American Ginseng (Panax Quinquefolius L.) with a Depressed Ginsenoside Proﬁle Does Not Affect Postprandial Glycemia

Home , American ginseng

European Journal of Clinical Nutrition (2003) 57, 243–248 ß 2003 Nature Publishing Group All rights reserved 0954–3007/03 $25.00 www.nature.com/ejcn ORIGINAL COMMUNCIATION Variable effects of American ginseng: a batch of American ginseng (Panax quinquefolius L.) with a depressed ginsenoside proﬁle does not affect postprandial glycemia

JL Sievenpiper1,2, JT Arnason3, LA Leiter1,2 and V Vuksan1,2*

1Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, Ontario; 2Clinical Nutrition and Risk Factor Modiﬁcation Centre, St. Michael’s Hospital, Toronto, Canada; and 3Department of Biology, Faculty of Science, University of Ottawa, Ottawa, Canada

Background: We have repeatedly reported that American ginseng (AG) with a specific ginsenoside profile significantly decreases postprandial glycemia. Whether this effect is reproducible using AG with a different profile is unknown. We therefore investigated the effect of a different batch of AG on glycemia following a 75 g oral glucose tolerance test (OGTT). Methods: Using a randomized, single blind design, 12 normal subjects (six males and six females, aged 31Æ 3 y, body mass index (BMI) 28Æ 2kg=m2) received 6 g AG or placebo 40 min before a 75 g OGTT. The protocol followed the guidelines for the OGTT, with venous blood samples drawn at 7 40, 0, 15, 30, 45, 60, 90 and 120 min. Ginsenosides in the AG were assessed by established methods for HPLC-UV. Results: Repeated measures analysis of variance demonstrated that there was no significant effect of the AG on incremental plasma glucose (PG) or insulin (PI) or their areas under the curve Indices of insulin sensitivity (insulin sensitivity index (ISI)) and release (DPI30-0=DPG30-0) calculated from the OGTT were also unaffected. The AG contained 1.66% total ginsenosides, 0.90% (20S)-protopanaxadiol (PPD) ginsenosides, and 0.75% (20S)-protopanaxatriol (PPT) ginsenosides, with the following key ratios: PPD:PPT of 1.2, Rb1:Rg1 of 8.1, and Rb2:Rc of 0.18. Conclusions: The present batch of AG was unable to reproduce the postprandial hypoglycemic effects we observed previously. Possible explanations for this discrepancy include marked decrements in total ginsenosides and the key ratios PPD:PPT, Rb1:Rg1, and Rb2:Rc. These data suggest that the ginsenoside profile of AG might play a role in its hypoglycemic effects. The involvement of other components cannot, however, be precluded. European Journal of Clinical Nutrition 2003; 57, 243 – 248. doi:10.1038=ejcn.1601550

Keywords: alternative medicine; ginseng; diabetes; postprandial; glucose; insulin

