Retrospective analysis of 486,956 Han Chinese from the Taiwan Precision Medicine Initiative evaluated prevalence and clinical impact of actionable pharmacogenetic (PGx) variants across 19 genes and 58 drugs, with focused outcome analyses for four gene–drug pairs. Findings confirm widespread PGx variation and statistically increased risks for some adverse events, but absolute excess risks were small and most variant carriers tolerated therapy.
0:00Welcome to Base by Base, the papercast that brings genomics to you wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. Imagine, um, for just a second, a perfectly normal trip to the pharmacy.
0:12Right, like an everyday errand? Yeah, exactly. You're picking up a common prescription. Maybe it's a painkiller for a bad knee or, you know, a daily statin to keep your cholesterol in check. Or even a blood thinner that your doctor recommended after a recent hospital visit.
0:26And you just take it, right? I take the pill and you fully expect it to do its job and make you better. But, well, what if your very own DNA? was secretly hardwired to reject that medication. Oh wow. That's a scary thought It is.
0:40What if, instead of healing you, your genetics caused a massive, severe adverse reaction? See, that's the thing. We operate under this, um, this assumption that modern medicine is basically one size fits all?
0:53You have a symptom, you get the standard drug for that symptom, and you just move on. But on a molecular level, I mean, the human body is vastly more complicated. Our cellular machinery dictates exactly how a chemical compound behaves once it enters the bloodstream.
1:07And the scaggering reality here is that nearly every single person on Earth is walking around right now with genetic booby traps for certain medications. You have them. I have them. Yeah, it's pretty universal.
1:18Our genetic codes are just filled with these subtle variations that completely alter how our bodies process the most common drugs on the market, but what really happens when doctors try to use our DNA to predict these drug reactions in the real world.
1:33And how could this change your next trip to the pharmacy? That is the heart of what we're looking at today. Today we celebrate the work of the Taiwan Precision Medicine Initiative. Researchers from Academia, Seneca, and 33 partner hospitals across Taiwan, who have advanced our understanding of pharmacogenetics and massive non-European populations.
1:51The sheer scale of their research is, frankly, mind boggling, but to really appreciate the paradigm shift they're proposing, I think we need to lay down the basics of pharmacogenetics, which is often just called PGX.
2:03Yeah, let's start there. Pharmacogenetics is the study of how your unique genetic makeup affects your biological response to drugs. When you swallow a pill. It doesn't just magically teleport to the target site and start working.
2:16Right, it's a whole process. Exactly. Your body has to actively process it. The liver is essentially this massive chemical manufacturing plant. It uses specific enzymes and transporters, all of which are built using the blueprints in your DNA, to break that drug down or convert it into an active form.
2:34And eventually clear it from your system, right? Spot on. So, if your DNA contains a variant, like a typo in that genetic blueprint, the resulting enzyme might be built incorrectly. Which means it might work way too fast, right?
2:47Like metabolizing a drug before it even has time to help you. Yes, or the opposite. It might work too slowly. And that's when the drug can build up to toxic levels in your bloodstream. Yes. So this biological reality has pushed major regulatory bodies.
3:01Um, like the US food and drug administration and the clinical pharmacogenetics implementation consortium. which is a mouthful. Yeah, we usually just call them CPIC. Right, CPIC. So they've actually issued strict clinical guidelines.
3:15For years, they've been telling clinicians to alter dosages or completely avoid specific drugs if a patient carries certain genetic variants. Because the risk of severe toxicity or total failure is just too high.
3:29Okay, let's unpack this. Because there is a massive blind spot embedded in those guidelines. There really is. And it's historical one. Yeah, the foundational studies that establish these genetic warning labels were generally pretty small, but more importantly, they were conducted almost entirely on populations of European descent.
3:46Which is a huge limitation. That's like trying to navigate the bustling streets of Taipei using a map drawn for London. You might recognize a few universal concepts like roads and intersections, but you are going to get hopelessly lost.
3:59And probably crash. Exactly, because you simply don't have the localized specific data for the environment you're actually in. If we connect this to the bigger picture, Applying European-based genetic guidelines to a global population without testing them thoroughly in local context, 1st is scientifically flawed and potentially dangerous.
4:18Right, because human populations differ genetically just due to millennia of migration and geographic isolation. Exactly. We know, for example, that certain genetic variants that severely impact the metabolism of klopodogril, which is a critical blood thinner, or as atheoprine, a really potent immunosuppressant.
4:34Those are significantly more prevalent in Asian populations than in European ones. Meaning an entire demographic might be at a much higher statistical risk for adverse drug reactions. And the global medical community wouldn't even fully grasp the reality of that risk if they only viewed the problem through the lens of European data.
