Better Starch Digestion = Lower Weight?

This paper was posted by Denise Minger on Facebook last night:

It has been known for some time that humans carry different numbers of copies of the genes responsible for producing amylase.  Amylase is the enzyme in human saliva that begins the breakdown of starches in the mouth.   Obligate carnivores, such as cats, lack salivary amylase.   The copy number of the AMY1 gene in humans range from 2 to 12 in this study (on FB, Denise states that it can vary from 1 to 15).  Want more information on this, try Chris Masterjohn's presentation from AHS12.  The amylase part is pretty near the beginning.

This graphic comes from around the 4 minute mark where Chris discusses how all of the chimpanzees have the same 2 copies but in humans we see varied duplication.  Roughly 10% or less of either the high starch or low starch human cultures carry only 2 or 3 copies, and it would appear that less than 5% of high starch cultures carry more than 10, while virtually none of the low starch do.

Of note, in the high starch cultures, roughly 70% have between 5 and 8 copies, while over 50% have just 4 or 5 copies in the low starch cultures.  

This graphic comes from Diet and the evolution of human amylase gene copy number variation, Perry  Here is the cumulative data per culture from that paper.  It is interesting that European American is designated as high starch.  I tend to think "high starch" when I think of traditional Thai diet (blogged on here).  But there's the rub, Euro-Ams have high gene copy number as a population.

Notably among the authors of this paper is Nathaniel Dominy, an actual anthropologist who has studied starch consumption in our ancestors and ... well ... watch this video.

OK ... so the new Nature paper.  They used data from several studies where they could compare obese vs. normal weight cohorts.  

The resuts shown graphically below, are a bit underwhelming.  We start with scatter plots of BMI vs. qPCR (a surrogate for gene copy number) in two studies.

Note that qPCR is a surrogate for gene copy number. The slopes of those lines are all but flat, 2-3 BMI units over the full range of qPCR. Next up the box plots for gene copy in two studies. Are there differences? Apparently statistically, yes. Are these meaningful? I'm less than convinced but please chime in in the comments section if I'm missing the obvious here!

Next up, the distribution for obese vs. normal weight (same designations as above) in these three studies.  There is a shift for LOWER copy number among obese.  

Again, I'm not really convinced that these are "significant" in any sort of meaningful way.  In all three studies, the peak gene copy is 5 for the obese and 5 or 6 for the normal weight, you have a little more with only 4.  How about 2 and 3?  Does this low copy number predispose towards obesity?   How about if you look at this the other way?  Only around 2.5-6% of the obese have only 2 copies, vs. about 2.5% normal weight.  Three copies?  We're up to 7.5-12% vs. 5-6%.  I always view such numbers and differentials through the lens of 1950's America vs. 2000's America and obesity rates.  Did our starch consumption change drastically?  Not really.  Did our genes change?  No.

Lastly we've got BMI vs. copy number in the two studies from the scatter plots.  Note that this is an "extreme closeup" on the y-axis.  A BMI differential of less than 2 points in either case over the range of copy numbers.

Oh wait!  If we flip the axes and plot just the "normal" range of copy numbers vs weight classifications we get a line.  This appears to be the "money plot" from this study.

If there is anyone who wants to delve deeper into the methodology used and results, please feel free to contact me or add your analyses to the comments here.  This is beyond the scope of my time and priorities at the moment.

However, here's my money quote:
Lower blood amylase levels have been observed in both obese humans and rats and have recently been associated with increased risk of metabolic abnormalities and reduced preabsorptive insulin release.

This DIRECTLY contradicts the Taubes Wrong Insulin Carbohydrate Hypothesis Of Obesity - TWICHOO.  