*Correspondence: V Vuksan, Clinical Nutrition and Risk Factor Introduction Modification Centre, St. Michael’s Hospital, 6 138-61 Queen Street East, While the use of herbal supplements has increased dramati- Toronto, Ontario, Canada M5C 2T2. cally in recent years (Eisenber et al, 1998), poor standardiza- E-mail: [email protected] Guarantor: V Vuksan. tion remains a serious concern. Numerous articles (Levin Contributors: JLS was the Phd student assigned to the project. He et al, 1997), editorials (Angelle Kassirer, 1998), and commen- wrote the grant, participated in the design, coordinated testing and taries (Fontanarosa & Lundberg, 1998; Hershfield, 1998; laboratory analyses, performed the statistical analyses, and wrote the Oppel, 1998) have called for more clinical safety and efficacy paper. JTA performed and provided interpretation of the ginsenoside analyses and revised intellectual content. LAL was a co-supervisor of testing of herbal products and the identification of active the project. He participated in the design and planning of the study, components to improve the basis for standardization. This supervised the patients medically, and revised intellectual content. call is of particular relevance to ginseng. The principal VV was the senior supervisor and principal investigator of the project. components used in ginseng authentication are its glycosidal He supervised all aspects of the research directly and provided final input into design, analyses, interpretation, and revision of intellectual saponins, called ginsenosides (Attele et al, 1999). Interspecies content. variation in ginsenosides has been reported to be as high as Variable effects of American ginseng JL Sievenpiper et al 244 15-fold in dried powder and 36-fold in liquid preparations glucose tolerance test (75 g OGTT; WHO Study Group, (Harkey et al, 2001). Intraspecies variation has been shown to 1994). Participants attended the Risk Factor Modification be of equal concern. Intrapsecies coefficient-of-variation Centre at St Michaelss hospital twice following a 10 – 12 h (CV), for example, can be extrapolated from a study of overnight fast. A minimum of 3 days separated each visit to American ginseng (AG) harvested at the same time from minimize carry-over effects. Each participant was instructed nine British Columbian farms (Li et al, 1996). It was calcu- to maintain the same dietary and exercise patterns the lated to range from 16 to 38% for the seven most common evening before each test and consume a minimum of 150 g ginsenosides and their totals and from 28 to 122% for of carbohydrate each day over the 3 days prior to the test. To ginsenoside-specific ratios used in its authentication. A ensure compliance, participants completed a questionnaire recent advisory by the Federal Drug Administration (FDA) detailing pre-session information about their diet and life- also recently reported that a ginseng preparation, marketed style patterns and submitted to measurements of their body as a diabetes remedy, contained the hypoglycemic agent, weight and total body fat, assessed by infrared-interactance sulfonylurea (Leiter, 2000). These inconsistencies challenge using a Futrex-50001 (Futrex Inc., Gaithersburg, MD, USA). the notion that the results seen with one ginseng prepara- Upon commencement of the OGTT, participants had a tion necessarily apply to others or even another batch of the catheter inserted into a forearm vein that was secured by same preparation. In this regard, whether other batches of tape and kept patent by saline. From this device a fasting AG with different ginsenoside profiles will be able to repli- 7 ml blood sample was obtained in a plasma tube. Either AG cate the postprandial glycemic reductions we repeatedly or placebo capsules were then administered with exactly observed with a single batch of AG is unclear (Vuksan et al, 300 ml of tap water. Participants gave another blood 2000a – c, 2001a,b). We therefore assessed the effect of a sample after 40 min. This was followed by consumption of second batch of AG from the same supplier on postprandial the 75 g oral glucose load (75 g Glucodex1, Technilab, glycemia in the present study. Quebec, Canada) over exactly 5 min. Additional blood samples were drawn using the same technique at 15, 30, 45, 60, 90 and 120 min after the start of the challenge. Adverse Methods symptom monitoring during each clinic visit and in the Participants intervening washout days ( 3 days) was assessed by sub- Healthy participants were recruited from faculty and stu- jective seven-point bipolar visual analogue scales. dents at the University of Toronto and hospital advertise- ments. Exclusion criteria included previously diagnosed dysglycemia, liver or kidney disease, major surgery in the Ginseng analyses last 6 months, < 18 or > 75 y of age, and morbid obesity Ginsenoside profiles of the AG and placebo used on the ( > 35 kg=m2). Twelve participants (six males and six females, present study were measured using standard techniques. aged 31Æ 3 y, body mass index (BMI) 28Æ 2kg=m2) were The ginsenosides comprise principally a family of steroids recruited who met these criteria. All gave informed written called dammarane-type triterpene glycosides with either consent. The study was approved by the Research Ethics (20S)-protopanaxadiol (PPD) or (20S)-protopanaxatriol Board at St Michael’s Hospital. (PPT) as the aglycone. The four main PPD ginsenosides

