Jump to content

Body-mass index and all-cause mortality


Todd Allen

Recommended Posts

I have not read the report, but there is a summary of this and another report from a generally reliable blogger on obesity issues here.

 

 

many observational studies have reported that people in the overweight or even obese categories sometimes experience better health outcomes than lean people.  This is the "obesity paradox".  I covered this concept in detail earlier this year (1).

Yet as with all observational methods, these findings are vulnerable to confounding-- and it can sometimes be profoundly misleading.  There are reasons to believe that confounding could be particularly relevant here.  First, people who are sick tend to lose weight, making leanness look more dangerous than it really is.  Second, cigarette smokers tend to be leaner than nonsmokers, and also much less healthy, also making leanness look dangerous.

Fortunately, there are ways to correct for these potential confounding factors, at least to some degree.  The research of Andrew Stokes has shown that when we do so, the obesity paradox goes away (1).  Two new studies strongly confirm that when confounding is minimized, there is no paradox.

Study #1: "BMI and all cause mortality: systematic review and non-linear dose-response meta-analysis of 230 cohort studies with 3.74 million deaths among 30.3 million participants" ... (2).

The thing I really like about this study is that they analyzed several subsets of the data, each of which was progressively less likely to be confounded.  ...  In all subjects, as well as current smokers, the lowest mortality level occurred at a body mass index (BMI) of 25, which is on the cusp of overweight.  Yet among people who have never smoked, the optimal BMI was 23-24.  Among people who had never smoked and who were healthy at baseline, the optimal BMI was 22-23.  And among people who had never smoked and were followed up for at least 20 years, the optimal BMI was 20-22!  ... The last analysis is a particularly powerful way of avoiding confounding due to existing illness.  If you're recording a person's weight right now and their risk of death in 20 years, it's likely that whatever kills them in 20 years is not impacting their weight right now.  So you get a cleaner assessment of the impact of BMI on health.  This assessment shows that it's much more dangerous to be obese than to be underweight.  You can see that in this graph of BMI vs. mortality from the paper:

 

BMI%2Band%2Bmortality.JPG

Link to comment
Share on other sites

Contrary to former reports of a "J−shaped" relationship between body weight and human mortality, when corrections are made for smoking and underlying disease the relationship is linear [NEW ENGLAND JOURNAL OF MEDICINE; Manson,JE; 333(11):677-685 (1995)].

 

Body Weight and Mortality among Women

http://www.nejm.org/doi/full/10.1056/NEJM199509143331101

 

 

RESULTS:

In analyses adjusted only for age, we observed a J-shaped relation between body-mass index and overall mortality. When women who had never smoked were examined separately, no increase in risk was observed among the leaner women, and a more direct relation between weight and mortality emerged (P for trend < 0.001). In multivariate analyses of women who had never smoked and had recently had stable weight, in which the first four years of follow-up were excluded, the relative risks of death from all causes for increasing categories of body-mass index were as follows: body-mass index < 19.0 (the reference category), relative risk = 1.0; 19.0 to 21.9, relative risk = 1.2; 22.0 to 24.9, relative risk = 1.2; 25.0 to 26.9, relative risk = 1.3; 27.0 to 28.9, relative risk = 1.6; 29.0 to 31.9, relative risk = 2.1; and > or = 32.0, relative risk = 2.2 (P for trend < 0.001). Among women with a body-mass index of 32.0 or higher who had never smoked, the relative risk of death from cardiovascular disease was 4.1 (95 percent confidence interval, 2.1 to 7.7), and that of death from cancer was 2.1 (95 percent confidence interval, 1.4 to 3.2), as compared with the risk among women with a body-mass index below 19.0. A weight gain of 10 kg (22 lb) or more since the age of 18 was associated with increased mortality in middle adulthood.

CONCLUSIONS:

Body weight and mortality from all causes were directly related among these middle-aged women. Lean women did not have excess mortality.

5XMlyRA.png

Link to comment
Share on other sites

Contrary to former reports of a "J−shaped" relationship between body weight and human mortality, when corrections are made for smoking and underlying disease the relationship is linear [NEW ENGLAND JOURNAL OF MEDICINE; Manson,JE; 333(11):677-685 (1995)].

 

Body Weight and Mortality among Women

http://www.nejm.org/doi/full/10.1056/NEJM199509143331101

 

We-e-e-el, you say "Contrary to former reports," but the report you're citing is from 1995, is just tone study, and involves a cohort of "just" 115,195 U.S. women enrolled in the Nurses' Health Study. The new report cited above is a massive "meta-analysis of 230 cohort studies with 3.74 million deaths among 30.3 million participants. The happy thing is that it performs many of the same kinds of adjustments and stratifications that Manson & co identified lo those many years back (and more — subanalysis of studies of 20 years' followup!), and come to a very similar (and, to we skinny bastids, reassuring) conclusion.

Link to comment
Share on other sites

Regarding cohort sizes and validity of studies based purely on statistical power, one needs to establish what is a valid sample. Unless you are studying every single individual, you per force must rely on a statistical sample. The question then becomes what size of sample makes it representative enough to allow for conclusions which are likely to be valid if extrapolated to the entire population. You can break that down by any number of criteria (age, race, class, health status, health behaviors etc.), but at some point, the sample is large enough to become statistically valid. In voting outcomes for U.S. elections, they found that a sample of 1050 randomly picked individuals as enough power, so that enlarging the sample further doesn't bring anything more to the table. It's sort of like maximum height - at some point bringing in more and more people is not going to suddenly find you a 10 foot tall man, so a smaller sample is absolutely valid and a larger sample does not make it "more" valid. It is my suspicion that if your study has 115,195 women it's going to be just as valid if you bring in 1 million or 30 million women. Here we are of course restricted to women, but at least we can assume it will be valid for women, period. I suppose one could try to make a case that it is dramatically different for men, but I suspect that any differences are not going to be that dramatic.

 

I think it's probably more informative (for our purposes) to examine BMI distribution in the elderly - 80 and above, and particularly the very old. Not just because we're interested in what BMI seems most common at such advanced ages, but also because it is at least theoretically possible that the BMI range that's best for survival at earlier ages may not be the same as the BMI range in the very old (for example, it's been noted that there is a common thread of people generally going up in their BMI ranges as they age - is that good? Bad? Indifferent?).

Link to comment
Share on other sites

Michael wrote:

I have not read the report, but there is a summary of this and another report from a generally reliable blogger on obesity issues here.

 

Apparently not only haven't you read the report, you haven't read about the report either. The discussion by your reliable blogger (on June 29th) was long before this brand-new meta-analysis [1] that Todd pointed to was published in the Lancet this week. In fact, like your reliable blogger, I discussed that other study (PMID 27146380) that he covered (and you mistakenly think is the new one) in this post on May 8th, including the same J-shaped mortality curve you (and he) posted above (although w/o the https so you may not have seen it... corrected now). So you may have read about that older study (from May) in my post as well, since it is a thread you started and apparently monitor, although perhaps not too carefully... ☺ 

 

But your confusion is not surprising, given the fact that the titles of the two papers are similar, and since they are both huge systematic reviews of virtually all studies done to date about the association between BMI and mortality. In fact, they almost entirely overlap in the the 230+ studies they analyzed. But it is definitely two different analyses, with the newer study Todd points to [1] including a few more studies in its data. In fact [1] isn't even in Pubmed yet. One other striking difference is the number of authors. The older review (PMID 27146380) had eight authors. This new one has too many authors to even bother counting, to say nothing of list in the abstract. In fact, the full author list is over two pages long, and included not in the free full text but in the appendix.