First of all, salivary amylase begins the process of the breakdown of complex carbs like starch.  But salivary amylase is not alone, it is joined by pancreatic amylase later down the line.   If one were deficient in amylase, they could not break down the starch fully and more would pass through -- like cellulose -- without being digested and absorbed.  Digestive enzymes are NOT metabolic enzymes.  Digestive enzymes act essentially "outside" the body (our gut if technically external to our organism) to break down foods into components that can be absorbed by the body.  If we don't absorb something, we can't use it for anything.  At best, this person's gut bacteria get access to a bit more starch, but whether this is good or bad would be a topic for another time.  In any case, low amylase would be akin to taking starch blockers.

If the obese got that way by absorbing carbs which spike insulin and trap their fat hopelessly in their fat cells, this is not consistent with the obese being less able to digest those carbs in the first place.   Amusingly enough, a student of Ayurveda is over on Denise's wall laughing at this, because his studies have led him to believe that poor digestion leads to excessive weight gain.  Oh well ...  chuckle.

What I do find interesting is the reduced insulin release.  This implies that there may be some sort of amylase feedback with the pancreas and salivary amylase may be critical to the early release of appropriate amounts of insulin.  Time and again, as glucose "tolerance" declines, it is the relative deficiency of early postprandial insulin that emerges early on.

Or ... Gary Taubes could update his "Japanese don't eat sugar and eat brown rice" explanation for that TWICHOO-defying culture to "the Japanese digest their starch better, making for more teaspoons of glucose than an elephant could handle, and that's why they don't get fat".  Oh wait.  That doesn't make any sense.  


Wuchtamsel said…
Uhhh, putting a video with Chris Masterjohn at the beginning of a post somehow reduces my motivation to read it close to zero...
Kevin Klatt said…
I'll take a closer look at the data when I a longer minute but have you read Mandel's paper on amylase copy number? I reference it in my post on amylase's evolutionary history ( They also suggested a pre-absorptive insulin release related to AMY copy #
Stephan Guyenet said…
Previous publications on AMY1 have shown that people with high copy number have a higher cephalic phase release of insulin when consuming starch, however their overall insulin release is approximately the same (trend toward slightly lower total insulin). They have much better starch tolerance than people with low copy number, yet glucose tolerance is the same. It's been hypothesized that AMY1 releases glucose oligomers in the upper alimentary tract upon starch consumption, which signals through glucose/oligosaccharide receptors, facilitating early insulin release and possibly insulin-independent disposal of glucose. I'd like to see the work replicated and extended, because aspects of it were not very convincing to me and the mechanism remains unclear.

So the finding that AMY1 copy number is related to obesity risk appears to have no bearing on the insulin-obesity hypothesis one way or another. AMY1 contributes relatively little to starch digestion relative to pancreatic amylase, and these differences probably have little or no effect on the amount of starch that ends up being absorbed. AMY1 is probably playing a regulatory role in this case, perhaps enhancing the signal that tells the brain carbohydrate has been ingested, so it can initiate the process of glucose disposal and metabolism.
eulerandothers said…
So, does the 'chew chew' diet (where each mouthful must be chewed 32 times) succeed because of the amylase in saliva, or the chewer becoming bored with chewing?
carbsane said…
Hi Stephan, nice to see you!

While it's not the strongest evidence against TWICHOO, it certainly goes against v. X.0 that was put forth in Why We Get Fat at least three times:

You’ll start secreting insulin (from the pancreas) even before you start eating—indeed, it’s stimulated just by thinking about eating. This is a Pavlovian response. It will happen without any conscious thought. In effect, this insulin is preparing your body for the meal you’re about to eat. When you take your first bites, more insulin will be secreted . And as the glucose from the meal begins flooding the circulation, still more is secreted. (Kindle Locations 1668-1671)

Here’s the chain of events:

1. You think about eating a meal containing carbohydrates.
2. You begin secreting insulin.
3. The insulin signals the fat cells to shut down the release of fatty acids (by inhibiting HSL) and take up more fatty acids (via LPL) from the circulation.
4. You start to get hungry, or hungrier.
5. You begin eating.
6. You secrete more insulin.
7. The carbohydrates are digested and enter the circulation as glucose, causing blood sugar levels to rise.
8. You secrete still more insulin.
9. Fat from the diet is stored as triglycerides in the fat cells, as are some of the carbohydrates that are converted into fat in the liver.
10. The fat cells get fatter, and so do you.

