http://informahealthcare.com/ijfISSN: 0963-7486 (print), 1465-3478 (electronic)

Int J Food Sci Nutr, 2014; 65(5): 629–636! 2014 Informa UK Ltd. DOI: 10.3109/09637486.2014.886187

STUDIES IN HUMANS

Effects of ingesting low glycemic index carbohydrate food for thesahur meal on subjective, metabolic and physiological responses,and endurance performance in Ramadan fasted men

Weileen Png1, Kalpana Bhaskaran2, Andrew J. Sinclair3, and Abdul Rashid Aziz1

1Sports Science Centre, Singapore Sports Institute, Singapore Sports Council, Singapore, 2Temasek Applied Science School, Temasek Polytechnic,

Singapore, and 3School of Medicine, Deakin University, Geelong, Victoria, Australia

Abstract

Purpose: To investigate the effect of low glycemic index (LGI) carbohydrate meal on subjective,metabolic and physiological responses, and endurance performance in the Ramadan fastedstate. Methods: During Ramadan, 12 Muslim men, in a randomized and crossover design, ingestedfor the sahur meal (i.e. last meal before commencement of the day’s fast), either LGI (glycemicindex¼ 37) or mixed (CON; �57) meal of equivalent macro-nutrient. At �12 h post-prandial,subjects completed a 60 min continuous run. Results: There were no significant differencesbetween the two meals for ratings in perceived satiety, fullness, appetite and mood states.During steady-state exercise, there were no significant differences in metabolic and physiologicalmeasures. In the time-trial, distance ran was significantly lower in LGI versus CON meal trial, butwith a corresponding lower perceived exertion in the LGI trial. Conclusion: Compared to CON,ingesting LGI as the sahur meal did not provide any metabolic, physiological or performancebenefits during endurance run performed 12 h post-prandial in Ramadan fasted state.

Keywords

Blood glucose concentration, exercise,perceived exertion, religious fast, satiety

History

Received 26 November 2013Revised 2 January 2014Accepted 3 January 2014Published online 13 February 2014

Introduction

The concept of glycemic index of foods was developed to allowthe comparison of different foods based on their physiologiceffects rather than their chemical composition (Jenkins et al.,1981). The glycemic index value of a food is calculated based onthe individual’s blood sugar response after ingesting the carbo-hydrate-containing food. This response is then compared to theresponse observed from ingesting equal amounts of a referencefood (i.e. 50 g of glucose; Jenkins et al., 1981). The area under theglycemic-response curve for each food is expressed as a percent ofthe mean response to the reference food taken by the samesubject, and the resulting values are averaged to obtain theglycemic index value for the food (Wolever et al., 1991). It hasbeen well established that foods with a high glycemic index value(�70) are those which are digested and absorbed rapidly causingmarked fluctuations in blood glucose concentration levels. Lowglycemic index foods (�55) are in contrast, digested and absorbedmore slowly and therefore produce gradual, relatively smallerfluctuations in blood glucose and insulin levels (Brand-Milleret al., 1996).

During the Islamic holy month of Ramadan, every able-bodied, healthy Muslims of adult-aged engage in the dailypractice of fasting during the daylight hours for a total of 30 days.Practicing Muslims, residing in the equatorial region, do not eat ordrink (i.e. total abstinence from food and fluids) from pre-dawn

until dusk (i.e. from �05:30 to 19:15 h), daily for about �13 h.In Ramadan, Muslims typically consume their main meals at twoprimary sittings: the first, called the sahur meal, is consumed justbefore the commencement of the day fast between �04:30–05:30 h, and the second meal, called iftar, is consumed immedi-ately upon the breaking of the day’s fast, at �19:15 h. There isevidence to indicate that high-intensity endurance exerciseperformance is adversely affected during Ramadan fasting,particularly when exercise was performed in the late afternoon,i.e. between �11–15 h post-ingestion of the sahur meal (Azizet al., 2010, 2012; Brisswalter et al., 2011; Chennaoui et al.,2009). Although the mechanism(s) for this impact is not known,there is clearly a need to develop specific strategies that couldhelp Muslims attenuate the impact of Ramadan fasting whenexercising in the fasted state (Aziz & Png, 2008; Burke & King,2012; Maughan et al., 2012).

Henceforth from a nutritional-exercise perspective, it seemsprudent to investigate the impact of ingestion of specificcarbohydrate foods prior to exercise during Ramadan. Forexample, low glycemic index carbohydrate foods, helps to providea more sustained energy release (Jenkins et al., 1981) throughoutthe post-absorptive fasting period and during the subsequentexercise. Indeed previous studies have indicated that pre-exerciseingestion of low glycemic index foods as compared to highglycemic index food enhanced endurance performance (Stevensonet al., 2005, 2006). Several mechanisms were proposed for thisobserved enhancement. One, ingesting low glycemic index mealmaintained a relatively stable blood glucose concentration levelsthroughout the endurance exercise duration (DeMarco et al.,1999; Thomas et al., 1991). Second, pre-ingestion of lowglycemic meal has been shown to result in lower carbohydrate

Correspondence: Abdul Rashid Aziz, Singapore Sports Institute,Singapore Sports Council, 230 Stadium Boulevard, Singapore 397799,Singapore. Tel: +65-9863 7282. Fax: +65-6500 5044. E-mail:abdul_rashid_aziz@ssc.gov.sg

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.

oxidation and substantial greater usage of fats during enduranceexercise (Wee et al., 1999; Wu et al., 2003). The former isclearly advantageous since this helps to limit the use of themuscle glycogen, i.e. sparing this premium substrates for lateruse. Third, in addition to exercise metabolic effects, ingesting lowglycemic index meal has also been shown to lessen hungersensations and stabilize blood glucose levels during the pro-longed post-prandial period, which could alleviate negativefeelings like moodiness, lethargy and subjective perceptions ofhunger and/or fatigue that are commonly associated with adverseexercise performance (Ludwig, 2000; Roberts, 2003). It istherefore not unreasonable to suggest that Ramadan fastingsubjects, whom with no access to food and fluid for up to 18 hduring the day, may also possibly benefit from ingesting lowglycemic index meal during endurance exercise, via directlymaintaining the fasted individual’s blood glucose levels lateduring exercise and using fats as primary fuel for muscular work,and/or indirectly, by increasing the individual’s subjective posi-tive feelings of satiety and extending the feeling of fullness (Aziz& Png, 2008; Maughan et al., 2012).

