Stroke is the second most common cause of death in the world and in China it has now become the main cause of death. It is also a main cause of adult disability and dependency. Acupuncture for stroke has been used in China for hundreds of years and is increasingly practiced in some Western countries. This is an update of the Cochrane review originally published in 2006 .

To determine the efficacy and safety of acupuncture therapy in people with subacute and chronic stroke. We intended to test the following hypotheses: 1) acupuncture can reduce the risk of death or dependency in people with subacute and chronic stroke at the end of treatment and at follow-up; 2) acupuncture can improve neurological deficit and quality of life after treatment and at the end of follow-up; 3) acupuncture can reduce the number of people requiring institutional care; and 4) acupuncture is not associated with any intolerable adverse effects.

We searched the Cochrane Stroke Group Trials Register (June 2015), the Cochrane Central Register of Controlled Trials (CENTRAL;Cochrane Library 2015, Issue 7), MEDLINE (1966 to July 2015, Ovid), EMBASE (1980 to July 2015, Ovid), CINAHL (1982 to July 2015, EBSCO), and AMED (1985 to July 2015, Ovid). We also searched the following four Chinese medical databases: China Biological Medicine Database (July 2015); Chinese Science and Technique Journals Database (July 2015); China National Infrastructure (July 2015), and Wan Fang database (July 2015).

Truly randomised unconfounded clinical trials among people with ischaemic or haemorrhagic stroke, in the subacute or chronic stage, comparing acupuncture involving needling with placebo acupuncture, sham acupuncture, or no acupuncture.

Two review authors independently selected trials for inclusion, assessed quality, extracted and cross-checked the data.

We included 31 trials with a total of 2257 participants in the subacute or chronic stages of stroke. The methodological quality of most of the included trials was not high. The quality of evidence for the main outcomes was low or very low based on the assessment by the system of Grades of Recommendation, Assessment, Development and Evaluation (GRADE).

Two trials compared real acupuncture plus baseline treatment with sham acupuncture plus baseline treatment. There was no evidence of differences in the changes of motor function and quality of life between real acupuncture and sham acupuncture for people with stroke in the convalescent stage.

Twenty-nine trials compared acupuncture plus baseline treatment versus baseline treatment alone. Compared with no acupuncture, for people with stroke in the convalescent phase, acupuncture had beneficial effects on the improvement of dependency (activity of daily living) measured by Barthel Index (nine trials, 616 participants; mean difference (MD) 9.19, 95% confidence interval (CI) 4.34 to 14.05; GRADE very low), global neurological deficiency (seven trials, 543 participants; odds ratio (OR) 3.89, 95% CI 1.78 to 8.49; GRADE low), and specific neurological impairments including motor function measured by Fugl-Meyer Assessment (four trials, 245 participants; MD 6.16, 95% CI 4.20 to 8.11; GRADE low), cognitive function measured by the Mini-Mental State Examination (five trials, 278 participants; MD 2.54, 95% CI 0.03 to 5.05; GRADE very low), depression measured by the Hamilton Depression Scale (six trials, 552 participants; MD -2.58, 95% CI -3.28 to -1.87; GRADE very low), swallowing function measured by drinking test (two trials, 200 participants; MD -1.11, 95% CI -2.08 to -0.14; GRADE very low), and pain measured by the Visual Analogue Scale (two trials, 118 participants; MD -2.88, 95% CI -3.68 to -2.09; GRADE low). Sickness caused by acupuncture and intolerance of pain at acupoints were reported in a few participants with stroke in the acupuncture groups. No data on death, the proportion of people requiring institutional care or requiring extensive family support, and all-cause mortality were available in all included trials.

From the available evidence, acupuncture may have beneficial effects on improving dependency, global neurological deficiency, and some specific neurological impairments for people with stroke in the convalescent stage, with no obvious serious adverse events. However, most included trials were of inadequate quality and size. There is, therefore, inadequate evidence to draw any conclusions about its routine use. Rigorously designed, randomised, multi-centre, large sample trials of acupuncture for stroke are needed to further assess its effects.