Associations of body mass index, fasting insulin and inflammation with mortality
nature.com
nature.com
If you look at a bunch of variables, including BMI, and then you remove the impact of well known negative health impacts of obesity (prediabetes/T2D), then you see that high BMI doesn't not correlate well with mortality. IMO, what this indicates to me is that (1) BMI is not a good indicator of obesity in this study. There are many healthy people with low body fat and high BMI, I am one of them. All you need to do is be tall and lift weights occasionally. If you remove the obesity related negative health signals, you also remove obesity(2) Being fat in and of itself is not the issue, the issue is prediabetes/T2D which is extremely reliably caused by obesity, and the treatment for prediabetes/T2D is weight loss.
(Leaving out the discussion: effects to environment, sedentary lifestyle, etc)
That suggests that there is no other material contributor to mortality from obesity[1], which is a somewhat unexpected result.
--
[1] it is possible that there are negative contributions to mortality associated with obesity which could potentially "hide" other positive contributors
Sure BMI can be more or less accurate depending on the person, but for those few who have actually rendered it inapplicable to them it's going to be blatantly obvious. No one will confuse them with being obese.
But he is an athlete with almost no fat (he is not aiming at muscles so he does not have a diet that minimizes fat, he is looking for sheer power and explosivity (gym, martial arts)). He is very built up (muscular) though.
So I wonder how the BMI can give any kind of reasonable indication when the fat level is not taken into account.
Without any consideration for fat BMI doe snot make much sense I guess.
I wouldn’t have been obese before, but I technically am now (with reduced bodyfat %).
Fat is not the only way to be heavy, body builders have very high BMI value yet have very different risk profiles.
Also, I'm still confused. My previous understanding was that bodybuilders are classified as obese without having excess body fat because they fit into that ratio because the ratio doesn't account for exceptions. Not that obesity actually incorporates more than just a ratio.
The point about body builders is that BMI as a metric assumes a fixed muscular profile, so doing (an healthy amount of) muscle training can wildly affect your BMI independently of your fatness.
If is still likely a good measure at a population level as very few people that have BMI > 30 are also fit. But you need to contextualize it to your own case.
For example, tall people tend to have higher BMI, as if 2 is not the right power to raise height to. But you barely need a 4-op calculator to calculate that. Good luck calculating BMI if the power were 2.18 or whatever gives best fit.
Measuring body fat is not exactly hard but much less straightforward and a touch unpleasant. If you're running a large study, it's much easier to measure BMI than anything else. At population level, I'd imagine correlation to obesity is easily 80-90%.
I guess it is also easier for doctors to say "your BMI is too high" rather than "you're too rounded".
` Higher fasting insulin and higher c-reactive protein confound the association between BMI and the risk of all-cause mortality. The increase in mortality that has been attributed to higher BMI is more likely due to hyperinsulinemia and inflammation rather than obesity.`
"available evidence in humans suggests that hyperinsulinemia precedes weight gain (rather than the other way around)"
"Higher fasting insulin and higher c-reactive protein confound the association between BMI and the risk of all-cause mortality. The increase in mortality that has been attributed to higher BMI is more likely due to hyperinsulinemia and inflammation rather than obesity."
It seems almost like a joke, but "hyperinsulinemia and inflammation" seems more or less identical with "chronic overeating", right?
https://www.ahajournals.org/doi/10.1161/circ.128.suppl_22.A1...
[Metabolic syndrome] doubles the risk of all vascular complications in patients often erroneously considered at lower CVD risk because of their normal BMI.
You can't imagine there are a few million people like that out of the billions in the world, and that we might want to give effective medical care to those people also?
> Likewise how do you treat these people to reduce their insulin swings in such a way that they... would fail to lose weight?
Is this an actual question asking about the state of the research, or do you think that it's unimaginable that science could ever be able to directly manipulate people's insulin levels?
I get the impression from most of the top level comments on this thread that some people would be upset if fat people's mortality could be reduced or normalized without weight loss. Like there's an urge to see fat people punished rather than happy and healthy.
The linked article is about an epidemiological measurement often used to flag health risks. And it's pointing out that the real causality is likely due to other factors. I'm pointing out that those other factors are hopelessly conflated with the first anyway, so we might as well keep measuring BMI and treating it as a risk factor.
