Low Carb or Low Fat? Let Science Settle the Debate

Part 1 | Part 2| Part 3 | Part 4 | Part 5

In part one of this series, I discussed how universal truths are a standard bit of rhetoric in the fitness industry. In part two, I offered some reasons for how the scientific literature has unwittingly supported this naïve view. Then in the latest installment, I pointed out how more enlightened practitioners have attempted to classify individuals based on somatotype, in order to better define an individual’s dietary needs.

Unfortunately, although the somatotype categorization is a quantum leap forward in establishing dietary recommendations (in my opinion), it doesn’t always work quite so perfectly in the real-world.

Why Can’t Life Be Simple?

Recall that according to somatotype theory, a short, squat individual is labelled an endomorph. Correspondingly, they should consume a high protein, high fat, type of diet.

While this recommendation often produces excellent results, occasionally it doesn’t work very well at all.

Why might this be the case?

Only one way to find out… off to dig through some original research to see if we can locate the answer…

Although somatotyping has been around for ages, there is a surprising lack of research linking body fat distribution and diet recommendations. However, I was able to come up with a couple of studies that shed some light on the issue.

Insulin Sensitivity and Weight Loss Success

Back in 2005, researchers in Colorado took two groups of over-fat females (BMI: 30-35 kg/m2) and stratified them based on insulin sensitivity. Subjects were included in the study if they were insulin sensitive (IS), defined as having a fasting insulin level of < 10 µU/mL, or insulin resistant (IR), as determined by a fasting insulin level of >15 µU/mL.

The goal of the study was to compare weight-loss after 16 weeks of dieting (400 kcal deficit/day) on diets of differing macronutrient composition. The two diets under investigation were:

  • HC/LF: 60% CHO, 20% fat, and 20% protein.
  • LC/HF: 40% CHO, 40% fat, and 20% protein.

So in total this study had 4 conditions:

  1. Insulin sensitive eating a high carb/low fat diet.
  2. Insulin sensitive eating a low carb/high fat diet.
  3. Insulin resistant eating a high carb/low fat diet.
  4. Insulin resistant eating a low carb/high fat diet.

Still following me?


So after 16 weeks, what did these researchers report?

Obes Res. (2005). Insulin Sensitivity Determines the Effectiveness of Dietary Macronutrient Composition on
Weight Loss in Obese Women, Human Physiology, 13:703–709.

First the good news: every group lost weight. Now the even better news: by tailoring diet advice to an individual’s physiological responses, it may be possible to produce substantially greater weight loss!

Among individuals with good insulin sensitivity, weight loss was nearly twice as effective when following a higher-carbohydrate diet. As I’ve pointed out previously, having good insulin sensitivity is critical if you are going to consume a high carbohydrate diet.

But this data also suggests that not only is having good insulin sensitivity critical when consuming a high carbohydrate diet, but that individuals with good insulin sensitivity actually respond better (in terms of weight loss) eating more carbohydrates!

Conversely, individuals with poor insulin sensitivity (i.e. they were insulin resistant), lost far more weight when they consumed a diet higher in fat.

I suspect weight loss would have been even greater in the insulin resistant group had the researchers  selected an even lower carbohydrate level. However, in the interests of performing a well-controlled study, protein intake needed to be kept constant.

What this suggests is that if you have poor insulin regulation (partially a function of your genetics, but also potentially as a result of your lifestyle), you should NOT consume a diet high in carbohydrates.

Both of these outcomes make perfect sense when we look at the underlying physiology, yet somehow the need for personalized diet recommendations based upon an individual’s biochemical and physiological make-up still remains a foreign concept to many in the nutrition field… <sigh>, maybe one day.

GT’s note: this study is just one more piece of concrete evidence that weight-loss advice consisting solely of calorie counting is of limited use. Clearly, we must also consider specific endocrine and metabolic responses to particular foods.

All of this looks quite promising, so what’s the problem?

Taking Typing One Step Further

The problem with findings such as this one is that it points to somatotyping as overly simplistic. Remember, all the individuals in this study had BMIs over 30. This means that had we used anthropometrical or even visual data, we would have misclassified a number of individuals.

