Hacker News Reader

Comment by carbocation | Hacker News Reader

Somehow LDL-cholesterol has come up in this thread about inflammation.

The article here offers a highly speculative opinion regarding the role of inflammation across many diseases. The luminaries in cardiology cited in this article ran trials which many of us consider to show that, rather than inflammation being important, any reason to start statins is a good reason.

The genetic data currently supports very little role for inflammation in important diseases like coronary artery disease, whereas there is crystalline evidence supporting the connection between LDL-cholesterol levels and mortality. Interestingly, when genetics are invoked and mere epidemiology is reassessed, there is no clear atheroprotective role of HDL cholesterol.

The treatment of statins is very much like the treatment of vaccines: dismissed in a pseudo-intellectual manner by people who know a lot (just not about the subject at hand).

View on HN

>there is crystalline evidence supporting the connection between LDL-cholesterol levels and mortality

The connection is very far from crystalline ([1],[2]) The strong association only seems to appear when considering groups with super high LDL-C levels, likely due to FHC. In any case, the typical approach, statins, has very little positive affect on individuals without pre-existing heart disease ([3],[4])

If vaccines had the same side effects and poor absolute risk reduction that statins have, the anti-vaxers might actually have a point.

[1] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3750440/

[2] http://www.crlcorp.com/wp-content/uploads/2013/06/JIMlipidsm...

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

[4]https://www.sciencebasedmedicine.org/statins-the-cochrane-re...

In controlled, randomized trials and in genetically-driven studies (the two most powerful tools that we have to assess causality), almost all factors that reduce LDL-cholesterol lead to reduced ischemic heart disease and death.

Your commentary about statins being primary useful for secondary prevention (in those who have already had a heart attack) was true until a couple decades ago, but there is now evidence demonstrating that statins reduce primary events (though at a high NNT depending on the population's risk). Naturally, the higher risk the population, the more benefit and lower NNT you will see from any intervention.

A classic genetic study, for example, is that of Hobbs and Cohen looking at PCSK9 variants [1]. People with loss of function mutations in PCSK9 have lower LDL-cholesterol. These people have an even lower risk of ischemic heart disease than you would expect based on the degree of LDL-cholesterol lowering. This is surmised to be due to the lifelong exposure to low LDL-C, though the actual reason is not currently known with certainty.

1 = http://www.nejm.org/doi/full/10.1056/NEJMoa054013

The real question is, though, is drug therapy something that should be pushed on otherwise healthy (non-FHC/high risk) individuals in order to reduce mortality. The PCSK9 research paper you linked has the percentage dead after 15 years (starting at an average of 55 yr/o) at 10.7% for those without the mutation and 8.3% for those with it.

Assuming they actually work identically to the mutation, maybe one could argue that the $$$ cost of the new PCSK9 targeting drugs will be worth it to save 2.4%, but we have zero long term data on their safety. There is certainly some evidence [1] that the risk for diabetes could be higher...much like the current statins.

I don't think we have anything even approaching a complete picture on the interplay between CHD, cholesterol, and overall mortality. Until the picture becomes clearer, it doesn't seem prudent to push drug therapy as strongly as it typically is.

[1] http://www.ncbi.nlm.nih.gov/pubmed/20026049

The coincidence of statins falling out of patent and the rush to develop PCSK9 drugs is not lost on me.

To rephrase the exact same data, 15-year mortality is ~30% higher among those with the normal variant compared to those with the cholesterol-lowering variant. That is an enormous mortality effect.

Based on the mechanism by which statins are postulated to cause diabetes (and LDLR-dependent mechanism), PCSK9 inhibitors may also cause diabetes, though it's not yet known. From a mortality perspective, this is irrelevant because we know that statins (and PCSK9 mutations) reduce mortality despite (probably) causing diabetes.

