-
EFFECT OF CARBOHYDRATE SOURCE ON POSTPRANDIAL
BLOOD GLUCOSE IN SUBJECTS WlTH TYPE 1 DIABETES USlNG
INSULIN LISPRO
Nadia H.J. Mohammed
A thesis submitteâ in confomity with the mquimments for the
ûegme of Master of Science,
Griduate Department of NutrStionaI Sciences Univemity of
Toronto
@ Copyright by Nadie ~ o h a ~ h e d 2001
-
National Cibrary û i b i i i u e nationale du Canada
Acquisitions and Acquisitions et ôibiiüg i-aphic Services
services bibiiographiques 395 Weliingtolr Street 395. rue Wdington
OnawaON KlAOW O(tawaON K 1 A W Canada CaMda
The author has granted a non- L'auteur a accordé une licence non
exclusive licence allowing the exclusive permettant à la National
Libraxy of Canada to Bibliothèque nationale du Canada de reproduce,
10- distrtiute or seil reproduire, prêter, distribuer ou copies of
this thesis in microfom, vendre des copies de cette thèse sous
paper or electronic formats. la forme de microfiche/nlm, de
reproduction sur papier ou sur fomat électronique.
The author retains ownersbïp of the L'auteur conserve la
propriété du copyright in this thesis. Neither the droit d'auteur
qui protège cette thèse. thesis nor substantial exûacts fiom it Ni
la thèse ni des extraits substantiels may be printed or otherwise
de celle-ci ne doivent être imprimés reproduced without the authw's
ou autrement reproduits sans son permission. autorisation.
-
To my parents
-
TABLEi OF CONTENTS page
ACKNOWLEDGEMENT
..........i....................................... iu
.......................................................... LIST
OF TABLES
LIST OF FIGURES
.........................................................
LIST OF APPENDICES
...................................................
ABSTRACT
..................................................................
vïii
CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW 1.1 Introduction
.....................................................
................................................. 1.2 Literature
Review 1.2.1 De~tionofDiabetesMe~tus .................... ... 1.2.2
Prevalence ................................................ 1.2.3
Classification ......................................*...... 1.2.4
Diagnosis ................................................. 1.2.5
Type 1 Diabetes Mellitus ................................
............................... 1.2.6 Type2DiabetesMelliais
1.2.7 Complications
............................................
.................. 1 -2.8 Monitoring Blood Glucose Control
..................... 1 -2.9 Intensive Blood Glucose Control
1.2.10 Treatment
................................................. ........ . 1.2.1
O 1 Non-f harmacological Therapy
.................... 1.2.10.1.1 Exercise 1 .2.10. 1.2
Nutritional Management ..
.......................... 1.2.10.2 Insulin'ïherapy 1.2.10.3
Lispro Insulin ......W.........-...........
.......... 1 -2.10.4 Lispro and Meal composition
................................ 1.2.11 TheGlycemicIndex
............................. 1 -3 Study Objectives and
Hypothesis
.................................... 1.3.1 Study objectives
.................................... 1.3.2 Study Hpthesis
CHAPTER 2
-
MATERIALS AND METHODS Subjects
............................................................
............................................. 2.1..1 Screening
..................................... Study Design and Protoc01
............................................. 2.2.1 Rationale
............................................... 2.2.2 Protocol
............................ 2.2.2.1 Hypogiycemia Test Meal
Preparation ...........................................
............................... 2.3.1 Instant Mashed Potato
.......................................... 2.3.2 White Bread
2.3.3 Spaghetti ..............................................
....................................... 2.3.4 Pearled Barley
............. ................... 2.3.5 . Pineapple Juice ...
........................................ Blood Glucose Analysis
............................................... S t a î i s t i
c a l ~ s i s
CHAPTER 3 RESULTS 3.1 Subjects
........................................................... 36 3 -2
Glycemic Response Data ...............................-.....Le 38
3.3 Hypogiycemic Outcomes .......................................
46
....................................... 3.3.1 Hypoglycemia 50
................................. 3.3.2 Low B l d Glucose 50
3.3.3 GI and Pg in Relation to Hypoglycemic
............................................. Outcomes
CHAPTER 4. DISCUSSION AND CONCLUSIONS 4.1 Discussion
......................................................... 58 4.2
Potentid Practical Implications ................................ 64
4.3 Future Research
........-........................................... 66 4.4
Conchsions ......................................................
66
REFERENCES ClTED 68
~Pl?ENDKEs 76
-
In the name of Allah, the most mercz~l, the most graciour
Fkst and forernost, I would like to thmk my hwband and e v q
member in my family. Without their unconditio~u~l love, patience
and
support, I would not have been able tu accornplsh this work
I'm gratefùl d thanml tu my supervisor Dr. Wolèver, for his
continuous encouragement and i m a l ~ ~ ~ b l e guihnce
throughout the
program- 1 would also like to thank Dr. Jenkns and Dr. Rao for
their valued
input. Bank you fo Dr. Singer for appraising the thesis and Dr.
El-Sohemy
for chairing my defense.
1 am irrdebted to al2 my subjects who volunteered to participate
in my
stuc&- ïlbanks to everybody in my Zab group, d special t h k
to Janet Vogt
for her heIpfu2 advice. Inmks to CZaudio in the Glycemic Index
Testing
Offce for hm technical support. Also greati'y appreciated is
thefiiendsh* of
everyone ut the Clinical Nuhition and Risk Factor Modzifiation
Centre at
St. Miehael's Hospital. A special thmrAyou tu Zènith Xu for her
kind heb.
1 would like to extend my thanh to the Faculty of Medicine in
Kuwait
University, for ofering thk scholmship.
mis thesis was made possible by a grantfiom Eli Lil&
company.
-
Table 2.1
Table 3.1
Table 3.2
Table 3.3
Table 3.4
Table 3.5
Table 3.6
Table 3.7
LIST OF TABLES
GI, Weigbt and Composition of Test Foods
M- Demographic and Biochemical Profile of the Sample at
srreening
Two-Way AnalysiP of Variana ofLncremental Bkxxi Glucose
Responses Mer Difllerent Types ofFood
Mean Bbod Ghicose Increments for Food Types vs. Time
Two-Way A d y s k of Variance of Area Under Giy- Curve (AUC) Mer
Werent Types of Food
Mean Area Under the Curve (AU0 of the Glycedc Response Curve for
Food Types
Number of Hypogiycemk and Low Blood Glucose Episodes per
subject
-
Figure 2-1
Figure 3-1
Figure 3.2
Figure 3.3
Figure 3-4
Figure 3.5
Figure 3 -6
Figure 3.7
LIST OF FIGURES
Tes te Day Protocol
Correlation between Ghicometer a d YS1 Blood Glucose
co~~centrations
Mean B W Ghicose Increments (by YSI) For Food Types vs. Time
Mean Blood Ghicose Concentrations (by YSI) For Dïfkent Test
Foods
Correhtion Between Mean AUC of Test Foods and Their Glycemic
Iodices
No. of Hypoglycemic and Low Bbod Ghrose Episodes per Food
Types
Correlation of No. of Hypoglycemic Episodes (by Glucometer a d Y
SI) with GI a d Pg
Correlation of No. of Low Bbod Glucose Episodes @y Glucometer
and YSI) with GI and Pg
-
LIST OF APPENDICES
AppendixA Consent Fomi
Appendix C Weight and Nuîrient Content of Standard DBmr Meal
Appendix D Data Fonn
Appendk E Demograpbic and Biocbemical Pro& of the Subjects
at s-
Appendk F TllSUlBl Types and Mean Doses Used By Subjects
-
.................. AD A ANOVA ..........-.. M T
.................. AUC ................. BMI ..................
CS11 ................. DCCT ....O..........- DM ...................
FBG .................
IFG ................... .................... IGT
L ....................... LDL
SBGM ............O.-.. SEM ....................
...................... U YS1 ...................
The Americaa Diabetes Association Analysis of Variance
AhnmeTransarninase Area Undet Giycemic Response Curve BodyMassIndex
Coiitinuous Subcutaneous IrisulEnmn Diabetes Control and
Complication Trial Diabetes Mellitus Fasting Blood G b s e Granis
GestatiodDiabetes Giycemic Index Hours Hemogiobh Alc
ImpairedFastmgGhicose Ingiiired Ghicose Tokrance -gram Litre
hw-Densii Lipopmtem Mü1snole per Litre Propoen of carbohydnie
absorbed as ghmse Correlationcoefficient SelfBbod Glucose
Monitoring Standard Error of the Mean units Yellow Spring
Instruments Adyzer
-
Effect of Carbohydrate Source on Postprandial Glycemia in Type 1
Diabetic Subjects Treated with
Lispro Insulin.
Master of Science, 2001 Nadia HJ. Mohammed
Graduate Department of Nutritional Sciences University of
Toronto
ABSTRACT
Treatment o f type 1 diabetes (TID) with iispro h s u h reduces
postprsndial
hyperglycemia, but because of its rapid onset of action,
postprandial hypoglycemia may
occur. To see ifgiycemic index (GI) and proportion of
carbohyârate absorbed as *se
(Pg) affecteci g iym-c responses and occurrence of hypogiycemia,
8 T1D subjects on
lispro were studied for 4 h on 5 &ys af€er o v ~ g h t
fàsts. Subjects took their usual
insulin dose and ate 50g carbohydrate tiom a starchy food (Pg=I;
instant potato G1=83,
white bread GI=71, spaghetti G P U , barley GI=ZS) or pineapple
juice (Pg=0.5; 01-46).
Glycemic responses differed significantiy for the diffèrent
foods and were close&
rehted to GI ( d . 9 8 , p
-
1. INTRODUCTION AND LITERATURE REViEW
1.1 INTRODUCTION
The therapeutic importance of tight blood glucose control is
well
established in both type 1 (1) and type 2 (2) diabetes. G d
blood glucose
control for people with type 1 diabetes depends on coordination
of insulin
doses, quantity and timing of food intake, and physical
activity.
