Paleolithic Diet Improved Metabolic Syndrome in Just Two Weeks

Wish I were here

A Paleolithic-type diet over two weeks improves several heart disease risk factors in folks with metabolic syndrome, according to Netherlands-based researchers.

The investigators wondered if the paleo diet, independent of weight loss, would alter characteristics of the metabolic syndrome. They seem to be the first scientists to do a study like this.

“Metabolic syndrome” may be a new term for you. It’s a collection of clinical features that are associated with increased future risk of type 2 diabetes and atherosclerotic complications such as heart attack and stroke. One in six Americans has metabolic syndrome, including almost one in four of adults. The most common definition of metabolic syndrome (and there are others) is the presence of at least three of the following characteristics:

• high blood pressure (130/85 or higher, or using a high blood pressure medication)
• low HDL cholesterol: under 40 mg/dl (1.03 mmol/l) in a man, under 50 mg/dl (1.28 mmol/l) in a women (or either sex taking a cholesterol-lowering drug)
• triglycerides over 150 mg/dl (1.70 mmol/l) (or taking a triglyceride-lowering drug)
• abdominal fat: waist circumference 40 inches (102 cm) or greater in a man, 35 inches (88 cm) or greater in a woman
• fasting blood glucose over 100 mg/dl (5.55 mmol/l)

These five criteria were identical to the ones used in the study at hand. But the study participants were required to have only two of the five, not three, for unclear reasons. I found no consensus definition elsewhere that would define metabolic syndrome as only two of the five conditions. Study participants ate either a paleo-style diet or a reference/control diet. Those eating the reference diet didn’t quite have the metabolic syndrome since they had a mean (“average”) of 2.7 metabolic syndrome characteristics. The paleo group had 3.7 characteristics.

How Was the Study Done?

Average age of the 34 study participants was 53 and they were generally healthy. None had diabetes, cardiovascular disease, or systolic pressure over 180 mmHg. Smokers were excluded. Mean body mass index was 32 (obese). Only 9 of the 34 subjects were men. Subjects were randomized to either a Paleolithic-type diet (n=18) or a “healthy reference diet based on the guidelines of the Dutch Health Council” (n=14). Efforts were made to keep body weight stable during the two-week study. Participants were nearly all caucasian.

All meals were home-delivered free of charge by a catering service.

The Paleolithic-type diet “…was based on lean meat, fish, fruit, leafy and cruciferous vegetables, root vegetables, eggs and nuts. Dairy products, cereal grains, legumes, refined fats, extra salt and sugar were not part of it.” [I like their version of the paleo diet.] Protein supplied 24% of calories while carbohydrate was 32% and fat 41%.

You can consult the full text of the published article for details of the Dutch Health Council diet. Calories were 17% from protein, 50% from carbohydrate, and 29% from fat. Alcohol isn’t mentioned at all.

Despite randomization, the paleo diet group had more metabolic syndrome characteristics than the reference diet group. For instance, 78% of the paleo group had elevated fasting glucose compared to 44% of the reference group. And 67% of the paleo group had low HDL cholesterol compared to just 13% of the reference group. These glucose and HDL differences were statistically significant. 39% of the paleo had high triglycerides compared to 19% of the others. Furthermore, the paleos’ average body weight was 98 kg (216 lb) compared to 86 kg (189 lb) in the others. The paleo group had 3.7 characteristics of the metabolic syndrome versus 2.7 in the other cohort.

Go John trail at Cave Creek Regional Park in Scottsdale, Arizona

Results

Compared to the reference diet, the paleo-style diet:

• lowered systolic pressure by 9 points and diastolic by 5
• total cholesterol fell by 0.52 mmol/l (20 mg/dl)
• triglycerides fell by 0.89 mmol/l (79 mg/dl)
• HDL cholesterol (good) rose by 0.15 mmol/l (6 mg/dl)
• body weight fell by 1.32 kg (3 lb)
• one metabolic syndrome characteristic resolved

No significant changes were seen in intestinal permeability ( by differential sugar absorption test on urine), salivary cortisol, and inflammation (hsCRP, TNFα).

Fasting plasma insulin and HOMA-IR fell in the paleo group but not the other.[Good news for folks with diabetes or prediabetes.] Yet the authors write, “Regarding glucose intolerance we did not find significant changes in our study.”

Fasting blood glucose for the group as a whole at baseline was about 1o8 mg/dl (6.0 mmol/l). Fasting glucose fell in both groups: 16 mg/dl (0.9 mmol/l) in the paleo group, 6 mg/dl (0.35 mmol/l) in the other. This was not a statistically significant difference between the groups. These numbers are from the text of the report; looking at the tables, I calculate different and less impressive reductions. The falls in fasting glucose from baseline were statistically significant for both diets.

