Diabetes and Diet:
A Type 2 Patient's Successful Efforts at Control
Substance Response Test on Author.
114 calories Shredded Wheat cereal
Blood glucoes in mg/dL
116 calories cheese omelet
(mainly protein and fat)
Food eaten at time zero
130 140 150
Time in minutes
Derek A. Paice
Copyright © 1997 by Paice & Associates Inc.
All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the
Paice & Associates Inc.
114 Rosewood Court
Palm Harbor, FL 34685
Neither the publisher, nor author, nor any person acting on their behalf (a) makes any warranty, express or
implied, with respect to the use of any information, method, or process disclosed in this book or that such
use may not infringe privately owned rights; or (b) assumes any liabilities with respect to the use of, or for
damages resulting from the use of, any information, method, or process disclosed in this book. This book
is not intended for use as a substitute for consultation with a qualified medical practitioner.
Note: Capitalized names of foodstuffs may be registered trade marks.
Library of Congress Catalog Card Number: 97-75726
Printed in the United States of America
Supplement added—January 2001
In 2003 this book was made available for viewing on the Internet. The contents are not to be construed as
offering any type of medical advice or recommendation.
1.1 Interpreting the graphs
2.0 MEASURING TECHNIQUES
2.1 Home tests
2.2 Laboratory tests
3.0 EXPERIMENTAL METHODS
3.1 Changing the body's performance
3.2 Changing the body's dietary input
4.0 EXPERIMENTS TO CHANGE THE BODY'S PERFORMANCE
5.0 EXPERIMENTS VARYING THE BODY'S DIETARY INPUTS
5.1 Experiments with equal calorie inputs
5.2 Experiments with various inputs and amounts
5.3 Response to snacks
5.4 Response to different breakfast meals
5.5 Short-term exercise effects
5.6 Effects of one type of "fast food," with and without wine
5.7 Effects of beer, wine, and distilled spirits
5.8 Foods that reduce blood sugar
5.9 Effects of improved blood glucose control on glucose tolerance tests
5.10 Effects of bread in consecutive and single meals
5.11 Comparing whole-wheat and pumpernickel bread
5.12 Effects of 7 years of type 2 diabetes on response to the same input
6.0 ANTICIPATING THE EFFECTS OF DIFFERENT FOODS
7.0 OTHER FACTORS
8.0 BLOOD ASSAYS
9.0 SUMMARY AND CONCLUSIONS
10.0 ON-LINE DIABETES SUPPORT GROUPS
11.0 MORE DATA ON SGI (Substance Glycemic Index)
11.1 Purpose and Benefits of the SGI
11.2 Individual Foods
11.3 Applying the SGI Table
11.4 Effects of Composite Foods
11.5 Combining Two Substances
11.6 Different Portion SGIs
11.7 Meal Glycemic Index
12.0 KEEPING A CHECK ON YOUR PROGRESS
13.0 TECHNICAL ASPECTS OF THE SGI
14.0 SOME LOW SGI RECIPES
APPENDIXES A-1 through A-6
In 1990 I noticed unusual tiredness after eating and I consulted with a doctor. She had me do a simple test to check for sugar in the
urine. The result was positive, and further testing established that I had type 2 diabetes. After that diagnosis I cut down on my sugar
input and generally watched my diet. Despite these lifestyle changes, however, after 5 years I noticed that my blood glucose readings
were getting higher. Also, side effects such as more frequent urination, and pain in the legs, were getting worse. On a routine visit
to my doctor I had a blood sugar level of 294 milligrams per deciliter (mg/dL). This is far too high, and I decided to get more
involved in my diabetes control. I set out to conduct a series of experiments on the effects of different foodstuffs. To me the results
were amazing and surprisingly straightforward. I found that my blood glucose levels were simply related to the type of food I ate.
I wondered why the simple relationships between blood glucose levels and the carbohydrates, protein, and fat that I ate hadn't been
explained to me earlier when I was first diagnosed and received "diabetes training." Also I wondered why these critical relationships
are not stressed in the magazines I read that deal with diabetes issues. Thinking that maybe the results were peculiar to me, with type
2 diabetes, I continued testing my sensitivity to various foodstuffs. The result was a number of graphs from which I could select
appropriate foods to eat. I also developed a means to determine whether my diabetes was getting better or worse. I made good
progress with this technique and was understanding the importance of different foodstuffs when Dr. Richard Bernstein's book Dr.