The Arnerican Diabetes ASSOC-iation (ADA) recommends
considering
total carbohydrate consumed rather than type of carbohydrae,
thus \
traditionally ins& dose has been adjusted based on the
amount of
carbohydrate in the mealS. However, meals may differ not only in
the
amount of carbohydrate they contain, but also in the source. The
Glycemic
Index (GI) is a classification of foods based on their relative
blood glucose
raising potential(3), which in tum depends on the rate at which
carbohydrate
is digested and absorbed (4). Oiir lab.has demonstrated that the
GI of foods
is the same for people regardes oftheir glucose tderance (5). We
have also
shown that lowering the GI of the diet improves b l d glucose
control in
both type l(6) and type 2 (7) diabetes. More recently we have
demonstrated
that day-to-day variation in both the amount of carbohydrate and
the GI of
the diet of subjects with type 1 diabetes influences the
glycemic control(8).
1
-
This suggests that to achieve optimal diabetic controi,
adjwtment of insulin
dose should be based on both the amount and source of
carbohydrate in the
meal to be consumed.
Recently, an analog of human insulin, lispro, is increashgly
used
because of its more rapid and physiologie response. This results
in reduced
postprandial blood glucose compared to reguiar bulin (9) and the
potentid
for improved blood glucose control(10). However there is the
potentiai for
postprandial hypoglycemia especially when a low carbohydrate
meal is
consumed (11). This Mcates that the dose of lispro needed
depends on the
amount of carbohydrate present in the meal. However, the= are no
studies
yet on how to adjust lispro for dzerent types of carbohydrate
foods.
The general purpose of the present study is to determine the
pattein of
blood glucose response produced by an equivalent amount of
different
carbohydrate f d s in type 1 diabetic subjects ushg insulin
lispro, and to
determine the influence of carbohydrate source on
lispro-induced
postprandial hypoglycemia The overall aim is to be able to
advise diabetic
people whether to adjust lispro dose for different types of
carbohydrate for
optimal glycernic control
-
1.2 LITRATURE REVIEW
1.2.1 Definition of Diabetes Mefitus
Diabetes Mellitus is defhed as a heterogeneous group of
metabolic
diseases characterized by the presence of hyperglycemia due to
defective
iasulin secretion, insulin action, or both.
1.2.2 Prevalence
Currently, diabetes has been diagnosed in 5% of Canadians or
1.5
million people (12). This number was expected to reach 2.2
million by the
year 2000 and 3 million by 2010 (13). Moreover, because U.S
statistics
demonstrated that for every person with diabetes there is
someone with
undiagnosed diabetes, these numbers most likely underestimate
the
prevalence of the disease. Assumùig that the same situation is
tme in
Canada, up to 10% of Canadian adults may currently have
diabetes.
1 -2.3 classincation
Diabetes is recently reclassified into five distinct types based
on the
pathogenesis rather than treatment (14). The vast majority of
cases f d into
two broad categories, which are type 1 a d type 2 DM (discussed
later).
-
Gestational diabetes (GDM) is any degree of glucose intolerance
that is
recognized during pregnancy.
A variety of relatively uncornmon conditions are listed under
"other
specifïc îypes". These consist mainly of specifïc genetically
defïned fonns of
diabetes (e.g. abnormalities of insulin or its teceptor) or
diabetes associated
with other diseases or drug use.
Impaired glucose tolerance (IGT) and mipaired fasting glucose
(IFG)
are metabolic stages interniediate between
established diabetes. They are mt clinical
are risk factors for Miire development
disease.
nomal glucose homeostasis and
entities in their own rather they
of diabetes and cardiovascular
1.2.4 Diagnosis
The Diagnosis (14) of diabetes is established by a fasting
plasma
glucose (FPG) value equal to or greater than 7.0 mmol L, a
random plasma
glucose greater than 11.0 m V L with symptom, or plasma glucose
value
in the 2-h sample of the oral glucose test (75g - glucose load)
greater than 11 .O mmoVL.
Diagnosis of IGT depends on a plasma glucose value in the 2-h
sample
of the oral glucose test (75g - glucose load) between 7.8-1
1.0.
-
Impaired fasting glucose (WG) has been established to iden*
another
intermediate stage of abnomal glucose homeostasis analogous to
IGT. IFG
is diagnosed by a fasthg plasma glucose value between 6.169.
1.2.5 Type 1 Diabetes MeIlitus
Type 1 DM commonly occurs in childhood and adolescence, thus
is
also known as Iwenile-onset diabetes. It represents about 5-10%
of all cases
of diabetes. It results fiom absolute deficiency of insulin
secretion due to
destruction of the B-cells of the pancreas, the etiology of
which is either an
autoimmune cell mediated or idiopathic.
1.2.6 Type 2 Diabetes Mefitus
Type 2 DM is the cornmonest type of diabetes. Approximately 3%
to
5% of the general adult population has unrecognized type 2
diabetes. It may
range ftom predominant insulin resistance with relative insulin
deficiency to
a predominant secretory defect with insuliÙ resistaiace. Studies
have
identified increased risk associated with older age, central
obesity, certain
ethnic back grounâ, physical inactivity, hïstory of GDM,
overt
cardiovascular disease, high fasting insulin levets, and IGT (1
5,16,17).
-
1.2.7 Complications
Diabetes is a serious M.th problem associated with acute and
chronic
complications. It often disables people in their middle years
and as a group,
people with diabetes die younger than those not affited by it
(18).
Major acute complications are hpglycemia and hyperglycemic
crises
including ketoacidosis and hyperglycemic-hypemsmolar states.
The chronic hyperglycemia of diabetes is associated with long
term
macro and microvascualr disease causiag damage, dysfunction,
and
ultimately failure of various vital or- especially the kidney,
eye, nerves,
heart and biood vessels,
Diabetes (both type 1 & type 2) is a major cause of coro-
artery
disease, which is the leading cause of death in Canada.
Morbidity and
mortaliîy rates are 2-to4fold higher than in age-and sex-matched
groups in
the non-diabetic population (19,20,21,22,23) it is &O a
leading cause ofnew
cases of blindness and kidney disease in adults.
1.2.8 Monitoring Blood Glucose Control
Daily Self Blood Glucose Monitoring (SBGM) has markedly
improved
the ability to acutely control blood glucose levels. It permits
recognition of
low levels of blood glucose before hypoglycernia occurs (24,25)
and allows
-
people to assess the effècts of diet, exercise and changes in
treatment
regimens.
People with type 1 diabetes often use pre-meal and bed time
tests, as
well as intermittent ps t meal testing to adjust insulin doses.
Testing for
glycated hemoglobin, as it refleck average blood glucose levels
of
approximately the previous three months, should be pe150-d
penodically
to assess long-term glucose control (26). A specific type of
glycated
hemoglobin, HbAlc, is a tool for assessrnent of the preceding 6
8 weeks
(27)-
1.2.9 Intensive Blood Glucose Control
Epidemiological studies and studies done in animal models of
diabetes
implicate hyperglycemia in the pathogenesis of long-terni
complications of
diabetes. Consequently the aim of diabetes management is to
maintain the
glycernic status as close to the no& range as safey
possible.
Intensive insulin therapy was applied in the Diabetes Control
and
Complication Triai @CCT) by multiple daily insulin injections
administered
as basal insulin with short-acting bolus insulin (basal-bolus)
adjusted
according to SBGM, meal composition and level of physical
activity.
Continuous subcutaneous insuiin infusion (CSII) using a pump is
an
alternative to multiple daily injections. DCCT achieved inipmved
b l d
-
glucose control, which in turn effectvely delayed the omet and
slowed the
progression of diabetic complications (1).
1 -2. 1 O Treatment
The primaty goal of therapy is to avoid acute &
long-terrn
complications. In addition, the person's quality of life &
overd sense of
weU-being should be considered. Depending on the type of
diabetes and the
therapy required, this objective may be more or less dSicult to
achieve
without acute adverse effkcts. Thus treatment must be tailored
individilally
based on medical & social factors. If medication is required
to achieve
diabetic controI, it should be optimized with regular physical
activity and
healthy eating.
Since my thesis is about people with type 1 diabetes, I'Il d y
focus
on them in my discussion.
1.2.1 O. 1 ~ o n - ~ h ~ c 0 1 0 g i c a l Therapy
While insulin therapy is essentid for M e in people with type
1
diabetes, it should be optimized with regular physical activity
and
nutritionally adequate food intake.
-
An active He style pmmotes cardiovascular fitness and well
being
(28). However the net effect of exercise on glycemic control is
unpredictable *
and varies among individu& with type 1 diabetes because many
variables
innuence glucose supply and utilkation. These include the state
of nutrition
and metabolic control before and at the onset of the exercise,
duration and
intensity of exercise, fitness of the patient, type, dose and
site of insulin
inj ec tion (29).
In general, low to moderate intensity exercise lowers glucose
levels
both during and d e r the activity, inc~iea~ing the risk of
hypoglycemic
episode. Conversely, intense exercise raises glucose levels and
can lead to
progressive hypergIycemia and even ketosis. Accordingly,
physical activity
may require the adjustment of insulin and carbohydrate intake
both before
and after the activity to prevent exercise-induced hypoglycemia
In addition,
fiequent blood glucose monitoring should aid any adjetment.
Obviously,
intensive diabetes management plus SBGM provide flexibility
in
appropriately modi@ing insulin for exercise (343 1).
The advantages of increased activity levels must be balanced
against
the risks due to diabetic complications or other medical
conditions; thus,
plans for physical activity should be individuaiized.
-
1 -2.1 0. 1 -2 Nutritional Management
As fàr as nutrition is concerned, it should contribute through
dietary
recommendations to irnprove glycemic control and avoid
short-term
symptoms fkom hypoglycemia and hypergiycemia.
In Canada, dietary recommendations for people with diabetes are
the
same as those for the gewral population, k d follow the
principles of
Canada's guideliws for healthy eating (32). ControUed fat intake
for the
prevention of cardiovascuiar diseases is the underlyhg principle
of the
recommendations. For adults with normal lipid levels and
reasonable weight,
the guidelines reconimend daily fat intake not to exceed 30% of
total da@
energy requirements, where sahirated and poly-msaturated fats
each not
exceeding IO%, with the remainder coming h m mono-unsahirated
fat (33).
Protein intake, preferably fiom vegetable sources, should range
fiom 10-
20% of total energy with daily intake of about 0.86glKg, which
is similar to
that of the general population. The remainder 50-60% of the
individual's
energy requirements should corne fiom dietary carbohydrates,
especiaiiy
those . unrefined, slowly absorbed, and rich in soluble fibers
(33).
