Obesity, Diabetes Linked to Cancers

http://health.usnews.com/usnews/heal...ty-diabetes-li
nked-to-cancers.htm

Obesity, Diabetes Linked to Cancers
Studies find effects on breast, prostate and colorectal tumors
By Ed Edelson
Posted 12/7/07
FRIDAY, Dec. 7 (HealthDay News) -- Obesity and diabetes -- risk factors so
often linked to heart disease -- can also affect the incidence and severity
of cancer, a collection of four new studies suggests.

The findings, presented Friday at the American Association for Cancer
Research's Sixth Annual International Conference on the Frontiers of Cancer
Prevention Research in Philadelphia, link weight gain and diabetes to a
number of malignancies, including breast, prostate and colorectal cancer.

"All of these are consistent with what we would expect with the occurrence
of each of these cancers or cancer survival," said Elizabeth Platz,
associate professor of epidemiology at the Johns Hopkins Bloomberg School
of Public Health. "Metabolic perturbations enhance certain cancers. Insulin
and other hormonal factors influence cell growth and make cells multiply."

Women with diabetes have a 50 percent increased risk of developing
colorectal cancer, according to the first study, by researchers at the
University of Minnesota. The group, led by Andrew Flood, assistant
professor of epidemiology and public health, followed more than 45,000
women enrolled originally in a breast cancer detection program for more
than eight years.

The increased incidence of colorectal cancer remained significant after all
possibly confounding factors were taken into account. While the reason for
the increased risk is not known, Flood said it could be due to the elevated
levels of insulin seen with diabetes.

High levels of insulin in diabetic women could explain a threefold higher
risk of death from breast cancer, said the second study, by researchers at
Yale University. They measured blood levels of C-peptide, a marker of
insulin secretion, in women in a long-term study of breast cancer. Over an
eight-year period, the women in the highest third of C-peptide levels had
twice the risk of dying from breast cancer, compared to women in the bottom
third, the researchers said.

Another study, by researchers at the Johns Hopkins Bloomberg School of
Public Health, found that weight gain after a diagnosis of invasive breast
cancer could significantly increase a women's risk of death from the
cancer.

The study of more than 4,000 women with breast cancer classified them by
body mass index, a ratio of weight to height. For obese women, the risk of
dying of breast cancer was 2.4 times greater than for women with a normal
body weight, a relationship that persisted when age, menopausal status and
smoking were taken into account.

Another Johns Hopkins study provided a possible explanation for the lower
risk of prostate cancer seen in men with diabetes. The researchers matched
264 men diagnosed with the cancer with a group of 264 cancer-free men,
measuring C-peptide levels in both groups.

Men with elevated blood levels of C-peptide when the study started were
one-third less likely to develop prostate cancer than those with lower
levels. Men with higher C-peptide levels had half the risk of developing
prostate cancer confined to the prostate.

The protective effect of those high levels could be due to the activity of
insulin in relation to the male hormone testosterone, Plantz said.
C-peptide derives from the same parent molecule as insulin, and insulin is
known to reduce the activity of testosterone, which stimulates the growth
of prostate cancer, she said.

The possible protective effect of insulin against prostate cancer could
offer a mirror image of the negative effect of estrogen -- the female sex
hormone -- in breast cancer, said Dr. Rexford Ahima, professor of medicine
at the University of Pennsylvania.

"We have known for years that women who are obese are at high risk of
breast cancer," Ahima said. "Fat tissue makes estrogen, which promotes
breast cancer. The frightening thing is that the more obese you are, the
greater the risk you have of dying of cancer. For every increase of 10
kilograms, 14 pounds, there is a 14 percent increased risk of breast cancer
death."

So, instead of thinking of obesity just as a risk factor for heart disease,
its effects on cancer must also be taken into account, Platz said. "In
general, it is a good thing to do to avoid obesity," she said. "That is a
message consistent with what we know about good health."

