Herbs have experienced an unprecedented surge in popularity (1). This has occurred in the absence of adequate safety and efficacy evidence, prompting calls for rigorous clinical assessments (2). Complicating these assessments is compositional variability. This is a concern with one of the most popular herbs, ginseng (3). The principal reference components, to which pharmacological effects have been attributed, are its ginsenosides (steroidal glycosides). We undertook a systematic quantitative analysis of the literature to assess the coefficient of variation (CV) in ginsenosides across species, assay technique, and ginsenoside type.
The PubMed (1966-present), EMBASE (1980-present), HealthSTAR (1975-present), Cochrane library (issue 2, 2002), and AGRICOLA (1979-present) databases were searched using “ginsenosides AND (chromatography OR HPLC OR HPTLC OR TLC OR LC OR DCC OR GC OR ELISA OR UV OR MS OR NMR OR ELSD)”. One-hundred eleven articles were identified. Two reviewers applied three inclusion criteria: publication quality: peer-reviewed; end point: quantitative ginsenoside concentrations; and ginseng type: dried derivatives of panax species roots. Thirty-two articles met these criteria, reporting ginsenoside concentrations for 317 ginseng batches.
A three-factor analysis was performed to assess the independent and interactive effects of species, assay technique, and ginsenoside type on the CV of ginsenoside concentrations using ANOVA (NCSS 2000; NCSS, Kaysville, UT). The CVs of ginsenoside concentrations were calculated as CV = SD/mean × 100% in a factorial block design. A blocking principle was applied to the data such that each level of each factor was crossed with each level of the other factors for the calculation of CV. Species was comprised of 10 levels of panax species, their preparations, and their varieties: Asian (Panax ginseng C.A. Meyer), Asian red (Panax ginseng C.A. Meyer [red]), Asian wild (Panax ginseng C.A. Meyer [wild]), Asian extract (Panax ginseng C.A. Meyer [extract]), American (Panax quinquefolius L.), American wild (Panax quinquefolius L. [wild]), American extract (Panax quinquefolius L. [extract]), Japanese (Panax japonicus C.A. Meyer), Pseudo (Panax pseudoginseng WALL), and Sanchi (Panax notoginseng [Burk] F.H. Chen) ginsengs. Assay technique was comprised of six levels of different assay techniques: high-performance liquid chromatography (HPLC)-ultraviolet (UV), gas chromatography (GC)-mass spectrometry (MS), HPLC-MS, diode counter current DCC, HPLC-differential refractometry DR, and HPLC-electrospray light-scattering detection (ELSD). Ginsenoside type was comprised of 21 levels of different ginsenoside indexes: protopanaxadiol (PPD) ginsenosides (Rb1, Rb2, Rc, Rd, and Rg3), protopanaxatriol (PPT) ginsenosides (Rg1, Rf, Re, and Rg2), and their sums (PPD, PPT, and total) and ratios (PPD:PPT, Rb1:Rg1, Rb2:Rc, Re:Rb1, Rc:Rb1, Rd:Rb1, Rb2:Rb1, Rf:Rb1, and Rg1:Re). The CV data calculated for each possible combination of levels from the three factors were pooled and then meaned for each level of each factor. As a result, CV data are means ± SD.
This systematic quantitative analysis of the literature demonstrated high CV in ginsenosides across species, assay technique, and ginsenoside type (26–103, 31–81, and 36–112%, respectively). These large ranges produced significant differences in each main effect (P = 0.00030, P = 0.014, and P = 0.00031, respectively), with differences in species sensitive to assay technique (P = 0.00011 for two-way interaction).
The high variability in ginseng identified by this analysis might have serious clinical sequelae. Variable pharmacological effects appear secondary to ginsenoside variability. We have shown in healthy humans that while two batches of American ginseng (cultivated Panax quinquefolius L.) (4–6) demonstrated similar acute postprandial glycemic-lowering efficacy, a third batch with a depressed ginsenoside profile was ineffective (4), whereas Japanese, Asian red, and Sanchi ginsengs had null effects (6) and Asian (6,7), American wild, and Siberian ginsengs (Eleutherococcus senticosus) (6) raised glycemia. These data suggest that the antihyperglycemic efficacy of ginseng might be as highly variable as its ginsenoside composition.
Although this makes a compelling argument for better standardization, there are mitigating factors. It is unclear which of the >30 ginsenosides or myriad of other principles should be targeted for an antihyperglycemic indication. There is also no universal ginsenoside assay. Until these issues are resolved, the reproducibility of ginseng’s composition, safety, and efficacy cannot be trusted. This conclusion likely holds true for other less well-studied herbal remedies used to treat diabetes.
Article Information
Funding for this systematic quantitative analysis of the literature was provided by a Grant for Applied Research and Education from the Canadian Diabetes Association. J.L.S. was supported by an Ontario Graduate Scholarship.
The authors thank Dr. Dennis V.C. Awang for invaluable intellectual input in the area of phytochemistry.
References
V.V. has received research and travel funding from The Ontario Ginseng Growers Association and Natural Factors Nutritional Products.
