Diurnal postprandial responses to low and high glycaemic index mixed meals

lable at ScienceDirect

Clinical Nutrition 33 (2014) 889e894

Contents lists avai

Clinical Nutrition

journal homepage: ht tp: / /www.elsevier .com/locate/c lnu

Original article

Diurnal postprandial responses to low and high glycaemic indexmixed mealsq

Michelle Gibbs a,*, Dawn Harrington a,d, Susan Starkey a,b, Peter Williams c,Shelagh Hampton b

aDepartment of Nutrition and Metabolism, School of Biosciences and Medicine, Faculty of Health and Medical Sciences, University of Surrey,Guildford GU2 7XH, Surrey, UKbDepartment of Biochemistry and Physiology, School of Biosciences and Medicine, Faculty of Health and Medical Sciences, University of Surrey,Guildford GU2 7XH, Surrey, UKcDepartment of Mathematics, Faculty of Engineering and Physical Sciences, University of Surrey, Guildford GU2 7XH, Surrey, UK

a r t i c l e i n f o

Article history:Received 6 November 2012Accepted 30 September 2013

Keywords:DiurnalGlycaemic indexInsulinGlucose

Abbreviations: GI, glycaemic index; HGI, high glycaemic index.q Data from this study was presented in abstract f

Summer meeting 2009 at the University of Surrey, G* Corresponding author. Tel.: þ44 (0) 1483 682532

E-mail address: [email protected] (M. Gibbs).d Department of Nutrition and Dietetics, Buckingha

Wycombe Hospital, Queen Alexandra Road, High Wyco2TT, UK.

0261-5614/$ e see front matter � 2013 Elsevier Ltd ahttp://dx.doi.org/10.1016/j.clnu.2013.09.018

s u m m a r y

Background & aims: Glycaemic index testing is conducted in the morning, however postprandial gly-caemia has a diurnal rhythm. The study aimed to evaluate the effect of glycaemic index on glucosetolerance at different times during the day.Methods: A randomised controlled crossover study was conducted in ten healthy participants after astandardised premeal and eight hour fast. Low (37) and high glycaemic index (73) meals, matched forenergy, available carbohydrate, protein and fat, were consumed at 08:00 h and 20:00 h. Blood sampleswere taken for 2 h postprandially.Results: Postprandial glucose area under curve showed effect with time of day after both meals (Lowp < 0.001, High p ¼ 0.003), and a trend (p ¼ 0.06) to higher glycaemic responses in the evening for lowglycaemic index meal. No differences were observed in insulin responses. Despite the calculated dif-ference in meal glycaemic index little difference was observed in morning responses, but differenceswere seen in the evening when insulin insensitivity is increasing, the glycaemic response increase wasproportionally greater for low glycaemic index meals.Conclusions: Low glycaemic index foods are of less value in glycaemic control in the evening than themorning. Consuming food late in the day has a detrimental metabolic impact irrespective of glycaemicindex.

� 2013 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

1. Introduction

Glycaemic index (GI) of foods is a method of classifying fooditems based on postprandial glycaemic responses.1 This responsehas been quantified and published tables of food items are nowavailable giving GI values for many food products.2 The GI of a foodindicates the rate of entry of glucose into the blood, a slower rise ofblood glucose is indicated by a low GI rating.1 The rate of entry of

caemic index; LGI, low gly-

orm at the Nutrition Societyuildford UK.; fax: þ44 (0) 1483686401.

mshire Healthcare NHS Trust,mbe, Buckinghamshire HP11

nd European Society for Clinical N

glucose into the blood affects many hormone and metabolic re-sponses which has implications in health and disease. There aremany studies that link low GI diets to improvements in variousdisease states, in particular type 2 diabetes and coronary heartdisease.3,4 Recently Livesey et al. carried out a systematic review ofthe literature and concluded that the low GI diets result in im-provements in health markers and that unavailable carbohydratealso plays an important role.5

Standardisation of the GI methodology for testing food items hasbeen extensively investigated with a standardised technique rec-ommended by the FAO/WHO (1998). This methodology has beenreviewed and standardisation suggested,6,7 to that end an interna-tional testing scheme has been established.8 Testing is carried out inthe morning, after an overnight fast, when insulin sensitivity isgreatest.9,10 Although standardisation of such techniques are essen-tial, food consumption happens throughout the waking hours and inthe case of shift workers when a person would otherwise be asleep.

utrition and Metabolism. All rights reserved.

