Insulin resistance is a spectrum
The history of low fat diets is riddled with crappy low fat food-like products.

Food quality matters.

Take a group of obese people and assess insulin sensitivity however you like: some researchers demand nothing less than a hyperinsulinemic-euglycemic clamp (Gold Standard), others are OK with insulin levels during an oral glucose tolerance test.

Next, divide the people up based on this — there are a few ways you  can do it.  You can: take the top half vs. the bottom half (a method  which includes everyone); take the top third vs. bottom third (excluding  the middle third); take the top quarter vs. bottom quarter (excluding  the middle 50%), etc.

THIS MATTERS because in referencing this topic, many people claim most obese are insulin resistant. They may be more insulin resistant than lean people, but even within obese people,  there’s a spectrum, and the spectrum matters in this #context.

In Gardner’s recent study, he chose the first path, ie, included all patients, and divided them in half based on insulin levels during an OGTT.

Nuance #1. half of their population had “insulin resistance,” by definition. Not: “most” of the overweight/obese; 

“half,” by definition.

They showed this half lost about 20% more if assigned to low carb (LC) than if they were assigned to low fat (LF); and the other half lost about 20% more if assigned to LF than LC.

This didn’t reach statistical significance:

“Small sample size.
There will be outliers.
Don’t overthink it.”

However, some markers of health significantly improved in people  assigned to their insulin-appropriate diet (LC for IR, LF for IS). And a  bigger, more statistically-powered follow-up study is underway, so we’ll have a clearer picture how this particular intervention pans out on a larger scale in the not-too-distant future.

Nuance #2. Gardner’s study was ad lib. Ie, no calorie restrictions in either group. Both diets were comprised of healthy whole foods. The history of low fat diets is riddled with crappy low fat food-like products. Replace the fat-free junk foods with healthy whole foods, people spontaneously eat less and lose weight (for whatever reason), regardless of whether insulin resistant or insulin sensitive.

Moving on. Now, you might be asking yourself: “how would this have  differed if they used a stricter definition of insulin resistance, eg, top quarter vs. bottom quarter?”

Insulin resistance is a spectrum

This was addressed by Cornier et al., albeit in calorie-controlled conditions, and it seems to have laser-focused on the target:

Insulin sensitive people lost ~80% more weight when assigned to LF,  whereas insulin resistant people lost ~50% more when assigned to LC  (statistical significance achieved).

Pittas et al. confirmed this with three [obvious] caveats: 1) similar to Gardner,  they included everyone and just divided them down the middle of insulin sensitivity; 2) they used low-glycemic load instead of LC per se; and 3) similar to Cornier, they controlled the calories.

Despite these differences, they found almost the exact same thing:

Lastly, we have Ebbeling et al., who: similar to Gardner, was ad lib (no calorie restrictions in either group); similar to Gardner and Pittas, included everyone and just divided them in half by insulin sensitivity; and similar to Pittas, used low glycemic-load instead of LC per se.

Despite these differences, they found almost the exact same thing.


1) choosing the insulin-appropriate diet seems to work better if you’re in the top vs. bottom quartiles, but it still works even if your just slightly more or less insulin sensitive than the next cat (think: good sleeper? sedentary or active? etc.)… and my personal theory is there are MANY more biological predictors specific for a wide variety of other interventions (eg, LIGHT therapy, high or low protein, SFA or PUFA, exercise, etc., etc.)… unknown unknowns?

2) it doesn’t matter if you’re a calorie-counter or prefer ad lib. Imo, ad lib studies are more valuable because I personally think calorie counting sucks on a number of levels: it’s difficult and takes a lot of experience to get it right; also, it’s tedious.

Lastly, please note everyone lost weight and got healthier in these studies, because, at least in part, the interventions stressed real whole foods. The LF diets included vegetables, legumes, and lean meats, and excluded fat-free food-like products, sweets, and added fats. The LC/LGI diets excluded junk foods and refined grains  (this is basically just the default in LC/keto/Paleo), and included nonstarchy vegetables, legumes, temperate fruits, and fattier cuts of  meat.

Criticisms and rebuttals:

Some might say everyone lost weight because their  diets simply got better. Perhaps, but that doesn’t explain why in these four different studies, there was a high level of consistency among the different effects of LC/LF diets in IR/IS populations.

Seriously though, fasting insulin levels declined in IS people  on a high carb diet but didn’t budge on a low carb diet. READ THAT SENTENCE AGAIN.  In IR people, insulin levels declined in both groups. These people are different on a biological level.

Some (eg, Kraft) might say almost everyone overweight is insulin resistant, but INSULIN RESISTANCE IS A SPECTRUM. In some studies, this was defined as the top 25% of the insulin resistance spectrum, in others, it meant half of all overweight people.

Some might say none of these studies tested true low carb diets.

Rebuttal #1: hypoenergetic low(ish) carb diets are probably closer to ketogenic than you think.

Rebuttal #2: Carbs: Low vs. Lower

Funded by Big Avocado  ??????

Low Fat Diet = starvation


The pro-keto argument: many LC vs. LF diet studies show modest benefit of LC overall in obese patients. This is simply a fact. But ten bucks says that if they did a post-hoc subgroup analysis on IS vs. IR patients at baseline, the results would support this whole line of thinking.  Maybe.

Did I miss anything?


That's all for now!

If you’re interested in setting up consultations, email me: [email protected].

Check out my Patreon campaign! Five bucks a month for full access and you can cancel if it sucks :-)

Affiliate links: Still looking for a pair of hot blue blockers? TrueDark is offering 10% off HERE and Spectra479 is offering 15% off HERE. If you have no idea what I’m talking about, read this then this.

Join Binance and get some cryptoassets or download Honeyminer and get some Bitcoins for free

20% off some delish stocks and broths from Kettle and Fire HERE

If you want the benefits of  ‘shrooms but don’t like eating them, Real Mushrooms makes great extracts. 10% off with coupon code LAGAKOS. I recommend Lion’s Mane for the brain and Reishi for everything else

Join with this link.

Start your OWN Patreon campaign!

calories proper

By becoming a patron, you'll instantly unlock access to 81 exclusive posts
By becoming a patron, you'll instantly unlock access to 81 exclusive posts