RÉSUMÉ
<p><b>OBJECTIVE</b>To exam the effect and safety of conventional acupuncture (CA) on cardiac arrhythmia.</p><p><b>METHODS</b>Nine medical databases were searched until February 2016 for randomized controlled trials. Heterogeneity was measured by Cochran Q test. Meta-analysis was conducted if I was less than 85% and the characteristics of included trials were similar.</p><p><b>RESULTS</b>Nine qualified studies involving 638 patients were included. Only 1 study had definitely low risk of bias, while 7 trials were rated as unclear and 1 as high. Meta-analysis of CA alone did not have a significant benefit on response rate compared to amiodarone in patients with atrial fibrillation (Af) and atrial flutter (AF) [relative risk (RR): 1.09; 95% confidence interval (CI): 0.79-1.49; P=0.61; I=61%, P=0.11]. However, 1 study with higher methodological quality detected a lower recurrence rate of Af in CA alone as compared with sham acupuncture plus no treatment, and benefits on ventricular rate and time of conversion to normal sinus rhythm were found in CA alone group by 1 study, as well as the response rate in CA plus deslanoside group by another study. Meta-analysis of CA plus anti-arrhythmia drug (AAD) was associated with a significant benefit on the response rate when compared with AAD alone in ventricular premature beat (VPB) patients (RR, 1.19, 95% CI: 1.05-1.34; P=0.005; I=13%, P=0.32), and an improvement in quality-of-life score (QOLS) of VPB also showed in 1 individual study. Besides, a lower heart rate was detected in the CA alone group by 1 individual study when compared with no treatment in sinus tachycardia patients (MD-21.84 [-27.21,-16.47]) and lower adverse events of CA alone were reported than amiodarone.</p><p><b>CONCLUSIONS</b>CA may be a useful and safe alternative or additive approach to AADs for cardiac arrhythmia, especially in VPB and Af patients, which mainly based on a pooled estimate and result from 1 study with higher methodological quality. However, we could not reach a robust conclusion due to low quality of overall evidence.</p>