Table 1Low and high glycaemic index meal composition and GI values.

GI Weightg

EnergykJ (kcal)

Proteing

Fatg

CHOg

Fibreg

Low GI mealKelloggsAll Bran

42 80 938 (224) 11.20 2.80 38.40 21.60

Semi-skimmed milk 11 150 313 (75) 5.40 2.70 7.2 0Plums 39 50 82 (19.5) 0.3 0.05 4.40 0.80Total 37 280 1330 (319) 16.9 5.6 50 22.4High GI mealKelloggsRice Krispies

82 49 779 (186.7) 2.94 0.49 42.63 0.49

Semi-skimmed milk 11 150 313 (75) 5.4 2.70 7.2 0Low fat mild cheese N/A 12 92 (22.1) 4.69 0.35 0.01 0Half fat mild cheese N/A 13 130 (31.2) 3.71 1.8 0.01 0Total 72 224 1313 (315) 16.7 5.3 49.9 0.5

M. Gibbs et al. / Clinical Nutrition 33 (2014) 889e894890

Time of day has an influence on the postprandial responses to ameal with elevated hormone and metabolic responses in the eve-ning compared to the morning.11 The effect of the meal prior totesting, the second meal effect, has been extensively reported11,12

and shown to have consequences on the response of the testmeal for impaired insulin response, and a cumulative effect on thepostprandial glycaemic rise.

Quantifying GI of individual food items is fairly easy to achievehowever problems start to occur when devising GI content of meals.Macronutrient composition has been shown to play a fundamentalrole in glycaemic responses to a particular meal. The main factorsthat modify the GI of a mixed meal are protein and fat. Both fat andprotein are reported to delay gastric emptying which is a majorfactor in controlling postprandial glycaemia.13,14 Insulin is increasedin the presence of protein15 and fat causes the release of incretinssuch as gastric inhibitory polypeptide and glucagon-like peptide-116,17 that also influence insulin secretion. Dietary fibre also plays arole, depending onwhether it is soluble or insoluble. Insoluble fibreshave no effect on GI of a food,18 but soluble fibres such a guar gumhave been shown to lower GI and insulin responses postprandially,19

this could be due to a slower gastric emptying rate and reducing therate of enzyme digestion by creating a physical barrier.

Foods and food products of known GI and meals of calculated GIare widely used in nutrition research, and conclusions are drawn asto the potential health effects without consideration for the diurnalpatterns of metabolic behaviour. This present study aimed toexamine the effect of altering the GI of meals, given to free livinghealthy young adults either in the early morning or in the evening,on postprandial glucose and insulin responses. We hypothesise thatlowand high GImeal will show elevated glycaemic and insulinaemicresponses in the evening, but that the time of day differencewill notbe of a similar proportion, so that the difference in metabolic benefitbetween the two GI meals will be reduced later in the day.

2. Materials and methods

2.1. Subjects

Ten healthy participants (9 females, 1 male) mean aged25.5 � 8.8 (mean � SD) years and a BMI of 21.9 � 1.7 kg/m2 wererecruited and asked to attend the clinical investigation unit at theUniversity of Surrey on four separate occasions. Subjects werenormotensive, non-smoking, with no history of diabetes mellitus,impaired glucose tolerance, metabolic disorders, gastrointestinaldisorders, coronary heart disease, liver and renal disorders, sleepdisorders, food allergies, drug/alcohol dependence, epilepsy, psy-chiatric disorders and free from any medication, except minor an-algesics and the contraceptive pill. Participants’ caffeineconsumption was 160 � 84 mg/day and their alcohol consumptionwas 6 unit/week (range 0e18 units a week).

