Trial Information

Effect of Low Glycemic Index Diets (With Canola Oil) on Glucose Control in Non-Insulin Dependent Diabetics

Overall design

- Sample size justification: The sample size requirement was determined based on the
ability to detect a HbA1c reduction of at least 0.4 units, significance of the test (α)
of 0.05, β of 2 and power (1-β) of 0.80 (The FDA criterion for drug effectiveness is a
HbA1c reduction of 0.3 to 0.4 units). Examination of recent study results for HbA1c of
the research group on type 2 diabetic subjects indicated a standard deviation of 0.32.
So the desired total number to complete this study is N=80, or 120 enrolled assuming a
30% attrition rate.

- Design: All subjects will be randomized to one of two 12-week treatments in a
two-treatment parallel design. Treatments: 1) low glycemic index dietary advice (e.g.
to eat intact grain cereals, Pita Break Finland Rye and Little Stream Quinoa breads,
parboiled rice, cracked wheat, pasta, peas, beans, lentils, and baked goods made from
legume flour), with particular emphasis on a canola oil containing bread; and a 2) high
cereal fiber diet emphasizing whole grains.

Duration: The study will consist of approximately two months recruitment and patient
selection, and a 3 month treatment period.

Study Details: Fasting blood samples are obtained at screening, week -2, 0, 2, 4, 8, 10 and
12 of each study period for glucose, lipids and oxidation products. HbA1c will be assessed
on all visits except week 2. Twenty-four hour urine for urinary C-peptide, isoprostane,
urea, creatinine and electrolyte analyses will be obtained immediately prior to the
beginning of the study and at the end of the 12 week treatment. At weeks 0 (baseline) and
12, endothelial function will be assessed. Following the 12 week treatment period, subjects
who wish to undertake the alternate treatment will be given appropriate instruction.

- Setting: Risk Factor Modification Center at St. Michael's Hospital.

- Participants/controls (selection and inclusion/exclusion criteria):

Patient Selection: Healthy non-insulin dependent diabetic men and women will be recruited by
newspaper advertisement, physician referral and the diabetic clinic at St. Michael's
Hospital.

- Interventions: Diets: Diets will be the subjects' diabetic diets modified as above.
Diet histories will be recorded at weeks 0, 2, 4, 8, 10 and 12. These diets will be
assessed for consistency by the dietitian in the subject's presence. Where necessary,
modifications in diet will be made to ensure weight maintenance.

- Privacy protection Subjects will have their data de-identified. Any databases with
personal health information will be password protected. Also, patient data that has
personal health information (e.g. patient charts) will be kept in locked cabinets, with
the door locked to the room with the cabinets, and the centre door locked after 6 pm.

- Confidentiality of data All subjects will be assigned an identification code to ensure
confidentiality. All data entered into files will use the identification codes.
Access to data files will be limited to the PI, statisticians, dietitians, students,
and data entry personnel working on the project.

The data obtained from the Endo-PAT test will be sent to a private statistical consultant in
the United States for analysis. The dataset will be de-identified and only the statistical
consultant and the research team will have access to this dataset.

-Identifiable data Subjects will be primarily identified by a code (number and letter)
assigned by the study statistician. A separate chart with routine clinical information is
maintained for contact with participants' family physicians.

DATA ANALYSIS:

a. Analysis method, including types of statistical or qualitative methods Results will be
expressed as means ± standard error. The differences between treatment means will be
assessed using the CONTRAST statement in SAS (Statistical Analysis Software) using -2 and
zero bloods as baseline to calculate change and weeks 8, 10 and 12 for the CONTRAST.
Pearson's correlations will be used to assess relationships between dietary compliance and
weight reductions with blood lipid changes.

For Optional Sub-study:

In our recent diabetes study which looked at the effect of mixed nuts on blood sugar control
(REB# 06-274), 44.8 % (35/78) of participants continued onto follow-up and 24% (19/78)
completed follow-up. All crossed from either the muffin (control) or half dose nut
supplement to the full dose nut (test) supplement. In a recent Portfolio study
(REB#04-056), after 24 weeks 62.5% of participants crossed from DASH diet (control) to
Portfolio diet (test). Assuming that in this study participants will also cross from the
control diet to the test diet, we will use a paired t-test with no adjustment for baseline
or sequence, as this is not a true crossover design.