(Always consult your physician before starting a new diet. This includes using the recommendations from our macro calculator.)

In the epidemic of non-communicable diet and lifestyle disease, type-2 diabetes [“diabetes”] has emerged as a particular scourge: the leading cause of death under 60-years of age, and an annual cost of 263-billion dollars in North America in 2013 (1).

The population-based research showing a significant rise in diabetes prevalence over the past 50-years mirrors environmental changes in nutrition habits, food availability, diet composition, and population activity levels (2). The clear environmental determinants of diabetes risk and prevalence in turn point to the potential for diet and lifestyle interventions to be effective in diabetes prevention or remission.

The Relationship Between Bodyweight and Risk for Diabetes

There is a clear association between Body Mass Index [BMI] and risk for diabetes, but the risk is not exclusively linear: BMI varies greatly at time of diagnosis, which indicates that the risk is associated with underlying metabolic complications, including insulin resistance and decreased pancreatic beta-cell [which secrete insulin] function (3).

For example, in metabolically healthy subjects who had the same BMI at baseline, those who progressed to diagnosis 13-years later displayed higher fasting blood glucose [FPG], postprandial glucose tolerance [OGT], and decreased insulin sensitivity at baseline (4).

While this demonstrates that the underlying metabolic dysfunction is a primary issue, the risk is not wholly inseparable from weight, and overweight [BMI ≥25.0 kg/m2] and obesity [BMI ≥30.0 kg/m2] increase risk of diabetes by 3-fold and 7-fold, respectively (5).

In parallel to this risk, weight loss may significantly improve diabetes prevention: a 5-10kg average weight loss is associated with a 50% reduction in risk (6). This is relevant both to extant diabetes prevention, and prevention of progression from “prediabetes” – which refers to 3 states: isolated impaired fasting glucose [IFG], isolated impaired glucose tolerance [IGT], and both IFG and IGT together (7) – to diabetes. It is also relevant to potential diabetes prevention in already diagnosed subjects.

Weight Loss for Prevention of Prediabetes Progression to Diabetes


There are 4 major trials across different populations showing diet and lifestyle interventions may be successful for diabetes prevention. In the Da Qing Study, prediabetic subjects in diet-only, exercise-only, or diet+exercise groups were successful in diabetes prevention by 22-29% on average (8).

Looking closer at the study, 38% of subjects in the diet-only group achieved diabetes prevention 6-years later compared to 60% in the control group, despite slightly more weight gain in the diet group (8).

In the exercise-only group, 43% of achieved diabetes prevention compared to 72% of controls, despite losing exactly the same amount of weight as controls (8). What this indicates is that diet and exercise influence the underlying metabolic complications – like insulin resistance and glucose tolerance – which may not be reflected by changes in BMI alone.

The preventative effect of diet and lifestyle interventions independent of BMI has been found in other trials. The Indian Diabetes Prevention Programme [IDPP-1] trial found a diet and lifestyle intervention was successful in diabetes prevention over 2.5-years, in a population with a high prevalence of progression to diabetes, despite no change in BMI (9).

This is consistent with the positive impact that diet and physical activity have on the prediabetic state, improving FPG and OGT and reducing risk of progression by increasing glucose tolerance and tissue insulin sensitivity (10).

However, while diet and exercise may improve FPG and OGT without a change in BMI, one of the issues with progression from prediabetes is continued decline in pancreatic beta-cell function and insulin secretion (11). This may be the crux of diabetes prevention through weight loss. In the Finnish Diabetes Prevention Study, diabetes prevention of 58% at 3-years follow-up in prediabetic overweight subjects corresponded to an average weight loss of 3.5kg compared to 0.8kg in the control group (12; 13).

Diabetes Prevention Program Study

In the Diabetes Prevention Program [DPP], obese subjects with prediabetes were randomly assigned to an intervention of either metformin, diet+exercise, or placebo: diabetes prevention was 58% in the lifestyle group, compared to 31% in the metformin group (14).

This was a significant study, as it showed that diet+lifestyle was superior to frontline pharmacotherapy for diabetes prevention. In analyzing the respective effects of weight loss, diet or exercise separately, the strongest factor associated with diabetes prevention was an average weight loss of 5kg over 3-years (15). Of particular note was that the effect was independent of diet composition: weight loss was the most important factor (15).

…the primary dietary factor in diabetes prevention is energy restriction is confirmed in other lines of research.

The challenge that is clear in the research is that failure to maintain weight loss may negate the ability of the intervention for diabetes prevention. In the DPP, fasting glucose levels returned to their prediabetic baseline after 3.5-years as subjects increased weight from their initial 7% body weight reduction to 4% (14).

In another study, an average of 2.6kg weight loss did increase insulin sensitivity but failed to restore beta-cell function (16). In the Da Qing Study long-term follow up 20-years later, 80% in the diet/lifestyle intervention group had progressed to diabetes compared to 93% of controls (17). And in the Finnish DPS, 10-year results showed that relapses in weight corresponded to deteriorations in glucose tolerance (18).

