Thursday, September 29, 2016

Flow Mediated Dilation: What does it mean?

Flow mediated dilation (FMD) is one of the last bastions of low fat dogma. FMD is particularly interesting as you appear to be able to prove almost anything with it. I suspect that results in general are very dependent on exactly how you set up a given experiment and how you report the numbers but most of the papers I've read take the background physiology as given and just supply you with the percentage change. Frequently FMD is enhanced under low fat weight loss conditions and obtunded under low carbohydrate weight loss or even after a single high fat meal.

Let's take a fairly typical study such as this one:

Benefit of Low-Fat Over Low-Carbohydrate Diet on Endothelial Health in Obesity

It uses the term "Atkins" and was published in 2008, a time when ketogenic dieting was less acceptable than nowadays. It's nice because the food intake was tightly controlled and the low carbohydrate group was in ketosis throughout. And it also gives me lots of numbers to play with from the results section.

Table 2 of the results is the most interesting. From the bottom line we can see that there is nothing wrong with the ability of the brachial artery to dilate, if you supply nitroglycerin as a nitric oxide source then there is no difference between groups at any time point for the ability of the artery to respond. What changes is the willingness (or ability/need) of the endothelium to generate the nitric oxide in situ to dilate the artery as flow increases in the low carbohydrate group compared to the low fat group. This shows in line three.

Nothing throughout Table 2 is statistically significant except for line three, that ability of the brachial artery to dilate after occlusion. It's a ratio of before and after occlusion, expressed as a percentage. This went down a little after six weeks ketogenic eating and up a little under low fat eating, giving p as less than 0.05, which means ketogenic eating is Bad For You. That's where the paper stopped. I carried on because the numbers making up the ratio are provided and they are interesting.

So next place to look is line two. This is the maximum diameter of the artery after occlusion is released. There is basically no difference in this parameter with time in either group. A little up and down but basically static.

Then there is line one. Line one shows the basal diameter of the artery in the LC group increasing linearly from 3.42mm to 3.58mm over time. The LF group does an up and a down, ending up dropping from 3.94mm to 3.85mm over the study.

[Addendum, see final conclusions: Does the upping and downing of the LF group reflect changes in FFA release with weight loss, high at 2 weeks and low at 6 weeks as the weight loss bottoms out for the LF group? We don't have enough numbers for weight change or FFA levels to check this, but it fits my ideas]

It is the behaviour of the resting diameter which differs between the groups on a background of a fairly static peak diameter. ie the peak diameter doesn't change much but LC/ketogenic eating progressively increases the basal diameter of the brachial artery. The artery doesn't need to dilate as much after 6 weeks of ketogenic eating to get to the maximal diameter needed to supply the post-occlusion hyperaemia of the forearm.

Conclusion: LC arteries are dilated at rest, they don't need to dilate much more to reach the "normal" size of a post occlusion hyperaemia distended artery.

You could leave it at that.

But I can't.

We have a set of diameters and a set of measured blood velocities. So I can crudely estimate the blood flow, aka the oxygen delivery. That's πr2 times the mean speed of the blood, not forgetting to convert the cm/s velocity to mm/s.

Under basal conditions the blood flow increases for the LC group over the six weeks, from 3326mm3/s to 3876mm3/s, up by 17%.

Basal blood flow for the low fat group decreased from 3939mm3/s to 3668mm3/s over the same period, that's down by 7%.

I have no idea whether these changes are statistically significant but they might well be clinically significant. If I had occlusive peripheral vascular disease I know what I would eat. But you knew I'd say that anyway!

Post occlusion hyperaemic blood flow for the LC group started at 6542mm3/s and increased to 7760mm3/s by six weeks, up by 19%.

For the low fat group the start was 8587mm3/s and this increased to 8752mm3/s, up by, err, 2%.

The gain for ketogenic eating, in peak hyperaemic blood flow, is ten times that of the low fat group. Both ended up about the same, but the LC group had a lower initial value. What would have happend on a level playing field at entry point? That we'll never know because it's probably set by the physical size of the forearms of the subjects...

Conclusion: As a percentage of entry point, the maximal post occlusion flow for LC shows a much greater increase over the six weeks of the study.

I could leave it at that.

But I can't.

Ketones work some magic on the redox spans of the ETC and on the energy yield of ATP hydrolysis. The ketogenic dieters were genuinely positive for urinary acetoacetate every day. So they actually need less oxygen, yet they had a greater oxygen supply to their forearm... Of course, we don't know the mixed venous oxygen tension but we've all read D'Agostino's rats and know it's probably high under ketones, if the arterial oxygen tension is anything to go by.

