2 April 2006

GI News Briefs

Memorable Meals
In recent years, a number of studies have shown that older people with type 2 diabetes have significantly greater risk of performing poorly in cognitive function tests such as recalling word lists. University of Toronto researchers say that a low-GI meal generally results in better verbal memory in the post-meal period, particularly in those who experience the greatest food-induced elevations in blood glucose levels, compared with a high-GI meal.


They reported their findings of a study involving 21 older adults (10 males, 11 females) with type 2 diabetes in the March 2006 issue of Diabetologia. Both the GI of the carbohydrate meal and individual differences in response to the meal contributed to the variation in consequent memory recall, they write. The scientists had set up the study to determine whether minimising the increase in blood glucose by eating a low GI meal (50 g pasta) rather than a high GI meal (50 g bread) would affect cognitive performance in the post-meal period. They found that:

  • Bread consumption, relative to pasta, resulted in both a higher area under the curve and worse delayed verbal memory (paragraph and word list recall) performance;
  • Performance following bread consumption was poorer than that following pasta on measures of working memory, executive function and auditory selective attention;
  • Sustained attention showed no sensitivity to type of carbohydrate food consumed.
Diabetologia, March 2006 (email: carol.greenwood@utoronto.ca)

Men Are Different
Men with cardiovascular disease may be at considerably increased risk for death even when their blood sugar level remains in the ‘normal’ range, suggests a new study by a team of scientists at UCLA and Cedars-Sinai Medical Centre in Los Angeles published in the American Journal of Epidemiology. Using extensive data from the Framingham study collected between 1948 and 1978, the scientists statistically analysed the connection between glucose (blood sugar) levels and death in patients with cardiovascular disease. ‘While our results show that glucose is a strong independent predictor of mortality,’ writes lead author Dr Sidney Port, ‘they also unequivocally show that the mortality/glucose relations are quantitatively different for men and women. Indeed for men risk rises very rapidly through the normal range and is flat thereafter. For women, risk is unrelated to glucose in the normal range, then rises quickly through the impaired range and at a much slower rate thereafter.’ The findings suggest that: ‘although 5.55 mmol/liter (= 100 mg/liter, normal) may be a useful mortality risk division (albeit with different implications for men and women), 6.99 mmol/litre (= 126 mg/litre, diabetic) is not, especially for men.
American Journal of Epidemiology, vol 163, No. 4

A Heart-healthy Diet
Foods based on soybeans have a beneficial role in our defence against heart disease.

photo: New England Journal of Medicine

There are three attributes of soybeans with the potential to reduce heart disease risk:
  1. soy protein
  2. anti-oxidant substances called isoflavones
  3. low GI carbohydrate
In a new study published in Nutrition, lead author Dan Lukaczer reports that postmenopausal women on a low GI diet (incorporating 30 grams of soy protein and 4 grams of phytosterols a day) showed a ‘significantly greater improvement’ in CVD risk factors compared with women on the AHA Step 1 diet. Scientists from the Functional Medicine Research Center of Mutagenics Inc., the company that manufactures the fortified soy beverage used in the study, carried out the 12-week study. The average age of the women in the study was 55, and all were clinically obese and at higher risk of CVD. Both groups started with high levels of total cholesterol, low-density lipoprotein (bad) cholesterol and triacylglycerols. The scientists collected 3-day diet records from the women at weeks 2, 4, 8 and 12. At the end of the program the low GI group showed statistically significant decreases in total cholesterol, low-density lipoprotein cholesterol and triglycerols say the authors. In addition, high-density lipoprotein (good) cholesterol rose significantly in the low GI group group, but not the AHA Step 1 diet group.
Nutrition (Vol 22, pp104–113)

Evaluating the value of GI …

‘The present results call into question the utility of GI and GL to reflect glycemic response to food adequately, when used in the context of usual diet’ write Elizabeth Mayer-Davis and her co-authors reporting their study in the British Journal of Nutrition. Using the glucose tolerance and insulin sensitivity data for 813 adults who participated in the Insulin Resistance Atherosclerosis Study, the researchers set out to evaluate the correlation of GI and GL with measures of glycaemia based on the responses the participants had given to that study’s food frequency questionnaire. They assigned mean GI values based on published data. Despite acknowledging the limitations of their study, their major finding was that ‘no statistically significant association was observed for GI in relation to fasting glucose, postprandial glucose, incremental postprandial glucose or A1c either at baseline [the start date] or at the follow-up examination.’
British Journal of Nutrition (2006), 95, 397-405

GI Group: This study has attracted an amazing amount of headline-making media mileage and we have received a number of inquiries from GI News readers. Here is a brief summary of our key concerns with the design of this study and its findings.
  1. The food frequency questionnaire used for Insulin Resistance Atherosclerosis Study has a poor correlation (just 0.37) with other measures of carbohydrate intake. What this simply means is that it doesn’t provide truly reliable data on how much carbohydrate the volunteers ate. Without accurate information on an individual’s carb intake it is unlikely they can confidently assess its GI or to calculate GL. Assessing food intake from food frequency questionnaires (FFQ) is subject to error and bias (it’s human nature for people to under-report some foods and overestimate others). This is why researchers always compare the estimates of nutrient intake from FFQs with an independent method such as a dietary record. A good study will have a correlation value above 0.5.

  2. No FFQ has been validated as truly reflecting the GI of a diet. For example, one question reads: ‘How often do you eat dark bread (including whole wheat, rye, pumpernickel, other high-fiber bread)’. Obviously the GI values of these various breads can vary greatly, yet a single GI value was assigned to all responses to this question.

  3. The cross-sectional, rather than prospective, nature of the Insulin Resistance Atherosclerosis Study. A better study would take either normal individuals or people with impaired glucose tolerance (IGT) and track changes in glucose metabolism over time and try to relate those changes to differences in food quality (such as the Harvard Nurses' Studies).


Melissa S. Green said...

The Mayer-Davis study really struck me as being not very well designed when I read press accounts of it. I'm very glad to read your response to it.

The accounts said also that the blood glucose of study subjects was measured only twice, which struck me as another weakness.