On the other band, the ADA nutrition recommendations stress
individualkation of diet based on the patient Westyle, and the
results of
clhical monitoring. It is recontmended that proteins, saturated
ht, a d poly-
-
unsaturated fat contribute a total of 30=4û% of daiiy ewrgy
intake, and the
remainllig 60-70% to come h m a combination of carbohydrates and
mono-
unsahuated fats (34). Daily fiber intake of at least 25-35gld is
reconmiended.
Soluble fiber intake has been associated with reduced blood
glucose
responses, and improved blood glucose control(35)
Several health agencies recornmend an h a s e of low-GI foods in
the
diets of individuals with diabetes (36,37,38). However, the ADA
has
questioned the clinical utility of the GI and recommends tbat
priority should
be given to the amount rather than the source of carbohydrate
(39). ,
On the contrary, Jenkins focused on the concept of 'spreading
the
nutrient load'. He suggested 4 factors that prolong the time of
nutrient
absorption fiom the gut, understanding of which is helpful in
the dietary
management of both diabetes and hyperlipidemïa. These are
increased food
fiequency (nibbling), viscous soluble fibers, low GI foods, and
enzyme
~ ï i t o r s of absorption (34). He added that m y higbfiber f
d s that lower
LDL cholesterol also have low glycemic indices, such as bkley,
beans, etc.
Therefore, exp1oration of low-GI f d s might be used to expand,
rather than
limit, the carbohydrate food choices of people with
diabetes.
-
1.2.10.2 Insulùi Therapy
Classically insuJin pfeparations can be classified according to
theii
time of omet and duration. In ascending order, these include
short acting
(Regular), intermediate acting W H , Lente) and long acting
(Ultralente).
Intermediate and long acting insuiin injecti011~ provide
appropriate basal
circulating insulin concentrations overnight and between meals.
Pre-meal
bolus injections of short acting are given to prevent an
excessive rise in
postprandial blood glucose immediately after meals (40).
Combinations of these insulins can be given in a variety of
protocols.
In the split-mixed protocol, mixture of short and longer acting
msulni is
administered twice a day, before breakEast and dirmer.
Altematively, in the
basal-bolus protocol, also known as multiple daily injections,
short-acting
insulin is given before each m e 4 while N'PH or Ultralente
injections
provide basai requirements (4 1).
The available short acting insulin preparations have vatious
shortcomhgs. The most evident is a delayed omet of action
and
uiappropriately long duration of action. Studies indicate that
the peak effect
of Regular insulin occurs h m 2 to 6 heurs d e r injection, and
its effect may
last as long as 16 hours (42). In addition, it was f o d diat
Reguiar insulin
-
should be given fiom 30 to 60 minides before d s to achieve
optimal
control of postpraadial glucose (43).
This is fàr tlom normal physiology where endogenous,
pancreatic
insului secretion is stimulated irnmediately after blood glucose
concentration
begins to rise following food ingestion, so that insulin
concentration peaks
fat around 60 minutes postprandial, then declines over 3 hours
to reach a
basefine level.
This departine in the pharmacokinetics of Regular insulin ftom
normai
pancreatic insului springs fiom the fact that injecteci human
insulin must be
in a mommeric fom before it c m be absorbed through the
capillary
membrane into circulation. However, the himian insulin molecule
has a high
tendency for seKassociation (44). When Regular insuiin solution
is stored in
vials or cartridges, insulin molecules are in a polymeric,
mostly hexameric
form to maintain stability. The dissociation rate into monomeric
molecules is
low at the injection site, hence the delay in peak insulin
concentration and
activity after subcutaneous injection (45). The end of the
insulin B chain
(specifically Proline at position B28) is the site responsible
for the self-
association of insului molecules.
Recently, bul in analogues have been developed which
overcome
many of these problems. They are very rapiidly absorbed so that
their onset
-
and peak action are closer to the injection time, which matches
more closely
the postprandial glucose excursion (46).
1.2.10.3 Lispro Insulin
[LYS (B28), Pro (B29)I-h~man msulin (Lyspro or Lispro (Humalog))
is
an insulin analog in which the natucal amino acid sequence of
the Bchain at
position 28 and 29 is inverted. These changes result in reduced
capacity for
self-association (47), and account for the monomeric behaviour
of Lispro in
solution (48), and for the faster phannacodynarnic action than
Regular
insulin when injected subcutaneously (49).
Lispro insulin beguis to work within 15 minutes after
subcutaneous
injection, peaks in about one hour, and has duration of 2 to 4
hours.
Compared to Reguiar insulin, Lispro insulin has a %ter
absorption and a
more rapid elimination, effectively produchg a shorter duration
of action,
which may offer an advantage over Reguiar insulin in the control
of blood
glucose after meals (49).
Studies addressing mealtime glycemic control using Lispro
iniifody
indicate that the postprandial rise in blood glucose is
signiscantly lower,
approximately 1 to 4 ~ l l t l ~ V L , than with human Regular
insulin (50,5 1) .The
best postprandial contml was obtained with an injection
immediately before
-
the meal. Even if Lispro is injected shortly after the meal, the
postprandial
glycemia is still well controlied (52,53).
The long-tem dycemic control as reflected by an improvement in
the
HbAlc level is better with insulin Lispro than with human
Regular insulin
(54), provided tbat an appropriate basal insuiin regimen is used
to take into
account its shorter diwtion of action (55).
In type 1 diabetic patients, the hypoglycemia rate was found to
be 12%
iower (particularly at night tirne) during treatment with Lispro
compared to
Regular insulin (56). In addition, rate of severe hypoglycemia
(defïned as
coma or requiruig glucagons or IV glucose), can be reduced by
30% with
insulin Lispro (57).
1.2.10.4 Lispro and Meal Composition
M. BU& (58) demonstrated the importance of meal composition
in
relation to lispro-induced Post-prandial hypoglycemia. He used
isocaloric
meals with different carbohydrate content. His expriment
revealed that pre-
prandial lispro has tendency for Post-prandial hypogiycemia in
the setting of
reduced carbohydrate iotake. This is consistent with other
studies showing
that amount of carbhydraate in individual f d s inauences
postprandial
-
blood glucose response (59). The author concluded that lispro
dose need to
be adjusted depending on meal composition to avoid
hypoglycemia
On the other han& M. Strachan (11) examined the effect of
meal
composition on postprandial glycemia to establish optimal time
for lispro
administration. He assumed that because meal composition affects
rate of
gastric emptying (60,61) which in tum affects timing of
post-prandial blood
glucose rise (62), meal composition has to be considered when
timing of
lispro administration is adjusted. So he concluded that for
meals with high
carbohydrate content, the optimal time for lispro administration
is pre-
prandial. However, for meals with hi& fàt content,
pst-prandial
administration is preferred to minsnize risk of pst-prandial
hypogiycemia.
A group of patients with type 1 diabetes on Lispro insulin
mêet
regularly at the Clinical Nutrition Center at St. Mchael's
Hospital, Toronto,
Ontario. They have reported that a h consuming some foods but
not others
(e.g. spaghetti, beans) thei. blood glucose &op initially,
followed by 'a late
rise (Wolever TMS, personai communication). Both beans and
spaghetti are
low-glycemic index f d s that are digested and absorbed slowly
resulting in
delayed pst-prandial blood glucose rise. So probably in case of
low-GI
carbohydrates, the blood glucose lowering action of iispro
precedes the
absorption of carbohydrate resulthg in ~vch desirable pattern of
blood
-
glucose response. Numerous studies have shown that pst-prandial
blood
glucose and insulin responses are intluenced by both the amount
of
carbohydrate consumed and its source (Brand et al. 1985, Jenkins
et al.
1985, Wolever et al. 1996 and others). SO, ifcarbohydrate f d s
differing in
terms of GI, nature, proportion of carbohydrate absorbed as
glucose (Pg),
etc. produce significantly different blood glucose responses,
then alteration
in either the dose andor timing of lispro administration may
allow people
with type 1 diabetes to achieve more acceptable pattern of blood
glucose and
reduce the risk of pst prandial hpglycemia.
1.2.1 1 The Glycemic Index
Systematic classification of f d s according to their
glycemic
responses was first undertaken by Otto and Nüdas (3) in an
attempt to
incorporate foods into diabetic diets in amounts inversely
proportional to
their glycemic responses. The glycemic Index (GI) was
developed
independently as ranking of foods based on theu postprandial
blood glucose
responses compared with a reference food. More precisely, it is
deîined as
the incremental area under the glucose ciirve of a 50g
carbohydrate portion
of a test food divided by the incremental area under the glucose
c w e of a
50g carbohydrate portion of a standard reference food (white
bread) and
17
-
multiplied by 100 (3). Several methods have been used to
calculate the area
under the glycemic-response cunre (63) and resuit in markedly
different
areas and GI values (3). Woiever and Jenkllis use the
incremental area for
calculations, where the area below the b a s e k glucose value
is not inchded
in the calculation (64). Area d e r the cuve (AUC) is calculated
by using
what has corne to be known as the 'the trapezoidal de ' . This
calculation for
AUC is well described in the literature (3). In 1981 Jenkms et
al (65)
published the fht lid of GI vahies for 62 f d s .
Studies have shown that both the amount and source of
carbohydrate
innuence the acute blood glucose response to meah in normal
(59), type
2(66) and type 1(67,68) diabetic subjects. Furthemore, Wolever
et al have
shown that in type I diabetes, coasistency in both the amount
and GI of
dietary carbohydrate is associated with ùnproved blood glucose
control
indicated by H b ~ l c (8). This supports adherence to a
consistent diabetic diet
plan. However, people with diabetes regard following a strict
diet a burden.
Liberalization of the diet is an alternatik that is suggested to
relieve tbis
bwden without adverse effects on glycemic control. To achieve
optimal
glycemic control with a Ii'beralized diet, education about
adjusting insulin for
both the amount and the GI of dietary carbohydrate is needed.
This is
consistent with a study showing that unlike in type 2 diabetes,
blood glucose
-
control in subjects with type 1 diabetes was not related to the
composition of
the diet in te- of diet GI and amount of Carbohydrate and fiber
intake (69).