(Rb1,Rb2, Rc, Rd) and three main PPT ginsenosides (Rg1, Re, Rf) were analyzed using a following HPLC-UV techniques Treatments developed for the American Botanical Council (ABC) Gin- Participants received either 6 g of AG or placebo in a rando- seng Evaluation Program (Fitzloft et al, 1998). The HPLC mized cross-over design. The AG capsules contained 3-y-old conditions included: chromatograph, Beckman HPLC Ontario grown, dried, ground AG root (Panax quinquefolius system; column, a reverse-phase Beckman ultrasphere C-18, L.) supplied by the same supplier, Chai-Na-Ta Corp., BC, 5 mm octadecylsilane, 250Â4.6 mm column; mobile phase, Canada, that provided the original batch for which we de-ionized water and acetonitrile; ﬂow rate, 1.3 ml=min; UV reported (Vuksan et al, 2000a – c, 2001a,b) a glycemic low- detection, a module 168 diode-array detector set at 203 nm.

ering effect. This AG was the same commercially available The ginsenoside standards for Rg1 and Re were provided by product, but from a different batch than the original. The Dr H Fong University of Illinois and the Rf, Rb1, Rc, Rb2,Rd placebo, on the other hand, consisted of identical capsules standards were provided by Indoﬁne Chemical Co., Somer- containing corn ﬂour. The energy, carbohydrate content and ville, NJ, USA. appearance of the placebo were designed to match that of the AG. Blood glucose analysis All samples were separated by centrifuge and the plasma Protocol immediately frozen at 720C pending analysis. Analyses of The protocol followed the World Health Organization glucose concentration of each sample were done by the (WHO) guidelines for the administration of a 75 g oral glucose oxidase method (Kadish & Hall, 1965) and the

European Journal of Clinical Nutrition Variable effects of American ginseng JL Sievenpiper et al 245 insulin concentration, by double antibody radioimmuno- meter (P ¼ 0.42, P ¼ 0.18). This was reflected in a lack of effect assay (Livery et al, 1980), at the Banting and Best Diabetes of AG on PG-AUC (P ¼ 0.96) and PI-AUC (P ¼ 0.58). It was Centre Core Laboratory, Toronto, Canada. also reflected in the lack of effect on derived indices of whole body insulin sensitivity (ISI; P ¼ 0.76) and first release insulin

secretion (DPI30-0/=DPG30-0; P ¼ 0.79). Panel (B) compares the Statistical analyses effect of 6 g of our original AG batch on incremental PG and Incremental plasma glucose (PG) and insulin (PI) curves were PI following a 75 g OGTT with a historical control (a 75 g plotted and the positive incremental area under the curve OGTT done alone previously). Two-way repeated measures (AUC) was calculated (Wolever et al, 1991). The whole body ANOVA demonstrated a significant effect of time insulin sensitivity index (ISI; Matsuda & De Fronzo, 1999) (P < 0.0001) on PG and PI, with significant independent and the early insulin secretion index, DPI30-0=DPG30-0 (Phil- (P ¼ 0.016) and interactive effects (P ¼ 0.046) of treatment lips et al, 1995), were also calculated from derived OGTT on PG only. Pairwise comparisons showed that 45 min, equations. ISI was calculated using fasting PG (FPG) and PI 60 min, 90 min and AUC PG were significantly decreased (FPI) with mean OGTT outcome, according to the formula (P < 0.05). This was reflected in a significant effect of AG on (Matsuda & De Fronzo, 1999): 10 000 divided by the square PG-AUC (P ¼ 0.02). Neither PI-AUC (P ¼ 0.51) nor whole root of ((FPGÂFPI)Â(mean-PGÂmean-PI)), where PG is body insulin sensitivity (ISI; P ¼ 0.49) nor first release insulin expressed in mg=dl (0.0551 mmol=l) and PI in mU=ml secretion DPI30-0=DPG30-0 P ¼ 0.67) were affected (Vuksan