2.2. Procedures

This study was conducted in accordance with the HelsinkiDeclaration and all procedures involving human participants wereapproved by the University of Surrey Ethic Committee (EC/2004/37/SBMS). All participants’ information was kept coded and held instrictest confidence. Informed written consent was obtained fromall participants prior to the start of the study.

2.3. Study meal

The meals were devised to minimise any preparation bias orvariation from the calculated GI. The low GI meal consisted of All-bran, semi-skimmed milk and a plum. High GI meals composition

was Rice Krispies, semi skimmedmilk, low and half fat cheese. Bothmeals were matched for calories and nutritional content with theexception of fibre. Table 1 shows the nutritional content and GI ofthe study meal. The low and high GI meals were designed to obtainas a wide difference in the calculated GI values. The values werecalculated using GI values from recognised tables 2 and the mixedmeal formula.20

2.4. Study protocol

The study was of a randomised, cross-over design. Each partic-ipant attended four study legs, two mornings and two evenings.Each study day was a minimum of two days apart over a four weekperiod. Time of day and order of meals was randomly assigned toparticipants. Participants were asked to refrain from heavy physicalexercise, caffeine and alcohol consumption for 24 h prior to andduring each study day. Participants were given a standardised pre-study meal which consisted of a vegetable lasagne and low fatyoghurt (2584 kJ (620 kcals), 14% protein, 53% CHO, 32.5% fat and GI37). In themorning studies, thismeal was consumed before 22:00 hthe previous evening and for the evening studies before 12:00 h onthe same day. No food was consumed after this except water whichwas allowed ad libitum throughout the study. Participants wereasked to keep a food and activity diary for 24 h prior to each studyleg to monitor compliance.

Participants arrived at the clinical investigation unit at 07:45 hfor the morning studies and provided one baseline blood sample bythe fingerprick technique using semi-automatic lancets (Accu-chekmulticlix lancets, Roche Diagnostics, West Sussex) and 300 ml bloodwas collected into fluoride oxalate microvettes (Sarstedt, Leicester).Prior to the study participants were familiarised with the lancettechniques and underwent a practice session. Participants werethen given 200 mls of water and either a low or high GI meal(Table 1) depending on the randomisation. The meal was started at08:00 h in the morning studies and consumed within 15 min.Further blood samples were collected 15, 30, 45, 60, 90 and 120minafter the commencing the meal.

An identical protocol was carried out for the evening study legswith the meals being consumed at 20:00 h.

The blood samples were separated immediately by centrifuga-tion at 4 �C (2500 rpm, 15 min) and plasma aliquoted into plainmicrovettes (Sarstedt, Leicester), and stored at �20 �C prior to in-sulin and glucose analysis, which was performed within 6 months.

3. Analysis

Plasma glucose was determined on an automatic analyser usinga glucose oxidase enzymewhich is immobilised ontomembrane on

Table 2Fasting period (time, hours) and blood glucose, (fasting value, peak value, value atstudy end) after low and high glycaemic index meal in the morning and evening[mean � SEM].

Parameter Low GI amMean � SEM

High GI amMean � SEM

Low GI pmMean � SEM

High GI pmMean � SEM

Fasted period (h) 11.85 � 1.45 11.55 � 1.6 8.4 � 0.49 8.28 � 0.18Fasting glucose

(mmol/L)4.71 � 0.15 4.40 � 0.24 4.65 � 0.17 4.78 � 0.18

Peak glucose(mmol/L)

7.83 � 0.42z 7.8 � 0.48* 8.3 � 0.22 9.54 � 0.42*z

Glucose at120 min(mmol/L)

4.92 � 0.29** 4.83 � 0.24*** 6.74 � 0.3** 6.11 � 0.59***

KEY significance level for paired symbols: z*p < 0.002, **p < 0.001, ***p < 0.05.

