Does Aduhelm mark the return of science-based medicine? [PODCAST]

Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes

“While we still lack a complete cause and effect model of Alzheimer’s disease, there is no doubt that the etiology is multifaceted and nonlinear. This accounts for the fact that it takes an incredibly long time for Alzheimer’s disease to develop. I mention this because one of the major problems with FDA approval has always been that it insists on utilizing the same framework and methodology of approving drugs designed to treat long-term diseases as it would with short-term diseases.”

Robert Trent is a graduate student who blogs at Medaphysics.

He shares his story and discusses his KevinMD article, “Does the FDA approval of aducanumab mark the return of science-based medicine?”

Did you enjoy today’s episode?

Please click here to leave a review for The Podcast by KevinMD. Subscribe on your favorite podcast app to get notified when a new episode comes out!

Do you know someone who might enjoy this episode? Share this episode to anyone who wants to hear health care stories filled with information, insight, and inspiration.

Hosted by Kevin Pho, MD, The Podcast by KevinMD shares the stories of the many who intersect with our health care system but are rarely heard from.

Transcript

Kevin Pho: Hi, and welcome to the show, where we share the stories of the many who intersect with our health care system but are rarely heard from. My name is Kevin Pho, founder and editor of KevinMD. Today we welcome back on the show Robert Trent. He is a biochemist, and he wrote the KevinMD article “Does the FDA approval of aducanumab mark the return of science-based medicine?” about aducanumab, or Aduhelm, the recently approved drug for Alzheimer’s. Robert, welcome back to the show.

Robert Trent: Happy to be here.

Kevin Pho: We’ll get into your article in a little bit, but for those who didn’t listen to our first episode together, can you just briefly share your story and journey to where you are today?

Robert Trent: I was originally trained as a biochemist. I had a lot of fun with that, but at the same time, I found the education for it quite tedious and boring, and I tried to find outlets for other types of patterns and see what it was I was actually learning at the substrate-independent level. From there, I actually got into complexity science. Then, from becoming an amateur student of complexity science, I really got into understanding and investigating ways of thinking and ways of approaching problem-solving. So that’s a little bit of the background behind this article.

Kevin Pho: So tell me about your background in complexity science and biochemistry, and those intersections with medicine and health care.

Robert Trent: I’m a little biased, right? Because if you’re a biochemist, you’re probably going to be thinking in terms of the molecules and all those types of interactions that are going on, and so sometimes you don’t have the option to see health care from the perspective of the big picture. A lot of biochemistry funding is associated with the NIH, associated with health care, stuff like that, so that pathway was kind of inevitable. I also worked a little bit in genetics before I realized that was not something I wanted to do. Props to people who do it, but it just wasn’t my thing.

So I think those things are kind of what led me to this point, really just trying to come up with a way to have whole-systems thinking, right? A way to see every little part, not just that one specific enzyme. If you do an advanced degree, it’s a pyramid, right? In undergrad, you know a little bit about a lot of things. In a master’s, you get to be more specialized. And then if you do a PhD, you’re the world expert on that one thing. I just don’t like that hyperspecialization. So I really reached out, and it’s totally amateur, right? I just read a lot of books about complexity science, and I find it just such an interesting field.

Kevin Pho: All right, so let’s transition into the KevinMD article that you wrote. It’s titled “Does the FDA approval of aducanumab mark the return of science-based medicine?” Before you get into that article, can you give my audience a little bit of context about aducanumab and the issues surrounding it, just to give a little bit of background to your article?

Robert Trent: Yeah. So aducanumab, or I’m just going to call it Aduhelm. I don’t understand the nomenclature that we’ve adopted for antibody therapy. This is, I think, the first time that a drug has been given FDA approval not based on what happens to the patient per se, but based on what’s going on at the molecular level, right? So this was a drug that was approved for Alzheimer’s patients not because it improved cognition or anything like that, but rather because it just cleared up that protein, that beta-amyloid protein that’s been implicated in a lot of Alzheimer’s cases. And that’s really kind of the first time, I think, that that’s happened, at least in a long while.

Kevin Pho: Tell me about the controversy surrounding that. So aducanumab, or Aduhelm, has a theoretical benefit, and I think that one of the reasons why the FDA approved it is because there are no other treatments for Alzheimer’s currently. So by having this potential theoretical benefit, it gave a conditional approval, which of course leads to a whole host of issues. So talk about some of the issues surrounding that.

Robert Trent: Yeah. Well, one of the issues, I think, is patient safety, which people are very concerned with. People don’t want to be someone who is not looking before we leave. We don’t want to be doing that per se. And in the specific details of that, obviously my background is not necessarily in antibody therapy and things like that, but the specifics behind that are just, well, we don’t know, right? They’re kind of just overwhelmed by this fear of the unknown kind of thing here.

And I think the benefits of it, and this is just my opinion, right, but I do think the benefits of it outweigh the risks, because at this point, right, we don’t have a complete cause-and-effect model for Alzheimer’s, right? It’s a very nonlinear thing, right? It’s not like TB. TB is a totally linear, one-to-one thing that happens, but Alzheimer’s, right, is nonlinear. It’s nonlinear, and it takes a very long time for Alzheimer’s disease to develop.

