Does lowering triglycerides reduce heart disease?
I’m kind of obsessed with triglycerides.
If you’ve been reading this newsletter for a while, you know that I think that most doctors don’t pay close enough attention to them.
High levels of triglycerides coupled with low levels of HDL cholesterol are the canary in the coal mine when it comes to metabolic syndrome.
And since metabolic syndrome is such a huge driver of chronic disease, it should follow that lowering triglycerides and raising HDL-cholesterol should reduce heart disease.
A new trial just presented at the recent American Heart Association conference aimed to answer that question.
The PROMINENT trial studied a medication that lowers triglycerides and raises HDL
The drug studied here is a medication called permafibrate.
Permafibrate belongs to a class of drugs known as fibrates - you may have heard of some its counsins, such as fenofibrate (Tricor) or gemfibrozil (Lopid).
Fibrates tend to lower triglyceride levels and raise HDL cholesterol, the exact problem that occurs in metabolic syndrome.
PROMINENT took patients with diabetes (the end stage of metabolic syndrome) who had elevated triglycerides and low HDL-cholesterol and randomized them to either permafibrate or placebo.
They followed the patients for about 4 years, and saw good reduction in triglyceride levels (about 26%):
They also saw a small rise in HDL cholesterol (about 5%):
But, unfortunately, despite making the biomarkers look better, permafibrate didn’t actually reduce the risk of heart attacks, strokes, or cardiac death.
The results here were as classically neutral as you’ll ever see:
One possible explanation here is that the benefit in triglyceride lowering and HDL raising was offset by a rise in LDL-cholesterol and apolipoprotein B levels:
But the simpler - and, I would argue, the more important - explanation is that the drug doesn’t work and that it doesn’t really matter why.
The biology here - especially when it comes to HDL - might just be way more complicated than simply looking at the amount in our blood.
There are two major lessons we should take from this study.
Lesson #1: targeting triglycerides and HDL with medical therapy are a great example of Goodhart’s Law
You may have heard of Goodhart’s Law before, which says that “when a measure becomes a target, it ceases to become a good measure.”
This is a constant issue in healthcare - we see that something is associated with better outcomes, and then we target it for improvement.
And the end result often isn’t better outcomes.
Goodhart’s law is usually thought to apply to metrics that measure quality in healthcare.
Think about waiting times for appointments (just ask the VA). Or patient satisfaction scores. Or even procedure related complications.
But this doesn’t just apply to quality metrics. It can also apply to treatments for things that are associated with better outcomes, but are probably just biomarkers of health.
Look at vitamin D. I’ve written before that I don’t think we should confuse a level of vitamin D obtained through sunlight and diet the same way that we look at a level of vitamin D obtained through supplementation because of the data on supplements to reduce disease kind of stinks.
The same thing probably goes for triglycerides and HDL cholesterol.
High triglycerides and low HDL are the downstream effect of dysfunctional metabolism and insulin resistance.
Targeting them directly for therapy doesn’t treat the underlying problem, it just treats the marker of the problem.
Lesson #2: you should have to prove that a treatment has an outcome effect before it becomes widely adopted and biomarkers are not enough
I’ve gotten into disagreements with my colleagues about whether we should be prescribing drugs like bempedoic acid or inclisiran for patients with elevated LDL-cholesterol levels.
The argument for these drugs is that they lower LDL-cholesterol and apolipoprotein B via a final stage mechanism of action that’s widely known to be effective: these drugs increase LDL receptors on the surface of liver cells, leading to more clearance of LDL from the bloodstream.
There is strong evidence that other classes of drugs that work via this mechanism - statins, ezetimibe, PCSK9 inhibitors - all lower cardiovascular risk.
But we don’t have outcomes data for these drugs, which means that we don’t know whether they reduce heart attacks, strokes, or death.
And so my argument is that when you don’t have data, you don’t really know whether the medication works.
The history of medicine is littered with drugs that *should have worked* but ultimately failed in clinical trials.
CETP inhibitors, drugs that raise HDL and lower LDL, are the quintessential example of this. These drugs are modeled after a genetic variation that seems to lead to longer lifespan.
There’s just one problem: torcetrapib, the CETP inhibitor that raised HDL and lowered LDL, unfortunately had the nasty side effect of increased risk of death.
A more benign example comes to mind in the realm of PCSK9 inhibitors, drugs that markedly lower LDL cholesterol levels and were similarly modeled off a genetic variation with significantly reduced risk of heart disease.
There are two PCSK9 inhibitors that have been approved, alirocumab and evolocumab.
They both do a great job of lowering LDL cholesterol. They both do a great job of lowering apolipoprotein B. And they both even lower risk of heart attack and stroke.
But alirocumab reduces the risk of death and evolocumab doesn’t.
Now, this may just be a statistical fluke, but it also may be a sign that the drugs don’t have the same biologic impact.
You only learn things about the outcomes that you care about by studying them directly.
A neutral trial like this is a wonderful step forward for science
I’ve written before that we get the type of evidence that we deem acceptable.
Trials like PROMINENT are an important step forward - they test the question of whether or not a drug does the thing that we want it to do.
While I may be a bit disappointed that the trial was ultimately neutral, I’m thrilled that it was done, because it’s only through rigorously testing treatments that we ever find out what really works versus what we hope will work.