- Email: [email protected]
Considerable temporal variability in glucose reference curves in humans for a year period
Available online at www.sciencedirect.com
Nutrition Research 28 (2008) 495 – 500 www.nrjournal.com
Considerable temporal variability in glucose reference curves in humans for a year period Alison J. Wallace a,⁎, Sarah L. Eady a , Russell S. Scott b , Jinny A. Willis b , Christopher M. Frampton c a
New Zealand Institute for Crop and Food Research Limited, Private Bag 4704, Christchurch, New Zealand b Lipid and Diabetes Research Group, PO Box 4710, Christchurch 8011, New Zealand c Christchurch School of Medicine and Health Sciences, PO Box 4345, Christchurch 8011, New Zealand Received 29 November 2007; revised 15 May 2008; accepted 16 May 2008
Abstract Glycemic glucose equivalent (GGE) is a measure of the blood glucose response to a defined portion of food. Their calculation requires the measurement of a standard glucose-response curve, with beverages containing 0, 12.5, 25, 50, and 75 g of glucose measured twice each. This study was designed to determine the stability of an individual's glucose-response curve measured every 3 months for a year and of their GGE estimates for 10 foods for that period. The blood glucose response to beverages containing 0, 12.5, 25, 50, and 75 g glucose and to 10 foods was measured for 16 healthy individuals. Capillary blood samples were collected fasting, then every 15 minutes for 1 hour, and every 30 minutes for at least 2 hours. The slopes and intercepts of the 4 glucose curves and the GGE of the 10 foods calculated using the available curves for each food was compared. The results showed considerable temporal variability in the slope (intraindividual coefficient of variation (CV) = 30%) and intercept (intraindividual CV = 40%) of the glucose curves. However, if GGE values were categorized into 3 groups (low GGE, ≤10; medium GGE, 10.01-19.99; and high GGE, ≥20), all but one food was consistently classified in the same category across the 4 glucose curves. In conclusion, it appears that if the exact GGE value is required, glucose curves should be repeated at least once every 3 months, but if foods are classed into general GGE categories, it may be possible to use the same glucose curve for a longer period. © 2008 Elsevier Inc. All rights reserved. Keywords: Abbreviations:
Glycemia; Glycemic index; Glycemic glucose equivalents; Reproducibility; Humans GGE, glycemic glucose equivalents; GI, glycemic index; WHO, World Health Organization; iAUC, incremental area under the curve.
1. Introduction Interest in the effect of particular foods on blood glucose levels and the possible contribution of glycemic properties to nutrition and health has increased with the realization of the potential outcomes on people with diabetes [1,2], on those at risk for developing diabetes and coronary heart disease [3-8], on performance in sports people [9,10], and on weight loss [11-15]. Traditionally, blood glucose response has been ⁎ Corresponding author. Tel.: +64 3 325 9638; fax: +64 3 325 2074. E-mail address: [email protected] (A.J. Wallace). 0271-5317/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.nutres.2008.05.006
measured using the glycemic index (GI), which is a glycemic response to a portion of food containing a defined amount of available carbohydrate (usually 50 g), expressed as a percentage of the response to the same amount of glucose. For foods containing the same amount of carbohydrate, the GI indicates what effect the food will have on an individual's blood glucose levels. However, because foods contain different amounts of carbohydrate, ranking foods by GI will not necessarily rank them according to the effect they have on blood glucose levels [16-18]. For this reason, the glycemic glucose equivalent (GGE) has been developed. Glycemic glucose equivalent is defined as the amount of
496
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
glucose in grams that would induce a glycemic response equal to that induced by a specified weight of a food [19,20]. For example, if a portion of food has a GGE of 6 g, it induces the same glycemic response as 6 g of glucose. Glycemic glucose response is food-based and equiglycemic in measurement, whereas GI is carbohydrate-based and equicarbohydrate in measurement [21]. The most accurate way to measure GGEs for a wide variety of foods with varying GGE values is to measure them relative to a standard glucose reference curve [22]. Glucose references of 0, 12.5, 25, 50, and 75 g are measured on more than one occasion in each individual. Brouns et al [23] published an extensive review on GI methodology, in which they suggested that the same reference glucose tests could only be used for a maximum of 3 months because no data are available on the impact of time on changes in response to the reference food. An individual's blood glucose response could vary over time, and therefore, because all foods are measured relative to the glucose reference, this could result in an incorrect GGE value being obtained. If this variation is systematic, this may not matter for the measurement of GGE values for foods because the glucose reference and food may change proportionally. Also, as noted by Brouns et al [23], no one has investigated how an individual's blood glucose response varies over time, and if their physiologic state remains the same (ie, no significant changes in medication, weight, or physical activity), their blood glucose-response curve may not vary much over time. Reliable estimation of GGE is essential if the concept is to be used accurately to describe a food's glycemic properties and its impact on health outcomes. This would then enable consumers to guide their food selections appropriately to lower the overall glycemic impact of their dietary intake assisting with improved blood glucose control, body weight management, and reduced cardiovascular and diabetic risk. The aim of this study was to examine the stability of an individual's glucose response using a standard glucose curve measured every 3 months for 1 year and identify the effect of any physiologic changes in an individual that may impact upon a reliable reference value being obtained for estimation of GGE values. The GGE values for a variety of foods for this period are also established. 2. Methods and materials 2.1. Participants Sixteen individuals (10 women and 6 men) were tested for a year. Guidelines for GI methodology state that a minimum of 10 participants in these types of studies provides a sufficient degree of power and precision to the analysis (23). None of the participants had diabetes according to the WHO classification (fasting glucose b6.2 mmol/L). The mean (range) age and body mass index of the group at the beginning of the study was 44 years (19-67