The timing of statins falling off patent is essentially uncorrelated with anything - there was a multi-year gap with atorvastatin off patent before the PCSK9 inhibitors were approved, which makes no sense if you would like to invoke a patent conspiracy. With the advent of genetically-informed early phase research, you will continue to see new drugs come out at an increasing frequency (reversing the trend over the past few decades of fewer approvals).

The link from cholesterol -> CHD -> mortality is as clear as anything in biology (clear enough for the basis for this work to have won a Nobel prize decades ago).

I think this is a bit unkind:

> dismissed in a pseudo-intellectual manner by people who know a lot (just not about the subject at hand).

I'm the one who's potentially sick here, and this (perhaps wrongly) smacks of appeal to authority. What I absolutely don't want to do is disregard medical opinion, but neither am I inclined to abdicate all responsibility for my own health and well being.

> The genetic data currently supports very little role for inflammation in important diseases like coronary artery disease, whereas there is crystalline evidence supporting the connection between LDL-cholesterol levels and mortality

I'd be very interested in learning more about this, do you have anything you could recommend? I'm fully aware that the majority of what I read on the internet is likely to be along the lines of "don't sweat it, eat natural and you'll be fine" and I'd very much appreciate some exposure to the opposite point of view too.

>> The genetic data currently supports very little role for inflammation in important diseases like coronary artery disease, whereas there is crystalline evidence supporting the connection between LDL-cholesterol levels and mortality

> I'd be very interested in learning more about this, do you have anything you could recommend?

I'll let the OP chime in if I'm wrong but I believe s/he is referring to the numerous genetic association studies[1] that have implicated LDL biology in coronary artery disease. Look at ref. 2 for a publicly-accessible review. The article details the genes that have been associated with coronary artery disease, none of which appear important for inflammation.

1. https://en.wikipedia.org/wiki/Genome-wide_association_study

2. http://circ.ahajournals.org/content/128/10/1131.full

You might want to read the book Why Zebras don't get Ulcers by Robert Sapolsky. It's about stress, all aspects of it. So it's not about inflammation, although inflammation means stress and in that regard it is covered - you might find it interesting.

Be prepared for lots of chemical terms, but don't let that hold you off - there's enough interesting stuff to read. It has lots of interesting information.

https://en.wikipedia.org/wiki/Why_Zebras_Don't_Get_Ulcers

In talking about pseudo-intellectual dismissals, I am appealing to the prior news.YC discussions about middlebrow dismissals, which would have been better wording. Sorry about that. How is what I mean different from an appeal to authority? Mostly it has to do with the type of evidence selected by statin "opponents", which tends to be of lower quality.

As for studies that I find instructive, here are a few:

People with loss of function mutations in PCSK9 have lower LDL-cholesterol. These people have an even lower risk of ischemic heart disease than you would expect based on the degree of LDL-cholesterol lowering. [1]

Mendelian randomization demonstrates that genetic scores for LDL cholesterol associate with reduced risk of myocardial infarction (suggesting causality, since cholesterol doesn't reverse-cause a genotype).[2]

IMPROVE-IT showed that a non-statin medication (ezetemibe) added to a statin regimen further reduced the risk of second heart attacks after suffering a first heart attack. This is probably the clearest experimental evidence to date that suggest it's not only statins that reduce cardiovascular risk, but LDL-cholesterol reducing agents more generally that lead to reduced risk of heart attack. [3]

This is a very contemporary summary of genetic and clinical trial data; there is a wealth of older data that I won't get into.

1 = http://www.nejm.org/doi/full/10.1056/NEJMoa054013

2 = http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3419820/

3 = http://www.nejm.org/doi/full/10.1056/NEJMoa1410489

[T]his (perhaps wrongly) smacks of appeal to authority.

Don't dismiss appeals to authority so easily. As it happens, not giving extra weight to experts in proportion to their track record is a provably suboptimal strategy in an adversarial setting (see e.g. [1]). So, in this sense, deferring to trained medical opinion might in fact be the most responsible thing to do.

