Attia email: LDL-C, blood pressure, and CVD risk

[Kenny4/4]

all things equal a lower LDL equates to a lower CHD-CVD risk.

The evidence suggests otherwise.

This 8-year study of 100,000+ people aged 50+ without preexisting diabetes or heart disease is the only high-powered study I have seen that focuses on the lipids of healthy older people. I am skeptical of the studies that support the high LDL -> disease link primarily because they include so many sick people. By excluding young people and those with preexisting diabetes, heart disease, or statin prescriptions, the study shows us a cross-section of people who navigated the contemporary nutritional minefield reasonably well to 1997. The subsequent experiences of this relatively insulin-sensitive population therefore seem more relevant to this question - how does LDL drive mortality when insulin is low?

It turns out that low LDL-C conforming to conventional recommendations correlates with greatly INCREASED mortality. LDL-C high enough to prompt many doctors to prescribe statins correlates with greatly REDUCED mortality. The effect is not small. As per Table III, here are the hazard ratios for LDL-C levels in mg/dL:

50-60 years
<97: 1
97-116: 0.57
116-154: 0.44
>154: 0.44

60-70 years
<97: 1
97-116: 0.67
116-154: 0.49
>154: 0.45

>70 years
<97: 1
97-116: 0.71
116-154: 0.60
>154: 0.52

So if you're between the ages of 60 and 70, don't have diabetes and are not on a statin, having an LDL-C in excess of 154 mg/dL puts you in a group whose members are only 45% as likely to die as those with supposedly ideal LDL-C.

So, "all things equal", higher LDL-C = longer life in relatively healthy populations. If you want to parse it further, the two studies I cited in my previous post above show that LDL-C has no relationship with CVD event risk as long as one is insulin sensitive (low trigs/HDL ratio). The disappearance of the correlation in metabolically healthy people tells me that LDL-C is a marker of CVD risk in metabolically dysfunctional people, not a cause. And the huge reduction in all cause mortality as LDL-C rises suggests that plentiful LDL may offer benefits for overall health even though it has no effect on CVD risk.

They left out Statin users thereby confounding the population pool they sampled from. This results in a sampling bias as the high cholesterol portion that has high cardiovascular risk would be left out. This makes the high cholesterol population that is drawn from skewed in favor of robust individuals. Lower cholesterol high risk individuals would not be being treated with statins.

If you look at the historical data the pre statin era cholesterol studies show a u shaped CHD curve at a much lower nadir 160-200. I believe the all cause curve is lower as well. I will look them up and post them they are large sample size studies.

I agree that cholesterol is not the be all do all and that other lipid ratios have more meaningful significance.

Again the drawing population is sample biased in the statin era.

[Kenny4/4]

I believe the studies were MRFIT, The Gas Company Employee one and others if my ApoE4 memory is working tonight. I will look it up in the AM. Men were more profoundly affected. Women did not see as much of an effect.

Again, I agree that total Cholesterol is not a good choice as ratios are a better way, many other variables play a great(er) role in CHD, and Very low to low cholesterol may be a marker for “other”associated Correlated risk factors like cirrhosis, pre mordialcancer or a direct cause of higher mortality. Generally speaking with similar variables high LDL cholesterol is associated with higher CHD and higher all cause mortality at middle age particularly with men. LDL/HDL are better metrics and other ratios and measurements like ApoB or ApoB/ApoA-I . The Framingham study really illustrates this.

[zc_hl]

We don't want infection or toxins, but if we do have them, we also want high LDL, right?

To be more precise, I was thinking that higher LDL could be suggestive of an infection, so, it's worth ruling out any infection just to make sure you're on the healthy side of higher LDL.

Indeed, infections seem to be involved in atherosclerosis:
Infectious burden and atherosclerosis: A clinical issue (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4097149/)

As you said, LDL binds pathogens and "deactivates" them.

Having both information, I can't understand how we can think that it's a good thing to lower LDL, except if the decrease stems from the clearing of the infection.