years) and 26 kg/m 2 (19-35 kg/m 2), respectively. The Canterbury Ethical Committee approved the study, and all participants gave informed consent. 2.2. Glucose references and test foods The glucose curves were measured every 3 months between April 2005 and May 2006, and the number of glucose curves obtained per person varied from 4 (n = 9) to 3 (n = 3) and 2 (n = 4). Measurements were carried out over the testing of a variety of different food products for a 12-month period thus the number of curves for each participant varied according to whether they participated in all testing periods or only participated in some of the trials. The primary limitation of studies of this nature is the retention of participants for long periods as availability is affected by changes in personal circumstances that cannot be predicted before commencement. The tests were designed to measure the blood glucose response for 2 to 3 hours after a test meal. Glucose beverages were supplied bottled as 75 g or 50 g of anhydrous glucose made up to 300 mL. The beverage containing 50 g of glucose was diluted with soda water in the ratio of 3:1 and 1:1 to make 300-mL drinks containing 12.5 and 25 g of glucose, respectively. This gave glucose beverages of 0 (soda water), 12.5, 25, 50, and 75 g. The GGEs for 10 foods were estimated from the 4 curves. The incremental area under the blood glucose curve (iAUC) for each of the foods was measured once for the 12-month study. Participants drank 250 mL of water with each of the foods. The foods measured were Sunreal muesli (100 g; Real Foods Ltd., Auckland, New Zealand) with Meadow Fresh skimmed milk (125 mL; Goodman Fielder, Auckland, New Zealand), Uncle Ben's parboiled rice (100 g; Masterfoods, Auckland, New Zealand), Kellogg's All Bran (50 g; Kelloggs Pty. Ltd., NSW, Australia) with Meadow Fresh skimmed milk (125 mL), Citrus Tree orange juice (250 mL), Ernest Adams fruit and nut loaf (100 g; Goodman Fielder), Ernest Adams apricot slice (100 g; Goodman Fielder), Uncle Toby's fruit twist muesli bars (100 g; Nestle New Zealand, Auckland, New Zealand), fresh parsnip (100 g), fresh pumpkin (100 g), and Wattie's frozen garden peas (100 g; Heinz Watties, Hastings, New Zealand). Not all participants were tested with all 10 foods; participants who contributed to each curve and whose data were subsequently used to calculate the GGEs are recorded in the tables. 2.3. Procedure The soda water and glucose beverages (0, 12.5, 25, 50, and 75 g) were each tested twice in each participant, at each period, with the exception of glucose time 3 when the 75 g of glucose was not measured. The decision was made to eliminate this test because none of the food products contains equivalent levels of carbohydrate thus rendering the results from this level of glucose not relevant to the study. The order in which the glucose references and foods were measured was randomly assigned and balanced so that
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
the glucose references were spread evenly over the course of the testing at each time point and the sequence for each individual varied. All tests were carried out at the Lipid and Diabetes Research Group facilities following a standard protocol. Participants were asked to eat a meal containing carbohydrates and to refrain from drinking alcohol the night before each test. The participants were asked to refrain from physical exercise on the morning of the test and to report to the clinic after having fasted. Capillary blood samples were taken using a lancet. A drop of blood was collected into a HemoCue cuvette, and blood glucose concentration was measured using a HemoCue glucose 201 analyzer (HemoCue, Helsingborg, Sweden). Each morning the instrument was checked using its own internal system to confirm that it was functioning correctly. This method uses glucose dehydrogenase and a micro spectrophotometric method for the determination of glucose [24]. A study has shown a very good correlation between the HemoCue analyzer and the Yellow Springs Instrument glucose oxidase analyzer (YSI 2300; Yellow Springs Instruments; Yellow Springs, OH, USA) [25]. The average of 2 fasting blood glucose concentrations, determined 5 minutes apart, was used as a baseline measure. Test products and glucose references were consumed within 15 minutes, and capillary blood samples were taken at 15, 30, 45, 60, 90, and 120 minutes after the person began consuming the test food or glucose reference. If the blood glucose concentration had not returned to within 0.2 mmol/L of the baseline concentration at 120 minutes, further blood samples were taken at 150 and 180 minutes after the start time. Participants were asked to remain seated for the duration of the tests with the exception of visits to the toilet. 2.4. Analysis of results 2.4.1. Calculation of iAUC The iAUC was calculated geometrically to a maximum of 180 minutes using the method described by Wolever and Jenkins [30] for each of the test foods for each participant. Areas where the curve dropped below baseline were excluded. 2.4.2. Calculation of slopes and intercepts for glucose curves The iAUCs for each of the 5 glucose concentrations at each time point were calculated as described previously. The iAUCs and the glucose concentrations (0, 12.5, 25, 50, and 75 g) were then log-transformed, and the slope and intercept of the iAUC vs glucose concentration curve determined. 2.4.3. Calculation of the GGE values of foods The curves describing the relationship between glucose beverage concentration and iAUC were used to estimate the glycemic response for all the foods. This was done by using the derived slope and intercepts from the individual's glucose-response curve to estimate the GGE from the individual iAUC for each food. The GGE for each food at
497
the specified portion size was then taken as the average GGE over the individuals for whom it was measured. 2.4.4. Statistical analysis The slopes and intercepts for each individual's glucoseresponse curve over the maximum of 4 times were tested for systematic changes using repeated measures analysis of variance [26]. A P b .05 was taken as statistically significant. In the absence of any clear systematic changes, the intraindividual coefficient of variation, reflecting “random” temporal variation over time, was calculated for the slope and the intercept. The impact of any changes in the slopes and intercepts on the actual GGE values was then determined by calculating the GGEs of a number of the foods using the available slopes and intercepts for each participant for whom the GGE of that food was measured. Given that GGEs are typically grouped into 3 categories rather than using the actual calculated value, the foods were assigned to 1 of the 3 categories, a value of 10 or lower being a low GGE, 10.01 to 19.99 being a medium GGE, and 20 or higher being a high GGE. The GGE category (low, medium, or high) of the average GGE of each food was compared between the 4 curves. 3. Results Table 1 shows the means and ranges for the slopes and intercepts calculated at the 4 time points. Although the average slopes and intercepts did not differ significantly over time (slopes F3,24 = 0.93; P = .44; intercepts F3,24 = 0.80; P = .51), the within-individual coefficient of variation representing the variability over the 4 times was 30% for the slopes and 40% for the intercepts, demonstrating considerable “random” variation in the glucose curves over time. Fig. 1A and B show the individual slopes and intercepts across the 4 glucose curves for all the individuals. Some individuals showed remarkably consistent results (participants 5, 7, 15, and 16), whereas others showed very variable results (participants 4, 10, and 14). Also evident in the figures is the considerable variation in the glucose-response curves between individuals. Table 2 shows the average GGE values for each of the 10 foods calculated at each time point using the glucose curve Table 1 Temporal variability in glucose reference curves for a 1-year period Glucose curve
No. of participants
Average slope (mmol/L per minute) (min, max)
Average intercept (mmol/L per minute) (min, max)
Time 1 Time 2 Time 3 Time 4
13 15 12 13
0.9 (0.5, 1.5) 0.9 (0.5, 1.2) 0.9 (0.4, 1.5) 0.7 (0.3, 1.1)
1.9 (0.2, 3.6) 2.0 (0.5, 3.1) 2.2 (0.2, 3.8) 2.6 (1.0, 3.8)
Average slopes and intercepts (mmol/L per minute) (including minimum and maximum) of the 4 glucose curves, across the participants who provided curves at each time point. Data are expressed as means with minimum and maximum values displayed.
