Don't fall in love with the biology
There’s a cliche that every first year medical student learns about the limitations of our current knowledge:
“Half of what you learn in med school is going to be wrong in 10 years, the problem is that we don’t know which half.”
This speaks to the way that the scientific method works: you design experiments that give you results that either increase or decrease the probability that your hypothesis is right.
There’s never an end to it - everything that we think we know, we only suspect to be true with varying levels of confidence.
If there’s a theme that I try to consistently highlight in this newsletter, it’s about the limits of what we know and the importance of remaining humble about the possibility that we might be wrong.
Uncertainty exists in many areas of medicine, even ones that people feel very confident about
I gave a lecture a couple of weeks ago about fish oil with regards to the prevention of cardiovascular disease.
It’s an area where I don’t think anybody really knows very much - while some data suggests that high dose EPA (one of the omega 3 fatty acids in fish oil) reduces the risk of heart disease, this isn’t a result I would report with much confidence.
There isn’t a ton of clarity about the magnitude (or presence) of effect, dose necessary to achieve it, or whether it really does anything on top of standard therapies for heart disease.
I had a slide at the end of the lecture with this graphic from an article in the Journal of the American College of Cardiology about the concept of “residual risk” for heart disease.
The concept of residual risk is often described as one of the keys to personalized medicine - in fact, the graphic below comes from an illustration that’s titled “Residual Risk and a Movement Toward Personalized Medicine:”
This is a review of strategies of cardiovascular prevention written by Paul Ridker, who is one of the biggest names in preventive cardiology.
The reason I bring up this graphic when I’m talking about having humility is how much of the data cited here is hypothetical rather than real.
Ridker cites randomized trial evidence for each of these areas, but there’s a big problem: some of those trials weren’t completed when this paper was published, and some actually ended up being negative trials that didn’t show benefit of the treatments, even though he cites them as evidence supporting this idea.
The flaws with this graphic reveal confidence that trials would show something that they didn’t show
The concept behind residual risk goes like this: once you’ve treated someone who has heart disease with aspirin and a statin, they can still have progression of disease related to risk that isn’t well controlled after the initial medical treatment - that’s “residual risk.”
Residual risk can come in a number of different forms: related to lipids, inflammation, or blood clotting.
Look at that graphic again, there are 5 subheadings, each with a biological pathway, and then listed clinical trials that showed a benefit of treating that specific area of risk.
I’m going to focus on the area of “Residual triglyceride risk” because I think it’s where this graphic gets the most wrong.
Ridker cites REDUCE-IT, STRENGTH, and PROMINENT as evidence that treating elevated triglycerides (usually in the setting of low HDL-C) lowers cardiovascular risk.
STRENGTH and PROMINENT were both negative trials.
I even wrote a whole newsletter on the fact that what we think we know about triglycerides causing heart disease hasn’t been born out by the data:
And then there’s REDUCE-IT. In my newsletter on fish oil, I wrote about the big problem with this trial - the authors compared Vascepa to mineral oil, which impairs statin absorption, worsens lipid profiles, and increases inflammation.
It’s hard to feel very confident that Vascepa is better than a neutral placebo even if it’s clearly better than a harmful one.
And so you’re left with a whole bunch of the residual risk graphic that doesn’t really have evidence to back it up.
You can take this criticism of this graphic a bit further:
The Lp(a) trials are ongoing and we don’t have results, so who really knows
The DAPT trial was neutral if you balance out the ischemic benefit with the bleeding risk
The CANTOS trial was positive but with such unimpressive data that Novartis gave up on getting FDA approval
Trials of some other anti-inflammatory therapies have also been unimpressive
The biology here is really cool, but we can’t let our current understanding of it outpace the evidence when we treat our patients
We know triglycerides, inflammation, and Lp(a) increase cardiovascular risk - the data there is clear.
But when it comes to the level of evidence supporting using them as targets of treatment, I’m not sure that we’re ready for prime time yet.
The triglyceride and HDL story is one I keep writing about because it’s so important (and it’s also a set of lab values that I see all the time).
HDL raising treatments and triglyceride lowering treatments have failed in clinical trials enough times that we can’t assume that drugs that make these numbers look better do anything to change the outcomes that we care about.
It’s probably because we don’t understand the biology well enough and the low HDL/high triglyceride combo is the result of something else that’s dysregulated so our treatments aren’t actually getting at the causal factor.
But ultimately, if something doesn’t work in a clinical trial, you can’t use it on patients, even if it makes a ton of theoretical sense that it should work.
The recognition that our understanding of the biology has significant limitations is vital.
The more I learn about this stuff, the less confident I feel about the certainty of much of my knowledge - and we should all need to have that humility, no matter how deep our expertise might be.