This isn’t intended to be disheartening to diabetes prevention, but it does indicate that, cumulatively, the research suggests that weight loss and maintenance of 5kg or ≥5% body weight is required for diabetes prevention from prediabetes (13; 15; 16).

Nutrition in Diabetes Prevention


While bariatric surgery can be an effective intervention for diabetes prevention (19), the fact that it is a surgical intervention and not a nutritional one means this article won’t touch on bariatric surgical procedures for diabetes prevention. Instead, the relevant focus will be on the evidence for nutrition in diabetes prevention.

However, the effects of bariatric surgery provide some clues as to the requirements for diabetes prevention through diet, as significant drops in blood glucose occur within days following surgery prior to any weight loss, in fact, occurring (20). This indicates that the sudden restriction of dietary energy is a primary driver of diabetes remission through surgery (20).

That the primary dietary factor in diabetes prevention is energy restriction is confirmed in other lines of research. Very-low-calorie liquid diets [VLCD] have been a focus of diabetes prevention in clinical settings, and are effective in inducing remission of diabetes.

In a trial in adults with diagnosed diabetes, a 600kcal per day VLCD normalized FPG after 1 week, and by the end of the 8-week intervention both beta-cell insulin secretion and insulin sensitivity had normalized (21).

Why VLCD Isn’t A Long-Term Approach

Part of the critique of VLCD is the ability to sustain long-term clinical effect; in this study subjects gained an average of 3kg in the 12-weeks after the intervention, however, they had lost 15kg during the intervention, and the modest weight regain did not correspond to increases in HbA1c [a marker of long-term blood glucose control] or liver fat (21).

The 15kg weight loss target from VLCD is consistent with the threshold for diabetes prevention observed post-bariatric surgery (22). However, it is clear from the VLCD research that there are responders and non-responders over longer-term maintenance, with one study showing that only 40% maintained remission over 6-months (23).

This is still very clinically significant, however, the caveat of the VLCD research is that it is performed under clinical supervision, and should only be undertaken in medical care.

In the VLCD research, one of the explanations for the return of insulin sensitivity is the decrease in fat in the liver and pancreas, and the effect of reducing circulating fatty acids (21; 23). This is important, as much of the recent focus on low-carb, high-fat [LCHF] diets led to some poor diet advice regarding fat intake circulating the internet [see: Coffee, Butter].

Don’t Dismiss Dietary Fat

However, dietary fats do play a role in the development of diabetes, and nutrition interventions for diabetes prevention need to provide appropriate fat balance. In prediabetic subjects, saturated fats increase FPG and high saturated fat intake may be as deleterious for insulin sensitivity as increasing body fatness (7).

In a controlled feeding trial in both normal glucose tolerant and prediabetic subjects, a high saturated fat intake led to increased whole-body insulin resistance (24).

On the other hand, polyunsaturated fats – in particular, omega-3 fatty acids – have a positive effect on glucose tolerance (24), and research shows replacing saturated fat with polyunsaturated fat reduces diabetes risk (25).

This is an important qualifier when it comes to research looking at diabetes prevention from LCHF diets: the diet setups may be higher in total fat, but they remain lower in saturated fat with emphasis placed on added fats from unsaturated sources (26; 27).

Thus, while there is nothing wrong with a higher total fat diet, the quality and balance of fat subtypes does matter and for optimal metabolic – not just cardiovascular – health, the diet should be lower overall in saturated fat and should emphasize unsaturated fats like olive oil, nuts, fish, and eggs (2).

VLCKD As An Intervention

There are certain proponents of carbohydrate-restricted diets as the primary dietary intervention for diabetes prevention (28). However, a complication with this research area is it is fraught with bias, and one major issue is the lack of any real definition of “low-carb”. So, let’s take a deeper look at the potential for a degree of carbohydrate restriction to be effective in diabetes prevention. The first issue is the degree of restriction: how “low” might one need to go?

Certain research suggests that very-low-carb-ketogenic diets [VLCKD] – with carbohydrate restriction to 20-50g or <10% total energy – are superior to standard very-low-calorie-diets for diabetes prevention (28).

However, the research cited in support of such a proposition didn’t control for calories in the VLCKD diet yet provided a 2,200kcal/d diet as a control diet (29). This is a major limitation, as VLCKD are noted to lead to spontaneous reductions in energy intake, an effect attributable to higher dietary protein intake replacing carbohydrates (30).

Thus, in the absence of the diets being truly controlled for energy intake, the study (29) was not comparing like with like in terms of impacts on glycemic control.

Low, Moderate or High Carb Diet


In a recent long-term trial over 1-year, obese subjects with diabetes were assigned either a low-carb diet of 14% carbohydrate [<50g/d] and 58% fat [with <10% saturated fat] or an isocaloric high-carb diet of 53% carbohydrate and 30% fat [also with <10% saturated fat] (27).