498
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500 Table 3 Temporal variability in glucose reference curves for a 1-year period Glucose Curve 1 2 3 4
n
Mean
Standard Error
13 15 12 13
78.542⁎ 78.011 † 77.162 75.121
0.852 0.78 0.91 0.864
Standard glucose reference curves of 0, 12.5, 25, 50, and 75 g performed on 16 individuals for a 1-year period. Average body weight over the 4 glucose curves of the participants who provided curves at each time point. Body weight is shown as mean values ± SE. P ≤ .05 across pairwise comparison of times. ⁎ P = .009 significantly different to glucose curve 4. † P = .019 significantly different to glucose curve 4.
but one instance, despite some changes in the mean values over time, the assignment to the 3 categories did not vary across the 4 assessments. Body weight data over the course of the study are shown in Table 3. Over the course of the study, there was a significant change in body weight with curves 1 (P = .009) and 2 (P = .019) differing significantly from curve 4 where body weight was lower. During the course of the study, there were no significant physiologic changes in physical activity or medication regimens observed in any of the participants. No extreme changes were made to any dietary regimens or daily lifestyles. 4. Discussion Fig. 1. Temporal variability in glucose reference curves for a 1-year period. Standard glucose reference curves of 0, 12.5, 25, 50, and 75 g performed on 16 individuals for a 1-year period. Changes in the glucose curve intercept (mmol/L per minute) (A) and in the glucose curve slope (mmol/L per minute) (B) for the 4 glucose curves for each individual. Values represent the mean of total number of individuals tested at each period.
for that time and the average of these 4 values. The assignment of a food to a GGE category is also shown. In all
This study showed some variability in the blood glucoseresponse curve for individuals for the 12 months of the study, with coefficients of variation of 30% for the slope of the glucose-response curve and 40% for the intercepts. This variability is consistent with the variability shown in individuals when measuring glycemic response from day to day, which is approximately 30% as demonstrated by our group and others [27,28]. The reason for this is that blood glucose response is affected by a multitude of factors that are
Table 2 GGE values and category classification for foods tested over a 1-year period Food
Ernst Adams apricot slice (100 g) Kellogg's All Bran (50 g) and 125-mL skimmed milk Ernst Adams fruit and nut loaf (100 g) Uncle Toby's fruit twist muesli bar (100 g) Sunreal muesli (100 g) and 125-mL skimmed milk Citrus Tree orange juice (250 mL) Parsnips, microwaved (100 g) Wattie's frozen garden peas (100 g) Pumpkin, microwaved (100 g) Uncle Ben's parboiled rice (100 g)
Average GGE value (mmol/L per minute) Curve 1
Curve 2
Curve 3
Curve 4
Average
12.3 M (n = 9) 16.1 M (n = 12) 13.9 M (n = 9) 29.9 H (n = 10) 12.4 M (n = 9) 11.2 M (n = 12) 6.4 L (n = 9) 3.6 L (n = 9) 4.8 L (n = 9) 19.5 M (n = 9)
12.8 M (n = 12) 15.9 M (n = 15) 18.0 M (n = 12) 30.5 H (n = 13) 12.0 M (n = 8) 12.2 M (n = 15) 6.2 L (n = 12) 3.8 L (n = 12) 4.0 L (n = 12) 18.5 M (n = 9)
9.6 L (n = 12) 12.9 M (n = 12) 11.8 M (n = 12) 22.1 H (n = 12) 10.6 M (n = 6) 10.2 M (n = 12) 5.1 L (n = 12) 3.1 L (n = 12) 3.6 L (n = 12) 19.9 M (n = 7)
12.7 M (n = 12) 15.4 M (n = 12) 16.6 M (n = 12) 31.6 H (n = 12) 12.0 M (n = 7) 11.3 M (n = 12) 5.4 L (n = 12) 3.2 L (n = 12) 3.5 L (n = 12) 18.8 M (n = 8)
11.8 M 15.1 M 15.2 M 28.5 H 11.8 M 11.3 M 5.7 L 3.4 L 3.9 L 19.1 M
Temporal variability in glucose reference curves for a 1-year period. Average GGE values and the classification into low, medium, and high GGE for the individuals who tested each food based on the 2 to 4 different glucose curves obtained for each individual. The number of individuals tested for that food at that time point is indicated in parenthesis. Data are expressed as mean values. Ranking of food into GGE category as follows: L = low GGE, ≤10; M = medium GGE, 10.01-19.99; and H = high GGE ≥20.
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
inherently variable from day to day, such as what the person ate the previous evening [29,30], the fitness of the participant [31], the amount of physical activity on the days preceding the test meal [32], and the exact length of the overnight fast [33]. All of these factors will affect the overall reproducibility of the glucose-response curve for longer periods. Variation in body weight may also affect an individual's response to glucose with weight loss being shown to increase insulin sensitivity and improve glucose tolerance [34]. For the course of this study, it was shown that participant's weight fell significantly for the period of glucose curve 4; however, it appears to have not had a significant effect on mean GGE values of the foods measured for the 1-year period. Despite this, there are limited data available on the impact of physiologic changes occurring in test subjects for the course of a testing period for their glycemic response to foods. As such, it is recommended that at least one test of the reference food should be performed at a 3-monthly interval to improve reliability of the GGE values obtained [23]. In conclusion, day-to-day variation in an individual's response to glucose produces differences of up to 30%. To provide accurate measurement of the GGE value of a food, this variability in the response to a standard glucose curve over time means that it is necessary to repeat the glucose standard curve for each individual regularly. This variability may also increase if there are more significant changes in the physiologic or pathophysiologic status of the test subjects; however, data to support this are lacking, and precise tracking of these parameters may prove to be a burden both timewise and financially for participants and investigators. For practicality and cost purposes, we recommend that the curve needs to be repeated every 3 months. However, if the GGE is being measured only for identifying which category a food falls into, that is, if the food has a high, medium, or low GGE, without assigning an exact value to the food, then it may be possible to use the glucose standard curve for a longer period of up to 1 year. Acknowledgment This study was funded by the New Zealand Foundation for Research, Science, and Technology (FfRST; Wellington, New Zealand) (contract C02X0401) and the Lifestyle Foods (NZ Institute for Crop and Food Research, Christchurch, New Zealand) industry partners. The authors thank the participants for their time. References [1] Brand-Miller J, Hayne S, Petocz P, Colagiuri S. Low-glycemic index diets in the management of diabetes. Diabetes Care 2003;26: 2261-7. [2] Brand-Miller J. Postprandial glycemia, glycemic index and the prevention of type 2 diabetes. Am J Clin Nutr 2004;80:243-4. [3] Salmeron J, Manson JE, Stampfer MJ, Colditz GA, Wing AL, Willett W. Dietary fiber, glycemic load, and risk of non-insulin dependent diabetes mellitus in women. JAMA 1997;277:472-7.
