So earlier today a friend asked me a very simple question: is being “fat” bad for you regardless of how much you exercise?
It has a simple answer: we don’t know.
There’s good evidence that very obese people are at greater risk of early death than people with “healthy” BMIs between 18 and 25. There’s some evidence that the excess risk is greater among young people, that is, that very obese 40-somethings are increasing their risk more than very obese 70-somethings, but this is a little shakier.
I’m reasonably certain the review I’m citing here made the news when it came out – I recall hearing that overweight people were actually at lower risk of death than “healthy” BMI people. I hadn’t heard, or remembered, about the distinction the authors discovered between people who are grade 1 obese (BMI 30-34) and those who are grade 2 or 3 obese (BMI 35+). It turns out that all the risk associated with obesity seems to be concentrated in the grade 2+ group, while the grade 1 group are at the same risk of death as people in the “healthy” range.
So it seems like you can be pretty big before you’re at higher risk – the table below shows what weight you have to be to reach the grade 2 obese group at various heights (you can calculate your own BMI here). And once you’re there, the increase in risk isn’t that substantial – it’s about a 30% increase in your risk of death in a given year, compared to people the same age and sex as you. So if you were an average Australian aged 40-44 in 2013, your risk of death that year was around 1 in 830. If you were grade 2 obese or larger, this would have risen to 1 in 640.
|Height / BMI||20 (“healthy”)||25 (overweight)||30 (grade 1)||35 (grade 2)|
Now, to the second part of my friend’s question, “regardless of how much you exercise”. Here’s where things get tricky.
Some people become obese in spite of already being quite active, and within the “obese” categories there are a wide range of body types. Most professional rugby players are “obese” according to BMI. Other people with BMI in the “healthy” range actually have quite high bodyfat percentages. Some people in the “obese” BMI range can easily lose excess bodyfat through light intensity but high volume exercise (e.g. by walking an hour a day). Some can’t. (And some, like the fit rugby players, haven’t got any to lose.) Some obese people have reasonably healthy diets, some don’t – and of course some people with “normal” BMIs live on beer, pizza, and cigarettes. All of this affects their risk of serious health events, including fatal events.
Here it’s important to draw a distinction between health and fitness. “Health” is a difficult thing to define, but when we talk about the health impacts of obesity we’re generally talking about reasonably hard outcomes – your risk of cardiac events, cancers, or premature death. “Fitness” on the other hand generally refers to cardiovascular fitness, which at one end of the scale is whether you’re capable of doing things like climbing a few flights of stairs or running for a bus, and at the other whether you’re likely to place in the Tour de France. In a sense, my friend’s question is whether or not these things are interchangeable – if you’re obese, but you’re no more out of breath on the fifth flight of stairs than someone half your size, can you really be said to be less healthy than them? If you’re a normal BMI but you have a high bodyfat percentage and your cardiovascular fitness is poor, are you at higher risk of early death than a fitter but much larger person? We don’t know.
In order to establish the relationship between obesity and mortality, the authors of the review I just mentioned analysed over 100 studies with a combined total of 2.9 million participants, of whom 270k eventually died. That is a lot of studies, a lot of people, a lot of deaths, and a lot of time and money. Even then, because death is relatively rare and this relationship is apparently pretty weak, there was a lot of variation in the results.
But this relationship in particular isn’t too hard to pin down – BMI is easy to measure, and although many people steadily gain weight through adulthood, people don’t tend to swing back and forth wildly between obese and underweight from one year to the next. Someone who is obese when you measured them at 45, and again at 50, and again at 55 was probably still obese when they died at 57. So you can be confident of who’s who in your study – who died with a high BMI, who died with a low BMI, and who’s still alive. The question is whether BMI itself is a useful measure of true risk – is everybody who’s grade 2 obese at the same excess risk of early death, or is BMI a poor proxy indicator for that risk?
Now, firstly, as I noted above, not everybody who is “obese” according to BMI actually has high bodyfat, but bodyfat is a lot harder to measure than BMI. As a consequence, the bodyfat percentage of most of the people in these studies is unknown, but probably there were people in both groups who had better or worse bodyfat than you might expect based on their BMI. As such, although you’ve got everybody’s BMI right, you might have got their risk of death due to their weight wrong – the healthy rugby players are dragging down the mortality in the “obese” group, while the sedentary, high bodyfat smaller people are dragging up the mortality in the low BMI group.
Another difficult thing to measure accurately is exercise, or of fitness, which is what we expect the relevant result of exercise to be. If you tell me you are 5’10 and 85kg, I can measure you to check that a minute later. If you tell me that you go to the gym three times a week, I have no way of verifying that without a lot of effort. And even if it’s true when you tell me that when you’re 55, I have no idea what you do in the intervening years until I see you again. Maybe you got the gym membership as part of a New Year’s resolution, and you’ll have stopped going next month. Maybe you’re about to get into marathon running and by next time I see you you’ll be twice as fit as you were. Anything could happen.
Barring lap-band surgery, your weight is not likely to change very radically between 3- or 5-year study follow-ups, but your exercise habits and consequently your fitness could be all over the place, and I would never know. If I don’t know, then I don’t know whether or not you were exercising regularly before you died, and so I don’t know whether I should count your death as a death among an exercising fit person or a sedentary unfit person. As such, I can’t tell whether obese people who exercise are at lower risk than obese people who don’t, because I don’t really know who’s who.
In order to accurately understand the relationship between being “fat”, being “fit”, and early death, we would need large numbers of large studies which accurately and regularly measured people’s bodyfat (not just their BMI), and their exercises habits (what they actually do – not just what they say they do). With the technology for both these types of measurements improving significantly, we may well get them eventually – fitbits and other digital activity monitors are getting cheaper and cheaper, and they can send information to research teams in real time, without us having to contact participants. Likewise, the technology to measure bodyfat will probably get easier, cheaper, and more accurate as time goes on. But death is rare, and so these studies will take decades to produce results.
Until then, we know that being grade 2-3 obese on the BMI scale is associated with increased risk of death. We don’t know whether that risk is concentrated in obese people with high bodyfat, or whether everybody with a BMI over 35 is at the same excess risk. We also don’t know whether people with lower BMI but higher bodyfat are at more risk, or whether people with high bodyfat but good fitness are at less risk. It seems reasonable to suspect that they are – but we won’t know for certain for quite some time yet.