Posted on

Red Light, Green Light: responses to cortisol levels in keto vs. longevity research

How a scientist interprets outcomes often depends on whether she thinks the outcome should be good or bad.
Cortisol levels make a good example.
In the context of low carb, ketogenic diets, the finding of slightly higher cortisol levels have been interpreted as a warning sign. In a recent post on our blog, Zooko and I attempt to explain why the “red light” that Boston Children’s Hospital gave low carb diets is not justified.

Boston Children's Hospital graphic (with our markup in black). Click for the original.

It is interesting to note that while cortisol sends up red flags for ketogenic diets, which mainstream medicine actively disapproves of, it sends up green flags in another context: longevity research.
Time and again I have come across glib statements in longevity papers saying that the beneficial, health-enhancing and lifespan-increasing effects of caloric restriction probably come in part from the moderately increased cortisol that is consistently seen in calorie restricted animals. The intuitiveness of the beneficial effects of cortisol is usually based on cortisol’s known anti-inflammatory action.
Here are just a few such quotes:

  • “The mechanisms responsible for calorie restriction–mediated beneficial effects on primary aging observed in rodents probably involve the metabolic adaptations to restriction itself, including… a modest increase in levels of circulating cortisol, which result in a reduction in systemic inflammation.”
    — Aging, adiposity, and calorie restriction. Fontana L, Klein S. JAMA. 2007 Mar 7;297(9):986-94.
  • “Even short-term DR can attenuate inflammation and affect metabolic and DNA repair pathways. Mechanisms by which DR suppresses peripheral inflammation include the elevation of glucocorticoids, lowering of glucose and activation of PPARs. Although the effects of DR are less understood in the brain, common pathways are emerging that link many normal aging inflammatory processes with age related diseases such as AD, cancer, diabetes and cardiovascular disease.”
    — Anti-inflammatory mechanisms of dietary restriction in slowing aging processes. Morgan TE, Wong AM, Finch CE. Interdiscip Top Gerontol. 2007;35:83-97.
  • “Glucocorticoids are yet another class of hormones that may contribute to the anticarcinogenic action of DR [101, 102]. Total and/or free glucocorticoid levels are increased by DR [103–105]. Glucocorticoids suppress cellular proliferation and enhance apoptosis in a number of cell types, including osteoblasts, lymphocytes and keratinocytes (for reviews, see Weinstein [106], Herold et al. [107] and Budunova et al. [108]). In humans, glucocorticoids are effectively used for treating lymphoid neoplasms [109]. Importantly, adrenalectomy abolishes the protective effect of DR on skin and pulmonary carcinogenesis, while glucocorticoid replacement restores this protection [110–112].
    — Can short-term dietary restriction and fasting have a long-term anticarcinogenic effect? Klebanov S. Interdiscip Top Gerontol. 2007;35:176-92.
  • Another mechanism by which CR may selectively exert it’s anti-inflammatory effects is via enhanced endogenous corticosteroid production (Sabatino et al. 1991). Chronic CR potentiates the diurnal elevation of plasma corticosterone. CR mice and rats have “moderately” but significantly higher daily mean plasma free corticosterone concentration than mice fed “ad libitum” throughout their lifespan…
    It is well known that the hypothalamic–pituitary–adrenal axis and glucocorticoids in particular are essential in limiting and resolving the inflammatory process (Sapolsky et al. 2000). Glucocorticoids have pleiotropic inhibitory effects on the immune system and inflammatory gene expression (Rhen and Cidlowski 2005). In addition, treatment with pharmacological doses of exogenous glucocorticoids has been used to block many inflammatory and autoimmune diseases, including rheumatoid arthritis, systemic lupus erythematosus, Graves’ disease, thyroiditis, glomerulonephritis, multiple sclerosis, and psoriasis.”
    — Neuroendocrine Factors in the Regulation of Inflammation: Excessive Adiposity and Calorie Restriction Luigi Fontana, Exp Gerontol. 2009; 44(1-2): 41–45.