This suggest that as long as insulin dose is adjusted to reflect
food intake, the
composition of the diet may not be important for optimizing
glycemic
control. This fact mostly applies to diabetics on intensive
insulin treatment.
Although it is generally agreed that different carbohydrate f d
s have
different glycemic effects, controversy exist about the
therapeutic
implications of this information for diabetes. High carbohydrate
diets are
recommended for individuais with diabetes and hyperlipidemia but
the type
of carbohydrate is likely to be important in determinùlg the
metabolic
response to such diets. Increasing carbohydrate intake with high
GI f d s
may increase blood glucose, insulin, and triglycerides
concentrati0~1~ (70).
However, increasing carbohydrate intake with low GI starchy f d
s may
allow carbohydrate intake to be increased witbut these unwanted
effects. In
long-term trials, low GI diets with no change in dietary fiber
content result
in modest but signincatlt improvements in overall blood glucose
control in
patients with type 1 diabetes as measured by glycated hemoglobin
(6).
Perhaps of greater therapeutic importance is the abiüty of low
GI diets to
reduce insului secretion and lower blood lipid concentration in
patients with
-
hypertriglyceridemia (3). h o , low GI starchy foods improved
glycemic
control and lowered sennn cholestero1 m type 1 diabetic children
(71).
1.3 STUDY OBJECTIVES AND HYPOTHESIS
1.3.1 Study Objectives
1.3.1.1 To detemine the pattern of b l d glucose response for
equivalent
a m t of carbohydrate foods with different Glycemic Index values
d e r a
standard dose of lispro in type 1 diabetic subjects.
1.3.1.2 To determine how equivalent amount of carbohydrate f d s
with
different Glycemic Index values idluence the pst-prandial
hypoglycemia
induced by Lispro.
1 -3 -2 Study Hypothesis
1.3.2.1 Significantly different blood glucose responses will be
elicited by
equivalent amounts of carbohydrate fiom different sources with
different
Glycemic Index values in subjects taking a standard dose of
lispro.
1.3.2.2 Postprandiai hypoglycemia will be elicited by lispro in
subjects
taking equivalent mmunts of carbohydrate h m different sources
more
fiequently with foods baving lower Glycemic Index a d lower
Proportion of
carbohybte absorbed as glucose.
-
2. MATERIALS AND METBODS
2.1 SrnJECTS
Volunteers were recniited by study invitation posters
distributecl in
Endocrinology clinics in 6 major hospitals in Greater Toronto
area and in
Clinical Nutrition Centre at St. Michael's hospital, and by
3-day newspaper
advertisement on three occasions.
Recruiûnent period lasted 10 months between Feb 2000-Nov
2000,
during which subjects were tespondhg sporadically and not in
groups.
A lecture was given to nurses and dietitians working in
Clinical
Nutrition Centre explainhg the purpose and procedures of the
study and
subject criteria to heQ in recruitment process.
A totd of 17 people were screened, 14 were eligible according to
our
study criteria, and 11 of these were recniited. They were 8
femaie and 3
male subjects. Three female subjects withdrew d e r 1-2 test
sessions for
personal reasons. So a total of 8 subjects each successfully
completed the
whole 5 tests. Each subject needed a period of 5-8 weeks to
complete the
study.
-
2.1.1 Screening
Subjects were interviewed pria to study commencement in the
screening session to assess suitability for participation.
EligibIe subjects
were male or non-pregnant females, 14-75 years of age, with type
1 diabetes
who intensively seKmanaged their diabetes using lispro insulin
for at least 3
months, the dose of which was adjusted based on the amount
of
carbohydrate in the meal. Usual fasthg blood glucose had to be
between 4
and 16.7 mmoVL, and HbAlc lower than or equal9.1%. Other
inclusion and
exclusion criteria are listed below:
Inclusion criteria:
14-75yrswithtype 1DM
On insulin lispro for at l e s t 3 months
Male or non-pregnant females
Moderate to good glycemic control WAlc 1 9.1%; normal 4.5-
5.8%).
Usual fasting blood glucose between 4 and 16.7 mmoYL
Willllig and able to coiriply with shdy protoc01 and give U i f
o d consent.
-
Exclusion criteria:
a History of severe hypogiycaemia (needs assisame of amther
person)
more than once withh the last 3 monthS.
a Use of dnigs other than insulin that affect carbohydnte
metabohm
(e.g. oral hypoglycemic agents, steroids). Use of stable doses
of beta-
blockers or thiazide diuretics for the treatment of hypertension
is
dowed.
Symptomatic gastroparesis or other gastrointeshl condition
affecthg digestion or absorption of nutrients.
a Use of drugs afEecting gastrointestinal motility or nutrient
digestion or
absorption.
a Surgery, infection or signincant cardiovascular event within
the last 3
months.
a Liver disease; severe rend failure or r d dialysis.
History of HIV infection or hepatitis.
Substance abuse.
Females already ptegnant or intendhg to become pregtiant driring
the
study, or sexuaily active with childbearbg potential not ushg
birth
control methods.
-
+ Lactating females less than 6 months after delïvery.
4 Participation in another medical, surjgical or
pharmaceutical
investigation.
Any other physical, mental or behavioral condition that may
make
participation of the subject dangerous to the subject's health
or that of
others, or affect the results obtained.
In the screening session subjects came fasting to Clinical
Nutrition
Centre at St. Mchael's Hospital. The study protocol was
described and
queries of the subjects were answered. Also they were given
instructions
about the standard dimer and snack for the night of the test.
Then subjects
who were willing to participate signed on a consent fonn
(Appendix A) and
were asked to give a fasthg bblood sample for biochemical
analysis of blood
glucose, iipids, HbAlc, AST, creatinine and ma. A snack was
given after
blood collection.
Mernards they answered a questionnaire about their medical
history
and home blood test results (Appendix B), their height and
weight were
measured and before they leave they were provided with the
standard d i r
for the f h t test.
-
2.2 STUDY DESIGN AND PROTOCOL
SLMichael's Hospital Research Ethics Board approved the
shidy
protocol. The study had a randombed, cross-over design.
Eligible subjects came to Glycemic Index Testing Office (55
Queen
St. East, 2d floor) on five occasions at weekly intends in the
moming
between 8:O and 8 3 0 am after 10 to 12 buis ovemight h t for a
4%hom
test (testing day protocol, Figure. 2.1). In each test day they
were studied
with one of the following 5 test foods: instant mashed potato,
white bread,
spaghetti, pearled barely and pineapple juice, each containing
50g glycemic
carbohydrate.
2.2.1 Rationale
The rationale for choosing the above test f d s was to study 4
starchy
carbohydrates with a 3-fold Merence in terms of their GIS to
elicit blood
glucose responses over a-wïde range. Pineapple juice was
included because
the carbohydrate it contaùis consist of sugars (sucrose,
glucose, and
hctose) that overall elicit a dwerent glycemic profile. These
foods also
m e r in the proportion of carbohydrate absorbed as glucose
(Pg). Starchy
foods have a Pg = 1 because they are totally absorbed as
giucose, while
-
& =-s-ri. FBG = Fasting Blood Glucose
Figure 2.1 Testing Day Protocol
-
pineapple juice bas a Pg = 0.5 because sucrose is absorbed as
half glucose
and half hctose (72).
2.2.2 Protocol
For dinner and snack (if n o d y coasumed) on the evening
before
each test day, they consumed a standard meal that was provided,
and nothing
else except for water. This meai consisted of a number of Choice
Bars
(Mead Johnson) and a- c m of Ensure F o d a (Ross, Abbott
Laboratones),
which is similar in ewrgy and carbohydrate content to their
normal dinner
(Appendix C: Weight and nutrient content of standard dinner). If
the subject
had to break his overnight fbt due to hypoglycemia, the test was
cancelled
and rescheduled for amther day.
Upon arriva1 on the moming of the test, the subject was weighed
and a
fasthg hger-prick capiiIary blood sample was obtained with
Autolet
lancets. If blood glucose was less than 4 or more than 16.7,
then the test was
cancelled for that day, and was rescheduled for another day. If
blood glucose
was in the above range, the test was carried out and the subject
took the dose
of insulin lispro that he or she would normally take before
eating 50g of
carbohydrates.-For each subject, the same dose of lispro was
repeated at the
same tirne (before the start of the test meal) on each of the
test days.
-
Within few (2-7) minutes d e r lispro administration, the
subject started
to eat a test meal containhg 50g available carbohydtate h m a
standard
portion of instant potato, white bread, spaghetti, pineapple
juice or pearled
bdey (GI, weight, and composition of test meals shown on Table
2.1). Test
meals were fed to subjects in random order.
Test meals were served with 250 ml water or tea (up to 50ml2% d
k
and 2 bags of non-nutrient sweetener are allowed if desired).
The drink
chosen remained the same for each subject for all subsequent
tests. Test
me& had to be consumed withh 10 min. Additional finger-prick
samples
were obtained haKhourly for 4 ho= after the start of the test
meal
(Appendix D: Data Form). Each capillary blood sample was split
into 2
aliquots; 3 drops of blood was saved in fluorocitrate tubes for
subsequent
glucose analysis using a YS1 glucose analyzer. Results of which
used for
study analysis. An additional drop was used for immediate
analysis by
glucorneter (Precision TM, Abbott Laboratories, Columbus, -OH).
The
glucorneter resdts were used to cietennine eligi'bility of
f&hg blood
glucose, for subject safety, and for analysis of hypogiycemic
outcornes.
During the test, subjects were expected to remain seated and
they were
not pennitted to smoke. M e r the test was over, a snack or
lunch meal was
provided if desired.
-
Table 2.1
GI, Weight and Composition of Test Foods
Weight
Moisture
Total Fat
Total carbohydrate
Available carbohYdrateB
Instant White Pineapple spaghettir P emled potato' 13readt ~uice
# ~ a r l e y ~
83 71 46 41 25
67.3 67** 373 72.3 79.6
6.1 9.1 1.5 11.1 8.4
4.5 NIA N/A 7.2 7.6
0.6 0.4 0.4 1.6 1.5
1.9 NIA NIA 0.5 1 .O
4.2 2.1 O 2.0 11.2
Weight and composition in grams.
