Personalized Nutrition II

More on Zeevi et al. (2015) (this is a follow-up to part 1)

I like this study a lot, or at least the fundamentals… or new tools that it might bring to the table.  Like, we know sleep and physical activity are important, and we know all calories aren’t created equal.  This study is the next level, showing there are even big differences in specific carb-rich foods depending on who’s eating them.

And more interestingly, if I’m interpreting the results of the intervention study correctly (which may not be the case), gut microbial responses to specific foods were very individualized… and predictable!

But first, the main part of the study — standardized meals (after overnight fast): 50g carbs from glucose, white bread, bread and butter, bread and dark chocolate, and fructose.  All repeated at least once (except fructose).  Everyone responded pretty similarly to fructose (little to no blood glucose spike), but a wide range of responses to glucose.

PPGR = PostPrandial Glucose Response


glucose and fructose






The range of PPGR to bread was ~15 to 79!

Again, here are some of the findings I found most interesting (besides the huge range in glycemic response to bread):



banana and cookie


Participant #468 has a consistently higher response to glucose than to white bread.  Participant #663 is the opposite.  And participant #445 is still winning.

I truly wonder if there’s a gut microbe (or something) that’s involved here…



This graph is interesting:


Carb content and PPGR


Different carb-rich foods are listed on the bottom.  The green boxes show that most of these data came from meals of roughly 25-35g carbs.  The blue lines show the average PPGR.




Rice and potatoes have a high glycemic index and accordingly, are more toward the right side of the figure (ie, higher PPGR); ice cream and dark chocolate have a lower glycemic index and appear on the left side of the figure (ie, lower PPGR).  But the ranges in response are HUGE (eg, see pita).



Part 2. the follow-up/intervention study: people were assigned a low or high PPGR diet for one week each.  Notably, not everyone’s low PPGR diet was the same, eg, the ‘healthy’ low PPGR diet for participant #445 could have cookies but not bananas, and the opposite for participant #644.


It worked:


Postprandial glucose response success


The ‘biome went wild, as expected, because of the non-uniformity of individual PPGRs (I think).  There are a few ways to consider these data: 1) find correlations between microbes and baseline PPGR; 2) did the species that were correlated with high/low PPGR increase/decrease on the high/low PPGR diet? …and 3) does this make any sense?

A lot of Bacteriodetes predicted a good response to bread and butter, whereas a lot of Proteobacteria predicted the opposite.

Some species negatively associated with T2DM, like Roseburia inulinivorans, Eubacterium eligens, and Bacteroides vulgatus, all increased following the good diet and decreased following the bad diet.


Differential glucose responses


One KEY point here is that many of these changes are in line with previous findings, suggesting we’re at least starting to crack the ‘biome puzzle.

And one COOL point is that, recall, the microbial changes on a good diet were similar regardless of whether said diet had cookies or not!  in other words (if I’m interpreting this correctly), if cookies spike your blood glucose, then you wouldn’t experience the beneficial changes in your ‘biome that participant #445 did.

Not sure if Dr. Deans was referring to this specifically, but it works:



Become a Patron!


Update: the ‘biome might even influence how we respond to cold exposure


calories proper


Be Sociable, Share!
  • Jim Catalano

    I want the Cookie Monster’s microbiome!!

    • hahaha

      or at least that of participant #445, cuz it looks like the Cookie Monster has a good deal of visceral adiposity 😛

    • erdoke

      Maybe he/she is ready to sell! ;o)

      • “insulin response is more important than BG”

        not really sure what you mean by this…

        insufficient insulin response ? hyperglycemia ? no bueno

        Check out Brownlee’s take on this:

        • Gabor Erdosi

          I meant that the insulin score of foods is more important in modern context than the glycemic index. These are not directly linked and normoglycaemia is often only achieved by hyperinsulinaemia that can remain hidden for years or even decades due to the focus on BG.
          I propose that “sufficient insulin response” occurs only when there is no need for supra-physiological levels to reach normoglycaemia (BG<90/5.0).
          frequent high insulin ?? no bueno
          PS: sorry, it's still me, just logged in with a different service on another machine…

          • weilasmith

            “…normoglycaemia is often only achieved by hyperinsulinaemia that can remain hidden for years or even decades due to the focus on BG.” exactly!!!!

          • weilasmith

            “..a potentially very long period of diabetic derailment (according to my definition and not to SOC’s) remains unrecognized.” you nailed it on the head YET again.

          • I understand your point, but hyperglycemia is largely responsible for diabetic complications…

            If your interested, I wrote a blog about the Insulin Index:

          • erdoke

            Thanks, I am aware of both the Banting lecture and nearly all recent blogs about insulin.
            I believe we face a definition problem here. Due to unrealistic characterization of “diabetes” and also “normoglycaemia”, a potentially very long period of diabetic derailment (according to my definition and not to SOC’s) remains unrecognized. As also included in the Brownlee talk/paper, a bit elevated BG together with high insulin (he says high FFA though) are able to induce the same complications, while chronically high insulin undoubtedly has its own negative effects on health.

  • Thomas Hemming Larsen

    Another great one about personalised nutrition! I don’t quite understand your conclusion about the biome and cookies. Do you mean that you can have cookies on a ‘good’ diet and still see positive changes (because the cookies are a non significant part of the diet)? I don’t understand how you from that can infer the section in italic. Do you mean that if you have a problem with cookies, a ‘good’ diet can’t change your biome?

    • honestly, I haven’t fully wrapped my head around it, either!

      If I read it correctly (and this could be a big “if”), carbs that don’t spike your glucose may positively impact the ‘biome. This could be bananas for some people but not others, depending on their own personal PPGR to bananas.

      • Thomas Hemming Larsen

        Ah okay, now I understand what you mean. If that’s the case it would be really cool 🙂

  • rs711

    The graph in figure F – what can me make of it given the massive error bars?

    • they’re not error bars bc there isn’t a statistical comparison in this figure — it’s more of a way to show the biological/physiological range of PPGR. See also the figures for glucose & fructose — more ways to show the range of PPGR.

      • rs711

        Ah i see, got it. i haven’t looked at the paper yet.

        Still not quite sure what to make of the upper range (meal carbohydrate content, grams).

    • Duh, the cookies were made with resistant starch.

  • 64 magpies

    I don’t know how pita ended up on the left side of the graph? With low fat and low fructose surely should be on right side?

    • good question

      An general, whole grain bread has a lower GI than refined white bread.

      Also, the figure isn’t showing GI, but rather PPGR to meals containing these foods as the predominant source of carbs. So on average (the blue dash), pita-containing meals elicit a low PPGR, although the huge PPGR range (the big box) suggests that some people exhibit high a PPGR — maybe the latter reflects people who eat them plain…

  • Pingback: Tedenski pregled #32()

  • Robert Rio

    To make it even more complicated what if the person reacts differently to bananas at one stage of his/her life, and differently at another? Bananas don’t spike the glucose at 25, do spike it at 55.

    • no doubt, especially (and this seems to be the case) if the ‘biome is involved