 

Interestingly, just two days ago Paul McGlothin teased this new BMI study [1] in his weekly CR Way newsletter as more reason to avoid getting too thin, saying:

 

A new paper is out reinforcing other findings that maintaining a higher BMI - 20 to 25 - did better for longevity. This reinforces what we've observed among CR Way travelers who have lived to greatly advanced age. Maintaining a higher BMI does not interfere with activating longevity signaling. Meanwhile, it may be protective of body structures, making disasters like osteoporosis-related fracture less likely.

 

This is obviously in line with my hobby horse of late - namely that serious CR probably won't beat a healthy, obesity-avoiding diet (by "serious" I mean CR that results in a BMI of ~18.5 or less). This is borne out by the uptick in the mortality graph Michael posted above (and I posted in May here) in folks with a BMI below 20, even restricting to studies with 20+ years of follow-up and limiting analysis to healthy, never-smokers.  In short, it appears to pay to be on the thin side, but not too thin.

 

But that conclusion was based on analysis of PMID 27146380 and other past studies of the association of BMI and mortality. What does Todd's new study [1] say on the topic? Let's find out, shall we?

 

First a little background from the intro, on the study and it's huge set of authors:

 

The Global BMI Mortality Collaboration was established to provide a standardised comparison of associations of BMI with mortality across different populations. It includes individual-participant data for 10·6 million adults in 239 prospective cohort studies in 32 countries, mainly located in Asia, Australia and New Zealand, Europe, or North America, about 4 million of whom were never-smokers without reported chronic diseases (mainly cardiovascular disease, cancer, or chronic respiratory disease) at recruitment and who were still being followed up 5 years afterwards.

 

Obviously this was a big meta-analysis, both in terms of number of authors, countries, studies and subjects included. And right off the bat you can see that this study had a relatively short minimum follow-up period of 5+ years, and so may be open to criticism regarding latent disease potentially skewing the results against the health benefits of being very thin. But more on that below.

 

Next, they winnowed down their initial study population of 10M people to ~4M never-smokers - so that's good. No confounders associated with thinness due to past or present smoking habits.

 

Interestingly, per Tom's speculation optimal BMI shifting with age, they found that as one ages, the optimal BMI for longevity increases, as illustrated in this figure, showing mortality vs. BMI curves for 35-49 year olds (left), 50-69 year olds (middle) and 70-89 year olds (right). 

 

 TgjXQSA.png

 

From the full text:

 

The [mortality] nadir depended on age, and was BMI=22 kg/m² for baseline age 35–49 years, BMI=23 kg/m² for baseline age 50–69 years, and BMI=24 kg/m² for baseline age 70–89 years.

 

This increase in optimal BMI with age is consistent with the strategy of packing on some pounds as one gets older to have more metabolic reserves to recover from the inevitable slings and arrows associated with advancing age, as discussed in depth on the optimal Late-Life BMI for Longevity thread.

 

The authors recognized that bad analysis in some prior studies unfairly penalized low BMI:

 

Compared with the strict primary analyses described above, crude analyses that ignored smoking and any effects of prior disease at baseline, and did not exclude the first 5 years of follow-up, yielded different (presumably substantially biased) results, with exaggerated HRs for underweight, inverted HRs for overweight, and less than half of the excess risk for grade 1 obesity suggested by the strict primary analyses (tables 1, 2, appendix p 43).

 

So the so-called "Obesity Paradox" does go away when you do the analysis right. But you'll notice that in all of the three age groups there was still a penalty for having a BMI below 20, even using their improved methodology. This held true across all countries (Figure 1, data now shown) and for both men and women. On the issue of gender, they found something interesting. For men, it was substantially worse to be too fat, or too thin, relative to women, as you can see from this graph showing BMI vs. mortality separately for each gender.

 

  6U2xfnR.png

 

Eyeballing the graph, it looks like for men, having a BMI in the 18.5-20 range increases mortality risk by over 50% relative to a BMI of 22.5 - 25. In fact, the mortality risk of the 18.5-20 BMI men was almost as high as men with a BMI in the 30-35 range! Being a healthy never-smoker with a BMI less than 18.5 almost doubled a man's mortality risk. Ouch!

 

Digging into the weeds of the Appendix, here is an interesting graph showing how various adjustments impacted the association between BMI and mortality:

 

Oyc17Ut.png

 

Look at the two pairs of points I've circled in green. They show the different between not excluding the first five years (red points) and excluding the first five years (black points) of follow-up from the analysis. As you can see, excluding the first five years makes the penalty for being thin smaller - presumably because of latent health conditions (like undiagnosed cancer) that caused people both to be thin at study onset, and to die within the first five years. But as you can see, the reduction in the thinness penalty was pretty modest from excluding the first five years, and additional reductions in the thinness penalty for longer initial exclusions would almost certainly be even more modest - i.e. longer initial exclusions would likely reduce the thinness penalty only a little bit more. This is borne out by the continued J-shape of the mortality curve in PMID 27146380 when data was limited to studies of healthy never-smokers with at least 20 years of follow-up.

 

In short, it appears to pay to be on the thin side, but not too thin, especially if you are a man.

 

And even that summary comes with some caveats and nuances (favoring being fatter) that should be considered, as Michael and I discussed at the recent conference, and I elaborated on recently here. In short - "too thin" for optimal longevity may be a moving target - i.e. optimal BMI may actually be increasing over time.

 

First, the longevity penalty for being overweight, obese, or even high-normal BMI may have actually gone down over the last several decades, as a result of improved medical treatments, particularly for CVD (e.g. stents and statins). So studies with a very long follow-up that show "it pays to be very thin" may be behind the times.

 

What such long follow-up studies are actually saying is "it used to pay to be very thin".  Decades ago it was good to be very thin because that was the only way you could minimize risk of dying from CVD. But these days, being thin provides less of a benefit, or more accurately, being fat provides less of a penalty, when it comes to longevity. And because being very thin carries with it some additional risks (due to fewer metabolic reserves to fight illness or recover from accidents) relative to carrying more meat on your bones, the trend has been towards a shift upwards in the best BMI for longevity over the last several decades.

 

Further, despite what Michael's "credible blogger" says, it seems to me just crazy to think that a latent condition (e.g. cancer) at the start of a study was making people thin, and contributing to their demise 15-20 years later, so the only way to see the benefits of being thin is to wait 20+ years of follow-up. The modest effect of excluding the first 5 years discussed above bears witness to the silliness of this "long follow-up required to really exclude latent conditions" nonsense.

 

The much more likely explanation for the attenuation of the thinness penalty in studies with very long follow-ups (besides a rising optimal BMI over the decades due to medical improvements) is the one I described here, using an analogy with robust beakers (fat people) and fragile test tubes (skinny people) and their survival prognosis in a laboratory. I really liked that analogy, and I think it sheds light on what's going on with optimal BMI.