The primary aim of this study therefore is to determine theeffects of ingesting a low glycemic index carbohydrate food as thesahur meal (i.e. last main meal before commencing of the day’sfast) on various subjective and metabolic markers, physiologicalresponses as well as the impact on the ensuring enduranceperformance in Ramadan fasted individuals. Based on the positiveeffects of low glycemic foods as pre-exercise meal cited in theabove-mentioned paragraphs, it was hypothesized that ingestinglow glycemic index food as the sahur meal will be advantageousduring exercise in the Ramadan fasted state.

Methods

Subjects

Twelve recreationally active male Muslim runners (age:27.9 ± 7.2 years; stature: 173 ± 4 cm; body mass: 65.0 ± 8.1 kg;maximal aerobic power (VO2max): 51.2 ± 9.2 ml kg�1 min�1; andmaximal heart rate: 189 ± 8 beats min�1) volunteered for thestudy. Participants ran regularly between 2 and 4 times week�1,averaging between 15 and 25 km week�1 during the last 3 monthsbefore the start of the study.

Experimental design

The study was undertaken to coincide with the actual Ramadanmonth where the daily fasting times was from �05:30 to 19:15 h(a total duration of �13.5 h). The study used a single-group,cross-over, counterbalanced, double-blind design comparing theparticipants’ subjective, metabolic and physiological responses,and endurance running performance during the Ramadan monthafter ingesting either low glycemic index (LGI) or normal mixed(CON) carbohydrate food as the sahur meal. Participants were notinformed of the real purpose of the study in order to minimize anypotential placebo effect. Instead, they were informed thatthe primary aim of the study was to examine if there werephysiological and performance differences between exercisingin the middle versus the end part of the Ramadan month.Further, both the participants and investigators who performedthe metabolic and physiological measures and tests did notknow which type of food (i.e. LGI or CON meal) wasadministered to the individual at any time point of the investi-gation. At the end of the study, participants were debriefed on thereal aim of the study and their results revealed. All participantssigned written consent and the study had the approval of theDeakin University’s Human Ethics Advisory Group, applicantnumber HEAG-H 130/09.

Each participant made three separate visits to the laboratory.The first visit was to determine the participant’s VO2max followedby a familiarization of the 60 min endurance run. The first visitwas conducted prior to the Ramadan month and participantsperformed the exercise in the non-fasted state. Visits 2 and 3 werethe experimental trials of either the ingestion of LGI or CONmeals. All experimental trials were conducted during the last 2weeks of the Ramadan month with the LGI or CON meal trialsconducted in random order among the participants, and with theinterval duration between the two food trials for each participantbetween 7 to maximum of 12 days apart. Participants wereinstructed to maintain their normal exercise routine throughoutthe study period and not undertake any strenuous exercise 24 hprior to their exercise trials. All exercise trials were conducted at�17:00 h, i.e. ±30 min for each individual’s LGI and CON mealtrials. All sessions were conducted in the laboratory underthermo-neutral conditions (20–23 �C temperature and 50–55%relative humidity).

Experimental meals

For each of the experimental trials, participant reported to thelaboratory at pre-dawn, at �04:30 h. Subject was then providedwith either a LGI or normal mixed (CON) meal (i.e. fried ricemeal) of the same macro-nutrient composition and 450 ml offluid. Participants consumed the provided meal within �20 minafter being served and thereafter, remained in the laboratory untiltime to exercise at �17:00 h. During this period, participantsengaged in light activities such as watching movies, reading,computing or limbering and stretching. They were also notallowed to take any daytime naps.

The carbohydrate content of the two intervention meals wascalculated based on 2.0 g carbohydrate per kg body mass. Theglycemic index value of the LGI and CON carbohydrate meal was36 ± 3.4 (energy content: 805 ± 102 kcal; carbohydrate120.8 ± 15.3 g; protein 40.3 ± 5.1 g; fat 17.9 ± 2.0 g) and �57(energy content: 791 ± 104 kcal; carbohydrate 118 ± 16.3 g; pro-tein 39.6 ± 5.2 g; fat 17.6 ± 2.3 g), respectively. The detailedingredients of both meals are listed in Table 1 and both mealslooked similar in appearance when presented to participants forconsumption. The LGI meal was tested in vivo in 10 healthysubjects using the internationally accredited ISO 26642 method-ology. The glycemic index of the CON meal was not tested

Table 1. Ingredients used in the preparation of the low glycemic index(LGI) and normal mixed meal (CON) carbohydrate fried rice meals.

Low glycemic index (LGI)(glycemic index¼ 36)

Normal mixed meal (CON)(glycemic index¼�57)

Basmati rice Jasmine rice, polishedPearled barleya –Red lentilsa –Broccoli, raw Broccoli, rawYard long beans, raw Yard long beans, rawCarrots, raw Carrots, rawPea, frozen, defrosted Pea, frozen, defrostedOil, canola Oil, canolaTomato, raw, diced, deseeded Tomato, raw, diced, deseededChicken thigh, raw, deskinned,deboned, cubed

Chicken thigh, raw, deskinned,deboned, cubed

Onion, garlic, dried chilli,shrimp paste

Onion, garlic, dried chilli,shrimp paste

Basmati rice, red lentil and pearly barley contributed to the majorcarbohydrate sources in the LGI meal. In the CON meal, the majorcarbohydrate source is jasmine rice. The basmati rice was cooked aldente by the pilaf method and vinegar was added while cooking todecrease starch gelatinization. The other ingredients were kept the samefor both of the meals.

