OBJECTIVETo assess the methodological quality of systematic reviews and meta-analyses regarding acupuncture intervention for stroke and the primary studies within them.

METHODSTwo researchers searched PubMed, Cumulative index to Nursing and Allied Health Literature, Embase, ISI Web of Knowledge, Cochrane, Allied and Complementary Medicine, Ovid Medline, Chinese Biomedical Literature Database, China National Knowledge Infrastructure, Wanfang and Traditional Chinese Medical Database to identify systematic reviews and meta-analyses about acupuncture for stroke published from the inception to December 2016. Review characteristics and the criteria for assessing the primary studies within reviews were extracted. The methodological quality of the reviews was assessed using adapted Oxman and Guyatt Scale. The methodological quality of primary studies was also assessed.

RESULTSThirty-two eligible reviews were identified, 15 in English and 17 in Chinese. The English reviews were scored higher than the Chinese reviews (P=0.025), especially in criteria for avoiding bias and the scope of search. All reviews used the quality criteria to evaluate the methodological quality of primary studies, but some criteria were not comprehensive. The primary studies, in particular the Chinese reviews, had problems with randomization, allocation concealment, blinding, dropouts and withdrawals, intent-to-treat analysis and adverse events.

CONCLUSIONSImportant methodological flaws were found in Chinese systematic reviews and primary studies. It was necessary to improve the methodological quality and reporting quality of both the systematic reviews published in China and primary studies on acupuncture for stroke.