And maybe someone in your audience knows the exact models that we can come up with that factor in the relationship between beta-amyloids, probably tau and those other proteins involved in that, and then time, and how all that stuff combined works together. But I’ve never come across any quantitative model that accompanies that. So if it takes a long time for Alzheimer’s to develop, well, it’s probably going to take a long time for any benefits from treatment to be displayed. Regardless, though, if it was me, right, I do not want plaquing in my brain, right? I do not want that cell death there. And so I think we need to start advocating for what’s best for the patient, and that’s just my thoughts on the issue.

Kevin Pho: Now, have there been cases where you have a theoretical benefit in a lab setting, however, in the real-world setting, those benefits don’t always translate? And I think that’s the concern a lot of people have regarding this approval. So what do you think about that?

Robert Trent: I’m not saying that the approval of this is 100 percent science-based medicine, good job, right? Unfortunately, because of ethics and things like that, right, and I’m saying that kind of playfully, we don’t work with well-controlled experimental conditions in clinical data, right? I mean, you don’t have genetically identical humans. You don’t control their lives, you don’t control their diet, and you don’t do any of this stuff, right? And so on top of that, we don’t even have a complete cause-and-effect model of Alzheimer’s.

Sure, but one of the problems that I have, just as a biochemist, as someone who works on the science side of things, not on the medicine side of things, is I’ve found that medicine and science have very different value systems. Medicine really values statistics and data more than it does mechanistic relationships. And I think the approval of Aduhelm, I’m going to call it that, marks a little bit of a shift in the value systems, a bit more towards the midpoint between those two deviations, right? And I think that’s a good thing, because there are numerous other examples of circumstances where, sure, yeah, right, we have chaos or complexity theory at work here, and we can’t make definitive predictions, but sometimes statistics and data can lead you down a blind alley.

Kevin Pho: Do you think that this approval of Aduhelm on a molecular level, like benefits on a molecular level, do you think that this is going to start a trend going forward with some other therapies in development?

Robert Trent: I would hope so. The thing that kind of frustrates me about the conversations I have with people who are trained in the biomedical sciences is that they really rely heavily on clinical data. David Gorski over at Science-Based Medicine has talked a lot about that, evidence-based medicine worshiping clinical trial data above everything else. And I think a solution to that is to start to maybe just think more in terms of mechanistic relationships, right? I really just want to emphasize that statistics can really lead you down a very, very blind alley, right? And without a theoretical foundation to guide you through that data, you can really get lost in the needle of that haystack.

To give an example, right, statistically speaking, patients who are obese who exercise live longer than patients who are not obese who don’t exercise. That’s one of the claims being made by a lot of people in the social sciences, and in science and evidence-based medicine; people are claiming that. However, right, we both know that if you’re obese, you develop metabolic syndrome, you get non-alcoholic fatty liver disease, all the things. There’s objective evidence that that stuff is bad.

And just to maybe veer a little bit off topic here, let’s compare two big data projects, one of which had a theory and the other of which didn’t, right? In Geneva, Switzerland, there’s CERN, with the Large Hadron Collider. We could have built a Superconducting Super Collider here in Texas, but Bill Clinton didn’t want that to happen. And there’s also something in the biomedical sciences called the Encyclopedia of DNA Elements, or ENCODE. These are two big data projects just trying to investigate relationships, right?

Physics has a lot of theories. Biology, the biomedical sciences, they don’t have a lot of theories. They just don’t, right? Natural selection is pretty much one of the key things there, but it’s inherently qualitative. It’s not really a quantitative, mechanistic thing. I’m going to get a lot of flak for that from the ecologists or systems biologists in the room. I’m not picking on you per se. I’m just talking about the people in the lab knocking out genes and saying, “Oh, I knocked out this gene, and that’s what caused it.”

But CERN produced massive amounts of data, 200 petabytes, I think, when last I checked; that’s how much they produced. I think Netflix is like two petabytes, so this is huge, right? ENCODE also worked with massive amounts of data. CERN was able to find evidence of the Higgs field really quickly through that 200 petabytes. They found evidence of the Higgs field because they had a theory to guide them through that data. ENCODE? You’ve probably never heard of ENCODE, because no one knows what the data from ENCODE means. No one. You can’t derive a function, right, a functional relationship, “this does that,” just from statistics.

Stuart Kauffman is a former physician who became a theoretical biologist. He worked at the Santa Fe Institute, and he’s written incredible books just about thinking; he’s kind of a philosophy guy. He talks about that a lot. If you look at just the statistical relationships among components in medicine, or components of a biological system, you can’t derive function from that. You will think, “OK, the heart exists to add space between the lungs, not to pump blood.” And so that’s the concern I have: We’re not focusing on functional relationships, and we’re not focusing on mechanisms that really rule out the fuzziness of statistics.