630 W. Png et al. Int J Food Sci Nutr, 2014; 65(5): 629–636

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.

in vivo, but was determined using the mixed meal calculations,which has previously been endorsed (Wolever & Jenkins, 1986).All meals were pre-prepared at the Temasek PolytechnicGlycemic Index Research Unit’s metabolic kitchen, under thesupervision of one of the authors, KB. Participants were furtherasked to record in a diary all the food and drink consumed in thelast 24 h prior to the exercise session. The records from the firstexercise trial was photocopied and returned to the participant.Participants were instructed to replicate the same amount ofenergy intake as closely as possible for the 24 h prior to their nexttrial (Jeacocke & Burke, 2010). Compliance was confirmed uponarrival to the laboratory for each experimental trial afterinspection of the individual’s food diary by the primary inves-tigator, WP.

Biochemical measurements at rest

Capillary blood samples via finger-prick were taken atcommencement of the day’s fast (at 0 min, before meal) and at15, 30, 60, 90, 120, 240, 360, 480, 600, 720th min after the mealhad been consumed, for analysis of blood glucose concentration.Whole blood glucose concentrations were determined using aglucose monitor (Advantage, Roche Diagnostics Corp.,Indianapolis, IN), and all measures were taken by the sametrained investigator.

Subjective ratings of satiety, hunger, satisfaction andfullness at rest

Participants were also instructed to rate their subjective feelingsof satiety, hunger, satisfaction and fullness before and afteringesting the two experimental meals. These questions wereadministered at �5 min before consuming the meal, and atintervals of 15, 120, 240, 360, 480, 600, 720th min marks post-absorptive. Participant’s perceived ratings was made on 100 mmvisual analogue scale with descriptors words anchored at each endthat expressed the two most extreme ratings, which has beenvalidated (Flint et al., 2000).

Maximal aerobic power (VO2max) test

Participants performed a standardized warm-up on treadmill(Venus, HP-Cosmos, Germany) with gradient set at 1% through-out the test. The initial speed for the VO2max test was between 7.0–9.5 km h�1, which was subsequently increased by 1.0 km h�1

every 90 s until volitional exhaustion. Respiratory gas exchangewas collected throughout test with an open circuit spirometrysystem (TrueOne 2400MMS, Parvomedics, UT) which wascalibrated prior to each trial. Participant’s VO2max was the highestvalue achieved over any 30 s period. Also, from the same test,participant’s running speed at 65% VO2max intensity wasdetermined graphically when plotting VO2 against runningspeed. The velocity obtained was subsequently used in theindividual’s LGI and CON meal trials for the first 30 min pre-loading run.

Endurance exercise performance

Prior to the exercise run, participants voided their bladders andprovided a urine sample for measurement of urine specific gravity(PR series, Atago, Tokyo, Japan) to assess the individuals’hydration status. Participants then provided a finger-prick capil-lary blood to determine blood lactate concentration (Lactate Pro,Arkay, Japan). Participants were then weighed (2-150-P, Mettler-Toledo, Albstadt, Germany) only in their running shorts to obtaintheir pre-exercise body mass. Participants also completed severalpre-exercise questionnaires relating to daytime sleepiness andcurrent mood state.

The 60 min run was conducted on the same treadmill withgradient set at 1% throughout the test to simulate energetic cost ofoutdoor running (Jones & Doust, 1996). Participants ran between7.0–9.0 km h�1 for 4 min followed by 6 min of self-stretching astheir standardized warmed-up. The continuous endurance runconsisted of a 30 min pre-loading run followed immediately witha self-paced 30 min time-trial (TT) run. In the pre-loadingrun, participants ran at a fixed speed equivalent to 65% ofthe individual’s VO2max (mean of 9.3 ± 1.4 km h�1 across the12 participants). For the TT, participants maximized the distancerun during the 30 min by manually adjusting their runningvelocity by increasing or decreasing the treadmill’s speedaccordingly, at their own discretion. During the TT run, partici-pants were provided with the elapsed time to complete the runbut were blinded to any feedback information on the distancecovered, heart rate (HR) and the prevailing treadmill speed.To ensure consistency across trials and subjects, no verbalencouragement was provided during the entire run but par-ticipants were clearly reminded and instructed to maximizetheir distance covered at the commencement of the TT run.At the end of the 60 min run, post-exercise blood glucose andlactate were measured. HR was recorded every minute through-out via a chest-strapped short-range transmitter (RS400,Polar Electro, Finland). Participants then dismounted the tread-mill, removed all other clothing except for the running shortsand were toweled dry of sweat and post-exercise body masswas measured. Subjective ratings of perceived exertion (RPE;1–10 scale) of the run were also administered after thepost-exercise body mass has been taken, about �15 minlater. The distance covered (in m) during the 30 min TT run wasused as the criterion measure of the individual’s enduranceperformance. The coefficient of variation for the distance coveredduring the 30 min TT run has been reported to be 1.6% (Oliveret al., 2009).

Metabolic measurements during exercise

During the pre-loading run (i.e. first 30 min run at 65% VO2max

intensity) of the 60 min run, between the 25th and 29th min mark,steady-state metabolic variables of volume of oxygen uptake (VO2)and carbon dioxide expired (VCO2) were obtained continuously for4 min, with the same respiratory equipment used in the VO2max test.However, only data during the last 2 min of the 4 min were averagedand used to calculate the rates of whole body carbohydrate and fatoxidation, using the stoichiometric equations of Jeukendrup &Wallis (2005), and assuming that protein oxidation was negligible:carbohydrate oxidation (g min�1)¼ 4.210�VCO2 – 2.962�VO2;and fat oxidation (g min�1)¼ 1.695�VO2 – 1.701�VCO2. Inaddition, participant’s blood glucose and lactate were measuredwith the same procedures as that for pre-exercise, between the 29thand 30th min marks.