4:53Right. So to correct this historical imbalance, the researchers behind the Taiwan precision medicine initiative undertook a project of staggering proportions. Yeah, staggering is the word. They analyze the genetics and medical histories of 486,956 Han Chinese individuals.
5:09That's nearly half a 1000000 people. The logistics of gathering and processing genomic data for half a 1000000 individuals across 33 different hospitals is just immense. They used advanced electronic medical records, EMRs, and cross-reference those clinical histories with custom design genetic tests.
5:26These are the S&P arrays, right? Specifically, the TPMV1 and TPMV2 arrays. Yes. And those arrays are designed to rapidly scan 100s of 1000s of single nucleotide polymorphisms or SNPs. And SMP is basically just a single letter typo in your DNA sequence.
5:44That's right And when that typo creates a distinct recognizable version of a gene, we call that an allele. So they cast this massive net across the genome, but then they narrow their focus down to 19 critical pharmacogens.
5:58These governed how the human body processes 58 commonly prescribed medications. But, you know, the really defining feature of this research is that it's retrospective, it was essentially a massive natural experiment.
6:09Oh, because when these half a 1000000 patients originally went to their doctors over the past decade, Genetic testing wasn't the standard of care. Exactly. The doctors were prescribing these drugs blindly, entirely unaware of the patient's underlying genetic risks.
6:23Wow. So the researchers were able to look back at the EMRs and see what actually happened in the real world when 100s of 1000s of people unwittingly took medications that their DNA suggested they shouldn't take.
6:34To extract meaningful insights from that mountain of data, they focused on 4 specific gene drug pairs. The massive, simple size gave them unprecedented statistical power for these categories. And those were Klopido Grill, the blood thinner.
6:47As atheoprane, the immunosuppressant, statins for cholesterol, and NSAIDS, which are your everyday pain and inflammation drugs like ibuprofen. Right. And the baseline genetic findings across the cohort were pretty shocking.
7:01Mind blowing, honestly. Out of nearly half a 1000000 individuals, 99.9% possessed at least one actionable PGX risk variant. Virtually the entire population. Yeah. Everyone carries a genetic marker that current guidelines flag is requiring a clinical adjustment.
7:18While a lot of this was driven by a really prevalent variant in a gene called VKORC1. The data still show that the average individual carried 4.3 clinically actionable risk variants. The implication there is wild.
7:30It means the average person is walking around with over 4 standard medications out there in the world that according to their DNA require a dose reduction, extra monitoring, or just a totally different therapy.
7:39Here's where it gets really interesting, because I had to push back on the implications of those raw numbers. Oh, how so? Well, if 99.9% of the population carries these variants and doctors have been blindly prescribing these 58 common drugs for years without checking DNA, the hospitals in Taiwan should be in a state of constant crisis.
7:59I see where you're going. We should see overflowing emergency rooms dealing with catastrophic adverse drug events every single day. But that isn't the reality on the ground. No, it's not. And you are hitting on the central paradox that the study eliminates.
8:14There is a profound disconnect between the statistical risk indicated by a genetic variant and the absolute risk. The actual number of patients who suffer tangible harm. Exactly. When we analyze those 4 key drug pairs, the nuance of human biology becomes incredibly clear.
8:29The genetic variants absolutely increase the statistical likelihood of an adverse event, but the absolute numbers remain surprisingly and reassuringly low. Okay, let's take a drug like Kloppit Ogrill. It's the standard blood thinner prescribed to prevent major heart attacks and strokes, especially after a patient gets a stent put in.
8:49The paradox here hinges on the fact that Kalapagogro is actually a pro drug. Which is a crucial biological distinction. When you swallow a pro drug, the chemical compound is completely biologically inactive, it cannot thin your blood in that state.
9:04So it has to be activated first. Right. Right. It must travel to your liver, where an enzyme encoded by the CYK 2C19 gene physically alters the molecule, converting it into its active form. And if you carry a loss of function variant for that gene, your litter manufactures a defective enzyme.
9:20It simply cannot activate the drug effectively. So a patient in that scenario is essentially taking a placebo while remaining in an imminent risk for a catastrophic blood clot. That sounds terrifying. It does.
9:31And the retrospective data confirmed that patients with this loss of functionalille did face a significantly higher relative risk for major adverse cardiovascular events or MAC. But the absolute reality is the critical takeaway here.
9:43Because 85.9% of the patients carrying that exact defective genetic allele tolerated the drug perfectly well. They never experienced a major cardiovascular event. Over 85% of patients with a genetically broken activation pathway were completely fine.
9:59That's amazing. The body must have redundant pathways, or maybe the baseline efficacy of the drug is just so strong that even a reduced active concentration is enough for most people. It's likely a combination effectors.
