A Low-Glycemic Index Meal and Bedtime Snack Prevents Postprandial Hyperglycemia and Associated Rises...

A Low–Glycemic Index Mealand Bedtime Snack PreventsPostprandial HyperglycemiaandAssociatedRises inInflammatoryMarkers, Providing ProtectionFrom Early but Not LateNocturnal HypoglycemiaFollowing Evening Exercise inType 1 DiabetesDiabetes Care 2014;37:1845–1853 | DOI: 10.2337/dc14-0186

OBJECTIVE

To examine the influence of the glycemic index (GI) of foods consumed afterevening exercise on postprandial glycemia, metabolic and inflammatory markers,and nocturnal glycemic control in type 1 diabetes.

RESEARCH DESIGN AND METHODS

On two evenings (∼1700 h), 10 male patients (27 6 5 years of age, HbA1c 6.7 6

0.7% [49.9 6 8.1 mmol/mol]) were administered a 25% rapid-acting insulin dosewith a carbohydrate bolus 60 min before 45 min of treadmill running. At 60 minpostexercise, patients were administered a 50% rapid-acting insulin dosewith oneof two isoenergetic meals (1.0 g carbohdyrate/kg body mass [BM]) matched formacronutrient content but of either low GI (LGI) or high GI (HGI). At 180 minpostmeal, the LGI group ingested an LGI snack and the HGI group an HGI snack(0.4 g carbohdyrate/kg BM) before returning home (∼2300 h). Interval sampleswere analyzed for blood glucose and lactate; plasma glucagon, epinephrine,interleukin-6 (IL-6), and tumor necrosis factor-a (TNF-a); and serum insulin,cortisol, nonesterified fatty acid, and b-hydroxybutyrate concentrations. Intersti-tial glucose was recorded for 20 h postlaboratory attendance through continuousglucose monitoring.

RESULTS

Following the postexercise meal, an HGI snack induced hyperglycemia inall patients (mean 6 SD glucose 13.5 6 3.3 mmol/L) and marked increases inTNF-a and IL-6, whereas relative euglycemia was maintained with an LGI snack(7.76 2.5mmol/L, P < 0.001) without inflammatory cytokine elevation. Bothmealtypes protected all patients from early hypoglycemia. Overnight glycemia wascomparable, with a similar incidence of nocturnal hypoglycemia (n = 5 for bothHGI and LGI).

CONCLUSIONS

Consuming LGI food with a reduced rapid-acting insulin dose following eveningexercise prevents postprandial hyperglycemia and inflammation and provideshypoglycemia protection for ∼8 h postexercise; however, the risk of late noctur-nal hypoglycemia remains.

1Faculty of Health and Life Sciences, Northum-bria University, Newcastle-upon-Tyne, U.K.2Institute of Cellular Medicine, Newcastle Uni-versity, Newcastle-upon-Tyne, U.K.3Applied Sports, Technology, Exercise and Medi-cine Research Centre, College of Engineering,Swansea University, Swansea, U.K.

Corresponding author: Daniel J. West, [email protected].

Received 21 January 2014 and accepted 18March 2014.

Clinical trial reg. no. NCT01794650, clinicaltrials.gov.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc14-0186/-/DC1.

© 2014 by the American Diabetes Association.See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

Matthew D. Campbell,1 Mark Walker,2

Michael I. Trenell,2 Emma J. Stevenson,1

Daniel Turner,3 Richard M. Bracken,3

James A. Shaw,2 and Daniel J. West1

Diabetes Care Volume 37, July 2014 1845

CLIN

CARE/ED

UCATIO

N/N

UTR

ITION/PSYC

HOSO

CIAL

There is a growing evidence base sur-rounding the wide range of health ben-efits from regular exercise for type 1diabetic patients (1). However, exerciseremains the most frequently identifiedspecific cause of severe hypoglycemia(2), the fear of which remains the pri-mary obstacle to patients wishing to en-gage in regular exercise (3).Strategies to combat exercise-induced

hypoglycemia, such as manipulatingexercise intensity (4), insulin dose, diet(5–7), or the order in which various ex-ercise types are undertaken (8), havepredominantly been tested for morningexercise (5–7). However, many individu-als prefer to exercise in the evening be-cause of study and work commitmentsor for social reasons. Unfortunately, ex-ercise in the evening is associated witha greater risk of postexercise hypoglyce-mia (4,9), with low blood glucose levelslikely to occur particularly nocturnally(2). Incorporation of evening exercisesafely into the lives of people with type1 diabetes is thus significantly hamperedby the lack of appropriate evidence nec-essary for informed self-managementstrategies.We recently demonstrated that meal-

time insulin adjustment, specifically, re-ducing the dose of rapid-acting insulinbefore and after exercise, is vital to min-imizing the risk of postexercise hypo-glycemia (5). However, little advicecurrently exists on optimal carbohy-drate types for patients with type 1 di-abetes who exercise (1). AmericanDiabetes Association guidance focuseson the quantity rather than on the com-position of the carbohydrate to be con-sumed after exercise (10). Consumptionof;5 g carbohydrate per kilogram bodymass (BM) is typically recommended formoderate-intensity exercise (11,12);however, the food composition is alsoan important consideration becausethe type of carbohydrate can exert ama-jor influence on postprandial glycemia indiabetic patients (13). Meals containingidentical macronutrient compositionsare digested and absorbed at varyingrates, producing a range of glycemic re-sponses (14). Carbohydrate-rich foodswith a low glycemic index (LGI) elicit amore gradual rise and fall in blood glu-cose levels compared with their highglycemic index (HGI) equivalents. As aresult, more favorable postprandial gly-cemic profiles have been shown after

ingestion of LGI foods in patients withtype 1 diabetes (15–17).