499
[4] Salmeron J, Ascherio A, Rimm EB, et al. Dietary fibre, glycemic load, and risk of NIDDM in men. Diabetes Care 1997;20:545-50. [5] Meyer KA, Kushi LH, Jacobs DR, Slavin J, Sellers TA, Folsom AR. Carbohydrates, dietary fiber, and incident type 2 diabetes in older women. Am J Clin Nutr 2000;71:921-30. [6] Liu SM, Willett WC, Stampfer MJ, et al. A prospective study of dietary glycemic load, carbohydrate intake, and risk of coronary heart disease in US women. Am J Clin Nutr 2000;71:1455-61. [7] Liu S, Manson JE, Stampfer MJ, et al. Dietary glycemic load assessed by food frequency questionnaire in relation to plasma high-densitylipoprotein cholesterol and fasting plasma triacylglycerols in postmenopausal women. Am J Clin Nutr 2001;73:560-6. [8] Frost G, Leeds AR, Dore D, Madeiros S, Brading S, Dornhurst A. Glycemic index as a determinant of serum HDL-cholesterol concentration. Lancet 1999;353:1045-8. [9] Burke LM, Collier GR, Hargreaves M. Glycemic index-a new tool for sports nutrition. Int J Sports Nutr 1998;8:401-15. [10] Walton P, Rhodes EC. Glycemic index and optimal performance. Sports Med 1997;23:164-72. [11] Speith LE, Harnish JD, Lenders CM, et al. A low glycemic-index diet in the treatment of pediatric obesity. Arch Pediatr Adolesc Med 2000; 154:947-51. [12] Ludwig DS. Dietary glycemic index and obesity. J Nutr 2000;130: 280S-283S. [13] Ludwig DS, Majzoub JA, Al-Zahrani A, Dallal GE, Blanco I, Roberts SB. High glycemic index foods, overeating and obesity. Pediatrics 1999;103:E261-6. [14] Ebbeling CB, Leidig MM, Sinclair KB, Hangen JP, Ludwig DS. A reduced-glycemic load diet in the treatment of adolescent obesity. Arch Pediatr Adolesc Med 2003;157:773-9. [15] Brand-Miller JC, Holt SHA, Pawlak DB, McMillan J. Glycemic index and obesity. Am J Clin Nutr 2002;76:281S-285S. [16] Monro JA. Redefining the glycemic index for dietary management of postprandial glycemia. Am J Clin Nutr 2003;133:4256-8. [17] Monro JA. Glycemic index-based classifications do not reflect the relative glycemic impact of foods. Proc Nutr Soc NZ 2001;24:96-103. [18] Monro JA. Food data for glycemic control. Food Sci Central 2003: 1-12. [19] Monro JA, Williams M. Concurrent management of postprandial glycemia and nutrient intake using glycemic glucose equivalents, food composition data and computer-assisted meal design. Asia Pac J Clin Nutr 2000;9:67-73. [20] Monro JA. Glycemic glucose equivalent: combining carbohydrate content, quantity and glycemic index of foods for precision in glycemia management. Asia Pac J Clin Nutr 2002;11:217-25. [21] Monro J. Assessing the glycemic potency of foods and its relevance to industry and the consumer. Int Rev Food Sci Technol 2005/2006 2006: 60-5. [22] Venn BJ, Wallace AJ, Monro JA, et al. The glycemic load estimated from glycemic index does not differ greatly from that measured using a standard curve in healthy volunteers. J Nutr 2006;136: 1377-81. [23] Brouns F, Bjorck I, Frayn K, Lang V, Slama G, Wolever TMS. Glycemic index methodology. Nutrition Res Rev 2005;18:147-71. [24] Banauch D, Brummer W, Ebeling W, Metz H, Rindfrey H, Lang H, et al. A glucose dehydrogenase for the determination of glucose concentrations in body fluids. Z Klin Chem Klin Biochem 1975;13: 101-7. [25] Stork ADM, Kemperman H, Erkelens DW, Veneman TF. Comparison of the accuracy of the HemoCue glucose analyzer with the Yellow Springs Instrument glucose oxidase analyzer, particularly in hypoglycemia. Eur J Endocrin 2005;153:275-81. [26] Bland M. An introduction to medical statistics. 2nd ed. New York: Oxford University Press; 1995. [27] Wallace AJ, Willis JA, Monro JA, Frampton CM, Hedderley DI, Scott RS. No difference between venous and capillary blood sampling and the Minimed continuous glucose monitoring system
500
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
for determining the blood glucose response to food. Nutr Res 2006; 26:403-8. [28] Wolever TMS, Vorster HH, Bjorck I, et al. Determination of the glycemic index of foods: interlaboratory study. Eur J Clin Nutr 2003; 57:475-82. [29] Robertson MD, Henderson RA, Vist GE, Rumsey RDE. Extended effect of evening meal carbohydrate/fat ratio on fasting and postprandial substrate metabolism. Am J Clin Nutr 2002;75:505-10. [30] Wolever TM, Jenkins DJ, Ocana AM, Rao VA, Collier GR. Secondmeal effect: low glycemic-index foods eaten at dinner improve subsequent breakfast glycemic response. Am J Clin Nutr 1988;48: 1041-7.
[31] Mettler S, Lamprecht-Rusca F, Stoffel-Kurt N, Wenk C, Colombani PC. The influence of the subjects' training state on the glycemic index. Eur J Clin Nutr 2007;61:19-24. [32] Malkova D, Evans RD, Frayn KN, Humphreys SM, Jones PR, Hardman AE. Prior exercise and postprandial substrate extraction across the human leg. Am J Physiol Endocrin Metab 2000;279:E1020-8. [33] Klein S, Sakurai Y, Romijn JA, Carroll RM. Progressive alterations in lipid and glucose metabolism during short-term fasting in young adult men. Am J Physiol 1993;265:E801-6. [34] Schäfer S, Kantartzis K, Machann J, Venter C, Niess A, Schick F, et al. Lifestyle intervention in individuals with normal versus impaired glucose tolerance. Eur J Clin Invest 2007;37(7):535-43.
