Twelve studies (n=1,134) were included: 10 RCTs (n=1,010) and 2 non-randomised controlled trials (n=124).
Two studies reported the randomisation method, five each reported blinding of the assessors and patients, three reported drop-outs, nine reported baseline comparisons of the treatment groups, one reported a check on the success of blinding, three reported power calculations and three reported intention-to-treat analysis.
Patients with generalised anxiety (4 RCTs plus 2 non-randomised controlled trials).
The only RCT (n=56) that used a sham acupuncture control treatment reported a significantly improved Clinical Global Impression score, clinical improvement at 10 weeks and an increased percentage of responders in the acupuncture group compared with the control group, but the majority of patients had minor depression rather than anxiety. Two RCTs (n=39 and n=296) that compared acupuncture with drug treatments reported no significant difference between treatments. One RCT (n=240) reported a significantly greater ‘cure rate’ in patients allocated to acupuncture plus EMG BFB than in those allocated to EMG BFB alone. One non-randomised trial (n=100) reported a significantly greater cure rate in patients allocated to acupuncture plus drug treatment than in those allocated to drug treatment alone, but the outcome measurement was subjective. There was insufficient information reported for the other non-randomised trial to draw any conclusions.
Patients with situational anxiety (6 RCTs).
All 4 RCTs that compared acupuncture with sham acupuncture (n=36, n=55, n=91 and n=67) reported some positive outcomes with acupuncture interventions compared with controls; three of these studies evaluated auricular acupuncture and the fourth evaluated auricular acupuncture and acupuncture at relaxation points. One RCT (n=90) reported no significant difference in anxiety between patients treated with acupuncture, diazepam and progressive relaxation. One RCT (n=40) reported that, based on State-Trait Anxiety Inventory scores, acupuncture was significantly more effective than diazepam.