Paleo Diet Improves Glucose Tolerance and Blood Cholesterol

  A Paleolithic diet improved metabolic status with respect to cardiovascular and carbohydrate physiology, according to a 2009 study at the University of California San Francisco.

Here are the specifics, all statistically significant unless otherwise noted:

  • total cholesterol decreased by 16%
  • LDL cholesterol (“bad cholesterol”) decreased by 22% (no change in HDL)
  • triglycerides decreased by 35%
  • strong trend toward reduced fasting insulin (P=0.07)
  • average diastolic blood pressure down by 3 mmHg (no change in systolic pressure)
  • improved insulin sensitivity and reduced insulin resistance; i.e., improved glucose tolerance


This was a small, preliminary study: only 11 participants (six male, three female, all healthy (non-diabetic), average age 38, average BMI 28, sedentary, mixed Black/Caucasian/Asian).

Baseline diet characteristics were determined by dietitians, then all participants were placed on a paleo diet, starting with a 7-day ramp-up (increasing fiber and potassium gradually), then a 10-day paleo diet.

The paleo diet: meat, fish, poultry, eggs, fruits, vegetables, tree nuts, canola oil [?], mayonnaise [?], and honey.  No dairy legumes, cereals, grains, potatoes.  Caloric intake was adjusted to avoid weight change during the study, and participants were told to remain sedentary.  They ate one meal daily at the research center and were sent home with the other meals and snacks pre-packed.

Compared with baseline diets, the paleo diet reduced salt consumption by half while doubling potassium and magnesium intake.  Baseline diet macronutrient calories were 17% from protein, 44% carbohydrate, 38% fat.  Paleo diet macronutrients were 30% protein, 38% carb, 32% fat.  Fiber content wasn’t reported. 

I’m guessing there were no adverse effects.


This study sounds like fun, easy, basic science: “Hey, let’s do this and see what happens!”

I don’t know a lot about canola oil, but it’s considered one of the healthy oils by folks like Walter Willett.  It sounds nicer than rapeseed oil.

I agree with the investigators that this tiny preliminary study is promising; the paleo diet (aka Stone Age or caveman diet) has potential benefits for prevention and treatment for metabolic syndrome, diabetes, and cardiovascular disease such as heart attack and stroke.

The researchers mentioned their plans to study the paleo diet in patients with type 2 diabetes.  No published results yet.

Are you working with a physician on a medical issue that may improve or resolve with the paleo diet?  Most doctors don’t know much about the paleo diet yet.  You may convince yours to be open-minded by trying the diet on your own volition—not always a safe way to go—and showing her your improved clinical results.  Or show her studies such as this.

I’m considering the paleo diet as a treatment for diabetes.  If interested, follow my progress at my PaleoDiabetic blog.

Steve Parker, M.D.

Reference:  Frassetto, L.A., et al.  Metabolic and physiologic improvements from consuming a paleolithic, hunter-gatherer type dietEuropean Journal of Clinical Nutrition, advance online publication, February 11, 2009.   doi: 10.1038/ejcn.2009.4


Filed under Uncategorized

9 responses to “Paleo Diet Improves Glucose Tolerance and Blood Cholesterol

  1. Hey, excellent info and report.

    This helps point the way for more healthly living for the ole hunter gatherer digestion/gene set in our 24/7 availability of high energy foods for all the other folks who are moving 2 ton stone blocks by hand for the pharoh’s tombs and edifices.

  2. “I’m considering the paleo diet as a treatment for diabetes”

    As you know from your ketogenic Mediterranean diet experiments, you can’t feed diabetics 38% carbs and get a good outcome. Eight percent carbs is more like it. I’m Type I and my HbA1c is 5.1% – normal – on Dr Bernstein’s 6% carb diet. At 15% carbs, my HbA1c was 8%, complications territory.

    The reason is that lowering dietary carbohydrate lowers glycogen stores in the liver which halves the liver’s output of glucose (; this prevents the fasting hyperglycemia which characterizes diabetes. Simple and effective, yet utterly counter to the prevailing dogma of diabetes management.

    Dr Malcolm Kendrick put it this way: “The reality is that over the years we have killed literally millions of diabetics by advising them to eat a high-carbohydrate diet and avoid fats. Only now is it being realized that previous advice was and remains useless, dangerous and scientifically illiterate”

    Don’t blink, Doc, so far you’re riding the tiger – make it a low-carb Paleo diet

    • Hello, Jonathan.
      Dr. Bernstein’s got a great program, no doubt. I’ve got his latest book edition on my desk now, waiting to be digested.
      Paleo diets range in carb content (as percentage of energy) from 20 to 45%. I agree completely that that’s too many carbs for most people with diabetes. Paleo carbs are generally high-fiber and low-glycemic index, however, which should help with glycemic control. Eaton and Konner have estimated that many ancestral diets had over 75 g of fiber daily! Modern Western humans eat about 15 g of fiber/day.


      • “Paleo carbs are generally high-fiber and low-glycemic index, however, which should help with glycemic control”

        For a diabetic, it’s not so much how much fiber is in the diet, it’s how much starch – more than the threshold of ketosis and the liver’s supply of glycogen is sufficient for the liver’s output of glucose to cause hyperglycemia. Each diabetic’s starch tolerance is the amount which keeps him or her in ketosis – any more and he or she is sliding down the razor blade of diabetic complications

        But this is not to say that the Glycemic Index is useless – I rely on it absolutely to pick my vegetables: I don’t eat any with a GI – no starch, no GI.

        You make me nervous when you talk about these ideas du jour like fiber and glycemic index – so politically correct yet so utterly ineffective in any but epidemiological studies. Diabetics aren’t dying of fiber deficiencies

  3. hii .enjoyed the report myslef and jim is right it does help point you in the right direction to a more healthly lifestyle .

  4. I agree that glycemic index has a lot of “issues.” A major one is that the index is determined by testing response to isolated ingestion of that one item. Most of us mix different foods together in meals, which will usually alter the glycemic effect of an individual food. Usually lowering its GI, I bet.

    Pure sugar has a GI, and is not a starch.

    My understanding is that dietary fiber tends to slow digestive processes. If true, it may reduce the blood sugar response to ingested fermentable carbohydrates. For instance, brown rice has a GI of 55; white rice is 64.


    • I’m not so sure, doc! The experiment has actually been done:

      When Dr Gerald Reaven tested mixed meals, he found that “The results indicated that the plasma glucose responses after the meals did not vary as a function of their glycemic potency in either the normal or NIDDM subjects. There were no significant differences in the plasma insulin responses for either group. These results indicate that the plasma glucose response to mixed meals did not vary as a function of the calculated glycemic potencies. Therefore, the glycemic response to a mixed meal was not predicted on the basis of the published values of the glycemic index of the individual carbohydrate foods included in the meal”

      Oatmeal raised my blood sugar explosively in spite of a lowish glycemic load and great quantity of soluble fiber. No one has demonstrated that fiber or eating according to the glycemic index does much for diabetic control, but Dr Bernstein has demonstrated that carbohydrate restriction to below the threshold of ketosis lowered the average HbA1c of 30 diabetics to 5.7%, whihc is within the normal range

      The idea that there are “good carbohydrates” for diabetics is a lethal anachronism.

      • Thanks for that link to the 1987 article, Jonathan. A quick reading of the abstract suggests, to me, that the glycemic index may be pretty worthless! Of course, I don’t have all the details, such as the size of the study, contrary results from other studies, composition of the three test meals, etc.


  5. For those not familiar with glycemic index, here’s an info page at