(6 pmol=l). The early insulin secretion index, DPI30-0=DPG30-0, et al, 2001b). was calculated as the change in PI from 0min to 30 min Table 1 shows the ginsenoside analysis for the placebo divided by the change in PG over the same period (Phillips capsules, the present batch of AG, and our original batch et al, 1995). Statistical analyses were then performed using (Vuksan et al, 2000a – c, 2001a,b). The composition of both the Number Cruncher Statistical System (NCSS) 2000 soft- the present and our original batch is consistent with panax ware (NCSS statistical software, Kaysville, UT, USA). Repeated quinquefolius L. Key ratios and the presence and absence of measures two-way ANOVA assessed the interactive and inde- various ginsenosides are used in the authentication of gin- pendent effects of treatment and time on incremental PG seng. A ratio of Rb1:Rg1 > 3 (Ma et al, 1996; Awang et al, 2000) and PI. One-way repeated measures ANOVA assessed differ- and ratios of Rg1:Re and Rb2:Rc < 1 have been shown to be ences in incremental PG and PI at each time point between indicative of Panax quinquefolius L. (Chan et al, 2000; Wang AG and placebo. This statistic also assessed differences in PG et al, 1999). The most powerful marker is the absence of Rf,

AUC, PI AUC, ISI and DPI30-0=DPG30-0. Paired Student t-tests a ginsenoside not found in AG and distinctive for Asian assessed differences in subjective ratings of symptoms for ginseng (Ma et al, 1996; Awang et al, 2000; Chan et al, 2000; both the clinical testing and washout periods. All results Li et al, 2000). These criteria were met by both batches of AG. were expressed as meanÆ s.e.m. and signiﬁcant at P < 0.05. However, within these criteria, there were still large differences between the two batches. The present batch of AG had 48% less total ginsenosides, 51% lower PPD:PPT ratio, at the

Results expense of PPD, 47% lower Rb1:Rg1 ratio, at the expense of All participants followed the study protocol safely and with- Rb1, and 28% lower Rb2:Rc ratio, at the expense of Rb2. out difficulty. A minimum of 150 g of carbohydrate was consumed over the 3 days prior to each test. Baseline anthro- pometry and fasting plasma glucose and insulin were main- Discussion tained. The participants consumed the placebo and AG The present study demonstrates a favorable side-effect profile capsules in the self-standardized amount of time and the of the present AG batch but with a lack of effect on post- 75 g Glucodex1 oral glucose load in the allotted 5 min for prandial indices of glycemia and insulinemia. Although each test. There were also no differences in reported symp- insufficient quantities of the original AG batch precluded a toms of bloating, belching, nausea, dizziness, headache, bioequivalence study, existing data from earlier studies pro- diarrhea, flatulence polyuria, insomnia, anxiety, numbness, vided a basis for indirect comparisons. The lack of efficacy light-headedness, thirst or cramping between placebo and observed with the present AG batch is in direct contrast to AG treatment during the clinic visits or the intervening this existing data. We noticed previously that 6 g of our washout period. original AG batch lowered the plasma glucose response to a Figure 1 shows the effect of the present batch of AG and 75 g OGTT significantly (P < 0.05) in eight nondiabetic sub- our original batch of AG (Vuksan et al, 2001b) on postpran- jects (Vuksan et al, 1991b). These findings are in addition to dial PG and PI responses. Panel (A) compares the effect of 6 g the reductions we reported with the original AG batch at of the present AG batch with placebo on incremental PG and doses from 1 to 9 g, following a capillary blood glucose PI following a 75 g OGTT. Two-way repeated measures protocol using a 25 g OGTT in people with (Vuksan et al, ANOVA demonstrated that there was a significant effect of 2000a,b) and without (Vuksan et al, 2000c, 2001a) diabetes. time (P < 0.0001) but no effect of treatment (P ¼ 0.56, Reasons for this disagreement in findings are not clear. P ¼ 0.69) on PG and PI, with no interaction for either para- The most compelling explanation is compositional differ-

European Journal of Clinical Nutrition Variable effects of American ginseng JL Sievenpiper et al 246