Thus, optimizing postexercise glyce-mia may be possible by manipulatingthe composition of foods and drinksconsumed during this time. The pro-tracted absorption rates of LGI foodsand drinks may be beneficial for reduc-ing postprandial hyperglycemia. How-ever, slower delivery of carbohydratesto postexercise musculature, and po-tentially slower rates of muscle glyco-gen replenishment following exercise(18,19), may increase the risk of postex-ercise hypoglycemia (11,20). Inversely,consuming HGI foods may promoteaccelerated muscle glycogen restora-tion (18,19), reducing the incidence ofpostexercise hypoglycemia (11,20).However, the need to reduce the insulin-to-carbohydrate ratio may be associ-ated with postprandial hyperglycemiafollowing ingestion of HGI foods (15,16),potentially leading to metabolic, hor-monal, and inflammatory disturbances(21–23). The aim of the current studywas to examine the influence of theglycemic index (GI) of a meal and sub-sequent bedtime snack consumedafter evening exercise on postprandialglycemia, metabolism, and circulatinginflammatory markers in addition tonocturnal glycemic control in type 1diabetes.

RESEARCH DESIGN AND METHODS

PatientsEligibility criteria were age 18–35years; a duration of diabetes .2 years;an HbA1c ,8.0% (64 mmol/mol); an ab-sence of diabetes-related complica-tions, including impaired awareness ofhypoglycemia; and insulin therapy alonewithout any other medication. Ten malepatients with type 1 diabetes were re-cruited (mean 6 SD age 27 6 5 years,BMI 25.5 6 0.9 kg/m2, duration of di-abetes 15 6 6 years, HbA1c 6.7 6 0.7%[49.9 6 8.1 mmol/mol], VO2peak 52 64 mL/kg/min). C-peptide was not mea-sured. All patients had been treated on astable basal-bolus regimen comprisinginsulin aspart and once-daily insulinglargine for a minimum of 6 months.Fifty percent of patients were injectinginsulin glargine in the morning and 50%in the late evening or before bed. Allpatients undertook regular and consis-tent exercise (participating in aerobic-based exercise for at least 30 min at a

time at least three times a week). Allpatients were familiar with carbohy-drate counting, administering 1.0 60.7 units of insulin aspart per 10 g ofcarbohydrate. All patients successfullycompleted the study.

Following approval from the local Na-tional Health Service Research EthicsCommittee, fully informed written con-sent was obtained from all patients. Pa-tients first attended the NewcastleNational Institute for Health ResearchClinical Research Facility exercise labo-ratory for a preliminary screening visit,as described by Campbell et al. (5), be-fore returning on three more occasions.On visit 1, peak cardiorespiratory pa-rameters were collected during thecompletion of an incremental–maximaltreadmill run protocol, as previously de-scribed (5,6). Computer randomizationwas then used to determine the se-quence of the two subsequent experi-mental visits.

Prelaboratory Phase

Continuous Glucose Monitoring

Patients were fitted with a continuousglucose monitor (CGM) (Paradigm Veo;Medtronic Diabetes, Northridge, CA)using an Enlite Sensor (MedtronicMiniMed; Northridge, CA) for a mini-mum of 48 h before attending the labo-ratory on each occasion. The ParadigmVeo provides real-time glucose profilesas part of an insulin pump. Patients didnot use the continuous subcutaneousinsulin infusion facility, however, butcontinued their usual basal-bolus regi-men. Glucose alerts were set at #3.5and$16 mmol/L during the pretrial pe-riod. The high glucose alert was discon-tinued once patients left the laboratoryafter the experimental trials. Sensorswere placed in the posterolateral ab-dominal region to minimize the physio-logical time lag between blood andinterstitial glucose (24). Insertion sitewas replicated across visits. During sen-sor wear, patients performed a mini-mum of four daily capillary bloodglucose tests (GlucoMen LX; MenariniDiagnostics, Berkshire, U.K.), enteringall values into the CGM device for cali-bration. Capillary glucose values and notCGM data were used to inform self-administered insulin aspart doses.Downloaded data were retrospectivelyprocessed and analyzed using CareLinkPro software (Medtronic Diabetes).

1846 Glycemic Index of Postexercise Food Diabetes Care Volume 37, July 2014

CGM data obtained from each patientwere complete; there were no missingdata streams in CGM recordings. Themean absolute difference between in-terstitial glucose and capillary blood glu-cose meter readings over both trials was1.4 6 1.1 mmol/L.

Diet and Activity Replication

Over the 24 h preceding main trial visits,patients replicated their diet (assessedusing weighed dietary recording sheets)and were instructed to maintain theirnormal insulin regimen, with basaldose standardized (dose, injection site,and time of injection) across trials. Dur-ing this time, patients used a pedometer(Omron Healthcare Europe B.V., Hoofd-dorp, the Netherlands) to record totalstep count. Avoidance of strenuous ex-ercise was required in the previous 48 h,with maintenance of similar activitypatterns between trials, which wereseparated by at least 7 days. On theday of the trial, patients were providedwith two standardized meals, a cereal-based breakfast (sugar-coated cornflakes, semiskimmedmilk, and peaches)equating to 1.3 g carbohdyrate/kg BM(5496 46 kcal) and a pasta-based lunch(pasta, tomato-based sauce, cheddarcheese, olive oil) equating to 1.3 gcarbohdyrate/kg BM (968 6 62 kcal).Meal composition was based on the ha-bitual dietary patterns of patients withtype 1 diabetes and current recommen-dations for exercise in diabetic patients(11,12). When combined with meals pro-vided in the laboratory during experimen-tal trials, total dietary intake across theday was calculated to constitute ;5.0 gcarbohdyrate/kg BM, with a macronutri-ent content consisting of 77% carbohy-drate, 12% fat, and 11% protein (11,12).