11. The fat stays in the fat cells until the insulin level drops.
(Kindle Locations 1784-1794)

Even before we begin eating, insulin works to increase our feeling of hunger. Remember, we begin secreting insulin just by thinking about eating (and particularly eating carbohydrate -rich foods and sweets), and this insulin secretion then increases within seconds of taking our first bite. It happens even before we begin to digest the meal, and before any glucose appears in the bloodstream. This insulin serves to prepare our bodies for the upcoming flood of glucose by storing away other nutrients in the circulation—particularly fatty acids. So our experience of hunger actually increases just by thinking about eating, and then it increases further with the first few bites we take. (Kindle Locations 2059-64)

So, by Taubes' assertions, *if anything*, the higher AMY1 should be fattening. I will check out that reference. I did come across two that were intriguing but I've yet to explore in depth:


These results suggest that after adjusting for BMI, low serum amylase is associated with decreased basal insulin levels and insulin secretion, as well as high insulin resistance. The nature of these associations remains to be elucidated in further studies.

2. ... maybe the sensation of viscosity reduces intake?

Interesting for sure.
carbsane said…
This AHS12 presentation was good.
carbsane said…
Probably more due to eating slower than anything else. Chewing would, if anything, enhance calorie absorption just as processing does.
Wuchtamsel said…
I'll never know, just would love to punch his face every time I see him...
He somehow reminds me of those annoying mormon missionaries. They don't only show the same dishonest salesman-smile but also a similar level of severe cognitive dissonance.

Yes, that's ad hominem to some degree! But somebody who is "interpreting" research in such a decieving way a s Masterjohn does really deserves any amount of ad hominem attacks in my opinion.
Stephan Guyenet said…
Hi Evelyn,

I would think the hypothesis would be most concerned with total insulin release rather than cephalic phase. I don't know what all the studies say, but the one I referenced above indicated similar total postprandial insulin release between low and high AMY1 groups. The study you linked to seems like it's more relevant to pancreatic amylase than salivary amylase.

I've always thought it was interesting when Taubes laid out his hypothesis (actually, stated more as fact than hypothesis) that cephalic-phase insulin makes us hungry by clearing fatty acids out of the bloodstream in response to seeing carbs. Is there any evidence for the following assumptions that underlie this hypothesis?

1) Cephalic phase insulin is released specifically when you're about to eat carbs, but not protein or fat. Has this been demonstrated? Not to my knowledge. If it has, I'd love to see the study.

2) Cephalic phase insulin has a significant impact on circulating fatty acid levels. CP insulin release is so small I wouldn't necessarily expect it to affect FFA levels. It hardly affects glucose levels-- dropping them by only a few mg/dL, an amount that has no impact on hunger. I'm not aware of data suggesting it reduces FFAs or lipolysis significantly.

3) Cephalic phase insulin release is causally related to the sensation of hunger. The idea here is that energy in the bloodstream drops, making us hungry. I'm quite certain there's no convincing evidence to support this, and it seems unlikely since CP insulin hardly has any effect on blood glucose.

I think Taubes may have been led astray on this issue by people who didn't really know what they were talking about but said things he wanted to hear. There is a lot of confusion on this issue because if you don't dig too deep, the ideas make intuitive sense, and some people (including researchers) have extrapolated far beyond the bounds of the evidence. People like to have concrete answers even where none are available.
Screennamerequired said…
Hey stephen. Are you still denying saturated fat raises TC and LDL?
I admire your persistance, but do you seriously still believe that?
Screennamerequired said…
Listen to his interview with jimmy moore. He starts his conversation somewhat rationally, but then gets caught up in a rant about how polyunsaturated fats are causing heart disease. From then he goes on a long political rant about how the government is objecting cholesterol to deprive us of cholesterol.

