245. Weight Loss for Treatment of Type 2 Diabetes (T2D)

Weight loss is an important aspect of the official policy for treating T2D. The recommendations are to increase the amount of fibre and to reduce fat, especially saturated fat. Specific advice includes:

  • Increase consumption of foods such as wholegrain bread and cereals
  • Choose foods that are low in fat – replace butter, ghee and coconut oil with low fat spreads and vegetable oil
  • Choose skimmed and semi-skimmed milk, and low fat yoghurts

If you have a BMI of 30kg/m2 or over you should lose weight by gradually by reducing your calorie intake and becoming more physically active. Losing 5-10% of your total body weight over the course of a year is considered to be a realistic initial target.

Doubts about the guidance on weight loss

There are two fundamental assumptions which underpin the recommendation to lose weight, namely:

  • Excess weight is a direct cause of T2D (and other disease).
  • Elimination of that excess will help to overcome T2D and improve health generally.

It is an article of faith that the BMI should be in the normal range of 20-25kg/m2. Hence anyone with a higher value is advised to reduce their weight in order to achieve “normality”. However this has been questioned by various researchers. In particular Katherine Flegal and her team at the US Center for Disease Control (CDC) have found that with respect to life expectancy, the optimum BMI range is 25-30 kg/m2 which according to the official approach in many countries is “overweight”. Her original report, published in 2005 (1) was based on information from the US National Health and Nutrition Examination Surveys (NHANES) with 36,859 participants, of whom 8,849 died. It was found that the lowest mortality was observed for those in the overweight BMI category. The highest death rates were in those who were underweight (BMI <18.5kg/m2) and with severe obesity (BMI >35 kg/m2). These conclusions have been confirmed by studies in Canada (2) and Norway (3). More recently the CDC team has conducted a meta-analysis based on almost 100 studies from the US, Canada, Europe, Australia, China, Taiwan, Japan, Israel, India and Mexico (4). There was data from 2.88 million participants and over 270,000 deaths. The results confirmed that the lowest death rates were in those who were overweight but also found that the death rates for those with obesity category I (BMI 30-35 kg/m2) were lower than for those with normal BMI.

It has to be recognized that these finding were not accepted with equanimity by everyone (5).

One of the prime concerns is that reports are likely to confuse the public and therefore undermine public policies to curb rising obesity rates. Some of the criticisms are related to decisions on which data should be excluded about individuals who may have lost weight because they smoke tobacco and/or suffer from illnesses and would therefore confound the results. However it has been shown that provided those with obvious weight loss have been excluded, any other any other limitations on the data used in the analyses have little impact on the final result (6, 7).

Relationship between BMI and death rates

As long ago as 1985, Reubin Andres used actuarial data to find out how the relationship between BMI and death rates changed with age. His findings indicated the weight for minimal mortality increased by about 2.5 kg per decade from the age of 40 years for both men and women (8).

Clearly there are serious doubts about the validity of the use of BMI to assess the optimum weight of an individual. It may well be that many people are being advised to lose weight when there is no reliable evidence that there will be a worthwhile benefit. The reality is that those who adopt the conventional approach to weight loss are rarely successful. In fact, they may actually do more harm than good. In a comprehensive review of the effectiveness of various strategies it was concluded that the efforts so far have been disappointing and that essential information is missing (9). Although some promising results have been achieved in clinical trials with carefully controlled conditions, attempts to replicate these in the wider community have not been successful. It is claimed that weight loss in the short term may be effective. Based on an evaluation of studies which continued for more than 12 months, the following points were noted:

  • After one year, any weight lost starts to be regained.
  • The weight loss achieved in the first year in most of the community programmes was less than that achieved in the prevention trials at the equivalent period. In other words it was probably too small to have much impact.

The conventional approach to weight loss is to reduce the input of calories by eating less and increasing the output by taking exercise. The reality is that:

  • It is extremely difficult to achieve weight loss by restricting calorie intake.
  • Of those who do succeed most regain the weight subsequently.
  • Paradoxically there are strong indications that weight loss in this way may actually be associated with an increase in all-cause mortality.

Further doubts about the benefits of weight reduction emerge from the Louisiana State University Hospital-Based Longitudinal Study (LSUHLS) which was a prospective study with diabetics which related death rates to BMI (10). A total of 34,832 (15,354 white and 19,478 black) patients with T2D who were between 30 and 94 years of age participated in the study. During a mean follow-up period of 8.7 years, 4042 subjects (1946 black and 2096 white) died. All-cause mortality (ACM) among blacks (11.5 per 1000 person-years) was lower than among whites (16.4 per 1000 person-years). It was found that among blacks, there was a significantly increased risk of ACM with baseline BMI <30 and ?35 kg/m2. This means that the lowest death rates were with those who were considered obese (category 1). For whites the highest death rates were reported for those with BMI <25 and ?40 kg/m2.while the lowest were also found in those who had a BMI of 30 to 34.9 kg/m2. This relationship did not change when patients with cardiovascular disease and cancer were eliminated from the analysis. Similarly removing those who died within 2 years of the start of the study there was no change. This relationship between BMI and mortality was found irrespective of age, sex, smoking status or type of diabetic medication. These results can be regarded as robust because there was a large sample size with a long follow-up time, and use of administrative databases to avoid differential recall bias.  If this study is accepted as reliable then it would be counter-productive to encourage diabetics to lose weight. There is certainly no convincing evidence that it would be beneficial but there are now strong indications that this may well be harmful.

