- Journal List
- J Nutr Sci
- v.3; 2014
- PMC4153099
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsem*nt of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more: PMC Disclaimer | PMC Copyright Notice
J Nutr Sci. 2014; 3: e1.
Published online 2014 Jan 7. doi:10.1017/jns.2013.33
PMCID: PMC4153099
PMID: 25191601
Amin Esfahani,1,2,3 Joanne Lam,1 and Cyril W. C. Kendall2,3,4,*
Author information Article notes Copyright and License information PMC Disclaimer
Abstract
Raisins are popular snacks with a favourable nutrient profile, being high in dietaryfibre, polyphenols and a number of vitamins and minerals, in addition to being rich infructose. In light of evidence demonstrating improvements in glycaemic control withmoderate fructose intake and low-glycaemic index (GI) fruits, our aim was to determine theGI, insulin index (II) and postprandial responses to raisins in an acute feeding setting.A total of ten healthy participants (four male and six female) consumed breakfast studymeals on four occasions over a 2- to 8-week period: meal 1: white bread (WB) (108g WB;50g available carbohydrate) served as the control and was consumed on two separateoccasions; meal 2: raisins (R50) (69g raisins; 50g available carbohydrate); and meal 3:raisins (R20) (one serving, 28g raisins; 20g available carbohydrate). Postprandialglucose and insulin were measured over a 2h period for the determination of GI, glycaemicload (GL) and II. The raisin meals, R50 and R20, resulted in significantly reducedpostprandial glucose and insulin responses when compared with WB(P<0·05). Furthermore, raisins were determined to be low-GI, -GLand -II foods. The favourable effect of raisins on postprandial glycaemic response, theirinsulin-sparing effect and low GI combined with their other metabolic benefits mayindicate that raisins are a healthy choice not only for the general population but alsofor individuals with diabetes or insulin resistance.
Key words: Raisins, Dried fruit, Glycaemic index, Glycaemic load
Abbreviations: GI, glycaemic index; GL, glycaemic load; iAUC, incremental AUC ; R20, raisins (20g available carbohydrate); R50, raisins (50g available carbohydrate); WB, white bread
Raisins are one of the most commonly consumed dried fruits, are eaten across the globe, andhave a unique nutrient profile that may confer distinctive health benefits when compared withother fruit. Raisins are a rich source of polyphenols and phenolic acids, which may serve asantioxidants and promote an anti-inflammatory environment with potential healthbenefits(,1–3). Raisins are also high in dietary fibre and prebiotics, such as inulin,which have been shown to produce a healthier colonic microflora profile in addition topossibly aiding weight management and reducing the risk of CVD(,4). A clinical study found that raisins as part of a healthy diet improvedblood lipids and reduced other risk factors for CVD(,5).
Raisins are also high in fructose, which has a low glycaemic index (GI). While concerns havebeen raised that fructose may have adverse metabolic effects and promote weight gain, a recentmeta-analysis(,6) demonstrated that moderate intakes of fructose may improve glycaemiccontrol, without harming cardiometabolic risk factors(,6). This is especially important in light of recent evidence demonstratingthat low-GI fruits may improve glycaemic and cardiovascular markers, including HbA1c and bloodpressure(,7).
Given that raisins are the most commonly consumed dried fruit, are high in fructose and thecontroversy surrounding the cardiometabolic effects of fructose, we investigated the effect ofraisins on postprandial glycaemia and insulinaemia in an acute feeding study.
Methods
Participants
Inclusion criteria included men or non-pregnant women aged 18–75 years who were in goodhealth. Individuals with a known history of AIDS, hepatitis, diabetes or a heartcondition, or individuals taking medication or with any condition that might makeparticipation dangerous to the individual or affect the results were excluded.
A total of ten participants were studied. Using the t distribution andassuming an average CV of within-individual variation of incremental AUC (iAUC) values of25 %, n 10 participants has 80 % power to detect a 33 % difference iniAUC with two-tailed P<0·05.
