More evidence (should you require it) that supports the notion that fat doesn’t induce diabetes, but carbs can

Variously on the site in the past I have attempted to highlight the science (never mind the common sense) that if you have diabetes and want to control it (or don’t want to get this condition in the first place), the part of the diet that generally needs reigning in comes in the form of carbohydrate (not fat). Diabetes is principally a problem of carbohydrate metabolism right. So, it makes sense that carb is the part of the diet that, primarily, diabetics need to be wary of, no? The carbohydrates most worth fingering in this respect are those that tend to cause substantial rises in blood sugar.

How disruptive a food is with regard to blood sugar levels is mainly determined by two things: the speed and extent to which it releases sugar into the bloodstream (its glycaemic index or ‘GI’), and how much we eat of it. Together, these two factors can are called a food’s ‘glycaemic load’ or ‘GL’. Basically, the higher the glycaemic load, the worse a food is for us (especially if we have diabetes and struggle with maintaining blood sugar levels in the healthy range).

With this in mind, I read with interest a study which assessed the relationship between dietary carbohydrate and risk of type 2 diabetes in women [1]. The study, published this month in the American Journal of Clinical Nutrition, assessed the relationship between a number of dietary factors and diabetes risk in more than 85,000 women over a 20-year period.

One of the dietary elements focused on in this study is what the authors termed the ‘low-carbohydrate-diet’ score. Basically, the less carbohydrate and more protein and fat individuals ate, the higher their score. The authors used this as a proxy for an individual’s adherence to the sort of low-carb diets (e.g. Atkins/South Beach/Protein Power) that individuals may have adopted in an effort to lose weight and improve their health. The basic question being asked by this study was, are such diets safe? Previous work by the same group had found that low-carb diets were not associated with an increased risk of heart disease (though some will insist that they are) [2]. Now, the authors were looking for any evidence that low-carb diets increased the risk of diabetes.

The long and short of it was that once ‘potential confounders’ such as body mass index and activity were taken into consideration, individuals with a higher low-carb-diet score were not found to be at increased risk of diabetes. Personally, the only thing I find in any way surprising about this finding is not that low-carb diets are not linked with an increased risk of diabetes, but that they appeared not to be associated with a reduced risk of this condition.

One of the criticisms I would have of the ‘low-carbohydrate-diet’ score is that it is a relatively crude measure of dietary quality. For instance, while it gives us some idea of what proportion of the diet comes from carbohydrate, it tells us nothing of the quality of those carbohydrates. And eating a diet very rich in white bread is likely to bring about quite a different result from eating a diet rich in broccoli, after all.

Fortunately, this study question collected and presented enough data for a more detailed appraisal to be possible.

In fact, this study analysed data relating to the glycaemic load of the diet and risk of diabetes. This is a good thing as the GL tells us not so much how much carb there is in the diet: it gives a guide to how much the diet is likely to upset blood sugar levels (which is far more useful).

And when confounding factors were accounted for, individuals with the diets of the highest glycaemic load were found to be about 2½ times more likely to develop type 2 diabetes compared to those consumed diets of lowest glycaemic load.

Analysis also revealed no link between animal fat intake and diabetes risk, and those with higher intake of vegetable fat were found to be at reduced risk of diabetes.

So-called epidemiological studies of this nature cannot be used to prove that a high GL diet causes diabetes. However, these findings are consistent with this notion. They are also consistent with, well, common sense.


1. Halton TL, et al. Low-carbohydrate-diet score and risk of type 2 diabetes in women. Am J Clin Nutr 2008;87:339-346

2. Halton TL, et al. Low-carbohydrate-diet score and risk of coronary heart disease in women. New Eng J Med 2006;355:1991-2002

8 Responses to More evidence (should you require it) that supports the notion that fat doesn’t induce diabetes, but carbs can

  1. Neil 13 February 2008 at 6:54 pm #

    Hi John,
    I feel that explanations of the Glycaemic Index should be precise, NOT refer to speed or height of the rise in blood glucose. I feel this only increases the confusion over the effects of carbohydrate on blood glucose.

    Measuring the GI from the University Of Sydney, home of the original research

    ” To determine a food’s GI rating, measured portions of the food containing 10 – 50 grams of carbohydrate are fed to 10 healthy people after an overnight fast. Finger-prick blood samples are taken at 15-30 minute intervals over the next two hours. These blood samples are used to construct a blood sugar response curve for the two hour period. The area under the curve (AUC) is calculated to reflect the total rise in blood glucose levels after eating the test food.
    The GI rating (%) is calculated by dividing the AUC for the test food by the AUC for the reference food (same amount of glucose) and multiplying by 100. The use of a standard food is essential for reducing the confounding influence of differences in the physical characteristics of the subjects. The average of the GI ratings from all ten subjects is published as the GI of that food. ”

    According to the University site,

    low GI = 0% to 55%
    medium GI = 56% to 79%
    high GI = 79% to 100%

    Personally I think these categories are artificial, after all a food with 79% of the effect of glucose is medium GI !!!!

    Why not divide it into equal thirds?? Obviously that would be politically incorrect as many allegedly medium GI foods would become recategorised (correctly in my view) as high GI, and many low GI would become medium GI.

    As you say in your essay, Glycaemic Load is a more useful concept than Glycaemic Index.