[Kenny4/4]

https://www.ahajournals.org/doi/pdf/10. ... .86.3.1046
Here is a paper from a very large pre-statin population 350,000 people. Look at page 5 of 15 it shows the risk ratios for all cause , CHD and other diseases. For men all cause mortality and CHD was positively correlated with cholesterol levels except at levels <160. The below 160 people if I remember correctly would account for 7% of the population and include a lot of hard living people-cirrhosis of the liver, maybe they smoke more eat less, have metabolic problems as it’s kind of hard in our American lifestyle to have cholesterol that low. Again-Cholesterol for men is positively correlated with morbidity in all except the very low cholesterol group.

[MarcR]

They left out Statin users thereby confounding the population pool they sampled from. This results in a sampling bias as the high cholesterol portion that has high cardiovascular risk would be left out.

The entire study screened out people with existing diabetes and heart disease, so the statin users in the study are generally healthy people. The primary driver of statin prescriptions for people without heart disease is high LDL-C, so it's logical to compare the mortality results for those statin users with those of the non-users who have high LDL-C.

As you can see from Table III, statin users in the three age groups enjoyed 66%, 63%, and 63% of the mortality risk of non-statin users with supposedly ideal LDL-C (under 97 mg/dL). This sounds good until you look at the statin refusers, those with high LDL-C who either didn't know or didn't care that they were supposed to be taking a drug to lower their LDL-C. Their risk was only 44%, 45%, and 52%!

Here is a paper from a very large pre-statin population 350,000 people.

Unlike the study I cited, which screened out people with preexisting diabetes and/or heart disease, this study included everyone. And because it relies solely on total cholesterol with no figures for HDL-C, triglycerides, and LDL-C, we can't see how the insulin-sensitive sliver group (trigs/HDL < 2) fared. As I noted above in this thread, Framingham data shows clearly that LDL-C is irrelevant when HDL-C is high and triglycerides are low.

[Kenny4/4]

They left out Statin users thereby confounding the population pool they sampled from. This results in a sampling bias as the high cholesterol portion that has high cardiovascular risk would be left out.

The entire study screened out people with existing diabetes and heart disease, so the statin users in the study are generally healthy people. The primary driver of statin prescriptions for people without heart disease is high LDL-C, so it's logical to compare the mortality results for those statin users with those of the non-users who have high LDL-C.

As you can see from Table III, statin users in the three age groups enjoyed 66%, 63%, and 63% of the mortality risk of non-statin users with supposedly ideal LDL-C (under 97 mg/dL). This sounds good until you look at the statin refusers, those with high LDL-C who either didn't know or didn't care that they were supposed to be taking a drug to lower their LDL-C. Their risk was only 44%, 45%, and 52%!

Here is a paper from a very large pre-statin population 350,000 people.

Unlike the study I cited, which screened out people with preexisting diabetes and/or heart disease, this study included everyone. And because it relies solely on total cholesterol with no figures for HDL-C, triglycerides, and LDL-C, we can't see how the insulin-sensitive sliver group (trigs/HDL < 2) fared. As I noted above in this thread, Framingham data shows clearly that LDL-C is irrelevant when HDL-C is high and triglycerides are low.

Sorry there Marc I assumed you were using the study that Feldman guy always quotes that leaves out Statin users(which makes the study flawed) Getting back to the logic of studies ...if you leave the statin users in you are again confounding as lower levels are high risk just as eliminating high risk from the high cholesterol group confounds the Feldman cited study. Then you take a middle aged to older middle age group and leave out those with heart disease and diabetes-that’s a pretty meaningless study.
I will look at the other one. I agree ratios ApoB-ApoA-1, HDL- Trig are great markers and you can run high LDL-C and have low risk buuuuut generally speaking higher cholesterol with all other factors equal including ratios =slightly higher risk of CHD. Look at the Mendelian randomization studies on cholesterol, they clearly show some pretty direct causation.

