Comment by virtuallynathan | Hacker News Reader

Comment by virtuallynathan | Hacker News Reader

Parent Full thread

virtuallynathan

Cholesterol / lipid science is a mess. First it was LDL-C that was bad, then that doesn’t actually correlate, then it’s LDL-P, oh wait, then it’s LDL size, and then it’s ApoB that’s going to “clog your arteries”. Once you dive into the “hard” science like pathology of atherosclerosis, you’ll find that the lipid hypothesis makes no sense.

We have drugs (PKCS9 inhibitors: https://www.nejm.org/doi/full/10.1056/NEJMoa1615664, bile acid sequesterants, CETP Inhibitors: https://www.nejm.org/doi/full/10.1056/NEJMoa1609581) that can lower LDL to 10-20mg/dL — the later even increased HDL. No impact on cardiovascular mortality. The studies all use composite endpoints.

What might cause Atherosclerosis?

- Bacteria

- Bacterial lipid byproducts

- Endothelial damage from toxins (lead, mecury, etc)

- Endothelial glycation from high blood sugar (Diabetics have a lot of CVD)

- Bad genetics resulting in excess hemodynamic stress at coronary branch points (fun fact: CVD only develops in a few specific spots)

- Autoimmune conditions (higher rates of CVD in this population)

What most of these things have in common is they trigger the activation of the immune system.

Just so other readers are aware, this view of cholesterol and lipidology is far outside the mainstream medical understanding of the topic.

LDL-c is bad. LDL-p (the particle number), LDL size, and ApoB (which is the protein part of LDL) are all very well correlated with LDL-c. So, while some people prefer one measurement or another, they're all proxies for one another.

We have lifelong exposure to our genetics-and-diet-induced cholesterol levels, whereas clinical trials run for a few years. The long-term follow up of the statin trials show mortality reduction. PCSK9 inhibitors reduce all-cause mortality, and the long-term follow-up of the PCSK9 trials may or may not show the same for cardiovascular specific mortality. However, everyone in the field thinks they will, because those with loss-of-function PCSK9 variants live longer with less cardiovascular disease. Also, adjudicating the cause of death is hard - autopsy, if done, overturns the diagnosis nearly 60% of the time[1].

Cholesterol isn't the only risk factor for cardiovascular disease, so other things (like autoimmune conditions) will contribute to risk as well, but this isn't an either/or phenomenon.

1 = https://www.ncbi.nlm.nih.gov/pubmed/12913844

> "PCSK9 inhibitors reduce all-cause mortality, and the long-term follow-up of the PCSK9 trials may or may not show the same for cardiovascular specific mortality."

Check the evalocumab trial I posted, ACM was actually slightly worse in the PCSK9 side of the trial.

> "LDL-c is bad. LDL-p (the particle number), LDL size, and ApoB (which is the protein part of LDL) are all very well correlated with LDL-c. So, while some people prefer one measurement or another, they're all proxies for one another."

Yes, this tends to be true, however discordance between them is possible - and generally not a good thing.

> "Cholesterol isn't the only risk factor for cardiovascular disease"

Sure, but I think the whole of the field has rabbit-holed on lipids for too long, ignoring actual causal factors.

> "Sure, but I think the whole of the field has rabbit-holed on lipids for too long, ignoring actual causal factors."

LDL cholesterol is by far the best established causal risk factor for coronary artery disease.

> "Check the evalocumab trial I posted, ACM was actually slightly worse in the PCSK9 side of the trial."

All cause mortality was significantly lower in Odyssey: https://www.nejm.org/doi/full/10.1056/NEJMoa1801174

> LDL cholesterol is by far the best established causal risk factor for coronary artery disease.

Is it causal, or is it a risk factor? I'd say at best it is "necessary but not sufficient". CVD risk calculators don't even take LDL-C into account, that's how good of a marker it is.

Mainstream medical understanding can move as we understand more about the nuances of the target populations. Here are some issues that I’m confused about:

- studies that incorporate statins don’t seem to segregate impact of anti-inflammatory effect of some statins vs cholesterol lowering.

- some docs are prescribing statins to healthy individuals based on only cholesterol scoring, also not taking into account more expensive but better imaging like coronary calcium scoring

- Are there studies that show proactive use of statins on otherwise healthy individuals lead to decreased overall mortality?

- there’s less question in my mind that a subpopulation with cardiovascular disease could benefit from statins but mainstream medical care seems far more aggressive that that, especially when dealing with asymptomatic elderly.

> "- studies that incorporate statins don’t seem to segregate impact of anti-inflammatory effect of some statins vs cholesterol lowering."

How would you tease that apart? CAD requires some amount of inflammation, since macrophages have to enter the blood vessel wall to become foam cells. So, it may be mechanistic that reducing LDL also reduces some markers of inflammation. If you have a method of testing statins' anti-inflammatory effects in some way that is independent of their LDL lowering effects, by all means someone would do that study.