Nutrition Research 28 (2008) 495 – 500 www.nrjournal.com
Considerable temporal variability in glucose reference curves in humans for a year period Alison J. Wallace a,⁎, Sarah L. Eady a , Russell S. Scott b , Jinny A. Willis b , Christopher M. Frampton c a
New Zealand Institute for Crop and Food Research Limited, Private Bag 4704, Christchurch, New Zealand b Lipid and Diabetes Research Group, PO Box 4710, Christchurch 8011, New Zealand c Christchurch School of Medicine and Health Sciences, PO Box 4345, Christchurch 8011, New Zealand Received 29 November 2007; revised 15 May 2008; accepted 16 May 2008
Abstract Glycemic glucose equivalent (GGE) is a measure of the blood glucose response to a defined portion of food. Their calculation requires the measurement of a standard glucose-response curve, with beverages containing 0, 12.5, 25, 50, and 75 g of glucose measured twice each. This study was designed to determine the stability of an individual's glucose-response curve measured every 3 months for a year and of their GGE estimates for 10 foods for that period. The blood glucose response to beverages containing 0, 12.5, 25, 50, and 75 g glucose and to 10 foods was measured for 16 healthy individuals. Capillary blood samples were collected fasting, then every 15 minutes for 1 hour, and every 30 minutes for at least 2 hours. The slopes and intercepts of the 4 glucose curves and the GGE of the 10 foods calculated using the available curves for each food was compared. The results showed considerable temporal variability in the slope (intraindividual coefficient of variation (CV) = 30%) and intercept (intraindividual CV = 40%) of the glucose curves. However, if GGE values were categorized into 3 groups (low GGE, ≤10; medium GGE, 10.01-19.99; and high GGE, ≥20), all but one food was consistently classified in the same category across the 4 glucose curves. In conclusion, it appears that if the exact GGE value is required, glucose curves should be repeated at least once every 3 months, but if foods are classed into general GGE categories, it may be possible to use the same glucose curve for a longer period. © 2008 Elsevier Inc. All rights reserved. Keywords: Abbreviations:
Glycemia; Glycemic index; Glycemic glucose equivalents; Reproducibility; Humans GGE, glycemic glucose equivalents; GI, glycemic index; WHO, World Health Organization; iAUC, incremental area under the curve.
1. Introduction Interest in the effect of particular foods on blood glucose levels and the possible contribution of glycemic properties to nutrition and health has increased with the realization of the potential outcomes on people with diabetes [1,2], on those at risk for developing diabetes and coronary heart disease [3-8], on performance in sports people [9,10], and on weight loss [11-15]. Traditionally, blood glucose response has been ⁎ Corresponding author. Tel.: +64 3 325 9638; fax: +64 3 325 2074. E-mail address: [email protected] (A.J. Wallace). 0271-5317/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.nutres.2008.05.006
measured using the glycemic index (GI), which is a glycemic response to a portion of food containing a defined amount of available carbohydrate (usually 50 g), expressed as a percentage of the response to the same amount of glucose. For foods containing the same amount of carbohydrate, the GI indicates what effect the food will have on an individual's blood glucose levels. However, because foods contain different amounts of carbohydrate, ranking foods by GI will not necessarily rank them according to the effect they have on blood glucose levels [16-18]. For this reason, the glycemic glucose equivalent (GGE) has been developed. Glycemic glucose equivalent is defined as the amount of
496
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
glucose in grams that would induce a glycemic response equal to that induced by a specified weight of a food [19,20]. For example, if a portion of food has a GGE of 6 g, it induces the same glycemic response as 6 g of glucose. Glycemic glucose response is food-based and equiglycemic in measurement, whereas GI is carbohydrate-based and equicarbohydrate in measurement [21]. The most accurate way to measure GGEs for a wide variety of foods with varying GGE values is to measure them relative to a standard glucose reference curve [22]. Glucose references of 0, 12.5, 25, 50, and 75 g are measured on more than one occasion in each individual. Brouns et al [23] published an extensive review on GI methodology, in which they suggested that the same reference glucose tests could only be used for a maximum of 3 months because no data are available on the impact of time on changes in response to the reference food. An individual's blood glucose response could vary over time, and therefore, because all foods are measured relative to the glucose reference, this could result in an incorrect GGE value being obtained. If this variation is systematic, this may not matter for the measurement of GGE values for foods because the glucose reference and food may change proportionally. Also, as noted by Brouns et al [23], no one has investigated how an individual's blood glucose response varies over time, and if their physiologic state remains the same (ie, no significant changes in medication, weight, or physical activity), their blood glucose-response curve may not vary much over time. Reliable estimation of GGE is essential if the concept is to be used accurately to describe a food's glycemic properties and its impact on health outcomes. This would then enable consumers to guide their food selections appropriately to lower the overall glycemic impact of their dietary intake assisting with improved blood glucose control, body weight management, and reduced cardiovascular and diabetic risk. The aim of this study was to examine the stability of an individual's glucose response using a standard glucose curve measured every 3 months for 1 year and identify the effect of any physiologic changes in an individual that may impact upon a reliable reference value being obtained for estimation of GGE values. The GGE values for a variety of foods for this period are also established. 2. Methods and materials 2.1. Participants Sixteen individuals (10 women and 6 men) were tested for a year. Guidelines for GI methodology state that a minimum of 10 participants in these types of studies provides a sufficient degree of power and precision to the analysis (23). None of the participants had diabetes according to the WHO classification (fasting glucose b6.2 mmol/L). The mean (range) age and body mass index of the group at the beginning of the study was 44 years (19-67