Bernstein's Diabetes Solution became available . His discussions tied all my experimental results together, and the critical
importance of carbohydrates became clear. It seemed there were no technical reasons why I could not enjoy excellent blood glucose
In discussions with others having diabetes I found that despite Dr. Bernstein's book, the vital link between carbohydrates and the
blood glucose level was still not universally recognized. Phrases like "your results may vary" are true in the sense that the response
to one type of food may vary from one person to another, but there are some universal basic relationships between food types and
the manner in which blood glucose is generated.
This book summarizes over 80 experiments with numerous graphs and charts, and it includes five blood assays over a 1-year period.
Pictures can sometimes be of great help in explaining basic ideas. I hope this applies to the glucose response graphs given in this
book. Figure 1-1 on page 3 is one of those graphical pictures that enabled me to understand the simple relationship between blood
glucose and food types. From that moment on I knew what tests I needed to make to get good control of my diabetes.
In mid 1997 the threshold at which diabetes should be suspected was lowered from a fasting blood plasma measurement of 140 to
126 mg/dL. The new blood plasma glucose level of 126 mg/dL corresponds to a whole-blood reading, as measured by many but not
all home BG (blood glucose) meters, of about 112 mg/dL. With these changes the number of people diagnosed with diabetes is
certain to increase dramatically. The good news is that this gives the opportunity to prevent the complications that can occur when
the disease is left untreated.
One cannot make dietary changes to improve blood glucose control without consideration of the effects on other parts of the body,
especially as regards the effect of medications. Thus any changes must be decided only after consultation with your physician.
Using results from my own experiments and reading books such as those by Dr. Bernstein, I have changed my eating habits, and my
own diabetes is now under excellent control without medication or insulin. A factor in blood analysis called HbA1c is used to
measure the average blood glucose level and indicates the degree of control. My HbA1c number went from 9.0% (poor control) in
1993 to 5.4% (excellent) in 1997. I look forward to many more years without the need for insulin injections. For those with type
1 diabetes and others in whom insulin injections are essential, I believe that an understanding of the body's responses to different
foods will aid in determining insulin doses. Medications and treatments are especially effective when they are tailored to the specific
needs of the individual. The testing methods described in this book will help you find your own specific response to different foods.
Detailed experiments were made over a 4-year period thus there is a time factor to be considered in evaluating some of the test results.
In 2000, the book was out of print, but requests for it were still being received. This edition is to support those requests, but it is more
than a reprint because it includes substantially more data, especially as regards the glycemic effects of different foods. For
convenience this additional information has been added in the form of a supplement to the original work.
I have had type 2 diabetes for 11 years and still manage to obtain good control simply by selection of appropriate foods with a low
SGI (substance glycemic index). In June of 2000 I had an HbA1c of 5.2%. I manage my control by paying careful attention to each
and every meal. Once it is accepted that good control is vital to one's health this careful attention to diet is not onerous. I usually eat
out at least once a week and even on a 15-day vacation in which I ate out every day I was able to select low SGI meals.
One of the ways in which I monitor my progress is simply by measuring fasting blood glucose (FBG) each morning. If it is higher
than my goal of 100 to 110 mg/dL (plasma) then I probably made a mistake in my eating the day before. For me the FBG seems to
be a good indicator of my type 2 diabetic condition. I speculate that the "Dawn Effect" in which the liver releases hormones and
glucose into the body each morning is in effect a miniature built-in daily glucose tolerance test. My FBG results reflect my response
to that daily test.
From the moment of my very first test I knew that in conjunction with my physician I could find better ways to control my diabetes,
and it is happening. I hope my experiments, described in this book, will be of similar help to you.
Derek A. Paice
Palm Harbor, Florida
The major source of technical assistance for this book is in the written works itemized in the reference section; however, informal
discussions through the Internet, the Diabetic mailing list at Lehigh University and America Online-bulletin boards, were invaluable.
The Lehigh University list is very active: more than 50 messages a day is not uncommon, and the information given there is so
helpful, ranging from detailed technical discussions to uplifting support for a member in need. Many people share their thoughts and
personal stories on these mailing lists and bulletin boards. We are indebted to them. It was there that I learned from Jenny McGrorey
that Merlot wine reduces blood glucose levels; also Dan Pipes reported, and Liz Tannenbaum confirmed, that onions had a similar
glucose-lowering effect. I used these observations for one of the experiments reported here.
Gretchen Becker, another on-line correspondent, encouraged me to document my test data in an appropriate book so that others might
benefit. She also provided very valuable technical discussion regarding diabetes and book-presentation issues. Lyle Taylor, a long-
time colleague, provided valuable suggestions after reviewing a much earlier version of this book. My personal physician monitored
my overall performance with routine checkups and blood assays.