It's pretty obvious that these ketogenic folks were uncoupling. They ate 100kcal more per day for 6 weeks than the LF group and lost more weight, more fat and less lean mass. Hard to believe I know, but I tend to believe the results in this paper because the numbers trash low fat eating again and again, while the authors are clearly lipophobes.

Uncoupling may be wasteful of oxygen but is advantageous for protecting our mitochondria, plus the excessive use of oxygen is offset by the oxygen sparing effect of ketones. With the onset of limited oxygen availability the ATP levels will drop and so uncoupling will stop, while ketones will continue to be oxygen sparing. Ketosis is a pretty good state to survive an occlusive episode.

No wonder ketogenic eating limits the need for vasodilation after brachial artery occlusion...

Conclusion: Ketogenic eating is the Bee's Knees for surviving ischaemic episodes.

That's pretty well it for this post. I'm about to stop, except to speculate as to why limited FMD is a good predictor of poor cardiovascular health.

Fat. Low FMD is a marker of using fat as a major metabolic fuel. Under fat oxidation your brachial artery will already be dilated so it won't dilate much more after an occlusion episode. If you are eating a diet of total crap based on carbohydrate it is quite possible to have elevated FFAs combined with elevated blood glucose, once your adipocytes have been stuffed so full that they won't take any more and leak FFAs despite elevated glucose and insulin. Innapropriately elevated FFAs is the disaster recipe, a hallmark of metabolic syndrome.

Elevated free fatty acids (and their oxidation) is Bad For You under metabolic syndrome but it is completely normal under ketogenic eating.

My overall conclusion: Reduced FMD is an epiphenomenon cause by fat oxidation. Fine when it's appropriate. Bad if you are overweight, insulin resistant and eating sugar.



James said...

Hi Peter,

Love it! Thanks again for your detailed and entertaining analysis!

I'm currently ~33g carbs / day from salads. Protein is set at 0.82 g/lb as per the findings of Bayesian BodyBuilding for optimal protein synthesis (I lift heavy things regularly in the gym).

Have never felt better and love what I'm eating (salads, tomatoes, cheese, beef, eggs, salmon, high-fat yoghurt, mascarpone, cashews, almonds).

GK said...

Ah. I had hoped you might comment on this study, having been aware of it for many years, just not knowing what to make of it.

NY said...

Elevated blood glucose and FFAs are bad. What about physiologic insulin resistance, even eating a small potato might raise blood glucose up to 8 for at least an hour or so. Is this bad then? No chips in beef dripping? Better to watch high glycemic carbs in addition to the total quantity in LC eating?

Peter said...

Gary, I've been thinking about FMD for many years but hadn't happened to come across enough numbers to try and work out what was going on.... This was a nice find.

NY, well, that's an interesting question and gets in to the realm of how low in carb is good, how low is optimal and what might the physiology be doing. I keep thinking that in VLC the shut down of insulin production is quite deliberate, which means that evolution favours hypoinsulinaemic hyperglycaemia for a day or two after digging up some roots when the mammoth meat ran out. Is this a problem? How serious is hyperglycaemia in the absence of hyperinsulianaemia? Does hyperglycaemia matter if ketones are inhibiting glycolysis? Is there any end to the number of questions we can think up?

Also, what may be physiologically perfect post prandially may be a problem when the post prandial period is continuous. Also what is ideal short term vs long term. Also.......


rheum101 said...

Hi Peter
Enjoyed this post (as with others!). Do you mind if I make some slides of your analysis of this paper. I am a Consultant Rheumatologist preparing a talk for my colleagues called "The diet that will save the NHS!" hifalutin but I think necessary. I will, of course make the appropriate citation.

Peter said...

Hi rheum, by all means. I'd appreciate if you checked the numbers as you went along, I'd used pi x d2 rather than r2 on first calculation....... There is some drift in the paper's terminology between basal, mean, peak and hyperaemic but I think I pulled all the correct numbers for the logic and using mean velocity for flow calculations etc.


JohnN said...

Taking another step from your calculation shows that the LC group's blood pressure decreases after 6 weeks while that of the LF group does the opposite (*).
Too bad that the study excludes the majority of the ketogenic dieters, those with total cholesterol > 200 mg/dL.
Thanks for a great debunking.
*: I simply compare the ratio of cross-section area increase to blood flow.

erdoke said...

FYI, FMD changes after different basal insulin treatment protocols.

Effect of basal insulin therapy on vascular endothelial
function and adipokine profiles in people with Type 2

Peter said...



R Cobb said...