Protocol
The study was open-label with a partial randomised, cross-over design using standard GImethodology (ISO 26642:2010; International Organization for Standardization). Eligibleparticipants were studied on four separate days over a period of 2–8 weeks with aninterval of no less than 40h and no more than 2 weeks between tests. On each test day,participants came to the clinic in the morning after a 10–14h overnight fast.Participants were asked to maintain stable dietary and activity habits throughout theirparticipation in the study. If any participant was not feeling well or had not compliedwith the preceding experimental conditions, the test was not carried out and wasrescheduled for another day. On each test occasion participants were weighed, and twofasting blood samples were obtained by finger-stick at 5-min intervals. Finger-stick bloodsamples were collected from hands warmed with an electric heating pad for 3–5min beforeeach sample. Blood samples were collected into two separate vials: one (two or three dropsof blood) for glucose analysis and the other (between six and eight drops of blood) forinsulin. After the second fasting sample was collected the participant was provided withthe test meal. At the first bite, a timer was started and additional blood samples weretaken at 15, 30, 45, 60, 90 and 120min. Before and during the test, a blood glucose testrecord was filled out with the participant's initials, identification number, date, bodyweight, test meal, beverage, time of starting to eat, time it took to eat, time andcomposition of last meal, and any unusual activities. During the 2h test, participantsremained seated quietly. After the last blood sample had been obtained participants wereoffered a snack and then allowed to leave.
The present study was conducted according to the guidelines laid down in the Declarationof Helsinki and all procedures involving human subjects/patients were approved by theWestern Institutional Review Board®. Written informed consent was obtained fromall participants before the start of the study.
Study meals
Each participant participated in a total of four breakfast study meals. Two test mealswere consumed: meal 1: R50, consisting of 50g available carbohydrate from raisins; andmeal 2: R20, consisting of 20g available carbohydrate from raisins, which is one standardserving (28g) of raisins. The control white bread (WB) meal, which provided 50gavailable carbohydrate, was consumed twice. The macronutrient profiles of the study mealsare provided in Table 1. The order of the testmeals was randomised.
Table 1.
Nutrient content of test meals
| Test meal | Abbreviation | Amount (g) | Protein (g) | Fat (g) | Total CHO (g) | Dietary fibre (g) | Available CHO (g) |
|---|---|---|---|---|---|---|---|
| White bread | WB | 108·0 | 9·3 | 0·8 | 52·0 | 2·0 | 50·0 |
| Raisins (50g CHO) | R50 | 69·0 | 1·7 | 0 | 53·4 | 3·4 | 50·0 |
| Raisins (one serving) | R20 | 28·0 | 0·7 | 0 | 21·7 | 1·4 | 20·3 |
Open in a separate window
CHO, carbohydrate.
Palatability
After consuming a meal, participants rated its palatability using a visual analogue scaleanchored at very ‘unpalatable’ at one end (0) and ‘very palatable’ at the other (100).Therefore, the higher the number, the higher was the perceived palatability of theproduct.
Blood samples
The finger-stick samples for glucose analysis were placed in a refrigerator and at theend of the test transferred to a –20°C freezer until analysed, which was performed within5 d. A YSI model 2300 STAT analyser (YSI Life Sciences) was used for glucose analysis. Forinsulin analysis, the microvette tubes were centrifuged and the serum transferred tolabelled polypropylene tubes and stored at –20°C before analysis. Insulin levels weremeasured using a Human Insulin ELISA Kit (Alpco Diagnostics).
Data analysis
Data were entered into a spreadsheet by two different individuals and the values comparedwith assure accurate transcription. Incremental areas under the glucose and insulinresponse curves (AUC), ignoring area below fasting, were calculated. For the purposes ofthe AUC calculation, fasting glucose and fasting insulin were taken to be the mean of thefirst measurement of the blood glucose concentrations and serum insulin concentrations attimes –5min and 0min. The GI and insulin index were calculated by expressing eachparticipant's AUC for the test food as a percentage of the same participant's mean AUC forthe two white bread controls. Values >2 sd above the mean were excluded.The blood glucose and serum insulin concentrations at each time, AUC, GI and insulin indexvalues were subjected to repeated-measures ANOVA examining for the main effects of testmeal and the meal×participant interaction. After demonstration of significantheterogeneity, the significance of the differences between individual means was assessedusing Tukey's test to adjust for multiple comparisons. Means differing by more than theLSD (least significant difference) were statistically significant, two-tailedP<0·05.
Glycaemic load (GL) was calculated using the formula: GL=GI×g of availablecarbohydrate in the portion.
Glycaemic index classification
Using the classification of Brand-Miller for the glucose scale, products with a GI of 55or lower are classified as being low GI; those with a GI of 56 to 69 are classified asmedium, while those with a GI of 70 or greater are classified as high GI.