> "- some docs are prescribing statins to healthy individuals based on only cholesterol scoring, also not taking into account more expensive but better imaging like coronary calcium scoring"

Coronary calcium scoring can be a good tiebreaker when someone's risk factors suggest that they should be on a statin, but they don't want to take a statin.

> "- Are there studies that show proactive use of statins on otherwise healthy individuals lead to decreased overall mortality?"

There are many primary prevention statin trials. They all involve people with at least one risk factor (high blood pressure, etc). I consider them trials in "healthy" people, since they don't yet have coronary artery disease, but if you meant people with zero risk factors, these wouldn't answer your question. Primary prevention trial examples: ASCOT https://www.ncbi.nlm.nih.gov/pubmed/14997313 and Jupiter https://www.ncbi.nlm.nih.gov/pubmed/18997196

> How would you tease that apart? CAD requires some amount of inflammation ...

This was my point. On one hand, statin supporters use these studies to say "cholesterol lowering" is warranted, but there's no quantification on how much protective effect is from its anti-inflammatory properties vs the cholesterol lowering aspects. The anti-inflammatory effect is also not tied to reducing LDL alone, but has separate mechanisms (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3394171/). Rather than try to isolate the cholesterol lowering aspect, I'd be interested in studies that compared a range of anti-inflammatory drugs (other than just low-dose aspirin) to a variety of statins that also differ in the degree of their anti-inflammatory effects. But as far as I know (and I'm definitely no expert in this field), such studies have not been done.

> Coronary calcium scoring can be a good tiebreaker when someone's risk factors suggest that they should be on a statin, but they don't want to take a statin.

I know a healthy guy who has sky high LDL-P, didn't want to take a statin, and then got CAC imaging with a score of zero. If he hadn't talked his cardiologist into that scan, he'd be on a statin with all the side-effects it entails. That to me seems at odds with the "do no harm" mantra, especially when the CAC imaging seems more directly indicative of underlying CVD in asymptomatic individuals. There are many examples of people with low cholesterol (not on statins) who have underlying CVD as well as people with high cholesterol who have no CVD.

> There are many primary prevention statin trials. They all involve people with at least one risk factor (high blood pressure, etc). I consider them trials in "healthy" people, since they don't yet have coronary artery disease

Similar to your critique of the OP research paper, I am troubled with studies that treat all asymptomatic individuals the same and don't assess state of atherosclerosis (i.e., a CAC score) and differentiate types of risks (high blood pressure vs high cholesterol). The question is whether high cholesterol itself _in the absence of underlying pathology_ leads to worse outcome. If we have no idea whether the starting population has underlying pathology to begin with, IMHO there's a problem with extrapolating those results to asymptomatic people with those risk factors. Other risk factors like HBP may actually be causal.

What you're describing sounds like exactly what the CAC score is good for. Most people don't mind taking a statin if guidelines indicate that they should. But if they really want to avoid a statin, then a CAC of zero is reassuring.

I'm suggesting that the reliance on a very fallible test (LDL Cholesterol) for asymptomatic individuals is bad. People could be started on statin treatment and yet have no underlying pathology. I understand CAC imaging is expensive, but so is decades of statin treatments and potential side effect issues.

My understanding of "the statin issue" is that in the 1970's drug companies found new chemicals to poison the liver's ability to manufacture cholesterol. They reasoned, "as cholseterol is a major part of the arteriosclerosis that hardens blood vessels, people whose bodies are depositing cholesterol where it shouldn't will benefit from poisoning their liver," and promoted the new patent medicines to medical professionals as anti-heart-disease medications.

The medical profession is rather vulnerable the promise of miracle cures to offer their patients, and prescribed the new anti-cholesterol drugs with gusto. Later studies have found the benefit of statin therapy to be minuscule: the "number needed to treat" to prevent one fatal cardiac event is rather high, on the order of 217 for low-risk patients [1]. That is, ~216 patients get to experience the side effects of artificially-reduced cholesterol levels while only 1 person avoids the problem the pill was prescribed to prevent.

In the holistic way of looking at the role of cholesterol, it is more important to think about why the body is manufacturing excess cholesterol and depositing it where it does. As you indicated above, inflammation is an important consideration. Another line of thinking is that there is a roadblock in the steroidogenesis pathway, and that supporting the body's effort to make steroids takes care of the elevated cholesterol levels too.

But our medical professionals tend to develop tunnel vision as they specialize, and assume their specialty's understandings of their domain is basically correct. The drug companies find it much more profitable to peddle anti-cholesterol drugs than to address patients' actual problems.

At least patients who are harmed by statins have the freedom to discontinue their anti-treatment.