I particularly like that last one, because of the use of the word “enhanced”, which connotes that the thing that has increased is surely a Good Thing.
Now of course, the longevity researchers can’t pick up on a significant difference between the animals that lived longer and those that didn’t and say: Well, the diet might be good for some things, but this difference warrants caution.
They can’t do that because the end outcome is definitive.
It’s true that there is a significant school of thought that says that longevity is a result of hormesis. The hormetic explanation amounts to “what doesn’t kill you makes you stronger.” That way a researcher gets to say that phenomenon X is bad for you, and that things that are bad for you improve your health. It is a way out of a paradox, as expressed nicely here:

“One well-known, exemplified response to stress is the hormonal increase in adrenal corticosterone levels in plasma during aging, where increases in these levels appear to be proportional to the degree of stress. Aged animals appear to have a diminished ability to attenuate the increase, causing the aged to have continually elevated plasma levels of corticosterones. These authors suggested that increased levels of corticosterone in aged rats result in hippocampal neuronal cell death, that is, the stage of exhaustion. However, this scenario in the glucocorticoid cascade hypothesis is obviously not applicable in the case of the CR paradigm, because CR results in an increased life span in spite of chronically elevated diurnal levels of serum corticosterone. This apparent contradiction makes the interrelation of glucocorticoid and aging far more complex than one might want to narrowly define it and needs other mechanistic explanations like stress resistance to resolve the disparity in responses.” — Stress resistance by caloric restriction for longevity. Yu BP, Chung HY. Ann N Y Acad Sci. 2001 Apr;928:39-47. (my emphasis)

As Carol Loffelmann recently said on twitter: Scientists who discover paradoxes should examine their initial assumptions.
So supposing we have a study showing that a group that looks healthier than the other groups in essentially every measure also has higher levels of cortisol.
We can reason, as Ebbeling et al. do, that since higher cortisol is associated with bad health outcomes, the ketogenic diet may be dangerous, despite the other measures.
Or we can reason, as the longevity researchers do (and as Zooko and I did), that since the group is healthier, higher cortisol must be exerting or reflecting a healthy process, and this may present a paradox that we as researchers have to resolve.
Allow me one further point:

The findings of higher cortisol in calorie restricted animals is itself a body of literature of relevance here.
Anyone finding that their intervention moderately elevates cortisol can and should now say: Higher cortisol levels are found in animals whose lifespans have been increased experimentally by dietary intervention, and so this finding in our intervention could be indicative of a longevity-inducing effect.

Posted on

Deeper ketosis without protein restriction

In my last update, I had talked about the trouble I was having staying at the β-hydroxybutyrate level I wanted while eating to satiety.
Not only was eating to hunger driving my ketone levels down, but higher ketone levels were correlating with irritability.
Hunger and irritability are not my style.
Besides, I have the intuition that something as healthy as ketosis should not entail health compromises.
That’s one reason I think the calorie restricted approach to ketogenic dieting, even in cancer, is likely to be wrong.
More on that another day.

I am delighted to report that for a couple of weeks now I have consistently been in the 1.5 — 3.5 mmol/L range without restricting the quantity of my food, or even trying to be careful about not passing the satiety mark.
Though I do, as always, emphasize fat in my foods, I am not limiting or even measuring my protein intake.
My body’s signals are clear and accurate and I don’t agonize over whether this bite would be the line between enough and too much.
There is no longer any correlation, as far as I’ve noticed, between irritability and higher ketone readings.

The trick seems to be exercise.

A few weeks ago, I made a few lifestyle changes at once.
(I don’t always have time for controls!)

  • I started getting up at 5:30 (amended to 5:00 several days later).
  • I gave up all but two small cups of coffee a day.
  • I started going to a weightlifting class at the local gym twice a week for an hour.
    I enjoy the 15 minute walk home.
  • I started running around the block once or twice a day in order to keep up with the 3-year-old riding his strider bike.
    That usually involves a little sprinting as we go the downhill direction, and walking or lightly jogging for the rest of the way.
    Some days I’ve also done a longer distance walk or bike ride.