0 40 80 120

4

6

8

10

Time (min)

Glu

co

se (m

mo

l/L

)

0 40 80 120

0

10

20

30

Time (min)

Ins

ulin

(p

mo

l/L

)

A

B

Fig. 1. Postprandial plasma glucose (A, N ¼ 10) and insulin (B, N ¼ 9) responses to thelow and high glycaemic index meals consumed morning or evening [mean � SEM] Dlow GI am, , high GI am, : low GI pm, - high GI pm.

M. Gibbs et al. / Clinical Nutrition 33 (2014) 889e894 891

all plasma samples (Yellow springs instrument (YSI) 2300, Farn-borough, Hants) with low, medium and high quality controls. Thelow and medium quality control are produced in house and valuesobtained running these samples repeated times. The high qualitycontrol is a 25 mmol/L standard supplied by YSI. The quality controlvalues in this study were as follows, low 4.44 � 0.08 mmol/L,medium 6.42 � 0.08 mmol/L and high 24.85 � 0.27 mmol/L. Thecoefficients of variationwere low 6.8%medium 4.6% and high 4.02%which reflect the variability of the assay, and samples weremeasured in singleton as is recommended in the International GItesting protocol.8 Plasma insulin was determined by chemilumi-nescent enzyme linked immunosorbent assay (Invitron Ltd, Mon-mouth, UK) and two quality controls were included on every ELISAplate. The quality control values were 12.3 � 0.31 pmol/L and124 � 4.2 pmol/L with a coefficient of variation 12% and 16%respectively.

3.1. Statistical methods

All results are presented as means and standard deviations orstandard error of the mean. All postprandial insulin data werestatistically analysed using Proc GLM repeated measure ANOVAanalysis in the SAS program (PROC MIXED; SAS version 9.1, SASInstitute, Cary, NC, USA). This analysis accounted for any missingvalues in the data sets which represented 8% of the sample. For allanalyses significant effects were followed up with post-hoc testsutilising least squares estimates of marginal means. Incrementalarea under the curve (IAUC) were calculated using the trapezoidalrule with the basal value removed, to assess the overall glucose orinsulin response, above the individual baseline, during the post-prandial period, and expressed as mmol/L.h or pmol/L.h.7 Baselineglucose and insulin levels, and IAUC, were assessed using ANOVAusing Statistica statsoft (Tulsa, USA). This statistical program wasalso used to carry out two factor repeated measures ANOVA(treatment, time) on the postprandial glucose data with Duncanpost hoc test used to ascertain where the differences occurred.Statistical significance was taken as p < 0.05. Due to insufficientsample volume samples fromone participant only nine participantswere analysed for insulin.

4. Results

No difference was observed between the male and femalesubjects, therefore all participants were included in the analysis.

4.1. Fasting period prior to each study

Table 2 shows the fasting period duration for all four meals. Oneway ANOVA showed a significant difference in fasted time for bothGI meal types between the morning and the evening (p < 0.001).

4.2. Fasting glycaemic responses

The mean plasma glucose fasting levels prior to morning mealsare shown in Table 2. One way ANOVA showed no significant dif-ference between fasting plasma glucose levels on all four occasions,indicating that all groups started from an equally fasted baseline.

4.3. Time of day effect on glycaemic response

Fig. 1A shows the 2 hour plasma postprandial glucose re-sponses after all four meals, plotting low and high GI in themorning and the evening. The time at which the meals wereconsumed had a significant effect on postprandial glucose re-sponses. Duncan post-hoc test after repeated measure analysisshowed a significant difference between morning and eveningglucose responses for both LGI (p ¼ 0.002) and HGI (p < 0.001)meals. There was also significantly greater mean peak glucoseresponse observed after HGI consumed in the evening mealcompared to the morning (Table 2).