Testing ProcedurePatients arrived at the laboratory in thelate afternoon (;1700 h), replicatingtheir start time across conditions. A12-mL resting venous blood samplewas taken of which 20 mL was used forthe immediate quantification of bloodglucose and lactate (Biosen C-Line; EKFDiagnostic GmbH, London, U.K.) and 10mL was analyzed for hemoglobin and he-matocrit (Hemo Control; EKF DiagnosticGmbH) used to correct for changes inplasma volume (25). The remaining sam-ple was measured by equal aliquots intolithium-heparin and serum (Vacuette;Greiner Bio-One GmbH, Kremsmunster,

Austria) separation tubes; centrifugedfor 15 min at 3,000 rpm at 48C; andstored at 2808C for retrospective anal-ysis of serum rapid-acting insulin analog(Invitron Insulin Assay; Invitron, Mon-mouth, U.K.) (see West et al. (6) fordetails of assay cross-reactivity), cortisol(Cortisol Parameter Assay Kit; R&DSystems, Roche Diagnostics, WestSussex, U.K.), nonesterified fatty acidsby colorimetric assay (RANBUT; RandoxLaboratories, London, U.K.), and b-hydroxybutyrate by D-3-hydroxybutyratekinetic enzymatic assay (RANBUT, witha lower limit of detection of 0.004mmol/L) and of plasma glucagon (Glu-cagon EIA; Sigma-Aldrich, St. Louis, MO),adrenaline (CAT ELISA; Eagle Biosciences,London, U.K.), interleukin-6 (IL-6) (Hu-man IL-6 Quantikine ELISA; R&D Sys-tems, Roche Diagnostics), and tumornecrosis factor-a (TNF-a) (Human TNF-aQuantikine ELISA; R&D Systems, RocheDiagnostics). The coefficient of varia-tion was ,10% for all biochemicalanalyses.

Immediately after the resting sample,patientswere administered a 25% (2.060.4 units) dose (i.e., a 75% reduction) ofinsulin aspart into the abdomen, withthe injection site standardized across tri-als as equidistant between the iliac crestand naval based on current recom-mendations (5,6,26). Patients thenconsumed a pre-exercise carbohydratebolus (sugar-coated corn flakes, semi-skimmed milk, and peaches) equatingto 1.0 g carbohdyrate/kg BM (423 637 kcal) within a 5-min period (5).

Patients remained at rest for 60 minfollowing consumption of the pre-exercise carbohydrate bolus. On 60 min,a blood sample was drawn immediatelybefore commencing 45 min of treadmill(Woodway, Weil am Rhein, Germany)running at a speed calculated to elicit70% of their VO2peak, an intensity fallingwithin recommendations of the Ameri-can College of Sports Medicine (27)for diabetic patients who exercise.Breath-by-breath respiratory parame-ters (MetaLyzer 3B; CORTEX, Leipzig,Germany) and heart rate (S810; Polar,Kempele, Finland) were continuouslyrecorded during exercise. Immediatelyfollowing cessation of exercise, a bloodsample was taken, with subsequentsamples taken at 15, 30, and 60 minpostexercise. At 60 min postexercise,patients were administered a 50%

(4.0 6 0.8 units) dose of insulin aspartinto the contralateral abdominal site tothe pre-exercise insulin aspart injectionsite (5). With this, in a random and coun-terbalanced fashion, patients were as-signed to consume one of two eveningmeals calculated to be of either LGI orHGI. Following this meal, patients con-tinued to rest with further blood sam-ples taken every 30 min for 180 min. Allpatients then consumed a trial-specificbedtime snack of either LGI or HGI. Pa-tients could drink water ad libitumthroughout. All patients received trans-portation home and were instructed tocontinue their usual basal insulin doseand replicate sleeping patterns as bestas possible across trials. Hypoglycemiawas defined as a blood or interstitial glu-cose concentration of#3.9mmol/L, andhyperglycemia was defined at $8.0mmol/L (5).

Meal Composition and Bedtime SnackAll meals were preprepared by the re-search team and comprised food toelicit either an HGI or LGI response.The beverage component of the mealand bedtime snack contained eitherHGI maltodextrin or LGI isomaltulose(Palatinose; BENEO, Mannheim, Ger-many) and was calculated to be a 10%solution. We calculated the GI of eachmeal using methods described byBrouns et al. (28) in 10 nondiabetic con-trol participants. Patients consumingthe LGI evening meal subsequentlyconsumed the LGI bedtime snack, andthose consuming the HGI evening mealconsumed the HGI snack. Both even-ing meals and bedtime snacks werematched for macronutrient contentand palatability and had negligiblefiber content (Supplementary Table 1).Bedtime snacks equated to 0.4 g car-bohdyrate/kg BM (29).

Calculation of Substrate OxidationDuring exercise, at 15 min before thepostexercise meal, and at 45, 105, and165 min following the postexercisemeal, expired gases were analyzed(MetaLyzer 3B). Substrate oxidationrates and energy expenditure weredetermined from VO2 and CO2 pro-duction values using stoichiometricequations (30).

Postlaboratory PeriodWhile wearing a CGM, patients contin-ued to self-record and replicate their

care.diabetesjournals.org Campbell and Associates 1847

diet throughout both trials using aweighed food diary. Patients were re-quired to report additional carbohy-drate ingestion and administration ofcorrective doses of insulin aspart, andinstructed to keep meal times as wellas insulin aspart and insulin and glarginedoses consistent across trials.

Data AnalysisStatistical analysis was performed us-ing PASW Statistics 18 software (IBM,Armonk, NY). A repeated-measuresANOVA on two levels (condition andtime) was conducted, with Bonferroni-corrected pairwise comparisons andpaired sample t tests used to examinetime and condition effects, respectively.Statistical significance was accepted atP # 0.05. Area under the curve was cal-culated using the methods described byWolever and Jenkins (31).