Note however the bounds in [1] also (indirectly) imply you are better off with a diverse ensemble of expert advice. So if you can afford it, always get a second (or third) reputable medical opinion. And weigh their advice accordingly.

[1] http://www.mit.edu/~9.520/spring08/Classes/online_learning_2...

This is correct. I reference one of the Mendelian randomization studies in my comment [1].

1 = https://news.ycombinator.com/item?id=10739687

It's hard to take the evidence at face value though, given the historically malevolent influence of big pharma on accurate publishing of trial data (SSRIs, vioxx etc), the significant links between the key trialists and the drug developers, and the lack of transparency in the data.

At this point the argument about statins can only be solved by opening the data.

What you are saying is true to the same extent that it is true about any drug. Statins are unexceptional in this regard, except that they have a lot more lines of evidentiary support than most drugs due to the importance of coronary artery disease.

Even ignoring statins, again, the genetic evidence paints a clear picture of the causal role of LDL cholesterol in coronary disease. The jury is largely still out on inflammation.

Agreed. However in the case of statins, the benefits for an individual are very slight, even if the population benefit is large. This means that this sort of uncertainty about side effects can make it rational for an individual to abstain from taking statins.

>The treatment of statins is very much like the treatment of vaccines

I don't think that's an accurate comparison. There's not really anyone in the healthcare industry that would argue against vaccines but there are plenty professionals with qualification (mds, pharmacists, etc) that are against the broad use of statins.

The treatment of statins is very much like the treatment of vaccines: dismissed in a pseudo-intellectual manner by people who know a lot (just not about the subject at hand).

This. It's rather unsettling to see so many well-meaning posters gleefully rejecting prescribed statins. They are playing with their health.

When the "system" becomes an adversary in one's fight for health, as a result of cost, restricted access to doctors and their time, and when an ocean of credible (and non-credible) information is available relating the mechanistically uncertain world of medication, it is not surprising to see many people feel as if they have looked behind the curtain in a wizard of Oz sense and begin reaching for alternatives. Statins are probably among God's greatest gifts to most people with high LDL as clinical trials show, but many in the HN community might be willing and able to make drastic lifestyle changes that other individuals are incapable of. Many entrepreneurial and data-driven types have the thick headedness to be that 1% of patients for whom lifestyle changes work, or at least we like to think so. In any case, starting a statin (in the case below) and then proving to oneself that one is supreme commander of diet and metabolism may be the optimal route. The issue lies in which pathway to choose for our attempt at supreme commandership--ketogenic, paleo, low trans fat, etc. Once again, the results are unclear in some trials and the mechanisms are as elusive as ever, which is why we reflexively search for explanations in the fey realms of inflammation, microbiome, epigenetics, and so forth. Addendum: Most clinical trials are aimed at helping the general population with a specific condition; it would be interesting to try clinical trials of diets in only highly educated and fairly affluent people who know python (usually not part of the subgroup analysis).

I had a long reply to your previous comment written out but you deleted before I could post and it got lost. But I guess what I was saying does boil down to this. I don't have a great deal of faith in the shallow, one size fits all, drug based approach I was offered. I do intend to look at wider options, and you're right that there's a lot of information available. However, I don't intend to do a Steve Jobs. If it looks like I've sustained damage to my arteries already or I can't find an intervention that works then I'll revert to the suggested route.

After I read your other comments I deleted the first post because I realized you are knowledgable about many issues I brought up and it was non-contributory, apologies. I wish you the best in keeping healthy!

Those guys at Harvard...

https://en.m.wikipedia.org/wiki/Paul_Ridker

They must think we still need to learn a lot more about the human body.

"This suggested that elevated cholesterol isn’t the only factor at work in cardiovascular disease, and that in some cases statins, acting as anti-inflammatory agents, could be used to treat the condition."

I noticed that you're a medical resident in Boston. I bet if you look back in 40 years, it'll look like we were in the Stone Age.

Reply on news.ycombinator.com