"Being overweight or obese puts men at a greater risk of dying prematurely than women," BBC News reports.
A survey of global trends found obesity was now second only to smoking as a cause of premature death in Europe. A study of almost 4 million people from 32 countries showed that being overweight (as well as being underweight) increases the risk of dying early, compared to people with a healthy weight. This is usually defined as having a body mass index (BMI) of between 18.5 and 24.9.
The study was designed to calculate the impact of BMI on chances of death in four geographical regions, free from the effects of confounding factors such as smoking or existing chronic disease.
Researchers calculated that, in Europe, 1 in 7 (14%) premature deaths could be prevented if people were a healthy weight, rather than overweight or obese. Men who were overweight were more likely to die early than women who were overweight.
The study does not prove that obesity causes early death, only that people who are overweight or obese are more likely to die earlier. Other factors such as diet, exercise, socioeconomic status and ethnicity may have an effect on people's individual risk, as well as their BMI.
That said, it does cast doubt on previous claims that it is possible to be "fat and fit", while also adding to evidence that a healthy weight plays an important role in the chances of living a long and healthy life.
The study was carried out by over 500 researchers from more than 300 institutions in 32 countries. It was co-ordinated by researchers at the University of Cambridge and was funded by grants from organisations including the UK Medical Research Council, British Heart Foundation, Cancer Research UK, National Institute of Health Research and US National Institutes of Health.
The story was widely covered in the UK media, with reasonably accurate reports.
Several articles quoted figures supplied by researchers in The Lancet's press release, which are not included in the main body of the report. While these figures (which look at the chances of death for men and women at different ages and different BMI levels) may well be true, they are not included in the main study, so we cannot verify their accuracy.
This was a meta-analysis of 239 cohort studies, carried out in four continents (Asia, Australia and New Zealand, Europe, and North America).
Meta-analyses pool data from lots of smaller studies, to give a more reliable overall figure. Cohort studies are good for showing links between factors (in this case BMI and death) but cannot show that one causes another.
A big group of researchers (more than 500 people) agreed to analyse large prospective studies of more than 100,000 people, which included data about BMI and mortality (death).
They excluded people who had ever smoked, people who'd been diagnosed with a chronic illness, and people who died in the first five years of the study. They then calculated the chances of having died during the study, for people in nine BMI categories, from very underweight to very obese.
Study centres on four continents used standardised analysis methods to ensure the results were as comparable as possible. They included studies with information about weight, height, age and sex, from a general population (not a group of patients with a particular disease), with records of deaths, and more than five years of follow up.
They deliberately excluded from their analysis all people in the studies whose records showed they had ever smoked, been diagnosed with a chronic disease or who died in the first five years. They also excluded people aged under 20 or over 90 at the start of the studies, or with a BMI under 15 or over 60 (the healthy range of BMI according to the World Health Organization (WHO) is 18.5 to 24.9).
The exclusion of people with chronic illness, who died within five years, or who had smoked, was because these things have an effect on people's BMI and can skew the results. For example, people who smoke often have a lower BMI, but are at increased risk of dying early, so that can mask the effect of a higher BMI.
Researchers then pulled together all the information to calculate the chances of death at different BMI levels, in different geographical regions and for different ages and sexes.
Researchers looked at the records of more than 10.6 million people from 239 studies, then narrowed their research down to 3.95 million people from 189 studies, after excluding smokers and people with chronic disease or who died within five years.
The pooled data showed that people with a BMI of 20 to 25 had the lowest chance of death. People with a BMI lower or higher than this had an increased chance of death. For overweight or obese people in Europe and east Asia, every additional five BMI points was linked to an additional 39% increase in their risk of death (hazard ratio (HR) 1.39, 95% confidence interval (CI) 1.34 to 1.43); the relative risk (RR) was slightly lower in the US and Australia.
Other notable results were:
The researchers say their results "challenge previous suggestions that overweight (25 to less than 30kg/m2) and grade 1 obesity (30 to less than 35kg/m2) are not associated with higher mortality, bypassing speculation" that excess fat might actually protect people who are otherwise healthy.
They say the size and rigour of their study provide a better estimate of the link between overweight and obesity than previous studies, which have been unable to adjust their figures fully to take account of the effects of smoking or pre-existing illness. They say their study supports efforts to combat overweight and obesity at all levels, worldwide.
The effect of being overweight or obese on length of life has been discussed a lot in recent years, mainly because of studies which seemed to show people might live longer if they have a BMI in the overweight range, and that even moderate obesity did not raise the risk of death.
However, this study suggests the previous findings were due to confounding factors – such as smoking and pre-existing diseases – which masked the link between BMI and length of life. The overall conclusion is that weight does matter, especially for men and younger people, who seem most affected by the link between BMI and early death.
The study has a number of strengths, including the vast amounts of data from a wide geographical area, and the researchers' use of a standardized protocol to exclude factors they thought might have confounded the results.
However, the use of BMI as a measure does exclude other factors that might have been important – for example, the fat to muscle ratio or the distribution of fat. People who carry fat around their waist (as many men do) are thought to be at higher risk of health problems than people who carry fat on their hips (as many women do).
Using BMI alone also means we don't know about people's overall health-related habits. For example, high BMI could be a sign of doing little exercise, or eating an unhealthy diet, both of which are likely to shorten life.
This means we cannot say that higher BMI is a cause of early death. But the study results make a strong case for higher BMI being linked to early death, across several geographical regions, among men and women, and at all levels of BMI.
It's worth noting that being underweight is also strongly linked to higher chances of dying early. The researchers found that even those at the lower end of the WHO's "healthy" spectrum – with a BMI of 18.5 to less than 20 – were at increased risk compared to people with a BMI of 20 to 25.