Thirty-four studies (n=1,719) were included in the review: 22 RCTs (n=1,115) and 12 NRSs (n=604).
Most of the studies were small (less than 30 participants) and most (59%) were considered to be of a poor quality (NRSs or 0 to 2 points on the Jadad scale). Nine studies reported using standardised preparations, while eight reported testing or provided ginseng profiles of the preparations.
Multiple-dose studies (6 RCTs, 2 NRSs): 2 RCTs found a significant reduction in blood-pressure compared with placebo. One of these found a significant reduction in systolic blood-pressure (4%) in patients with type 2 diabetes mellitus compared with placebo, while the other found a significant reduction in diastolic blood-pressure (3%) in patients with physical or emotional stress compared with the control group. Three studies demonstrated slight elevations in blood-pressure (1 to 4%), two compared with placebo and one compared with baseline, but these were not statistically significant.
Single-dose studies (1 RCT, 3 NRSs): 2 acute single-dose studies (1 RCT, 1 NRS) found greater reductions in systolic and diastolic blood-pressure (8 to 11%). When the participants in one of these studies were examined in a subsequent RCT with placebo control, a much smaller reduction was shown (1 to 4%).
Lipid (3 RCTs, 6 NRSs).
Overall, the results were inconsistent. Five studies found a statistically significant improvement in one or more lipid parameters compared with baseline or an active control group.
Blood glucose. Single-dose studies (9 RCTs): one research group performed these 9 acute single-dose studies. Five studies conducted with Panax quinquefolius found reductions in postprandial blood glucose levels (9 to 39%) in patients with and without diabetes. Three acute studies conducted with Panax ginseng found either no change in postprandial blood glucose or an increase in postprandial blood glucose levels. Multiple-dose studies (4 RCTs, 2 NRSs): 4 studies found significant reductions in blood glucose outcomes compared with placebo. Reductions of 7 to 10% in fasting blood pressure (3 studies), 4 to 8% in glycosylated haemoglobin (2 studies) and 13% in blood glucose (1 study) were observed. No other statistically significant between-group differences were found.