Recently stumbled on carb backloading and wondering how this might relate if at all. My understanding is that you eat very low carb for 1 week and then once a week or so in the evening you eat a large amount of carbs for 6 hours or so -most beneficial post work out. I guess folks do this daily as well- very low carb all day and in the evening post workout carb load. According to DH Kiefer shifting carb ingestion to the evening when insulin/cortisol are low after very low carb all day or week helps you shunt available calories into muscle development bypassing fat storage and apparently helps with leptin and ghrelin responsiveness (something that down regulates with extreme ketogenic diet)

I tried it but I might be to far down the ketogenic path for it to be of much use- I just don't have a taste for sugar or carbs anymore. But I am interested in lifting weights like I used to and hoping this might be the trick. But wondering how this might intersect with your post about hyperglycemia etc. post tuber ingestion. My take is that if you are 30g or less of carb per day that your body resets itself relatively quickly after carb ingestion. If not relevant- disregard. thx R

altavista said...

Peter, have you done the EBT (sp?) test again? Interested to see evolution from the previous one.

Anonymous said...

Thanks Peter

By the way, you are going to love the latest anti-fat / anti -'Paleo' findings from the Charles Perkins Institute at Sydney Uni...:


Puddleg said...

I read an interesting explanation of this somewhere.
Insulin (in insulin sensitive) promotes vasodilation to cope with increased blood glucose, i.e. to both dilute it, and get it where it needs to go.
In the same way that insulin promotes fluid and sodium retention and increases extracellular fluid volume; it's about getting the carrier mix right for different fuel mixes. Glucose is water soluble, glycation is risk at higher concentrations, so volume expands to accommodate it, and insulin-stimulated eNOS is part of this adjustment.

raphi said...

Funny how many ways data can be interpreted when you're not selling statins !

marksuave25 said...

Peter, great always.

marksuave25 said...

Peter, great always.

Peter said...

Hi Peter!

I wanted to get this information into your hands....I have been seeing this trend for years in the fat adapted athletes I work with including myself but as a "Run & Gun & Get 'er done" kind of guy never collected data....knew is was part of the process and not only benign but what the body is suppose to do....Dave Feldman and I connected in March and he has been obsessively collecting data on himself and thinking about this and so we connected as I was one of the few who "got it" .....take a read Here:

or listen to the podcast here:

Feel free to email me at

Stan Bleszynski said...

Hi Peter,
Different subject - did you look into the mechanism behind metformin? I came across a study (2) where diabetics on metformin had lower risk of dying, including heart disease, than a control group of non-diabetics, not using metformin. What is suprising, is that there are more studies (1) where the effect of metformin seems to mimick the effects of the ketogenic diet, including all the health benefits such as lower CVD risk, better endothelial reaction, better immune resistance, reducing metabolic syndrome, hyperglycemia, hyperinsulinemia, PCOS, helping with congestive heart failure, lower cancer risk and a few other benefits. There is mention of stimulation of AMPK activity and switching over from mitochondrial respiratory cycle Complex I to II. However, in spite of my gut feeling that there should be a connection, I was not able to identify any biochemical link with ketone metabolism. Any thoughts?
Stan (Heretic)



1. "Cellular and molecular mechanisms of metformin: an overview"

2. "Can people with type 2 diabetes live longer than those without? A comparison of mortality in people initiated with metformin or sulphonylurea monotherapy and matched, non-diabetic controls."


Commercial Loans.
Personal Loans.
Business Loans.
Investments Loans.
Development Loans.
Acquisition Loans .
Construction loans.
Business Loans And many More:

Contact Us At :

Unknown said...


nice analysis, thank you.
However - I wonder if it's not an invitation to take risks. Surely ketosis switches the body to a different mode, and economizes the organism to deal more efficiently with the fat.
So far so good.

But still lower FMD is an independent factor for cardiovascular problems (

Perhaps ketosis makes you bulletproof, nevertheless.

But we do not really know, do we? We do not have the larger data - so takin the keto step (in the light of FMD) is taking a step into uncertainty.

Puddleg said...

I don't know. for one thing, the risk reduction associated with lower 2-hour insulin or lower TG/HDL (which we see with low carb in the original study) probably beats the risk reduction from higher FMD in those studies, especially in the bigger, less confounded studies. And they do say
"However, the presence of heterogeneity in study quality, the remaining confounding
factors, and publication bias in the available literature
prevent a definitive evaluation of the additional
predictive value of brachial FMD beyond traditional
cardiovascular risk factors."
Also, most of the studies seem to involve pre-existing disease, medicated populations, and so on. I see chronic heart failure, peripheral vascular disease, stenting, hypertension in the reference list. These are the "convenience cohorts" and the association is strongest in these groups - but the few other groups also look like populations with high rates of medication and MetSyn.