RESULTS

Prelaboratory PhaseGlycemic control was comparable overthe 24 h before the patients’ arrival atthe laboratory for both experimental tri-als (CGM mean glucose: HGI 7.9 6 2.2mmol/L, LGI 7.9 6 2.2 mmol/L, P =0.465; total interstitial glucose area un-der the curve: HGI 11,277 6 3,208mmol/L/min, LGI 10,971 6 3,186mmol/L/min, P = 0.215). Dietary intake

was also similar during the 24 h beforeboth trials. There were no differences intotal energy consumed (HGI 2,143 6673 kcal, LGI 2,358 6 668 kcal, P =0.508), with similar contributions fromcarbohydrates (HGI 516 11%, LGI 46610%, P = 0.896), fat (HGI 30 6 9%, LGI32 6 11%, P = 0.301), and protein (HGI20 6 5%, LGI 22 6 10%, P = 0.556). Thetotal amount of insulin administered(HGI 26 6 13 units, LGI 26 6 14 units,P = 0.609) and levels of activity (HGI6,949 6 105 steps, LGI 7,041 6 118steps, P = 0.372) were comparableover the 24 h before each trial.

Laboratory PhaseThere was a significant time effect (P ,0.001), condition effect (P = 0.05), andcondition3 time interaction for absoluteblood glucose concentrations (P, 0.001)(Fig. 1). Blood glucose values were com-parable before the standardized pre-exercise carbohydrate bolus and insulininjection and after the 1-h pre-exerciserest period during both experimental tri-als (Fig. 1). Serum insulin and all otherhormone andmetabolite levels were sim-ilar at rest and immediately before exer-cise (P . 0.05) (Fig. 2A and B, Table 1).

Patients exercised at a similar inten-sity (%VO2peak: HGI 776 0.09, LGI 7460.09, P = 0.352; peak heart rate: HGI 8066, LGI 79 6 7, P = 0.631). Patients ran

at a velocity of 10.1 6 1.0 km/h, com-pleting 7.6 6 0.7 km and expending7186 143 kcal. Similar peak lactate lev-els were elicited immediately postexer-cise (HGI 4.1 6 2.4 mmol/L, LGI 4.2 62.7 mmol/L, P = 0.137) (Table 1). Exer-cise induced a similar decrease in bloodglucose from pre-exercise concentra-tions (HGI 25.4 6 1.6 mmol/L, LGI26.8 6 1.3 mmol/L, P = 0.733) (Fig.1), such that immediately followingthe cessation of exercise, blood glucosevalues were comparable to baseline un-der both conditions (P = 0.304) (Fig. 1).There were no incidences of hypoglyce-mia during exercise, with all patientscompleting the exercise protocol onboth occasions. Immediately beforethe postexercise meal, serum insulinconcentrations were similar to restingconcentrations (P . 0.05) (Table 1), aswere all other hormone, metabolite,and cytokine levels (P . 0.05) (Fig.2A–C, Table 1).

Postexercise InterventionSerum insulin peaked similarly at 60 minfollowing the postexercise meal beforedeclining under both conditions, withconcentrations returning to resting val-ues at 180 min (P . 0.05) (Table 1).Blood glucose levels increased over the180 min after both postexercise meals,but this was significantly attenuated

Figure 1—Time-course changes in blood glucose from rest, during exercise, and over 3 h postexercise. Data are presented as mean6 SD (error bars).-, LGI; ◆, HGI; , and ◇, significant difference from premeal concentrations (P# 0.05). *Significant difference between conditions (P# 0.05).Shaded area indicates exercise; dashed line indicates postexercise meal intervention. Note that the test meal and insulin were administeredimmediately following rest and 60 min postexercise.


under LGI compared with HGI (P, 0.05)(Fig. 1). Over this time, all patients wereprotected from hypoglycemia underboth conditions. However, all patientswere exposed to hyperglycemia afterthe HGI meal, whereas this was limitedto four patients after LGI. Moreover, hy-perglycemia was less pronounced(mean peak blood glucose: LGI 8.8 63.1 mmol/L, HGI 15.9 6 3.8 mmol/L)and tended to be only transient (timespent while hyperglycemic: LGI 81 643 min, HGI 165 6 32 min) followingthe LGI meal. On leaving the laboratory,blood glucose remained significantlygreater after the HGI meal (HGI 12.7 64.8 mmol/L, LGI 7.5 6 2.5 mmol/L, P =0.004) (Fig. 1), with more patients leav-ing the laboratory hyperglycemic (n = 9HGI; n = 4 LGI).

Counterregulatory hormonal andmetabolic responses are presented inTable 1 with inflammatory cytokineand b-hydroxybutyrate responses shownin Fig. 2A–C. There were no differencesin serum b-hydroxybutyrate concen-trations between the two experimen-tal trials (P . 0.05) (Fig. 2A). Followingthe postexercise meal, IL-6 and TNF-aconcentrations significantly increasedfrom rest and premeal concentrationsin the HGI trial and were significantlygreater than LGI during the postpran-dial period (P , 0.05) (Fig. 2B and C).During this period, concentrations inthe LGI trial were significantly lowerthan baseline measures (P , 0.05)(Fig. 2B and C). There were no differen-ces in substrate oxidation responsesduring the postexercise meal period ofboth trials, with carbohydrate (HGI14.5 6 4.1 g/h, LGI 14.7 6 4.0 g/h, P =0.927) and lipid (HGI 3.06 01.2 g/h, LGI3.16 1.2 g/h, P = 0.809) oxidation ratessimilar.

Postlaboratory Phase

Late Evening

After leaving the laboratory, interstitialglucose concentrations in the HGI trialwere significantly greater than LGI in thetime before sleep (Fig. 3), with individ-ualized mean peak interstitial glucoselevels higher (HGI 18.3 6 4.1 mmol/L,LGI 13.9 6 2.5 mmol/L, P = 0.009).