Plasma glucose returned to fasting levels by the end of the studyperiod after both HGI and LGI morning meals but not followingboth evening meals (Table 2) there we no significant differencebetween the glucose levels after both morning or evening mealshowever it is interesting to note that HGI pm glucose levels appearto be falling faster than the LGI pm glucose levels.

Fig. 2. Incremental area under the curve for postprandial glucose (N ¼ 10) and insulin(N ¼ 9) after low and high glycaemic index meals [mean � SEM].

M. Gibbs et al. / Clinical Nutrition 33 (2014) 889e894892

A time of day effect was also observed in IAUC (Fig. 2), with asignificantly greater IAUC in the evening compared to morningIAUC for both for HGI (p ¼ 0.003) and low GI meals (p < 0.001).

4.4. Effect of GI on glycaemic response

One way ANOVA of the IAUC showed that the difference be-tween postprandial plasma glucose responses for the LGI and HGImeals in themorningwas not quite significant but indicated a trend(142 � 26 versus 218 � 33 p ¼ 0.06), whereas evening glucoseresponses to LGI and HGI meals were not significantly different(309 � 26 versus 351 � 44 mmol/L.h p ¼ NS), suggesting that theeffect of low GI on glycaemic response is diminished whenconsumed in the evening. The percentage increase in the post-prandial glucose response betweenmorning and the eveningmealswas 197% � 35 for LGI and 117% � 13 for HGI meals. This suggeststhat HGI meals will always elicit a similar high glycaemic response,but LGI meals show a greater variation from morning to evening,and are not consistently low GI.

4.5. Insulinaemic responses

Morning and evening plasma postprandial insulin responsesafter low and high GI meals are shown in Fig. 1B.

4.6. Fasting insulin responses

One way ANOVA showed no significant difference betweenfasting plasma insulin levels on all four occasions (LGIam0.51 � 0.3 pmol/L, LGI pm 0.25 � 0.004 pmol/L, HGIam1.03� 0.5 pmol/L, HGI pm 0.29� 0.03 pmol/L)indicating that allgroups started from an equally fasted baseline.

4.7. Time of day effect on insulinaemic response

As with the glucose responses a time of day effect was observedon postprandial insulin responses. LGI meal postprandial insulinresponses in the morning and evening were significantly different(p ¼ 0.036) with higher levels being observed in the morning.However the HGI meal postprandial insulin responses showed theconverse with a trend toward being higher in the eveningcompared to the morning (p ¼ 0.067).

Postprandial plasma insulin levels where still significantlyelevated at the end of the study compared to basal levels, aftermorning low GI meal (p ¼ 0.031), evening low GI meal (p ¼ 0.007)and evening high GI meal (p ¼ 0.034), but not after the morninghigh GI meal.

4.8. Effect of GI on insulinaemic response

Duncan post hoc test showed a significant difference in thepostprandial insulin responses after the LGI and HGI meals in theevening (p ¼ 0.011) with no differences observed in the morningresponses (Fig. 1B).

No significant differences were observed in GI content for IAUCfor postprandial insulin responses.

5. Discussion

Postprandial glycaemic responses to a meal are increased in theevening and remain elevated at the end of the study. This obser-vation has been well documented by ourselves and other re-searchers21,12 this finding is confirmed herewith both high and lowGI meals and suggests that the GI ascribed to a food by morning GItesting may not be accurate over the duration of the day. It is

unclear why insulin levels return to baseline only with the morninghigh GI meal, it may be that the short duration of the postprandialperiod has not elucidated the full response with all meals or be anartefact of the small sample size.

