Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Cholesterol and mortality: Evidence of a U-shape Relationship

nature.com

Post reply on HN
Print view View original

41–50 of 66 posts

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#41

Post by myle »

stavros wrote: post #9
jeffbee wrote: post #8

Earlier quoted context omitted.

170 would be in the lowest quartile for American adults, no? Seems like an unreasonable limit.

I'm in Greece, but it seems fairly unreasonable here as well. Other labs have the limit at 200 (it varies per lab), but 170 seems too low.

Greeks tend to have higher "normal" cholesterol levels. In other labs I have seen in Greece, they write as the limit 200.

I know of someone with TC around 240 and the doctor didn't prescribe statins. Not a medical advice.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#42

Post by thaumasiotes »

radford-neal wrote: post #35
sooheon wrote: post #14

This is not a new finding. It's common enough to have a name: the cholesterol paradox ( https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6374572/ ). Also replicated in a more recent n=100k Denmark study: https://www.bmj.com/content/371/bmj.m4266 These are all observational studies, meaning cause and effect is not established. It's relevant that people with some cancers will have very low cholesterol (the cancer apparentl…

"...there are no strong theories for the diseases caused by low cholesterol..." Well, the body has a whole mechanism for manufacturing cholesterol, so it must be good for something. It seems obvious that there's some level at which low cholesterol must cause death. A U-shaped cholesterol-mortality curve is pretty much guaranteed to exist. The only question is where the minimum is.

> Well, the body has a whole mechanism for manufacturing cholesterol, so it must be good for something. It seems obvious that there's some level at which low cholesterol must cause death.

Agreed.

> A U-shaped cholesterol-mortality curve is pretty much guaranteed to exist.

Yes, this is a necessary consequence of the first observation. Mortality will go to 100% if cholesterol is too low or too high.

> The only question is where the minimum is.

But I don't think this is right. You can calculate a mortality-by-cholesterol-level curve for the population, and it will have a minimum somewhere. But it doesn't follow that it would be good for you, even in a sense restricted to all-cause mortality, if you moved your personal cholesterol level closer to the minimum value on that curve. Most people in the population aren't like you.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#43

Post by tails4e »

stavros wrote: post #11
pocw wrote: post #2

Can someone tl;dr? Not that the paper is too long, but I think I'd need to go back to university and major in biology or statistics to understand what they discovered and more importantly what it might mean.

Here's the hazard graph for total cholesterol: https://www.nature.com/articles/s41598-018-38461-y/figures/2 230 mg/dL is associated with the lowest mortality, and anything below or above that increases mortality.

Is it correct that 230 mg/dl is 5.6mmol/l? We've always been told that we must keep the number below 5 at all costs for health, but this seems to be as bad as being above 6.3. Am I reading this correctly? If so, current medical advice is way off.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#44

Post by twofornone »

>This suggests that a low cholesterol level should be interpreted as a biomarker of illness severity.

What if we are just looking at genetic variations? I feel like medical science in general does not appreciate the magnitude of confounding likely caused by unevenly distributed alleles of all sorts, which would explain a lot of the conflicting papers. And this comes from a reluctance to assign people to genetic groups - because they would almost certainly map at least somewhat to what we call races, and that would have other, potentially uncomfortable implications. It's a deeply seated bias that is so totally normalized as to be virtually undetectable by those who suffer from it, but in some ways it holds back extremely beneficial progress in medical acience.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#45

Post by kuhewa »

twofornone wrote: post #44

>This suggests that a low cholesterol level should be interpreted as a biomarker of illness severity. What if we are just looking at genetic variations? I feel like medical science in general does not appreciate the magnitude of confounding likely caused by unevenly distributed alleles of all sorts, which would explain a lot of the conflicting papers. And this comes from a reluctance to assign people to genetic group…

I reckon accounting for generic variation in observational studies is more common then you are letting on, see mendelian randomisation. Also, in this type of research, race isn't the 'elephant in the room' that it seems you are implying, I mean it is already one of the most common stratifications used and disparities across racial lines in health outcomes are well recognised.