[1] http://www.thennt.com/nnt/statins-persons-low-risk-cardiovas...

CAC imaging is expensive and exposes people to (a very small amount of) radiation. Statins are safe and well tolerated. Side effects are treated by temporarily stopping the medication. In people with an appropriately elevated risk, they are cost effective for risk reduction. I.e., they keep people alive longer. So, I don't agree that CAC screening is necessary or helpful prior to initiating therapy in most cases. CAC imaging is a good choice when there is equipoise or disagreement.

> In people with an appropriately elevated risk, they are cost effective for risk reduction.

And this is where this is disagreement. What is appropriately elevated risk? In the case of asymptomatic high cholesterol individuals, especially those without additional risk factors beyond cholesterol, (1) there are flawed studies that include people with underlying pathology together with people who have none to assess statin "benefit", and (2) there are flawed studies that show statins might actually be detrimental with regard to overall mortality. Mainstream medicine currently is fine with #1 (your "they keep people alive longer") and disregard #2 ("they are inversely associated with mortality").

I completely disagree with your dismissal of statin side effects and definite overplaying of the radiation dose of CAC scanning. Nobody has complete understanding of the full side-effect profile of long-term statin treatment let alone all the different types of statins. You have a point with expense, but lets not kid ourselves about the relative side effect profile of long-term statins vs a CAC scan. There's also the management advantages knowing CAC score vs just using LDL.

http://imaging.onlinejacc.org/content/10/3/304 "The presence of CAC is both a sensitive and specific indicator of atherosclerosis, while the total CAC score is a strong predictor of both short- and long-term clinical outcomes. A zero CAC score has become the most definitive predictor of low risk, mild CAC score elevations are indicators for initiating optimal prevention strategies, and high CAC scores may indicate the need for more aggressive management and follow-up. Furthermore, the test has become inexpensive and, with current imaging technologies, is associated with radiation exposure that has decreased to 1.0 to 1.5 mSv per study and to even lower radiation exposure with the newer low-dose imaging protocols (34)."

Average person receives about 3 mSv per year from natural sources. The annual dose for flight attendants is 1.5-1.7 mSv. Mammograms and CAC imaging is similar and less than what flight attendants get in a year.

As I have been doing throughout this thread, I'll continue to agree that a zero CAC score is a really valuable piece of prognostic data, and that the radiation dose is very small. But I'll continue to disagree that we need to scan every adult who qualifies for a statin. And I don't agree that the statin literature is a set of flawed studies. The statin literature is robust - this topic is one of the best-studied topics in medicine. Within medicine, this topic isn't controversial. Even the conservative USPSTF recommends statins for primary prevention above some cutoff: https://www.uspreventiveservicestaskforce.org/Page/Document/...

Re: there's robust literature on statins and it's settled. We're not talking about statin studies on CVD patients. We're talking about asymptomatic treatment. I'm not aware of any study that partitioned asymptomatic high-cholesterol individuals on underlying pathology, e.g., by CAC score, and then seeing if statins were of any benefit to the zero score group. That's because we wouldn't treat those individuals. Instead, all studies (used to argue for treatment of asymptomatic high-LDL) I'm aware of lump "high-risk" individuals together, give them all statins, and then say the outcome is improved for the whole group.

So on one hand, you argue the OP article is fatally flawed because the low-cholesterol group contains high-risk people given statins. On the other hand, you don't concede that the studies lumping all asymptomatic "high risk" individuals together suffer from the same issue. The EISNER study started with "an older population with baseline risk factors" and found 52% had CAC abnormality. I would be interested to see what percentage of asymptomatic high cholesterol people have CAC zero scores and would be treated with statins anyway under current guidelines.

There is a categorical difference between the OP, in which failure to account for statins creates the observation, and a clinical trial that successfully reduces mortality even though the event rate in the control arm wasn’t 100%. That difference is so great that there is almost no relationship between those scenarios.

You don’t know how many people in the low-cholesterol group were on statins in OP article, correct? And despite you completely downplaying it with “the event rate in the control arm wasn’t 100%”, you don’t know how many people in the high-cholesterol group in asymptomatic studies had zero CAC scores (e.g., 50% from EISNER). In both cases you are drawing a conclusion based on incomplete partitioning.

There’s also no consideration of non-CVD mortality and especially morbidity in most of these studies. Despite your unequivocal statement that (all) statins are safe, there’s still a lot we don’t know about long-term use including effects on cognition (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5830056/).

> LDL-c is bad.

Is it really? For example:

Ravnskov, Uffe, Michel de Lorgeril, David M. Diamond, Rokuro Hama, Tomohito Hamazaki, Björn Hammarskjöld, Niamh Hynes, et al. 2018. “LDL-C Does Not Cause Cardiovascular Disease: A Comprehensive Review of the Current Literature.” Expert Review of Clinical Pharmacology 11 (10): 959–70. https://doi.org/10.1080/17512433.2018.1519391.