Results
A total of ten participants (four male and six female) with a mean age of 39 (sd11) years and an average BMI of 26·4 (sd 6·2) kg/m2 completed the study.
Within-subject variation of reference food
The mean within-subject CV of the iAUC values after the two repeated WB tests was17·0±3·6 % and was thus considered technically satisfactory (average intra-subjectvariation of less than 30 %).
Palatability
Palatability scores are presented in Table 2.The subjective palatability of the R50 and R20 meals was higher than that of the WBcontrol. However, this difference did not reach statistical significance.
Table 2.
Palatability, glycaemic index (GI), GI category, glycaemic load (GL), GL categoryand insulin index
(Mean values with their standard errors)
| Palatability (mm) | GI | Insulin index | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Test meal | Abbreviation | Mean | sem | Mean | sem | GI category* | GL | GL category | Mean | sem |
| White bread | WB | 63·0a | 10·0 | 71·0a | High | 35·5a | High | 71·0a | ||
| Raisins (50g CHO) | R50 | 75·0a | 6·0 | 49·0b | 4·0 | Low | 24·5b | N/A | 38·0b | 3·0 |
| Raisins (one serving) | R20 | 72·0a | 6·0 | N/A | N/A | 9·9c | Low | N/A | ||
Open in a separate window
CHO, carbohydrate; N/A, not applicable.
a,b,c Mean values within a column with unlike superscript letters weresignificantly different (P < 0·05).
* Category from GI Factor (Atkinson et al.(,27)).
Postprandial glucose response and glycaemic index
Postprandial incremental glucose levels after the R50 meal were significantly higher thanthose after the WB meal at 15 and 30min. At 60, 90 and 120min, however, the postprandialincremental glucose levels after R50 were significantly lower than after WB (Fig. 1). iAUC were significantly lower after bothraisin meals than after WB (Fig. 2). The final GIand GL values are presented in Table 2.
Fig. 1.
Postprandial glucose responses to three meals containing 50, 50 and 20g ofavailable carbohydrates from white bread (○), raisins (•) and raisins (∆),respectively. Values are means, with standard errors represented by vertical bars.a,b,cMean values at a specific time point with unlike letters weresignificantly different (P<0·05).
Fig. 2.
Incremental AUC (iAUC) for glucose after consumption of three meals containing 50,50 and 20g of available carbohydrates from white bread (WB), raisins (R50) andraisins (R20), respectively. Values are means, with standard errors represented byvertical bars. a,b,cMean values with unlike letters were significantlydifferent (P<0·05).
Postprandial insulin response and insulin index
There was no significant difference between the WB and R50 meals in incrementalpostprandial insulin levels at 15 and 30min. However, insulin levels were significantlylower at 45, 60 90 and 120min with the R50 meal compared with WB (Fig. 3). iAUC were also significantly lower with raisins comparedwith the WB control (Fig. 4). The final insulinindex values are presented in Table 2.
Fig. 3.
Postprandial insulin responses to three meals containing 50, 50 and 20g ofavailable carbohydrates from white bread (♦), raisins (
)and raisins (■), respectively. Values are means, with standard errors represented byvertical bars. a,b,cMean values at a specific time point with unlikeletters were significantly different (P<0·05).
Fig. 4.
Incremental AUC (iAUC) for insulin after consumption of three meals containing 50,50 and 20g of available carbohydrates from white bread (WB), raisins (R50) andraisins (R20), respectively. Values are means, with standard errors represented byvertical bars. a,b,cMean values with unlike letters were significantlydifferent (P<0·05).