Portion size and nutrient composition baçed on a d y s k (TMS
Wolever). * Portion size and nutrient composition based on food
tables. * GI values nom literature (Foster-Powell K et al 1995) **
Weight of fiou.. * Available carbohydrate = total carbohydrate (by
difference) -total dietsry nkr.
-
To minnnize withh- subject variation, data were only included
for
tests when the subject's fastmg plasma glucose on the study days
varied by
no more than 6 mmoVL. Tests outside this range were
repeated.
2.2 -2.1 Hypoglycemia
According to our protocol, hpglycemia was recorded if blood
glucose dropped below 3.0 mmoi/L at the time of blood s a m p b
based on
the glucometer or the YS1 analyzer retrospectively, or if at any
time the
subject experienced imtohbIe symptoms of hypoglycexnia. At this
point the
test was terminated and depending on the severity of hmycaemia,
subject
was treated with oral dextrose, *ose Tablets or simp1y H cup of
juice and
something to eat if required. Then blood glucose was monitored
until it was
recovered-
Time to hypoglycemia was recorded exactly as minutes at
which
hypoglycemia (based on glucorneter-or symptoms) occurred nom the
start of
the test food or retrogradely as discovered by YS1 blood d y z e
r .
The occurrence of 'low'blood glucose was defined as a reaâing
less
than 4.0 mmoVL by glucometer or YS1 anaiyzer.
-
2.3 TEST MEAL PREPARATION
Portion sizes of white bread, potato, spaghetti, and barley
were
determined based on d y s i s of moisture, ash, protein, fat and
total dietary
fiber with available carbohydrate calculated by diffe~ence
(Wolever TMS).
Portion size of pineapple juice was determined based on food
Tables.
2.3.1 Instant Mashed Potato:
2 5 W boilùig water and '/z tbs salt was added to 67.3g of
potato,
stirred mtil unifody hydrated. It was pqared just before
servïng.
2.3 -2 White Bread
White bread was baked in 525g loaves containhg 250g
carbohydrates.
250g of warm water was mixed with 334g of d-purpose flour (Robin
hood),
7g sucrose, 6.5g Quick Rise Instant Yeast, and 4g of salt. The
ingredients
were placed into an automatic bread d e r (BlackBcDecker)
according to
instructions and mixed, kneaded and baked over a 2-hour p e n d
Loaves
were cooled at room temperature for 1 hr and weighed. C ~ s t
ends were
sliced off and discarded. The remainder of the loaf was cut into
105g
portions contalliing 50g carbohydrate, packed into plastic
freezer bags and
fiozen. Prior to consumption, bread was thawed in microwave
oven.
-
2.3.3 Spaghetti
72.30 of plain spaghetti (No Name Spaghetti) was boiled in a
covered
pot in about 175ml-salted water for 15 min mtil all water was
absorbed. It
was prepared the nigM before the test, kept in ikidge and micro
waved before
consumption.
2 -3 -4 Pearled Barley
79.6 g of pearled barley m. Goudas) wai boiled in a covered pot
in about 250ml-sdted water for about 2(hnin until aU water was
absorbed. It
was prepared the night before the test, kept io fiidge and micro
waved before
c onsumption.
2.3.5 Pineapple Juice
373 g of unsweetened pineapple juice (Dole), (approximately
1%
cups), was served after reiiigeration.
2.4 BLOOD GLUCOSE ANALYSIS
Blood giucose measurements were obtained ushg an automatic
glucose analyzer (Yellow Springs Instruments YS1 Myzer) , which
&es
the glucose oxidase technique (73). A standard glucose solution
(10mmoVL)
was used to calibrate the equipment. Ali the saniples were
andyzed ody
after the reading of the standard solution gave values in the
range of 9.8-10.2
-
mmoVL. This standard was run before each set of samples
belonging to a
single test meal of each subject.
When D-glucose in blood sample makes contact wÏth the
immobilized
enzyme glucose oxidase, it is rapidly oxidized pduciog hydrogen
peroxide
(H202). The H G is, in turn, oxidized at the platinum d e ,
producing
electrons. Thus, a dynamic equilibrium is achieved when the rate
of H202
production and the rate at which it leaves the immobiîized emyme
layer are
equivalent and is indicated by a steady state tesponse. The
electron flow is
linearIy proportional to the to the steady state of H202
concentration and
therefore, also to the concentration of glucose (74).
As with direct oxygen measurement, whole blood or hemolysed
blood
interferes with methods rneasuritlg H202 directly. This problem
is
circumvented with the YS1 analyzer by enclosing the enzyme in a
semi
permeable polycarbonate envelope. This allows measurement in
whole
blood, senmi or plasma
-
2.5 STATXSTICAL ANALYSIS
Lotus 123 97-dtion database software was used. Results were
expressed as means +/- SEM. Mean incremental areas mder the
glucose
curves, ignoring any areas below the fast& level were
calculated (3).
The GI for each test food was calculated for each subject using
the
incrementd area under glycemic response cinve for 50 g of
carbohydrate
fiom that test food and expressed as a percentage of the
response to 50 g of
carbohydrate f?om white bread. Calculated GI values were divided
by 1.4 to
be expressed by glucose scale (GI of glucose = 100). AUC of
white bread
for one of the subjects was zero, and therefore food GIS
pertaining to his
tests could not be calculated. As a result, these were omitted
fkom the
calculation of mean GIS of foods, and n was considered to be 7
rather than 8.
Mean blood glucose increments and mean areas mder the
glucose
response curve for food types were compared ushg two-way
analysis of
variance. The Newman-Keuls method was used to adjust for
multiple
~~mparisons.
Hypoglycemia and low blood glucose happened kquently based
on
both glucorneter and YS1 ivralyzer. Least squares linear
regression analysis
was perfonned to evaluate how hypoglycemia, low blood glucose,
and time
to hypoglycemia correlate with several independent fktors such
as
-
fkequency at which subject went hypogLycemic/low, kting blood
glucose
concentration, food GI, and food Pg. Multiple regression d y s i
s was a h
perfiormed to evaluate the correlation between hyjmglycemic
outcomes; time
to hpgiycemia; and low blood giucose, aed the abve fàctom
when
combined. Differences were considered statistically significant
when p
-
3. RESULTS
3.1 SUBJECTS
8 subjects, 3 males and 5 females, each completed 5 test
sessions.
Their age ranged between 14 and 74, with a mean of 36.75 years.
Mean
duration of diabetes was 13 years, with a range of 2-32 yem.
They were on
insulin Lispro for at least 3 months, with HbAlc raaging between
0.05 and
0.087. Main subject description is summarized in Table 3.1
(fully detarled in
appendix E).
All 8 subjects were using lispro and human insulin in a
multiple
injection (basal-bolus) regimen. Seven subjects were treated
with
intermediate acting insulin to provide basal needs. Two of these
were on
Lente Uisuiin and five were on NPH insulin (3 taking it twice
daily, 2 ody in
the evening). The remahhg 1 subject was on continuous
subcutaneous
insulin infusion (CSII by pump releasing lispro at a basal rate
of 1.2 Uh).
The mean of their total evening insulin dose was 15.9 f 4.5 U
(range 6 to 44
U). The mean of their total monhg dose was 8.6 f 2.7 U (range 6
to 44 U).
The mean of their total lispro bolus intake was 8.3 f 2.5 U
(range 2.6 to 24)
for the evening and 6.6 f 2.7 U (range 2 to 25) for the moniing.
Appendix F
describes the insulin treatment of the subjects.
-
Table 3.1
Main Demographic and Biochemical Profile of the Sample at
Screening
BMI
FPG (mrnoUL)
- Values (where applicable) are means f SEM * Indicates gJycated
hemoglobin (Normal range 0.035 - 0.065)
-
The subjects completed a total of 40 tests. An additional 5
tests were
perfonned, as repetitions to those where the subject's fhsting
blood glucose
varied by more than 6 mmoYL on the study days according to our
study
protocol. One test couldn't be repeated because the subject
could not attend.
3.2 GLYCEMIC RESPONSE DATA
Blood glucose concentrations were measured by glucometer as well
as
by YS1 anaIyzer. Linear correlation showed that readings
obtained fiom
these two methods are siificantly correlateci, where the slope
eqiials 1.05 f
0.01 and the correlation coefficient (r) = 0.97 at p
-
- S m
25- --- Line of Identity
0 /
0
I I
8 a 2 2 0 - 0 8 % g 8 g w E 15-
8 , 1 s > 10- m u &
5-
O " I I I I 1 O 5 10 15 20 25
Bbod Gkrocse Concentration by Glucometer (mroYL)
Figure 3.1
Correlation between Glucometer and YS1 Blood Glucose
Concentrations
-
+PdatQ +Wb Bread +- Pineapple Juice *Spaghe(G -0- Bariey
Figure 3.2 Mean Blood Glucose Increments @y YSI) for Food
Types vs. Time
-
Figure 3.3
Mean Blood Glucose Concentrations (by YSI) For Different Test
Foods
-
Pineapple juice shows a high early blood glucose increment peak
followed
by rapid decline.
Two-way ANOVA of blood glucose increments vs. tirne for food
types
(Table 3.2) shows a significant effect of food type (P4).05) h m
30 to 180
minutes, and a highly si@cant effect of subject (P
-
Table 3.2
Two-Way Analysis of Variance of Incremental Blood Glucose
Responses After Different Types of Food
30min
Source SS MS F P- Val Foods 234.9 4 58.7 11.2
-
C0nt' d: Two-Way A d p i s of Va- of focrementa1 B h d Ghmse
Respo- After Different Types of Food
Source SS Q F P-Val Foods 147-6 4 36.9 6.6
-
Table 3.3
Mean Blood Glucose Increments for Food Types vs. Time
Pearled Barky Spgkttr* Pineapple White Bread Illstant Potuto J i
e
- VaIues are means f SEM expressed as nmiol/L. - Cornparison of
al1 means: means sharing same letter superscript are not
sign%cantiy Fiifferent. Means with differenî superscripts are s
igdbdy Merent at f l .05.