There could be some effect of lower carbohydrate contribution from cream, or from something disruptive in the coffee itself.
However, I feel fairly confident that the driving factor is the exercise.

Lyle McDonald: Effects of exercise on ketosis

Way back in 2002, I got my hands on a copy of Lyle McDonald’s The Ketogenic Diet.
It was out of print at the time, and was acquired for me magically by Zooko, in honour of our second anniversary.
(Thank you, Sweetheart!)
Back then, it was one of the few resources available for studying ketogenic physiology.
Lyle McDonald’s purpose in writing his book was to promote Cyclic Ketogenic Dieting for bodybuilding, and dispel myths associated with it.
It is fairly technical, and well-referenced, but it does not presuppose detailed knowledge about specific biochemical pathways, so it’s also accessible.

In it he shows that high intensity exercise (weight training or interval training) is a quick route to establishing ketosis, because it uses up glycogen stores.
In the short term, however, high intensity exercise can decrease ketosis by inhibiting free fatty acid (FFA) release into the bloodstream.
He also emphasizes the utility of low-intensity aerobic exercise, both for lowering glycogen and for increasing FFA for the liver to make ketones with.
Low intensity aerobic exercise is very effective in establishing ketosis, but it takes a long time to deplete glycogen that way.

His bottom line recommendation, then, for establishing ketosis quickly, is to do a high intensity workout to deplete glycogen, followed by 10-15 minutes of low intensity aerobics.

This is precisely what I’ve been doing every day!
Lifting and then walking home, sprinting and then a fast walk around the block, or a long distance, low intensity walk or ride, all qualify as efficient ketosis enhancement.

This is working for me without recourse to hunger-inducing protein restriction.

Fat loss without muscle gain necessarily implies a caloric deficit even though caloric deficit does not necessarily result in fat loss.
Analogously, inducing ketosis through exercise and carbohydrate restriction may well be resulting in a naturally lower protein intake for me.
I don’t really care that much.
I’m eating as much as feels good to me of foods that make me well, and it is no longer interfering with my health goal of being in ketosis.

Whether and to what degree this affects my body composition is not yet clear.
My clothes are fitting better.
I guess I ought to buy a scale.

Posted on

Update on Ketosis and Weightlifting

Well, that was a really long month. In seriousness, writing frequently is hard for me, not because I’m not constantly brimming with things to say — trust me, I am — but because I have so many other things going on. In, particular, it turns out that raising three children is a lot of work, even when you are well. I also released some software as part of the effort to finish my long-stalled Master’s degree. My hobby site, The Ketogenic Diet for Health, which we started as a way to make progress on writing a book about ketogenic therapies, gets almost no love at all.

So, let me briefly tell you what I learned from actively deepening ketosis, and reintroducing myself to weightlifting.

Deeper Ketosis

I had decided not to measure some things that could be informative, such as protein or calorie intake, and just to focus on eating only when hungry. I started making a lot of fatty broths, and drinking that if I felt what I perceived to be a habitual desire to eat, rather than hunger. I took smaller portions to make my clean-the-plate tendencies less detrimental.

This was effective for increasing blood ketones, and as those went up, I lost weight and fat, at least according to my Tanita scale. I lost about 6 pounds over the course of 2 months. That doesn’t sound like a lot, but I don’t have a lot of weight to lose, and I wasn’t in deep ketosis all the time.

Then I ran out of measuring strips, and soon after that I found my scale in a mysteriously broken state.

Odd things

Here are some particular things I noticed during the process:

  • Even though I am already keto-adapted to some degree, I found it hard to stay in the zone I was aiming for (1.5 – 2.5 mmol/L) consistently, because some days eating to hunger meant eating a lot of protein, which drove the ketones down. In other words, staying in deep ketosis for long periods was effortful.

This is strange to me, because theoretically, living off my ample fat stores should allow me to feel satiated indefinitely. As long as I am getting enough protein for body maintenance and blood sugar, why would I ever feel hungry? But I do get hungry, and I do get symptoms like brain fog if I go too long without eating, even when I’ve been in deep ketosis for days.

