of acupuncture (odds ratio, 0.68; 95% confidence interval, 0.47–0.98). There was no statistical difference in death or institutional care between the 2 groups (odds ratio, 1.06; 95% confidence interval, 0.63–1.79). Severe adverse events occurred in 7.6% and 8.3% of patients in the 2 groups, respectively.
Conclusions—Acupuncture seemed to be safe in the subacute phase of ischemic stroke. If the potential benefits observed are confirmed in future larger study, the health gain from wider use of the treatment could be substantial.
Clinical Trial Registration—URL: http://www.chictr.org/en/. Unique identifier: ChiCTR-TRC-11001353.
Worldwide, stroke is the second common cause of death and first leading cause of adult disability.1 In China, there are 1.5 to 2 million new strokes per year and stroke has become the first leading cause of death and disabilty.2,3 The proportion of patients being dependent in daily life is still high at discharge even in those with intensive treatment including rehabilitation.4 In Western medicine, treatment added in standard rehabilitation in the early postacute phase of stroke that might promote the process of recovery such as selective serotonin reuptake inhibitors and constraint-induced movement therapy shows great promise.5,6 Acupuncture which has been used for treating stroke for >1000 years might act in a similar way. Experimental studies have showed its potential beneficial effects in post stroke rehabilitation.7–10 However, randomized controlled trials have yet to show convincing evidence of benefit in acute and subacute ischemic stroke.11–13 In the earlier trials conducted outside China, 2 studies showed that acupuncture can improve motor function and independence in activities of daily living (ADL),14,15 whereas 2 trials using sham treatment as control did not find evidence of benefit.16,17 Although, by contrast, many Chinese trials reported positive results, the conclusions were not regarded as robust, as a result of small sample size and methodological weaknesses as follows: the use of outcome measures that were not internationally recognized; unclear methods of randomization and allocation concealment; the lack of long-term follow-up; and publication bias.12,13,18 Therefore, we designed this randomized controlled trials with a larger sample, and scales to assess efficacy and safety of acupuncture. The hypothesis was acupuncture, in the immediate postacute phase of ischemic stroke, can improve patients’ independence in ADL over a longer follow-up.
Study Design and Participants
This study was designed as a multicenter, single-blinded, randomized controlled trial. The study was divided into 2 phases: pilot study from April 2001 to February 2002 and main study from February 2002 to February 2004. The study was registered in Chinese Clinical Trial Register (ChiCTR-TRC-11001353) with protocol approved by the ethics committee of West China Hospital of Sichuan University.
Patients aged from 18 to 85 years and both sexes were eligible if they had an acute ischemic stroke and could be enrolled between 3 to 10 days after stroke onset. The ischemic stroke was diagnosed according to the World Health Organization criteria for stroke and all patients had brain computed tomographic or magnetic resonance imaging scan to exclude intracerebral hemorrhage.19 Only patients with limbs paralysis were included. The extent of paralysis had to be such that patients should be unable to walk or eat or dress without assistance.
The criteria included the following: (1) dependency in ADL before present stroke, (2) inability to complete scheduled treatment course, (3) infection in acupuncture sites, (4) severe aphasia or unconsciousness making it difficult to comprehend the study or cooperate in evaluation, and (5) other severe complications or comorbidities such as heart/renal function failure.
The sample size calculation was based on assumption of a ≥10% reduction in the absolute risk of death or dependency in the acupuncture group compared with controls (30%) at 6-month follow-up. Considering possible loss of follow-up (10% to 20%), ≥800 patients (400 patients in each group) were required to test the hypothesis (α=0.05; β=0.80).
Randomization and Allocation Concealment
The list of treatment allocations were produced centrally by computer-generated random numbers and the randomization sequence was stratified by participating centres. The article revealing the treatment allocation was placed in sequentially numbered, sealed opaque envelopes. After informed consent was obtained, the envelopes were opened sequentially and acupuncture therapists were notified if participants were allocated into acupuncture group.
Patients in both groups received routine ischemic stroke treatment, such as antithrombotic medications and prevention of complications. Rehabilitation was also performed in centres where it was available. Background medical care could differ from center to center but the protocol required all trial patients in a given center to receive the same background care irrespective of treatment allocation.
