The New York Times reports:
The study randomly assigned 5,145 overweight or obese people with Type 2 diabetes to either a rigorous diet and exercise regimen or to sessions in which they got general health information. The diet involved 1,200 to 1,500 calories a day for those weighing less than 250 pounds and 1,500 to 1,800 calories a day for those weighing more. The exercise program was at least 175 minutes a week of moderate exercise.
But 11 years after the study began, researchers concluded it was futile to continue — the two groups had nearly identical rates of heart attacks, strokes and cardiovascular deaths.[…]
The study participants assigned to the intensive exercise and diet program did lose about 5 percent of their weight and managed to keep it off during the study.
I’m impressed they did an 11-year study; it’s rare to see any weight-loss based study continue beyond six months, frankly, which is to say that most weight-loss studies are completely useless.
The next thing that occurred to me is that five percent weight loss isn’t really what the diet and exercise advocates promise, is it? Paul Campos had the same thought:
Note that, perhaps out of necessity, the study defined “significant long-term weight loss” quite modestly, since on average this intensive lifestyle intervention produced a little less than a 5 percent reduction in body mass (i.e., a 160-pound “overweight” woman in the program weighed, on average, 152 pounds after four years of participation). In other words, what the study proved yet again is that lifestyle interventions don’t produce much, if anything, in the way of long-term weight loss.
My third fault is wondering what this means to me, as a person with type two diabetes? I’m on the medication, and don’t find it unpleasant (I just seem to be a lucky person who doesn’t get side effects from Metformin). I kind of like exercise; even if it doesn’t make me live longer, it’s somewhat enjoyable and makes me feel better, and may offer benefits as I get older like extending my mobility. But is trying to watch what I eat worthwhile at all? Because I gotta tell you, it’s not fun.
Via.
Agreed. Not fun.
Diabetes runs on both sides of my family, but I am blessed not to have it myself (knock wood). I’m not surprised by this study. But honestly, I’m kind of sad about it. I mean, all that work for nothing, or almost nothing?
According to one model of medicine I am at risk, because of my height/weight ratio – 5’6″, 230lbs. However, I eat well, exercise, am an active member of my community, maintain friendships, and follow a spiritual practice. My weight hasn’t fluctuated more than ten pounds in the last ten years, and I consider that a win. I know from experience the stress of yo-yoing is a bigger deal than maintaining and doing all the above.
So, keep the faith and keep on doing whatever you’re doing that feels right. It sounds like that’s all you can rely on, if I am interpreting this study correctly!
Well, there are two different ways to “watch what you eat” that unfortunately get conflated all the time: you can watch what you eat, and you can watch how much you eat. Evidence for the latter as a means of either increased health or maintained weight-loss is basically nil for all but the most severe and lifelong of restrictions.* But as for what you eat–that can make a big difference even if the number of calories is the same. I don’t have diabetes but I do have genetic insulin weirdnesses and I know that I feel much better when I’m restricting my sugar intake: more energetic on average, definitely more evenly energetic instead of spikes and down times, fewer food cravings, less digestive trouble.
I think, actually, that this is the worst part of the diet industry: it erases the difference between health and thinness. You can do things to make yourself healthier that don’t involve weight loss at all, and you can do things to make yourself weigh less that make you manifestly unhealthier–but far too many people, including doctors, have been convinced that the number on the scale has as much effect on their health as the numbers that actually matter (like hemoglobin A1C…).
*I remember an article from some researchers who tracked down people who’d kept off significant amounts of weight for a long period of time, and one retired person basically said, this takes so much effort that I don’t think I could do it if I was still working. And even that was just for maintained weight loss, not health, because the numbers of people who manage it are too small to draw conclusions about their health relative to people who never lost the weight.
This is a good point. There are also two different ways to watch *what* you eat. You can either moralize about it and have “forbidden” foods (which you then crave), or you can give yourself permission to eat whatever you want, and then pay close attention to the consequences, good and bad, to see if it’s worth it to you. (The second approach is called eating competence–Ellyn Satter and Michelle Allison have tons of info on this.) I know Barry knows this, but for everyone else’s reference: The Fat Nutritionist.
Having an idea what foods have what effects on your body is only part of it. You also have to have an overall relationship with food that’s sane and doesn’t trigger feelings of deprivation.
Eat what you want and die like a [fully-privileged member of your society].
I wrote about this recently.
