If losing weight is to be justified then there should be convincing evidence that there will be definite benefits to health for those who actually succeed in reducing their body weight.

In an investigation in which 2453 men and 2739 women aged 45 to 74 were monitored initially in the period 1971-1975 and followed up until 1987, it was found that there was a strong association between weight loss and the risk of death from cardiovascular disease among both men and women with BMI in the range 26-29….the overweight category. For men this risk had increased by a factor of 2 and for women it had increased by a factor of well over 3 compared with those who did not lose weight. Similar results were obtained if the analysis was restricted to non-smokers (1).

In the Honolulu Heart Study, 6537 Japanese-American men aged 45 to 68 in 1965 and living in Hawaii, were monitored from 1973 to 1988 during which time there were 1217 deaths. The results are shown in Table 1. It is quite clear that there has been a notable increase in the death rate of those who lost weight. By contrast, those who gained up to 4.5kg had a reduced death rate while those who gained more than that did not experience any increase in mortality rate when compared with those whose weight did not vary(2).

TABLE 1 RELATIVE DEATHS AND CHANGE IN BODY WEIGHT

  ALL-CAUSE MORTALITY  RELATIVE RISK
Weight loss >4.5 kg 1.21
Weight loss 2.6-4.5 kg 1.29
No change 1.00
Weight gain 2.5-4.5 kg 0.83
Weight gain >4.5 kg 0.99

 

A study conducted in Southern California monitored 140 men and 90 women with and without diabetes aged between 40 and 79 when the work commenced in 1972-1974. The follow-up was done after 12 years. It was found that those who lost 4kg or more had an increased death rate compared those who did not lose weight (Table 2). Similar results were obtained after excluding cigarette smokers, those who were depressed, had a low BMI and those who died within 5 years of losing weight(3). 

TABLE 2 CHANGE IN DEATH RATE AS A RESULT OF LOSING WEIGHT

Non-diabetic men +38%
Non-diabetic women +76%
Diabetic men +266%
Non-diabetic women +65%

 

Because it is so common for weight that has been lost to be regained many people subject themselves to “weight cycling” which is losing weight and then regaining it repeatedly. A major national project, the Nurses’ Health Study, has been conducted in the USA in which 46,224 women were monitored over a 4-year period. It was found that:

  • 78% intentionally lost weight (between 2.25 and 4.5 kg) at least once
  • 41% had similar weight losses at least twice
  • 20% had intentionally lost 4.5 kg three times during the 4-year period(4)

Another study from the USA followed 8479 men and women from 1971 to 1992. The respondents were classified as:

  • Stable weight non-obese
  • Stable weight obese
  • Weight gain
  • Weight loss
  • Weight fluctuation

All cause mortality(ACM) in the weight fluctuation group was 83% higher than in the stable non-obese group even after controlling for pre-existing disease, initial BMI and excluding those of poor health or those who were incapacitated. For those in the weight loss group the ACM was increased by a factor of over 3 when compared with the stable non-obese group. It is also highly relevant that the stable obese did not show any increase in mortality rates (5).

As part of the Helsinki Businessmen Study, 1815 healthy men born between 1919 and1934 were monitored between 1974 and 2000, during which time 909 died. On the basis of BMI at in 1974 and in 2000 the participants were allocated to 4 different groups viz:

  • A. Normal weight in 1974 and in 2000 (n=345)
  • B. Overweight in 1974 and in 2000 (n=494)
  • C. Normal weight in 1974, overweight in 2000 (n=136)
  • D. Overweight in 1974, normal weight in 2000 (n=139).

The results (Table 3) show that the death rate for those who lost weight (Group D) was approximately double that of the other 3 Groups, in which the death rates were similar. Despite the fact that a number of different factors have been used in the calculations the results are essentially the same.  If these results can be extrapolated to men generally then it would mean that it is only weight loss which is linked to any substantial increase in mortality. They also provide further confirmation that those who are overweight as determined by BMI do not have a higher death rate than those assessed as normal (6).

TABLE 3 HAZARD RATIOS FOR ALL-CAUSE MORTALITY BETWEEN 2000 AND 2006 IN THE DIFFERENT GROUPS BASED ON BMI HISTORY

Group A B C D
Model  1 1.00(reference) 1.2 1.0 2.0
Model  2 1.00 1.2 0.9 2.0
Model  3 1.00 1.2 0.9 2.0
Model  4 1.00 1.1 0.7 1.9

 

Note: The factors used in making the making the calculations above are shown below.

Model 1: adjusted for age.

Model 2: adjusted for age and smoking in 1974.

Model 3: adjusted as above plus perceived health in 1974.

Model 4: as above plus reported diseases (systemic hypertension, diabetes, memory disturbances,

cerebrovascular disorders, coronary heart disease, congestive heart failure, pulmonary disease,

musculoskeletal disease, and cancer) in 2000.

A review of studies, which combined the results from a range of investigators concluded that:

“…the highest mortality rates occur in adults who have either lost weight or gained excessive weight. The lowest mortality rates are generally associated with modest weight gains”(7).

Although this review was published almost 20 years ago there have been numerous investigations since which effectively confirm this conclusion. On the other hand there is a complete dearth of research to support the current official advice being promulgated in the UK and many other countries that those who are in the overweight BMI category will derive health benefits from body weight reduction.

REFERENCES

  1. E Pamuk et al (1993) 119 (7,Part2) pp.744-748
  2. C Iribarren et al. (1995) New England Journal of Medicine 233 (11) pp 686-692
  3. N M Wedick (2002) Journal of the American Geriatric Society 50 (11) pp 1810-1815
  4. A E Field et al (1999) Epidemiology (1999) 150 pp573-579
  5. Vanessa A Diaz (2005) Journal of Community Health 30 (3) pp 153-165
  6. T E Strandberg et al (2009) European Heart Journal 30 (14) pp 1720-1727
  7. R Andres et al(1993) Annals of Internal Medicine 119 (7) pp 737-743.