Dr Bernstein not only reviewed the original book, but encouraged me to publish it and I am indebted to him for that. I trust that he
will find the information in the supplement equally worthy of being published. Dr. Jennie-Brand Miller, lead author of the original
the G.I. Factor, published in the USA as The Glucose Revolution, has from time-to-time offered helpful comments as also has my on-
line colleague Rick Mendosa.
If there are any inadequacies in this book they are mine alone.
Finally, I want to acknowledge the contribution of my wife Joan who willingly undertook a glucose tolerance test to generate data
for a nondiabetic response curve. Her contributions to my well being by helping develop low glycemic recipes, three of which are
included in this book, have been invaluable. Also, her tolerance and understanding of all those early-morning experiments made this
In managing diabetes the patient holds the key. Physicians and others can explain effects and make recommendations, but as yet there
is no magic pill to cure diabetes. Once diabetic patients accept responsibility for their care, however, I believe good blood glucose
(BG) control is possible for many.
The body includes numerous complex chemical control systems, many of which are adaptive. One of the body's systems controls
BG levels when ingested food is converted into glucose. The hormone insulin is produced by beta cells in the pancreas and helps
transport glucose into various body cells. Type 2 diabetes, also known as non-insulin-dependent diabetes mellitus (NIDDM), occurs
when inadequate insulin is produced, or the body's cells have a reduced ability to use the insulin that is produced (insulin resistance),
or combinations of these two limitations. Ways of combating these limitations are addressed.
In engineering it is common to characterize the performance of a machine by establishing what happens to the output for various
inputs while other parameters remain nearly constant. The same characterization process can be applied to the human body. I am
a 68-year old research engineer who has had type 2 diabetes for more than 7 years. By simply treating my body as a machine and
determining input (food) to output (BG level) I was able to characterize my basic glucose control system and from that knowledge
take steps to maintain good BG control. Tests on a single diabetic person have limited statistical significance because of individual
variations; however, as with any machine, some of the basic issues are readily defined.
Subjecting the body to various food inputs and measuring the resulting BG level versus time is here defined here as a substance
response test. This method can be used by the patient to evaluate his or her own glucose response to specific stimuli using a simple
BG meter. The results for any one patient are unique; however, some general conclusions can be drawn from my specific results.
For example, I found that carbohydrates had a rapid and major impact on my BG levels, whereas protein had a smaller and much
slower effect. Fat seemed to have no direct effect. The results agree perfectly with the low-carbohydrate diet recommendations of
R. K. Bernstein, M.D. . I also found that fiber reduced the rate at which the BG level increased, and fat when present in a large
amount reduced the effects that the carbohydrate in bread had on my BG level.
Knowledge of my test results should help others gain a better understanding of how their own bodies may respond to different food
inputs; and the test procedure I use has widespread application. However, results from my experiments will likely be somewhat
different from those with type 1 diabetes and different from those with more severe type 2 diabetes. Thus your results may vary.
My tests highlight some of the fundamental dietary issues that those with diabetes must address in controlling their BG level each
and every day. It is much more than simply counting calories.
The numerous experiments reported (a) show the dramatic effects of diet and different foodstuffs, (b) highlight the concept of
glycemic index, and (c) show deteriorating BG responses when good control is not maintained.
The graphs are powerful aids to learning; from them I was able to characterize my body's response to food and then maintain good
BG control. Using the test results in this book I reduced my hemoglobin A1c measurement (HbA1c), which reflects long-term
average BG levels, from 9.0% in 1993 to 5.4% in 1997.
At the outset I needed some sort of baseline reference to determine the extent of my diabetes and to gauge any changes that occurred.
A goal for the body's BG response was obtained from tests on a nondiabetic person, and then a search was made for substances that
might help me, a type 2 diabetic person, to emulate that response. Tests were made on over-the-counter and prescription substances;
however, some substances caused unacceptable gastric upsets, and nothing was found to improve my basic body performance. Much
more helpful results were obtained by determining responses to different foodstuffs.
The key for excellent control of my diabetes is in the data presented here. I hope these results are equally helpful to others and their
1.1 Interpreting the Graphs
his book summarizes over 80 experiments and includes more than 20 graphs to highlight important results. It has been said that a
picture is worth a thousand words. This is only true of engineering "pictures" if you know how to interpret them. At the risk of
explaining something that is already clear to many readers, I'll discuss how the pictures I present can be obtained and understood.