One thing that’s interesting about this is that the calorie intake of the people in the study is actually pretty comparable to what some long-term radical weight loss maintainers do. (They seem to be exercising a bit less and eat less calories.) But they’re “only” maintaining a 5% loss. After I wrote that I came across something (here) talking about the Minnesota Starvation Experiment, and realized that they’re also eating slightly less food than the people in the Minnesota Starvation experiment, and the people in that experiment lost about 25% of their original weights on average, not 5%. Is this because it’s even harder for diabetics to lose weight than other people? Is it because of how long the experiment went on? If the diabetics were mostly female (I don’t know) could that account for the difference? (The last two would account for the difference from the Starvation Experiment comparison, but not for the radical maintainer comparison.) Individual differences in response to weight loss attempts can’t account for the big average difference in weight loss.
Edit: fixed HTML
Caution: unsupported speculation ahead…
Stress hormones (i.e. cortisol) raise blood sugar. So obsessively watching what you eat, especially if it’s not fun, may be doing more harm than good. Do you have a decent endocrinologist who can talk to you about this sort of thing?
It seems hard to believe that more than a decade of “rigorous diet” of 1200-1500 calories a day, and regular exercize didn’t cut more than 5% of their body-weight. I wonder if that diet was rigorous on paper, or in practice, human beings aren’t perpetuum mobiles thus if calories in are less than calories burned they *will* loose weight.
But lots of studies show that it’s really hard to get people to change their calorie-intake significantly over long periods. “They where unable to follow the diet” is very different from “they followed the diet but it did not work”.
Interesting post, Amp. I find this news very confusing (as I suspect many people do). I think most of us are familiar with the notion that sedentary people are at high risk for a number of adverse health conditions, including “heart attacks, strokes and cardiovascular deaths.” I’ve seen a flurry of references of late to the particular risks faced by people who spend an extended amount of time sitting, with some noting that people’s life expectancies are reduced by 22 minutes for every day (or whatever; I don’t remember the exact metric) spent sitting.
Assuming that these studies and the study you’re referencing have both been done in a scientifically sound manner, I’m trying to come up with possible explanations that would be consistent with both:
1. The ‘sedentary studies’ were just classic causation/correlation confusion. Sedentary people do have worse outcomes, but that’s because of factors unrelated to their being sedentary. They have ‘bad’ genes or maybe a sedentary childhood that produces irreversible risks later in life, or to put it in a slightly different way, people who exercise tend to be people who are healthy.
2. Exercising does produce real health benefits for most people, just not for overweight diabetics.
3. Exercise alone would have produced health benefits for the overweight diabetics in the study if they hadn’t been simultaneously subjected to a weight loss regimen which induced stress hormones and subsequent blood sugar ‘turbulence’ (as per Dianne’s speculation).
I want to be clear that I haven’t been following this issue closely at all, and I’m just relying on a casual memory of news reports that I’ve seen. With that caveat, I don’t think #1 can be (entirely) correct because I think I’ve seen references to studies that show improved health outcomes for people who change from sedentary to somewhat more active lifestyles. However, it would be very interesting if this is just the impression created by the way correlation studies have been reported, and that in fact there has been no study showing a benefit from such a change.
Too many diet advocates still believe in the above myths – and that weight is a simple matter of input and output. But real human bodies are far more complex systems. From the New England Journal of Medicine (emphasis added):
It’s spelled “lose,” by the way.
It really *is* that simple: If your intake is less than your consumption, your weight will go down.
But it does not follow that slightly reducing calorie-intake will reduce weight, because as you point out atleast some people have a metabolism that tends to adjust, so when they eat somewhat less, the result is that their calorie-consumption *also* becomes somewhat less and the end-effect is no change in weight.
But in this study they also exercized 3 hours a week more than before, and the number of calories in their diet is very low, especially for the men, and metabolism-adjustments still aren’t magic, a 240 lbs man who exersize for 180 minutes a week and eats 1500 kcal/day without losing weight seems highly unlikely.
I’m going to assume you mean “if your output is less than your consumption,” here.
Right, but we assume that we control all aspects of both of those variables, which we don’t. You don’t consciously control how much energy your body extracts from a given amount of food, your body temperature and energy levels, or what each specific calorie gets used for.
We act like each bit of fat or protein or sugar can only either be burned or stored. In fact, the body is slightly more complicated than a bunsen burner. Fat might be stored, it might be converted to energy, or it might be used to insulate neurons, transport vitamins, or repair cell membranes. Protein, similarly, rebuilds muscles and cell membranes.