Discussion
The present study demonstrates that raisins are a low-GI and -insulin index fruit thatprovides a favourable postprandial glucose and insulin response. In terms of postprandialglucose response, raisins elicited a swifter response compared with WB for the first 30min.This, however, was followed by a sharp decline and an overall lower AUC for glucose whencompared with WB (P<0·05), which is commonly observed with otherfruits. This postprandial glucose response pattern may be explained by the high sucrosecontent of raisins. The sucrose would be rapidly digested and the glucose rapidly absorbedrelative to starch. However, the fructose, which is responsible for 50 % of the availablecarbohydrate content of raisins, would not contribute to the rise in blood glucose. Evidencefrom other studies suggests that the benefits of fructose on glycaemic control may extendbeyond simple replacement of glucose. Moore et al.(,8) demonstrated that the addition of only 7·5g of fructose, levels whichare slightly lower than the fructose content of one serving of raisins, to 75g of glucoseas part of an oral glucose tolerance test significantly lowered the glucose response whencompared with 75g glucose with no added fructose(,8). A potential mechanism of action for this improved glycaemic responsewith fructose ingestion may be enhanced hepatic glucose uptake. Fructose ingestion increasesthe hepatic concentrations of fructose-1-phosphate (first product of hepatic fructosemetabolism), which in turn competes with fructose-6-phosphate for binding to gluco*kinaseregulatory protein (GKRP). This leads to the release of gluco*kinase (rate-limiting enzyme inthe hepatic metabolism of glucose) from GKRP, causing hepatic metabolism and further uptakeof glucose and thus lower postprandial glucose concentrations(,8,9). This glycaemic advantage with moderate intakes of fructose over glucoseis not a novel finding, and has been reported in both healthy individuals and patients withdiabetes in the 1970s and 1980s(,10–12). A recent meta-analysis put this link into perspective by demonstratingthat small doses of fructose<10g/meal or<36g/d can significantly improveserum levels of HbA1c and fasting glucose levels(,6). Furthermore, this daily intake level was not associated with anyadverse metabolic effects that have been linked to high intake of fructose such asdyslipidaemia(,6).
Also of interest is the low GI of raisins as determined by the present study (49 based onthe glucose scale). A previous study by Jenkins et al.(,13) reported the GI of raisins to be 64. However, this study was conductedon only six subjects. More recently a study by Kim et al.(,14) reported GI values of 49 in sedentary individuals, 49 in individualswith prediabetes and 55 in aerobically trained adults. These results are very similar to theGI value determined for raisins in the present study. The health benefits of low-GI fruitwere demonstrated in a recent secondary analysis of a clinical intervention that showed thatlow-GI fruit consumption as part of a low-GI diet was associated with statisticallysignificant reductions in HbA1c, systolic blood pressure and overall CHDrisk(,7). The original randomised clinical trial assessed the effects of a low-GIv. a high-fibre diet on glycaemic control in patients with type 2diabetes and included fruit intake advice as part of the dietaryintervention(,15). The secondary analysis included 152 patients and demonstrated that theGI of fruit was an independent predictor of HbA1c reduction and that the lowest quartile ofGI intake led to the greatest reduction in HbA1c(,7). It is important to note that in this study grapes were consideredhigh-GI foods (GI>90 based on the bread scale). However, the present study suggeststhat raisins have a low GI (GI<70 based on the bread scale). The present study alsodemonstrated that both serving sizes of raisins studied (69 and 28g) are low-GL foods. Thebeneficial effects of low-GI and -GL foods on diabetes and risk of CVD have beendemonstrated by a number of large cohort studies(,16–18). Lastly, the type and amount of fibre present in raisins should not beoverlooked as another component that may account for the lowered glycaemic response.Overall, the present findings support the notion that incorporation of raisins as part of ahealthy, low-GI diet in patients with diabetes or impaired glucose tolerance can potentiallyimprove glycaemic management and provide additional cardiovascular benefits.
The present study also demonstrated that raisins lead to a lower postprandial insulinresponse when compared with WB. This insulin-sparing effect may also be in part due to thefructose content of raisins. Fructose is not an insulin secretagogue and, unlike glucose,does not require insulin for cell entry(,19). The insulin-sparing effects of fructose have been demonstrated in anumber of other studies(,20–22). While long-term impacts of raisins on insulin control require furtherinvestigation, the present study suggests that raisins, through acute postprandialinsulin-sparing effects, may be a healthy food choice in patients with insulin resistance ordiabetes.
The major limitation of the present study, as with all acute feeding studies, is theinability to translate these acute findings to long-term benefits. However, at least interms of the beneficial effect of fructose on glycaemic management, previous studies haveshown that these effects are sustainable over a longer period of time(,23,24). Another shortcoming is the sample size. While the use of ten subjectshas been validated by a number of studies, nevertheless this sample size reduces the studyprecision and may lead to exaggerated associations.