Figure 1 Differential effects of American ginseng (AG) batches with different compositions: (A) the postprandial glycemic and insulinemic effects for the present AG batch. The line plots and bars represent the incremental change and AUC for plasma glucose and insulin following placebo (d)or6gAG(n) taken 40 min before a 75 g oral glucose tolerance test (75 g OGTT) in 12 healthy, nondiabetic subjects. (B) The postprandial glycemic and insulinemic effects for our original AG batch used in ﬁve previous studies (Vuksan et al, 2000a – c, 2001a). The line plots and bars represent the incremental change and AUC for plasma glucose and insulin following a 75 g OGTT done alone previously (n) or with 6 g AG taken 40 min before the test (d) in eight healthy, nondiabetic subjects. P-values are for repeated measures one-way ANOVA adjusted with the Tukey – Kramer procedure. Data are meanÆ s.e.m.

ences. Although the ginsenoside profile confirmed that the 70 kg person that we fed in the present study. This suggests present AG batch was also Panax quinquefolius L, there were that these ginsensoides might have been present in insuffi- marked decrements in ginsenosides compared with our ori- cient quantities in the present batch of AG to have similar ginal batch. It is possible that specific active ginsenosides hypoglycemic related effects. The limitation of this argu- were below the efficacy range or, in the case of the ratios, ment is that although the relative proportion of ginsenosides favored counter-regulatory effects of ginsenosides present in was less in the present batch of AG, the dose administered higher relative proportions. was double or greater than that found to be efficacious with There is some evidence to support a role of ginsenoside our original batch of AG (Vuksan et al, 2000a – c, 2001a). The differences in variable glycemic effects. The efficacy range for implication is that for even those ginsenosides which were various ginsenosides can be high. The 3 main PPD ginseno- 50% lower, the quantity administered would be roughly

sides measured on the present study, Rb1,Rb2, Rc, have been equivalent. shown only to be able to increase glucose transport at higher If the present AG was not administered at a quantity levels ( > 1.0 mM) in sheep erythrocytes (Hasegawa et al, below the efﬁcacy range of its active ginsenosides, then

1994). It was also with a high dose of Rb2 (10 mg=kg) that counter-regulatory effects of speciﬁc ginsenosides must be an increase in glycolytic enzymes and a decrease in gluco- considered. A water extract of Asian ginseng, expected neogenic enzymes were observed in rat liver (Yokozawa et al, to have lower PPD:PPT ratios (Cui, 1995), inhibited non- 1984, 1985). The 10 mg=kg dose used in these latter studies is insulin stimulated 2-deoxy-D-[2-3H]glucose (2-DG) uptake in 100-fold higher than the 0.077 mg=kg equivalent for a 3T3-L1 adipocytes at the highest dose of 10 000 mg=ml and

European Journal of Clinical Nutrition Variable effects of American ginseng JL Sievenpiper et al 247 Table 1 Ginsenoside proﬁles be absorbed (Yamamoto, 2001). In the absence of measurements of these components in our ginseng batches, this still Contenta leaves open the possibility of their interaction in effects. Present Original In conclusion, different batches of AG from the same Ginsenosides Placebo AG batch AG batch supplier can give differential results. Although ginsenoside

(20S)-protopanaxadiols (PPD) (% w=w) differences are probably driving effects, other components

Rb1 — 0.65 1.53 such as quinquefolans and ginsenans might share some Rb2 — 0.02 0.06 responsibility. This situation warns against the lack of gin- Rc — 0.11 0.24 seng standardization. It also warns against standardization Rd — 0.12 0.44 (20S)-protopanaxatriols (PPT) (%) w=w) that is without basis. That is, the standardization of total