Mood state and daytime sleepiness

Participants completed the validated Brunel University MoodState (BRUMS) which comprised 24 items of descriptors for sixdifferent mood subscales of anger, confusion, depression, fatigue,tension and vigor (Terry et al., 1999). Raw scores for each moodsubscale were summed and converted to T-scores for statisticalanalysis. Participants also rated their degree of sleepiness via the9-point Karolinska Sleepiness Scale, which has previously beenvalidated (Kaida et al., 2006).

Statistical analysis

Statistical analysis was performed using the SPSS version 15(Chicago, IL). No prior power analysis was performed, but the

DOI: 10.3109/09637486.2014.886187 Glycemic index foods and Ramadan fasting 631

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.

number of participants studied was justified based upon thefact that almost all studies that have examined the effects ofGI meals had less than 12 (Donaldson et al., 2010). Pairedsample t-test was used to compare all variables of interestbetween the LGI and CON trials. For variables with a timecomponent (e.g. velocity, HR, subjective ratings, BRUMS, etc.) atwo-factor analysis of variance on repeated measures with the trial(LGI and CON meal) and time (rest time or exercise time) as thebetween-group and within-group factors. If a significant maineffect was found, paired sample t-tests were used to detect wherethe differences occurred with Bonferroni post-hoc tests. The levelof statistical significance was set at p50.05. To illustrate themagnitude (meaningfulness) of differences between LGI andCON meal trials, effect sizes (d) calculations were utilized.A modified scale was used for the interpretation of d; trivial:50.2; small: 0.2–0.6; moderate: 0.6–1.2; and large: 1.2–2.0(Hopkins et al., 1999).

Results

Endurance performance, metabolic markers andphysiological responses

All subjects completed the 30 days of intermittent Ramadanfasting. Table 2 shows that the distance covered and averagevelocity over the 30 min TT run were significantly higher in theCON compared to that performed in the LGI meal trial (bothp50.05). Figure 1 shows each individual’s endurance perform-ance in the two meal trials. The difference in the distance ranbetween meal trials is equivalent to mean 2.6 ± 3.3%; however,

the effect size difference in magnitude is deemed trivial, largelydue to the higher variability in performances between the subjects(Figure 1). Participants’ running velocity and HR averaged every5 min across the TT run tended to be statistically significant: maineffects of trial (F1,11¼ 4.44 and 4.16, p¼ 0.059 and 0.066,respectively; Figures 2 and 3). Both variables did indicate,however, a main effect of time (F5,55¼ 30.39 and 55.20, bothp¼ 0.001).

Table 3 shows measures taken prior to the start of the 60 minendurance run. Table 4 shows the metabolic markers andphysiological responses during the steady-state 30 min pre-loading run. Table 5 shows all the measures undertaken at thecompletion of the 60 min exercise run, i.e. end of the TT run.Except for pre-exercise blood lactate concentration (Table 3) andpost-exercise RPE (Table 5), there were no statistical significantdifferences between LGI and CON meal trials in all othervariables measured and calculated at pre-, mid- and post-60 minendurance run.

Figure 2. Velocity (averaged every 5 min) during the 30 min time-trialrun (second part of the 60 min endurance exercise) in the low glycemicindex (LGI; glycemic index¼ 37) and normal mixed (CON; glycemicindex¼�57) carbohydrate meal exercise trials (N¼ 12).

Figure 3. Heart rate (averaged every 5 min) during the 30 min time-trialrun (second part of the 60 min endurance exercise) in the low glycemicindex (LGI; glycemic index¼ 37) and normal mixed (CON; glycemicindex¼�57) carbohydrate meal exercise trials (N¼ 12).

Figure 1. Distance covered during the 30 min time-trial run (second partof the 60 min endurance exercise) in the low glycemic index (LGI;glycemic index¼ 37) and normal mixed (CON; glycemic index¼�57)carbohydrate meal exercise trials (N¼ 12).

Table 2. Mean distance covered and average velocity for the 30 min time-trial component of the �60 min endurance running performed at 12 hpost-prandial in the low glycemic index (or LGI; glycemic index¼ 37) ornormal mixed (or CON; glycemic index¼�57) carbohydrate mealexercise trials (N¼ 12).

CON LGI p Value d, interpretation

Distancecovered (m)

5612 ± 998 5486 ± 943* 0.03 0.13, trivial

Average velocity(km h�1)

11.2 ± 2.0 10.9 ± 1.8* 0.03 0.19, trivial

d¼ effect size; *significantly different between LGI and CON meal trial,p50.05.

632 W. Png et al. Int J Food Sci Nutr, 2014; 65(5): 629–636

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.

Blood glucose concentration

Figure 4 shows the blood glucose concentration pattern of bothmeal trials throughout the post-absorptive 12 h fasting period andthroughout the 60 min exercise trials. There was significant maineffects of treatment (F1,11¼ 7.85, p¼ 0.017) and interactions(F11,121¼ 5.78, p¼ 0.001). Blood glucose concentrations peakedafter �45 min post-prandial in both meal trials, with significantlyhigher peak value observed in CON meal trial (p50.01).Thereafter blood glucose concentration levels declined moregradually in LGI compared to that of CON as the fasting periodprogressed. At the 360th min mark, mid-way between the sahurmeal and commencement of the run, blood glucose in the LGImeal trial was significantly higher than in the CON trial(p50.05). It is also important to highlight that there were nostatistical significant differences in blood glucose between the twomeal trials at the start, mid- and end of the 60 min endurance run(all p40.05; Figure 4).

Subjective ratings of satiety, hunger, satisfaction andfullness

Although there was a main effect of time for ratings of subjectivesatiety, hunger, satisfaction and fullness (data not shown; allF ratio values were p50.01), more importantly, all variablesshowed no main effect of meal or interactions (all F ratios valueswere p40.05).

Mood state and daytime sleepiness

There were no significance differences between LGI and CONmeal trial for any of the mood descriptors in the BRUMS (data notshown; all variables p40.05). There was also lack of significantdifference for the daytime sleepiness in participants between theirLGI and CON meal trials as reflected by the KarolinskaSleepiness Scale (p40.05; Table 3).