This is done using an example. Figure 1-1 is a continuous graph that enabled me to understand the simple relationship between the
BG level and food types. Along the bottom horizontal line is a scale marked in minutes of time, and along the left vertical line is a
scale marked in milligrams per deciliter (mg/dL) of BG, such as measured from a drop of blood obtained by pricking your finger.
At the start of a test, time is considered to be zero and the food or substance being tested is eaten as rapidly as possible. During the
course of digestion the BG rises and its value is determined using the BG meter. Readings are taken at appropriate intervals, say 15
minutes or less, and the results recorded. You can see individual results in figure 1-1, where they are represented in one case by small
circles and in the other by small squares. Each of those points is fixed in relation to the time and mg/dL lines. For example after 15
minutes a small circle shows that the BG level was about 113 mg/dL in the test with Shredded Wheat. When all the points are
obtained they are interconnected by straight lines to give a graph (picture) showing how BG varies with time. A dashed line
Another way to present a "picture" of the measurements is to use a bar graph in which the height of the bar shows the measurement.
This type of chart may be familiar to many. It is often used to present financial data, but it is more difficult to display multiple sets
of data on the same graph. Figure 1-2 is an example of this type of graph using data I had recorded over the years on my fasting BG
The fasting values presented in this graph are only from single test points, but they give a picture that shows how improvements had
occurred in 1997 after a lower-carbohydrate diet was implemented
2.0 MEASURING TECHNIQUES
2.1 Home Tests
Responses to different stimuli were primarily monitored using an Accu-Chek Advantage home-test meter. This meter measures the
glucose concentration in whole blood and is calibrated to read whole-blood values. Because of this it gives results about 12% lower
than the data obtained from laboratories, which separate the blood plasma with a centrifuge and then measure the glucose in the blood
Accuracy is affected by a number of variables, which are discussed by the manufacturer. A few home meters, such as the Glucometer
Elite, are purposely calibrated to try to match the plasma readings obtained by laboratories, so it is important to know what your meter
reads. In this book all readings are whole-blood readings unless indicated otherwise. Most of the data relate to the BG level after
a given stimulus. The peak BG and the time to reach it are easily determined. There is no specific definition for what constitutes
an acceptable maximal BG value, but my goal was to prevent my BG from exceeding 140 mg/dL (157 mg/dL plasma) at any time.
Although individuals may vary, the BG level at which appreciable amounts of glucose appear in the urine (glycosuria) is about 180
mg/dL (plasma). My goal is 13% less than this value.
2.2 Laboratory Tests
Glucose becomes chemically incorporated into proteins, including hemoglobin in the red blood cells, at a rate determined by the BG
level. This process (glycosylation) provides a way to determine the average BG level over a period of about 120 days, the lifetime
of a red blood cell. The HbA1c value measures this glycosylation and is useful for determining whether the average BG level is being
controlled satisfactorily. Because it provides a 120-day averaged result, it does not change significantly over a short period. For this
reason some people recommend waiting 6 to 8 weeks between HbA1c tests . Other methods to measure average BG levels are
available; for example the fructosamine test measures the average BG level over about a 3-week period. The relationship between
the HbA1c test and the assessment of quality of control may vary from one laboratory to another. Table 1 gives an example.
Table 1 Comparison of Blood Glucose Control Levels *
Quality of Control
Average Blood Glucose
Less than 6.4
Less than 128
6.4 to 7.6
128 to 152
7.6 to 8.8
152 to 176
More than 8.8
Greater than 176
* This is only a guide. Different laboratories may use different numbers.
3.0 EXPERIMENTAL METHODS
3.1 Changing the Body's Performance
To determine the effects of medications or dietary supplements, I first determined my body's basic response to a glucose stimulus.
This glucose response test was then repeated after some period of time, usually 2 to 3 weeks, during which the medication or the
supplement under test was used. Although this was the usual procedure, another test that monitored fasting BG over a similar-
duration test period was also implemented in one of the supplement evaluations. Exercise effects were evaluated in a similar fashion.
3.2 Changing the Body's Dietary Input
This approach was found to be very enlightening. For results to be compared, tests are done from the same starting conditions. Most
of my tests were carried out from the early morning fasting condition. For those who experience significantly increased BG levels
in the morning (dawn phenomenon), a different time might be necessary. The procedure was to eat the test substance and measure
the BG level at appropriate intervals for a period of about 2 hours. By drawing the graph as the test proceeds, one can anticipate an
appropriate time for the next reading before too large a change in BG occurs. The final graph was analyzed to determine the peak
BG level (usually occurs after 30 to 75 minutes), and in some cases an average BG level over the 2-hour period was determined.