 

In a nutshell, in the short term, it sucks being thin because thin people are more fragile and therefore more likely to die if they get sick or injured. But in the long run, it pays to be thin since you're accumulating less obesity-related damage (i.e. clogged arteries, diabetes, etc) which eventually catches up with fat people and makes them more prone to dying than thin people in their later years. As a result, it is natural for studies with a very long follow-up to show an advantage for being thin relative to being fat, since with a very long follow-up the penalty for being fat has a chance to catch up with the fatties.

 

For those of you more visually inclined, here is a schematic graph of my explanation for why short-term follow-up studies look bad for thin folks and long-term follow-ups look good for them (note it is only meant to be a schematic - the slopes, shapes and positions of the curves are not meant to be accurate):

NXcquQt.png

 

Just now I thought of another reason that could explain why it looks better to be thin in longer follow-up BMI-vs-mortality studies.

 

In nearly all prospective studies like these, BMI is measured once at the beginning of the study and perhaps a second time in the early years of follow-up, if they are lucky. Over time, the BMI of the initially-thin people will gravitate upwards, as most people pack on the pounds as they age. So if there really is a consistent & persistent penalty for being thin, it's effect on the data will be attenuated with longer follow-up since the initially thin people won't be as thin (or therefore, as death-prone) as the study progresses and as they slide upwards into the higher, less death-prone BMI categories. In other words, they were thin at the beginning of the study, but not by the end many years later, so being thin at the beginning doesn't look as detrimental.

 

Overall, it does not appear to me that there is any data supporting the idea that being a healthy, very thin never-smoker is any better than being a healthy, normal-weight never smoker. If anything the evidence points in the opposite direction - in line with my "serious CR won't beat..." hypothesis.

 

But I will acknowledge that the diets of most people (thin and chubby alike) are on average pretty bad in all these studies of average people, so it is a bit hard to generalize to healthy-eating thin folks like us.

 

--Dean

 

---------

[1] Lancet, 10.1016/S0140-6736(16)30175-1

 

Body-mass index and all-cause mortality: individualparticipant-data meta-analysis of 239 prospective studies in four continents

 

The Global BMI Mortality Collaboration*

 

Summary
Background Overweight and obesity are increasing worldwide. To help assess their relevance to mortality in diff erent
populations we conducted individual-participant data meta-analyses of prospective studies of body-mass index (BMI),
limiting confounding and reverse causality by restricting analyses to never-smokers and excluding pre-existing
disease and the fi rst 5 years of follow-up.
 
Methods Of 10 625 411 participants in Asia, Australia and New Zealand, Europe, and North America from 239 prospective
studies (median follow-up 13·7 years, IQR 11·4–14·7), 3 951 455 people in 189 studies were never-smokers without
chronic diseases at recruitment who survived 5 years, of whom 385 879 died. The primary analyses are of these deaths,
and study, age, and sex adjusted hazard ratios (HRs), relative to BMI 22·5–<25·0 kg/m².
 
Findings All-cause mortality was minimal at 20·0–25·0 kg/m² (HR 1·00, 95% CI 0·98–1·02 for BMI 20·0–<22·5 kg/m²;
1·00, 0·99–1·01 for BMI 22·5–<25·0 kg/m²), and increased signifi cantly both just below this range (1·13, 1·09–1·17
for BMI 18·5–<20·0 kg/m²; 1·51, 1·43–1·59 for BMI 15·0–<18·5) and throughout the overweight range (1·07,
1·07–1·08 for BMI 25·0–<27·5 kg/m²; 1·20, 1·18–1·22 for BMI 27·5–<30·0 kg/m²). The HR for obesity grade 1
(BMI 30·0–<35·0 kg/m²) was 1·45, 95% CI 1·41–1·48; the HR for obesity grade 2 (35·0–<40·0 kg/m²) was 1·94,
1·87–2·01; and the HR for obesity grade 3 (40·0–<60·0 kg/m²) was 2·76, 2·60–2·92. For BMI over 25·0 kg/m²,
mortality increased approximately log-linearly with BMI; the HR per 5 kg/m² units higher BMI was 1·39 (1·34–1·43)
in Europe, 1·29 (1·26–1·32) in North America, 1·39 (1·34–1·44) in east Asia, and 1·31 (1·27–1·35) in Australia and
New Zealand. This HR per 5 kg/m² units higher BMI (for BMI over 25 kg/m²) was greater in younger than older
people (1·52, 95% CI 1·47–1·56, for BMI measured at 35–49 years vs 1·21, 1·17–1·25, for BMI measured at
70–89 years; pheterogeneity<0·0001), greater in men than women (1·51, 1·46–1·56, vs 1·30, 1·26–1·33; pheterogeneity<0·0001), but similar in studies with self-reported and measured BMI.
 
PMID: 27423262
DOI: 10.1016/S0140-6736(16)30175-1
Link to comment
Share on other sites

But I will acknowledge that the diets of most people (thin and chubby alike) are on average pretty bad in all these studies of average people, so it is a bit hard to generalize to healthy-eating thin folks like us.

 

Yes. A few years back, the obesity paradox was extensively discussed on the list. One tentative mechanism to explain it was that the fatter people ate more in general than thin people - but since everyone's diet (apart from calories) was equally SAD (Standard American Diet), a diet that is low in several critical nutrients, those who eat more (fatties) have less of those critical deficiencies, whereas the super thin are very frankly malnourished. I don't know to what degree I buy this idea, what with studies showing that in some ways overweight and obese people are not simply eating more of the super thin person's diet, but that the weight gain is driven by "empty" calories such as more sweetened soda etc. - lacking in nutrients and perhaps even exacting a penalty of having to process f.ex. excess sugars with fewer digestive resources. But then again, if excess mortality starts to climb again with more overweight and obese, that would argue in favor of seeing the super thin and the optimal diet folks as having pretty much the same diet (except amounts, of course), so no extra penalty for "empty" calories - the penalty such as it is, would then apply simply in absolute amounts of deficiencies, as empty calories would be then more threatening to the super thin on account of less margin of safety wrt. deficiencies. 

 

And right on time, Al posted this (thanks Al!) interesting study - PMID: 27406257

 

Energy intake, energy required and mortality in an older population

 

Abstract

 

Objectives 

 

The present study evaluated the association between energy intake, energy required and mortality in older adults.

 

Design 

 

A cohort study with a mean of 10·67 (sd 4·74) years of follow-up. Participants completed a 24 h dietary recall. Energy required per day was computed by BMR. Deaths through 2006 were identified from the National Death Index. A Cox regression was used to estimate the hazard ratios (HR) of quantiles of energy intake and energy required on all-cause and CVD mortality, adjusting for demographics, socio-economic status and co-morbidity.

Setting The National Health and Nutrition Examination Survey (NHANES) III, 1988–1994.

 

Subjects 

 

A total of 4846 participants aged 60 years or above were analysed.

 

Results 

 

Within the follow-up period, there were a total of 2954 deaths (61·0 %), 51·9 % were caused by CVD. Relative to those in quartile 1 of energy intake, only quartile 4 was associated with all-cause mortality and CVD mortality with HR of 0·86 (95 % CI 0·77, 0·96, P=0·006) and 0·76 (95 % CI 0·65, 0·89, P=0·001), respectively. On the other hand, relative to those in quartile 1 of energy required, all quartiles of participants had a lower risk of all-cause mortality and CVD mortality. The interaction effects between energy intake and energy required with all-cause and CVD mortality were insignificant (P=0·70 and 0·61, respectively).