While the potential benefits of moderate consumption of fructose on glucose control havebeen overshadowed by the adverse outcomes, especially on serum lipids(,23,25,26), associated with overconsumption and over-utilisation of high-fructosecorn syrup in the everyday diet, the benefits of fructose as a component of whole fruitsshould not be overlooked. Raisins are popular snacks that are readily accessible at areasonable price. Their nutrient profile, being high in antioxidants, dietary fibre,prebiotics, vitamins and minerals, indicate that they could contribute to overall health.While long-term studies are needed, the present study demonstrates that in addition to theaforementioned benefits, raisins can acutely improve postprandial glycaemic control and, asa low-GI food, may serve as a healthy snack, when used in moderation, in the diets ofhealthy individuals and for those with diabetes or impaired glucose tolerance.
Acknowledgements
The present study was supported by Sun-Maid Growers of California, Kingsburg, CA, USA. Theauthors wish to thank Dr Arianna Carughi, Health & Nutrition Research Coordinator,Sun-Maid Growers of California, for assistance with the study.
C. W. C. K. provided the establishment of funding, study design, data gathering andmanuscript preparation. A. E. and J. L. were involved with data gathering and manuscriptpreparation.
C. W. C. K. has received research grants, travel funding, consultant fees, honoraria or hasserved on the scientific advisory board for Abbott, Advanced Food Materials Network, AlmondBoard of California, American Peanut Council, American Pistachio Growers, Barilla,California Strawberry Commission, Canadian Institutes of Health Research, Canola Council ofCanada, Danone, General Mills, Hain Celestial, International Tree Nut Council, Kellogg,Loblaw Brands Ltd, Nutrition Impact, Oldways, Orafti, Paramount Farms, Pulse Canada,Saskatchewan Pulse Growers, Solae, Sun-Maid Growers of California and Unilever. A. E. and J.L. have no conflicts of interest.
References
1. Lambert JD, Hong J, Yang GY, et al. (2005) Inhibition of carcinogenesis bypolyphenols: evidence from laboratory investigations. Am J ClinNutr 81, Suppl. 1,284S–291S [PubMed] [Google Scholar]
2. Erdman JW, Balentine D, Arab L, et al. (2007) Flavonoids and heart health:Proceedings of the ILSI North America flavonoids workshop, May 31–June 1, 2005,Washington, DC. J Nutr 137, Suppl.,718S–737S [PubMed] [Google Scholar]
3. Dubick MA & Omaye ST (2001) Grape wine and tea polyphenols in themodulation of atherosclerosis and heart disease In Handbook ofNutraceuticals and Functional Foods, 2nd ed., pp.101–130 [Wildman REC, editor]. Boca Raton, FL:CRC Press [Google Scholar]
4. Todd S, Woodward M, Tunstall-Pedoe H, et al. (1999) Dietary antioxidants and fiber in theetiology of cardiovascular disease and all-cause mortality: results from the ScottishHeart Health Study. Am J Epidemiol 150, 1073–1080 [PubMed] [Google Scholar]
5. Bruce B, Spiller GA & Farquhar JW (1997) Effects of a plant-based diet rich in wholegrains, sun-dried raisins and nuts on serum lipoproteins. VegNutr Int J 1, 58–63 [Google Scholar]
6. Sievenpiper JL, Chiavaroli L, de Souza RJ, et al. (2012) ‘Catalytic’ doses of fructose maybenefit glycaemic control without harming cardiometabolic risk factors: a smallmeta-analysis of randomised controlled feeding trials. Br J Nutr 108, 418–423 [PMC free article] [PubMed] [Google Scholar]
7. Jenkins DJA, Srichaikul K, Kendall CWC, et al. (2011) The relation of low glycaemic indexfruit consumption to glycaemic control and risk factors for CHD in type 2diabetes. Diabetologia 54, 271–279 [PMC free article] [PubMed] [Google Scholar]
8. Moore MC, Cherrington AD, Mann SL, et al. (2000) Acute fructose administrationdecreases the glycemic response to an oral glucose tolerance test in normaladults. J Clin Endocrinol Metab 85, 4515–4519 [PubMed] [Google Scholar]