Patients in the treatment group received additional traditional Chinese acupuncture which was started 3 to 10 days after stroke onset, 5 times per week for 3 to 4 weeks (in the pilot study) or 3 weeks (in the main study). Before the trial started, all acupuncturists in participating centres received training of standardized acupuncture procedure. The acupoints with corresponding needling methods used in the present study were selected on the basis of “Xing Nao Kai Qiao” acupuncture methods which was originally developed by the acupuncture research group in the Medical College of Tianjing.20 The main acupoints include Renzhong-DU26, Neiguan-PC6 (both sides), and Sanyinjiao-SP6 (paretic side). Auxiliary acupoints include Baihui-DU20, 4 acupoints on the paretic side (Zusanli-ST36, Fenglong-ST40, Taichong-LK3, and Chize-LL5), and 2 acupoints on the both sides (Fengchi-GB20, Qihai-RN6). The Huato sterile acupuncture needles (1.5–4 cm in length and 0.38–0.42 mm in diameter) were used in the present study. Patients were in the supine position when they were treated. The needling methods for each acupoint have been published in the article of pilot study in the online-only Data Supplement.21
Demographic and clinical characteristics were recorded at the time of enrollment. During treatment period, adverse events which may be related to acupuncture were monitored and reported to the principal investigator who would decide whether acupuncture should be stopped. All patients who did not adhere to the allocated treatment were followed up in full and their data were included into analysis according to intention-to-treat principle.
The severity of neurological deficits at baseline and at the end of scheduled treatment periods or at discharge was assessed with Scandinavian Stroke Scale. This scale includes 10 items with total of score ranging from 0 to 60 points (higher points indicate better neurological function). At the end of 3 and 6 months after randomization, patients were followed up by telephone interview. ADL were evaluated by using Barthel Index (BI) which involves 10 personal activities with a maximum of 100 points (higher BI score indicates better function in ADL). The evaluation of both neurological deficits and ADL were performed by investigators who were unaware of treatment allocation.
Primary outcomes in this trial were defined as the following: (1) death/disability at 6 months: disability was defined as BI of ≤60 points; (2) death/long-term institutional care at 6 months (in the hospital, nursing home, or rehabilitation house). As sample size/power calculation may vary when it is based on different primary outcomes, sample size calculation based on outcome “death/disability at 6 months” was preset in this study. Secondary outcomes include the following: (1) case fatality; (2) changes of neurological deficit score at the end of scheduled treatment periods or at discharge. Severe adverse events occurred in both groups and adverse events directly related to acupuncture were analyzed to assess acupuncture safety.
Statistical analyses were performed by using SPSS for Windows statistic software (version 13.0). χ2 test for categorical variables, t test for continuous variables with normal distribution, and nonparametric test for skewed distribution were used to detect difference in baseline characteristics between the 2 groups. For the outcome analysis, odds ratio [OR] with corresponding 95% confidence interval [CI] were calculated for dichotomous data including death, death or disability, and death or long-term institutional care, whereas t test was used to assess difference in changes of neurological deficits score between the 2 groups. The level of significance was set at 5% in the comparison, and all statistical testing was 2-sided. Subgroup analysis was preset to compare patients receiving ≥10 sessions of acupuncture with control group.
Total of 862 patients from 40 hospitals were enrolled in this study; of them 427 were allocated acupuncture and 435 control (Figure 1). There was no significant difference in baseline characteristics between the 2 groups except more patients with dysphagia and dysarthria in control group than in acupuncture group (Table 1). Two hundred sixty-seven (62.5%) patients in acupuncture group completed all scheduled sessions of acupuncture (≥15 sessions), whereas 54 (12.6%) received 10 to 14 sessions, 59 (13.8%) received 5 to 9 sessions, 24 (5.6%) received 1 to 4 sessions, and 23 (5.4%) did not receive any acupuncture treatment after randomization. The main cause of discontinuing acupuncture was early discharge from hospital (112 patients, 82.4%). The concomitant therapies were comparable between the 2 groups (Table I in the online-only Data Supplement).
At 6 months, 40 (9.4%) patients in acupuncture group and 39 (9.0%) patients in control group were lost to follow-up. Fewer patients had died or were dependent in ADL in the acupuncture group (113/410, 27.6%) than in control group (137/405, 33%) at 3 months, the tendency continued at 6 months (80/385 [20.7%] versus 102/396 [25.8%]), but the difference was not statistically significant (OR, 0.75; 95% CI, 0.54–1.05; Table 2). There was also no significant difference in death or institutional care between the 2 groups at 6 months (8.0% versus 7.6%; OR, 1.06; 95% CI, 0.63–1.79). The results above did not change significantly when last observation carried forward analysis was conducted.
