Twenty-two studies were included in the evidence synthesis section of the review: 2 RCTs (n=100), 5 CCTs (n=266) and 15 pre-test post-test studies (n=1,262)). A further 40 studies were included in the appendices as they did not satisfy the inclusion criteria reported above, but their results have still been presented by the review authors.
There were several methodological limitations in the design of the studies (lack of randomisation, blinding and comparison groups) and small sample sizes. The studies were also heterogeneous in the herbs and formulas evaluated and in their method of preparation.
Two RCTs evaluated Coccinia indica and holy basil, respectively, and showed a favourable benefit in levels of fasting and post-prandial blood glucose compared with control.
The results of the CCTs showed a favourable effect on reducing fasting blood sugar (2 studies of gymnema and one of C. tamala), while 2 studies showed no statistically significant benefit (one of Coccinia indica and one of Eugenia jambolana). Four studies measured post-prandial blood sugar, of which two showed a favourable effect. Two studies of gymnema measured haemoglobin and both showed a favourable effect.
The majority of pre-test post-test comparison studies showed a benefit associated with treatment.
No serious adverse events were reported, although they were not always evaluated in the included studies and may, therefore, have been under-reported.