Seven studies (n=368) were included: 3 randomised controlled trials (RCTs) (n=214) and 4 uncontrolled studies (n=154).
One of the 3 RCTs was of high quality (score 5 points); the other two were of poor quality, scoring only 1 point each (for randomisation). The uncontrolled studies did not report withdrawals and drop-outs and scored zero points. Major methodological limitations of the included studies were small sample size, the lack of a statistical comparison between treatments, use of unreliable or subjective outcome measures, lack of consistent protocols and poor reporting.
The one high-quality RCT (n=90) found that ear acupuncture significantly reduced pain intensity, on a VAS, compared with placebo ear acupuncture at day 30 (P=0.02) and day 60 (P<0.001) of a 2-month trial. The remaining poor-quality RCTs (n=48 and n=76) found mixed results: one found no significant difference between body acupuncture and control, while the other found body acupuncture improved chest pain.
Three of the 4 uncontrolled studies (n ranged from 10 to 92) found pain relief with acupuncture; the other study found no effect on pain.
Three studies reported mild or no adverse effects; the other 4 studies did not mention adverse effects.