Nocturnal Glycemic Control

During sleep, falling glucose levels wereevident under both conditions such thatconcentrations became comparable 8 hafter exercise (P . 0.05) (Fig. 3). Five

Figure 2—Time-course changes in serum b-hydroxybutyrate concentration (A), plasma IL-6concentration (B), and plasma TNF-a concentration (C). Data are presented as mean6 SD (errorbars).-, LGI;◆, HGI;, and◇, significant difference from premeal concentrations (P# 0.05).*Significant difference between conditions (P # 0.05). Shaded area indicates exercise; dashedline indicates postexercise meal intervention. Note that the test meal and insulin were admin-istered immediately following rest and 60 min postexercise.


patients during the LGI trial and fiveduring the HGI trial experienced noctur-nal hypoglycemia. Some patients expe-rienced multiple bouts of hypoglycemia(n = 10 HGI; n = 8 LGI). Mean interstitialglucose nadir was similar between con-ditions (HGI 3.66 1.9mmol/L, LGI 3.461.7 mmol/L, P = 0.650), as was timespent in hypoglycemic (P = 0.569), eu-glycemic (P = 0.705), and hyperglycemic(P = 0.765) ranges (Fig. 3).

Next-Day Glycemic Responses

On waking, interstitial glucose levelswere comparable (HGI 8.5 6 2.8mmol/L, LGI 8.3 6 2.8 mmol/L, P =0.614) (Fig. 3), and glycemia remainedsimilar between conditions for the re-mainder of the 24-h postexercisewindow (P . 0.05). During the postla-boratory period, total energy consumed(HGI 7196 256 kcal, LGI 6866 289 kcal,P = 0.774) was similar, with contributionfrom carbohydrate (HGI 72 6 17%, LGI64 6 26%, P = 0.767), fat (HGI 20 615%, LGI 226 22%, P = 0.834), and pro-tein (HGI 8 6 10%, LGI 14 6 19%, P =0.548) also similar between conditions.Activity patterns for 24 h after exercisewere comparable (HGI 6,0866 94 steps,LGI 6,478 6 112 steps, P = 0.369).

CONCLUSIONS

The aim of this study was to determinewhether manipulating the GI of foodsand drinks consumed following eveningexercise could modulate postprandialglycemia and metabolism to provideprotection from postexercise hypergly-cemia and hypoglycemia in patientswith type 1 diabetes. To our knowledge,this study is the first to show that con-sumption of LGI food under conditionsof reduced rapid-acting insulin dose af-ter evening exercise improves postpran-dial glycemia, reducing hyperglycemiaand concentrations of circulating in-flammatory markers in combinationwith providing protection from hypogly-cemia for;8 h after exercise. However,beyond this time, risk of late-onset noc-turnal hypoglycemia persists, regardlessof the GI of the postexercise meal andbedtime snack.

We recently demonstrated the im-portance of reducing the rapid-acting in-sulin dose administered with the mealafter as well as before exercise to extendthe period of protection from postexer-cise hypoglycemia (5). Now, we demon-strate that under these conditions, the

Table

1—Se

rum

insu

linco

unterreg

ulatory

horm

onalandbloodmetabolite

resp

onse

sto

postexe

rcisemealsofva

ryingGIs

ANOVAPvalue

Rest

60Exer.

015

30Prem

eal

3060

90120

150

180

TT*C

Serum

insulin

(pmol/L)

HGI

1266

471376

582016

112†

1386

621286

591056

491796

101‡

1806

97‡

1436

68‡

1276

69‡

956

5394

656

,0.001

0.992

LGI

1246

521506

642036

108†

1376

641256

571026

421726

99‡

1746

97‡

1446

57‡

1266

54‡

1056

4998

647

Plasmaglucagon

(pg/mL)

HGI

7306

280

5916

286†

6826

316

7606

360

7686

383

8336

332

9536

437†

9226

437†

8726

410†

7986

399

6906

284‡

6696

286‡

,0.001

0.306

LGI

7336

290

6116

296†

6586

309

7926

369

8186

403

8166

355

9476

481†‡

9376

451†‡

9076

421†

8626

398

8066

340

8406

394

Plasmaadrenaline

(nmol/L)

HGI

0.096

0.05

0.156

0.10

0.556

0.35†

0.356

0.37†

0.156

0.10

0.156

0.12

0.176

0.07

0.166

0.11

0.116

0.12

0.096

0.06

0.116

0.09

0.086

0.08

0.013

0.497

LGI

0.086

0.06

0.156

0.08

0.546

0.40†

0.286

0.32†

0.146

0.11

0.146

0.12

0.156

0.05

0.136

0.10

0.116

0.07

0.116

0.06

0.106

0.08

0.076

0.05

Serum

cortisol

(nmol/L)

HGI

0.176

0.09

0.186

0.07

0.286

0.06†

0.336

0.13†

0.246

0.09†

0.196

0.08

0.146

0.07

0.146

0.08

0.136

0.06‡

0.116

0.05†

‡0.086

0.03†‡

0.086

0.05†

‡,0.001

0.099

LGI

0.176

0.09

0.156

0.05

0.246

0.11†

0.326

0.16†

0.236

0.12†

0.186

0.10

0.136

0.06‡

0.126

0.05‡

0.106

0.04†‡

0.106

0.05‡

0.106

0.03‡

0.096

0.04‡

Serum

NEFA

(mmol/L)

HGI

0.186

0.15

0.126

0.10

0.256

0.19

0.356

0.20†

0.436

0.37†

0.536

0.49†

0.376

0.25

0.246

0.20‡

0.246

0.25‡

0.276

0.36‡

0.276

0.29‡

0.356

0.43

0.011

0.514

LGI

0.276

0.24

0.186

0.10

0.276

0.23

0.346

0.23†

0.336

0.24

0.396

0.32†

0.396

0.33†

0.276

0.22

0.246

0.16

0.246

0.16

0.276

0.18

0.306

0.16

Bloodlactate

(mmol/L)