Post hoc ordinal analysis was performed to assess if there was favorable shift in distribution of scores on BI with acupuncture treatment, and a nonsignificant trend was tested (pooled OR for shift to higher BI, 1.15; 95% CI, 0.88–1.50; P=0.296; Figure 2). Preplanned subgroup analysis showed that only when patients who received ≥10 sessions of acupuncture were included into analysis, a significant difference was detected in death or dependency at 6 months (57/298 [19.1%] versus 102/396 [25.8%]; OR, 0.68; 95% CI, 0.47–0.98; number needed to treat 15) but not for other outcomes (Table II in the online-only Data Supplement).
Death occurred in 6.2% and 4.5% of patients in the 2 groups at 6 months, respectively (OR, 1.39%; 95% CI, 0.74–2.60). At the end of the scheduled treatment period, there was significantly greater change of Scandinavian Stroke Scale score from baseline in acupuncture group (9.70±7.85) than in control group (7.57±12.58; P=0.03), which indicated that patients in acupuncture group had greater reduction in their neurological deficit.
Severe adverse events occurred in 7.6% of patients in acupuncture group and 8.3% in control group. The most common adverse event directly related to acupuncture is pain, moderate or severe pain occurred in ≈9.4% patients; 1.9% of patients discontinued acupuncture because of severe pain (Table 3).
In this study, we assessed efficacies and safety of acupuncture used in subacute stage of stroke; acupuncture seemed to be a safe treatment but did not show definitive benefits of reducing death or dependency in ADL, although a trend of better outcomes was observed in acupuncture group at 3 and 6 months. Also, acupuncture did not prove to be efficacious in reducing death or long-term institutional care. However, it was noted that there was more improvement of neurological deficits in whole acupuncture group and significant reduction of death or dependency in subgroup receiving ≥10 sessions of acupuncture (OR, 0.68; 95% CI, 0.47–0.98).
The effect of acupuncture on death or disability observed in this study is consistent with 3 recently published studies with long-term follow-up and 2 systematic reviews of acupuncture in post stroke rehabilitation.13,22–25 The recent systematic review included 56 randomized trials published before 2009; of the evaluated, 46 used acupuncture within 1-month post stroke and only 10 took functional outcomes; pooled OR of 38 trials was in favor of acupuncture (OR, 4.33; 95% CI, 3.09–6.08).25 Of the 56 included trials, 32 trials were published after 2004, most of them (27 trials) were conducted in China. None of the trials conducted outside China reported positive results on primary outcome, whereas all Chinese trials except 1 showed positive results in favor of acupuncture. In the Cochrane systematic review focusing on efficacy of acupuncture in acute and subacute phases of stroke, meta-analysis of 4 trials with a total of 373 patients showed that there was a borderline significant trend toward fewer patients being dead or dependent (OR, 0.66; 95% CI, 0.43–0.99) in acupuncture group than in control group.13 The meta-analysis showed more aspiring result when data of this study was merged (OR, 0.72; 95% CI, 0.55–0.93; Figure 3). It is suggested that acupuncture might be efficacious in reducing disability but whether the effect is any moderate or less, sample size of this trial is still under power to detect it. In fact, except recombinant tissue-type plasminogen activator thrombolysis within 3 hours after stroke onset, there are few interventions in acute/subacute stage of stroke which showed ≥10% of absolute risk reduction of death or dependency. To clarify this issue, a larger randomized controlled trial with sample size of >1340, according to the result of updated meta-analysis, is needed. If the potential benefits above are confirmed, the health gain from wider use of the acupuncture after stroke could be substantial.26
This study has strengths. First, it is the largest randomized controlled trial of acupuncture for stroke at the moment. Second, when compared with most published Chinese acupuncture trials, this study was a real randomized trial with adequate allocation concealment, long-term follow-up and blinding assessment. Third, we used disability in ADL or death as primary outcome which have been preferred in trial of acute stroke but just adopted in a few previous trials of acupuncture for stroke. We also evaluated the efficacy of acupuncture on death or long-term institutional care. The use of coprimary outcomes may increase the chance of positive results in favor of the evaluated treatment. Therefore, adjustment of P value for significance (eg, P<0.025 rather than 0.05) should be considered. In this trial, the adjustment of P value for significance has no real impact on the trial’s conclusion because the P values calculated for the 2 primary outcomes were >0.05. Finally, “Xing Nao Kai Qiao” needling method used in this trial is one of the most commonly used acupuncture program in China. Before this trial began, an uncontrolled observational study of 2336 patients with stroke suggested that it was potentially efficacious in improving neurological deficits.27 Thus, we actually assessed efficacy of acupuncture program which is widely accepted and most likely proved to be efficacious.
