#1257

The 2024 Chinese paper with millions of people adjusted for HDL and yet still found a U curve.

Also: there are probably different types of HDL and not all of them might be good.

What I’d love to see if a 3d chart with LDL (or ApoB) and HDL (or total) for the x and Y axes and the risk in z. Adjusted for everything else.

1 Like
#1258

To be fair I focused on the TC part of the CONUT score but it’s only part of it and the rest might be as important or even more. I just assumed that people here try to eat a reasonable diet as a first step.

I would love that too as well as PCA models too.
That said genetic factors are probably a very large part of the issue too. At least they are for my N of 1.

1 Like
#1259

Most of the cohort studies suffer from one glaring weakness: they do not report repeated measurement of LDL/ApoB.

ASCVD and the associated mortality is a lifelong chronic condition. But there are a variety of conditions and interventions that arise in the short term (relatively speaking - so months or a few years) and impact LDL.

If a study is effectively cross-sectional (measuring LDL at one point in time) instead of longitudinal (measuring LDL repeatedly over several years) it will be difficult to distinguish pathologies causing low LDL as by product, from low LDL due to long term good health. Or even people doing a short-term diet from those doing it long-term.

Observational data is quite fuzzy for LDL, as there is so much confounding at play, which you do not have for many other blood-markers (e.g. fasting insulin is a lot less fuzzy).

At the very least - if the only studies to make your point are observational - you should come up with a potential mechanism (backed up by published science), by which low LDL is causal for massive and frequent mortality. Because that’s your U-curve.

3 Likes
#1260

Dying people always have low blood sugar, low cholesterol and a low bmi.

#1261

Are you sure of this? Maybe some that suffer and can’t eat regularly but there is other people that are otherwise healthy and even overweight and they pass way in couple days.

#1262

Some conditions that cause low LDL and are causal for increased mortality:

  • malnutrition

  • a range of cancers

  • infections

  • sterile inflammation/autoimmune

  • certain hormonal conditions

  • people taking LDL lowering drugs, because they already had stroke/heart attack or are otherwise a high risk of ASCVD-death; sure they will live longer than ASCVD-victims not taking drugs - but not longer than people without previous stroke/MI

You really have to have a very complete dataset at hand, doing repeated measurements over several years to hope to disentangle that - and infections and sterile inflammation might well not be reported unless it’s extreme cases.

But we DO have RCTs and MR. Therefore: if you just want to dismiss these and focus on the fuzzy observational studies: please provide a plausible mechanism. I.e. how can low LDL on it’s own be so deadly (and what cause of death), that it doesn’t just negate the very positive effects size on ASCVD mortality - but raises mortality on top of that. Where are these bodies killed by low LDL en mass? I’m genuinly interested.

4 Likes
#1263

Some advice to for the kind of studies you want to have a look at to make the point “low LDL on it’s own increases mortality by a large effect size”:

I hope I don’t need to spell it out but: you can’t use these same observational studies to argue that point. They show associations and suffer from reverse causality or unobserved variables (and there are many variables at play with LDL). Also notably most of them measure LDL once, so you don’t acutally know if low LDL was sustained over the follow up period.

You can look at RCTs and try to figure out, if the intervention group got higher mortality in other areas than ASCVD than controls - and which likely are not due to side effects of the drugs themselves. The same goes for Mendelian randomization, though you gotta trust the authors, that the gene-variations themselves are not causing increased mortality in other areas (instead of low LDL).

In addition you can cite in-vitro studies - though in that case it would be studies showing added LDL or cholesterol to be beneficial to kill cancer or germs in the petri-dish. But: an in-vitro study on it’s own is useless (their dud rate in medical science is astronomical) and needs to be discussed jointly with at least animal studies or RCTs/MR.

Remember: it’s established by very high quality evidence (RCTs with longterm follow up, MR, mechanisms extensively understood), that low LDL/ApoB is highly protective against the number 1 killer in the developed world. Therefore you need to find causes of death that are also very common and are promoted by low LDL. The effect size gotta be massive as well: nullifying the proven mortality benefits for ASCVD + increasing mortality on top of that.

3 Likes
#1264

Increased infection and cancer rates would’ve been detected in studies on older adults. None of them showed an increased acm, cancer mortality or risk of infection.

#1265

Thanks for posting this. There are so many podcasts these days, it’s hard to keep up. I’d never heard of the guest, Mike Mutzel, but what he said about lipids mirrors almost exactly what I said in my thread, The “LDL” Kingpin Theory - One Ring To Rule Them All? about a month ago. As he mentioned, the Lean Mas Hyper Responder study is getting these concepts more exposure. As Matt said everyone is different and the widely broadcast and accepted rule that “high LDL-C will cause atherosclerosis” is probably true for most of the general public who are for the most part metabolically unhealthy but Matt agreed that a smaller subset of lean, healthy and athletic (active) people, eating a good diet might be the exception. I would add that you should be regularly testing your metabolic levels - blood pressure, HbA1c, and HsCRP and your plaque levels because high ApoB does certainly raise your risk. Mike Mutzel (along with many others) said that Triglyceride/HDL ratio is also important although Tom Dayspring lately has been downplaying this.

There was also a good section in that Optispan Podcast about epigenetic tests, particularly TruDiagnostics DunedinPace (and the Rejuvenation Olympics) and SymphonyAge, as well as GlycanAge. Matt said that Optispan was experimenting with those although he thinks they are oversold to the consumer. (that starts at 46 min)

3 Likes
#1266

One question I find interesting is, if ldl has a specific and important “delivery” function as described in the podcast, then would people with variant psk9 have likely evolved alternative ways of fulfilling that function.