Discussion

The primary aim of this study was to examine the effects ofingesting low glycemic index carbohydrate foods during the sahurmeal on metabolic markers and physiological responses, and itssubsequent effect on endurance running performed 12 h post-prandial in the Ramadan fasted state. The results showed thatingestion of LGI sahur meal (glycemic index¼ 37), generally ledto similar impact on perceived satiety, metabolic markers andphysiological responses throughout the 12 h post-prandial periodto that of ingesting a normal mixed sahur meal (CON; glycemicindex¼�57); but distance covered during the 30 min TT run wassignificantly lower in the LGI meal trial. Thus the main finding ofthe study was that ingesting LGI versus normal meal for sahur ledto a decrease in endurance performance. This finding however,must be viewed with caution given the following circumstancesand observations: (i) the small number of subjects tested, (ii) thetrivial nature of the magnitude of the difference in exercise

Table 4. Metabolic markers via gas exchange, and physiological responses during the first 30 min pre-loading run at 65% VO2max

intensity performed at 12 h post-prandial after the ingestion of either low glycemic index (LGI; glycemic index¼ 37) or normal mixed(CON; glycemic index¼�57) carbohydrate meal (N¼ 12).

CON LGI p Value d, interpretation

Respiratory exchange ratio (au) 0.89 ± 0.04 0.88 ± 0.03 0.07 0.29, smallCHO oxidation (g min�1) 1.78 ± 0.45 1.68 ± 0.43 0.11 0.22, smallFat oxidation (g min�1) 0.41 ± 0.16 0.45 ± 0.13 0.06 0.17, trivialHeart rate (b min�1) 159 ± 11 158 ± 8 0.72 0.11, trivialBlood glucose (mmol l�1) 5.2 ± 0.4 5.1 ± 0.3 0.77 0.29, smallBlood lactate (mmol l�1) 2.7 ± 1.1 2.7 ± 1.3 0.61 0.01, trivial

d¼ effect size; au¼ arbitrary units; CHO¼ carbohydrate; all gaseous exchange and heart rate data were averaged over the last 2 min of4 min of continuous data taken.

Table 5. Measures taken at the end of the 60 min of endurance running performed at 12 h post-prandial after the ingestion of either lowglycemic index (LGI; glycemic index¼ 37) or normal mixed (CON; glycemic index¼�57) carbohydrate meal (N¼ 12).

CON LGI p Value d, interpretation

Blood glucose (mmol l�1) 7.4 ± 1.8 7.1 ± 1.3 0.47 0.19, trivialBlood lactate (mmol l�1) 7.7 ± 2.4 7.3 ± 2.5 0.13 0.16, trivialRatings of perceived exertion (au) 8.8 ± 1.4 8.3 ± 1.7* 0.03 0.32, smallSweat loss (ml) 1126 ± 290 1132 ± 280 0.74 0.02, trivialBody mass loss (% of pre-exercise) 1.8 ± 0.5 1.8 ± 0.4 0.86 0.01, trivial

d¼ effect size; au¼ arbitrary units; *significantly different between LGI and CON meal trial, p50.05.

Table 3. Measures taken at pre-exercise of �60 min endurance running performed at 12 h post-prandial after the ingesting either the lowglycemic index (LGI; glycemic index¼ 37) or normal mixed (CON; glycemic index¼�57) carbohydrate meal (N¼ 12).

CON LGI p Value d, interpretation

Body mass (kg) 63.3 ± 7.6 63.6 ± 8.0 0.19 0.03, trivialUrine specific gravity (au) 1.018 ± 0.005 1.020 ± 0.005 0.25 0.44, moderateBlood glucose (mmol l�1) 4.4 ± 0.4 4.6 ± 0.4 0.15 0.50, moderateBlood lactate (mmol l�1) 1.8 ± 0.4 1.5 ± 0.5* 0.02 0.67, moderateKarolinska sleepiness scale 3.8 ± 1.5 3.8 ± 1.9 0.21 0.01, trivial

d¼ effect size; au¼ arbitrary units; *significantly different between LGI and CON meal trial, p50.05.

DOI: 10.3109/09637486.2014.886187 Glycemic index foods and Ramadan fasting 633

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.

performance between meal trials (Table 2 and Figure 1), (iii) thelack of significant differences in many of the metabolic andphysiological (Tables 4 and 5) during exercise between the mealtrials and (iv) the significantly lowered RPE at the end of the TTrun in LGI versus CON meal trial (Table 5).

In this study, it was initially hypothesized that consuming LGImeal would be advantageous, both metabolically and perform-ance-wise, compared with consuming the CON meal. Thishypothesis was based on the premise that low glycemic indexmeal would be slowly digested during the post-prandial periodand the delayed release of glucose to the working muscles willhelp sustained blood glucose levels during exercise, and thiswould then affect a superior performance in the subsequentendurance run. This is supported by the previous work ofStevenson et al. (2005, 2006). In the 2006 study, subjects were fedwith either a low or high glycemic index meal for 24 h periodbefore embarking on a time-to-exhaustion run at 70% VO2max

performed 12 h post-prandial the last pre-exercise meal. The studyfound that the ingestion of low glycemic index foods increasedendurance capacity and this was primarily due to the increased infat oxidation during the run (Stevenson et al., 2006). In contrast toour initial hypothesis, this study’s result showed a poorerendurance performance when ingesting LGI food (Figure 1 andTable 2) and although fat oxidation during the run tended tohigher in the LGI versus CON meal trial, the difference was trivial(Table 4). In addition, blood glucose concentration was notsignificantly different between meal trials before, during and afterthe 60 min endurance run (Figure 4).