  2. Paul Anderson 14 February 2008 at 1:08 pm #

    I wonder if diabetes, and obesity, are not so much a product of defective carbohyrdrate meatbolism as a product of defective fat metabolism.

    As I understand it, in a healthy individual, the insulin response to the carbohydrate portion of a meal removes the resulting glucose from circulation within approximately two hours: the glucose either being used as energy, to restore glycogen stores with any surpus stored as fat in the adipose tissue. Between meals the fat becomes the body’s main source of energy and, all being well, weight remains at an equilibrium healthy level. (a BMI of approximately 22).

    It seems that in diabetics (type 2’s) and the overweight (insulin resistant) there is an exaggerated insulin repsonse to carbohydrates that has the effect of trapping fat in the adipose tissue, and prolonging the period during which insulin, and eventually blood glucose, remain above a normal fasting level.

    So it would appear that the exaggerated insulin repsose triggered by high carb diets (both GI and load) initially disturb fat metabolism and eventually leads, in many cases, to high elevated blood glucose and diabetes.

    Its the inability to access stored body fat that promotes hunger and increases appetite. This situation is aggravated by a diet high in “healthy” complex carbohydrates.

    This is how I interpret Gary Taubes recent work. Obesity is a product of a malfunctioning fat metabolism which is probably initiated by over consumption of refined carbohydrates – and perhaps insufficent fat and protein. Eating unprocessed foods based upon protein, fats and low GI veg (and fruit) would in my opinion greatly reduce the incidence of diabetes and obesity and represents the best way of controlling the former and reversing the obesity crisis.

    This diet seems to greatly improve satiety. I would throw in cod liver oil, a multi vitamin (without iron) and an exercise programme for good measure. I suspect that Taubes may be correct in his belief that exercise will not generate weight loss but I do believe it can help indirectly by enhancing insulin sensitivity. Exercise may well be one of those classic chicken and egg issues. But in any case, a moderate exercise programme which includes some aerobic, anaerobic and flexibility exercises have a host of benefits over and above weight loss. Oh, and regular does of sunshine as the icing on the cake.

    Paul Anderson.

  3. Sue 15 February 2008 at 2:32 am #

    Well said Paul – my belief too.

  4. Hilda 15 February 2008 at 4:44 pm #

    Dr Briffa, I read the Pubmed abstract from the study and it said in the results that ‘a higher carb consumption was associated with an icrease in the risk of diabetes’ but conclusion was more cagey. Data interpretation is often difficult. Hilda

  5. Hilda 15 February 2008 at 4:52 pm #

    Another problem with the GI diet or using GI as a means of choosing food is that a high sugar food such as a Mars bar might come lower in the GI scales as it is combined with fat so the sugar goes into the bloddstream more slowly that just sugar on its own. White bread has a higher GI than sucrose. But a sandwhich with lots of chicken, eg, overall would have quite a low GI so it all depends on how they ate their food. Pasta has a lowish GI but is pure starch. I think the upshot is eat to eat meals with protein at every meal rather than bowls of pasta.

    Incidentally lady on TV trying to eat a really healthy diet : processed cereal for brekfast, pasta for lunch and dinner, cereal in the evening! Where do people get then info from?

  6. Liz 16 February 2008 at 6:16 pm #

    I agree that high carbs are the problems especially in a diet that will exaggerate the glycaemic load (e.g. low fat/protein).

    And it’s certainly the case that once someone is obese enough to have central adiposity the fat cells take over and have their own little hormone factory that stimulates further insulin production and insulin resistance.

    How sad then that standard dietary advice is still low fat. I’ve had a number of weight loss clients since the new year, all teetering towards type 2 diabetes or suffering from it. And all of them proudly say “I’ve had yogurt but it was low fat” and other similar comments. It’s almost impossible to get these people to eat nuts and seeds, oily fish or olive oil. They have become obsessed with calories and low fat.

    A calorie analysis shows that most of them on a diet are eating too few calories and are pushing their metabolism towards starvation mode. They need to eat more, eat better quality foods and stop starving themselves of fat. This more natural type of diet seems to work much better for appetite control as well.

    It’s a real act of faith for some people to abandon low fat and trust their eating instincts around food, so it’s useful to have info like this to show how sensible fat eating is OK.

  7. Tiggy 18 February 2008 at 3:08 am #

    Much as I’d love to put my PCOS and insulin resistance down to birth, I do remember eating rather too much chocolate as a child and being very sedentary. Also my mother overfed me – she overfeeds everyone. I expect I could have stayed slim had that been my priority earlier in life. Once you build up the adipose cells though, you can’t get rid of them.

    I have some of those Japanese seaweed pills (Fucothin) that are supposed to destroy adipose tissue, but I have no idea if it’s working. I fear the task is too great for it!

    There are slim people with insulin resistance too. I never quite understand the bit about fat round the ‘middle’. Do they mean the waist or the belly? I don’t have a big waist at all, but then I’m half Asian, well Middle Eastern. I’m never sure what ‘Asian’ means when it’s referred to in medical literature. I wrote to the NHS as they’d used the term on their website and they weren’t sure either.

  8. None Given 20 February 2008 at 8:32 pm #

    The belly fat can be internal, packed around your organs. Someone who hasn’t yet gained a lot of weight may not notice it.

Leave a Reply