years) and 26 kg/m 2 (19-35 kg/m 2), respectively. The Canterbury Ethical Committee approved the study, and all participants gave informed consent. 2.2. Glucose references and test foods The glucose curves were measured every 3 months between April 2005 and May 2006, and the number of glucose curves obtained per person varied from 4 (n = 9) to 3 (n = 3) and 2 (n = 4). Measurements were carried out over the testing of a variety of different food products for a 12-month period thus the number of curves for each participant varied according to whether they participated in all testing periods or only participated in some of the trials. The primary limitation of studies of this nature is the retention of participants for long periods as availability is affected by changes in personal circumstances that cannot be predicted before commencement. The tests were designed to measure the blood glucose response for 2 to 3 hours after a test meal. Glucose beverages were supplied bottled as 75 g or 50 g of anhydrous glucose made up to 300 mL. The beverage containing 50 g of glucose was diluted with soda water in the ratio of 3:1 and 1:1 to make 300-mL drinks containing 12.5 and 25 g of glucose, respectively. This gave glucose beverages of 0 (soda water), 12.5, 25, 50, and 75 g. The GGEs for 10 foods were estimated from the 4 curves. The incremental area under the blood glucose curve (iAUC) for each of the foods was measured once for the 12-month study. Participants drank 250 mL of water with each of the foods. The foods measured were Sunreal muesli (100 g; Real Foods Ltd., Auckland, New Zealand) with Meadow Fresh skimmed milk (125 mL; Goodman Fielder, Auckland, New Zealand), Uncle Ben's parboiled rice (100 g; Masterfoods, Auckland, New Zealand), Kellogg's All Bran (50 g; Kelloggs Pty. Ltd., NSW, Australia) with Meadow Fresh skimmed milk (125 mL), Citrus Tree orange juice (250 mL), Ernest Adams fruit and nut loaf (100 g; Goodman Fielder), Ernest Adams apricot slice (100 g; Goodman Fielder), Uncle Toby's fruit twist muesli bars (100 g; Nestle New Zealand, Auckland, New Zealand), fresh parsnip (100 g), fresh pumpkin (100 g), and Wattie's frozen garden peas (100 g; Heinz Watties, Hastings, New Zealand). Not all participants were tested with all 10 foods; participants who contributed to each curve and whose data were subsequently used to calculate the GGEs are recorded in the tables. 2.3. Procedure The soda water and glucose beverages (0, 12.5, 25, 50, and 75 g) were each tested twice in each participant, at each period, with the exception of glucose time 3 when the 75 g of glucose was not measured. The decision was made to eliminate this test because none of the food products contains equivalent levels of carbohydrate thus rendering the results from this level of glucose not relevant to the study. The order in which the glucose references and foods were measured was randomly assigned and balanced so that
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
the glucose references were spread evenly over the course of the testing at each time point and the sequence for each individual varied. All tests were carried out at the Lipid and Diabetes Research Group facilities following a standard protocol. Participants were asked to eat a meal containing carbohydrates and to refrain from drinking alcohol the night before each test. The participants were asked to refrain from physical exercise on the morning of the test and to report to the clinic after having fasted. Capillary blood samples were taken using a lancet. A drop of blood was collected into a HemoCue cuvette, and blood glucose concentration was measured using a HemoCue glucose 201 analyzer (HemoCue, Helsingborg, Sweden). Each morning the instrument was checked using its own internal system to confirm that it was functioning correctly. This method uses glucose dehydrogenase and a micro spectrophotometric method for the determination of glucose [24]. A study has shown a very good correlation between the HemoCue analyzer and the Yellow Springs Instrument glucose oxidase analyzer (YSI 2300; Yellow Springs Instruments; Yellow Springs, OH, USA) [25]. The average of 2 fasting blood glucose concentrations, determined 5 minutes apart, was used as a baseline measure. Test products and glucose references were consumed within 15 minutes, and capillary blood samples were taken at 15, 30, 45, 60, 90, and 120 minutes after the person began consuming the test food or glucose reference. If the blood glucose concentration had not returned to within 0.2 mmol/L of the baseline concentration at 120 minutes, further blood samples were taken at 150 and 180 minutes after the start time. Participants were asked to remain seated for the duration of the tests with the exception of visits to the toilet. 2.4. Analysis of results 2.4.1. Calculation of iAUC The iAUC was calculated geometrically to a maximum of 180 minutes using the method described by Wolever and Jenkins [30] for each of the test foods for each participant. Areas where the curve dropped below baseline were excluded. 2.4.2. Calculation of slopes and intercepts for glucose curves The iAUCs for each of the 5 glucose concentrations at each time point were calculated as described previously. The iAUCs and the glucose concentrations (0, 12.5, 25, 50, and 75 g) were then log-transformed, and the slope and intercept of the iAUC vs glucose concentration curve determined. 2.4.3. Calculation of the GGE values of foods The curves describing the relationship between glucose beverage concentration and iAUC were used to estimate the glycemic response for all the foods. This was done by using the derived slope and intercepts from the individual's glucose-response curve to estimate the GGE from the individual iAUC for each food. The GGE for each food at
497
the specified portion size was then taken as the average GGE over the individuals for whom it was measured. 2.4.4. Statistical analysis The slopes and intercepts for each individual's glucoseresponse curve over the maximum of 4 times were tested for systematic changes using repeated measures analysis of variance [26]. A P b .05 was taken as statistically significant. In the absence of any clear systematic changes, the intraindividual coefficient of variation, reflecting “random” temporal variation over time, was calculated for the slope and the intercept. The impact of any changes in the slopes and intercepts on the actual GGE values was then determined by calculating the GGEs of a number of the foods using the available slopes and intercepts for each participant for whom the GGE of that food was measured. Given that GGEs are typically grouped into 3 categories rather than using the actual calculated value, the foods were assigned to 1 of the 3 categories, a value of 10 or lower being a low GGE, 10.01 to 19.99 being a medium GGE, and 20 or higher being a high GGE. The GGE category (low, medium, or high) of the average GGE of each food was compared between the 4 curves. 3. Results Table 1 shows the means and ranges for the slopes and intercepts calculated at the 4 time points. Although the average slopes and intercepts did not differ significantly over time (slopes F3,24 = 0.93; P = .44; intercepts F3,24 = 0.80; P = .51), the within-individual coefficient of variation representing the variability over the 4 times was 30% for the slopes and 40% for the intercepts, demonstrating considerable “random” variation in the glucose curves over time. Fig. 1A and B show the individual slopes and intercepts across the 4 glucose curves for all the individuals. Some individuals showed remarkably consistent results (participants 5, 7, 15, and 16), whereas others showed very variable results (participants 4, 10, and 14). Also evident in the figures is the considerable variation in the glucose-response curves between individuals. Table 2 shows the average GGE values for each of the 10 foods calculated at each time point using the glucose curve Table 1 Temporal variability in glucose reference curves for a 1-year period Glucose curve
No. of participants
Average slope (mmol/L per minute) (min, max)
Average intercept (mmol/L per minute) (min, max)
Time 1 Time 2 Time 3 Time 4
13 15 12 13
0.9 (0.5, 1.5) 0.9 (0.5, 1.2) 0.9 (0.4, 1.5) 0.7 (0.3, 1.1)
1.9 (0.2, 3.6) 2.0 (0.5, 3.1) 2.2 (0.2, 3.8) 2.6 (1.0, 3.8)
Average slopes and intercepts (mmol/L per minute) (including minimum and maximum) of the 4 glucose curves, across the participants who provided curves at each time point. Data are expressed as means with minimum and maximum values displayed.
