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 apparently using it as fuel source), which reverts to normal in remission. Other medical conditions can lower serum cholesterol, from IBD to chronic liver disease (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3271321/).
If low cholesterol is inherently mortality inducing, lowering it directly in otherwise healthy people should cause more deaths. This is a testable hypothesis. The Denmark study found that people on statins (cholesterol lowering) did not die more, even when they were on the lower side of the U curve.
There have also been countless RCTs with statins (or other cholesterol reduction drugs) being the only controlled intervention, finding that low TC improves total mortality.
A long term study tracking change in total cholesterol found that people who always had low TC (genetically) did not have a higher risk of death, but those whose TC suddenly dropped from normal to low had a higher risk of cancer: "We observed a significant risk of cancer, noncardiovascular noncancer, and all-cause mortality in men whose TC levels changed to low (<180 mg/dL) from the middle (180 to 239 mg/dL) level at baseline." (https://www.ahajournals.org/doi/10.1161/01.cir.92.9.2396)
These points together point to the low cholesterol -> high mortality link as being a biomarker of underlying disease, not a direct causal link.
(Info taken from this video: https://www.youtube.com/watch?v=CxX51n2Z0vc)
---
edit: skimmed the paper, authors mention everything in this post but have some counterpoints. They point to other longitudinal studies where consistently low cholesterol was a risk factor for death (a point against the sudden lowering due to cancer hypothesis). There are many studies like this, but the problem is while we have shown the link b/w cardiovascular death and high cholesterol through RCTs, there are no strong theories for the diseases caused by low cholesterol, while we know cancer and chronic liver disease cause low cholesterol. It could just be a marker of general frailty in old age.
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.
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.
It's good for all sorts of things, for example building cells - cell membranes are made of cholesterol, which is why our body produces a shitload of it.
Alternative hypothesis: high cholesterol is correlated (a small part of TC is from eating) with some adverse medical condition that statins are effective against. From this cohort, statin users measure lower cholesterol and die less. That lower cholesterol is always desirable and we should stop eating eggs would make absolutely no sense.
However, logically my post does not entirely rely on knowing the precise in vivo mechanism of statins and their causal health effects. We only need to see that in RCTs, lowered cholesterol doesn't cause increased mortality as these observational studies would suggest. That is a pretty direct test of the null hypothesis.
edit re your alternative hypothesis: what is this underlying adverse medical condition, and how is it resolved by statins? Every source I can find agrees statins work straightforwardly by inhibiting biosynthesis of cholesterol in the liver.
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%, 34%, 31%, 20%, and 13%, respectively, in the range < 200 mg/dL (P < 0.001 for each age group). TC had U-curve associations with mortality in each age-sex group. TC levels associated with lowest mortality were 210–249 mg/dL, except for men aged 18–34 years (180–219 mg/dL) and women aged 18–34 years (160–199 mg/dL) and 35–44 years (180–219 mg/dL). The inverse associations for TC < 200 mg/dL were stronger than the positive associations in the upper range.
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.
Implies a causal relationship. In reality, it's much more likely that causality points the other way, and an underlying disease like cancer is causing low cholesterol: https://www.ahajournals.org/doi/10.1161/01.cir.92.9.2396
Meaning: lower cholesterol is not always correlated with lower mortality.
Cholesterol is an essential building/repair block that our body makes, with very little absorbed from diet.[1] Since cholesterol responsible for repairing damaged tissue in the body[2], how do we know know that inflamed, unhealthy individuals don't generate high cholesterol amounts for repair and not from an unhealthy diet? .... Time will tell.
[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3257636/#:~:tex....
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.
Manganese also affects cholesterol. https://pubmed.ncbi.nlm.nih.gov/3705654/ "a favourable influence of manganese was shown which effected a decrease of the cholesterol content of serum, liver and aorta and inhibited the entry of lipids into the aorta. The influence of the manganese on various enzymes as well as a manganese-copper interaction are discussed as possible causes." Manganese is one of those minerals in constant decline from birth when Reactive Oxygen Species starts, before then we all get about 9months to build up levels in the body, provided the mother is healthy and has an healthy diet.
And there isnt much discussion on the content of HDL. https://en.wikipedia.org/wiki/High-density_lipoprotein#Overv...
And on the point of macrophage accumulation, no mention of how vit K1 keeps calcium out of tissue which prevents elastic fibres from hardening and thus causing their own problems.
Thing is whatever science looks at can be portrayed in a positive or negative light.
So it appears that you can drive LDL levels quite low without harm.
We've seen this story play out time and time again, with everything from vitamin D to HDL/LDL. Low levels of vitamin D are related to everything from depression to heart disease to diabetes, but a recent study gave people over 300% of their daily value of vitamin D every single day for 10 years and found zero improvement on a giant list of health metrics. The same with LDL/HDL: bad LDL/HDL numbers give you dramatically higher risk for heart disease and stroke, so people designed a drug that improves those numbers, but of course it does nothing for your actual health.
We need to fix the actual, underlying causes of disease - lifestyle and diet - not hack the various biomarkers higher or lower until the numbers look good. I'm actually working on a startup to help with this, so if anyone's curious (about using it or just to discuss) my email's in my profile.
Also about diet remember that just in the recent history we are shifted, without really scientific proofs behind, from saying "eggs are good, eat them at least one a day" to "egg are bad, limit them"; "blue fish are the protein of the see, eat them much" to "blue fish are pollutants collectors of the see, avoid them| etc, mostly following some lobbyist PR campaign dressed as science so just talking about diet it's complex.
Again, the old school when medicine was not that much a business say: all animal foods help lower the LDL, so eggs, fish, cheese, meat help. Starchy foods (from potatoes to beans, rice, ...) do the opposite. Taking a global look from countries with more animal protein based diets and others with starchy foods based diets confirm that well, but such simple observations is considered flawed because there are too many variables in the middle, while no one really criticize modern studies that actually tend to fail proving anything, just claiming...
Goodhart's law is an adage often stated as "When a measure becomes a target, it ceases to be a good measure". (Wikipedia)
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.
The real barrier to a genetic understanding is with few exceptions, the affect of genes is really complicated.
Instead of basic linear / bell/U curve type ways of describing impact of biomarkers, are new ways of describing complex biomarker relations emerging into the vernacular? e.g. optimums within clusters/sets of biomarkers