HGI

1.06

0.7

1.16

0.9

4.16

2.4†

2.16

1.5†

1.36

1.1

1.06

0.9

0.76

0.6

0.96

0.6

0.86

0.4

0.76

0.5

0.56

0.4

0.46

0.4†

0.001

0.129

LGI

0.96

0.2

1.06

0.6

4.26

2.7†

1.76

0.4

1.26

0.3

1.06

0.2

0.86

0.2†

1.06

0.3

1.16

0.3

1.26

0.2‡

0.66

0.2

0.56

0.2

Dataaremean6

SD.Testmealandinsulin

wereadministeredim

med

iatelyfollowingrestandpremealsam

plepoints.Exercisecommen

ced60

minafterrest.C,condition;Exer.,exercise;NEFA,nonesterified

fatty

acid;T,time.

*Significantlydifferentbetweenconditions(P

#0.05).†Sign

ificantlydifferentfrom

rest.‡

Sign

ificantlydifferentfrom

premeal.


composition of the postexercise mealplays an important role in modula-ting postprandial glycemia. Blood glu-cose concentrations were significantlygreater with the HGI than with the LGIpostexercise meal and snack, conse-quently exposing all patients in the HGIcondition to hyperglycemia during thelaboratory observation period. Con-versely, the incidence of hyperglycemiawas reduced by 60% after the LGI condi-tion (40%) versus the HGI condition(100%). Indeed, in the affected patients,hyperglycemia was less pronounced andtended to be only transient and shortlasting after LGI meals. Despite clearpostprandial differences in glycemia be-tween the two conditions, all patientswere still protected from hypoglycemiaduring their time in the laboratory. Pres-ently, there are relatively few dietaryguidelines to assist individuals with type1 diabetes in managing postexercise gly-cemia. However, we have shown that byconsuming an LGI postexercise meal anddrink, postprandial hyperglycemia can bereduced without exposure to hypoglyce-mia. This observation is important be-cause the aim of diabetes managementis to normalize blood glucose concentra-tions (32), especially when incorporatingexercise into patients’ lives (1).Given the potential for such large dif-

ferences in postprandial glycemia with

this strategy, we examined this impacton metabolic, hormonal, and inflam-matory measures. This is importantbecause regular exposure to meta-bolic, hormonal, or inflammatory dis-turbances could significantly influencelong-term diabetes-related complica-tions in patients who regularly exercise(26). Here, we show that meal GI hassignificant implications for postprandialcirculating inflammatory markers; spe-cifically, we demonstrate for the firsttime with nonclamp techniques andreplication of free-living conditionsthat TNF-a and IL-6 were dramaticallyincreased following an HGI meal. Anotherwise comparable LGI meal com-pletely prevented rises in these inflam-matory cytokines. The clinical relevanceof these findings should not be under-estimated because offsetting hypergly-cemia and inflammation is important forpreventing early pathogenic diabetes-related complications (23). Additionally,b-hydroxybutyrate concentrations didnot rise significantly during either con-dition (Fig. 2A), remaining similar topremeal and resting concentrations.Basal insulin dose remained unchanged,and despite a reduction in rapid-actinginsulin dose, circulating insulin con-centrations remained sufficient for asuppression in b-hydroxybutyrate pro-duction (33) and to drive ketone body

disposal (34). Concentrations duringboth trial conditions were well belowthose deemed clinically significant(.1.0 mmol/L) (22).

When type 1 diabetic patients exer-cise in the evening, consumption of acarbohydrate-based snack before bedis recommended (29). Blood glucosewas typically within the euglycemicrange before the consumption of thebedtime snack following LGI but stillhyperglycemic following HGI (LGI ;7.5mmol/L, HGI ;12.2 mmol/L). Outsideformal studies, patients within normalblood glucose range before bed oftenchoose to raise blood glucose concen-trations by consuming a carbohydrate-based snack (29) due to fear of nocturnalhypoglycemia (35). However, patientsin the hyperglycemic range before bedmay be tempted to administer correctiveinsulinunits,which in anexercise-inducedinsulin-sensitized state (36,37) is likelyto cause a rapid fall in glucose duringthe night. Avoidance of the bedtimesnack, and hence missing a valuablesource of carbohydrate before sleep, islikely to exacerbate the risk of nocturnalhypoglycemia. Despite large differencesin blood glucose concentrations beforebed following HGI and LGI in the currentstudy, levels fell in both conditions, be-coming comparable at 3 h after consum-ing the bedtime snack, with similar rates

Figure 3—Time-course changes in interstitial glucose throughout the postlaboratory period. Data are presented as mean6 SD. Red trace, HGI; bluetrace, LGI. *Interstitial glucose area under the curve is significantly different between conditions (P # 0.05). Vertical lines indicate nocturnal ordaytime periods. End of nocturnal period indicates when patients awoke.