Kevin Pho: We’re talking to Robert Trent. He’s a biochemist. He wrote the KevinMD article “Does the FDA approval of aducanumab mark the return of science-based medicine?” Robert, what’s the response that you got from the KevinMD article, and did you receive any pushback from some of the readers who read it?

Robert Trent: I think Evell Snoats, I don’t know if I’m pronouncing that right, made a really interesting comment. Thank you for commenting on this. He talks about “great ideas, mechanistically, don’t always pan out for a number of reasons. This is also the first I’ve seen that social sciences influenced the development of EBM,” and he says that my link has nothing to do with that assertion. Can I support that statement?

Yes. The thing that I’m saying here, and really, the first person that I’m going to say is quite controversial, but Dr. Sally Satel out at Yale College of Medicine wrote a book basically about how the social production of disease and public health belief systems can kind of lead things down a blind alley, right? You get things like belief in homeopathy and other stuff like that. That’s the first thing that I’ve seen.

And the point that I’m making, though, Evell, is that both the social sciences and evidence-based medicine place, in their value system, a higher connotation on statistical data, a probabilistic framework, rather than a science, which places a mechanistic framework higher up in that hierarchy. I’m not talking about all or nothing, like, “OK, it all has to be mechanistic,” or, “It all has to be statistical.” There is room between the two. But I’m saying that in the hierarchy of value systems, maybe we need to bring statistical data down a little bit and bring mechanistic data, or mechanistic relationships, up a little bit further when we go into our considerations.

Kevin Pho: And what’s the proper balance? Is it a case-by-case basis based on the situation and the particular drug in development?

Robert Trent: Well, this is actually where I think complexity science comes to the rescue, right? If you can develop quantitative models for the relationships that you’re working with, right? And some things aren’t easily done, right? It is kind of a case-by-case basis. But you have some things that you can quantify, right, models that you can quantify, like drug dosaging, for example, right? That is heavily predicated on, frankly speaking, I think we know all the dominant variables that are involved with drug dosaging, right? You have the size of your cardiovascular system. It branches off at so many points, right? There’s a finite distance that things can diffuse from a capillary, all that stuff there. And so I think if we were to work more on getting quantitative relationships, quantitative models, we could probably develop a way to balance those two things out.

But I do think ultimately it has to kind of come on a case-by-case basis. It has to go off of really how good the data is, how good the experimental data is, because without a viable control group, without well-controlled experimental settings, and then good, testable metrics, right, it’s really, really, really hard to do that.

Just to give one example, there are many intelligent physicians out there who are very smart people who say, “Hey, you know what? PPI usage doesn’t increase the risk of developing C. diff.” But there’s also a lot of data out there that says that the optimum pH range if you want to grow C. diff is within that range of about pH five or six, which is what you get if you’re taking PPIs. That means that every enzyme in that C. diff bacterium is in an optimal conformation, and so those are the great conditions you need for growth.

And so I think that to say, well, the probability is low that if someone takes a PPI they’re going to develop C. diff, that doesn’t make sense to me. Because, one, we have to look at how we test for C. diff, right? We don’t test for C. diff until someone actually develops the condition, which can lay dormant in patients. If they ingest an endospore, it can lay dormant for very long periods of time. And because I used to work in diagnostics, I can say that PCR for C. diff is so sensitive that some physicians aren’t even using it anymore, and they’re using it in contrast with other cytotoxic assays or things like that, right, just to verify if it’s actually C. diff, because it’s so sensitive it can’t differentiate colonization from infection.

And those minor details, right, that’s not explicit in the data, right? It’s not explicit when you look at the probability, how that probability was derived and developed. So I’m very skeptical of that, and I think it’s definitely something that we need to look into. We need to figure out how we’re going to do it. But I think complexity science is, and I guess I’m just biased, right, I think complexity science is going to pave the way. For the things that can be quantified, it would be in the framework of complexity science.

Kevin Pho: And my final question: What’s your take-home message that you want to leave with the KevinMD audience?

Robert Trent: Be skeptical of statistics, and be skeptical of data. If you really want to know more, I guess I would say Charles Seife, who is one of my favorite science writers out there. He wrote a book called Proofiness: The Dark Arts of Mathematical Deception, and in it he basically just breaks down a lot of the, maybe, showboating manipulation that people are doing with data.

I see a lot of physicians who are very busy, who don’t have time to read the methods of things. They don’t have time to read; they just go straight to the conclusion of a publication on PubMed or on NCBI or something like that. So maybe just be a little bit more skeptical if there’s not a control group, or if the test of statistical significance isn’t explicitly clear about what they were doing, right? That kind of stuff would be my take-home message. And maybe give a little bit more credence to mechanistic relationships.

Kevin Pho: And how can people reach you?

Robert Trent: Yeah, I have a website called Medaphysics.com, and that’s “med” like med, not “meta.” On there, you can find a lot of additional writings. I’ve done a lot of educational content on a whole wide range of topics in biochemistry, immunology, epigenetics, and things like that.

Kevin Pho: Robert, thank you so much for coming back on the show and sharing your time and insight.

Robert Trent: Thanks.

Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes