Good to be fat?

“Now doctors say it’s good to be fat” was the front page headline of The Independent today. The newspaper added: “After years of anti-obesity public health advice, a major new study causes an outcry by concluding that the overweight live longer.”

This coverage was based on the results of a US study that looked at the causes of death between 1971 and 1994, and categorised the people who died according to their BMI. This new statistical analysis found that overweight people in the study were less likely to die of cancer and heart disease than those of a healthy weight.

Although The Independent did give an accurate, balanced account of the study's findings within its report, the front page coverage may have given the false impression that being fat has been found to be beneficial. It is important to point out that the study also found that overweight people were more likely to die from diabetes and kidney disease, and that obese people were more likely to die from all the diseases the study examined.

Rather than being interpreted as a guide to how we should behave as individuals, this study should be considered in a more useful context - as being an aid to those who must make informed decisions when forming national policies. At the current time, it seems sensible not to change healthy eating and regular exercise behaviour based on these findings.

Where did the story come from?

Dr Katherine Flegal and colleagues from the Centers for Disease Control and Prevention in the US carried out this research. The data were collected by the National Center for Health Statistics, and both organisations approved the report before publication. There was no external funding. The study was published in the peer-reviewed Journal of the American Medical Association (JAMA).

What kind of scientific study was this?

This study combined the results from three large US surveys on the number and cause of death from three periods between 1971 and 1994. A further part of the study compared the results of the three studies with each other.

The researchers estimated the number of “excess deaths” associated with three weight groups; underweight, overweight and obese. These groups were defined by the person’s Body Mass Index (BMI) a measure of obesity calculated using a person’s height and weight. Underweight was defined as a BMI less than 18.5; overweight as BMI 25 to 29.9; and obese as BMI more than 30.

The researchers defined “excess deaths” using a “cause-specific attributable fraction”; this being a measure of how many deaths from a particular cause could be prevented if a specific risk factor was removed; in this case being underweight, overweight or obese.

The risk of death in the three abnormal weight categories was compared to the risk in the normal weight category (BMI 18.6 to 24.9). The difference was used to estimate the excess deaths that could be attributable to each cause, for example coronary heart disease, lung cancer, diabetes and kidney disease or respiratory disease.

In a separate “balanced” analysis the authors only included mortality data collected in the first 15 years, reasoning that basing the estimated excess deaths on the total follow up of all three surveys might give misleading results. This was because the earlier studies had longer follow up (about 30 years) and because the participant’s weight may not have been measured since they were first enrolled in the survey.

What were the results of the study?

The researchers found there to be a complex relationship between the weight groups and death.

People in the underweight group were more likely to die from causes not related to cancer or cardiovascular disease (CVD).

Surprisingly, being overweight was not associated with more deaths from cancer or cardiovascular disease. However, people in the overweight group were substantially more likely to die from diabetes and kidney disease. Conversely, these people were significantly less likely to die from other causes that were not cancer or CVD-related. Overall, being overweight was associated with a significant decrease in deaths from all causes.

People in the obese group were associated with more deaths from CVD, from some cancers (such as cancers of the colon, breast, oesophagus, uterus, ovary or pancreas), and from diabetes and kidney disease combined. Obesity showed little or no association with other cancers (such as lung cancer) and with other causes of death. Overall, obese people was associated with more deaths which were mainly due to cardiovascular disease.

What interpretations did the researchers draw from these results?

The authors suggest that the data indicate that the association of BMI with mortality varies considerably by cause of death. They refer to earlier results from previous studies which found a similar trend. One explanation they offer for this is that "Overweight... may be associated with improved survival during recovery from adverse conditions, such as infections or medical procedures, and with improved prognosis for some diseases. Such findings may be due to greater nutritional reserves or higher lean body mass associated with overweight."

In other words, it is possible that being “above normal weight for your height” may not prevent you getting the disease in the first place but may improve your chances of surviving when you do.

What does the NHS Knowledge Service make of this study?

This is a well-conducted study where the authors have succeeded in presenting complex information well. The study is likely to cause a re-evaluation of advice and may temper the current emphasis on encouraging weight loss for the overweight group (those with a BMI of 25 to 29.9).

However, before relaxing entirely on diet we should bear in mind several limitations of the study:

The study did not look at any outcomes other than death; there are quality of life issues and non-fatal diseases that could possibly be prevented by losing weight. The study did not address these.
The number of excess deaths associated with obesity that result from cardiovascular disease, cancer or other causes have progressively reduced over the three surveys, despite the average BMI for the population increasing over this time. There are other improvements in individual behaviour, heath services access, and technological advances with medication and surgery that partly explain this reducing mortality; and the links between these have not been evaluated in these surveys.
It is also not possible to reliably assess from the study report, how weight and height were measured in the participants to give a calculation of BMI. For instance, if these were based on the participants’ self report, they may be estimated and therefore introduce some inaccuracies to the data.

Sir Muir Gray adds...

The length of our life is only one measure. The journal this research appeared in also published another study that showed that disability is now even more common in obese older people. The headline of the editorial in this journal asked “Is Disability Obesity's Price of Longevity?”

Our aim is to add life to years, not just add years to life.