Seventeen RCTs (n=1,445) were included.
The overall methodological quality of the trials was poor. Five trials included both children and adults, no trials presented a sample size calculation, one trial described the method of randomisation, and withdrawals were described in only 2 trials. Nine of the 17 trials were double-blind. Fourteen of the 17 trials scored 3 or less out of 5 on the Jadad scale.
Traditional Chinese herbal medicine (6 trials).
All scored 1 on the Jadad scale. Gingko biloba (1 trial, concentrated gingko leaf liquor 15 g thrice daily) gave a clinically relevant improvement at 8 weeks in FEV1, which was significantly greater than placebo (p<0.05). L. wallichii (1 trial, 10 mL thrice daily) gave a significantly improved FEV1 at one month compared with baseline (p<0.01), but this was not clinically relevant. The results compared with the control were not reported. The SBR method (1 trial) reported a significant but not clinically relevant increase in FEV1 compared with baseline over 2 weeks (p<0.01). The results compared with the control were not reported. The RKISP method in conjunction with steroid treatment (1 trial) gave a significant and clinically relevant increase in FEV1 over 4 to 6 months when compared with baseline, and a significant but not clinically relevant increase was seen in the control group (steroid treatment alone, p<0.05). No comparison between the treatment and control groups was reported. IKPA tablets (1 trial, 5 tablets thrice daily) were found to improve FEV1 significantly more than the control of beclomethasone dipropionate over 3 months (p<0.05). Improvements in both groups from baseline were significant and clinically relevant. The Wenyang Tonglulo mixture (1 trial, 30 mL twice daily) improved FEV1 significantly more than the control of oral salbutamol and inhaled beclomethasone over 8 weeks (p<0.05). Improvements in both groups from baseline were significant and clinically relevant.
Traditional Indian herbal medicine (8 trials).
P. kurroa (1 trial, P. kurroa root powder 300 mg thrice daily) showed no significant benefit over placebo. S. xanthocarpum and S. trilobatum were compared with salbutamol or deriphylline in one trial. A significant increase in FEV1 from baseline was found in both herb groups, although this was less than the increase in the control groups. B. serrata (1 trial, B. serrata gum resin) gave a significantly greater increase in FEV1 than placebo (p<0.0001). Only 3 of the 5 trials of T. indica reported between-group comparisons. One of these favoured T. indica over placebo for FEV1 improvement (p<0.01), symptom scores (p<0.05) and medication usage (p<0.01) at 4 weeks. The second showed no difference compared with placebo at 16 days for lung function, but a significant improvement in nocturnal dyspnoea (p<0.01). The third showed no significant differences compared with placebo at 3 weeks.
TJ-96 (1 trial).
Symptomatic improvement and patient perceptions were reported to be significantly better in the treatment group than the control group (p<0.01). The results of lung function tests were not reported.
Marihuana (1 trial).
No significant difference from placebo was found in the group which smoked a 2% tetrahydrocannabinol preparation for 2 hours. However, a significant decrease in airway resistance (10 to 13%, p<0.05), compared with placebo, was seen 1 to 4 hours after ingestion of a 2% tetrahydrocannabinol capsule.
Dried ivy leaf extract (1 trial, 35 mg).
In children, at 3 days, there was no significant improvement in FEV1 compared with placebo, but a significant decrease in airway resistance (23.6%, p=0.0361).