 

Conclusions 

 

Independent of energy required, higher energy intake was associated with lower HR of both all-cause and CVD mortality in older adults.

 

 

But what about the very old? Dean pointed to the "optimal late-life BMI for longevity" thread where the BMI distributions compiled by Michael Lustgarten seem to be, quote: "19.3-24.4 kg/m2, with an average BMI of 21.8" - not exactly even in overweight category, FWIW, even allowing for any BMIs creeping upward with age.

 

It is just possible - although I don't know if particularly plausible - that maybe it's not enough to restrict energy intake to not get overweight, but you actually have to have a superior diet than SAD to reach late 90's and into centenarian territory - in other words, if it were just a matter of not getting overweight to reach a very old age, there would be a lot more of the very old around. Instead, only a subgroup of the "not overeaters" make it to such an advanced age, because on the whole, their diets are better than SAD, and they don't suffer from the same nutritional deficiencies on a smaller energy allotment (i.e. they consume a more nutrient dense diet with a superior calories/nutrients ratio). If that is so, then one could at least make a theoretical argument that the super thin would also survive into such an old age with f.ex. BMIs of 18.5 and slightly below, but that would necessitate super-dense nutrient diet that simply doesn't happen for any appreciable number of people (freaks like us excepted!). In other words, only CRON folks (rather than straight DR energy restriction) can make it to very old age with sub 19 BMIs.

 

But is there any evidence at all that the very old have a diet that in any way differs from SAD except in quantity? I'd be interested in studies such as PMID:25446984

 

Metals in plasma of nonagenarians and centenarians living in a key area of longevity.

 

Abstract

 

The concentration of calcium (Ca), cobalt (Co), copper (Cu), iron (Fe), magnesium (Mg), manganese (Mn) and selenium (Se) in plasma of 76 nonagenarians (mean age, 89.0±6.3 years), 64 centenarians (mean age, 101±1 years) and 24 middle-aged subjects as controls (mean age 61.2±1.1 years), was determined by sector field inductively coupled plasma mass spectrometry. All the subjects lived in Sardinia, an Italian island, that has the higher prevalence of centenarians than in other European countries. A comparison among the three classes of age showed a significant depletion of Ca, Co, Fe, Mn and Se (all p<0.001) in nonagenarians and centenarians with respect to controls. In particular, the geometric mean (GM) values of Ca, Co, Fe, Mn and Se were: 94.1 μg/ml, 0.46 ng/ml, 1314 ng/ml, 2.47 ng/ml and 111 ng/ml in controls; 87.6 μg/ml, 0.22 ng/ml, 815 ng/ml, 1.07 ng/ml and 88.9 ng/ml in nonagenarians; 87.0 μg/ml, 0.29 ng/ml, 713 ng/ml, 1.27 ng/ml and 81.9 ng/ml in centenarians. The highest inverse relationship with age was observed for Fe (p<0.001; ρ=-0.352) and Se (p<0.001; ρ=-0.417). This trend was also observed when data were sorted by gender. On the other hand, Cu and Mg levels in plasma remained substantially unchanged during aging. As regards Cu, it was significantly higher in females than in males in controls (GM, 1294 ng/ml vs. 1077 ng/ml; p=0.012), in nonagenarians (GM, 1216 ng/ml vs. 1081 ng/ml; p=0.011) as well as in centenarians (GM, 1226 ng/ml vs. 1152 ng/ml; p=0.045) and in hypertensive subjects with respect to healthy people (GM, 1215 ng/ml vs. 1129 ng/ml; p=0.021). These data can be used to enhance knowledge and support the research on: i) metals involved in aging in areas with high rates of human longevity; ii) variables (gender, lifestyle habits and health status) as critical determinants in aging; and iii) mineral intake and supplementation at older age affecting the healthy aging.

 

What's interesting to me in this study, is that the subjects were all from the same area, which cuts down on the number of confounders wrt. ethnic background, cultural factors etc. I don't know how many far-going conclusions one can derive from this, but this is the direction I think is fruitful for exploration.

 
Link to comment
Share on other sites

Tom,

 

Thanks for your great post. I'm going to divide my response into two parts. In this post, I'm going to address the part of your post that deals directly with BMI. Then I'm going to use the second, more general and extremely interesting half of your post as a jumping-off point for a new thread  - Should We Mirror the Long-lived to Live Long? which addresses the important topic of what we can (and cannot) learn from the very old.

 

You wrote:

One tentative mechanism to explain [the Obesity Paradox] was that the fatter people ate more in general than thin people - but since everyone's diet (apart from calories) was equally SAD (Standard American Diet), a diet that is low in several critical nutrients, those who eat more (fatties) have less of those critical deficiencies, whereas the super thin are very frankly malnourished. I don't know to what degree I buy this idea, what with studies showing that in some ways overweight and obese people are not simply eating more of the super thin person's diet, but that the weight gain is driven by "empty" calories such as more sweetened soda etc. - lacking in nutrients and perhaps even exacting a penalty of having to process f.ex. excess sugars with fewer digestive resources.

 

Yes - there certainly could be two opposing mechanism that creates a "sweet spot" for longevity in the low-middle of the BMI range. Eat too little of a SAD diet and people may become malnourished and die early. Eat too much of a SAD diet and people may become obese from the empty calories and die early.

 

Moral - It's best not to eat a SAD diet, either way!

 

And right on time, Al posted this (thanks Al!) interesting study - PMID: 27406257

 

[Aside: Here is a question for someone (perhaps Michael). Should we start using DOIs rather than PMIDs as our way of pointing to papers? I see PubMed is providing them with every paper, and they are the most effective way of search sci-hub.cc. Here is the DOI FAQ. Maybe DOIs should become our new default?]

 

But onto PMID: 27406257 (aka DOI: 10.1017/S1368980016001750) which I'll call [1] for short ☺.

 

It was a study (once again) of the NHANES III cohort, which began in 1988. They followed ~5000 people ages 60+ for an average of 10 years, correlating energy intake and energy requirement at baseline (both assessed via questionnaires) with subsequent mortality. They calculated energy requirement as follows:

 

Energy required per day was computed as BMR times physical activity factor using the revised Harris–Benedict equation (ref) as follows. First, BMR was computed using the following equations:

 

dzbpSqs.png

 

The energy required per day was then computed as BMR times 1·2, 1·375, 1·55, 1·725 or 1·9 if the physical activity frequency was <1, 1–3, 4–5, 6–7 or >7 times per week. Both energy intake and energy required were grouped into quartiles.

 

So their estimate of energy requirement was a function of both basal metabolic rate and activity level, as you'd expect.

 

Although probably largely underreported and therefore to taken with a big grain of salt, the reported energy intake for the four quartiles of men and women combined were:

  • Q1: < 1144         kcal/day
  • Q2: 1144 - 1545  kcal/day
  • Q3: 1545 - 2040  kcal/day
  • Q4:  > 2040        kcal/day (average ~2300kcal/day)

Surprisingly, the average BMI varied little across the four quartiles of energy intake, with values of 27.4, 27.1, 26.9 and 26.8 for Q1-Q4 of energy intake, respectively. Interesting that the BMI trend was modestly down with increasing energy intake. 