So, yes, if you could accurately measure the energy your specific body can take from a specific bit of food, you could identify the amount at which your body would be forced to lose weight. You don’t get to decide that the weight lost should be fat rather than muscle, and you also don’t get to decide which functions the body will ignore or perform non-optimally in order to conserve fat stores.
Only if you assume that food and exercise are the only variables. What’s their body temperature? How much are they sleeping? Are they losing muscle while gaining fat? There are a lot of functions that a body can perform suboptimally to conserve fat stores, especially when you put it under famine conditions. That’s a survival mechanism.
http://danceswithfat.wordpress.com/2011/03/06/the-calories-incalories-out-myth/
http://danceswithfat.wordpress.com/2011/06/02/calories-incalories-out-science-says-no/
As it happens, we are able easily and fairly accurately measure how many calories a person burns, doubly-labeled water methods are accurate to within about 3% and that is more than good enough for these kinds of experiments.
It’s trickier to know how many calories a person has eaten. Unless the person lives in a jail, and eats only carefully prepared food in carefully measured portions, the error-bars will be huge, and I suspect they’ll tend to systematically go in the direction of under-reporting. (people will more often forget they ate something than they’ll mistakenly report that they ate something when they did not)
You are ofcourse entirely correct that we cannot control what kind of mass is lost – in principle you could successfully eat less than you burn, and end up losing muscle while keeping the fat. In that case you’d notice your weight dropping -more- than expected though, because a kg of muscle has much less energy than a kg of fat does.
Listen, I totally agree that most diets are completely useless, and that a person can go on a diet, eat less, yet remain the same weight.
But they had over 5000 participants – and 5% was the *average* loss for the strict-diet group. That is flat out unbelievable. My bet would be that a huge fraction of the participants did not actually stick with anything close to the prescribed diet – because that *is* very very hard, and something that most people can’t do. (if most people could cut 1000kcal/day from their diet while exersizing regularly, we would not be having the obesity-rates we actually have)
Amp, the headline is a little overblown.
Also, we’re looking at an NYT interpretation of the study – not the study itself. That said, what the article reports is not that diet and exercise “didn’t do squat”; it reports that it didn’t significantly lower the risk of certain cardiovascular problems. (I’m also a little concerned at the conflation of “overweight and obese”.)
It would also not surprise me in the least if one of ballgame’s points is correct: healthier people are more prone to exercise.
YES! Can’t believe no-one has mentioned this yet, but there’s more to health as a diabetic than heart attacks and strokes and cardiovascular deaths and ******* weight loss. Poorly controlled blood sugar leads to very serious complications: low blood sugar can lead to short-term problems (ketosis, coma etc.) and high blood sugar contributes to long-term problems, including retinopathy (leading to blindness) and neuropathy (which can lead to amputations). Sorry to seem scary but the longer you have diabetes, the more the risk of the long-term complications gets more and more significant.
Watching what you eat, in particular the carbohydrate content, is a huge contribution towards keeping blood sugar under control, and exercise is a hugely useful tool as well, particularly in Type 2 diabetics who are dependent solely on their pancreas for insulin supplies (which may be inadequate). Exercise mops up blood sugar immediately, reduces appetite, and lowers insulin resistance for some hours after the exercise itself. That’s quite apart from its other fitness and health benefits.
Recommended: the Blood Sugar 101 site (can be found by googling).
ballgame @8: a sedentary lifestyle and 3 hrs/week of moderate exercise are not inconsistent with one another. If you sit at a desk for eight hours a day and don’t move, then get up and walk or jog for 35 minutes a day after work, you’re still living a sedentary lifestyle; the recommended way to combat it is not to exercise more in discrete chunks, but rather to make sure you stand up and walk for a couple of minutes at least once per hour.
Can’t find the articles just now, but I’m pretty sure some of the sedentary-lifestyle research was done in a way to probe causation, by taking a group of healthy people, splitting them in two, and making half the group sit for a lot of the day for some period of time, and then measuring various health indicators.
Gunnar @12:
To maintain weight loss, studies generally find that people have to eat ~20% fewer calories than people of the same weight who had never been heavier. (Here’s the most reliable article about that research that I could find during a quick search although it’s about 10 years old and the first part is not really about that.) So it’s not a surprise to me that a caloric intake we would associate with weight loss doesn’t continue to produce weight loss after some significant decrease has already been achieved (again using the usual medical definition of significant, ie 5-10% of starting body weight).
Also, your last sentence isn’t evidence, since it relies on the same belief that makes you object to the weight loss stats in the first place. That being said, if this was self-reported food intake (and it would almost have to be for a study of this size) then yes, there are known biases with self-reported numbers…but, I suspect, still not nearly enough to make them indistinguishable from the non-strict-regimen group.
[edited:getting rid of spare adverbs]
Sure, initially. But muscle mass is one of the many factors that affects BMR. Lose muscle tissue, and that drops too.
There have been smaller studies done where food intake was directly observed, though. Before assuming that the problem is with self-reporting, it would be worth checking to see whether the average weight loss matches what’s been reported with smaller, more controlled studies. If it’s similar, then it’s very unlikely to be a case of “Oh, they’re just eating way more than they realize.”
It would also be worth identifying *how much* self-reporting is likely to be off. If, say, the average dieter underestimates their caloric intake by 300 calories, a 240-lb person eating 1800 calories per day “should” have a 600 calorie daily deficit and lose a pound every 5 days, or 73 pounds in the first year alone. Even once they drop below the point where 1800 calories is a deficit, “calories in-calories out” would predict that they could maintain somewhere around 180 lbs.
So for dieter error to explain the results, they’d have to be off by a lot more than that.
Ballgame wrote:
One thing that may matter is how outcomes are being measured. A lot of studies will measure for improvements in risk factors, such as blood pressure and cholesterol levels, and if those numbers improve that is reported as an improved outcome, and a statistic for improved mortality can be inferred from those numbers. An advantage to doing a study this way is that it’s a lot less expensive, since you can measure for improved blood numbers in six months or a year, unlike measuring for mortality, which generally involves having to study the same people for a great many years.
However – as this study the Times reported on seems to show – sometimes the connection between the blood numbers and long-term mortality may not be as clear-cut as the researchers assume.
Mythago – Agreed. I try to avoid hyperbole, but I make an exception for post titles.
The thing is, it’s hard to take seriously “refutals” which throw around words like ‘science’ yet obviously lacks even basic knowledge of results that are old and well-established. For example, the blog-post above titled “Calories in/Calories out – science says no” contains nuggets such as this:
“First, it turns out that accurately calculating the calories out side of the equation is at best an awfully indirect science producing questionable results.”
Which is simply false. You can easily and accurately measure the calories-out side of the equation by something as simple as letting each participant drink a cup of doubly labeled water, then do isotope-analysis on a sample of their urine or blood or saliva or whatever one day later and 15 days later – basic math will then tell you how many calories they burned in the intermediate 14 days. The method has been known since 1950, and is accurate to within +- 3% or thereabouts.
Yes it’s hard to lose weight and keep it off. The reason is not that the human body violates basic physics.
I hadn’t heard of that method, so that’s good to know. It’s also not super-relevant to the study in question, or to general conversations about calories-in/calories-out unless that method was used on the participants in the study. Ideally every two weeks on everyone, but at least a few times throughout the study on a reasonable sample to verify that the calorie numbers they were given were appropriate, and that they remained appropriate as their BMRs dropped—which is known to occur when people are on diets. I see no evidence from the NYT article that this was the case, but if the study itself says otherwise, please, let me know.
Your assumption that they should have lost more weight and were therefore erroneously reporting what they ate relies on the assumption that their calories used didn’t drop enough to make up the difference.
Also, it’s not “violating basic physics” to say that there are more variables that are commonly accounted for. You can’t always, as the physics major joke goes, say a cow is a sphere to make the math easier.
The actual study ain’t available, so it’s really hard to say what they even tested, much less what should have happened.
If you put a large (5000+) group of overweight people on a 1200 – 1500 calorie diet and 3 hours of moderate exersize a week, and they actually stick to that program, you’d expect average weight-loss of about 2 pounds a week. If actual results differ significantly from that, there’s only two possible explanations.
Either these people (not a few “oddities”, but the *average* person) have metabolisms that are very different from the norm.
Or else, for whatever reasons, essentially all of them failed to stick with anything resembling the plan.