The equating of saturated fat was an important strength of the study, controlling for a nutrient which could impact on insulin resistance (7; 24). In addition, even though the diets differed in carbohydrate content, the study controlled for the glycemic index to minimize the differential effects of simple vs. complex carbohydrates (27).

After 1-year, weight loss was similar in both groups, as were reductions in FPG and HbA1c; the low-carb group did reduce diabetes medications and triglycerides to a greater degree (27).

There are two aspects which emerge from this study: 1) the importance of glycemic index, i.e. carbohydrate quality, and; 2) whether very low carb <50g/d is in fact required for improving glycemic control in diabetes prevention.

In particular, the greater degree of carbohydrate restriction in the short-term over 3–6-months led to greater reductions in HbA1c; over 1-year there was no significant difference between low-moderate or high carb diets (31).

In relation to the former, the research shows that high fiber, low GI diets are associated with diabetes prevention and in extant diabetes, low GI diets lead to greater reductions in HbA1c (2). Thus, relevant to the discussion of the exact amount of carbohydrate in the diet is the stipulation that the type of carbohydrate be complex, unrefined, high fiber, low GI carbohydrates.

The latter question is, however, the divisive one: is “very-low” required, or can more moderate carb diets achieve diabetes prevention?

A recent meta-analysis provides the fairest representation of the state of the evidence overall in specific relation to diabetes prevention, and compared diets with 45-60% carbohydrate [high-carb] with diets <45% carbohydrate [low-to-moderate carb] from randomized controlled trials – average intake in the low-moderate trials was 30% (31). A strong feature of this meta-analysis was the inclusion of studies which quantified carbohydrate intake by percentage, with 4 studies reporting on actual intake in grams (31).

Long-Term Study Results

The results indicate that low-moderate carb diets led to greater reductions in HbA1c over 1-year than high-carb diets (31). In particular, the greater degree of carbohydrate restriction in the short-term over 3–6-months led to greater reductions in HbA1c; over 1-year there was no significant difference between low-moderate or high carb diets (31). This meta-analysis also confirmed that there was a greater reduction in medication over 3–6-months in low-moderate carb diets (31), an effect observed in other trials (27).

One of the salient features of this meta-analysis – which is a consistent observation in the literature – is that whatever the percentage of initial carbohydrate restriction, carbohydrate intake amongst subjects incrementally increases when analyzed over the longer term (>12-24m).

This should be considered in light of the greater drop-out rates in low-carb diet groups in the included studies (31). This suggests that practically for diabetes prevention, the degree of restriction of CHO <20-30% of calories in neither necessary for treatment effect over the long term nor achievable in free-living settings.

Ultimately, the contentions of marked superiority to very low-carb diets are simply not borne out in high quality randomized controlled trials – improvements in glycemic control occur with a reduction in energy <45% (31).

In a recent trial, prediabetic obese adults were randomized to a diet of either 30% protein, 30% fat and 40% carbohydrate vs. a control diet of 15% protein, 30% fat and 55% carbohydrate: after 6-months, the higher protein/lower carb group had achieved 100% remission to normal glucose tolerance compared to only 33% of subjects in the standard diet (26).

The overall weight of the literature certainly supports reducing carbohydrate intake from >50% to <45%; it does not support any real need to go <20% for therapeutic effect, and <30% appears to be practically unsustainable in free-living conditions.



Let’s recap the foregoing paragraphs with a synopsis of the research:

Nutrition interventions can be successful in diabetes prevention;

Weight loss is a fundamental, overarching goal in both pre-diabetic and diabetic states;

In “prediabetes” – impaired fasting blood glucose, impaired glucose tolerance, or both FPG + IGT, the primary goal is weight loss and maintenance of 5-10kg or 5-10% baseline weight;

In extant diabetes, very-low-calorie liquid meal replacement diets may be employed to induce remission, but only under clinical supervision;

Very-low-carb-ketogenic diets do not appear to be necessary, but may certainly be strategically employed in the short-term (3-months) to achieve greater glycemic control;

Low-moderate carb diets, defined as energy intake from carbs 20-45%, maybe the appropriate intervention generally going over 3–6-months. The lower intake does not appear to be necessary over the long-term;

The carbohydrate type should be unrefined, whole grain, complex [i.e. low GI] carbohydrates;

Protein is favored as the ideal replacement nutrient for carbs;

The fat composition should emphasize added unsaturated fats from plant sources and oily fish.

It is generally acknowledged that diabetes prevention can be achieved through diet and lifestyle change. However, the interventions in research often involve intensive in-person counseling, diet guidance, supervision, and in some cases prepared meals.

If you’re dealing with prediabetes or extant diabetes, make sure to work with a Registered Dietician or regulated nutritionist legally entitled to give medical nutrition advice, in conjunction with your medical supervisors.