There were 3 limitations in this study. The first limitation is poor compliance of patients with allocated treatment, only 62.5% of patients in acupuncture group received all scheduled sessions of acupuncture, and 6% of patients in control group withdrew and required acupuncture treatment after randomization. The reason for this may be related to the Chinese culture background: most Chinese patients think that acupuncture is useful in improving neurological deficits, so they would request for the therapy if they did not have satisfactory recovery from routine treatment. The second limitation is a little high proportion of loss of follow-up (9.4% in acupuncture group and 9.0% control group). The 2 limitations above made the results more uncertain although sensitivity analysis excluding patient who crossed over to the other group, and analysis by last observation carried forward did not show significant change in the results on primary outcomes. Considering the result of subgroup analysis and main cause of discontinuing acupuncture (early discharge from hospital), 2 weeks for acupuncture treatment period may be an appropriate choice in future trials, especially in China. The third limitation is that reasons for noninclusion of screened patients were recorded only in a proportion of participating centers, which made it difficult to assess generalizability of the trial’s results. However, data from center with the largest number of recruited patients (West China Hospital of Sichuan University) indicated that ≈45% of screened patients were excluded for ineligibility and 40% were not included because of unavailable informed consent.
Although sham acupuncture used is the main controversy in design of acupuncture trial, it seems to be reasonable to use sham acupuncture as control to evaluate real or specific effect of acupuncture. However, sham acupunctures which have been used in some previous studies are not feasible in China because patients are well acquainted with acupuncture and recognize the special sensation of “Teh Chi.” Furthermore, sham acupuncture may not be truly inert and may have underlying physiological effects, and hence may not be a proper control group.28,29 To enable us to design a large pragmatic trial, we decided not to use sham acupuncture as control in this study. In contrast to previous Chinese trial in stroke, we used binary outcome as primary outcome (dependency or independency according to BI score) because the systematic review suggested that placebos had no significant effects on objective or binary outcomes in clinical trials.30
In conclusion, this study assessed efficacy of acupuncture used as adjunct to standard care in the early stage of stroke, the result does not support its use in the routine clinical practice from Western medicine perspective. However, a trend toward fewer patients with dependency in acupuncture group was noted, especially in those receiving >10 sessions of acupuncture. Trials with larger scale are needed to confirm the potential benefit of acupuncture for post stroke rehabilitation in future.
We would like to thank all doctors and patients who participated in this study for their cooperation. We also would like to express our gratitude to Professor Peter Sandercock of Western General Hospital, Edinburgh, United Kingdom, and Professor KS Wong of Prince of Wales Hospital, Hongkong for useful comments on the article of this study.
Sources of Funding
This study was funded by the State Administration of Traditional Chinese Medicine of the People’s Republic China.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.007659/-/DC1.
- Received October 4, 2014.
- Revision received March 18, 2015.
- Accepted March 18, 2015.
- © 2015 American Heart Association, Inc.
- Chen Z
- Chen Z.
- Jia HK,
- Zhang MJ,
- Guo YL.
- Sze FK,
- Wong E,
- Or KK,
- Lau J,
- Woo J.
- Gosman-Hedstrom G,
- Claessson L,
- Klingenstierna U,
- Carlsson J,
- Olausson B,
- Frizell M,
- et al
- Johansson BB,
- Haker E,
- von Arbin M,
- Britton M,
- Långström G,
- Terént A,
- et al
- Shi XM
- Shi XM.
- Zeng XR,
- Liu M,
- Wu B,
- Zhang SH,
- Rao P,
- Wu RF,
- et al
- Sze FK,
- Wong E,
- Yi X,
- Woo J.
- Wu P,
- Mills E,
- Moher D,
- Seely D.
- Zhang S,
- Li N,
- Liu M.
- Shi XM.
- Bao T,
- Cai L,
- Snyder C,
- Betts K,
- Tarpinian K,
- Gould J,
- et al