Although overall not statistically significant, it was apparentthat participants who consumed the CON meal had a faster meanrunning velocity than in the LGI meal trial for the first 20 min ofthe 30 min TT run (Figure 2). Perhaps, the higher running speedsin CON meal trial was driven by the relatively higher metabolicrate, and this is indirectly supported by the slightly greatercarbohydrate and lower fat oxidation during the pre-loadingexercise (Table 4) and coupled with a higher post-exercise bloodlactate concentrations observed in CON meal trial at the endof the TT run (Table 5). In addition, fasted participants in theCON meal trial seemed to be able to optimize their exertion levelsas evidenced by the significantly greater psycho-physical exertion

(p¼ 0.027, Table 5) and slightly higher exercise HR throughoutthe TT run (Figure 3). While it must be highlighted that many ofthe above-mentioned data between CON versus LGI meal trialswere not statistically significant or showed trivial to smalldifferences, they do seem to support that fasted participants wereable to maximize their exercise performance relatively better,after the ingestion of the CON meal, relative to ingesting the LGImeal. Indeed, in previous studies that have demonstrated theoutcome of an enhanced endurance performance with theingestion of pre-exercise carbohydrate relative to ingesting aplacebo, participants’ physiological measures during the subse-quent exercise such as carbohydrate metabolism, HR, lactate andRPE were consistently observed to be relatively higher (Febbraioet al., 2000; Sparks et al., 1998; Wong et al., 2009); vis-a-vis tothat similarly observed in this study. It should be highlightedthough that in the studies cited, the duration between the ingestionof the pre-exercise meal to the commencement of exercise were,however, of much shorter duration to that in the present study(from 45 min to 4 h versus 12 h, respectively).

It would be interesting to speculate the possible reasons and/ormechanisms for the relatively poorer endurance performance inthe LGI to that of CON meal trial in the present investigation. Thesuperior endurance performance in the CON meal trial isindirectly supported by previous studies that have demonstratedthe advantageous of ingesting high glycemic index foods(Donaldson et al., 2010; Nicholas et al., 1997). The rate ofmuscle glycogen storage is predominantly influenced by therelative concentrations of insulin and the available glucosesubstrates, and thus ingesting carbohydrate sources with a highglycemic index will likely lead to an accelerated and higher levelsof muscle glycogen synthesis. Indeed, it had previously beenshown that muscle glycogen storage was substantially greater withthe consumption of a high glycemic index meal when compared toconsumption of a low glycemic index diet (glycemic index: 108versus 71, respectively; Burke et al., 1993), and that the ingestionof a high glycemic index meal led to 61% greater muscle glycogensynthesis rate relative to low glycemic index meal during the first6 h of exercise recovery (van Hall et al., 2000). Hence, in thisstudy, we speculate that ingesting the CON as compared to LGI,as the sahur meal, was beneficial in ensuring relatively greater

Figure 4. Blood glucose concentration during the 12 h post-prandial period and 60 min endurance run exercise in the low glycemic index(LGI; glycemic index¼ 37) and normal mixed (CON; glycemic index¼�57) carbohydrate meal exercise trials (N¼ 12). *p50.05; **p50.01.

634 W. Png et al. Int J Food Sci Nutr, 2014; 65(5): 629–636

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.

storage of endogenous glycogen during the passive 12 h fastingperiod which then led to greater contribution of glycogenolysis asfuel substrate during the subsequent endurance run. The former isindicated by the much earlier and higher peak blood glucoseconcentration observed in CON versus LGI (Figure 4), while thelatter is evident by the marginally higher, albeit not statisticallydifferent, carbohydrate oxidation in CON versus LGI trial duringthe 30 min steady-state exercise (Table 4). While this study’s datagenerally indicated that the magnitude of differences betweenmeal trials was mainly trivial to small, it should also be noted thatsuch small changes may however be significant in top-levelcompetitions where the gap between winners and losers are verysmall (Donaldson et al., 2010; Hopkins et al., 1999).

In much of the research work on carbohydrate foods,contrasting glycaemic responses were commonly observed overpost-prandial period of between 2–4 h (Donaldson et al., 2010).The present study assessed the subjects’ glycaemic responses for12 h duration and a similar trend was observed. The LGIcompared to the CON meal, resulted in a lower blood glucoseconcentration during the first hour post-prandial period; thereafterblood glucose declined at a more gradual rate and remained fairlystable between the 6–8 h marks (Figure 4). From this time pointonwards, there were no significant differences in the bloodglucose concentration between the LGI and CON meal; whichsuggest that the impact of LGI had diminished as a result of theprolonged fasting duration. Nonetheless, an interesting observa-tion was the significantly higher blood glucose in the LGI versusCON meal trial between the 4–6 h mark post-prandial (p¼ 0.041;Figure 4). Given that relatively higher blood glucose concentra-tion may enhance exercise performance by providing a higherglucose uptake to the working muscles (Karamanolis et al., 2011;McConell et al., 1994; Moore et al., 2010), it is therefore notunreasonable to speculate that exercise performance performed atthis time-point may well be advantageous or beneficial in the LGImeal trial as compared to that of the CON meal trial. Thisobservation clearly merits further investigation into the effects ofLGI meal in the Ramadan fasted state, particularly if the fastingMuslim individual plans to exercise or compete in the latemorning or early afternoon period, at �4 to 6 h post-sahur mealingestion.

Previous research has also indicated that low, as compared tohigh glycemic index foods, generally promote satiety and suppresshunger due to differences in their blood glucose and hormonalresponses (Ludwig, 2000; Roberts, 2003). However, this study’sresults on ratings of subjective perception of hunger, satisfactionand fullness showed no significant differences between the LGIand CON meal trials; although there were variations in thesevalues over the 12 h post-prandial passive period. The lack ofimpact of LGI meal on these subjective measures is nonetheless,in line with previous studies (Liu et al., 2012; Reynolds et al.,2009). It is however important to highlight that none of theparticipants reported of any gastrointestinal discomfort through-out the post-prandial period as well as during exercise wheningesting the prepared LGI meal.