More information

On Sat, 8 Dec 2007, Kenny wrote:

> FRIDAY, Dec. 7 (HealthDay News) -- Obesity and diabetes -- risk factors so
> often linked to heart disease -- can also affect the incidence and severity
> of cancer, a collection of four new studies suggests.

Did they check for obesity vrs diabetes to make sure they both apply, or
did they just lump the two for dramatic effect?

> Women with diabetes have a 50 percent increased risk of developing
> colorectal cancer, according to the first study, by researchers at the
> University of Minnesota.

And what about obese women who are not diabetic?

> High levels of insulin in diabetic women could explain a threefold higher
> risk of death from breast cancer, said the second study, by researchers at
> Yale University.

Again, diabetic risks, what about obesity risks?

> The study of more than 4,000 women with breast cancer classified them by
> body mass index, a ratio of weight to height. For obese women, the risk of
> dying of breast cancer was 2.4 times greater than for women with a normal
> body weight, a relationship that persisted when age, menopausal status and
> smoking were taken into account.

Ah, finally... I'm glad they took into account other factors, but what
about diabetes? How many of the 4,000 women are diabetic also?

You see, that's the problem with studies like this. Diabetes continues to
be a known problem when it comes to cancer. However, not every obese
person has diabetes, and not every diabetic is obese. I for one was
checked not too long ago for diabetes, and I am glad to say that I don't
have it.

How does obesity affect cancer independently of diabetes? Every specific
reference made in this story is in reference to diabetes, not obesity.
The ONLY time they give a result based on obesity, they do not account for
diabetes. Yet another worthless study.

The Master <tardis@nospam.sdf.lonestar.org.nospam> wrote in
news:Pine.NEB.4.64.0712101404290.19685@sdf.lonesta r.org:

<...>
>> The study of more than 4,000 women with breast cancer classified
>> them by body mass index, a ratio of weight to height. For obese
>> women, the risk of dying of breast cancer was 2.4 times greater
>> than for women with a normal body weight, a relationship that
>> persisted when age, menopausal status and smoking were taken
>> into account.
>
> Ah, finally... I'm glad they took into account other factors,
> but what about diabetes? How many of the 4,000 women are
> diabetic also?
<...>


If you really want to be nit-picking :-) , you should note that they
talk about *dying of breast cancer. It would be interesting to know if
late discovery and problems with treatment are included in that
number.


Lisbeth.

----
The day I don't learn anything new is the day I die.

*What we know is not nearly as interesting as *how we know it.



--
Posted via a free Usenet account from http://www.teranews.com

On Mon, 10 Dec 2007, Lisbeth Andersson wrote:

>> Ah, finally... I'm glad they took into account other factors,
>> but what about diabetes? How many of the 4,000 women are
>> diabetic also?
>
> If you really want to be nit-picking :-) , you should note that they
> talk about *dying of breast cancer. It would be interesting to know if
> late discovery and problems with treatment are included in that
> number.

And family history...

The assumption on a study like that is that the sample size is such that
it negates those types of unknowns. The non-diabetic skinny sample should
ideally have the same late discovery issues as the fat diabetic group...

So, lets look at just raw numbers for a second... 4,000 women total for
the "obesity link" study, that didn't take into account for diabetes. How
many females are there in the world? Of those, how many now have breast
cancer? I honestly don't know, but I would be amazed if 4,000 would even
be 1% of that figure. I would guestimate that 4,000 is much less then 1%.

That's the problem with population based statistics, the sample size is
ALWAYS so frickin small, the "margin for error" is inane. Might as well
say Ted Kennedy represents all of Congress.

In article <20071208005613.CFAF94E55B@outpost.zedz.net>, Kenny wrote:

<In short, study results indicating higher cancer rates in those with
diabetes>

Were the extra cancers caused by the diabetic condition, or by the
overweightness/obesity that is a major to essential contributing factor to
mosty diabetes?

- Don Klipstein (don@misty.com)

In article <slrnflrs7m.7sl.don@manx.misty.com>,
Don Klipstein <don@manx.misty.com> wrote:
>In article <20071208005613.CFAF94E55B@outpost.zedz.net>, Kenny wrote:

><In short, study results indicating higher cancer rates in those with
>diabetes>

> Were the extra cancers caused by the diabetic condition, or by the
>overweightness/obesity that is a major to essential contributing factor to
>mosty diabetes?

Or is it even due to the presence of the gene for
Syndrome X, even if it is not expressed? Crude
statistical techniques CANNOT separate out the
various causes, but only report association.
It is not the case that people who have regained
a "normal" weight by dieting are the same as those
whose weight is that naturally.

Finding causes in complicated situations is not easy,
and the relevant experiments cannot even be done on
people. Not only is correlation not causation, but
it can even go in the opposite direction.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hrubin@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558

Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
> Don Klipstein <don@manx.misty.com> wrote
>> Kenny wrote

>> <In short, study results indicating higher cancer rates in those with diabetes>

>> Were the extra cancers caused by the diabetic condition,
>> or by the overweightness/obesity that is a major to
>> essential contributing factor to mosty diabetes?

> Or is it even due to the presence of the gene
> for Syndrome X, even if it is not expressed?

Only a subset of cancers have a genetic link.

> Crude statistical techniques CANNOT separate
> out the various causes, but only report association.

They are quite effective for working out what has a genetic component tho.

> It is not the case that people who have regained a "normal" weight
> by dieting are the same as those whose weight is that naturally.

Thats arguable with cancer risk.

> Finding causes in complicated situations is not easy, and
> the relevant experiments cannot even be done on people.

There is more than just experiments available with questions like that.

> Not only is correlation not causation,
> but it can even go in the opposite direction.

And we can still work out the basics like the association
of smoking with various medical problems anyway.

>><In short, study results indicating higher cancer rates in those with
>>diabetes>
>
>> Were the extra cancers caused by the diabetic condition, or by the
>>overweightness/obesity that is a major to essential contributing factor to
>>mosty diabetes?
>
> Or is it even due to the presence of the gene for
> Syndrome X, even if it is not expressed? Crude
> statistical techniques CANNOT separate out the
> various causes, but only report association.
> It is not the case that people who have regained
> a "normal" weight by dieting are the same as those
> whose weight is that naturally.
>
> Finding causes in complicated situations is not easy,
> and the relevant experiments cannot even be done on
> people. Not only is correlation not causation, but
> it can even go in the opposite direction.
> --
> This address is for information only. I do not claim that these views
> are those of the Statistics Department or of Purdue University.
> Herman Rubin, Department of Statistics, Purdue University
> hrubin@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558

You heard it from the actuary, guys. Now, just let things be. All the
insults you hurl at ppl you'll never meet IRL don't seem to be making any
kind of impact.
Stop the corporate greed, and you might see a turn around in the way we all
look, but since you want to see all fat ppl gone anyway, just let us choose
the way we take ourselves out. It might be slow or quick, but if it pains
your eyes to see this, you can always choose to look away.

On Tue, 11 Dec 2007, "Charmander" <tarrah@personals-scripts.com> wrote:
>Stop the corporate greed, and you might see a turn around in the way we
>all look, but since you want to see all fat ppl gone anyway, just let us
>choose the way we take ourselves out. It might be slow or quick, but if
>it pains your eyes to see this, you can always choose to look away.

Nice k00k rant, chubs.

On Tue, 11 Dec 2007, Herman Rubin wrote:

> Finding causes in complicated situations is not easy,
> and the relevant experiments cannot even be done on
> people. Not only is correlation not causation, but
> it can even go in the opposite direction.

That's what many have been saying all along (myself included). Yet when
it comes to the politically charged hate for one class of unpopular
people, cool heads and rational thought rarely come into play.

In article <5s840jF1784ndU1@mid.individual.net>,
Rod Speed <rod.speed.aaa@gmail.com> wrote:
>Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
>> Don Klipstein <don@manx.misty.com> wrote
>>> Kenny wrote

>>> <In short, study results indicating higher cancer rates in those with diabetes>

>>> Were the extra cancers caused by the diabetic condition,
>>> or by the overweightness/obesity that is a major to
>>> essential contributing factor to mosty diabetes?

>> Or is it even due to the presence of the gene
>> for Syndrome X, even if it is not expressed?

>Only a subset of cancers have a genetic link.

This is by no means clear, as more and more proteins
are being found which affect cancers.

>> Crude statistical techniques CANNOT separate
>> out the various causes, but only report association.

>They are quite effective for working out what has a genetic component tho.

Not really.

>> It is not the case that people who have regained a "normal" weight
>> by dieting are the same as those whose weight is that naturally.

>Thats arguable with cancer risk.

Why? If the basic metabolic processes differ in
the two populations, cancer risk should be different.

>> Finding causes in complicated situations is not easy, and
>> the relevant experiments cannot even be done on people.

>There is more than just experiments available with questions like that.

Demographic data are not reliable here; as Oliver Wendell
Holmes said when asked how to live to a ripe old age,
"Choose your grandparents carefully."

>> Not only is correlation not causation,
>> but it can even go in the opposite direction.

>And we can still work out the basics like the association
>of smoking with various medical problems anyway.

Sometimes and sometimes not. In this case, there were
even reasonable indications that hot gases and tobacco
byproducts could irritate the lungs. We have no such
indication between weight and the growth of cancers.

Medical studies indicate that those who use strong
methods to lose weight are not medically "normal",
unlike those who naturally weigh less. Even the
"normal" weight for a person is totally unclear.


--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hrubin@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558

Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
> Rod Speed <rod.speed.aaa@gmail.com> wrote
>> Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
>>> Don Klipstein <don@manx.misty.com> wrote
>>>> Kenny wrote

>>>> <In short, study results indicating higher cancer rates in those with diabetes>

>>>> Were the extra cancers caused by the diabetic condition,
>>>> or by the overweightness/obesity that is a major to
>>>> essential contributing factor to mosty diabetes?

>>> Or is it even due to the presence of the gene
>>> for Syndrome X, even if it is not expressed?

>> Only a subset of cancers have a genetic link.

> This is by no means clear,

Yes it is.

> as more and more proteins are being found which affect cancers.

And more and more environmental causes of cancer are found too.

Most obviously with the main one, smoking. Which is now recognised
to produce a variety of other serious fatal medical problems as well.

>>> Crude statistical techniques CANNOT separate
>>> out the various causes, but only report association.

>> They are quite effective for working out what has a genetic component tho.

> Not really.

Yes, really. Its how the major genetic effects are worked out after all.

>>> It is not the case that people who have regained a "normal" weight
>>> by dieting are the same as those whose weight is that naturally.

>> Thats arguable with cancer risk.

> Why?

Because that has clearly been established with other fatal medical problems like diabetes.

> If the basic metabolic processes differ in the two populations,

They dont. The reason for the vast bulk of the differences in weight
seen are just that the morbidly obese are actualy stupid enough to
keep shovelling many more calories into their mouths than they burn.

Thats why stomach banding and stomach stapling etc works.

> cancer risk should be different.

Easy to claim, hell of a lot harder to actually substantiate that claim.

Even if the appetite difference or will power does have a
genetic component, that doesnt necessarily mean that that
genetic component has any effect what so ever on cancer risk.

>>> Finding causes in complicated situations is not easy, and
>>> the relevant experiments cannot even be done on people.

>> There is more than just experiments available with questions like that.

> Demographic data are not reliable here;

Didnt say anything about demographic data.

> as Oliver Wendell Holmes said when asked how to live
> to a ripe old age, "Choose your grandparents carefully."

That sort of glib over simplification doesnt alter the fact that its clear
that as the diet of particular groups change over time, or due to
migration, there is a clear effect of the changed diet on lifespan etc.