of nocturnal hypoglycemia thereafter.This finding indicates the patients are atrisk for late-onset nocturnal hypoglyce-mia despite the consumption of a bed-time snack, with a predicted nadir .8 hpostexercise (5,9) and regardless of theGI of the snack or blood glucose levelsbefore bed.So that we could investigate the im-

pact of the GI of evening meals andsnacks, patients consumed enough car-bohydrates (consuming 2.6 g/kg BMduring the evening) to cover the costof the bout of exercise, with patientsusing ;1.7 g/kg in total during exerciseand with total daily carbohydrate intakematching current recommendations[;5.0 g/kg (11,12)], thus establishing apositive carbohydrate balance. Despiteconsuming sufficient carbohydrates forthe recovery of muscle glycogen postex-ercise and perhaps even consumingmore carbohydrate than is typical, hypo-glycemia was still encountered late afterexercise in the early hours of the morn-ing. Thesefindings direct attention towardthe role of basal insulin administration inavoiding nocturnal hypoglycemia afterevening exercise. Considering that once-daily insulin glargine administration is as-sociated with a glucose nadir 4–14 h later(38,39), not only basal insulin dose but alsothe timing of administration may be ofparticular importance.It is important to consider that the

patients in this study were treatedon a basal-bolus regimen; therefore,outcomes may differ in patients usingcontinuous subcutaneous insulin infu-sion therapy. However, the strategieswe used are likely to carry practicaland useful implications, so we suggestthat patients tailor these strategies ac-cording to their own treatment regimenand exercising habits. Indeed, thisshould not detract from the importanceof the findings because this study showsfor the first time that consuming LGIfoods and drinks in tandem with re-duced rapid-acting insulin dose fol-lowing evening exercise can play animportant role in normalizing glycemia,preventing postprandial hyperglycemiaand inflammation, and protecting pa-tients from postexercise hypoglycemiafor up to 8 h. The clinical utility of thesefindings is clear because foods that pa-tients habitually consume can be easilyexchanged with foods that offer thesame macronutrient content but are of

an LGI (e.g., substituting particular typesof breads, strains of rice, pastas, andpotatoes or sports drinks with variouscarbohydrate compositions) to facilitatemore desirable postprandial glycemicresponses. However, it does not seemthat carbohydrate type or total carbohy-drate intake alone are factors in the de-velopment of late-onset hypoglycemiabecause patients may still be exposedto nocturnal hypoglycemia followingevening exercise. Future research onbasal insulin adjustment will determinewhether late-onset nocturnal hypogly-cemia following evening exercise canbe avoided while harnessing the bene-fits of consuming LGI foods with a re-duced rapid-acting insulin dose duringthe postexercise period.

Acknowledgments. The authors thank thestudy participants for their time, effort, andcommitment and the research team at theNational Institute for Health Research ClinicalResearch Facility, Newcastle University, forassistance.Duality of Interest. This study was partiallyfunded by BENEO. BENEO reserves the exclusiveright to use the results and data for regulatoryproceedings. Instant drink powders containingmaltodextrin or isomaltulose (Palatinose) werespecially developed for this study and producedby BENEO and Sudzucker AG (Central Depart-ment Research, Development, and Technologi-cal Services). CGM equipment was provided byMedtronic Europe. Neither party had any role inthe design of this study or the preparation of themanuscript. No other potential conflicts of in-terest relevant to this article were reported.Author Contributions. M.D.C. and D.J.W.contributed to the study concept and design,researched data, and wrote the manuscript.M.W. and J.A.S. contributed to participantrecruitment and reviewed and edited the man-uscript. M.I.T. and E.J.S. contributed to the datacollection and reviewed and edited the manu-script. D.T. contributed to the CGM data anal-ysis and reviewed and edited the manuscript.R.M.B. contributed to the CGM data analysisand preparation of the manuscript. D.J.W. is theguarantor of this work and, as such, had fullaccess to all the data in the study and takesresponsibility for the integrity of the data andthe accuracy of the data analysis.

References1. Chu L, Hamilton J, Riddell MC. Clinical man-agement of the physically active patient withtype 1 diabetes. Phys Sportsmed 2011;39:64–772. Taplin CE, Cobry E, Messer L, McFann K,Chase HP, Fiallo-Scharer R. Preventing post-exercise nocturnal hypoglycemia in children withtype 1 diabetes. J Pediatr 2010;157:784–788, e13. Brazeau AS, Rabasa-Lhoret R, Strychar I,Mircescu H. Barriers to physical activity amongpatients with type 1 diabetes. Diabetes Care2008;31:2108–2109

4. Davey RJ, Bussau VA, Paramalingam N, et al.A 10-s sprint performed after moderate-intensityexercise neither increases nor decreases the glu-cose requirement to prevent late-onset hypogly-cemia in individuals with type 1 diabetes.Diabetes Care 2013;36:4163–41655. CampbellMD,WalkerM, TrenellMI, et al. Largepre- and postexercise rapid-acting insulin reduc-tions preserve glycemia and prevent early- butnot late-onset hypoglycemia in patients with type1 diabetes. Diabetes Care 2013;36:2217–22246. West DJ, Morton RD, Bain SC, Stephens JW,Bracken RM. Blood glucose responses to reduc-tions in pre-exercise rapid-acting insulin for 24 hafter running in individuals with type 1 diabetes.J Sports Sci 2010;28:781–7887. West DJ, Stephens JW, Bain SC, et al. A com-bined insulin reduction and carbohydratefeeding strategy 30 min before running bestpreserves blood glucose concentration after ex-ercise through improved fuel oxidation in type 1diabetes mellitus. J Sports Sci 2011;29:279–2898. Yardley JE, Kenny GP, Perkins BA, et al. Re-sistance versus aerobic exercise: acute effectson glycemia in type 1 diabetes. Diabetes Care2013;36:537–5429. McMahon SK, Ferreira LD, Ratnam N, et al.Glucose requirements to maintain euglycemiaafter moderate-intensity afternoon exercise inadolescents with type 1 diabetes are increasedin a biphasic manner. J Clin Endocrinol Metab2007;92:963–96810. Evert AB, Boucher JL, Cypress M, et al. Nu-trition therapy recommendations for the man-agement of adults with diabetes. Diabetes Care2014;37(Suppl. 1):S120–S14311. Riddell MC, Perkins BA. Type 1 diabetes andvigorous exercise: applications of exercise phys-iology to patient management. Can J Diabetes2006;30:63–7112. Perry E, Gallen I. Guidelines on the currentbest practice for the management of type 1 di-abetes, sport and exercise. Pract Diabetes Int2009;26:116–12313. Parillo M, Riccardi G. Dietary carbohydratesand glucose metabolism in diabetic patients.Diabete Metab 1995;21:391–40114. Jenkins DJ, Wolever TM, Taylor RH, et al.Glycemic index of foods: a physiological basisfor carbohydrate exchange. Am J Clin Nutr1981;34:362–36615. Nansel TR, Gellar L,McGill A. Effect of varyingglycemic index meals on blood glucose controlassessed with continuous glucose monitoring inyouth with type 1 diabetes on basal-bolus insulinregimens. Diabetes Care 2008;31:695–69716. Parillo M, Annuzzi G, Rivellese AA, et al. Ef-fects of meals with different glycaemic index onpostprandial blood glucose response in patientswith type 1 diabetes treated with continuoussubcutaneous insulin infusion. Diabet Med2011;28:227–22917. West DJ, Morton RD, Stephens JW, et al.Isomaltulose improves post-exercise glycemiaby reducing CHO oxidation in T1DM. Med SciSports Exerc 2011;43:204–21018. Jensen TE, Richter EA. Regulation of glucoseand glycogen metabolism during and after ex-ercise. J Physiol 2012;590:1069–107619. Jentjens R, Jeukendrup AE. Determinants ofpost-exercise glycogen synthesis during short-term recovery. Sports Med 2003;33:117–144


