Transcription
Two things can be true at the same time, depending on how you look at them. In this video, I'm going to delve into some of the nuances of ApoB in all-cause mortality, focusing on two studies in an attempt to make the Steelman case that ApoB is relevant to all-cause mortality. Even where it's not, it still might be more important than you think.
Warning: this video is going to be provocative, but really it's just an appetizer for what's to come—a 5-hour-long podcast between me and Simon Hill, the proof of which, after watching this video, I think you're going to be provoked to listen to. Yes, all 5 hours.
[Music]
Welcome to my channel. Stay curious!
First, for a tiny bit of background, I want to distinguish ApoB from a related measure, LDL cholesterol. ApoB is a lipoprotein, and ApoB-containing lipoprotein particles are the family of lipoprotein particles that include LDL particles. LDL cholesterol is the cholesterol content of LDL particles, and LDL particles constitute the most common ApoB-containing lipoprotein particles in the bloodstream.
When it comes to cardiovascular risk assessment, there's now a trend towards focusing on ApoB as opposed to LDL-C, since ApoB is a marker of LDL particle number, which is better than LDL cholesterol, which is cruder. By way of analogy, it's kind of like thinking about BMI, which is like LDL cholesterol, versus body composition on a DEXA scan. The body composition is more relevant to health; the BMI is cruder, but overall, at a population scale, they tend to run the same. So, LDL-C and ApoB tend to run together at a population scale. Where they diverge a little bit, ApoB is better, but for the sake of this video, you can consider them more or less the same if it's easier for you.
Again, for the sake of this video, there is controversy. Now, getting back to our main point over the relationship between ApoB and, by extension, LDL and all-cause mortality (or death by any cause), some people note a J-shaped relationship between ApoB and all-cause mortality, whereby very low levels of ApoB might not necessarily be better because you see very low levels associated with higher all-cause mortality.
But to appreciate whether this may or may not be a fair conclusion, we have to ask another question: What is the primary driver or primary predictor of all-cause mortality? The answer is metabolic vulnerability.
So, the next question: What is metabolic vulnerability? Metabolic vulnerability is a marker of dysfunctional metabolism, broadly inclusive of malnutrition, inflammation, and malnutrition-inflammation complex syndrome. Malnutrition and inflammation tend to synergize negatively in the body, and it's only very recently that we're starting to quantify metabolic vulnerability.
A really stellar paper that is central to our discussion was published by O.T. Vosol last year (2023) in The Lancet Healthy Longevity, and it showed that metabolic morbidity, a multimarker MVX score, dominated as a predictor of all-cause mortality.
Now, as an aside, the MVX score, the metabolic vulnerability score, was composed of GlyA (which is a systemic marker of inflammation), small HDL particle count, citrate, and three branched-chain amino acids: valine, isoleucine, and leucine. The specifics of why these six markers combined are such an effective multimarker for metabolic vulnerability is beyond the scope of this discussion, but just take it as these markers combined constitute a pattern—a signature of dysfunctional lipid metabolism.
Strikingly, the MVX metabolic morbidity score dominated beyond not just cholesterol for predicting all-cause mortality (5-year all-cause mortality), but factors like age. It was much stronger than even age at predicting all-cause mortality and was generalizable across different groups: the young and the old, male and female, different races, etc. Long story short, metabolic vulnerability quantified by this MVX score is a powerful predictor of all-cause mortality, i.e., death.
Now, why is this so important? Well, consider this: if I drop a pebble into a still lake, you'll see a ripple, right? Of course. But what if I drop the same pebble into white water or a raging stormy sea? Do you see the ripple? Well, of course not; it's lost in the raging stormy sea. This gets to the idea of a signal-to-noise ratio. If metabolic vulnerability is a really great predictor of all-cause mortality, it's like the white water or the raging sea with respect to all-cause mortality.
If you think of ApoB as the pebble with respect to all-cause mortality, dropping the pebble in the white water means you're not going to see a signal. However, that does not mean that one, ApoB isn't important in mortality or can't be important in mortality, and two, that ApoB can't be important in other ways.
Two points we'll hit on now, focusing first on mortality and ApoB. I now want to turn to another paper published by Lee et al. in 2022, a year before the Vosol paper characterizing the MVX score. We'll get to why that's relevant in a minute, and I want to jump straight to figures 5A and 5D and juxtapose them because you'll see there's a very interesting contrast.
