Finally got to follow up the links on Lyle's page.
1.
Pittas et al in "A Low-Glycemic Load Diet Facilitates Greater Weight Loss in Overweight Adults With High Insulin Secretion but Not in Overweight Adults With Low Insulin Secretion in the CALERIE Trial" compared two diets over 6 months.
The "High GL" diet was 60% carbs with an estimated daily average Glycemic Index (GI) of 86 (!) and a glycemic load (GL) of 116 g/1,000 kcal.
The "Low GL" version was 40% carbs with an estimated GI of 53 and GL of 45 g/1,000 kcals.
We can see that the low GL version is still a moderate carb diet, and that the GI of the high GL version is very high. Healthy women and men aged 24–42 years with a BMI of 25–29.9 kg/m2 and fasting plasma glucose <100 mg/dl were recruited.
Subjects were grouped according to their insulin value at 30 min (INS-30) after a 75-g oral glucose tolerance test. So we have four groups - high and low GL diets in each of high and low INS-30 insulin levels (above and below median value). All food was provided over 6 months and all groups lost weight with no significant differences in energy intake between groups.
The most statistically significant difference was between the weight loss of the Low Glycemic load diet between the two insulin groups - the high insulin level (high insulin response to glucose test) group lost more weight on the Low GL diet than the low INS-30 group.
A significant difference (at the 5% level) was also seen in the difference in response of the high INS-30 group to the two diets - they did better on Low GL.
The difference in response to the two diets among the low INS-30 group was not statistically significant, nor was the difference in performance of the High GL diet between the two insulin level groups.
In summary - if you have a good insulin response to glucose test then the evidence says you'll lose more weight on a Low GL diet. If you have a poor insulin response it may not matter, and on a High GL diet it doesn't matter what your insulin sensitivity is.
With N=8 subjects in each group and mixed gender the study is going to struggle to detect significance in modest differences, so the above outcomes should be taken in context. The usual call for further studies
Finally, the often used HOMA-R insulin sensitivity assessment did not find any statistical significant effects - "We also examined whether baseline HOMA-R predicted weight change, and we found no diet × HOMA-R interaction."
To be continued....