 

What they found was fascinating, and strongly supports my contention about calories, exercise and longevity. Here it is in cool 3D graphical form, with some blue labels I've added for clarity:

 

dGbHsr3.png

 

Pretty much self explanatory, don't ya think?! If not, I'll spell it out.

 

There was no a benefit to eating fewer calories in this 60+ age cohort. In fact, the data goes the other way, particularly for CVD mortality (right graph) - i.e. the more calories you were eating at baseline, the less likely you were to die over the subsequent 10 years. In fact, if we ignore energy requirement for a moment, that was indeed true:

 

Relative to those in quartile 1 of energy intake, only quartile 4 was associated with all-cause mortality with HR of 0·86 (95 % CI 0·77, 0·96, P=0·006).

 

In other words, people who ate the most calories (Q4) were 14% less likely to die during follow-up than those eating the least (Q1). Actually, digging into the supplemental material, the all-cause mortality advantage for eating the most (Q4) relative to the least (Q1) was even greater (19%) after adjusting for "age, sex, BMI, race, education level, smoking status, hypertension, type 2 diabetes, and high C-reactive protein level." 

 

But the real mortality advantage came from being active. The most active folks were just shy of 40% less likely to diet than the least active folks, even after adjusting for the same set of potential confounders.

 

In fact, the lowest mortality of all was seen in active folks who ate the most calories (front right bar in both graphs above). Again we see no advantage to being active while restricting calories relative to being active and eating a lot. As you can see, if anything being active and eating little (i.e. the classic CR lifestyle) was worse than being active and eating the most calories, especially relative to CVD mortality.

 

Here is how the authors put it:

 

Energy intake and energy required were independently [and inversely] associated with all-cause and CVD mortality with an insignificant interaction effect (P=0·70 and 0·61, respectively). In other words, regardless of the level of energy required, higher energy intake was associated with lower HR of both all-cause and CVD mortality.

 

Discussing implications, the authors say (my emphasis):

 

[R]esults of the current study suggest that the benefits of nutritional support [encouraging seniors to eat more] override its disadvantages in this population. Our study also showed that energy intake was negatively associated with both all-cause and CVD mortality, regardless of the level of energy requirement; our study did not support our hypothesis that energy requirement modifies the association between energy intake and mortality. The study results did not support the argument that lower energy intake may be beneficial to health.

 

On the topic of this thread, BMI, they say:

 

Our findings showed that the people still alive at followup had higher BMI compared with those who died in both bivariate analysis (Table 2) and after adjusting for demographics, socio-economic status and health conditions (results not shown).

 

This is right in line with the "thin people are more fragile" hypothesis to explain why it's better to have meat on your bones, at least over 60. And note, even the skinniest folks in this cohort weren't that skinny: "only 2.13% had BMI < 18.5". So it wasn't like the people in the lowest quartile of energy intake (or BMI) were wasting away at death's door. In fact excluding the first 3 years of follow-up, and correcting for known health conditions didn't change the results, suggesting at least tentatively that lack of appetite (and hence low energy intake) was not due to either diagnosed or undiagnosed health conditions. 

 

And what goes for BMI, also went for calories - i.e. the more the better:

 

Increased energy intake was observed in the fourth quartile (~9625 kJ/d (~2300 kcal/d)) as a protective factor to reduce all-cause and CVD mortality after adjusting for confounding; this was further evident in the null interaction of energy intake and energy required with mortality, in which energy intake higher than ~9625 kJ/d (~2300 kcal/d) was protective regardless of the amount of energy required.

 

So you didn't even have to be more active to benefit from extra calorie intake, although being active helped reduce mortality risk even more.

 

And as Michael likes to say: N.B. (note bene - or note well) these folks weren't Adventists - just average Jills and Joes. It was beneficial to eat more calories, even if those calories were in the form of a standard American diet.

 

I'd say this is pretty strong support for my contention that serious CR won't beat a healthy, obesity-avoiding diet and lifestyle. In fact, this study quite strongly suggests exactly the opposite for seniors, i.e. even a not-so-healthy, obesity-avoiding but calorie-replete diet coupled with an active lifestyle will trounce CR when it comes to longevity.

 

--Dean

 

-------

[1] Public Health Nutr. 2016 Jul 13:1-7. [Epub ahead of print]

 
Energy intake, energy required and mortality in an older population.
 
Lee PH(1), Chan CW(1).
 
 
[see Tom's post above for abstract]
 
DOI: 10.1017/S1368980016001750 
PMID: 27406257 
Link to comment
Share on other sites

Hi ALL!

 

Dean, the really bad confounder in all the material in this thread is:  We are very different from ANY of the categories being discussed -- we not only eating less (CR), but we are carefully being sure to obtain adequate nutrition (AN).  The "AN" part of "CRAN" is not being considered in any of these studies -- therefore they are meaningless to us.

 

  --  Saul

Link to comment
Share on other sites

Hmmm... People who are less active and lack appetite may of course just be depressed, sickly, weaker genetically, on and on. How do you control for that? Even eliminating the actual sick factor by using data after three years would not account for generally weaker, sicker types who are not suffering from a chronic illness but are just lacking vitality. I think Saul has a point here. Not so much supporting CR as supporting the fact that vitality effects life expectancy and possibly the Michael Rae "slop factor hypothesis"

Link to comment
Share on other sites

Saul, 

... The "AN" part of "CRAN" is not being considered in any of these studies -- therefore they are meaningless to us.

 

I think you give too much credit to the benefits of a healthy CR diet. Neither a healthy diet nor CR make you invincible. We can debate whether its rate is attenuated, but aging continues, even for CR folks.

 

What the evidence clearly shows it that it's bad to be too thin when you're old, independent of how your got there (via good genes, good luck or a good diet & lifestyle), and independent of the number of candles on your b-day cake when your body starts to break down, which it inevitably will, even for people eating a healthy CR diet.

 

Put another way - what PMID 27406257 shows is that eating more of even a pretty crappy diet is beneficial when one is old - likely in large part because it helps avoid frailty and increase one's robustness to illness and injury.

 

If the benefits outweigh the costs of eating more of even a crappy diet in old age, why in heck would eating more of a good diet be bad for you? It would seem eating more of a good diet would not only help you avoid frailty, but also prevent loading your system with things that harm it, like saturated fat & refined sugar.

 

Do you really think the negligible (if any) longevity benefits of sticking with CR into old age would really offset the cost in terms of increased frailty? If you do, you're even less rational than I give you credit for, and that's saying a lot! ☺

 

We'll see if you're still singing the same tune when you've caught pneumonia or broken a hip and have been stuck in a hospital for month, as your already thin body wastes away due to lack of appetite and refusal to eat the diet of white toast and jello they allow you.

 

Even our MIA CR Society President Brian acknowledged in the Optimal late life BMI thread that based on the available evidence, it would likely be good for a skinny (BMI ~18.5) long-term CR practitioner to put on some pounds when they get older.

 

Mike, 

Hmm... People who are less active and lack appetite may of course just be depressed, sickly, weaker genetically, on and on. How do you control for that? ... 

 

Not so much supporting CR as supporting the fact that vitality effects life expectancy...

 

I agree with the importance of vitality as predictors of healthy and longevity.

 

But disciplined seniors (and even some disciplined juniors!) might also lack vitality because they are blindly laboring under the apparently false notion they are doing themselves a favor by continuing to restrict calories and then trying to avoid getting too thin (due to both CR and age-induced nutrient malabsorption - esp protein) by limiting energy expenditure.

 

In short, it's not all about "calories, calories, calories", especially when you get into advanced age. Quite the contrary. The lesson for seniors from PMID 27406257 seems to be: eat more and do more in order to live more, both in terms of quantity and quality of life.

 

--Dean

Link to comment
Share on other sites

Admin Note: I moved TomB's post (with permission) about Luigi e al's latest paper on CR & inflammation from this thread to the Main CALERIE II Results thread, since the study was the latest installment of the ongoing analysis of that very important study of human CR. Please follow the link above to continue the discussion. 

 

--Dean

Link to comment
Share on other sites

[Admin Note: Sigh... I've moved this post by Saul back from the CALERIE study thread reference above, since unlike Tom's, this post by Saul is clearly a continuation of our BMI discussion on this thread. Saul, I hope you don't mind... --Dean]

 

Dean,
I think that you're too pessimistic about the value of a CRAN diet, including in older adults (I'm 75.) Your supposition of hip fractures doesn't work, if Dexa says different (I'm on Forteo again, and my Dexa haven't been better since before starting CR [over 20 years ago].)
And vigorous daily exercise keeps muscles in good shape.
The important study just released also doesn't jive with your assumptions.
-- Saul

Link to comment
Share on other sites

Saul wrote:

I think that you're too pessimistic about the value of a CRAN diet, including in older adults (I'm 75.) Your supposition of hip fractures doesn't work, if Dexa says different (I'm on Forteo again, and my Dexa haven't been better since before starting CR [over 20 years ago].)

And vigorous daily exercise keeps muscles in good shape.

 

Bless your heart Saul. You are a poster child for exactly the point I'm making.

 

Let's look at the facts. At age 75 and despite vigorous exercise and an extremely healthy self-described "CRAN" diet, you nevertheless need to be taking a powerful drug (Forteo) to manage your CR-induced late-onset osteoporosis to avoid a hip fracture.

 

Do you really consider your profile to be a glowing endorsement for continuing to practice serious CR into one's elder years? Apparently you do, again prompting me to question your rationality...

 

But I'm glad to hear you think the combination of CR, Forteo & daily exercise has made you invincible and immortal. Good luck with that...

 

The important study just released also doesn't jive with your assumptions.

 

Sigh... Once again Saul your are completely mistaken. You appear to be parroting Tom, and neither of you appear to have actually read the new study, or remember previous paper in this series.  

 

As I discuss in this post, the "important study just released" to which you refer, in which Luigi Fontana et al document more results from the CALERIE study, is in complete agreement with my argument that most of the benefits of CR can be achieved via a healthy, obesity-avoiding diet, without serious calorie restriction.

 

--Dean

Link to comment
Share on other sites

Hi Dean!

 

I had severe osteo way before going on CR -- probably from restricting Vitamin D intake -- in those days, the FDA claimed that itwas dangerous to take over [a very low dose], since it is a fat soluble vitamin.

 

:(xyz

 

So, CR wasn't the (primary) source of my osteo -- although it's certainly true that CR is a negative for osteo.

 

Forteo has brought me up from an extremely severe osgteoporosis to (approaching) osteopena.

 

  --  Saul 

Link to comment
Share on other sites

  • 4 weeks later...

All,

 

Al posted another new study [1] (thanks Al!) showing that a low BMI sucks for mortality risk, especially if you are a guy. These researchers followed about 100K men and 20K women in China, ranging in age from 18 to over 90 (avg age ~50) for about 7.5 years. They correlated baseline BMI with subsequent mortality.

 

As usual, they found a U-shaped mortality curve. The lowest mortality risk in men was seen in those with a BMI of 24-28. But in fact, it was really only the underweight (BMI < 18.5) folks who showed poorer survival. Here are the survival curves (combining men and women) as a function of years of follow-up (left) and hazard ratio as a function of BMI (right):

 

Zu6dkQV.pngAny9Edu.png

 

It appears underweight men (BMI < 18.5) were 37% more likely to die during follow-up than normal weight men. Conversely, overweight men, which these researchers characterized as a BMI of 24-28, had a 7% survival advantage relative to normal weight men (BMI 18.5-24):

 

Using proportional hazards (Cox) regression analysis, hazard ratios (HRs) of all-cause mortality in underweight, overweight, and obesity were 1.37 (1.19–1.57), 0.93 (0.88–0.97), and 1.03 (0.96–1.10), respectively[, relative to normal weight]. 

 

Folks like Michael, Al, and Saul will likely point to the fact that no exclusion was made for early mortality, and the skinny folks probably weren't eating impeccable diets like we do, and therefore probably weren't very healthy. Almost all of that is true, and clearly represent limitations of this study. The exception is the "probably weren't very healthy" part. In fact, by virtually all of the commonly identified markers of good health, the really skinny folks appeared to be in better shape at baseline than any of the heavier groups. Here is the table of baseline characteristics. I've highlighted in green the characteristics where the underweight folks look better than any of the other groups:

t9G6jJD.png

 

See what I mean? By every biometric measure and every demographic/lifestyle factor except exercise (skinny folks were lazier), the skinniest folks looked healthier on paper than any of the chubbier groups. And yet they were nonetheless dramatically more likely to die during follow-up.

 

To me this is more evidence suggesting that being very skinny makes you more fragile and increases your risk of dying, even if your lifestyle and biomarkers show you to be healthy.

 

--Dean

 

 

----------

[1] Medicine (Baltimore). 2016 Aug;95(32):e4327.

 

Association between body mass index and mortality in a prospective cohort of Chinese adults.

 
Sun H, Ren X, Chen Z, Li C, Chen S, Wu S, Chen Y, Yang X.
 
 
Abstract: Obesity is associated with an increased risk of diabetes mellitus, hypertension, and coronary artery disease; however, the relation between body mass index (BMI) and the risk of all-cause mortality is controversial. We prospectively examined the relationship between BMI and all-cause mortality in 123,384 Chinese men and women who participated in the Kailuan health examination study from 2006 to 2007 and 2008 to 2009. Cases included 6218 deaths (5770 men and 448 women) that occurred during a mean follow-up period of 7.39 years. Relative risk was adjusted for factors such as age, serum lipid levels (ie, triglyceride, high-density lipoprotein, and low-density lipoprotein cholesterol), history of smoking and drinking, and physical activity, as well as a medical history of hypertension, diabetes, myocardial infarction, and stroke. Within the cohort, the lowest risk of all-cause mortality was seen among persons with a BMI of 24 to 28 kg/m2 in male, and the risk was elevated among persons with BMI levels lower or higher than that range. Moreover, all-cause mortality was greatest in the group with a BMI of <18.5 kg/m2. In contrast, in female, a high BMI was associated with increased mortality, and a BMI of <18.5 kg/m2 was associated with the lowest risk. Further, a U-shaped association was seen between BMI and the risk of death from any cause among men and women, even after adjusting for confounding factors. In conclusion, underweight was associated with a substantially increased risk of all-cause mortality in males. The excess risk of all-cause mortality with a high BMI, however, was seen among females.
 
PMID: 27512844
Link to comment
Share on other sites

Actually, the relative lack of exercise compared to higher BMI is a significant problem. There's the old conundrum about exercise - people who exercise are healthier than those who don't, but is it because healthier people have the energy and capacity for exercise because they are healthier to begin with? This is the obverse of that - these skinny people may not exercise because they don't have the energy or health to do so. The other explanation - provided by Dean - is that skinnier people are lazier... which is absurd on the face of it - what would make skinny people lazy? Think about it - it's silly. If anything, one might hypothesize that skinny people are less lazy, because it's thanks to more activity that they stay skinny, they move about, instead of the more sedentary couch potatoes (sedentary is correlated with higher BMI). Therefore I reject the idea that skinny people are lazier. It seems highly possible that instead they are less able to exercise due to less vitality and health. They didn't get skinny because they exercised more - they got skinny because they were sick. Fact: they are skinnier. Question: why? There are two ways: exercise more, eat less (or combo of both). One thing we know for sure: they don't exercise more. In fact, they exercise LESS. Now, you have a serious problem - you are getting skinnier despite husbanding your energy through less exercise (compared to higher BMI)... and that's because you are sick. That's not proof by any means, but it's a highly, highly plausible hypothesis. That one variable - exercise - seriously muddies the water in the neat narrative Dean has spun here. Once again we see that being skinnier is (possibly) a proxy for hidden health problems.

 

As to smoking/drinking less - again, we have the old conundrum. One of the reasons it's been suggested that teetotalers are worse off health/lifespan-wise than light/moderate drinkers in many studies, is because those studies don't account for the alcoholics who became teetotalers and also because sicker people don't engage drinking - or cut back on unhealthy habits when they feel worse or get a diagnosis. This is a frequent confounders in supplement studies (people take up vitamins when they feel sick, or hope for a cure), diet studies (people up the quality of their diets when they feel sicker), and health habits (people engage less in unhealthy habits when feeling sicker). Again, this is not proof by any means, but another hypothesis which shows that you can't take the data of health habits as telling you much about outcomes due to confounding.

 

One could go through virtually all your highlighted data showing purportedly better health markers for the skinny folks and poke holes in this. Take for example higher HDL. But that is simply wrong - higher HDL might very well be an indicator of a shorter lifespan:

 

High Density Lipoprotein Cholesterol and the Risk of All-Cause Mortality among U.S. Veterans

 

"Conclusions 

Our results show a U-shaped relationship between HDL cholesterol and risk of all-cause mortality across all eGFR categories. The risk is modified by eGFR and cardiovascular disease."

 

Lower LDL in some studies have shown to be a hidden cancer indicator etc., etc., etc.

 

Bottom line, while suggestive, this is by no means dispositive. A lot more work is necessary to disentangle the confounders.

Link to comment
Share on other sites

Tom,

 

 

One thing we know for sure: they don't exercise more. In fact, they exercise LESS. Now, you have a serious problem - you are getting skinnier despite husbanding your energy through less exercise (compared to higher BMI)... and that's because you are sick. That's not proof by any means, but it's a highly, highly plausible hypothesis. That one variable - exercise - seriously muddies the water in the neat narrative Dean has spun here. Once again we see that being skinnier is (possibly) a proxy for hidden health problems.

 

Whose spinning narratives here? It seems to me you are speculating pretty wildly that being skinny without increased physical activity is a sign of hidden health problems. I'll certainly admit that is one possibility, but given their better health markers and lifestyle practices across the board, it seems as plausible to speculate that they were skinny because they were restrained eaters and/or had a naturally fast metabolism. With a skinny phenotype, they may just not think they need to exercise quite as much, since the reason many folks engage in physical activity is to avoid gaining weight.

 

Moreover, the skinny folks didn't report engaging in regular physical activity that much less frequently than the heavier groups (13.2% for skinny folks vs. ~15% for chubbier folks). The reason I mentioned it was that it was the one metric by which the skinny folks weren't better than all the other groups - not that they were dramatically worse (i.e. much more sedentary).

 

So Tom I really think you are reading far too much into the tiny amount less physical activity reported by the skinny folks.

 

Speaking of reading too much into things - do you actually read the studies you point to, like your HDL and veteran study, rather than just look at the abstract? If you looked at the full text, you'd see how tiny the "U-shape" was for all-cause mortality with higher HDL, particularly for people with normal kidney function (eGFR). Here is the table from the full text:

 

rO9wLdT.png

 

I've highlighted the hazard ratios for normal eGFR folks with various HDL levels. As you can see, the blip in increased mortality with HDL > 50 is tiny.

 

Bottom line, while suggestive, this is by no means dispositive. A lot more work is necessary to disentangle the confounders.

 

Agreed. Too bad such disentanglement will never happen, at least in our lifetimes.

 

--Dean

Link to comment
Share on other sites

Yes, I read the study - and furthermore, the study authors certainly don't think it's a "tiny blip" - rather, they think it's clearly and significantly a U shape... in the words of one of the authors:

 

"However, the previous studies are limited in that the number of patients in those cohorts is relatively small compared with what a big data approach enabled us to see in our new research," he said. "Big data allow a more nuanced examination of the relationship between HDL cholesterol and risk of death across the full spectrum of HDL cholesterol levels."

Research data showed a relationship between HDL cholesterol levels and mortality as a U-shaped curve with the risk of death increased at both ends of the spectrum. "Too low and too high are both associated with higher risk of death," Al-Aly said.

Whether maintaining intermediate HDL cholesterol levels may increase longevity will need to be explored in future studies, Al-Aly said."

 

I also reject the suggestion that the key is the "tiny amount less physical activity" - not only is it not tiny, but again, to stress, these are skinny folks. Perhaps you could make a case that this is insignificant and tiny if it were dealing with equivalent BMIs, or even higher BMIs - after all, if a higher BMI person engaging in less exercise you don't demand an explanation as to why, they have higher BMI while exercising less - because it seems self-explanatory. The problem you have is that you have to explain not just how it is that someone engages in less exercise but has the same BMI (which would be bad enough), but how it is that they engage in less exercise but have LOWER BMI.

Furthermore, even if we 100% dismiss the role of exercise in energy balance and BMI, and put the differences all down to energy intake - eating less and therefore having lower BMI, one could argue they are eating less because they have less appetite due to being sick.  

The other factor is that BMI is too blunt a number. What if the higher BMI folks have more muscle than the skinny ones and therefore their higher BMI is a marker of better health, musculature (perhaps due to more exercise), whereas it is a well-known phenomenon that there exist the "skinny fat" of low BMI but high body fat and low muscle mass (very unhealthy phenotype) - now, obviously this does not apply to the much higher BMI range (as it's not plausible that the population studied is composed of body builders) - but in a very narrow range of BMI it might be a small factor, small but enough to tip the balance. 

In short, I don't believe this study is a strong argument one way or another about the relative health of BMI ranges except perhaps at the extremes - for example, what if the group of below 18.5 BMI was mostly composed of below 15-16? 

Link to comment
Share on other sites

Tom,

 

You (and the authors of the HDL study) can bluster until you're blue in the face about how these effects aren't tiny, but the statistics don't lie. The increased mortality risk for so-called "high" HDL for people with normal kidney function was 3-5%. The reduction in reported regular physical activity was less than 2%. I'd say these are tiny differences, especially in comparison to the >40% increased risk of mortality observed in the very skinny folks relative to the overweight (BMI 24-28) folks.