This brings me to a second point,

  • The reason I don’t like to go higher than about 2.5 mmol/L is that I start to get irritable when ketosis gets too high. I figure this must be a blood sugar effect. I haven’t measured blood sugar and ketosis together consistently enough to be sure of a pattern, but other people have reported an inverse relationship, and we found scientific support which we reported here.

Dr. Ede also reported great discomfort in very deep ketosis, although I notice that her blood sugar is not nearly as low as mine gets at similar ketone levels. When my ketosis is above 3mmol/L, my blood sugar gets close to 65mg/dL.

It’s possible that I am simply not sufficiently keto-adapted, but I have suspicions that there is more going on. I’ll get to that in my next post. With any luck that will be sooner than 4 months from now.

Weightlifting

My first idea was to do short sessions every day, rotating parts so that major muscles would get worked a couple of times per week. There were two reasons for this. First, it was very hard for me to get more than 10 minutes of uninterruted time in a day while looking after a 3 year old. Second, sometimes it seems easier to make a habit stick if it is a daily one. Relatedly, missing a day was not devastating.

This worked only okay. I could feel strength gains, but not tremendous ones. I’m not sure why.

Then my schedule changed. Morning preschool sessions started up again after a long winter break, and I found a wonderful person to care for my son on some afternoons, too. I started working out my whole body twice a week, and this seemed to be much more effective. I don’t have a good way of measuring the results. I don’t even have a scale, though scales don’t show recomposition well. Nonetheless, I was feeling stronger, and Zooko seemed to think he could see a difference.

Unfortunately, a couple of weeks ago, as has happened to me many times in the past, a few of life urgencies resulted in missed sessions, and then I lost the habit.

A new scheduling plan

While reading a post by Cal Newport, my favourite productivity blogger, I realized that trying to make every day or even every week the same for the sake of habit building is not necessarily helpful, and in cases like this, it seems harmful.

So my new plan is to take one week at a time. At the beginning of the week I’ll schedule the things I want to get done around all the perpetual idiosyncrasies that make up my life. In other words, instead of planning to work out every Monday and Thursday afternoon, and then falling off the wagon because this week I had to meet the principal on Thursday afternoon, I’m going to be more adaptive.

Ultimately, I need to find ways of operating that promote consistency over a chaotic, ever-changing life situation. I have too much going on to benefit from “bikini bootcamp” style interventions that require me to focus on nothing but getting slim for 10 weeks.
This probably means that it will take me more than 10 weeks to reach my goal, though.

Bottom Lines

Although I could see the beginnings of progress in fat loss from deepening ketosis (which for me amounts to eating less), there were serious sustainability issues that I didn’t anticipate and don’t understand. Some ideas to follow.

Similarly, weightlifting seemed to be having a positive effect, but I need to figure out how to make it happen consistently in the face of constant uncertainty and chaos. This is no different from the struggles I am facing in other arenas of my life, including graduate studies, book writing, and blogging.

Posted on

The proverbial last 10 pounds

Even though I’m happy with the health effects of my diet, and my weight is stable in a healthy bracket, I still consider myself a little overfat.
Having eliminated most plants from my diet, and given that I believe that weight loss arrived at by enduring hunger is likely to reflect loss of lean mass, not fat,
one might think there was little left for me to try, but that’s not quite true.
There are two obvious changes I could make.
Possibly to the detriment of science, but in the interest of speed and because of the plausibility of synergy, I’ve decided to try both at once.

More Ketosis

The first involves deepening my level of ketosis.
Most of the times that I have measured the level of β-hydroxybutyrate in my blood, it has come out surprisingly low, maybe 0.2 to 0.4.
This is much less than the Phinney-and-Volek-recommended nutritional ketosis range that starts at about 0.5 mMol.
According to them, benefits increase until about 3.0 mMol.
Here is the graphic from their book, which I borrowed without permission from Jimmy Moore.
I hope he doesn’t mind.

I really don’t know what the y-axis means, but it’s clear what their recommendations are.

