Keto vs. Mediterranean diet: which is better for blood sugar control?

Foods high in animal protein

A new study from Drs. Christopher Gardner, Lucia Aronica, and colleagues concludes that a whole foods keto diet and “Mediterranean Plus” diet both equally lower blood sugar in people with diabetes and prediabetes. However, they also caution that the keto diet “dangerously” raises LDL and may lead to nutrient deficiencies. 

The study was recently published in the American Journal of Clinical Nutrition, but the authors’  conclusions may not hold up upon closer inspection. 

The researchers randomly assigned 33 adults with type 2 diabetes or prediabetes to 12 weeks  on one of two diets. The first was a whole foods keto diet — with daily 20 to 50 grams of carbs, 1.5 grams/kilo of protein, at least three servings of non-starchy vegetables, and the remainder from fat. The second diet was what they called a “Mediterranean Plus” diet based on the Mediterranean pyramid but with no added sugar or refined grains. Both diets discouraged processed foods and encouraged whole foods. After the 12 weeks, the 33 subjects crossed over to the other diet. 

Interestingly, both groups ate about 300 calories less than their baseline diet. So, both diets proved to be higher-satiety diets (meaning participants felt fuller for fewer calories).

The authors reported that the primary outcome, HbA1c (a measure of blood sugar), was not significantly different between the groups, decreasing by 9% in the keto diet and 7% in the Med Plus diet. 

However, those on the keto diet successfully reduced their diabetes medications more than those on the Med Plus diet. Therefore, the keto diet showed better HbA1c reduction when accounting for the difference in medications. In addition, continuous glucose monitoring (CGM) data revealed that the keto diet led to a greater reduction in average glucose and more time spent in the normal blood glucose range than the Med Plus diet.

One conclusion is clear: both diets were improvements from the baseline diet. They both led to improved blood sugar, weight loss, and satiety.

But what about the concerns? The authors reported that LDL “dangerously” rose 10% on the keto diet. But was it really a dangerous change? Triglycerides went down on the keto diet, as we would expect. And as we saw in 2018 with the Virta Health trial, on average, LDL went up 10%. However, the calculated cardiac risk score went down 12%. 

How could that be? In the Virta study, other blood lipid markers improved: Apo B (a better measurement than LDL for predicting heart risk) didn’t change, and small LDL particles decreased, VLDL decreased, and the TG:HDL ratio decreased. 

The current trial from Gardner and colleagues didn’t report Apo B, small LDL, or VLDL. Therefore, they assume any rise in LDL must be dangerous. However, given the Virta Health data, at a minimum, we should likely conclude that we need more data before assuming harm. 

The other area of concern, the authors state, is the decreased intake of thiamine, folate, and vitamin C. But consuming less doesn’t mean being deficient long term. And it’s unclear if this has any clinical implication. So, again, the conclusion should likely be that we need more data rather than an assumption of harm.

So, what can we conclude from the study?

  • Both keto and Med Plus improve blood sugar, with keto being slightly better
  • Both keto and Med Plus help reduce caloric intake and are therefore higher-satiety diets

Sounds like a good place to start!


  1. 36 comments removed
  2. Heather Vaughan
    Good to know!
  3. Sam Phel
    Until the 1940s, the Tibetans were eating a heavily meat-based diet, a Tibetan Nomadic Diet, and people were living into their 90’s and 100s, lean and strong. Memory loss was almost unheard-of unless owing to severe head injuries.
    With exposure to the contemporary way of eating poor quality foods, almost every Tibetan household now has a member that's either hypertensive, diabetic, or obese and dying early with multiple organ failures.
    The low-Carb way of eating doesn't make you immortal, but it's the most compatible with the human body and its nutritional requirements.
    Thank you DIET DOCTOR!
  4. Helena Bergström MD
    Wonderful News!
  5. Karl
    It seems like a well-designed study with many important outcomes, although some of the concerns raised are, as you correctly say, not based on any data. It is a great pity that the no sizing of LDL-C particles were done - they even mention the fact that other studies show a benefit of low-carb diets on LDL:

    " Increases in LDL-C and decreases in TG are common effects of low-carb diets, and have been linked with a shift in LDL-C particle size associated with decreased cardiovascular risk(45-48). However, since we did not fractionate LDL by size, we cannot dismiss the potential harms of the observed LDL-C increase on WFKD".

    This conclusion deserves criticism - surely they can't appropriate harm to LDL-C when other studies have shown benefit? I.e. why are they assuming harm, when one could equally be assuming benefit? I also agree that the assumption of harm around thiamine, vitamin C and folate is not data-based. It really would be great if the team from Diet Doctor could write a letter to the editor of the American Journal of Clinical Nutrition to raise concern about these specific conclusions - though on the whole I think it is a very well-done study, and it clearly shows the benefit of the ketogenic (and Mediterranean) diets for diabetics and pre-diabetics. The slight apparent bias against a ketogenic diet without clear evidence is a pity!

  6. Anna
    Great to know . Tnx
  7. Alina
    Thank you for the good news!
  8. HU Bar
    How nice that there can be a scientific bickering within the Mediterranean and Low Carb communities while millions are continuing down nutritional suicidal pathways. “My car is cleaner than your car” could be a self serving, “look at me” argument versus promoting just washing your car. Adequacy is sufficient?

    Either approach can drastically improve the lives of any who have the courage to accept the change.

    That stated, please continue sharing studies and simplified interpretations of the results.

  9. Luis
    Interesting article but still showing biased views. It is clear from the Virta studies that both triglycerides and VLDL decrease despite an increase in total LDL. In my personal experience with ketogenic lifestyle that observation holds true as well as a significant deprescription of medication. I am in my third year of Keto, and my HbA1c holds below 5.7% and raking only Metformin. Both keto and Mediterranean approaches result in both caloric and carbs restriction which improve glycemic control. It is not surprising that glycemic control be more robust with further carbohydrate restriction, one crucial factor in glycemic homeostasis which rests on four legs: liver glucose production (neoglucogenedis and glycogenolysis), periferal insulin resistance (muscle and adeocytes), impaired secretion or altered functionality of insuline, altered secretion or functionality of glucagon.
  10. Steph
    Would love to know more specifics on the diets. I would also like to see this study done on only pre-diabetics with no diabetic medications.
  11. Karen Oyama MD
    Thank you for your review. I am continually bothered by the use of LDL since most studies and physicians use the calculated LDL which is nearly meaningless. I’m very glad you pointed that out and highlighted the other cardiovascular risk factor improvements.
  12. Rob Oh, MD
    thanks, good review. agree with ApoB measurements and reduction in med use as an argument for keto vs. med plus. however, this provides an intriguing use of Mediterranean!

Leave a reply

Reply to comment #0 by

Older posts