Rg1 — 0.08 0.1 ginsenosides or specific ginsenosides seen in some extracts Re — 0.67 0.83 and preparations (Cui, 1995) is of little benefit if the efficacy Rf — 00 of the standards is unknown. Together these concerns high- Total (% w=w) — 1.66 3.21 Ratios (% w=w:% w=w) light the need for greater research to understand the role of PPD:PPT — 1.2 2.44 ginsenosides, quinquefolans, ginsenans, and other isolated Rb1:Rg1 — 8.1 15.3 components and the effect their variability has on safety and Rb2:Rc — 0.18 0.25 efficacy. In this regard, more detailed animal and in vitro Rg :Re — 0.12 0.12 1 work to determine the hypoglycemic activities if AG and its aDetermined by HPLC-UV analyses (Fitzloff et al, 1998). The present AG batch components is required, with the ultimate goal of identify- represents the AG batch tested in the present study and the original AG batch ing candidate components for clinical safety and efficacy represents the AG batch shown to be efficacious in our previous studies testing. Such efforts will increase the ability of researchers to (Vuksan et al, 2000a – c, 2001a,b). perform interpretable, generalizable studies, the consumer to select consistently efficacious and safe products, and health significantly inhibited insulin stimulated 2-DG uptake com- practitioners to assess possible therapeutic benefits and pared with control in a dose-related manner (Hong et al, adverse ginseng – drug interactions. Until then, advocacy of 2000). These data suggest a possible inhibition of GLUT-1 at AG for any of its indications is premature. high doses and the insulin-dependent GLUT-4 with increas- ing doses. PPTs were also observed to inhibit [14C]-a-mean glucose uptake in a dose dependent manner at doses from 10 Acknowledgements to 100 mg in cultured rabbit renal proximal tubular cells in We thank MuscleTech Research and Development Incorpo- which the sodium coupled glucose transporter, SGLT-1, rated, Mississauga, for generously providing the main fund- mediates glucose uptake (Han et al, 1999). As the relative ing for this study. We also thank Chai-Na-Ta Corp., Langley, proportion of PPT was increased in the PPD:PPT ratio in the BC for providing the AG and placebo capsules. This work was present AG batch studied, this might have had an opposing presented as a poster at the 19th International Symposium on effect on glucose transport mechanisms. Taken together, the Diabetes and Nutrition, Diabetes and Nutrition Study Group AG ginsenoside profile should be considered as a possible (DNSG) of the European Association for the Study of Diabetes mediator of its effects. (EASD) in Dusseldorf, Germany 28 June – 1 July, 2001 (Sie- Other components alone or together with ginsenosides venpiper et al, 2001). JLS was in receipt of an Ontario might also have influenced the results. For example, the Graduate Scholarship during the conduct of this work. peptidoglycans (panaxans from Asian ginseng and quinque- VV has received research and travel funding from Chai- folans from AG) and polysaccharides (ginsenans) from gin- Na-Ta Corp., Langley, BC and MuscleTech Research and seng have shown pharmacological activity. Of these only the Development Incorporated. former have shown glycemic effects. Panaxans A-L (Konno et al, 1984, 1985; Oshima et al, 1985, 1987) and quinquefolans A-C (Konno et al, 1985) have shown hypoglycemic References effects with varying potencies when administered as intra- Angell M & Kassirer JP (1998): Alternative Medicine — the risks of perintoneal injections in both normal and alloxan induced untested and unregulated remedies. New Engl. J. Med. 339, 839 – hyperglycemic mice. However, their influence is dubious. 841. Whether the main route of commercial administration, oral Attele AS, Wu JA & Yuan CS (1999): Ginseng pharmacology: multiple administration, will result in the same effects in humans is constituents and multiple actions. Biochem. Pharmacol. 58 1685 – 1693. uncertain. Digestive gut and colonic microfloral enzymes Awang DVC (2000): The neglected ginsenosides of North American would likely degrade these components and those that ginseng (panax quinquefolius L.). J. Herbs Spices Med. Plants 7, 103 – escaped digestion would be impermeable across the enter- 109. ocytes. Both inevitabilities would preclude any activity. Chan TWD, But PPH, Cheng SW, Kwok IMY, Lau FW & Xu HX (2000): Differentiation and Authentication of Panax ginseng, There is, nevertheless, data that suggest that functional Panax quinquefolius, and Ginseng Products by Using HPLC=MS. amounts of some intact peptides and peptidoglycans can Analy. Chem. 72, 1281 – 1287.