>>> Not only is correlation not causation,
>>> but it can even go in the opposite direction.

>> And we can still work out the basics like the association
>> of smoking with various medical problems anyway.

> Sometimes and sometimes not.

Mostly we can.

> In this case, there were even reasonable indications that
> hot gases and tobacco byproducts could irritate the lungs.

Irrelevant to whether the act of smoking has a clear statistical
correlation with a range of serious and fatal medical conditions.
Thats just the detail of how smoking does that.

> We have no such indication between weight and the growth of cancers.

Yet.

> Medical studies indicate that those who use strong
> methods to lose weight are not medically "normal",
> unlike those who naturally weigh less.

And those that lose weight just be not shovelling as many
calories into their mouths are perfectly medically 'normal' as
long as they dont over do that and end up starving themselves.

> Even the "normal" weight for a person is totally unclear.

Sure, but thats irrelevant to whether there is a hell of a lot more than
JUST experimentation that useful in determining cause and effect.
The other obvious approach is to use natural variations in behavour
and see what effect that is having on the medical condition being
considered, particularly when that variation is due to effects like
the change in society behaviour over time and migration etc.

In article <5sao46F18dmkpU1@mid.individual.net>,
Rod Speed <rod.speed.aaa@gmail.com> wrote:
>Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
>> Rod Speed <rod.speed.aaa@gmail.com> wrote
>>> Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
>>>> Don Klipstein <don@manx.misty.com> wrote
>>>>> Kenny wrote

>>>>> <In short, study results indicating higher cancer rates in those with diabetes>

>>>>> Were the extra cancers caused by the diabetic condition,
>>>>> or by the overweightness/obesity that is a major to
>>>>> essential contributing factor to mosty diabetes?

>>>> Or is it even due to the presence of the gene
>>>> for Syndrome X, even if it is not expressed?

>>> Only a subset of cancers have a genetic link.

>> This is by no means clear,

>Yes it is.

>> as more and more proteins are being found which affect cancers.

>And more and more environmental causes of cancer are found too.

>Most obviously with the main one, smoking. Which is now recognised
>to produce a variety of other serious fatal medical problems as well.

>>>> Crude statistical techniques CANNOT separate
>>>> out the various causes, but only report association.

>>> They are quite effective for working out what has a genetic component tho.

>> Not really.

>Yes, really. Its how the major genetic effects are worked out after all.

>>>> It is not the case that people who have regained a "normal" weight
>>>> by dieting are the same as those whose weight is that naturally.

>>> Thats arguable with cancer risk.

>> Why?

>Because that has clearly been established with other fatal medical problems like diabetes.

>> If the basic metabolic processes differ in the two populations,

>They dont. The reason for the vast bulk of the differences in weight
>seen are just that the morbidly obese are actualy stupid enough to
>keep shovelling many more calories into their mouths than they burn.

The current careful evidence indicates that this is
not the whole problem. There are different metabolic
effects; I went in one year from being unable to gain
weight to having problems with weighing too much. In
fact, may obese people eat little.

>Thats why stomach banding and stomach stapling etc works.

It reduces weight, but also has other effects.

>> cancer risk should be different.

>Easy to claim, hell of a lot harder to actually substantiate that claim.

>Even if the appetite difference or will power does have a
>genetic component, that doesnt necessarily mean that that
>genetic component has any effect what so ever on cancer risk.

>>>> Finding causes in complicated situations is not easy, and
>>>> the relevant experiments cannot even be done on people.

>>> There is more than just experiments available with questions like that.

>> Demographic data are not reliable here;

>Didnt say anything about demographic data.

>> as Oliver Wendell Holmes said when asked how to live
>> to a ripe old age, "Choose your grandparents carefully."

>That sort of glib over simplification doesnt alter the fact that its clear
>that as the diet of particular groups change over time, or due to
>migration, there is a clear effect of the changed diet on lifespan etc.