Never mind his theories about PUFA's and heart disease are not backed by evidence.
Screennamerequired said…
I must also say, from an observer, it's almost funny that Denise is doing everything she can to pander to the low carbers. I understand it from a financial point of view, considering she has published her book with mark sissons company, but it still surprises me that a girl that got so deceived by the the raw food vegans can do such a 180.
Wuchtamsel said…
OK, an interview with Masterjohn AND Moore is even MORE impossible to watch... if that is possible... ;-)
But yeah, those guys are really excellent when it comes to pulling BS out of their *****.
carbsane said…
Uggh ... my browser keeps crashing ... sorry for delayed response.

It's hard to figure out who did the leading and who followed astray on this. Taubes did devote a fair amount of ink in GCBC to this notion of "internal starvation" and that is part and parcel of Eades', Fat Head/Naughton and Lustig's schticks. As the song goes, you eat carbs, insulin traps fat making those calories unavailable to the rest of your cells, they run out of glucose and you get voraciously hungry and eat more. It's in every Fat Head production and one of Lustig's Skinny on Obesity videos.

However the cephalic thing was new to WWGF. He had to replace the "you can't store fat without carb to make alpha glycerol phosphate" with something.

I haven't looked into cephalic insulin in a long time, but I remember being underwhelmed when I did some years back when this notion was floated in LC circles. So I don't know that there's any "there" there in this instance, I'm playing devil's advocate a bit. IF there is anything to gene copy -> amylase levels -> insulin === obesity, it is diametrically opposed to the claims.

There is no evidence whatsoever of cepahlic insulin clearing the bloodstream of fatty acids in anticipation of a meal. If anything, early insulin (probably later than cephalic) may serve to increase efflux temporarily so as to enhance proper fat uptake to adipocytes. (browse to quote on FIAT)

"People like to have concrete answers even where none are available."

So true! My favorite response to this these days.
Wuchtamsel said…
Sorry, but this is just that - wrong!
You either never took a single look at the literature or got deceived by those that have their share in misreporting the results of published data on that subject, just like Masterjohn for example. I guess Masterjohn e.g. never lost a single word about the nurses health study, that clearly showed the lowest(!) CVD incidence in the high omega 6 folks (HR~0.4).
charles grashow said…

"The authors speculated that the slightly weaker associations in the
20-year analysis may be a consequence of favorable trends in fat
consumption; from 1980 to 1998, mean intakes of total fat, saturated
fat, monounsaturated fat, and trans fat decreased in this cohort by 26%, 40%, 28%, and 27%, respectively, whereas polyunsaturated fat intake increased by 6%."
Hello_I_Love_You said…
CarbSane, that first link you provided is a study of serum amylase rather than amylase copy numbers. However, that is an intriguing study and the results are revealing. You see that the two groups identified as low amylase and high amylase display virtually no disparity in HbA1c, FBG, BP, GFR, HDL, 1H PP BG, Cr, etc. On blood tests given by your PCP or GP, they will pass in flying colors. But if you pore over the low amylase group's results you see deepening insulin resistance: statistically significant higher BMI, significantly higher 2H PP BG of (151 vs. 119), lower fasting and 1H PP insulin but higher 2H PP insulin, and higher leptin.

The distinguishing markers are really 1H PP insulin, and 2H PP BG. If you think about it, this makes sense. The low amylase is probably a marker of IR to the point where insulin response begins to fail, where the insulin response curve starts sloping downward and diverges from the insulin resistance curve, which is flatlining. You're prediabetic and a long way from insulin depletion but the ball gets rolling and your insulin secretion is impaired.

So low fasting insulin doesn't always mean your BG control is hunky dory! It could mean deepening IR, especially when the other markers line up.
Wuchtamsel said…
As long as there ARE no significant RCTs the epidemiological data is the best we have and can't just be ignored!
The article you use as evidence is a great example of "analyzing" completely underpowered joke-trials that wouldn't have been able to prove anything anyhow.
ZM said…
Wait, so you cite an observational study that cannot accurately evaluate specific pufas and on top of it is susceptible to massive confounding, yet call far better controlled trials capable of
evaluating specific pufas as "joke-trials". You are not making any
sense, and should read the arguments put forth by Ramsden et al a
little more carefully.