 

The problem you have is that you have to explain not just how it is that someone engages in less exercise but has the same BMI (which would be bad enough), but how it is that they engage in less exercise but have LOWER BMI.

 

Hmmm... let me think... Perhaps they friggin' eat less!?!?

 

one could argue they are eating less because they have less appetite due to being sick.  

 

You could try to argue that. But that is why I pointed out the fact that for all intents and purposes, the skinny folks appeared at baseline to be healthier than the heavier folks.  

 

What if the higher BMI folks have more muscle than the skinny ones and therefore their higher BMI is a marker of better health, musculature (perhaps due to more exercise), whereas it is a well-known phenomenon that there exist the "skinny fat" of low BMI but high body fat and low muscle mass (very unhealthy phenotype) 

 

Your singing to the choir on this one. It could very well be the skinny folks had less muscle and were "skinny fat" - eating like a bird and not doing much exercise. Unfortunately, this is how at least some people used to interpret the "calories, calories, calories" mantra for CR.

 

Now hopefully we know better. Rather, to minimize mortality risk, it is important to maintain muscle mass and good cardiovascular performance (i..e. VO2Max), and to do so requires engaging in quite a bit more low-intensity aerobic exercise and eating quite a bit more to support that exercise, than would otherwise be necessary.

 

--Dean

Link to comment
Share on other sites

I tell you what - here's how we could settle this: instead of trying to repeal the second law of thermodynamics as an explanation for how bodies that expend less energy through exercise none the less retain less mass than bodies that expend more energy, or speculate about energy intake with no data to back it up, I'd be quite happy to accept the premise if we could just limit the comparison to BMI cohorts of less than 18.5 vs higher that both have the same exercise amounts. Surely they must have a subset that meets that criteria in a big enough group - so just select those who exercise the same amount. This would get around the claim that the different exercise levels are due to different health status. Works for me, should work for you. And the data should be available given the sample size. Deal? 

Link to comment
Share on other sites

Ordinarily, I'd say it's a matter of a quick email to the authors of the study to see if they have the raw data, but given that this is a Chinese study, I conceded this is unlikely. Rather, I'm suggesting that if you can cite a study that shows same activity levels across BMI rates correlated to all-cause mortality, I'll accept the result for skinny folks excess mortality because it'll take the health status confounder off the table. And btw., note, this is a very generous reading - I even let the HDL issue slide.

Link to comment
Share on other sites

  • 3 weeks later...

More Evidence for the Obesity Paradox

 

Well, here is another one of those studies [1] (thanks to Al) which shows it doesn't hurt to be fat if you're metabolically healthy.

 

Researchers followed ~4500 older men and women (avg age 74) for 11 years to see who died as a function of BMI and metabolic risk factors.

 

They stratified BMI in the usual categories, 18.5 - 24.9, 25 - 29.9, 30 - 34.9, 35+. For risk factors they said high risk folks were people with two or more of the following health conditions at baseline:

  • High triglycerides (≥150 mg/dL)
  • Low HDL (< 40 mg/dL for men, <50 for women)
  • High blood pressure (≥ 130/85 mm Hg) or hypertension diagnosis 
  • High fasting glucose (≥ 100 mg/dL) or diabetes diagnosis

Interesting they didn't count high LDL as a risk factor, although they did include it one of the covariant models in the table below.

 

Here were the baseline characteristics for subjects stratified by BMI and whether they were metabolically healthy or not (< 2 of the above risk factors = healthy), indicated by the "yes/no" columns under each BMI category. I've also added (in blue) the percent of people in each BMI category who were metabolically healthy at baseline:

eByKCSd.png

As you can see from the blue, the heavier the BMI category the lower the percentage of people who were metabolically healthy at baseline. But if you were in the healthy class at baseline (< 2 risk factors), your health parameters looked pretty similar, regardless of which BMI category you were in, pretty much by definition.

 

So who died?

 

Here is the table with all-cause mortality results stratified again by weight category and good metabolic health status (yes/no):

 

29Ao7HX.png

 

It seemed to me their "Model 2" was the fairest, although they all showed pretty much the same thing.

 

What I've highlighted above is the all-cause mortality hazard ratios for people from each BMI category who were healthy at baseline. As you can see, assuming you were one of the lucky few who were metabolically healthy at baseline, you were better off being heavy than you were to if you were in the normal BMI category (18.5-25). In fact, metabolically healthy obese folks (BMI 30-35) were 40% less likely to die during follow-up than normal-weight individuals. The classic "obesity paradox".

 

You might object - what about thin (former) smokers? Mightn't they have increased the mortality rate in the normal weight group? Apparently not, as highlighted in this table:

 

Fr2z3Na.png

As you can see, even among never smokers, metabolically healthy obese folks had a nearly 50% lower risk of dying during follow-up than did the normal weight folks.

 

It seems to me the three most plausible explanations for this seemingly paradoxical result are:

  1. Heavier people are more robust to the slings and arrows of old age as a result of having extra meat on their bones.
  2. The healthy fat folks might have freakishly good genes enabling them to maintain their health despite obesity, which also give them a survival advantage.
  3. The healthy fat folks might be pursuing a healthier lifestyle (e.g. exercising) in order to maintain their health despite their obesity, giving them a survival advantage.

But whatever the explanation(s), it's interesting to see once again that being fat in old age is not a disadvantage if you enter old age with a healthy metabolism. In fact a fat and healthy phenotype looks like it might be a distinct advantage relative to being normal weight and healthy in old age.

 

--Dean

 

--------------

[1] Journal of Nutrition in Gerontology and Geriatrics  Volume 35, 2016 - Issue 3, Page 161-176 | Published online: 25 Aug 2016

 

Metabolic Health Status and the Obesity Paradox in Older Adults
 
Feon W. Cheng , MPH, RD, CHTS-CP, Xiang Gao , MD, PhD, Diane C. Mitchell , MS, RD, Craig Wood , MS, David D. K. Rolston , MD, Christopher D. Still , DO & Gordon L. Jensen , MD, PhD
 
 
ABSTRACT
 
The explanation for reduced mortality among older persons with overweight or class I obesity compared to those of desirable weight remains unclear. Our objective was to investigate the joint effects of body mass index (BMI) and metabolic health status on all-cause mortality in a cohort of advanced age. Adults aged 74 ± 4.7 (mean ± SD) years at baseline (n = 4551) were categorized according to BMI (18.5–24.9, 25.0–29.9, 30.0–34.9, and ≥35.0 kg/m2) and the presence or absence of a metabolically healthy phenotype (i.e., 0 or 1 risk factors based on a modified Adult Treatment Panel III). Metabolically unhealthy was ≥2 risk factors. There were 2294 deaths over a mean 10.9 years of follow up. Relative to metabolically healthy desirable weight, metabolically healthy overweight or class I obesity was not associated with a greater mortality risk (HR 0.90; 95 CI% 0.73–1.13 and HR 0.58; 95 CI% 0.42–0.80, respectively) (P-interaction <0.001). Results remained consistent in rigorous sensitivity analyses. The “obesity paradox” may be partially explained by the inclusion of metabolically healthy overweight and obese older persons, who do not have elevated mortality risk, in population studies of BMI and mortality.
 
KEYWORDS: Prospective cohort study, obesity, metabolic health, elderly, mortality, 
 
DOI: 10.1080/21551197.2016.1199004
PMID Not available
Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...