Nearly all the statistical analysis focused on comparing the paleo diet group to the reference diet group.

My Comments

Overall, I’m not very pleased with this study. My biggest problems are 1) the unfortunate randomization that created dissimilar experimental groups,  2) the use of two diet protocols, 3) some of the study participants didn’t even have metabolic syndrome, and 4) as is typical for paleo diet studies, not many experimental subjects were involved.

The randomization led to significant differences in the metabolic syndrome patients in the two diet groups. I’m puzzled why the authors don’t comment on this. It’s a problem with clinical studies involving low numbers of participants. Ideally, you want to apply the two different diets to groups of people that are as similar as possible. These groups weren’t that similar.

The investigators’ main goal was to study whether a paleo-style diet, independent of weight loss, alters characteristics of the metabolic syndrome. Then why introduce another variable, the Dutch Health Council diet? Is it the gold standard for treating metabolic syndrome? Has it even been used to treat metabolic syndrome? The authors don’t tell us. And why not restrict participation to subjects who meet the common international definition of metabolic syndrome (at least three of the five characteristics)? Why not just take all your subjects and switch them from their standard Netherlands diet to the paleo diet? That would increase your statistical power, and would have avoided the randomization mis-match in which some in the reference diet group didn’t even have metabolic syndrome.

Here we’ve got two different experimental groups, and we’re applying a different diet to each group. The results are going to be messy and difficult to interpret. It’s always better if you can alter just one variable.

Since the paleo and reference diet cohorts were so different at baseline, why not make it easy to simply compare the paleo diet group’s “before and after numbers”? Maybe the analysis is there and I’m just not smart enough to see it.

There weren’t enough men in the study to tell us anything about the paleo diet in men with metabolic syndrome.

The statistical analysis was difficult for me to read and understand. There’s a good chance I’ve missed or misinterpreted something.

This paleo diet reduced fasting blood sugar significantly, making me think it may help in management of diabetes and prediabetes.

I estimate that as much as a quarter of the experimental subjects didn’t even have metabolic syndrome, so the study title is a bit of a misnomer.

This paleo diet did result in resolution of one metabolic syndrome characteristic, which is a good thing. So you could say the diet improves metabolic syndrome, even resolves it in some folks if it drops their metabolic syndrome characteristics from three to two. It predominantly helps lower blood pressure and triglycerides, and reduces excess weight modestly. In white women. Compared to the healthy Dutch Health Council diet.

If I had metabolic syndrome, I’d do something about it in hopes of lowering my future risk of diabetes and atherosclerotic complications. Standard physician advice is to lose excess weight and exercise regularly. There’s no consensus on diet yet. I think carbohydrate restriction is important. If the study at hand is reproducible in a larger study population, the paleo diet is a reasonable approach. Dietitian Franziska Spritzler has a great review of nutritional management of metabolic syndrome at her blog. The Mediterranean diet supplemented with nuts helps improve metabolic syndrome. The Spanish Ketogenic Mediterranean Diet may cure metabolic syndrome.

Steve Parker, M.D.

Update: I took a fresh look at this study as if it were simply a paleo diet trial involving 18 subjects who had metabolic syndrome. If I’m interpreting Table 5 correctly, and I think I am, these are the statistically significant changes after two weeks:

• abdominal circumference decreased by 3.1 cm
• systolic and diastolic blood pressures dropped by 8.5 and 8, respectively
• fasting glucose dropped by 0.4 mmol/l (7 mg/dl)
• fasting insulin fell
• HOMA-IR decreased (less insulin resistance)
• total cholesterol decreased from 220 to 193 mg/dl (5.7 to 5.0 mmol/l)
• LDL-cholesterol decreased from 135 to 124 mg/dl (3.5 to 3.2 mmol/l)
• triglycerides decreased from 168 to 89 mg/dl (1.9 to 1.0 mmol/l)

HDL cholesterol was unchanged.

The fall in AUC (area under the curve) for insulin approached but didn’t reach statistical significance (p=0.08)

Body weight fell from 98 kg (216 lb) to 95.3 kg (210 lb) but I found no p value. HDL-cholesterol was unchanged (the higher HDL I mentioned above is only in comparison to the reference diet, in which HDL fell)

All of these changes (except the lack of change in HDL-chol) would tend to promote health in someone with metabolic syndrome, prediabetes, or overweight type 2 diabetes.

Reference: Boers, Inga, et al. Favorable effects of  consuming a Palaeolithic-type diet on characteristics of the metabolic syndrome: a randomized controlled pilot-study. Lipids in Health and Disease. 2014 Oct 11;13:160. doi: 10.1186/1476-511X-13-160.