Substance response test results helped me choose appropriate foods for my diet.
4.0 EXPERIMENTS TO CHANGE THE BODY'S PERFORMANCE
Figure 4-1 illustrates the response of a person without diabetes and myself (with type 2 diabetes) to a 12-gram glucose stimulus
obtained by ingesting glucose tablets. In this experiment, we each ate the 12 grams of glucose in less than 4 minutes. You can see
that initially my BG increased almost 60% faster than that of the nondiabetic person. Also it took my body 32 minutes to start
reducing the BG, whereas the nondiabetic person started to reduce her BG after only 23 minutes. Finally it is seen that the average
BG for me is much higher than that of the nondiabetic person. To improve my body's response I wanted to reduce the peak and the
2-hour average BG to more closely match that of the nondiabetic person.
Trying to change the body's performance was unsuccessful. The data presented here are focused over a 12-month period, but test
data taken 7 years ago are similar except there has been a general worsening of performance. The diabetic condition is indeed
stubborn. This section contains a brief summary of experiments in which I tried to improve my basic body performance.
Prescription medications were attempted, but in each case I suffered gastric upsets that prevented a full 2-week trial from being
completed. These medications included glyburide, metformin (Glucophage), and troglitazone (Rezulin.) All of these medications
have been reported in the literature to reduce the HbA1c results, and Rezulin (withdrawn from the market in 2000) appears to change
the body's response function in a very desirable way by reducing insulin resistance. Further testing with these drugs is left for others.
Mineral supplements like chromium, magnesium, and zinc had no discernible effect, and vanadyl sulfate, which was found to reduce
my fasting BG level, was discontinued because it caused me unexplained chest pains.
In this phase of testing to change my body's basic performance, I found it beneficial to incorporate a high-fiber cereal such as
Kellogg's All-Bran into my diet. This had the effect of slowing down the rate at which my BG increased and also the peak BG. The
effect is presumably due to some sort of filtering action, and I emphasize that the All-Bran was not ingested at the same time as the
glucose: thus improvement is caused by fiber ingested at least 24 hours earlier. Whether this is truly changing the body's performance
or is in effect a diet change is debatable. Measurements during this period indicated that the increased fiber also helped reduce my
Figure 4-2 compares the performance in response to a glucose stimulus before and after All-Bran was added to my diet. To emphasize
the results and eliminate any differences in fasting BG, only the changes in BG are plotted in this figure. The results were sufficiently
encouraging for me to add All-Bran and later Bran Buds to my diet.
Acarbose, a prescription medication taken before eating, is touted to achieve similar or more pronounced reductions in BG levels for
"double sugars" (disaccharides) like sucrose or complex carbohydrates like starch, but not for "single sugars" (monosaccharides) like
glucose [3, 4]. However, no tests were made with acarbose.
Conclusions from the experiments shown in figure 4-2 are summarized as follows:
1. The peak BG increase was reduced by 18%.
2. The average BG increase was reduced by 29% over a 2-hour period.
3. The initial rate at which the BG increased was reduced by about 20%
(similar results were found after 32 days on the All-Bran breakfast diet).
5.0 EXPERIMENTS VARYING THE BODY'S DIETARY INPUTS
This section describes my responses to various foodstuffs. It consists primarily of graphs that are labeled to indicate the substances
under test. Each test is briefly discussed. The actual data points are given and connected by straight lines. In some of the graphs
this gives rise to small perturbations, rather than a smooth curve. No explanations are attempted for these perturbations. They are
believed to be real and not simply the result of meter errors. However, it is easy to visualize a smooth curve interpolated between
the points to give the average BG, which is the prime interest. The main focus of the work took place over a period of about 12
months and involved testing many different foodstuffs. There is a chronological factor to be considered in the results because as time
progressed I observed improvements in my glucose system response because of the integrated effects of a different diet. Thus, for
example a BG response in November 1996 will not appear as favorable (low) as one made in July 1997. It transpires that the low-
carbohydrate diet used to control the BG level on a day-to-day basis also improved my body's basic response to glucose. Thus the
low-carbohydrate diet had a compounding effect.
The results given in figure 5-12 cover a span of 7 years. These data illustrate gradual deterioration of the transient BG response when
control is not especially good. The increase in fasting BG level is also evident. The eventual effects of good control on the fasting
BG level are also apparent in this figure.
5.1 Experiments With Equal Calorie Inputs
Figure 5-1 shows the effect of different substances with a similar calorie content.