Also, claiming that trials are
"underpowered" is poor excuse because many diet trials involving omega-3
have indeed found significant benefits whereas the ones involving
omega-6 either found significantly increased risks or indications of
harm. In fact, if you are going to make the "underpowered" argument, the
correct argument would be that one of the omega-6 trials found
significant harm, but the other omega-6 trials found indications of harm
but may have been underpowered to find statistically significantly
increased risks. So when used correctly it actually supports my case.
Sanjeev Sharma said…
that would make sense
Screennamerequired said…
The evidence against polyunsaturated fats in the role of heart disease is stronger than the evidence regarding saturated fat? Seriously?
Screennamerequired said…
The science of oxidised/modified lipoproteins are not as sorted out as the bloggers speculations likes people like masterjohn make.
ZM said…
No one said that it was proven, not even Masterjohn. However, the idea that LDL per se is harmful is not sorted out either, so let's not pretend that it is. The point is that the evidence for the harm of modified lipoproteins (including oxidative modifications) is stronger than the evidence for unmodified lipoproteins.

BTW, Masterjohn is not just a blogger, and his views are in line with many other scientists who believe that oxidative modifications are the major problem.
carbsane said…
Chris is currently working for Kummerow
ZM said…
Yes. In clinical trials, n-6 pufa is associated with increases in risk compared to saturated fat whereas the opposite is found for n-3 pufa.
carbsane said…
I love a good fat fight :-)
Screennamerequired said…
Rarely is anything fully "sorted out" but even Chris admits high cholesterol is a problem. If you really want to play the game of what other scientists in the field believe you'll find that Chris is on the fringe. He seems mainly concerned with defending the nutritional recommendations of the WAPF.
ZM said…
It makes no sense to criticize the role of modified cholesterol while claiming that unmodified cholesterol is harmful, since the former has more evidence behind it than the latter. If Chris believes that high cholesterol is a problem (as in FH), a high total/HDL ratio as a problem, and oxidative modifications as causally involved, then how is that "fringe"?
Michal Piják said…
Personally, I think, that increased risk of obesity in persons with low amylase can be
explained by a reduced responsiveness of the islet B-cells to incretins. Incretins are insulinotropic factors of the gut released by nutrients and stimulating insulin secretion in physiological concentrations in the presence of elevated blood glucose levels. The incretin
effect refers to the phenomenon of oral glucose eliciting a higher insulin response than intravenous glucose at identical plasma glucose profiles. It is conveyed by the two insulinotropic incretin hormones: glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP). Type 2 diabetes mellitus (T2DM) has been shown to be characterised by an almost abolished incretin effect.

The incretin effect was shown to be affected in subjects who had impaired glucose
tolerance and who were therefore at high risk for developing type 2 diabetes. This
observation could imply a primary role for the reduced incretin effect in type
2 diabetes, but on the other hand, the finding could also represent an early
consequence of the chronic mild hyperglycemia of impaired glucose tolerance. However,
recent study suggests that reduced incretin effect in type 2 diabetes is a consequence of the diabetic state rather than a primary event leading to type 2 diabetes. (

Nevertheless, the primary role of reduced increting effect due to low amylase cannot be excluded. This assumption is supported by the fact, that proper assimilation of nutrients
stimulates secretion of GIP and GLP.1. ( Importantly, there is also evidence that genetic variants may influence the incretin effect and possibly also the responsiveness to incretin-based therapy in some patients. (

Though the specific process by which salivary amylase stimulates PIR and affects glucose homeostasis remains unclear, the authors of the recent study offer several possibilities:

One possibility is that the production of glucose and/or maltose through amylolytic activity in the oral cavity signals the body to prepare for incoming starch and the ensuing glucose. The sugars would bind lingual T1R2-T1R3 sweet taste receptors (26) and/or glucose transporters in taste receptor cells (27). Because the amount of glucose produced by salivary amylase is too low to be consciously tasted and maltose is only weakly sweet tasting, the stimulation of these taste receptors would not be expected to activate perceptible sweet taste (28).