The former could easily be tested with doubly labeled water. I know this study didn’t do that, what I’m saying is I strongly suspect if they had, they’d have found a burn-rate significantly higher than the calorie-intake. (and thus either these people on the average have magical bodies, or else the self-reported intake is inaccurate)
I guess I would say that watching what you eat is worthwhile if it improves your quality of life and your sense of well-being. When you have diabetes, you have kind of a special opportunity to find out how food affects your body, by measuring your blood sugar directly, or even just feeling the sensations of your blood sugar changing. At least for those reasons, I would think that doing things to eat well would be worthwhile. However, “doing things to eat well” might look quite different from “going on a strict diabetic diet,” and also it might look different from one person to another. Diabetes is really individual, and food affects different people in strikingly different ways. Some people seem okay just doing basic stuff like eating meals at regular times, eating carbs mixed with protein, fat, and fibre, and taking their meds consistently. Other people have to be a lot more careful because their blood sugar is more reactive, or because they have a different medication or insulin regimen. You didn’t really ask me, but I would say that regardless of what this study says (and I am looking forward to reading the full text when it comes out), you should look at how your body responds, and how you FEEL when you eat, and then do what seems best for you.
My understanding is that the body metabolizes food super-efficiently when it needs to; there’s something in the range of a 20% uptick (i.e. if you normally eat 2000 kcal/day to stay level weight, your body can actually learn to sustain the same weight on something like 1600 kcal/day.) Obviously this means that relatively minor changes in diet won’t tend to do a whole lot for weight.
But nonetheless, if people metabolize fewer calories than they burn, they will lose weight. If you combine
-[low calorie diet, purportedly low enough to lose weight] and
-[long term study] and
-[nobody is losing weight]
then something appears to be wrong in a “how does that make sense?” kind of way: either the diet kcal aren’t low enough, or the results are short term, or possibly there’s some other issue, whatever it is. Or, possibly, the study is fine but the information is inaccurately reported (not like THAT ever happens, right?)
This can have a lot to do with the underlying study design and statistics. For example, it may be that a 251-pound man who starts eating 1500 kcal/day has a very different %age change from stable diet than does a 125-pound woman who starts eating 1200 kcal/day.
For instance, maybe they’re all overweight and obese, and the typical o/o person doesn’t have the metabolism of a “typical” non-o/o person.
Or maybe they’re all diabetics taking strong medications, and in some way the typical diabetic on medication doesn’t have the exact same metabolism response as that of a typical non-diabetic.
Or maybe they’ve all been on low-calorie diets, and the human body responds to low-calorie diets by altering its metabolism.
Or maybe they’ve all been on low-calorie, moderate-exercise diets for over a decade, and the human metabolism adjusts to that in ways that short-term studies of diet and exercise are unable to reflect.
Or maybe the participants are lying about their food intake, and the researchers, despite being some of the most prominent weight-loss researchers in the country, somehow never considered that factor and didn’t do any testing of the validity of participant self-reports. It’s possible, although we won’t know until the study comes out.
Or maybe it’s a combination of factors.
None of the above ideas are “magic,” and that you jump to that kind of sneering is not only annoying (because you’re being condescending), it only makes sense if we assume that we know everything and there’s not any possibility of our beliefs being changed by new data.
The fact is, long-range studies like this are extremely rare. We don’t know much at all, empirically, about how the body’s metabolism responds to long-term calorie restriction. I’d be very cautious about assuming that studies of people who may or may not be overweight, who may or may not be diabetic, and who probably have not been on a program like this for a decade or more, have results which are directly applicable to this sample.
Also, not that it’s relevant, but doubly labeled water has been used on humans since the early 1980s, although it’s true it was used on small animals since the 50s. When applied to individuals, some studies have found that DLW isn’t as accurate as you’re claiming (pdf link):
I agree with the plausibility all your alternate considerations, Ampersand.
Another reason I don’t find, “Maybe they’re all just lying!” very compelling: in most studies like this, the subjects get more monitoring of what they eat and more help/encouragement/etc. in sticking to their plan than most dieters do. If they’re still unable to stick to the plan with all that additional help, then I think it’s not realistic to expect others to follow it. If these researchers (who surely are more familiar with the likelihood of underreporting than we are) couldn’t design a fair test of a program for weight loss maintenance, should we be making excuses for them? Or should the lesson be that it does not work, whether the reason for not working is that sticking to the program doesn’t work or that basically no one could stick with the program?
Sure, it’s interesting from a scientific point of view whether the problem was with the program not working when followed or that people couldn’t follow the program. But it’s not like one makes the program a failure and the other doesn’t.
Guys, your hypotheses are so exhausting. It’s way easier to just believe that fatties get fat because they eat differently than thin people, and to stigmatize them as noncompliant (as well as awkward, ugly, and unattractive). I mean, “Fat people lie about what they’re eating” is such an obvious explanation, why even do science about it?
“Eat Right, Stay Fit, and Die Anyway”
–George Carlin’s “Book Club”–