This study further investigated postprandial glucose and insulinresponses to low and high GI meals in the early or later part of theday in order to establish if the type of GI is equally affected by timeof day. This increase in the glycaemia responses in the evening wasgreater proportionally after the low GI meal leading to the sug-gestion that low GI foods are of more value in glycaemic control inthe morning than in the later part of the day when the benefit ofglycaemic control is lost, so consuming a meal late in the day has adetrimental metabolic impact irrespective of the GI of the meal.These responses to the different GI meals, which were matched formacronutrients and calories, could be due to changes in insulinsensitivity which have been report to decrease during the day.9

Furthermore the presence of counter regulatory hormones suchas glucagon, cortisol, ghrelin and growth hormone that influenceinsulin secretion22 and have been shown to have circadianrhythms,23 and thus may also be affecting the diurnal variation inglycaemic response.

The high GI meal resulted in postprandial glycaemia responsesthat were higher and elevated for longer period compared to thelow GI meal. Postprandial hyperglycaemia has been linked as anindependent risk factor of type 2 diabetes and cardiovascular dis-ease.24 Therefore the type of GI consumed in the evening may haveimportant implication in the long term health of the individual.

The conditions the meals were consumed in the morning andevening meals were as similar as achievable but the fasting periodwas significantly shorter in the evening compared to the morningmeals due to the constraints of normal eating times. Howeverfasting glucose level were comparable on all four occasions. Insulinlevels rose faster in the morning than the evening after the low GI

M. Gibbs et al. / Clinical Nutrition 33 (2014) 889e894 893

meal, but also fell faster, suggesting a tighter insulin response. Thismay be an effect of the diurnal variation in insulin resistance or dueto the different fasting periods prior to the meals, or different ratesof gastric emptying later in the day as reported by Goo et al.25 Thecomposition of the lowGImeal included fibre from fruit in the formof a plum and bran and whole fruit have been reported to flattenthe glycaemic responses.26 The meal was identical on both occa-sions, but the composition of the meal may have contributed, inpart, to the rise in insulin, via the intrinsic sugars in the fruit.

The meal prior to the test meal was standardised in order toreduce any variation in hormone and metabolic responses whichhave been shown to occur with a second meal.11,12,27 Researchershave reported that feeding meals rich in slow release carbohydrateimproved both fasting and postprandial blood glucoses levels thenext morning.28 The pre-meal had a typical amount of fibre for aready-meal and this may have contributed to postprandial re-sponses of the test meal, but was the same on both occasions.Another factor is that as one pre-meal was consumed in the eve-ning and other at lunch time an unidentified diurnal effect of thepre-meal, may have also contributed to differences observed inpostprandial responses of the test meal. A full 24 h or more ofcontrolled pre-study period may be recommended for futurestudies.

Even though the GIs of both mixed low and high GI meal wereestimated to be significantly different, this did not result in signifi-cantly different postprandial responses although the area undercurves were greater after high GI meals in the morning and evening.This could have been due to using incorrect GI values of individualfood items when calculating GI from the tables, as this has beenreported by other researchers,7 however the meal used was verysimple and therefore this is an unlikely error. Another explanationcould be other nutritional components of the meals, since proteinand fats are known to affect GI of the meal, however the levels werematched in bothmeals.13,14 The level of carbohydratewasmatched inthe twomeals but the fibre contentwas greater in the low GImeal asis frequently the case for carbohydrate based low GI meals since it isone of the factors that lowers GI in foods.

The limitation of this study is that the low and high GI mealswhere not tested prior to instigating themain study as clearly otherfactors affect GI of a meal, for example gastric emptying andstimulation of gastrointestinal hormone (such as glucose depen-dent insulinotropic hormone) are involved.19,29,30

In summary, low and high GI meals eaten in the early part of theday result in better glycaemic control compared to eating later inthe day. Postprandial control after both low and high GI mealsdeteriorated at the later end of the day, however the low GI meal isaffected to a greater extent, so as to lose any metabolic benefit fromthe low GI. The low GI meal still remains the meal of preference inthe early part of the day to reduce glycaemic divergence fromfasting, although from this study it appears that the benefit of lowGI meals is reduced by the evening. This observation needs to bestudied in larger meals sizes with different nutritional composi-tions, and in a larger group to elucidate any population variation, toidentify the best diet advice for glycaemic control late in the day, orif the better advice is to eat earlier.