Using changes over time, unless they are drastic, takes
a large number of years to substantiate, especially if
there are other random effects. I mean decades here.

>>>> Not only is correlation not causation,
>>>> but it can even go in the opposite direction.

>>> And we can still work out the basics like the association
>>> of smoking with various medical problems anyway.

>> Sometimes and sometimes not.

>Mostly we can.

>> In this case, there were even reasonable indications that
>> hot gases and tobacco byproducts could irritate the lungs.

>Irrelevant to whether the act of smoking has a clear statistical
>correlation with a range of serious and fatal medical conditions.
>Thats just the detail of how smoking does that.

>> We have no such indication between weight and the growth of cancers.

>Yet.

>> Medical studies indicate that those who use strong
>> methods to lose weight are not medically "normal",
>> unlike those who naturally weigh less.

>And those that lose weight just be not shovelling as many
>calories into their mouths are perfectly medically 'normal' as
>long as they dont over do that and end up starving themselves.

They show the same metabolic effects as "normal" people who
are starving themselves. In most, but not all, cases, those
who lose a lot of weight get a "starving" metabolism.

>> Even the "normal" weight for a person is totally unclear.

>Sure, but thats irrelevant to whether there is a hell of a lot more than
>JUST experimentation that useful in determining cause and effect.
>The other obvious approach is to use natural variations in behavour
>and see what effect that is having on the medical condition being
>considered, particularly when that variation is due to effects like
>the change in society behaviour over time and migration etc.

Natural variations can only show correlation, and there are
many known cases of correlation going in the opposite direction
to causation.


--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hrubin@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558

Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
> Rod Speed <rod.speed.aaa@gmail.com> wrote
>> Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
>>> Rod Speed <rod.speed.aaa@gmail.com> wrote
>>>> Herman Rubin <hrubin@odds.stat.purdue.edu> wrote
>>>>> Don Klipstein <don@manx.misty.com> wrote
>>>>>> Kenny wrote

>>>>>> <In short, study results indicating higher cancer rates in those with diabetes>

>>>>>> Were the extra cancers caused by the diabetic condition,
>>>>>> or by the overweightness/obesity that is a major to
>>>>>> essential contributing factor to mosty diabetes?

>>>>> Or is it even due to the presence of the gene
>>>>> for Syndrome X, even if it is not expressed?

>>>> Only a subset of cancers have a genetic link.

>>> This is by no means clear,

>> Yes it is.

>>> as more and more proteins are being found which affect cancers.

>> And more and more environmental causes of cancer are found too.

>> Most obviously with the main one, smoking. Which is now recognised
>> to produce a variety of other serious fatal medical problems as well.

>>>>> Crude statistical techniques CANNOT separate
>>>>> out the various causes, but only report association.

>>>> They are quite effective for working out what has a genetic component tho.

>>> Not really.

>> Yes, really. Its how the major genetic effects are worked out after all.

>>>>> It is not the case that people who have regained a "normal" weight
>>>>> by dieting are the same as those whose weight is that naturally.

>>>> Thats arguable with cancer risk.

>>> Why?

>> Because that has clearly been established with
>> other fatal medical problems like diabetes.

>>> If the basic metabolic processes differ in the two populations,

>> They dont. The reason for the vast bulk of the differences in weight
>> seen are just that the morbidly obese are actualy stupid enough to
>> keep shovelling many more calories into their mouths than they burn.

> The current careful evidence indicates that this is not the whole problem.

Rubbish. There is no other viable explanation for the
dramatic increase in the incidence of morbid obesity and
just gross overweight right thruout the modern first world.

> There are different metabolic effects;

Yes, but that was true 100 years ago too.

And you never saw mobid obesity or even gross overweight
in concentration camps except with a few of the guards either.

> I went in one year from being unable to gain
> weight to having problems with weighing too
> much. In fact, may obese people eat little.

Very few eat little in fact.

>> Thats why stomach banding and stomach stapling etc works.

