In reply
Ritter and Dembicki in their letter and Cheng in his identify very important points of discussion. One concern was that some of the confounding variables inherent in the testing of plant-based medicines were not controlled in our study.1 Another was that some components of ginseng might interact with various drugs, resulting in undesirable outcomes.
Although the species (American ginseng [Panax quinquefolius L]), plant part (root), age (4 years), preparation (dried ground root), manufacturer lot (same), and growing location (Ontario) of the ginseng used in our study were controlled, Ritter and Dembicki are correct that we neglected to characterize the active components. Several active ginseng components that may affect carbohydrate metabolism have been implicated in data from animal studies. These include ginsenosides (saponins), the main markers used for species identification and standardization.2 Other components include panaxans (peptidoglycans) and the water- and methanol-soluble fractions of Panax ginseng CA Meyer, DPG-3-2, and EPG-3-2, respectively.2 Because of the documented variability of these components among and within ginseng species,3 it is unclear whether the glycemic-lowering effect of the ginseng in our study would occur for other species or even for related American ginseng products. We therefore have addressed the need for further characterization of active components in subsequent ginseng studies.