Although a somatotype-based diet recommendation may account for underlying physiology better than the alternative (one size fits all recommendations), it certainly isn’t fool proof.

Thankfully, we may be on the cusp of some major breakthroughs in our understanding of diet individualization!

Earlier this year, researchers out of Stanford University presented a re-analysis of data from the A to Z weight loss study that showed dramatically greater weight loss occurs when matching individuals to diets based on simple genetic markers.

Originally, the A to Z weight loss study was designed simply as a year-long comparison of the effectiveness of the Atkins (very low carb), Zone (40% carb, 30% protein, 30% fat), Ornish (very low fat, <10% of total calories) and LEARN (a lifestyle behaviour-change program) diets.

Much like any reasonable diet study, researchers randomly allocated their 311 overweight, premenopausal females subjects to one of the four diet conditions and tracked results over the 12-month time span.

At the study’s conclusion, the weight loss results were as follows:

  • Atkins: -4.7 kg (95% confidence interval: -6.3 to -3.1 kg)
  • Zone: -1.6 kg (-2.8 to -0.4 kg)
  • LEARN: -2.6 kg (-3.8 to -1.3 kg)
  • Ornish: -2.2 kg (-3.6 to -0.8 kg)

Had the researchers stopped their analysis right there, one could conclude that the Atkins diet is the best diet for weight loss, because that’s what the data suggests.

Fortunately, the researchers collected DNA swabs on 101 of these women. Tests were performed on these DNA samples and researchers were able to identify 3 distinct genotypes:

  • a low carbohydrate diet responsive genotype (LCG, n=61)
  • a low fat diet responsive genotype (LFG, n=35)
  • a balanced diet responsive genotype (BDG, n=5)

When the data was re-analyzed with this concept of bio-individuality in mind (discussed in part 3 of this series), a very different story emerged.

Nelson MD, Prabhakar P, Kondragunta V, Kornman KS, Gardner C. Genetic Phenotypes Predict Weight Loss Success: The Right Diet Does Matter. (Oral Presentation #4). Presented at the American Heart Association’s Joint Conference – 50th Cardiovascular Disease Epidemiology and Prevention and Nutrition, Physical Activity and Metabolism – 2010, March 2-5, 2010, San Francisco, CA. Available online here.

When classified by correct vs. incorrect genotype, correctly matched individuals lost 2-3x more weight (~13.2 lbs vs. 4.6 lbs) than individuals who were placed in the wrong diet condition.

In other words, when an individual who is genetically predisposed to poor carbohydrate metabolism was placed on the low carbohydrate Atkins diet, they lost a ton of weight. A similar result was found when individuals with a strong carbohydrate metabolism were placed on the Ornish diet, they lost a lot of weight.

However, when an individual with a genetic predisposition for favourable carbohydrate metabolism was given the Atkins diet, results were predictably abysmal; same story occurred when individuals with poor carbohydrate metabolism were placed on the Ornish diet.

If you’d like to look over some of physiological rationale behind the genetic test used in this study: click here.

Once again, solid evidence that by correctly matching a diet to an individual’s biochemical make-up, demonstrably greater weight loss can be achieved.

Individualization in the Real World

Although this kind of research is exciting, we are still a good ways away from having easy, affordable access to these kinds of genetic tests.

We also need more research to verify whether the population breakdown of 60% low-carb genotypes, 35% high-carb genotype and 5% balanced genotype is one that holds true for the North American population as a whole, or whether it was unique to this study.

Regardless, results such as these definitely support the efforts that practitioners have made to stratify people by diet type.

At the current time, no one can say for certain whether the optimal breakdown will ultimately involve 3, 4, 6 or even 12 different metabolic types; however, it’s abundantly clear our answer is NOT “one size fits all”!

Of course, identifying what an individual should eat is only half the battle, convincing them to change their dietary habits is an entirely different challenge.

In the final installment of this series, I’ll let everyone try a tool that I’ve had success using to help match individuals to an optimal diet approach.

Stay tuned for that!

Till next time, train hard and eat clean!