The real barrier to a genetic understanding is with few exceptions, the affect of genes is really complicated.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#46

Post by raverbashing »

sooheon wrote: post #14

This is not a new finding. It's common enough to have a name: the cholesterol paradox ( https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6374572/ ). Also replicated in a more recent n=100k Denmark study: https://www.bmj.com/content/371/bmj.m4266 These are all observational studies, meaning cause and effect is not established. It's relevant that people with some cancers will have very low cholesterol (the cancer apparentl…

I really wonder if that measure (lower levels) is not correlated with veganism, and there might be a cause/consequence confusion there.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#47

Post by amelius »

thaumasiotes wrote: post #42
radford-neal wrote: post #35

Earlier quoted context omitted.

"...there are no strong theories for the diseases caused by low cholesterol..." Well, the body has a whole mechanism for manufacturing cholesterol, so it must be good for something. It seems obvious that there's some level at which low cholesterol must cause death. A U-shaped cholesterol-mortality curve is pretty much guaranteed to exist. The only question is where the minimum is.

> Well, the body has a whole mechanism for manufacturing cholesterol, so it must be good for something. It seems obvious that there's some level at which low cholesterol must cause death. Agreed. > A U-shaped cholesterol-mortality curve is pretty much guaranteed to exist. Yes, this is a necessary consequence of the first observation. Mortality will go to 100% if cholesterol is too low or too high. > The only question…

> Yes, this is a necessary consequence of the first observation.

Not necessarily. The cholesterol could be redundant.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#48

Post by imtringued »

stavros wrote: post #7
onlyrealcuzzo wrote: post #3

Earlier quoted context omitted.

When you have TC (total cholesterol) When you have TC > 200 mg/dL - each 1mmol/L increases mortality. The confidence interval was high. > In the age groups of 18–34, 35–44, 45–54, 55–64, 65–74, and 75–99 years, each 1 mmol/L higher TC increased mortality by 14%, 13%, 8%, 7%, 6%, and 3%, respectively (P < 0.001 for each age group), for TC ≥ 200 mg/dL, while the corresponding TC changes decreased mortality by 13%, 27%,…

So basically the 170 mg/dL TC limit that my lab shows is very suboptimal, and I should go for 230 instead?

Imagine you have container ships named HDL and it is loaded with containers called triglycerides. You don't know how many containers you have. You just know you have this many container ships. Do you have too many containers? Do you have too many ships? That question cannot be answered with just one absolute number.

This is why the HDL to triglycerides ratio exists.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#49

Post by thaumasiotes »

amelius wrote: post #47
thaumasiotes wrote: post #42

Earlier quoted context omitted.

> Well, the body has a whole mechanism for manufacturing cholesterol, so it must be good for something. It seems obvious that there's some level at which low cholesterol must cause death. Agreed. > A U-shaped cholesterol-mortality curve is pretty much guaranteed to exist. Yes, this is a necessary consequence of the first observation. Mortality will go to 100% if cholesterol is too low or too high. > The only question…

> Yes, this is a necessary consequence of the first observation. Not necessarily. The cholesterol could be redundant.

No, that's not the case. If your natural weight is 150 pounds, but you've gotten to the point where you contain 700 pounds of cholesterol, you'll die.

Re: Cholesterol and mortality: Evidence of a U-shape Relationship

#50

Post by imtringued »

tmaly wrote: post #5

Am I the only one that seems to think there is some correlation is not causation here with the numbers?

Consider the following scenario. "Bad chemical compound" (which is a close variant of "good chemical compound") exists in a concentration of 1% when your total level of "good chemical compound" exceeds Y.

Therefore a high level of "good chemical compound" correlates with the harmful effects of "bad chemical compound". Does "good chemical compound" cause the harmful effects? No. Does it correlate? Yes. Now "good chemical compound" is the evil compound that must be lowered.

Now imagine that there is a product that has a high degree of "bad chemical compound". So suddenly there are lots of people with low levels of "good chemical compound" but they suffer from the very same harmful effects that higher levels would predict! "Bad chemical compound" has been found in the product itself and therefore the product was banned.

However, the fact that "good chemical compound" isn't the problem hasn't been recognized because detecting "bad chemical compound" in products is very easy due to the extremely high concentration but difficult (read expensive) to detect in the blood due to the much lower concentration. So the only "reliable" indicators that exist must be based around correlations because measuring the "bad chemical compound" isn't done.

That was a long winded way of saying oxidization of cholesterol causes inflammation and the formation of plaque in blood vessels. Detecting oxidized LDL is possible but almost no lab does it. Instead you get spurious correlations, the most reliable of which is HDL to triglycerides.

Post reply on HN