Second, the mechanism may also involve binding of short-chain oligosaccharides by the putative polysaccharide receptor, hypothesized to enable identification of starch-rich foods (29).

Finally, it is also possible that hormones orvincretins (e.g., glucagon-like peptide-1) are peripherally released by lingual taste cells into the blood stream in response to carbohydrates, stimulating insulin release from the pancreas during the PIR period. „

Excerpted from :
Shaun Clark said…
I am not convinced of any general relationship at this point although I can tell you that I consume what is considered a very high saturated fat diet, via Coconut Butter and Coconut Oil. When I started I had a baseline as I have kept complete medical records, including tests, starting over 15 years ago after a work accident. About 5 years post recovery I decided I wanted to hike the Appalachian Trail. Which I did. Along the way I was amazed at how much weight I shed and my the astoundingly clear and elevated mental status I was in.

Afterwards the weight crept back and mental status devolved. I am not saying I was depressed, just :( normal feeling. So I decided to find out and replicate what I had experienced and attained on my hike.

The start was abysmal, going through all the snarling diet guru v diet guru publications I was quickly irritated and decided to go out on my own. After about 3 years of playing around with diets, all the while having blood tests done every month (I have a great doc and because of my surgical history it's no problem to have authorization, plus I now utilize the company Theranos for testing. Anyway, I noticed some trends and kept tinkering. I eventually ended up with my own "ketogenic" style diet that I like. I stopped all sugar and subsequently discovered that any cravings for sugar/sweets disappeared. I don't crave bread or grains, again at first I did but like I said it disappeared. The same with large meat intake, large meals, salt....etc. So maybe this whole focus on diet type, which to me is just a money orientated business even though some of the diet gurus start off with good intentions, it eventually devolves into manipulative practice to enable their own product and book marketing.

My Hba1c and lipid profiles plus liver panels just amaze my doctor especially after I told her the specifics of my diet. Just anecdotal I realize but I have no dog in this fight. My story it that I just wanted to become healthy again after my accident and shed the gut that over 15 surgeries had put on me. I also have an intense interest in my much elevated mood and even higher mental motivation following changing my diet. In summary, I shed ~50 lbs in 6 months with only moderate exercise, walking 2.5 miles morning and evening. I am not a Lauric or Capric acid champion and have read the studies and I yawn. I just know what worked for me. Some more an

I think the confounding factors to be too numerous to pigeonhole any one factor in obesity and related disease from diet. I do consider an active lifestyle to be the only "ubiquitous" requirement to health.

Additional anecdotal info: My family, on both sides, do not have cancer and the only disease has been emphysema from smoking. We have huge family reunions with many 80 year olds present. In the past three generations we have had 6 centenarians. I only include those how were still mentally acute and physical mobile. So my very unscientific view is that there are many factors at play and I think that with the increasing databases available for meta analysis and the increase in gene testing, research and understanding we are going to find more answers and even more questions.
angelstarchgr said…
Your post is very nice. useful information. Thanks for sharing.
I have a page on Starch for Food Adhesives
Visit my page here
Food Grade Dextrin Manufacturers
Blogger said…
On EasyHits4U you can earn free advertising credits by viewing other website-ads from a account base of over 1,200,000 accounts. Earn advertising credits fast with a view for view model.
α-Amylase is a protein enzyme EC that hydrolyses alpha bonds of large, alpha-linked polysaccharides, such as starch and glycogen, yielding glucose and maltose. It is the major form of amylase found in Humans and other mammals. It is also present in seeds containing starch as a food reserve, and is secreted by many fungi. α amylase