> It reduces weight, but also has other effects.

Irrelevant to whether its the rigorous scientific evidence
that obesity is an excess of calories over what you burn.

You also get the same effect with concentration camps too where
the intake of calories is drastically reduced in a situation where the
individual has no control over the amount of calories they injest.

>>> cancer risk should be different.

>> Easy to claim, hell of a lot harder to actually substantiate that claim.

>> Even if the appetite difference or will power does have a
>> genetic component, that doesnt necessarily mean that that
>> genetic component has any effect what so ever on cancer risk.

>>>>> Finding causes in complicated situations is not easy, and
>>>>> the relevant experiments cannot even be done on people.

>>>> There is more than just experiments available with questions like that.

>>> Demographic data are not reliable here;

>> Didnt say anything about demographic data.

>>> as Oliver Wendell Holmes said when asked how to live
>>> to a ripe old age, "Choose your grandparents carefully."

>> That sort of glib over simplification doesnt alter the fact that its clear
>> that as the diet of particular groups change over time, or due to
>> migration, there is a clear effect of the changed diet on lifespan etc.

> Using changes over time, unless they are drastic,
> takes a large number of years to substantiate,

Irrelevant to whether its still rigorous scientific evidence.

> especially if there are other random effects.

There arent with obesity.

> I mean decades here.

Irrelevant to whether its still rigorous scientific
evidence that doesnt involve experimentation.

>>>>> Not only is correlation not causation,
>>>>> but it can even go in the opposite direction.

>>>> And we can still work out the basics like the association
>>>> of smoking with various medical problems anyway.

>>> Sometimes and sometimes not.

>> Mostly we can.

>>> In this case, there were even reasonable indications that
>>> hot gases and tobacco byproducts could irritate the lungs.

>> Irrelevant to whether the act of smoking has a clear statistical
>> correlation with a range of serious and fatal medical conditions.
>> Thats just the detail of how smoking does that.

>>> We have no such indication between weight and the growth of cancers.

>> Yet.

>>> Medical studies indicate that those who use strong
>>> methods to lose weight are not medically "normal",
>>> unlike those who naturally weigh less.

>> And those that lose weight just be not shovelling as many
>> calories into their mouths are perfectly medically 'normal' as
>> long as they dont over do that and end up starving themselves.

> They show the same metabolic effects as "normal" people
> who are starving themselves. In most, but not all, cases,
> those who lose a lot of weight get a "starving" metabolism.

No they dont as long as they only reduce their calorie intake to a
little less than they burn or increase what they burn by exercise etc.

>>> Even the "normal" weight for a person is totally unclear.

>> Sure, but thats irrelevant to whether there is a hell of a lot more than
>> JUST experimentation that useful in determining cause and effect.
>> The other obvious approach is to use natural variations in behavour
>> and see what effect that is having on the medical condition being
>> considered, particularly when that variation is due to effects like
>> the change in society behaviour over time and migration etc.

> Natural variations can only show correlation, and there are many
> known cases of correlation going in the opposite direction to causation.

And while that is true, its a relatively rare effect.

You claim about the significance of the fact that experimentation
often isnt feasible with humans is just plain wrong. While that
certainly makes it harder to prove the proposition, its still perfectly
possible, most obviously with the effect of smoking on health.

Its just as true of obesity and how the individual ends up obese.

On Fri, 13 Dec 2007, Herman Rubin wrote:

>> They dont. The reason for the vast bulk of the differences in weight
>> seen are just that the morbidly obese are actualy stupid enough to
>> keep shovelling many more calories into their mouths than they burn.
>
> The current careful evidence indicates that this is
> not the whole problem. There are different metabolic
> effects; I went in one year from being unable to gain
> weight to having problems with weighing too much. In
> fact, may obese people eat little.

There are trolls that you can talk to, come close to actually having a
discussion with them. Then there is Rod Speed... He is in a class all
his own, thank the gods... He's the only person I have in my kill file
from this group.