Conflict of interests

The authors have no conflict of interest or financial interest todeclare.

Acknowledgements

The authors thank the Surrey GI Testing Service for financialsupport.

The authors’ contributions are as follows: MG was involved inthe concept and design of the study, analysis and interpretation,and preparations of the manuscript. DH was responsible foracquisition of participants, instigating and running clinical aspects,data collection and analysis. SS performed the insulin assays. PWcarried out the SAS analysis. SH was involved in the concept anddesign of the study, analysis, and interpretation and preparations ofthe manuscript.

No Sponsor had any role in the study design or the manuscript.

References

1. Jenkins DJ, Wolever TM, Taylor RH, Barker H, Fielden H, Baldwin JM, et al.Glycemic index of foods: a physiological basis for carbohydrate exchange. Am JClin Nutr 1981;34:362e6.

2. Foster-Powell K, Holt SH, Brand-Miller JC. International table of glycemic indexand glycemic load values: 2002. Am J Clin Nutr 2002;76:5e56.

3. Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucoseand incident cardiovascular events. A metaregression analysis of publisheddata from 20 studies of 95,783 individuals followed for 12.4 years. DiabetesCare 1999;22:233e40.

4. Barclay AW, Petocz P, McMillan-Price J, Flood VM, Prvan T, Mitchell P, et al.Glycemic index, glycemic load, and chronic disease riskea meta-analysis ofobservational studies. Am J Clin Nutr 2008;87:627e37.

5. Livesey G, Taylor R, Hulshof T, Howlett J. Glycemic response and healtheasystematic review and meta-analysis: relations between dietary glycemicproperties and health outcomes. Am J Clin Nutr 2008;87:258Se68S.

6. Wolever TM, Jenkins DJ, Jenkins AL, Josse RG. The glycemic index: methodologyand clinical implications. Am J Clin Nutr 1991;54:846e54.

7. Brouns F, Bjorck I, Frayn KN, Gibbs AL, Lang V, Slama G, et al. Glycaemic indexmethodology. Nutr Res Rev 2005;18:145e71.

8. Wolever TM, Brand-Miller JC, Abernethy J, Astrup A, Atkinson F, Axelsen M,et al. Measuring the glycemic index of foods: interlaboratory study. Am J ClinNutr 2008;87:247Se57S.

9. Beck-Nielsen H, Pedersen O. Diurnal variation in insulin binding to humanmonocytes. J Clin Endocrinol Metab 1978;47:385e90.

10. Lee A, Ader M, Bray GA, Bergman RN. Diurnal variation in glucose tolerance.Cyclic suppression of insulin action and insulin secretion in normal-weight, butnot obese, subjects. Diabetes 1992;41:750e9.

11. Ribeiro DC, Hampton SM, Morgan L, Deacon S, Arendt J. Altered postprandialhormone and metabolic responses in a simulated shift work environment.J Endocrinol 1998;158:305e10.

12. Hampton SM, Morgan LM, Lawrence N, Anastasiadou T, Norris F, Deacon S,et al. Postprandial hormone and metabolic responses in simulated shift work.J Endocrinol 1996;151:259e67.

13. Gentilcore D, Chaikomin R, Jones KL, Russo A, Feinle-Bisset C, Wishart JM, et al.Effects of fat on gastric emptying of and the glycemic, insulin, and incretinresponses to a carbohydrate meal in type 2 diabetes. J Clin Endocrinol Metab2006;91:2062e7.

14. Karamanlis A, Chaikomin R, Doran S, Bellon M, Bartholomeusz FD, Wishart JM,et al. Effects of protein on glycemic and incretin responses and gastric emptyingafter oral glucose in healthy subjects. Am J Clin Nutr 2007;86:1364e8.