Again I’ll look at your other study but my glancing at it shows some vague selection bias. Statin users who went on and stayed on vs people who said no thanks or those who dropped- how do you stratify that sample proportionately for risk? Common sense tells us the worried high risk stays on them and the “other means” exercise, diet, supplements goes off or never goes on and maybe their problem wasn’t that large to begin with. The study smells pretty bad already

[NewRon]

https://www.docsopinion.com/2014/07/17/ ... hdl-ratio/

https://www.cambridge.org/core/journals ... 5ABBCBB598

https://www.docsopinion.com/about/

However, maybe we should get back to blood pressure, there's been a bit of thread drift!

[MarcR]

Great point, NewRon. I split the topic at the point that the link to Attia's email, which was about CVD risk, not Alzheimer's, was introduced.

[MarcR]

Parsing your words ...

leaves out Statin users(which makes the study flawed)

Exclude statin users = "study flawed"

if you leave the statin users in you are again confounding

Include statin users = "confounding"

By this logic, all lipid studies conducted since the inception of statin sales are useless. I don't agree. I think evaluating statin users as an independent group is ideal since they are subject to a broad range of effects that confound any attempt to compare them to unmedicated individuals.

You have contended repeatedly that isolation of statin users ensures that the remaining high LDL-C individuals are subject to selection bias in the direction of lower CVD risk. I continue to disagree. High LDL-C has long been the primary driver of statin prescriptions, so the typical statin user also had high LDL-C prior to medicating. What then distinguishes high LDL-C people who take statins from high LDL-C people who don't? I think it's much more likely that those who accept statin prescriptions engage in behaviors that bias them toward lower risk:

- annual checkup
- comply with doctors' orders
- follow recommended nutrition and exercise guidelines
- higher income (can afford health care and prescription drug costs)

The remaining high LDL-C people, those with less than half the all cause mortality risk of the group with supposedly ideal LDL-C levels, are more likely to be less health-conscious people who don't visit their doctors as often, resist prescriptions, and ignore or are unaware of recommended lifestyle guidelines. That group therefore would be biased toward higher risk!

Then you take a middle aged to older middle age group and leave out those with heart disease and diabetes-that’s a pretty meaningless study.

It's only meaningless for those of us with heart disease and/or diabetes! As someone who suffers from neither, I'm absolutely fascinated!

P.S. Regarding those statin effects, from Statin Adverse Effects: A Review of the Literature and Evidence for a Mitochondrial Mechanism:

Abstract

HMG-CoA reductase inhibitors (statins) are a widely used class of drug, and like all medications have potential for adverse effects (AEs). Here we review the statin AE literature, first focusing on muscle AEs as the most reported problem both in the literature and by patients. Evidence regarding the statin muscle AE mechanism, dose effect, drug interactions, and genetic predisposition is examined. We hypothesize, and provide evidence, that the demonstrated mitochondrial mechanisms for muscle AEs have implications to other nonmuscle AEs in patients treated with statins. In meta-analyses of randomized controlled trials (RCTs), muscle AEs are more frequent with statins than with placebo. A number of manifestations of muscle AEs have been reported, with rhabdomyolysis the most feared. AEs are dose dependent, and risk is amplified by drug interactions that functionally increase statin potency, often through inhibition of the cytochrome P450 (CYP)3A4 system. An array of additional risk factors for statin AEs are those that amplify (or reflect) mitochondrial or metabolic vulnerability, such as metabolic syndrome factors, thyroid disease, and genetic mutations linked to mitochondrial dysfunction. Converging evidence supports a mitochondrial foundation for muscle AEs associated with statins, and both theoretical and empirical considerations suggest that mitochondrial dysfunction may also underlie many non-muscle statin AEs. Evidence from RCTs and studies of other designs indicates existence of additional statin-associated AEs, such as cognitive loss, neuropathy, pancreatic and hepatic dysfunction, and sexual dysfunction. Physician awareness of statin AEs is reportedly low even for the AEs most widely reported by patients. Awareness and vigilance for AEs should be maintained to enable informed treatment decisions, treatment modification if appropriate, improved quality of patient care, and reduced patient morbidity.