498
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500 Table 3 Temporal variability in glucose reference curves for a 1-year period Glucose Curve 1 2 3 4
n
Mean
Standard Error
13 15 12 13
78.542⁎ 78.011 † 77.162 75.121
0.852 0.78 0.91 0.864
Standard glucose reference curves of 0, 12.5, 25, 50, and 75 g performed on 16 individuals for a 1-year period. Average body weight over the 4 glucose curves of the participants who provided curves at each time point. Body weight is shown as mean values ± SE. P ≤ .05 across pairwise comparison of times. ⁎ P = .009 significantly different to glucose curve 4. † P = .019 significantly different to glucose curve 4.
but one instance, despite some changes in the mean values over time, the assignment to the 3 categories did not vary across the 4 assessments. Body weight data over the course of the study are shown in Table 3. Over the course of the study, there was a significant change in body weight with curves 1 (P = .009) and 2 (P = .019) differing significantly from curve 4 where body weight was lower. During the course of the study, there were no significant physiologic changes in physical activity or medication regimens observed in any of the participants. No extreme changes were made to any dietary regimens or daily lifestyles. 4. Discussion Fig. 1. Temporal variability in glucose reference curves for a 1-year period. Standard glucose reference curves of 0, 12.5, 25, 50, and 75 g performed on 16 individuals for a 1-year period. Changes in the glucose curve intercept (mmol/L per minute) (A) and in the glucose curve slope (mmol/L per minute) (B) for the 4 glucose curves for each individual. Values represent the mean of total number of individuals tested at each period.
for that time and the average of these 4 values. The assignment of a food to a GGE category is also shown. In all
This study showed some variability in the blood glucoseresponse curve for individuals for the 12 months of the study, with coefficients of variation of 30% for the slope of the glucose-response curve and 40% for the intercepts. This variability is consistent with the variability shown in individuals when measuring glycemic response from day to day, which is approximately 30% as demonstrated by our group and others [27,28]. The reason for this is that blood glucose response is affected by a multitude of factors that are
Table 2 GGE values and category classification for foods tested over a 1-year period Food
Ernst Adams apricot slice (100 g) Kellogg's All Bran (50 g) and 125-mL skimmed milk Ernst Adams fruit and nut loaf (100 g) Uncle Toby's fruit twist muesli bar (100 g) Sunreal muesli (100 g) and 125-mL skimmed milk Citrus Tree orange juice (250 mL) Parsnips, microwaved (100 g) Wattie's frozen garden peas (100 g) Pumpkin, microwaved (100 g) Uncle Ben's parboiled rice (100 g)
Average GGE value (mmol/L per minute) Curve 1
Curve 2
Curve 3
Curve 4
Average
12.3 M (n = 9) 16.1 M (n = 12) 13.9 M (n = 9) 29.9 H (n = 10) 12.4 M (n = 9) 11.2 M (n = 12) 6.4 L (n = 9) 3.6 L (n = 9) 4.8 L (n = 9) 19.5 M (n = 9)
12.8 M (n = 12) 15.9 M (n = 15) 18.0 M (n = 12) 30.5 H (n = 13) 12.0 M (n = 8) 12.2 M (n = 15) 6.2 L (n = 12) 3.8 L (n = 12) 4.0 L (n = 12) 18.5 M (n = 9)
9.6 L (n = 12) 12.9 M (n = 12) 11.8 M (n = 12) 22.1 H (n = 12) 10.6 M (n = 6) 10.2 M (n = 12) 5.1 L (n = 12) 3.1 L (n = 12) 3.6 L (n = 12) 19.9 M (n = 7)
12.7 M (n = 12) 15.4 M (n = 12) 16.6 M (n = 12) 31.6 H (n = 12) 12.0 M (n = 7) 11.3 M (n = 12) 5.4 L (n = 12) 3.2 L (n = 12) 3.5 L (n = 12) 18.8 M (n = 8)
11.8 M 15.1 M 15.2 M 28.5 H 11.8 M 11.3 M 5.7 L 3.4 L 3.9 L 19.1 M
Temporal variability in glucose reference curves for a 1-year period. Average GGE values and the classification into low, medium, and high GGE for the individuals who tested each food based on the 2 to 4 different glucose curves obtained for each individual. The number of individuals tested for that food at that time point is indicated in parenthesis. Data are expressed as mean values. Ranking of food into GGE category as follows: L = low GGE, ≤10; M = medium GGE, 10.01-19.99; and H = high GGE ≥20.
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
inherently variable from day to day, such as what the person ate the previous evening [29,30], the fitness of the participant [31], the amount of physical activity on the days preceding the test meal [32], and the exact length of the overnight fast [33]. All of these factors will affect the overall reproducibility of the glucose-response curve for longer periods. Variation in body weight may also affect an individual's response to glucose with weight loss being shown to increase insulin sensitivity and improve glucose tolerance [34]. For the course of this study, it was shown that participant's weight fell significantly for the period of glucose curve 4; however, it appears to have not had a significant effect on mean GGE values of the foods measured for the 1-year period. Despite this, there are limited data available on the impact of physiologic changes occurring in test subjects for the course of a testing period for their glycemic response to foods. As such, it is recommended that at least one test of the reference food should be performed at a 3-monthly interval to improve reliability of the GGE values obtained [23]. In conclusion, day-to-day variation in an individual's response to glucose produces differences of up to 30%. To provide accurate measurement of the GGE value of a food, this variability in the response to a standard glucose curve over time means that it is necessary to repeat the glucose standard curve for each individual regularly. This variability may also increase if there are more significant changes in the physiologic or pathophysiologic status of the test subjects; however, data to support this are lacking, and precise tracking of these parameters may prove to be a burden both timewise and financially for participants and investigators. For practicality and cost purposes, we recommend that the curve needs to be repeated every 3 months. However, if the GGE is being measured only for identifying which category a food falls into, that is, if the food has a high, medium, or low GGE, without assigning an exact value to the food, then it may be possible to use the glucose standard curve for a longer period of up to 1 year. Acknowledgment This study was funded by the New Zealand Foundation for Research, Science, and Technology (FfRST; Wellington, New Zealand) (contract C02X0401) and the Lifestyle Foods (NZ Institute for Crop and Food Research, Christchurch, New Zealand) industry partners. The authors thank the participants for their time. References [1] Brand-Miller J, Hayne S, Petocz P, Colagiuri S. Low-glycemic index diets in the management of diabetes. Diabetes Care 2003;26: 2261-7. [2] Brand-Miller J. Postprandial glycemia, glycemic index and the prevention of type 2 diabetes. Am J Clin Nutr 2004;80:243-4. [3] Salmeron J, Manson JE, Stampfer MJ, Colditz GA, Wing AL, Willett W. Dietary fiber, glycemic load, and risk of non-insulin dependent diabetes mellitus in women. JAMA 1997;277:472-7.
