We appreciate the comments of Dr. Plotnikoff and his colleagues. Their letter addresses important issues that warrant clarification and further discussion.

Our data showed that American ginseng (Panax quinquefolius L.), a commonly used herb in the United States, reduces the anticoagulant effect of warfarin. Limitations on the length of our article prevented us from providing information about other types of ginseng, which were discussed in our previous publications (1, 2). We did not extrapolate our data on American ginseng to other species of ginseng. Our study, however, provides an example of drug–herbal interaction. Herbal medicines, such as American ginseng, have beneficial effects (1). But herbs contain pharmacologically active constituents that may also interact with drugs. This interaction is particularly important if the drug has a narrow therapeutic index, such as warfarin. A slight alteration to warfarin's effect may have clinical consequences. To prevent the possible clinical consequences of herbal therapies, more controlled trials of drug–herbal interaction are needed.

The major active constituents of ginseng are ginsenosides. Dr. Plotnikoff and his colleagues point out that the American ginseng in our study has an Rb₁/Rg₁ ratio of approximately 6 to 1, a ratio higher than usual. We believe, however, that this ratio falls within the range of 5.5 to 9.6, a range found in previous reports (3, 4). Nonetheless, we acknowledge the variability of the chemical composition of botanicals. The composition of herbal products can vary from manufacturer to manufacturer and from lot to lot from the same manufacturer (4). Data from our laboratory found significant variation in ginsenoside content between measured samples (5). Cultivation conditions such as soil, temperature, moisture, period of cultivation, and harvest season can change total ginsenoside concentration, as well as the percentage of individual ginsenosides. Asian ginseng (Panax ginseng) has a different ginsenoside profile than American ginseng. Whether Asian ginseng interacts with warfarin remains to be tested.

Since animal studies and clinical trials performed in the past to test the effects of ginseng often used preparations with variable phytochemical content, results of these studies are difficult to compare. Lack of standardized processing methods contributes to the variations in a ginseng product. Some herbal manufacturers have tried to standardize products to fixed concentrations of selected chemical constituents. The benefit of this effort is uncertain, however, because herbs may achieve their effects through the combined or synergistic actions of different constituents. Thus, future investigation toward standardization of preparations is clearly needed.