Background

Sensory stimulation via acupuncture has been reported to alter activities of numerous neural systems by activating multiple efferent pathways. Acupuncture, one of the main physical therapies in Traditional Chinese Medicine, has been widely used to treat patients with stroke for over hundreds of years. This is the first update of the Cochrane Review originally published in 2005.

Objectives

To assess whether acupuncture could reduce the proportion of people with death or dependency, while improving quality of life, after acute ischemic or hemorrhagic stroke.

Search methods

We searched the Cochrane Stroke Group trials register (last searched on February 2, 2017), the Cochrane Central Register of Controlled Trials Ovid (CENTRAL Ovid; 2017, Issue 2) in the Cochrane Library, MEDLINE Ovid (1946 to February 2017), Embase Ovid (1974 to February 2017), the Cumulative Index to Nursing and Allied Health Literature (CINAHL) EBSCO (1982 to February 2017), the Allied and Complementary Medicine Database (AMED; 1985 to February 2017), China Academic Journal Network Publishing Database (1998 to February 2017), and the VIP database (VIP Chinese Science Journal Evaluation Reports; 1989 to February 2017). We also identified relevant trials in the Chinese Clinical Trial Registry (last searched on Feburuary 20, 2017), the World Health Organization (WHO) International Clinical Trials Registry Platform (last searched on April 30, 2017), and Clinicaltrials.gov (last searched on April 30, 2017). In addition, we handsearched the reference lists of systematic reviews and relevant clinical trials.

Selection criteria

We sought randomized clinical trials (RCTs) of acupuncture started within 30 days from stroke onset compared with placebo or sham acupuncture or open control (no placebo) in people with acute ischemic or hemorrhagic stroke, or both. Needling into the skin was required for acupuncture. Comparisons were made versus (1) all controls (open control or sham acupuncture), and (2) sham acupuncture controls.

Data collection and analysis

Two review authors applied the inclusion criteria, assessed trial quality and risk of bias, and extracted data independently. We contacted study authors to ask for missing data. We assessed the quality of the evidence by using the GRADE approach. We defined the primary outcome as death or dependency at the end of follow‐up .

Main results

We included in this updated review 33 RCTs with 3946 participants. Twenty new trials with 2780 participants had been completed since the previous review. Outcome data were available for up to 22 trials (2865 participants) that compared acupuncture with any control (open control or sham acupuncture) but for only six trials (668 participants) that compared acupuncture with sham acupuncture control. We downgraded the evidence to low or very low quality because of risk of bias in included studies, inconsistency in the acupuncture intervention and outcome measures, and imprecision in effect estimates.

When compared with any control (11 trials with 1582 participants), findings of lower odds of death or dependency at the end of follow‐up and over the long term (≥ three months) in the acupuncture group were uncertain (odds ratio [OR] 0.61, 95% confidence interval [CI] 0.46 to 0.79; very low‐quality evidence; and OR 0.67, 95% CI 0.53 to 0.85; eight trials with 1436 participants; very low‐quality evidence, respectively) and were not confirmed by trials comparing acupuncture with sham acupuncture (OR 0.71, 95% CI 0.43 to 1.18; low‐quality evidence; and OR 0.67, 95% CI 0.40 to 1.12; low‐quality evidence, respectively).

In trials comparing acupuncture with any control, findings that acupuncture was associated with increases in the global neurological deficit score and in the motor function score were uncertain (standardized mean difference [SMD] 0.84, 95% CI 0.36 to 1.32; 12 trials with 1086 participants; very low‐quality evidence; and SMD 1.08, 95% CI 0.45 to 1.71; 11 trials with 895 participants; very low‐quality evidence). These findings were not confirmed in trials comparing acupuncture with sham acupuncture (SMD 0.01, 95% CI ‐0.55 to 0.57; low‐quality evidence; and SMD 0.10, 95% CI ‐0.38 to 0.17; low‐quality evidence, respectively).

Trials comparing acupuncture with any control have reported little or no difference in death or institutional care at the end of follow‐up (OR 0.78, 95% CI 0.54 to 1.12; five trials with 1120 participants; low‐quality evidence), death within the first two weeks (OR 0.91, 95% CI 0.33 to 2.55; 18 trials with 1612 participants; low‐quality evidence), or death at the end of follow‐up (OR 1.08, 95% CI 0.74 to 1.58; 22 trials with 2865 participants; low‐quality evidence).

The incidence of adverse events (eg, pain, dizziness, faint) in the acupuncture arms of open and sham control trials was 6.2% (64/1037 participants), and 1.4% of these (14/1037 participants) discontinued acupuncture. When acupuncture was compared with sham acupuncture, findings for adverse events were uncertain (OR 0.58, 95% CI 0.29 to 1.16; five trials with 576 participants; low‐quality evidence).

Authors' conclusions

This updated review indicates that apparently improved outcomes with acupuncture in acute stroke are confounded by the risk of bias related to use of open controls. Adverse events related to acupuncture were reported to be minor and usually did not result in stopping treatment. Future studies are needed to confirm or refute any effects of acupuncture in acute stroke. Trials should clearly report the method of randomization, concealment of allocation, and whether blinding of participants, personnel, and outcome assessors was achieved, while paying close attention to the effects of acupuncture on long‐term functional outcomes.