499
[4] Salmeron J, Ascherio A, Rimm EB, et al. Dietary fibre, glycemic load, and risk of NIDDM in men. Diabetes Care 1997;20:545-50. [5] Meyer KA, Kushi LH, Jacobs DR, Slavin J, Sellers TA, Folsom AR. Carbohydrates, dietary fiber, and incident type 2 diabetes in older women. Am J Clin Nutr 2000;71:921-30. [6] Liu SM, Willett WC, Stampfer MJ, et al. A prospective study of dietary glycemic load, carbohydrate intake, and risk of coronary heart disease in US women. Am J Clin Nutr 2000;71:1455-61. [7] Liu S, Manson JE, Stampfer MJ, et al. Dietary glycemic load assessed by food frequency questionnaire in relation to plasma high-densitylipoprotein cholesterol and fasting plasma triacylglycerols in postmenopausal women. Am J Clin Nutr 2001;73:560-6. [8] Frost G, Leeds AR, Dore D, Madeiros S, Brading S, Dornhurst A. Glycemic index as a determinant of serum HDL-cholesterol concentration. Lancet 1999;353:1045-8. [9] Burke LM, Collier GR, Hargreaves M. Glycemic index-a new tool for sports nutrition. Int J Sports Nutr 1998;8:401-15. [10] Walton P, Rhodes EC. Glycemic index and optimal performance. Sports Med 1997;23:164-72. [11] Speith LE, Harnish JD, Lenders CM, et al. A low glycemic-index diet in the treatment of pediatric obesity. Arch Pediatr Adolesc Med 2000; 154:947-51. [12] Ludwig DS. Dietary glycemic index and obesity. J Nutr 2000;130: 280S-283S. [13] Ludwig DS, Majzoub JA, Al-Zahrani A, Dallal GE, Blanco I, Roberts SB. High glycemic index foods, overeating and obesity. Pediatrics 1999;103:E261-6. [14] Ebbeling CB, Leidig MM, Sinclair KB, Hangen JP, Ludwig DS. A reduced-glycemic load diet in the treatment of adolescent obesity. Arch Pediatr Adolesc Med 2003;157:773-9. [15] Brand-Miller JC, Holt SHA, Pawlak DB, McMillan J. Glycemic index and obesity. Am J Clin Nutr 2002;76:281S-285S. [16] Monro JA. Redefining the glycemic index for dietary management of postprandial glycemia. Am J Clin Nutr 2003;133:4256-8. [17] Monro JA. Glycemic index-based classifications do not reflect the relative glycemic impact of foods. Proc Nutr Soc NZ 2001;24:96-103. [18] Monro JA. Food data for glycemic control. Food Sci Central 2003: 1-12. [19] Monro JA, Williams M. Concurrent management of postprandial glycemia and nutrient intake using glycemic glucose equivalents, food composition data and computer-assisted meal design. Asia Pac J Clin Nutr 2000;9:67-73. [20] Monro JA. Glycemic glucose equivalent: combining carbohydrate content, quantity and glycemic index of foods for precision in glycemia management. Asia Pac J Clin Nutr 2002;11:217-25. [21] Monro J. Assessing the glycemic potency of foods and its relevance to industry and the consumer. Int Rev Food Sci Technol 2005/2006 2006: 60-5. [22] Venn BJ, Wallace AJ, Monro JA, et al. The glycemic load estimated from glycemic index does not differ greatly from that measured using a standard curve in healthy volunteers. J Nutr 2006;136: 1377-81. [23] Brouns F, Bjorck I, Frayn K, Lang V, Slama G, Wolever TMS. Glycemic index methodology. Nutrition Res Rev 2005;18:147-71. [24] Banauch D, Brummer W, Ebeling W, Metz H, Rindfrey H, Lang H, et al. A glucose dehydrogenase for the determination of glucose concentrations in body fluids. Z Klin Chem Klin Biochem 1975;13: 101-7. [25] Stork ADM, Kemperman H, Erkelens DW, Veneman TF. Comparison of the accuracy of the HemoCue glucose analyzer with the Yellow Springs Instrument glucose oxidase analyzer, particularly in hypoglycemia. Eur J Endocrin 2005;153:275-81. [26] Bland M. An introduction to medical statistics. 2nd ed. New York: Oxford University Press; 1995. [27] Wallace AJ, Willis JA, Monro JA, Frampton CM, Hedderley DI, Scott RS. No difference between venous and capillary blood sampling and the Minimed continuous glucose monitoring system
500
A.J. Wallace et al. / Nutrition Research 28 (2008) 495–500
for determining the blood glucose response to food. Nutr Res 2006; 26:403-8. [28] Wolever TMS, Vorster HH, Bjorck I, et al. Determination of the glycemic index of foods: interlaboratory study. Eur J Clin Nutr 2003; 57:475-82. [29] Robertson MD, Henderson RA, Vist GE, Rumsey RDE. Extended effect of evening meal carbohydrate/fat ratio on fasting and postprandial substrate metabolism. Am J Clin Nutr 2002;75:505-10. [30] Wolever TM, Jenkins DJ, Ocana AM, Rao VA, Collier GR. Secondmeal effect: low glycemic-index foods eaten at dinner improve subsequent breakfast glycemic response. Am J Clin Nutr 1988;48: 1041-7.
[31] Mettler S, Lamprecht-Rusca F, Stoffel-Kurt N, Wenk C, Colombani PC. The influence of the subjects' training state on the glycemic index. Eur J Clin Nutr 2007;61:19-24. [32] Malkova D, Evans RD, Frayn KN, Humphreys SM, Jones PR, Hardman AE. Prior exercise and postprandial substrate extraction across the human leg. Am J Physiol Endocrin Metab 2000;279:E1020-8. [33] Klein S, Sakurai Y, Romijn JA, Carroll RM. Progressive alterations in lipid and glucose metabolism during short-term fasting in young adult men. Am J Physiol 1993;265:E801-6. [34] Schäfer S, Kantartzis K, Machann J, Venter C, Niess A, Schick F, et al. Lifestyle intervention in individuals with normal versus impaired glucose tolerance. Eur J Clin Invest 2007;37(7):535-43.