In which case, surely we should be wary of using Mendelian randomization studies to encourage use of psk9 inhibitors to target “lower the better”.

(ducks back down into metaphorical trench)

1 Like
#1267

…but MR utilize a broader spectrum of gene variants - not just PSCK9 relevant ones…

…and you still would need to make a mechanistic argument, what kind of pathologies (wide spread in the population - so much that you can’t possibly miss it) are promoted by that. And why the RCTs with statins/PCSK9 find better mortality the lower the LDL (down to 50mg/dl and below)

#1268

The argument would apply equally to all genetic variations examined using MR.

I think the burden of proof is on the intervention, not the other way round. I’m merely saying that if LDL has positive functions (which the podcast suggests/describes) then Mendellian Randomization studies can’t be relied on completely to support pharmaceutical interventions because genetic variants which lead to low ldl are likely to have evolved alongside compensatory mechanisms. In order words a PSCK9 variant person may not need to worry about the loss of function from low LDL, whereas someone taking PSCK9 inhibitors might

That’s the trouble, the RCTs often don’t show lower all cause mortality the lower the LDL (down to 50mg/dl and below). They show fewer cardiovascular deaths but not lower ACMortality.

#1269

This may not be measuring what you think it is measuring.

The group of people who don’t take their medicine is different in many respects than the group of people who do take their medicine. The group who takes it is probably more likely to exercise, watch their diet, have stable employment, etc. etc. etc.

This is measuring the differences between groups as well as the impact of the drug. Disentangling those is a challenge.

1 Like
#1270

I’m sorry, but that’s just wrong. It’s okay to admit that you are not well informed about the LDL/APoB literature. But that’s no excuse for making up your own “facts”.

ALL the statins trials for high-dose therapy (that is: large degree of LDL-reduction) that were designed to test for ACM (necessary number and duration), showed lower all cause mortality. Not just cardiovascular mortality.

Exactly because of that the standard of proof for new LDL-medication is so high: it used to be untill the late 90s, that all you gotta do is to lower LDL and events, to get your drug approved as first line treatment. Statins changed that: now you have to test for all cause mortality as well, because high dose statins were so successful in their respective trials. By the way one reason, why Bempedoic Acid will never replace statins.

The most notable study for a statin is JUPITER (and the reason why Rosuvastatin often is the statin of first choice):
https://www.ahajournals.org/doi/full/10.1161/circoutcomes.109.868299

And ODYSSEE is using PSCK9 in addition to high dose statin to lower all-cause-mortality even further (in addition/beyond high dose statin):
https://www.nejm.org/doi/full/10.1056/NEJMoa1801174

And no: the burden of proof is no longer on the side of the scientific consensus, based on RCTs, their longterm follow-up and MR. This side of the debate provided high quality data to argue that lower LDL (from 80mg/dl to below 50mg/dl) is on average better than “normal” or “high” LDL - including better for ACM.

You are the one claiming based on fuzzy observational trials (I laid out their shortcomings in my previous posts - the single measurement of LDL alone makes them already questionable), that low LDL is actually bad. Okay. Then where is the evidence, given that RCTs, their longterm follow-up and MR say the opposite? These observational trials are hypothesis forming - they are not conclusive evidence! In particular for LDL, with so many factors influencing it at play.

At the very least: provide a common cause of death (let’s say: 5%-10% share of ACM), that is promoted by low LDL. If you can’t even do that: why are we still having this debate?

7 Likes
#1271

I’m sure there is a study with a small sample size somewhere where the statin arm had a higher incidence of being shot than the placebo arm.

1 Like
#1272

Sadly Jupiter isn’t clear evidence for all cause mortality benefits. And certainly not for “down to 50mg/dl and below”. I’m not sure why you think it is - the treatment arm was above 50mg/dl. And it’s not without very clear shortcomings:

Findings: The all-cause mortality in JUPITER was more than twice that of the average of primary prevention studies, matching well only with specific trials designed in diabetics (ASPEN or CARDS), early hypertension studies (ALLHAT-LLT) or a trial in patients with acute coronary syndromes (PROVE IT). Since the ‘play of chance’ is unlikely to explain these discrepancies due to excellent baseline match, excess death rates and all-cause mortality rates in both JUPITER arms must be questioned

Which facts are you referring to?

I think you’ve mxed me up with someone else. I’m not making any claims based on observational trials.

I think the burden of proof should always be on the intervention (“first do no harm”) but you seem to be getting things confused. I am very clearly talking about the lack of evidence for going below 50mg/dl LDL.

Given the lack of “all cause mortality” evidence from RCTs to go below 50 mg/dl - and given the fact that we know LDL serves a transport function in the body. I’m not sure why we would aim for below 50mg/dl.
Whether or not we’ve identified a “common cause of death” from extremely low LDL seems irrelevant at this point.

4 Likes
#1273

We all have low levels of LDL cholesterol as children up until the age of 12. This is when we, as humans, grow the most developmentally.

Considering how CVD is the number one killer, and I have never heard of a child having a heart attack, I think I’ll follow their cue and keep my LDL and ApoB at around 50.

3 Likes
#1274

Very interesting, wasn’t aware.

#1275

About 50 is my goal too, but there seem to be a lot of people arguing to go lower. And I’m not sure what evidence is driving it.

The childhood data is really interesting - I wonder whether the hormonal role of cholesterol means it needs to be higher in teens/adults. Pure speculation, but just a thought.

1 Like
#1276

Historically - for those who are interested - here’s the original controversy with JUPITER (a lot of back and forth since, of course):

https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/416101

Quote:

" Conclusion The results of the trial do not support the use of statin treatment for primary prevention of cardiovascular diseases and raise troubling questions concerning the role of commercial sponsors."