15. Nuttall FQ, Mooradian AD, Gannon MC, Billington C, Krezowski P. Effect ofprotein ingestion on the glucose and insulin response to a standardized oralglucose load. Diabetes Care 1984;7:465e70.

16. Morgan LM, Hampton SM, Tredger JA, Cramb R, Marks V. Modifications ofgastric inhibitory polypeptide (GIP) secretion in man by a high-fat diet. Br JNutr 1988;59:373e80.

17. Gniuli D, Calcagno A, Dalla Libera L, Calvani R, Leccesi L, Caristo ME, et al. High-fat feeding stimulates endocrine, glucose-dependent insulinotropic poly-peptide (GIP)-expressing cell hyperplasia in the duodenum of Wistar rats.Diabetologia 2010;53:2233e40.

18. Jenkins DJ, Popovich DG, Kendall CW, Rao AV, Wolever TM, Tariq N, et al.Metabolic effects of non-absorbable carbohydrates. Scand J Gastroenterol Suppl1997;222:10e3.

19. Gatenby SJ, Ellis PR, Morgan LM, Judd PA. Effect of partially depolymerized guargum on acute metabolic variables in patients with non-insulin-dependentdiabetes. Diabet Med 1996;13:358e64.

20. Wolever TM, Jenkins DJ. The use of the glycemic index in predicting the bloodglucose response to mixed meals. Am J Clin Nutr 1986;43:167e72.

21. Carroll KF, Nestel PJ. Diurnal variation in glucose tolerance and in insulinsecretion in man. Diabetes 1973;22:333e48.

22. Bendtson I, Rosenfalck AM, Binder C. Nocturnal versus diurnal hormonalcounterregulation to hypoglycemia in type 1 (insulin-dependent) diabeticpatients. Acta Endocrinol (Copenh) 1993;128:109e15.

23. VanCauterE, Plat L, ScharfMB, Leproult R, CespedesS, L’Hermite-BaleriauxM,et al.Simultaneous stimulation of slow-wave sleep and growth hormone secretion bygamma-hydroxybutyrate in normal young men. J Clin Invest 1997;100:745e53.

24. Gao W, Qiao Q, Tuomilehto J. Post-challenge hyperglycaemia rather thanfasting hyperglycaemia is an independent risk factor of cardiovascular diseaseevents. Clin Lab 2004;50:609e15.

M. Gibbs et al. / Clinical Nutrition 33 (2014) 889e894894

25. Goo RH, Moore JG, Greenberg E, Alazraki NP. Circadian variation in gastricemptying of meals in humans. Gastroenterology 1987;93:515e8.

26. Jenkins DJ, Srichaikul K, Kendall CW, Sievenpiper JL, Abdulnour S, Mirrahimi A,et al. The relation of low glycaemic index fruit consumption to glycaemiccontrol and risk factors for coronary heart disease in type 2 diabetes. Dia-betologia 2011;54:271e9.

27. Wolever TM, Jenkins DJ, Ocana AM, Rao VA, Collier GR. Second-meal effect:low-glycemic-index foods eaten at dinner improve subsequent breakfast gly-cemic response. Am J Clin Nutr 1988;48:1041e7.

28. Axelsen M, Lonnroth P, Lenner RA, Taskinen MR, Smith U. Suppression ofnocturnal fatty acid concentrations by bedtime carbohydrate supplement intype 2 diabetes: effects on insulin sensitivity, lipids, and glycemic control. Am JClin Nutr 2000;71:1108e14.

29. RaynerCK,SamsomM, JonesKL,HorowitzM.Relationshipsofuppergastrointestinalmotor and sensory function with glycemic control. Diabetes Care 2001;24:371e81.

30. Nauck MA, Baller B, Meier JJ. Gastric inhibitory polypeptide and glucagon-likepeptide-1 in the pathogenesis of type 2 diabetes. Diabetes 2004;53(Suppl. 3):S190e6.