20. MacDonald MJ. Postexercise late-onsethypoglycemia in insulin-dependent diabeticpatients. Diabetes Care 1987;10:584–58821. Wallace TM, Matthews DR. Recent ad-vances in the monitoring and management ofdiabetic ketoacidosis. QJM 2004;97:773–78022. Laffel L. Ketone bodies: a review of physiol-ogy, pathophysiology and application of moni-toring to diabetes. Diabetes Metab Res Rev1999;15:412–42623. Rosa JS, Oliver SR, Flores RL, et al. Alteredinflammatory, oxidative, and metabolic re-sponses to exercise in pediatric obesity andtype 1 diabetes. Pediatr Diabetes 2011;12:464–47224. Keenan DB, Mastrototaro JJ, Zisser H, et al.Accuracy of the Enlite 6-day glucose sensor withguardian and Veo calibration algorithms. Diabe-tes Technol Ther 2012;14:225–23125. Dill DB, Costill DL. Calculation of percentagechanges in volumes of blood, plasma, and redcells in dehydration. J Appl Physiol 1974;37:247–24826. Rabasa-Lhoret R, Bourque J, Ducros F,Chiasson JL. Guidelines for premeal insulindose reduction for postprandial exercise of dif-ferent intensities and durations in type 1 diabeticsubjects treated intensively with a basal-bolus

insulin regimen (ultralente-lispro). Diabetes Care2001;24:625–63027. Clinical Exercise Testing. In ACSM’s Guidelinesfor Exercise Testing and Prescription. Walter R,Gordon NF, Pescatello LS, Eds. Baltimore, Ameri-can College of Sports Medicine, 2010, p. 114–14128. Brouns F, Bjorck I, Frayn KN, et al. Glycae-mic index methodology. Nutr Res Rev 2005;18:145–17129. Hernandez JM, Moccia T, Fluckey JD,Ulbrecht JS, Farrell PA. Fluid snacks to help per-sons with type 1 diabetes avoid late onset post-exercise hypoglycemia. Med Sci Sports Exerc2000;32:904–91030. Jeukendrup AE, Wallis GA. Measurement ofsubstrate oxidation during exercise by means ofgas exchange measurements. Int J Sports Med2005;26(Suppl. 1):S28–S3731. Wolever TM, Jenkins DJ. The use of the gly-cemic index in predicting the blood glucose re-sponse to mixed meals. Am J Clin Nutr 1986;43:167–17232. Thomas D, Elliott EJ. Low glycaemic index,or low glycaemic load, diets for diabetes melli-tus. Cochrane Database Syst Rev 2009;(1):CD00629633. McGarry JD, Foster DW. Ketogenesis. In:Ellenberg and Rifkin’s Diabetes Mellitus, Porte

D, Sherwin RS, Editors. 1996, New York,McGraw-Hill, 1996, p. 19–2834. Balasse EO, Fery F. Ketone body productionand disposal: effects of fasting, diabetes, andexercise. Diabetes Metab Rev 1989;5:247–27035. Cryer PE, Childs BP. Negotiating the barrierof hypoglycemia in diabetes. Diabetes Spectr2002;15:20–2736. Mikines KJ, Sonne B, Farrell PA, Tronier B,Galbo H. Effect of physical exercise on sensitiv-ity and responsiveness to insulin in humans. AmJ Physiol 1988;254:E248–E25937. Maarbjerg SJ, Sylow L, Richter EA. Currentunderstanding of increased insulin sensitivityafter exercise - emerging candidates. Acta Phys-iol (Oxf) 2011;202:323–33538. Ashwell SG, Gebbie J, Home PD. Optimaltiming of injection of once-daily insulin glarginein people with Type 1 diabetes using insulinlispro at meal-times. Diabet Med 2006;23:46–5239. Thomas RM, Aldibbiat A, GriffinW, CoxMA,Leech NJ, Shaw JA. A randomized pilot study inType 1 diabetes complicated by severe hypogly-caemia, comparing rigorous hypoglycaemiaavoidance with insulin analogue therapy, CSIIor education alone. Diabet Med 2007;24:778–783


A Low-Glycemic Index Meal and Bedtime Snack Prevents Postprandial Hyperglycemia and Associated Rises...

Documents

Transcript of A Low-Glycemic Index Meal and Bedtime Snack Prevents Postprandial Hyperglycemia and Associated Rises...