Focusing first on 5A, what they're looking at here is the unadjusted model, whereby they're looking at the association between ApoB on the x-axis (whereby more to the right is a higher ApoB) and all-cause mortality on the y-axis (where higher is more death, higher all-cause mortality). You see a J-shaped curve, whereby at very low levels of ApoB, you see more mortality, and that might lead some to think, "Oh, lower ApoB is worse for all-cause mortality."
But wait! What if you then account for what is a proxy for metabolic vulnerability? They actually didn't use the MVX score because, again, it wasn't characterized for a year after this paper was published, but they adjusted for nutritional factors. What you see when you do that is a transformation of the J-curve into a line. You see a new relationship whereby lower ApoB continues to associate with lower all-cause mortality, lower death.
This is really important if you think about it because if you think about an intervention that might change—let's imagine an intervention that just magically changes ApoB as a singular change in an individual—that individual's metabolic vulnerability is going to be held constant. So, the line in that secondary graph where you make the adjustment may be more important when considering therapy on an individual level. That really shines a new light on this, doesn't it?
Another really important point I want to make is that the predictors of fatal and non-fatal cardiovascular events are distinct. For example, the metabolic vulnerability score is really great at predicting fatal events but not good at predicting non-fatal events. Of course, the non-fatal events are still very relevant to quality of life and health span. If you have a non-fatal stroke and end up impaired, is that relevant? You didn't die, but of course, it's still relevant.
So, just because a marker, say ApoB, isn't that strong at predicting all-cause mortality, and there might be something stronger like metabolic vulnerability, doesn't mean that ApoB isn't important for other things, including predicting non-fatal events, which are still important to health, health span, and quality of life.
So, those are really the main points I wanted to make in this video. But for those who want to stick around, the nuance ninjas, I want to make a couple of adjunctive points.
Extra nuance point 1: The physiological patterns we observe in things like the MVX score inform metabolism. It's not per se that the differences in branched-chain amino acids, which are included in the MVX score, are themselves protective or harmful, but the composite, the fingerprint, the signature tells us something about the underlying metabolism. That's really important since metabolic health is so important to overall health.
What I think we really should be looking at is patterns—things like the MVX score, but also the LMHR (lean mass hyper-responder) triad, which constitutes the high LDL, the high HDL, and the low triglyceride patterns seen in some people who go low carb. My interest in that, similar to the MVX score, is not that I think the high HDL is protective or the low triglycerides are protective, but the pattern together tells you something unique about the underlying physiology, and that can be very interesting and informative to risk and outcomes. We need to study that more.
So, point one: Physiological patterns and signatures inform us about metabolism, and we need to look at things in context.
Now, point two: Interventions come with side effects. It was nice in this video, and it's a useful exercise to think about the abstraction of, "What if I just snap my fingers and lower ApoB? What would that do to my risk?" But in the real world, you have real patients with comorbidities and complex conditions, and interventions are required to have an effect on a biomarker. You can't really just snap your fingers and affect one biomarker. You can't snap your fingers and lower ApoB; you need to do something—a dietary intervention or take a medication.
In individual patient cases, these interventions can come with side effects or harms or unknown long-term consequences because we don't have enough long-term data to understand them. So, I think it's useful to make the Steelman argument for ApoB lowering with respect to all-cause mortality, but also recognize humbly that there's a lot we don't know about the interventions required to exert ApoB lowering. Just consider that on an individual patient basis; there may be more complexities in determining what might be prudent for an individual with respect to care.
Of course, that should be the realm of patient-physician discussions. And don't get health information or health advice off social media, YouTube, Twitter, etc. That goes without saying.
Now, finally, to wrap up, I want to give a hat tip to two people. One is Simon Hill, the proof. We actually just sat down for a 5-hour-long podcast that was a tremendous discussion. If you know a little bit more about me and Simon, you might think that we will be quote adversarial, but sitting down with him for 5 hours, man, I think this guy is going to be a lifelong friend. We got on great, and I think that there was so much nuance we unpacked in those 5 hours. I really, really hope that you listen to that episode. If you liked this video, you're going to like the nuance that we get into in that particular episode, so I'll link it below when it drops.
And two, I want to give a hat tip to Professor William Cromwell, Bill Cromwell, who really turned me on to all the literature I just discussed with you and has really mentored me and taught me. What I'm trying to do now is translate the lessons I've got from him to you, so hopefully I was able to do that to some extent. But I'd really like to credit Bill Cromwell for all I've learned on this topic.
Um, Simon and Bill, hats off to you. You're both great gentlemen, and everyone, have a lovely day.
[Music]