Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Disclaimer: The Parliamentary Library does not warrant or accept liability for the accuracy or usefulness of the transcripts. These are copied directly from the broadcaster's website.
Rethinking BMI and mortality -

View in ParlViewView other Segments

Norman Swan: On last week's Health Report we had a soothing story for anyone carrying a few extra kilos. Danish researchers looked at the Danish population over 30 years, comparing their body mass index to the point at which their risk of dying prematurely was at its lowest. And they found that the BMI with the lowest risk had shifted upwards by over three points from just under a normal BMI to the overweight range. And in fact when you looked at the BMI of 30, which is obesity, there was no added risk at the beginning of the obese range.

Now, my story infuriated Dr Rosemary Korda, an epidemiologist and chronic disease researcher at the Australian National University, and Rosemary's ready to vent in our Canberra studio. Welcome to the Health Report.

Rosemary Korda: Thank you, Norman. That's strong words!

Norman Swan: Just, you know, opening it up for you! But you don't believe this study.

Rosemary Korda: Well, I think there's little dispute that obesity disposes us to poorer health outcomes. We know that it increases heart disease, diabetes, stroke, some cancers, and consequently higher mortality rates. I think the issues around whether there's an association with…when you look at the association between BMI and mortality, some studies have found that in fact mortality's lower in the overweight BMI range than the normal range, as this Danish study did. And many researchers, including myself, argue that these findings really come about because of methodological issues rather than the fact that there's really this what they call the 'obesity paradox'. And I guess that importantly we shouldn't get too excited about one study. We need to look at all the evidence together. We need to look at large numbers to tease out the different effects and find out what factors apart from BMI might be affecting this relationship.

Norman Swan: This isn't the only study to have found it. There was one by a woman called Katherine Flegal which suggested that when you looked at obese people in the United States over 30 years they did seem to be getting healthier. So it's not just one study.

Rosemary Korda: Well, in terms of getting healthier, we're all dying at older ages, so that goes across the spectrum. So whether you're in the normal range or the higher range death rates are going down, you can see that. It's about the relative death rates we're talking about here.

And I think the most reliable evidence really comes from large studies where they've pooled individual level data. So you've got pooling of large prospective studies, because again, you need large numbers to tease out the real effects from some of the confounding effects.

Norman Swan: So just to explain this, the Danish study was a cross-section 30 years ago, then another two cross-sections over a period of 30 years rather than the same people in 1970 followed through to 2013.

Rosemary Korda: Yes, some of them overlap in the early cohorts. Although with each cohort they did follow them over time. So for instance the last cohort they followed for about four and a half years on average, and then looked at the death rates in people who were normal weight, overweight, obese. But when you actually pool the large data together, and there are two international studies that have been published in the last few years, one in The Lancet which looked at over 900,000 people they pooled individual data for. Another in The New England Journal of Medicine where they pooled data from one and a half million people. And what happens when you do that, you can see—and you control very closely for things like smoking and whether you have pre-existing illness—is that mortality is lowest in the normal BMI range, not in the overweight range. And I think very clearly it shows that as your BMI goes up from the normal range you get a gradual increase in mortality rates.

Norman Swan: So how did they get it so wrong?

Rosemary Korda: Well I think, again, it's not the only study to show what they again refer to as this obesity paradox. And I think it's a complex issue. And because there are so many other factors affecting mortality apart from body mass index, you actually need large numbers of people to tease out these other effects. You need to control very closely for age, you need to control very closely for pre-existing illness. You have to be very careful about the smoking effect. We know smokers tend to be leaner and die earlier, for instance. We know…

Norman Swan: So that could give you a bias…

Rosemary Korda: You can get biased results. People who are chronically ill lose weight, so you need to account for the fact that people could actually be losing weight and getting sick rather than the other way round.

So there's a lot factors, and I just touched on a few. And the problem is, if you take one study it can be very hard to tease those out, because you need very large numbers. Hence why these large, prospective what we call pooled data studies are very powerful, because they can bring data together for a lot of people.

There's another one coming up in The Lancet in a forthcoming journal which is under embargo, but that actually includes data from 239 studies across four continents, 4 million people, and they show very much the same results, that in fact the BMI associated with the lowest mortality is in the normal weight range.

Norman Swan: That means I've got to give up on the danish on the basis of the Danish study. So back to the diet.

Rosemary Korda: Again, it's just one study.

Norman Swan: Rosemary Korda, thanks for coming on to correct the record.

Rosemary Korda: Thanks for having me.

Norman Swan: Dr Rosemary Korda, who is an epidemiologist and chronic disease researcher at the Australian National University. And that's the Health Report for this week. If you want to hear more or download a podcast, go to our website, which is got to via Radio National, very popular, and I'm also starting to do video, so why don't you download one of those as well. See you next week.