European Journal of Clinical Nutrition Variable effects of American ginseng JL Sievenpiper et al 248 Cui JF (1995): Identification and quantification of ginsenosides in Matsuda M & DeFronzo RA (1999): Insulin sensitivity indices various commercial ginseng preparations. Eur. J. Pharm. Sci. 3,77– obtained from oral glucose tolerance testing: comparison with 85. the euglycemic insulin clamp. Diabetes Care 22, 1462 – 1470. Eisenberg DM, Davis RB, Ettner SL, Appel S, Wilkey S, Van Rompay M Oppel L (1998): Evaluating unconventional therapies. Can. Med. & Kessler RC (1998): Trends in Alternative Medicine use in the Assoc. J. 159, 759. United States, 190 – 1997: Results of a Follow-up National Survey. Oshima Y, Konno C & Hikino H (1985): Isolation and hypoglycemic J.A.M.A. 280, 1569 – 1575. acitivity of panaxans I, J, K, and L, glycans of Panax ginseng roots. Fitzloff JF, Yat P, Lu ZZ, Awang DVC, Arnason JT, van Breeman RB, J. Ethnopharmac. 14, 255 – 259. Hall T, Blumethal M & Fong HHS (1998): Perspectives on the Oshima Y, Sato K & Hikino H (1987): Isolation and hypoglycemic quality control assurance of ginseng products in North America. activity of quinquefolans A, B, and C, glycans of Panax quinque- In Advances in Ginseng Research: Proceedings of the 7th Interna- folium roots. J. Nat. Prod. 50,188– 190. tional Symposium on Ginseng, H Huh, KJ Choi, YC Kim ed. pp Phillips DI, Clark PM, Hales CN & Osmond C (1995): Understanding 138 – 145. Seoul: Korean Society of Ginseng. oral glucose tolerance: comparison of glucose or insulin measure- Fontanarosa PB & Lundberg GD (1998): Alternative medicine meets ments during the oral glucose tolerance test with specific measure- science. J.A.M.A. 280, 1618 – 1619. ments of insulin resistance and insulin secretion. Diabetes Med. 12, Han HJ, Park SH, Koh HJ, Nah SY, Shin DH & Choi HS (1999): 931. Protopanaxatriol ginsenosides inhibit glucose uptake in primary Sievenpiper JL, Arnason JT, Xu Z, Leiter LA & Vuksan V (2001): cultured rabbit renal proximal tubular cells by arachidonic acid Differential effects of American ginseng (Panax quinquefolius L.) release. Kidney Blood Press Res. 22,114– 120. With different compositions on postprandial glycemia. Diabetes Harkey MR, Henderson GL, Gershwin ME, Stern JS & Hackman RM Nutr. Metab. 14, 169 – 170 (abstract). (2001: Variability in commercial ginseng products: an analysis of Vuksan V, Sieveniper JL, Koo VYY, Francis T, Beljan-Zdravkovic U, Xu 25 preparations. Am. J. Clin. Nutr. 73, 1101 – 1106. Z & Vidgen E (2000a): American ginseng reduces postprandial Hasegawa H, Matsumiya S, Murakami C, Kurokawa T, Kasai R, glycemia in nondiabetic and diabetic individuals. Arch. Intern. Ishibashi S & Yamasaki K (1994): Interactions of ginseng extract, Med. 160, 1009 – 1013. ginseng separated fractions, and some triterpenoid saponins with Vuksan V, Stavro MP, Sievenpiper JL, Beljan-Zdravkovic U, Leiter LA, glucose transporters in sheep erythrocytes. Planta Med. 60, 153 – Josse RG & Zheng Xu (2000b): Similar postprandial glycemic 157. reductions with escalation of dose and administration time of Hershfield NB (1998): Herbal medicine: show me the proof! Can. American ginseng in type 2 diabetes. Diabetes Care 23, 1221 – Med. Assoc. J. 158, 1689 – 1690. 