-
The mean mas under glucose response curve (Am) for food
types
were examined. Two-way ANOVA of AUC (Table 3.4) shows that
the
variation in AUC for food types can be explained by the
signincant effects
of both food type @ < 0 .05) and subject (P
-
Table 3.4
Two-Way Analysis of Variance of Area Under Glycemic Curve (AUC)
Mer DifEerent Types of Food
Source
Foods
Subject
Error
Total
p p p p p
1214845.8 4 30371 1.45 5.5976174
-
Table 3.5
Mean Area Under the Glycemic Response Curve (AUC) for Food
Types
Food Tpes AUC
-
Pearled Barley
Spaghetti
Pineapple Juice
White Bread
Instant Potato
- Vahies are means * SEM expssed as mmolmia/l 0 - - Cornparison
of d means: means shering same letter superscript are not q @ b n t
f y
different. M~eans with différent mpema@ts are signiscantiy
different at @.OS.
-
Glycemc Index
Figure 3.4
Correlation Between Mean AUC of Test Foods and Their Glycemic
Indices
-
blood glucose G . 0 m V L ) , time to hypoglycemia, and low
blood glucose
(Le. < 4.0 mmoYL).
Based on YSI, all 8 subjects experienced hypoglyrcemia compared
to
ody 5 subjects based on the glucometer. Out of the 40 tests,
hypoglycemia
occurred more fkequently based on the YS1 (17/40, Le. 42.5%)
than the
glucometer (12/40, Le. 3W). This is consistent with our
observation that
YS1 readings were on average lower than glucometer readings for
the same
blood samples. The fhquency at which each subject had
hypoglycemia
ranged fiom 0.2 to 0.8 (i.e. 1 of 5 to 4 of 5 tests) with both
methods of
measurement. Hypoglycemia occurred with aU food types at times
ranghg
fkom as early as 87 minutes, up to the end of the 4-hour test
period (240
minutes).
3-3-2 LOW BLOOD GLUCOSE
AU subjects at some point had low blood glucose based on
glucometer
and YSI. Similar to hypoglycemia, episodes of low *ose occurred
more
fkequently based on YS1 (22140) than glucometer (20/40). The
fhquency at
-
which each subject had low blood glucose ranged nom 0.2 to 1
(le. once to
aIl5 tests) on both YS1 and glucorneter.
Table 3.6 summarizes number of hypoglycemic and low glucose
episodes per subject based on glucometer a d YSI. The number
of
hypoglycemic and low glucose episodes based on food type is
illustrated in
Figure. 3.5 as detected by YS1 and glucometer. Episodes of
hypoglycemia
and low glucose occmed most hquently with pkapple juice.
3.3.3 GI AND Pg IN RELATION TO HYPOGLYCEMIC OUTCOMES
The number of hypogiycemic and low glucose episodes was
plotted
agakt food GI and Pg in Figure 3.6 and Figure 3.7, respectively.
Resuits
fiom linear regression showed that food GL was not significantly
related to
number of hypogiycemic or low blood glucose crpisodes. However,
Pg was
inversely related to number of hypoglycemic (YS0 and low blood
glucose
(glucometer) episodes @ < 0.05). Nevertheless, the number of
points (types
of test meals) is small to draw conclusions.
Simple correlation was perfo~ned to obtain the relationship
between
the occurrence of hypoglycemia, low blood glucose and GI and Pg
of the
foods. There was no signincant correlation between food GI
and
hypoglycemia or between food GI and low blood glucose by YS1
or
-
Table 3.6
Number of Hypoglycemic and Low Blood Glucose Episodes per
subject
Numbers represent No. of episodes occurring in 5 tests.
Hypoglycemia
LowBlood Glucose
Subjects
Glucorneter YS1 Glucorneter YS1 -
2
3
1 2 O
1
4
4
3 O
2
4 O
3 1 2 1 4 1 4 1 3
3 1 3 4 4 1 5 1
5 6 1
1
7 4
5
1
1
% T o t a l 12 17 20 22
-
O Barley Pineapple Juice Potato Z Spaghetti White Bread
Test Foods
ü P - Low Blood Glucose
O Barley Pineapple Juiœ Potato Ir' Spaghetti White Bread
Test Foads
Figure 3.5 No. of Hypoglycemic and Low Blood Glucose
Episodes per Food Types
-
Figure 3.6
Correlation of No. of Hpglycemic Episodes (by Glucorneter and
YSI) with GI and Pg.
-
"Law" (meter) vs. GI
"Low" (meter) vs. Pg a 1
Figure 3.7
"Law" (Ys9 vs. GI 8 1
Correlation of No. of Low Blood Glucose Episodes (by Glucorneter
and YSI) with GI and Pg.
-
glucometer. However, Pg was significantly related to
YSL-based
hypoglycemia (r = 0.46, slope = - 1-13, P
-
Table 3.7
Multiple Regression Analysis
Dependent 1 Variables Fas ting r Subject BIood GI pg G k o s
e
- -
Occurrence of Hypogiycemia §*
Glucometer
Occurrence of Low Glucose $*
Glucometer
YS1
Time to Hypoglycemia *
YSI*"
* ~efined as blood glucose < 3 nmoYL or symptoms of
hypogiYCenga ~enned as blood ghrose < 4 inmVL.
* Number of observations = 40 ** Numbet of observations = 12 ***
Number of observations =16
3
Values are X coefkients
r
-
(slope) fobwed ôy p vahies. NS = not signiscant.
-
4. DISCUSSION AND CONCLUSIONS
4.1 DISCUSSION
The purpose of this study was primarily to compare the pattern
of
blood glucose response for carbohydrate f d s with Merent GI
values in
type 1 diabetics on Lispro insulin. This was examined in two
ways. The h t
method was plotting the measuced blood glucose increme~~ts vs.
time for
each food to obtain the correspondhg curves. Another method
was
cornparhg mean AUCs for Metent f d s .
The results obtained h m both methods indicate that the source
of
carbohydrate significantly Muences blood glucose responses in
type 1
diabetic subjects using insulin lispro. We believe that our
study design was
strong enough to detect real differences, which were not due to
error or
chance. These foods dEered in their glucose response because
their GI
values dfler over a wide range, i .e. their carbohydrates are
digested and
absorbed at different rates depending on thek nature (starch vs.
sugar) and
chernical structure (amylose: amylopectin ratio). Among starchy
test foods,
barley and spaghetti, which have relatively low GI values (25
and 41
respectively) produced significantly lower glucose increments
compared to
bread and potato which have relatively high GI values (71 and
83
-
respectively) between 60 and 120 minutes, as shown in Table
3.3.
Furthemore, the mean AUC for potato was significautly greater
than that
for spaghetti and barley (Table 3.5). The mean AUCs for test f d
s were
signifïcantly related to the respective GI values (Figure 3.4)
obtained h m
the lïterature (these GI values were determined in normal and
type 2 diabetic
subjects). This incücates that the GI is valid in this group of
subjects with
type 1 diabetes using lispro.
In the case of pineapple juice, GI done caowt explain its
blood
glucose response pattern. Pineapple juice contains sugars that
dser fiom
starchy f d s in 2 ways. Sugars are absorbed more rapiâly
compared to
unprocessed starchy carbohydrate. Therefore, although it has a
GI lower (GI
= 46) than 2 of the starchy f d s , the initial (30 minutes)
mean blood
glucose increment for pineapple juice was greater than al1 the
starchy foods
(Table 3.3). However, shortly after that, its incremental blood
glucose fêUs
rapidly and, by 90 minutes, it becomes no longer different fiom
that of
barley (Gr = 25). This is probably because sucrose, the main
coristituent of
pineapple juice is half glucose and half hctose. Fmctose, in
him, has a very
small effect in raising blood glucose, which lends pineapple
juice its Pg
value of 0.5, and exp1abs the rapid deche in blood glucose
incremental
curve.
-
Despite having different Pg, mean AUC for pineapple juice
retained its
median location among other test foods (Table 3.5) because AUC
is rehted
to GI rather than Pg.
Our finding regardhg the signifïcance of source of carbohydrate
on
glycemic response contradicts with current ADA Recommendations
(1994)
on dietary cahhydrates, which state: "From a cfinical
perspective, fbt
priority should be given to the total amount rather than the
source of
carbohydrate consumed". One reason for the disagreement maybe
that, for
this study, we selected carbohydrate sources with a wide range
of GI, which
we expected to produce a wide range of glucose responses.
However, in
practice, cornmon foods nom difEerent sources may not have
large
Herences in GI and, thus, would not be expected to affect
giucose
responses (Wolever et al. 1994). Thus, source of carbohydrate
will affect
postpmdial glucose responses only if the GI values of the
different f d s
differ by a sutnciently large amount. In addition, other factors
such as
number of subjects king tested and variation between and
within-subjects
may all affect the ability to detect difEerences in blood
glucose responses to
different f d s .
Another aspect, which may infiuence the validity of our
conclusions, is
the use of individual foods rather tban mixed meah in our
experiment. Some
-
studies showed that diBeremes in source of carbohyclrates make
no
merences to glycemic responses in the context of mixed me&
because of
the potential effects of fat and protein on glucose response.
However, faulty
methods were used in a numbet of the studies cited to support
this position
(Wolever and lenkùis. 1986). In addition, there are other
studies showkg
that GI retains its predictive ability in the mixed d setting
(Collier et
d.1986). Furthemore, alterhg diet GI was fomd to improve overall
blood
glucose control in diabetes (Wolever et al. 1992a).
Apart fiom GI and Pg, methodological factors may also
înfiuence
profoundly the interpretation of glycemic response data (3).
Factors that
might innuence blood glucose value obtained include: method of
blood
sampiing subject characteristics (e.g. body fatness, glucose
tolerance status,
insulin resistance); dose and timing of lispro and long-acting
insulin; degree
of diabetic control and particularly the fasting blood glucose
value on the
day of the test.
Om results indicate that subject factors had a significant
effect on
blood glucose increments throughout the test (Table 3.2) and on
the area
under the glycemic response curve (Table 3.4). This is because
subjects
m e r fiom each other. In addition, variability could arise fkom
within-
subject variation. That is, when a group of subjects test a food
once, the
-
glycemic response obtained in each subject may Vary considerably
(3) which
doesn't necessarily mean real dif fe~e~:es between the
subjects.
Our measurement of glucose response was based on capillary
whole
blood, because it is simple, non-invasive a d f d a r to our
patients. In
addition, it was suggested that the use of capillary compared to
venous blood
is more precise, as the glycemic responses are gceater in
capillary biood.
