Macro Ratios & Glucose Management: Eating Lower GI Carbs and Higher Protein Alone is Less Effective in Blood Sugar Normalization Than You May Have Thought

Macronutrient ratios matters, but food quality does, too. And so do exercise, laziness, sleep, .... the list is endless, so can we be surprised that modulating GI and protein is not helping much?

We all "know" that going low(er) carb is good for your blood glucose management, right? Nice! So, we probably don't need studies like the Marleen A. van Baak's conducted only recently, right? I mean, carbs are bad! So why would we even be interested to hear how much the subjects of the Diogenes Study lost on one of four different diets with varying protein content and glycemic index? Or who would want to know what the subjects' 24-h glucose profiles on one out of four diets differing in carbohydrate content by 10 energy % and glycemic index by 20 units during three days. No one would like to know that, right!?

In view of the fact that you're still there, I suppose that I was wrong and you are interested in the effects of different marconutrient ratios and types of carbohydrates (low vs. high glycemic index) on weight loss and glucose management. Let's take a closer look at the design of the DiogenesTrial (learn more), then.
The Diogenes Study was a field study, in the course of which subjects were randomized into 5 different diets: lower protein (LP)/lower GI(LGI), higher protein (HP)/LGI, LP/higher GI (HGI),HP/HGI and a control diet according to national recommendations for a healthy diets.

• The intended difference in protein content of the LP and HP diets was 10%–12% of energy intake.
• The intended difference in glycemic index between the LGI and HGI diets was 15. 

All diets were reduced in fat (<30 energy%) to make room for significant differences in the CHO to PRO ratios. All subjects had to keep food diaries and monitor their glucose levels continuously.

Table 1: Examplary diet composition in the Diogenes Trial (van Baak. 2014)

As some of you will probably remember (I've written about the Diogenes Studyand its results in the Facebook News, previously), the total amount of weight the subjects lost from the start of the Diogenes Study was 8.8 ± 8.4 kg. The subsequent weight regain since the start of the randomized diet intervention was 1.6 ± 7.3 kg, with no significant differences between the diet groups (in spite of a marginal advantage for those who kept eating a high(er) protein diet).
The analysis of the three-day dietary records showed the expected differences in carbohydrate content and glycemic index between the diet groups, but the differences were smaller than intended (a bummer in view of the fact that they were not large to begin with). Much in contrast to what our initially phrased prejudice would say, there were no significant differences in mean 24-h, daytime or nighttime glucose concentrations between the diet groups.

Table 2: Self-reported dietary macronutrient composition in the field study (a) and macronutrient composition of the diet in the lab study (b) (van Baak. 2014)

Moreover, the absence of at least a deviation in standard deviations (SD) signifies that the lack of statistical significance is not due to extreme outliers who may not have adhered to their dietary prescription (macronutrient make-up / reported / see Table 2).

"Post hoctesting did not reveal differences between the LP/HGI diet and any of the other diets for the glucose parameters studied. Adjustment for BMI, HOMA-IR (Homeostatic Model Assessment of Insulin Resistance) index, minutes of exercise performed by the subjects, total weight loss or weight regain did  not change this outcome (data not shown)." (Van Baak. 2014)

Still, in a relatively uncontrolled scenario like the one in the Diogenes Study it's always possible that someone cheated twice, i.e. he deviated from the diet and reported to have stuck to it to the literal "T". Against that background, the results of van Baak's three-day follow up in a very controlled scenario is all-the-more interesting.

Figure 1: Parameters of glucose homeostasis means (left) and standard deviations (right) on the different diets in the lab study (van Baak. 2014)

It's a study in which we see the same non-significant effects of the macronutrient ratio on ability of the study participants to keep their glucose in check. With goodwill you could say that protein seems to buffer the glucose excursions, but with p-values of p >> 0.05, this effect is everything but statistically significant.

To find an effect that is statistical significant we do thus have to take a look at the continuous overlapping net glycemic action over 1 h periods (CONGA1) values over 24 h on the different diets in the lab study, where the high protein + low GI group finally shows the statistically significant advantage we've been waiting for all along (see Figure 2). Still, as the scientists point out, overall their data does not support the hypothesis that "glucose concentrations would be highest on the reference diet (LP/HGI) compared to all other diets and especially compared to the HP/LGI diet." (van Baak. 2014). As the author points out, the study did yet provide support for the hypothesis regarding glucose variability was found, since both 24-h and daytime variability of glucose concentrations were lower on the HP/LGI diet than on the LP/HGI diet.
Reference:

• van Baak, Marleen. "24-Hour Glucose Profiles on Diets Varying in Protein Content and Glycemic Index." Nutrients 6 (2014):3050-3061