10:10Now, let's contrast that failure to activate a drug with the opposite problem. A failure to clear a drug from the system. We see this with the immunosuppressant as atheoprin. Right, which is used for severe autoimmune conditions and organ transplants.
10:24It's metabolism relies heavily on enzymes produced by the NUDT 15 and TPMT genes. They're responsible for breaking down the drugs toxic byproducts. So if a patient carries high risk variants in these genes, The clearance mechanism just fails.
10:38The toxic metabolites accumulate rapidly in the bone marrow, which leads to severe hair loss, and a precipitous life-threatening drop in white blood cells. Which is known as neutropenia. The study certainly found a strong statistical association linking those variants to the toxicity.
10:53But again, the absolute clinical numbers tell a much softer story. When looking at the patient's flagged as high risk by their DNA, Only about 21% actually experience these severe adverse events. And while a one in 5 chance of severe toxicity sounds alarming, Context is everything.
11:09In the normal control group, patients with healthy functioning alleges for those genes, about 17%, experience severe adverse events anyway. Oh, wow. So the genetic variant only raises the risk profile from 17% to 21%.
11:23That is a measurable increase, but it is far from a biological guarantee of toxicity. Exactly. And the narrative surrounding statins is even more striking. Millions of people take statens daily to manage cholesterol.
11:35They rely on specific transporter proteins, encoded by the ABCG2 and SLCO1B1 genes, to pull the drug out of the bloodstream and into the liver. Well, it does its actual work. Right. But if those transporters are genetically defective, the statin remains circulating in the blood, eventually seeping into muscle tissue.
11:52This causes the primary known risk of statins, which is severe muscle pain and a dangerous muscle breakdown called myopathy. And current pharmacogenetic guidelines often recommend lowering the dose or just avoiding certain statens entirely if a patient carries these defective transporter variants.
12:08Yet within this massive cohort of half a 1000000 individuals, over 98% of the high risk patients never developed any muscle issues whatsoever. 98% of the patients flashing a genetic red flag were completely asymptomatic.
12:24I mean, the human body's ability to tolerate these pharmacological mismatches is astounding. really is. Finally, we look at the everyday painkillers, the NS ides, like ibuprofen or solicoxet. These are processed by the CYP 2C9 gene, and the primary severe risk is gastrointestinal bleeding.
12:40In the case of NSides, the genetic data barely registered as a clinical factor. The retrospective analysis clearly demonstrated that if a patient suffered GI bleeding after taking an NSAID. Their everyday clinical realities were overwhelmingly to blame, rather than their DNA.
12:57Meaning things like underlying comorbidities like diabetes or hypertension, or taking other medications concurrently, like diuretics. Those were vastly stronger predictors of a severe bleeding event than the presence of the CYP 2C9 genetic variant.
13:11In that scenario, the genetics are merely a whisper, while the patient's overall clinical health is a shout. What's fascinating here is the incredibly delicate balancing act this data creates for practicing physicians.
13:23Yeah, it's not straightforward at all. Not at all. PGX testing is undeniably a vital tool for illuminating hidden biological risks, but it is absolutely not a crystal ball, blindly adhering to genetic guidelines like avoiding a medication, simply because a test flags a variant can inadvertently inflict severe harm on a patient if the broader clinical context is ignored.
13:42That concept requires careful consideration. Because if a genetic test flags a specific drug as high risk for toxicity or failure. I mean, the intuitive response is to simply swap it out for an alternative medication.
13:55But the flaw in that intuition is assuming the alternative medication carries a universally lower risk profile. Let's return to the blood thinner. Clopadugral. Okay. Imagine a 75-year-old patient who carries the loss of function allele.
14:10A doctor, looking at the genetic test, might decide to avoid clipodogal, and instead prescribe an alternative, more potent blood thinner like decogriller to ensure efficacy. Right. That makes logical sense on paper.
14:22But those stronger alternatives carry a significantly higher baseline risk of spontaneous life-threatening gastrointestinal or intracranial bleeding, particularly in elderly populations. So, but trying to save the patient from a statistically elevated risk of a drug failing to activate, the doctor might actively trigger a severe bleeding event with the alternative drug.
14:43Precisely. And the same logic applies to the statin paradox. The primary side effect we worry about with statin's muscle pain is entirely reversible. If a high risk patient begins a statin regimen and develops myopathy, they can cease the medication, and the muscle tissue generally recovers within weeks.
15:01But conversely, if a physician never prescribes the statin in the 1st place, due to a genetic red flag, that patient remains at a high risk for a fatal myocardial infarction. Something the statin could have easily prevented.
15:14Exactly. A genetic risk variant is really like a caution, slippery when wet sign on a winding road. I love that analogy. Thanks. It is a vital piece of information. It means you should drive carefully, pay close attention to your surroundings, and perhaps check your tire tread.