Given that I eat practically no carbohydrate, this low measurement is probably a reflection of my protein intake.
Although the scientific evidence available on the inverse relationship between protein intake and ketosis in keto dieters is distinctly lacking, it appears to be true anyway.
(I intend to write a post about that on my ketogenic science blog, but don’t hold your breath — it’s complex, and I’m a busy woman.)

I have chosen to leave the evidence of the link between my diet and my blood ketones somewhat unscientific, in that I am not tracking my food.
I have tracked my food extensively in the past, and the problem with it is that the kinds of food I prefer to eat are not easy to analyze, because they aren’t uniform.
For example:

  • One piece of ribeye is fattier than another, especially if at this meal I’m eating the cap and at that meal I’m eating the center.
    Moreover, I often buy large, untrimmed pieces and leave the fat on.
    My steaks look more like this,

    than this,

    and so I distrust nutritional databases.

  • How do I know how much of the bacon grease absorbed into my portion of the scrambled eggs, and if there are also pieces of bacon in it, do I need to take them back out and measure them separately or assume I got an amount proportional to what went in?
  • The nutritional composition of homemade broths are anyone’s guess, particularly when there are bits of meat in it.

Because of this complication, when I did that tracking last time, I felt restricted to a subset of foods, and even then the imprecision bothered me.
So I decided not to do it this time.

Nonetheless, I still get to develop my own intuitive sense of what’s going on by taking ketone measurements.
My basic strategy is that when I feel hungry, I take what seem to me small portions of meat, relative to what I had been eating, and eat lots of fat for satiety.
If I start feeling more hunger and fat doesn’t relieve it, I eat a little more meat.
What it lacks in definition, it at least makes up for in simplicity.

I started this on Monday, but the next day I was disappointed to discover that I had carelessly broken my Precision Xtra, and so I couldn’t actually measure blood ketones.
But I found a link to order a free one.
It arrived yesterday, and I am delighted to report that I had a blood ketone measurement of 3.1!
That’s significantly higher than any previous measurement, and I retested later with very similar results, so I don’t think it was an error.

Lifting

The other tack I’m taking is to start lifting weights again.
It’s been maybe a year and a half since the last time I was regularly lifting.

Weight lifting is probably like aerobic exercise, in that by itself it doesn’t lead to fat loss.
For example, take a look at this typical, recent study.
Its abstract points out that the participants who did a combination of aerobic exercise and weight training for 30 minutes, 5 days a week, lost “significantly more” fat than those who did only one of those.
The way the study is written, and the data reported, distract from the fact that these people lost only about 3.3 pounds over 3 months of this intervention.
This is in people who need to lose some 60 pounds of fat to be healthy.
Just imagine yourself 60 pounds fatter than you want to be, committing to 5 days a week of probably commuting to a gym and working out week after week, and coming back 3 months later to find that you have lost 3.3 pounds.
It’s not something I would tout as particularly effective.
Assuming it kept going at that rate (which is unlikely), it would take four and a half years to get to a healthy weight.
Certainly, four and a half years down the road, it would be better to have lost fat at that rate than not to have, but again,
the long-term effects are unlikely to be linear.

The reason for this is that weight composition is a function of your hormonal state, and the effect of the food you eat is by far the dominating factor.

If you combine resistance training with a fat loss diet, though, the effect is to at least preserve lean body mass, and possibly also increase fat loss.
This is true even of non-ketogenic diets.
Unfortunately, there are very few studies that compare ketogenic diets with and without resistance training.
This review reports on two relevant studies.
The first compares what it calls low-fat and low-carb diets with or without resistance training.
Even though the “low-carb” diet was nowhere near ketogenically low (38:30:32 percent of carbohydrate to
protein to fat as opposed to 61:18:26 in the low fat group), the results are striking:

In men the difference between either diet alone and that diet + resistance training meant at least as much fat loss but also a gain in lean mass instead of a loss. Women lost lean mass either way, but lost very little in the lower carb + resistance training situation.

In the same paper, Volek reports the results of a similar study he performed in men only.
This time the low carb condition was actually low carb.
(They say