1226. Hong SJ, Fong JC & Hwang JH (2000): Effects of crude drugs on Vuksan V, Stavro MP, Sievenpiper JL, Koo VY, Wong E, Beljan- glucose uptake in 3T3-L1 adipocytes. Kaohsiung J. Med. Sci. 16, Zdravkovic U, Francis T, Jenkins AL, Leiter LA, Josse RG & Xu Z 445 – 451. (2000c): American ginseng improves glycemia in individuals with Kadish AH & Hall DA (1965): A new method for the continuous normal glucose tolerance: Effect of dose and time escalation. monitoring of blood glucose by measurement of dissolved oxygen. J. Am. Coll. Nutr. 19, 738 – 744. Clin. Chem. 11,869– 875. Vuksan V, Sievenpiper JL, Xu Z et al (2001a): American ginseng Konno C, Sugimyama K, Kano M & Hikino H (1984): Isolation and (Panax quinquefolius L.) attenuates postprandial glycemia in a hypoglycemic activity of panaxans A, B, C, D, and E glycans of time, but not dose, dependent manner in healthy individuals. Panax ginseng roots. Planta Med. 50, 434 – 436. Am. J. Clin. Nutr. 73, 753 – 758. Konno C, Murakami M, Oshima Y & Hikino H (1985): Isolation and Vuksan V, Sievenpiper JL, Xu Z, Beljan-Zdravkovic U, Jenkins AL, hypoglycemic activity of panaxans Q, R, S, T and U, glycans of Arnason JT, Bateman R, Leiter LA, Josse RG, Francis T & Stavro MP Panax ginseng roots. J. Ethnopharmac. 14,69– 74. (2001b): Ginseng and diabetes: A new way to use an old medicine. Leiter LA (2000): Alternative and ‘Alternative’ diabetes therapies. C.J.D.C. 25(Suppl 2), 111 – 120. C.J.D.C. 24,21– 22. Wang X, Sakuma T, Asafu-Adjaye E & Shiu GK (1999): Determination Levin JS, Glass TA, Kushi LH, Schuck JR, Steele L & Jonas WB (1997): of ginsenosides in plant extracts from Panax ginseng and Panax Quantitative methods in research on complementary and alter- quinquefolius L. by LC=MS=MS. Anal. Chem. 71, 1579 – 1584. native medicine. A methodological manifesto. NIH Office of WHO Study Group (1994): Prevention of diabetes mellitus. WHO Alternative Medicine. Med. Care 35, 1079 – 1094. Technical Report series no. 844. Geneva: WHO. Li TSC, Mazza G, Cottrell AC & Gao L (1996): Ginsenosides in roots Wolever TMS, Jenkins DJA, Jenkins AL & Josse RG (1991): The and leaves of American ginseng. J. Agric. Food. 44, 717 – 720. glycemic index: methodology and clinical implications. Am. J. Li W, Gu C, Zhang H, Awang DV, Fitzloff JF, Fong HH & van Breemen Clin. Nutr. 54, 846 – 854. RB (2000): Use of high-performance liquid chromatography- Yamamoto A (2001): Improvement of transmucosal absorption of tandem mass spectrometry to distinguish Panax ginseng CA biologically active peptide drugs. Yakugaku Zasshi 121, 929 – 948. Meyer (Asian ginseng) and Panax quinquefolius L. (North Amer- Yokozawa T, Kobayashi T, Kawai A, Oura H & Kawashima Y (1984): ican ginseng). Anal. Chem. 72, 5417 – 5422. Stimulation of lipid and sugar metabolism in ginsenoside-Rb2 Liverey JH, Hodgkinson SC Roud HR & Donald RA (1980): Effect of treated rats. Chem. Pharm. Bull. 32, 2766 – 2772. time, temperature and freezing on the stability of immunoreactive Yokozawa T, Kobayashi T, Oura H & Kawashima Y (1985): Studies on LH, FSH, TSH, growth hormone, prolactin and insulin in plasma. the mechanism of the hypoglycemic activity of ginsenoside-Rb2 Clin Biochem 13, 151 – 155. in streptozotocin-diabetic rats. Chem. Pharm. Bull. 33,869– 872. Ma Y, Mai L, Malley L & Doucet M (1996): Distribution and proportion of major ginsenosides and quality control of ginseng products. Chin. J. Med. Chem. 6,11– 21.

European Journal of Clinical Nutrition