Prolonged acute fast can lead to athletes experiencing hypo-glycemia (defined as blood glucose concentration of 53.5 mmol�1; Mondazzi & Arcelli, 2009) which could impact bothphysical and cognitive performance (Faye et al., 2005). Theingestion of the LGI meal may also placed the fasted subjects atsome risk to hypoglycemia given the food’s minimal disruption toblood glucose levels and the prolonged duration of the post-prandial period (412 h). This study, in this regard, to the best ofour knowledge, is the first to monitor blood glucose concentrationlevels of Ramadan fasted individuals throughout the daytime fastperiod. There are several noteworthy observations. First, ofthe 288 blood glucose data points taken throughout the resting

post-prandial period, only two data points (0.7%) were 54.0 mmol�1. These two glucose values of 3.6 and 3.7 m mol�1 wereobtained, interestingly, from the same participant at the 12 h markpost-prandial his LGI and CON meal trials, respectively.However, it is important to note that the same participant didnot complain of any light-headedness throughout his exercise trialand that his blood glucose concentration reverted to normal levelsduring the exercise (blood glucose at the end of his 30 min steady-state run was 4.6 and 4.8 m mol�1 during his LGI and CON mealtrial, respectively). Overall, these results indicated that thereseemed to be minimal risk for hypoglycemia, if Muslimindividuals have had consumed their sahur meal, within thehour just before commencing the day’s fast.

There are several methodological limitations of this study.First, as previously mentioned, the small number of samples mayhave underpowered the study to detect statistically significantdifferences between interventions, particularly in regards to therunning velocity and HR data. Second, the glycemic index valueof �57 of the CON meal may be deemed as ‘‘medium’’ levelglycemic index meal; and thus the exclusion of a high glycemicindex meal (�70) trial is a major limitation. It is unclear if a muchhigher glycemic index meal will result in a much more superior(i.e. larger effects size difference) endurance performance to thatof LGI meal trial. Third, only the last main meal ingested(i.e. sahur meal) before the exercise trials were controlled. It maybe argued that the impact of ingesting different types of food priorto the sahur meal could have compounded the current findings,although a previous study has indicated that such accumulativeeffects are at best, minimal (Ning et al., 2009). Finally, this studydid not measure any clinical blood markers such as insulin, freefatty acids and epinephrine which would have allowed moreinsight into the mechanisms of the effects of ingesting LGI orCON meals in the Ramadan fasted state.

Conclusion

This study is the first attempt to examine a specific nutritionalstrategy to try to mitigate the adverse effects of Ramadan fastingon endurance exercise performance. The study’s result showedthat when ingesting either LGI (glycemic index¼ 37) or normalmixed (CON; glycemic index¼�57) carbohydrate for the sahurmeal during Ramadan, there were minimal differences in fastedindividuals’ subjective ratings of hunger and appetite, metabolicmarkers, and physiological responses during the 12 h post-prandial rest period and in the subsequent steady-state run at65% VO2max. The distance ran during the 30 min TT was,however, higher in the CON compared to the LGI meal trial. Thisindicates that the ingestion of LGI as the sahur meal did notprovide any clear metabolic, physiological or performancebenefits during endurance run performed 12 h post-prandial inthe Ramadan fasted state. Given the trivial differences in exercisemetabolism and exercise performance, as well as the lowerperceived exertion observed in the LGI compared to the CONmeal trial, further research is required to examine the efficacy ofingesting either low or high glycemic index meals before any firmrecommendation on the appropriate type of glycemic carbohy-drate food can be made to fasted Muslims who plan to exercise orcompete during Ramadan.

Declaration of interest

The authors declared that there is no conflict of interest that should bedisclosed in relation to the contents of the manuscript.

References

Aziz AR, Png W. 2008. Practical tips to exercise training during theRamadan fasting month. ISN Bull 1:13–19.

DOI: 10.3109/09637486.2014.886187 Glycemic index foods and Ramadan fasting 635

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.

Aziz AR, Chia MYH, Low CY, Slater GJ, Png W, Teh KC. 2012.Conducting an acute intense interval exercise session during theRamadan fasting month: what is the optimal time of the day?Chronobiol Int 29:1139–1150.

Aziz AR, Wahid MF, Png W, Jesuvadian CV. 2010. Effects of Ramadanfasting on 60 min of endurance running performance in moderatelytrained men. Br J Sport Med 44:516–521.

Brand Miller J, Foster-Powell, K, Colagiuri S. 1996. The GI factor – theglycaemic index solution. Sydney: Hodder and Stoughton.

Brisswalter J, Bouhlel E, Falola JM, Abbiss CR, Vallier JM, HausswirthC. 2011. Effects of Ramadan intermittent fasting on middle-distancerunning performance in well-trained runners. Clin J Sports Med 21:422–427.

Burke LM, King C. 2012. Ramadan fasting and the goals of sportsnutrition around exercise. J Sports Sci 30:S21–S31.

Burke LM, Collier GR, Hargreaves M. 1993. Muscle glycogen storageafter prolonged exercise: effect of glycemic index of carbohydratefeedings. J Appl Physiol 75:1019–1023.

Chennaoui M, Desgorces F, Drogou C, Boudjemaa B, Tomaszewski A,Depiesse F, Burnat P, et al. 2009. Effects of Ramadan fasting onphysical performance and metabolic, hormonal, and inflammatoryparameters in middle-distance runners. Appl Physiol Nutr Metab 34:587–594.

DeMarco HM, Sucher KP, Cisar CJ, Butterfield GE. 1999. Pre-exercisecarbohydrate meals: application of glycemic index. Med Sci SportsExerc 31:164–170.

Donaldson CM, Perry TL, Rose MC. 2010. Glycemic index andendurance performance. Int J Sport Nutr Exerc Metab 20:154–165.

Faye J, Fall A, Badji L, Cisse F, Stephan H, Tine P. 2005. Effects ofRamadan fast on weight, performance and glycemia during training forresistance. Dakar Med 50:146–151. [in French].

Febbraio MA, Chiu A, Angus DJ, Arkinstall MJ, Hawley JA. 2000.Effects of carbohydrate ingestion before and during exercise on glucosekinetics and performance. J Appl Physiol 89:2220–2226.