9. Hawkins M, Gabriely H, Wozniak R, et al. (2002) Fructose improves the ability ofhyperglycemia per se to regulate glucose production in type 2diabetes. Diabetes 51, 606–614 [PubMed] [Google Scholar]
10. Akerblom HK, Siltanen I & Kallio AK (1972) Does dietary fructose affect the control ofdiabetes in children? Acta Med Scand Suppl 542, 195–202 [PubMed] [Google Scholar]
11. Crapo PA, Kolterman OG & Olefsky JM (1980) Effects of oral fructose in normal, diabetic,and impaired glucose tolerance subjects. Diabetes Care 3, 575–581 [PubMed] [Google Scholar]
12. Akgün S & Ertel NH (1980) A comparison of carbohydrate metabolism aftersucrose, sorbitol, and fructose meals in normal and diabetic subjects.Diabetes Care 3, 582–585 [PubMed] [Google Scholar]
13. Jenkins DJ, Wolever TM, Taylor RH, et al. (1981) Glycemic index of foods: aphysiological basis for carbohydrate exchange. Am J Clin Nutr 34, 362–366 [PubMed] [Google Scholar]
14. Kim Y, Hertzler SR, Byrne HK, et al. (2008) Raisins are a low to moderate glycemicindex food with a correspondingly low insulin index. Nutr Res 28, 304–308 [PubMed] [Google Scholar]
15. Jenkins DJ, Kendall CW, McKeown-Eyssen G, et al. (2008) Effect of a low-glycemic index or ahigh-cereal fiber diet on type 2 diabetes: a randomized trial.JAMA 300, 2742–2753 [PubMed] [Google Scholar]
16. Liu S, Willett WC, Stampfer MJ, et al. (2000) A prospective study of dietaryglycemic load, carbohydrate intake, and risk of coronary heart disease in USwomen. Am J Clin Nutr 71, 1455–1461 [PubMed] [Google Scholar]
17. Salmeron J, Ascherio A, Rimm EB, et al. (1997) Dietary fiber, glycemic load, and riskof NIDDM in men. Diabetes Care 20, 545–550 [PubMed] [Google Scholar]
18. Salmeron J, Manson JE, Stampfer MJ, et al. (1997) Dietary fiber, glycemic load, and riskof non-insulin-dependent diabetes mellitus in women. JAMA 277, 472–477 [PubMed] [Google Scholar]
19. Bantle JP (2009) Dietary fructose and metabolic syndrome anddiabetes. J Nutr 139, 1263S–1268S [PMC free article] [PubMed] [Google Scholar]
20. Crapo PA, Scarlett JA & Kolterman OG (1982) Comparison of the metabolic responses tofructose and sucrose sweetened foods. Am J Clin Nutr 36, 256–261 [PubMed] [Google Scholar]
21. Teff KL, Elliott SS, Tschop M, et al. (2004) Dietary fructose reduces circulatinginsulin and leptin, attenuates postprandial suppression of ghrelin, and increasestriglycerides in women. J Clin Endocrinol Metab 89, 2963–2972 [PubMed] [Google Scholar]
22. Chong MFF, Fielding BA & Frayn KN (2007) Mechanisms for the acute effect of fructose onpostprandial lipemia. Am J Clin Nutr 85, 1511–1520 [PubMed] [Google Scholar]
23. Bantle JP, Swanson JE, Thomas W, et al. (1992) Metabolic effects of dietary fructosein diabetic subjects. Diabetes Care 15, 1468–1476 [PubMed] [Google Scholar]
24. Osei K, Falko J, Bossetti BM, et al. (1987) Metabolic effects of fructose as anatural sweetener in the physiologic meals of ambulatory obese patients with type 2diabetes. Am J Med 83, 249–255 [PubMed] [Google Scholar]
25. Stanhope KL, Schwarz JM, Keim NL, et al. (2009) Consuming fructose-sweetened, notglucose-sweetened, beverages increases visceral adiposity and lipids and decreasesinsulin sensitivity in overweight/obese humans. J Clin Invest 119, 1322–1334 [PMC free article] [PubMed] [Google Scholar]
26. Swanson JE, Laine DC, Thomas W, et al. (1992) Metabolic effects of dietary fructosein healthy subjects. Am J Clin Nutr 55, 851–856 [PubMed] [Google Scholar]
27. Atkinson FS, Foster-Powell K & Brand-Miller JC (2008) International tables of glycemic index andglycemic load values: 2008. Diabetes Care 31, 2281–2283 [PMC free article] [PubMed] [Google Scholar]
Articles from Journal of Nutritional Science are provided here courtesy of Cambridge University Press