That is, greater absolute differences between f d s and greater
heterogeneity
between means d o w detedion of small differences in glycemic
responses to
diflierent foods, and thus more experimental power is obtained
(61).
Both glucorneter and YS1 d y z e r were used to masure
capillq
glucose concentration. The former measures it in plasma; the
latter in whole
blood. Glucose concentration is lower in red blood cells
compared to
plasma; therefore, our hding that glucose concentration
measurements
using glucorneter were greater was not unexpected.
The second objective of the study was to determine if the source
of
carbohydrate expressecl as GI and Pg c m predïct the occurrence
and timing
of hypoglycemia induced by iispro.
Results h m simple and multiple tegression d y s i s showed that
the
tendency of haviag hypoglycemia postprandially was inversely
mlated to Pg
(O* in YSI-based hypoglycemia). Similarly, the occurrence of low
glucose
-
was sigdicantly related to Pg (YS1 and giucorneter). GI did not
contribute
signincantly to the aforementioned correlations.
Om explanation for the finding that Pg predicts occurrence
of
postprandial hypoglyceniia and low glucose is that Pg detemiines
the
amount of carbohydrate absorbed as glucose, which in tum
directly
influences its avdability in blood. The results of blood
response studies
(Lee and Wolever) indicate that the less glucose consumed, the
quicker the
blood glucose drops to baselb. Accordingiy, the occurrence
of
hypoglycemia and low glucose is more likely expected with
carbohydrate
foods having lower Pg (Le. more sucrose) regardless of their GI
values.
GI did not predict occurrence of hypoglycemia and low glucose.
Low
GI carbohydrates produce relatively lower glucose increments and
smaller
AUCs, yet because they are absorbed slowly their glucose
increments are
sustained for a longer tïme than the high GI carbohydrates. This
may be why
we found no significant tendency for low GI to be related to
hypoglycemia.
On the other hand, other fkctors could be responsible for
missing a
signincant correlation between GI and hpglycemia such as the
high
variability of giucose responses in type 1 diabetes. More power
may be
needed to obtain a signifiant result.
-
When time to hpgiycemia was assessed, it was not found to be
signiscantly related to Pg. However, it was directly related to
GI and FBG
(YS1 only; in multiple regression only, F0.05). Therefore,
should
hypoglycemia happen, it wiU occm eariier with foods hahg lower
GI
values and in subjects baving lower basehe glucose.
4.2 POTENTUU, PRACTICAL IMPLICATIONS
The current fïnding regardiig the signincant effect of the
source of
carbohydrate (GI and Pg) on postprandai glycemia is relevant to
every day
life of type 1 diabetics, especially those who are intensively
treated with
lispro. One possible practical miplication of this shdy is that
education
about insulin adjustment for the GI and Pg of dietary
carbohydrate may be
needed to achieve optimal glycemic control. This is especially
important
with diabetics following libersrlized diet. We have shown that
high GI
starchy carbohydrates producecl on average greater glycemic
responses up to
3 hours *postpratldidy compared to those with low GI. So, it may
be
beneficial to advise diabetics to arlminrstei 0 a relatively
larger dose of lispro
for higher GI carbohydrates Furthemore, spaghetti, which is low
GI, slowly
digested carbohydrate showed unfavorable siycemic response after
lispro in
the sense that blood glucose dmpped early for a while before it
became
-
sustained for the rest of the c w e . So there is a potential
for early
postprandial hypoBlyda, although we could not demonstrate it.
So,
alteration in either the dose or timing of lispro by
admiiiistering a smaller
dose or admmiste~g the dose a& eatiog may aliow patients to
achieve a
more acceptable pattern of blood glucose.
One of the fkdhgs was thet f i t juices are more readily
absorbeci
caused an initial greater and fater rise in postprandial glucose
than starchy
carbohydrate. Furthemore, they are more Iürely to predispose
to
hypoglycemia than starchy carbohydrates. Thmefore, if juices are
to be
taken, then tighter contro1 of blood glucose may be achieved if
a larger dose
of lispro taken preprandially. If this is the case it may be
necessary to take a
snack to prevent late postprandial hpglycemia
-
4.3 FUTURE RESEARCH
One area for fiÙtrae research is to examine the effect of
carbohydtate
source on postprandial glycemia in the context of mixed rneals
rather than
individual carbohydrates (as m our study) to look at the
possible
confounding variable effects of fat and protein. This way, the
validity of our
conclusions in normal meal conditions can be d e t e e ,
A M e r step would be to examine the glycemic response to
dif5erent
GI meals with different doses of lispro. Variation in timing of
lispro
injection can also be tried. These procedures would be helpful
in
detemiinùig how best to adjust the dose of lispro for diffierent
carbohydrate
sources.
4.4 CONCLUSIONS
There were 2 objectives outlined in this thesis. The primary
objective
was to compare postprandial glycemic response profile for
equivalent
arnounts of carbohydrate foods k m different sources with
different
glycemic index (GI) values aAet a standard dose of lispro in
type 1 diabetic
subjects. The data supported my first hypothesis that
signiscantly different
blood glucose responses will be elicited by quivalent amounts
of
carbohydrate fkom dierent sources having different GI
vaiues.
-
M y second objective was to determine how equivalent amounts
of
carbohydrate foods with Merent GI values influence the
postprandial
hypoglycemia induced by lispro. W e hpthesized that
postprandial
hypoglycemia will be elicited by lispro in type 1 diabetic
subjects taking
equivalent arnounts of carbohydrate firom different sources more
fkequently
with foods having lower GI d lower proportion of carbohydrate
aôsorbed
as glucose (Pg). The resuhs supported the hypothesis for tbe Pg
but not for
the GI. However, GI innuencecl the timing of postprandial
hypoglycemia
-
REFERENCES
The Diabetic Control and Complication Trial research Group. The
effect of intensive treatment of diabetes on the development and
progression of long-term complications in - insulindependent
diabetes mellitus. New Eng j Med 1993;329:977-86.
UK Prospective Diabetes Study Group. Intensive blood glucose
control with SUlphonyIureas or insulin compared with conventional
treatment and risk of complications in patients with type 2
diabetes (UKPDS 33). Lancet 1998;352:837-53.
Wolever TMS, Jenkins DJA, Jenkins AL, Josse RG. The glycemic
Index: Methodology and clinical implications. Am J Clin Nutr 1991
;54:84654.
Jenkins DJA, Ghafari H, Wolever TMS, Taylor RH, Barker HM,
Fielden 6 Jenkins AL, Bowling AC. Relationship between the rate of
digestion of f d s and postpcandial glycemia. Diabetelogica
l982;22:450-55.
Wolever TMS, Jenkins DJA, Josse RG, Wong GS, Lee R. The glycemic
index: Similarity of values derived in insulindependent and non
insului-dependent diabetic patients. J Am Coll Nutr
1987;6:295-305.
Collier GR, Giudici S, Kalmusky J, Wolever TMS, Helman G, Wesson
V, Ebrlich RM, Jenlùns DJA. Low glycemic index starchy foods
improve glucose control and lower senmi cholesterol in diabetic
children. Diab Nuh Metab 1988; 1 : 1 1-1 9.
Wolever TMS, Jenkins DJA, Vuksan V, Jeakias AL, Wong GS, Josse
RG. Betieficial eff i t of low-gh/emic index diet in overweight
NIDDM subjects. Diabetes Care 1992; 1 5:S6266.
Wolever TMS, H . S, Chiasson JL, Josse RG, Leiter LA, Rodger NW,
Ross SA, Ryan EA. Day-to-day consistency in amount and source of
carbohyârate intake associated with improved b l d
-
glucose control in type 1 diabetes. J Am Coll Nutr
L999;18(?):242- 47.
Burge MR Castille KR, Schade DS. Meal composition is determinant
of lispro-induced hypoglycemia in IDDM. Diabetes Care l997;ZO:
152-55.
Ebeling P, Jaasson PA, Smioth U, Lalli C, Bolli GB, Koivisto VA.
Strategies toward improved contr01 during insulin lispro therapy in
IDDM: Importance of basai insulùl Diabetes Care lWT2O: 1287-89.
Strachan MWJ, Frier BM. Optimal time of administration of
insulin lispro: Importance of meal composition. Diabetes Care
1998;2126- 31.
Tan H, Maclean DR. Epidemiology of diabetes meIlitus in Canach C
h Invest Med 1995; l8:24O6.
Tan a Daneman D, Lan DCW, et al. Diabetes in Canada: Strategies
towards 2000. Toronto: Canadian Diabetes Advisory Board;
19979.3.
Report of the Expert Cornmittee on the Diagnosis and
Classification of Diabetes Mellitus. Diabetes Care l997;2O: 1-1
5.
Charles MA, Fontbonne A, Thibult N, et al. Risk factors for
NIDDM in white population. Paris Prospective Study. Diabetes 1991;
40:796-9.
Eastman RC, Cowie CC, Harris MI. Undiagnoseci diabetes or i m p
d -ose tolerance and cardiovasc& risk Diabetes Care l997;2O:
127-8.
Tuomilehto J, Knowler WC, Zimmet P. Rimary prevention of non-
insului-dependent diabetes meilitus. Diabetes Matab Rev
1992;8:339-53.
Diabetes in America 2d ed. Washington= National Diabetes Data b
u p , National Iastaute of Diabetes and Digestive and Kïdney
Diseases, National htiitute of He* 1995.
-
19. Fuller JEE, Elford J, Goldblatt P, et al. Diabetes
mortality: new lïght on an underestimated public health problem.
Diabeteloica l983;24:3364l.
20. Garcia W, McNamara PM, Gordon T, et al. Morbidity and
mortality in diabetics in the Fmmhgham population. Sixteen year
foilow-up study. Diabetes l974;23: 105-1 1.
21. Jarrett J. Mortality in diabetes. Q J Med l99O;75:4
13-4.
22. Stamler J, Vaccaro O, Neaton JD, et al. Diabetes, other risk
Iactors, and 12-year wdiovascular moaality for men screened in the
Multiple Risk Factor Intervention TM. Diabetes Care 1993;16:434-
44.
23. Krolewski AS, Kosiski El, Warram JH, et al. Magnitude and
determinanis of comnary artery disease in juvenileonset, iusuiim
dependent diabetes mellitus. Am J Cardio 1987;59:750-5.