Flint A, Raben A, Blundell JE, Astrup A. 2000. Reproducibility, powerand validity of visual analogue scales in assessment of appetitesensations in single test meal studies. Int J Obes Rel Metab Disord 24:38–48.

Hopkins WG, Hawley JA, Burke LM. 1999. Design and analysis ofresearch on sports performance enhancement. Med Sci Sports Exerc31:472–485.

Jeacocke N, Burke LM. 2010. Methods to standardize dietary intakebefore performance testing. Int J Sport Nutr Exerc Metab 20:87–103.

Jenkins DJA, Wolever TMS, Taylor RH, Barker H, Fielden H, BaldwinJM. 1981. Glycemic index of foods: a physiological basis forcarbohydrate exchange. Am J Clin Nutr 34:362–366.

Jeukendrup AE, Wallis GA. 2005. Measurement of substrate oxidationrate during exercise by means of gas exchange measurements. Int JSports Med 26:S28–S37.

Jones AM, Doust JH. 1996. A 1% treadmill grade most accurately reflectsthe energetic cost of outdoor running. J Sports Sci 14:321–327.

Kaida K, Takahashi M, Akerstedt T, Nakata A, Otsuka Y, Haratani T,Fukasawa K. 2006. Validation of the Karolinska sleepiness scaleagainst performance and EEG variables. Clin Neurophysiol 117:1574–1581.

Karamanolis IA, Laparidis KS, Volaklis KA, Douda HT, Tokmakidis SP.2011. The effects of pre-exercise glycemic index food on runningcapacity. Int J Sports Med 32:666–671.

Liu AG, Most MM, Brashear MM, Johnson WD, Cefalu WT, GreenwayFL. 2012. Reducing the glycemic index or carbohydrate content ofmixed meals reduces postprandial glycemia and insulinemia over theentire day but does not affect satiety. Diabetes Care 35:1633–1637.

Ludwig DS. 2000. Dietary glycemic index and obesity. J Nutr 130:280S–283S.

Maughan RJ, Zerguini Y, Chalabi H, Dvorak J. 2012. Achieving optimumsports performance during Ramadan: some practical recommendations.J Sports Sci 30:S109–S117.

McConell G, Fabris S, Proietto J, Hargreaves M. 1994. Effect ofcarbohydrate ingestion on glucose kinetics during exercise. J ApplPhysiol 77:1537–1541.

Mondazzi L, Arcelli E. 2009. Glycemic index in sport nutrition. J AmColl Nutr 28:455S–463S.

Moore LJ, Midgley AW, Thurlow S, Thomas G, McNaughton LR. 2010.Effect of the glycaemic index of a pre-exercise meal on metabolism andcycling time trial performance. J Sci Med Sport 13:182–188.

Ning B, Brown RC, Venn BJ, Williams SM, Green TJ. 2009. The effect ofthe fat and carbohydrate contents in the evening meal preceding GItesting on GI. Eur J Clin Nutr 64:224–226.

Nicholas CW, Green PA, Hawkins RD, Williams C. 1997. Carbohydrateintake and recovery of intermittent running capacity. Int J Sport Nutr 7:251–260.

Oliver SJ, Costa RJS, Laing SJ, Bilzon JL, Walsh NP. 2009. One night ofsleep deprivation decreases treadmill endurance performance. Eur JAppl Physiol 107:155–161.

Reynolds RC, Stockmann KS, Atkinson FS, Denyer GS, Brand-Miller JC.2009. Effect of the glycemic index of carbohydrates on day-long (10 h)profiles of plasma glucose, insulin, cholecystokinin and ghrelin. Eur JClin Nutr 63:872–878.

Roberts SB. 2003. Glycemic index and satiety. Nutr Clin Care 6:20–26.Sparks MJ, Selig SS, Febbraio MA. 1998. Pre-exercise carbohydrate

ingestion: effect of the glycemic index on endurance exerciseperformance. Med Sci Sports Exerc 30:844–849.

Stevenson EJ, Williams C, Mash LE, Phillips B, Nute ML. 2006.Influence of high-carbohydrate mixed meals with different glycemicindexes on substrate utilization during subsequent exercise in women.Am J Clin Nutr 84:354–360.

Stevenson E, Williams C, Nute M, Swaile P, Tsui M. 2005. The effect ofthe glycemic index of an evening meal on the metabolic responses to astandard high glycemic index breakfast and subsequent exercise inmen. Int J Sport Nutr Exerc Metab 15:308–322.

Terry PC, Lane AM, Lane HJ, Keohane L. 1999. Development andvalidation of a mood measure for adolescents. J Sports Sci 17:861–872.

Thomas DE, Brotherhood JR, Brand JC. 1991. Carbohydrate feedingbefore exercise: effect of glycemic index. Int J Sports Med 12:180–186.

van Hall G, Shirreffs SM, Calbet JA. 2000. Muscle glycogen resynthesisduring recovery from cycle exercise: no effect of additional proteiningestion. J Appl Physiol 88:1631–1636.

Wolever TMS, Jenkins DJA, Jenkins AL, Josse RG. 1991. The glycemicindex: methodology and clinical implications. Am J Clin Nutr 54:846–854.

Wolever TM, Jenkins DJ. 1986. The use of the glycemic index inpredicting the blood glucose response to mixed meals. Am J Clin Nutr43:167–172.

Wee SL, Williams C, Gray S, Horabin J. 1999. Influence of high and lowglycemic index meals on endurance running capacity. Med Sci SportsExerc 31:393–399.

Wong SH, Chen YJ, Fung WM, Morris JG. 2009. Effect of glycemicindex meals on recovery and subsequent endurance capacity. Int JSports Med 30:898–905.

Wu CL, Nicholas C, Williams C, Took A, Hardy L. 2003. The influenceof high-carbohydrate meals with different glycaemic indices onsubstrate utilisation during subsequent exercise. Br J Nutr 90:1049–1056.

636 W. Png et al. Int J Food Sci Nutr, 2014; 65(5): 629–636

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Tor

onto

on

08/1

1/14

For

pers

onal

use

onl

y.