This is consistent with a study involving 4786 men aged between 40 and 59 years who participated in the British Regional Heart Study (11). BMI was recorded in 1992 and 1996. During the follow-up period of 7 years there were 858 deaths from all causes (381 CVD and 477 non-CVD causes). It was found that weight loss as a result of ill-health or physician’s advice was associated with a small increase in mortality, largely owing to death from non-CVD causes. Furthermore these men tended to have a greater prevalence of disease and tended to be heavier than those who lost weight intentionally as a result of personal choice. It is likely that these men had been prompted to lose weight as a result of having already developed disease. It was also found that when weight had been lost intentionally, irrespective of underlying reason, there was no benefit in terms of CVD mortality. This is in agreement with many other studies which suggest that in these older men who are aged 56 to 75 years, once atherosclerotic vascular disease is established, it is not easily reversed.

In California, an increase in ACM was observed in both men and women, including some with T2D, who lost 4kg or more over a 12-year period (12). In the men with T2D the death rate was over twice that in those who did not lose weight.

We simply have to face the reality that the relationship between weight/BMI and good health is much more complex than was recognised by those who devised the current the guidelines. Despite the emphasis on weight loss as a therapy for diabetics, there is no evidence to demonstrate that it has been successful. Further doubt about the credibility of this approach is that many diabetics have a BMI which is <25 kg/m2.

The conventional method of losing weight is based on reducing the consumption of calories and increasing the output by taking more exercise. With respect to diet, invariably this means lowering the amount of fat and increasing that of the complex carbohydrates as indicated in the advice cited above. The fundamental problem is that this only adds fuel to the fire because it is likely to INCREASE the glucose level in the blood. Hence MORE insulin has to be produced which of course only makes things worse. Since one of the key functions of insulin is to direct the glucose to the liver, where it is converted into fat, and then stored, it follows that it is virtually impossible to lose weight with this type of diet. In practice, the only way to lose weight is to starve, which explains why so few are successful. Not surprisingly, most of these re-gain the weight lost within a relatively short time and probably finish up heavier than they were initially. One of the major weaknesses in in these studies on weight loss by means of calorie reduction has been the failure to take hunger into account. In an article in the New York Times, Gary Taubes comments on a study which compares a diet which is low in fat with one which is low in carbohydrates (13) and notes that no attempt was made to allow for the possibility that the subjects might be hungrier on one diet than the other. In fact, it has been established that one of the benefits of a diet which is restricted in carbohydrates and high in fat is that satiety is readily achieved. Weight can be lost without having to count calories.


The advice to reduce weight by calorie reduction just does not work as shown by many different studies and, of course, the experience of so many individuals. It simply does not make sense to continue a strategy which has been such an abject failure. On top of all this, it reinforces the view that T2D is restricted to those who are overweight/obese. The hard reality is that many with T2D have a BMI that is regarded as “normal” and even if the approach did work it would not be applicable to these patients.

Essentially, the development of policies for coping with T2D has been a total disaster that is confirmed by this examination, which has identified a number of fundamental flaws in the thinking behind them.


  1. K L Flegal (2013). Journal of the American Medical Association 293 (15) pp1861-1867
  2. Heather M Orpana et al(2010). Obesity18 (1) pp214-218
  3. H T Waaler (1984). Acta Medica Scandinavica supplement 679 pp 1-56
  4. K L Flegal (2013). Journal of the American Medical Association 293 (15) pp1861-1867
  5. V Hughes (2013). http://www.nature.com/news/the-big-fat-truth-1.13039#/b2
  6. R Andres (1999). Obesity Research 7 (4) pp 417-419
  7. D B Allison et al (1999). Obesity Research 7 (4) pp 342-352
  8. R Andres et al (1985). Annals of Internal Medicine 103 (6, part 2) pp 1030-1033
  9. R Kahn and M B Davidson (2014). Diabetes Care 37 (4) pp 943-949
  10. W Zhao et al (2014). Circulation 130 (24) pp 2143-2151
  11. S G Wannamethee et al (2005). JAMA Internal Medicine 165 (9) pp 1035-1040
  12. N M Wedick (2002). Journal of the American Geriatric Society 50 (11) pp 1810-1815
  13. Gary Taubes (2015). New York Times 29 August



Scroll to Top