15:30It does not mean you should abandon your vehicle on the shoulder and attempt to walk the rest of the way in the dark. Exactly. The sign is there to inform your navigation, not paralyze your journey. The clinical utility of a genetic variant must always be weighed alongside a patient's age, their comprehensive medical history, their daily lifestyle, and the stark reality of whether a genuinely safer, equally effective alternative drug actually exists.
15:55Furthermore, to fully grasp the landscape of pharmacogenetics, we must acknowledge the inherent limitations of even a study of this unprecedented scale. Because, as massive as a half 1000000 person cohort is, the retrospective data collection wasn't flawless.
16:12No, it had gaps. The electronic medical records only captured data from the specific hospital where a patient was actively enrolled in the initiative. Oh, right. So if a patient suffered an adverse reaction and sought treatment at a competing hospital network across town, that critical data point was entirely invisible to the researchers.
16:30Exactly. Similarly, tracking over-the-counter NSAI consumption is virtually impossible in a retrospective model, since patients self medicate without prescription records. That makes sense. Beyond the fragmented clinical data, there is a substantial limitation rooted in the genetic testing technology itself.
16:46The study relied on S&P arrays, which Excel, finding single letter typos. But this meant they completely missed one of the most critical pharmacogens in the human body. The CYP2D6 gene, which is involved in processing roughly 20% of all prescription medications on the market.
17:04And the CYP2D6 gene is notorious for highly complex structural variations. Yeah, we aren't just talking about a single missing letter of code here. We're talking about massive chunks of DNA being duplicated, inverted, or entirely deleted, right?
17:19Precisely. Imagine an instruction manual where entire chapters are photocopied and inserted 3 times, or whole sections are ripped out and pasted back in reverse order. Standard SMP arrays cannot reliably read those massive structural anomalies.
17:32Wow, so the researchers had to exclude CYP 2D6 from their analysis entirely. They did. Furthermore, the EMR data could not explicitly track the highly dynamic phenomenon known as pheno conversion. Ah, vino conversion.
17:44That is one of the most fascinating mechanisms in pharmacology. It's the biological reality that external environmental factors like your daily diet, your liver health of the other routine drugs you are taking can temporarily override and rewrite how your genes actually express themselves.
18:01A patient might possess the pristine genetics of a rapid drug metabolizer, but if they drink a large glass of grapefruit juice, or take a common daily antidepressant that chemically blocks their liber enzymes, they suddenly act biologically identical to someone with a defective, poor metabolizing genetic allele.
18:19That's wild. They're observable phenotype converts, completely independent of their underlying DNA sequence. Yes, your morning routine can essentially override your genetic code by the time you take your evening medication.
18:30That introduces a staggering layer of complexity to personalized medicine. I mean, to truly capture the full picture of the patient's risk, the field needs to evolve past basic SNP arrays and move toward whole genome sequencing.
18:43Which can read those massive structural variations. Right. While seamlessly integrating comprehensive nationwide health databases to capture every hospital visit, prescription, and ideally lifestyle factor.
18:55The future of precision medicine relies on building a holistic, continuously updating biological profile of the patient, rather than isolating a single strand of DNA and treating it as biological destiny.
19:06So what does this all mean? If we distill this unprecedented study of half a million individuals down to its core? Here is the takeaway. Karmacogenetics is an undeniably powerful, necessary tool for the future of medicine, proving definitively that our genetic variations influence how our bodies interact with pharmaceutical compounds.
19:27It's undeniably a huge leap forward. Absolutely. However, discovering a genetic risk variant is not an absolute predictor of harm, it's merely one piece of a complex biological puzzle. That genetic data must be carefully weighed alongside a patient's clinical history, their environment, and the pragmatic reality of whether a safer, better drug alternative actually exists in order to truly optimize their care.
19:49The disparity between our genetic blueprints and our clinical realities is a humbling reminder of human resilience. It also introduces a compelling new dynamic to consider. What does this mean for the future of personalized medicine when we have to accept that our DNA provides us with complex probabilities rather than simple absolute certainties?
20:10It really makes you wonder how long it will be until your local pharmacist doesn't just ask about your allergies, but needs a detailed log of exactly what you ate for breakfast, and what herbal supplements you take before they can safely hand over a standard prescription.
20:24The pharmacy of the future might look a lot more like a biochemical interrogation. This episode was based on an open access article under the CCBY 4.0 license. You can find a direct link to the paper and the license in our episode description.
20:38If you enjoyed this, follow or subscribe in your podcast app and leave a 5 star rating. If you'd like to support our work, use the donation link in the description. Now stay with us for an original track created especially for this episode and inspired by the article you've just heard about.
20:53Thanks for listening and join us next time as we explore more science base by base.