Are there other things we should measure? Of course there are. Propose a test and let's look at the tradeoffs. I'm just saying let's not stop treating BMI as informative, because it still is.
There's not a good reason to think that those people are not already getting effective medical care. Medical practitioners are interested in their patients' health and longevity. Obesity and good long-term health outcomes have negative correlations. Here is one of many studies:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3963496/
If there is such a population comprising even 1% of the world population (and that figure has almost no backing in any literature unless you include age 18-30 in that cohort, who are broadly too young to have seen longitudinal health effects from obesity affect their health yet), it's unfortunate that occam's razor may impede the quality of their care, but this is very much a hypothetical question without any meaningful evidence to the contrary.
> Is this an actual question asking about the state of the research, or do you think that it's unimaginable that science could ever be able to directly manipulate people's insulin levels?
Uh, that's what insulin pumps and injectable insulin do. We really don't have to imagine it. But beta cells in the pancreas can literally get worked to death in type 2 diabetes, and type 1 diabetics never produced enough (or any) in the first place. So exogenous insulin is administered because without it (and sometimes with it), glucose will acidify the blood and cause neuropathy or ketoacidosis in the worst cases, and it impedes healing, so reduced feeling causes a small injury which heals more slowly to lead to an enormous ulcer and infiltration which the patient still may not feel, until amputation is a necessary lifesaving measure.
Science can directly manipulate lots of hormone levels, but it all has side effects versus letting your body manage itself, most of them profoundly negative.
> I get the impression from most of the top level comments on this thread that some people would be upset if fat people's mortality could be reduced or normalized without weight loss. Like there's an urge to see fat people punished rather than happy and healthy.
I don't think anyone in this thread wants to see fat people punished instead of happy and healthy. Rather, it's that obesity and being happy and healthy are incompatible in the long term, and this study doesn't change that at all. There's no moralizing or proselytizing in this thread, just a "hey dude, this isn't gonna work out in the long run, and you shouldn't fool yourself or others that it is".
If this were a study indicating that lung cancer actually wasn't linked to smoking, but was instead due to some environmental effect, and people were out here saying "wait, wait -- the study says that it was actually elevated levels of blood CO that caused it, but the number of people sucking on exhaust pipes is so low it doesn't move the needle, and this doesn't change COPD, heartattack, etc", nobody would be saying 'it's like there's an urge to see smokers punished rather than happy and healthy".
I used this as an example because the obesity rate in the US is that kind of public health crisis, with side effects that bad. Impotence, amputation, and blindness (among others) are nothing to minimize.
It's not surprising to find these causes are significant factors in obesity-linked mortality, but it is moderately surprising, at least to me, to find that they're the only significant ones.
In recent years there has been a back-and-forth discussion about whether being overweight is bad for your health. Intuitively one might think that being overweight is bad. But in 2015, for example, there was this article, which reflected a story making the rounds at the time:
https://qz.com/550527/obesity-paradox-scientists-now-think-t...
What I take this study to be saying is that obesity in itself is not really the problem, but it is (presumably) highly correlated with high A1C measurements and inflammation, which do indicate a poor state of the endocrine system (i.e., Type 2 diabetes), which can in turn lead to heart disease and stroke if not treated.
I'm hoping that someone who knows more than I do will poke some holes in what I have just written so as to give you a better answer.
That seems like a place where the article should maybe have a citation or two.
This takeaway is basically "the increase in mortality that has been attributed to fluid in the lungs is likely due to a lack of blood oxygenation rather than being submerged underwater".
There exist obese individuals that are not positive for those markers, and there exist non-obese individuals that are.
In a clinic, that might look like more proactively monitoring an obese patient for these indicators but not focusing as much on their obesity if those are doing well. Likewise, it might look like adding these tests to a 5 year physical (or whatever) for non-obese patients, since they appear to be especially suggestive of concern.
Yes, non-obese individuals with type 2 diabetes exist, though they are rare. Obese individuals who are not prediabetic or already have pathological type 2 diabetes are unicorns.
Patients are already having their blood sugar monitored in a clinical setting, especially diabetic ones or ones likely to be prediabetic. And those patients are... obese.
There is no getting around his fact.
Non-obese patients also have this checked regularly. Clinicians frequently ask for fasting the night beforehand to check blood values. They are checking for this. Nothing here is new, even remotely.
However, both elevated CRP and higher fasting insulin are correlated with obesity, so controlling for these variables seems misleading.
Well: https://en.wikipedia.org/wiki/Multicollinearity
Summary: highly correlated variables don't fundamentally mess up the inference (coefficient estimates are still unbiased), the estimation becomes numerically unstable, however, and individual coefficients might be messed up because of that. This is reflected in the standard errors, so you're more likely to miss effects (type II error).
And what if they don't eat that much? They die or what?
Yes a lot of crash diets fail, for obvious reasons, but once a diet has stabilized the cravings will go away.
It does for many people mean they will need to cut out or dramatically cut down the eating out, drinking and instant meals. Most people don't want to do that.
If we agree on this, then your argument isn't that people are destined to obesity, but rather that once they're obese, they have patterns of behavior that are extremely hard to break over a long term basis. That isn't the same as saying that weight loss is a futile endeavor, it's just saying that people who are obese are very often going to fail to stick to the patterns of behavior that resulted in weight loss.
It is not hard for the body to discourage unnecessary caloric expenditure in response to a notable caloric deficit. Indeed, doing just that is a basic famine survival instinct.
Now given the decidedly non-famine conditions, eventually things should restablize, but it does mean for a person at caloric equilibrium, reducing caloric intake by say 500 Calories, will often not result in a 500 calorie deficit, but a smaller one for for quite some time until the body adjusts to this being the new normal.
In the mean time, the person likely feels like shit. Furthermore, If there actually is food available, the body and subconscious mind is doing everything it can to encourage the person to eat more.
This substantial resistance of the body to attempts to run meaningful caloric deficits for an extended period are a not insignificant portion of why "dieting" often fails. People don't stick with it, since it majorly sucks.
Appetite changes likely play a more important role than slowing metabolism in explaining the weight loss plateau since the feedback circuit controlling long-term calorie intake has greater overall strength than the feedback circuit controlling calorie expenditure. Specifically, it has been estimated that for each kilogram of lost weight, calorie expenditure decreases by about 20–30 kcal/d whereas appetite increases by about 100 kcal/d above the baseline level prior to weight loss31. Despite these predictable physiologic phenomena, the typical response of the patient is to blame themselves as lazy or lacking in willpower, sentiments that are often reinforced by healthcare providers, as in the example of Robert, above.
The data shows that there is a better measure than BMI which doesn't carry all the stigma, so why not use that one?
Taking any of the more accurate measurements of body composition requires effort, unlike BMI. It's easy and lazy and not yet acknowledged as an ineffective practice so doctors do it. It's also not always wrong, just because BMI and obesity do correlate.
I don't think it'll reduce stigma or whatever of fat people, but I do agree that if someone is fat but their insulin is normal maybe a doctor can pay attention to something else like if they came in for an allergy test or something they may not need the diet/exercise spiel. Similarly if someone's thin but their insulin is crazy its time to talk diet/exercise.
Otherwise if your insulin is high, you need to diet and exercise. Measuring blood insulin is a lab procedure - you go to a collection lab, you get jabbed, they mail it for testing.
Measuring BMI requires stepping on a scale and knowing how tall you are.
There are contradicting posts in this very thread that say high BMI and obesity aren't necessarily the same thing, because apparently you can be tall and be only mildly into lifting and suddenly you have a high BMI. If that's the case [I'm not a lifter] then it makes total sense to me to track resting blood glucose, not obesity, because it's simply the more accurate measurement.
Very unlikely. BMI is a standardized model. Unless you are dramatically out of the parameters (you're 7'2" or you are an olympic lifter) it is pretty accurate.
You can also use calipers in that case - you need someone to help you because you can't reach it yourself but it's very fast and accurate.
Again, insulin testing is expensive, invasive and time consuming. You might get insulin tested once a year.
It just seems much simpler to me to take an actually accurate test once a year and prescribe purely based on the accurate testing. That way we also get skinnyfat people.
You might ask him why he didn't give you any feedback earlier so you could have fixed the issue. "Oh, well I only give feedback once a year since it is hard to do".
You need regular (weekly, biweekly) feedback on your diet.
It also means that if we can control people's reactions to sugar, their weight becomes irrelevant (or possibly even protective.) Losing weight is usually the way to fix your sugar, but plenty of people who aren't fat have sugar problems, and sugar problems might not be reversed by weight loss in a particular individual.
There are a plethora of studies correlating a change in diet with improved blood markers for fasting blood glucose and insulin response. Sugar "problems" are resolved by diet, and resolving diet goes along with reversing weight loss. It cannot be any other way.
There is no evidence that obese BMIs are negatively correlated to mortality, only the opposite. Obese BMIs still come with increased arterial plaque, complications in anesthesia and surgical healing for medical interventions, fatty liver disease, ocular pressure, lung function, cardiovascular health, and so on.
Obesity is up with smoking as a cause of preventable death. Let's not try to put a pretty face on it. A spade is a spade.
Insulin resistance/high fasting glucose and obesity are inseparable peas in a pod.
Then my father is irrelevant, being diagnosed with prediabetes at 6', 165lb.
So is it really irrelevant numerically, or is that hyperbole? Does your irrelevance include millions of people?
-----
edit:
> There is no evidence that obese BMIs are negatively correlated to mortality, only the opposite.
TFA mentions that it is negatively correlated when you control for two things that are positively correlated with BMI. So, while the study doesn't say that high BMI reduces mortality (it in fact observes the opposite), nobody claimed it did.
But in finding the negative BMI correlation when controlled for insulin, it does make a very good case that
> Obese BMIs still come with increased arterial plaque, complications in anesthesia and surgical healing for medical interventions, fatty liver disease, ocular pressure, lung function, cardiovascular health, and so on.
are largely irrelevant to the mortality of the obese relative to the thin, and inflammation and insulin resistance are decisive.
BMI is a population-level metric, and population-level analyses of health risks, from an epidemiological perspective to inform healthcare providers and public health officials of risks, frankly, don't care about your father.
Yes, it's irrelevant numerically, and it's irrelevant numerically because there's little evidence that any significant portion of the population (much less millions of people) falls into this category, whereas we are one of the fattest countries on Earth, which has absolutely no need to minimize or downplay the health effects of that.
----
Edit (ugh):
> TFA mentions that it is negatively correlated when you control for two things that are positively correlated with BMI.
Kinda the point is that TFA is mathematical gaming which doesn't yield any actionable results for healthcare providers.
> So, while the study doesn't say that high BMI reduces mortality (it in fact observes the opposite), nobody claimed it did.
You did, a bit:
> It also means that if we can control people's reactions to sugar, their weight becomes irrelevant (or possibly even protective.)
There is a range in which increased BMI provides increased survivability from catastrophic events, since the body has energy reserves and some protection from trauma, but then it drops off a cliff. It's not going to be "irrelevant" for health outcomes, much less protective very far outside of the overweight range.
> are largely irrelevant to the mortality of the obese relative to the thin, and inflammation and insulin resistance are decisive.
Yeah, ok. I guess a bunch of things which every significant health organization on Earth calls out as leading causes of mortality and linked to obesity are actually due to inflammation and insulin resistance, which must also have an effect on how hard it is to push a tube into an obese patient's throat, to find a vein to get a line in, the amount of pressure their weight exerts on their lungs, the amount of medication it takes to keep them anesthetized, how much fat accumulates in their liver, plaque buildup inside arterial walls, etc. Definitely just inflammation.
That's encouraging because I've been able to fix the hyperinsulinemia and inflammation, but not all of the excess BMI. Maybe fit & fat aren't mutually exclusive.
https://www.youtube.com/watch?v=wadKIiGsDTw
Taking someone sick with high inflammation and putting them in a hospital bed is terrible for their recovery.
We have created an environment devoid of it and we reap the consequences: increase amount of chronic autoimmune diseases and many other things.
Personally I had to deal with eczema every winter that I got rid of with a simple infrared heater. A friend of mine was dying of IBS in the hospital, he say f*ck it if I am going to die it will be under the sun. He checkout of hospital (Canada), went to mexico and healed himself while ditching the medication.
The OP actually says that BMI is negatively correlated with mortality if you control for insulin levels, so that would mean that thin people with high fasting blood sugar would have more mortality than fat people in the same situation.
Weight gain is caused by an overconsumption of calories, barring a narrow range of medication side effects.
Insulin tells your body to uptake glucose into cellular bodies. If glycogen stores aren't depleted (in the liver from normal/basal metabolism, in muscular tissue from activity) and there's nowhere for it to go, it's converted via de novo lipogenesis and stored in adipose tissue. If there's no free room, new fat cells are created. Lipogenesis and the creation of new adipose tissue are not instantaneous processes.
Insulin resistance is because your body has continued to see that blood glucose is high because it cannot be taken up quickly enough, so it sends out more insulin. As high circulating levels become more normal, the alpha channel on the insulin receptor requires more insulin to activate a signal transduction pathway to open the GLUT4 transporter for glucose.
Very strictly, excess blood glucose causes insulin resistance. That excess blood glucose is caused by caloric overconsumption, which causes obesity. It is not the inverse.
Carbohydrates have poor satiety, so when the stomach is empty (fats and proteins have a longer processing time in the stomach), ghrelin is released, which stimulates hunger, and the person eats again. Probably more simple carbohydrates, possibly before the previous ones have been taken up by cells.
Caloric overconsumption -> obesity
Insulin resistance is a side effect here, one which is seen because simple carbohydrate consumption is part and parcel of obesity in the western world. It's very hard to become clinically obese without cheap sugar. The same mechanisms (overconsumption of calories) cause obesity and insulin resistance, so they're two sides of the same coin.
> Higher fasting insulin and higher c-reactive protein confound the association between BMI and the risk of all-cause mortality. The increase in mortality that has been attributed to higher BMI is more likely due to hyperinsulinemia and inflammation rather than obesity.
Isn’t high BMI interchangeable with obesity in most studies?
You're letting the word "are" carry a lot of weight where "often" or "can be" might be more accurate.
They're correlated with obesity, which means that they're more likely to be present in obese patients. But they exist independent of it: not being present for all obese patients and sometimes being present for non-obese patients. And since they're measurable on their own, it's kind of a big deal to draw them out as obesity-associated factors more closely correlated with all-cause mortality than obesity as a whole.
Can you more specifically characterize what you mean by that and what your references are?
Because you're taking a statistical correlation and a hypothesized mechanism and turning it into logical entailment. It's hard to believe that even a extremely high correlation is anywhere near 100% but I would love to see real data saying otherwise.
I think I took it to mean you saying the study hadn't meaningfully narrowed down mortality risk to less noisy factors than obesity, when I see now that you may have just been helping the prior commenter see that obesity itself is still correlated with mortality through these. Probably because I can't even grok how they got to that idea. :)
This is still a useful study though! It's useful to know that the cause is almost entirely through those factors. Though this is also an epidemiological study with no randomization and some of the hazard ratios are not particularly high.
Waist measurement is a stronger measure but still flawed.
That's basically tautological. Outliers are outliers because they don't fit the model.
for folks like yourself with T2D, noting changes. Keep on keeping on (i.e. follow current medical advice)
No, it is not. We should be very clear that high BMI is, in fact, life shortening.
High BMI is, in fact, strongly correlated with atherosclerosis. It is, in fact, strongly correlated with decreased lung volume and risk pulmonary edema. It is correlated with clot formation and stroke. It is correlated with fatty liver disease. It is correlated with decreased bone density, which leads to a heightened risk of throwing embolism-causing clots if a fall leads to a break. It is correlated with a variety of cardiac events, mostly around hardening valves and fat buildup in cardiac tissue causing lessened output while there is more circulating blood volume. It is correlated with difficulties administering anesthesia or intubating patients in case of emergency. It is correlated with...
Being obese is still horrifyingly unhealthy even if high fasting glucose is just a proxy for some of those factors (which it is not, as this study is terrible).
~20.3 here
and we chose not to justify or explain those parameterizations in any way, because what we probably actually did was diddle SAS until the model did what it was supposed to.
yawn.
"Don't be snarky."