You can find fringe material in any topic.

This review summarized the evidence in a way that most of the rest of the field would not. Fringe views are valuable, but unless they can actually refute the underlying hypothesis or provide a better model, the work is not very impactful. Particularly so with a review article.

Do the statin follow-up studies correct for participants seeing a doctor more often? Even if the statins do nothing I would expect mortality to be lower among the group that takes them just by having more interactions with doctors.

You're correct, which is why these trials are placebo controlled and double blinded.

Hi there, by "this view", I take it you mean the article, not what virtuallynathan was saying. Is that correct?

If LDL-c correlated with LDL-p, those two metrics would have been similarly good as CVD predictors, but they aren’t, LDL-c is basically useless. (Don’t have sources atm, can link later if you’re interested)

This is simply incorrect.

https://www.tandfonline.com/doi/full/10.1080/17512433.2018.1...

Well, the study in the posted article examined Total Cholesterol (TC), not LDL, and the conflicting results you say were found in different studies were all about LDL, which suggests there really is some involvement of LDL in CVD. I wonder whether you might have misinterpreted these studies as conflicting when they are merely complementary in nature.

Additionally, I think you will not find many studies that claim that high HDL is bad for you, whereas you will find many that claim it's good (because it protects from CVD). I can't say I see much of "a mess" here- just ordinary, progressing scientific understanding of a complex system.

Are there really any studies that attribute atherosclerosis to bacteria, or some of the other conditions you describe? If so- are there many of those and are they good quality studies? I am very skeptical of your comment's suggestion that the immune system may be involved in atherosclerosis.

My comment is more general, not specifically in response to the specifics of the article on TC. FWIW, TC is probably not a great measure on an individual basis, but sort of interesting population-wide.

Yes, folks agree high HDL is good, high LDL is bad. However, drug trials to raise HDL didn't work so well.

> "Are there really any studies that attribute atherosclerosis to bacteria, or some of the other conditions you describe?"

Yes! Many!

- "Accelerated Atherosclerosis in Autoimmune Rheumatic Diseases" https://www.ahajournals.org/doi/full/10.1161/CIRCULATIONAHA....

- "Deposition and hydrolysis of serine dipeptide lipids of Bacteroidetes bacteria in human arteries: relationship to atherosclerosis"

- "The immune system in atherosclerosis" https://www.nature.com/articles/ni.2001

- "Is atherosclerosis an autoimmune disease?" https://bmcmedicine.biomedcentral.com/articles/10.1186/1741-...

- "Atherosclerosis and the role of immune cells" https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4391004/

> TC is probably not a great measure on an individual basis, but sort of interesting population-wide.

I keep getting the impression that using TC as a guide the ratio of number needed to treat vs morbidity reduced is kinda dismal.

Thank you for the links. My understanding is also that TC is not a very good predictor of cardiovascular health, but I'm not any kind of expert on the matter, or even close.

I thought I'd double down on reading medical books and the more I read the less I knew. The deeper the fuzzier.

This is a relevant video to some of the stuff you said:

https://www.youtube.com/watch?v=vBKa6d6j9_8

For the life of me I really don’t understand why the regulators haven’t clamped down on statins. Sure, it lowers your cholesterol, but unless if you’re in a very specific population it doesn’t move your all cause mortality risk one iota. All while costing the patient money and causing some pretty nasty side effects too.

Probably because most data disagree with this comment. There's a ~20% relative risk reduction in vascular events & mortality per ~40 mg/dL decrease in LDL.

The CTT is probably the best summary of these data: https://www.cttcollaboration.org/

> "~20% relative risk reduction in vascular events & mortality per ~40 mg/dL decrease in LDL."

Is that a composite endpoint of vascular events AND mortality?

From TheNNT:

"It takes 5 years of daily statin therapy to achieve a 1.6% chance of avoiding a heart attack, and a 0.37% chance of avoiding a stroke. Most disappointing, statins seem unable to prevent death in this group. And most concerning, the drugs may increase diabetes, a serious and life-altering disease." (Primary Prevention)

http://www.thennt.com/nnt/statins-for-heart-disease-preventi...

That would be the same CTT that makes money from the makers of Statins and has been accused of being insufficiently transparent around their studies[0].

0 http://www.drjohnm.org/2014/06/growing-doubt-on-statin-drugs...

> but unless if you’re in a very specific population it doesn’t move your all cause mortality risk one iota

Have you got a link to some relevant research?

Sure do! I typically defer to Mark Sisson for this kind of stuff, who’s very good at linking in relevant research.

https://www.marksdailyapple.com/the-evidence-continues-to-mo...

Reply on news.ycombinator.com