Thursday June 25 2009
The study does not suggest obesity has health benefits
The Daily Telegraph reported that “a few extra pounds helps you live longer.” According to the newspaper, new research shows that overweight people live longer than their slender peers. It said that results defy the common belief that staying slim is the secret to a long, healthy life.
While Canadian researchers have found that a group of overweight people had a lower rate of death than people in an ideal weight group, they stress that their research should not be interpreted to mean that slim people should gain weight to get into the overweight category.
The reason for this apparently contradictory finding is unclear. As this was an observational study, it is impossible to say what caused the association. Also, the study did not measure other factors, such as socioeconomic status and fitness, that could have been involved in the association. Despite news reports on this study, people should aim to maintain a healthy weight through exercise and good dietary choices.
Where did the story come from?
This research was carried out by Heather Orpana, a social scientist and researcher from Statistics Canada in Ottawa, and colleagues from elsewhere in Canada and the US. The study was supported by grants from the National Institutes of Health in the US, Aging and Diabetes and Digestive and Kidney Diseases, plus a research grant from the Canadian Embassy in Washington. The study was published in Obesity, a peer-reviewed medical journal.
What kind of scientific study was this?
This was a longitudinal/cohort study that analysed survey data and death rates in a Canadian population.
The researchers explain that a clear risk of mortality is associated with obesity (defined by the World Health Organization as a Body Mass Index, or BMI, of 30 or over), but that the risk of mortality associated with being overweight (BMI of 25 to 30) is ambiguous. They wanted to test the link between BMI and death from any cause in a sample of Canadian adults.
The researchers say that studying this weight group is especially important since the number of people who are classified as overweight is rising throughout the world. It is well known that obesity is linked to increased rates of type 2 diabetes, heart disease, high blood pressure, stroke, gallbladder disease, some forms of cancer, osteoarthritis and psychosocial problems. It is also a risk factor for early death.
The researchers analysed data from the National Population Health Survey, a study that has been conducted by Statistics Canada every two years since 1994/5. In this survey, follow-up data was available up to 2006/07. In total, 17,276 members of private households were selected for analysis from the 1994/5 survey. One participant from each household selected was asked to participate and 86% of them agreed to take part.
Participants who were over 25 (12,455 people) were included in the present analyses. The researchers excluded 109 women who were pregnant and those with missing data on BMI or smoking status. The final sample size was 11,834 people.
Any deaths were counter-checked against the Canadian Deaths Database up to 31 December 2005. Deaths after this date could not be confirmed, but the researchers say that the reporting of deaths before this date had been accurate.
The participants reported their own height and weight, which were then used to calculate their BMI. The researchers analysed the data using models that allowed them to adjust for various other factors that could also affect the risk of death, such as age, sex, self-reported smoking status, physical activity frequency and alcohol consumption.
What were the results of the study?
Approximately equal numbers of men and women were involved and more than half the sample was under 45 years of age.
Within the studied sample, 1,929 deaths were observed during the follow-up period of 115,225 person-years. (Person-years is a measure calculated by multiplying years of follow-up by the number of people followed and offers a better comparative measure of rates in studies which follow many people over time.)
In the models that adjusted for socio-demographic factors and health behaviours:
- The risk of death for people who were underweight was increased (relative risk [RR] was 1.73, 95% confidence interval [CI] 1.25 to 2.39).
- The risk of death for those with a BMI of 35 or over (obesity class II+) was also slightly increased (RR was 1.36, 95% CI 1.00 to 1.85).
- Those who had a BMI between 30 and 35 did not show an increased risk (RR was 0.95, 95% CI 0.72 to 1.18).
- When compared to those in the normal weight category, overweight individuals (BMI from 25 to 30) had a lower risk of death (RR was 0.83, 95% CI 0.72 to 0.96).
What interpretations did the researchers draw from these results?
The researchers say that, in the Canadian population studied, there was a significantly increased “risk of mortality over 12 years of follow-up among individuals in the underweight and obesity class II+ categories”.
They say that being overweight was associated with a significant protective effect compared to those in the ideal weight category.
What does the NHS Knowledge Service make of this study?
This study collected survey data from a large number of people and followed them up for 12 years on average. It adds to the information available regarding the degree of risk for people in different weight categories and has the advantage of being a longitudinal study of a population over time.
In some respects, the follow-up time of 12 years was relatively short. With a high proportion of younger people in the survey, it is likely that the causes of death were those that lead to premature mortality, rather than weight-related lifestyle illnesses that might cause deaths in a selection of older people. As cause of death is not reported, it is not possible, from this study, to say more about a relationship between disease, weight and death in this population.
There are several other limitations which should be considered when interpreting the data:
- Due to the sampling design, the participants were similar to the people in an average Canadian household but may not necessarily represent those of other countries.
- The researchers collected a limited number of “control variables” (factors in the sampled group that could also potentially increase the risk of death). For this reason, the researchers say that caution should be taken when inferring that their results prove that being overweight reduces the chance of death.
- Height and weight were collected through reports from the participants, but it is accepted that respondents have a tendency to underestimate their weight and/or overestimate their height. The researchers say that they used a correction factor developed by Statistics Canada to adjust for this, but the workings of this correction model were not reported in detail.
Overall, this large study was unable to control for confounding factors – those other aspects of life that may influence the link between weight and early mortality. A better understanding of how weight acts through socioeconomic, fitness and other intermediate factors is required before firm conclusions can be drawn from this type of study.