24. Cox DI, Kovatcheu BP, Juüan DM, et al. Frequency of severe
hypogiycemia in insulin-riependent diabetes mellihis can be
predictednom self-monitoring blood glucose data J Clin Endocrino1
Metab 1994;79: 165962.
25. SchifFRn A, Suissa S. Predicting nocturnal hypoglycemia in
patients with type 1 diabetes treated with continuous subcutaneous
insulin infusion. Am I Med 1987;82:1127-32.
26. Inada M, Oishi M, Nishikawa M, Kurata S, Imura- H. Chical
evaluation of measiuing glycosylated hemoglobin levels for
assessing the long-tenn blood glucose control in diabetics.
Endocrino1 Jpn. l980;27:4 1 1-5.
27. Kilpatrick ES, Maylor PW, Keevil BG. Biological cariation of
glycated hemoglobm: Ins>lications for diabetes screening and
monitoring. Diabetes Care l998;2 1 261-4.
28. Pate RR, Pmtt M, Blair SN, et al. Phrysical activity and
public healtk A recommendation h m the for Disease Control and
-
Prevention d the American College of Sport Medicine. $.MA
1995;273:402-7.
Ha Lean MEJ. Technical Review: Recommendations for the
nutritional management of patients with diabetes mellitus. Eumpian
Journal of CLinical Nutrition l997;52:467-8 1.
American Diabetes Association. Diabetes meIlitus and exercise.
Diabetes Care 1998;21(Suppl1):540-4.
Wasserman DH, Zinman B. Exercise in individuals with IDDM.
Diabetes Care 1994; 17:924-37.
Health and W e l f h Canada. Action Towards ... Canadian
Guidelines for Healthy Eating and ~ e c o k n d e d Strategies for
Implementation. The Report of the CommUnicatio~plementatiom
Conmittee. Ottawa: W s t e r of Supply and Services Canada;
1990.
Health and WeKare Canada- Nutrition reconmendations. The Report
of the Scientific review Cornmittee. Ottawa: Canadian Govenunent
Publications Centre; 1990:Cat No. H49-42/lg!JO E.
American Diabetes Association Position statement. Nutrition
recommendations and principles for people with diabetes mellitus.
Diabetes care 1994; 175 19-22.
CRC Handbook of Dietary Fiber in Humui Nutrition 2& ed. Baca
raton, FL: CRC Press; 1 993:lll-52.
Diabetes and Nutrition Study Group of the Europian ~ssociation*
for Study of Diabetes: recomrnendations for nutritional -nt of
patients with diabetes meIlitus. Diab Nutr Metab l995;8: 186-9.
World health Organization: Diet, Nutrition, and the prevention
of chronic diseases. Technical Report Series. l99O;797: 1-204.
Canadian Diabetes Association: Guidelines for nutritional
management of diabetes m e b in the 1990's (Positive Statement).
Beta Release 1989; 13%-17.
-
39. American Diabetes Association: Nutritional Recommendations
and principles for people with diabetes meIlitus (Positive
Statement). Diabetic Care 1995; 18(Suppll): 16-9.
40. 1998 Clinical practice guidelines for the management of
diabetes in Canada. CMAJ 1998;159(Suppl8),5 16.
41. Reeves ML, Seigler DE, Ryan EA, Skyler JS. Glycemic control
in insulùi dependent diabetes mellitus: comparison of outpatient
intensined conventional therapy with continuous subcutaneous
insulùi infusion. Am J Med 1982;72:673-80.
42. Bhasker R, Ghon MCY, Field JB: Tirne-action charecteristics
of regular and NPH insulin-treated diabetics. J Clin Endocrin Metab
1980;50:475-80.
43. Lean MJE, Ng LL, Tennison BR Intemal between insulin
injection and eathg ni relation to blood glucose control in adult
diabetics. Br Med J 1985;29O: 105-8.
44. Brange J, Owens DR, Kang S, Volund A. Monomeric insulin and
their experimental and clinical implications. Diabetes Care 1990; 1
3:923-54.
45. Ter Brack EW, Woodworth JR, Biamhi R, et al. Injection site
effects on the pharmacokinetics and glucodynamics of insuiin lispro
and regular insulin. Diabetes care 19%; 19: 1437-40.
46. Kang S, Brange J, Burch A, Volund A, Owens D. Subcutaaeous
mSuim absorption cxp1aMed by insulin's physicochemical properties:
Evidence fiom absorption studies of soluble human insulin and
insulin analogues in humans. Diabetes Care 1991;14:942-8.
47. DiMarchi RD, Mayer JP, Fan L, Brems DN, Frank BH, Green LK,
Hofnnan JA, Howey DC, Long HB, Shaw WN, Shields JE, Slieker LJ, Su
KSE, S d e U KL, Chance RE. Synthesis of a kt-acting insulin based
on structural homology with msulin-lilre p w t h -or 1 in peptides
Symposium. ESCOM 1992:26-8.
-
48. Brems DN, Alter LA, Beckage MJ, Chance RE, DiMarchi RD,
Green LK, Long HB, Pekor Shields JE, Frank BH. Altering the
association properties of insulin by amino acid replacement.
Protein Eng 1992;5:527-33.
49. ~ o w e ~ DC, Bowsher RR, Brunelle RL, Woodworth JR.
~ys(B28),Pro(B29)]-Human Insulin= A rapidly absorbed analogue of
human insuün. Diabetes 1994;43:369-40.
50. Anderson JHJr, B ~ l l ~ ~ l l e R, Koinsto VA, Trautmann
ME, Vignati S, DiMarchi R. The Multicenter Insulùi Lispro Study
Group. Improved meai time treatment of diabetes meIlitus using
insulin analogue. Clin nier 1997; 19:62-72.
51. Anderson JIZ Brunelle RI,, DiMarchi R, Koivisto VA, Pfiibner
A, Trautmm ME, Viguati L. The Multicentre Insulin Lispro Study
Group: Reduction of p o s t p d a l hypogly~emia and fiquency of
hypoglycernia in IDDM patients on insulin analogue treatment.
Diabetes 1997;46:265-70.
52. Schemthaner G, Wein W, Sandholzer K, Equiluz-Bruck S, Bates
P, Birkett MA. Postprandial insulin lispro: a new therapeutic
option for type 1 diabetic patients. Diabetes Care
1998;21:570-3.
53. Schernthaner G, Wein W, Sandholzer K, equiuz-Bmck S, Birkett
M. Postprandial use of insulin lispro: A new therapeutic option in
the treatment of type 1 diabetic patients? 9Abstract).
Diabetologica 1996;39:A24.
54. Ebeling P, Jansson P-A, Smith ULF, Lalli C, Belli GB,
Koivisto VA. Strategies towards improved control uring insuiin
lispro therapy in IDDM: Importance of basal insuiin, Diabetes Care
l997;20(8): 1287-9.
55. Koivisto VA. The human insulin analogue insulin lispro. Am
Med 1998;30260-6.
56. Anderson JH Jr., Brunelle RL, Koivisto VA, PiÙtzner A, T r a
m ME, Vigoati L, DiMarchi R Reduction of postprandial
-
hypoglycemia and ~~y of hypoglycemia in IDDM patients on
insuiin-analogue treatment: Multicentre InsuIin Lispro Shidy Group.
Diabetes 1997;46:265-7O.
57. Brunelle RL, Llewlyn J, Anderson JH Jr., Gale EAM, Koivisto
VA. Meta-analysis of the effect of insulin lispro on severe
hypoglycemia in patients with type 1 diabetes. Diabetes Care
1998;21(10).
58. Burge Castillo KR, Schade DS. Meal composition is a
determinant of lispro-induced hypoglycemia in IDDM. Diabetes care
l997;20(2).
59. WoIever TMS and Bolognesi C. Source and amornt of
carbohydrate affect postprandial glucose and insulin in normal
subjectsJ Nutr 1996; 126-2798-806.
60. Kong M-F S-C, Macdodd IA, Tattersall RV. Gastric emptying in
diabetes. Diabet Med 1996; 13: 1 12-9.
61. Macdonald IA. Physiological regdation of gastrying emptying
and glucose absorption. Diabet Med 19%; 13(SupplS):S 1 1-5.
62. Horowitz M, Edelbtoek MAL, Wishart JM, Straathof JW.
Relationship between oral glucose tolerance and gastric emptying in
nomal healthy subjects. Diabetologia 1993;36:857-62.
63. Wolever TMS. How important is prediction of glycemic
response? Diabetes Care 1989; l2:59l-3.
64. Wolever TMS, Jenkins DJA. The use of glycemic index in
predicting the blood glucose response to mixed me&. Am J C h
Niitr l986;43: 167-72.
65. Jenkins DJA, Wolever TMS, Taylor RH, et al. Glycemic Index
of f d s : a physiological b i s for cahhydrate exchange. Am J C h
Nutr 198 1 ;34:362-6.
66. Wolever TMS, Jenkins DJA, V - V, icAtm\en L, Jeakias AL,
Josse RG. Variation in d fat does not affect the relative blood
-
glucose response of spaghetti in subjects with type 2 diabetes.
Diab Nutr Metab 1992;5:191-7.
67. Rasmussen O. Dose-dependency of the glycemic response to
starch- rich meals in insulùi dependent diabetic subjects: Studies
at constant insulinemia. Metabolism 199334237 10- 1 3.
68. Hennanser K, Rasmussen O, A d k d J, Winter E, Schmitz O.
Glycemic effects of spghettï and potato wnsumed as part of mmed
meal on TDDM patients. Diabetes care 1987; IO:4Ol6.
69. Wolever TMSy Nguyen P-M, Chiasson J-L, H a t JA, Josse RG,
Pairnason C, Rodger NW, Ross SA, Ryan EA, Tan MH- Relatiomhip
between habitual diet and blood glucose and lipids in nominsulin
dependent diabetes. Nutt Res 1995; lS:843-57.
70. Reeves GM. Dietary therapy for non-insulin-dependent diaetes
rnellitus. N Eng J Med l988;3 l9:862-4?.
71. Collier GR, Giudici S, Kalmusky J, Wolever TMS, H e h G,
Wesson V, Ehrlich RM, Jenkms DJA. Low glycemic index starchy foods
improve *ose control and lower