Pharmacogenomics: Why People Respond Differently

Same dose, two completely different outcomes

About one in twelve people carry a CYP2D6 gene variant that makes them "ultra-rapid metabolisers" , their livers chew through certain compounds so quickly that standard doses barely register. Their neighbour might carry a loss-of-function variant in the same gene and experience the opposite: slow clearance, prolonged exposure, effects that linger far past what the prescriber expected. Same molecule. Same milligram count. Radically different pharmacology.

That's pharmacogenomics in a nutshell. It's the study of how an individual's genetic makeup shapes the way their body processes, responds to, and eliminates substances. And whilst the field has been formally recognised since the early 2000s , the Human Genome Project gave it a real platform , the underlying observation that people vary wildly in drug response is older than modern medicine itself.

The enzymes at the centre of it all

Most of pharmacogenomics comes back to a family of enzymes called the cytochrome P450 system, or CYPs. These are the liver's main metabolic machinery for breaking down foreign compounds, including most pharmaceuticals and many plant-derived molecules. The genes encoding CYP enzymes; CYP2D6, CYP2C9, CYP2C19, CYP3A4, and others, are among the most polymorphic in the human genome. That means there are dozens of known variants, and the version you carry has real consequences for pharmacokinetics: the absorption, distribution, metabolism, and elimination of a substance.

Broadly, people are classified as:

• Poor metabolisers, reduced or absent enzyme function; substances clear slowly.
• Intermediate metabolisers; somewhere between poor and normal.
• Normal (extensive) metabolisers, the reference population used in most clinical trials.
• Ultra-rapid metabolisers, multiple gene copies or highly active variants; substances clear very quickly.

These categories aren't evenly distributed. CYP2C19 poor metaboliser status, for instance, is present in roughly 2–3% of Northern European populations but in 12–23% of some East Asian populations, according to data published in the PharmGKB knowledge base. This is part of why ancestry-informed pharmacogenomic testing is an active area of research; population-level dosing guidance often reflects the demographics of the original trial cohort, which has historically skewed heavily toward Northern European participants.

I'll admit I got genuinely annoyed reading through some of the older trial methodology on this. The representation gaps are substantial, and they matter clinically.

Beyond metabolism: receptors, transporters, and targets

CYP enzymes are only part of the picture. Pharmacogenomics also covers genetic variation in:

• Drug transporters, proteins like P-glycoprotein (encoded by the ABCB1 gene) that shuttle substances across cell membranes, including the blood-brain barrier. Variants here affect how much of a compound reaches its target tissue.
• Drug targets themselves, receptors, ion channels, enzymes. If the receptor a drug is designed to bind has a structural variant, binding affinity can change significantly.
• Phase II conjugation enzymes; NAT2, TPMT, UGT1A1 and others that tag metabolites for excretion. Variants in UGT1A1, for example, are clinically significant in oncology pharmacology.

This is where it gets interesting for cannabinoid pharmacology specifically. The endocannabinoid system, the body's own network of lipid-signalling molecules and receptors, has measurable genetic variation. The CNR1 gene encodes the CB1 receptor, and single nucleotide polymorphisms (SNPs) in CNR1 have been associated with differences in receptor expression and binding characteristics in published neuroimaging studies. The CNR2 gene, encoding the CB2 receptor, similarly shows population-level variants. What this means for the pharmacology of plant-derived cannabinoids is still an open research question; the clinical data is early-stage, the studies are mostly small, and it would be wrong to draw firm conclusions.

The fatty acid amide hydrolase (FAAH) enzyme, responsible for breaking down endogenous compounds like anandamide, also carries a well-characterised variant. The FAAH C385A polymorphism reduces enzyme activity, meaning people who carry it have higher circulating anandamide levels at baseline. That's genuinely interesting mechanistically, even if the downstream implications for plant-derived cannabinoid pharmacology aren't fully characterised yet.

First-pass metabolism and bioavailability: where genetics enters early

Genetic variation hits pharmacokinetics at the very first hurdle. Oral administration of any compound means it passes through the gut wall and then the liver before reaching systemic circulation; what's called the first-pass metabolism effect. CYP3A4 and CYP3A5 activity in gut enterocytes and hepatocytes varies enormously between individuals, some of it genetically determined, some of it influenced by inducers and inhibitors in food, co-medications, and the microbiome.

The result is that bioavailability for many oral compounds spans a wide range across individuals, not because one person swallowed it differently, but because their metabolic machinery processed it differently. This is worth understanding before attributing a "non-response" purely to the formulation or the dose.

Pharmacogenomics in the Australian regulatory context

In Australia, pharmacogenomic testing is a recognised but still-evolving part of clinical practice. The TGA doesn't directly regulate genetic tests used for prescribing decisions the way it does therapeutic goods, though the Australian Commission on Safety and Quality in Health Care has published guidance on genomic testing in healthcare settings.

For Schedule 4 (prescription-only) and Schedule 8 (controlled) medicines, prescribers are already required to make individualised assessments under the Poisons Standard. Pharmacogenomic data is increasingly one input into that, particularly in psychiatry and oncology, where CYP2D6 and CYP2C19 genotyping is now part of some clinical guidelines. The Special Access Scheme and Authorised Prescriber pathways, used for substances like medicinal cannabis, operate within this same framework of individualised clinical decision-making.

Honestly, the conversation around pharmacogenomics in Australian clinical practice is more advanced in some specialities than others. Oncology has moved relatively quickly. General practice is catching up. The gap between what the science suggests is possible and what's routinely implemented in the clinic is still wide, and that's not a criticism of individual practitioners; it's a structural issue of translation time and reimbursement.

What this means for plant-based medicine research

For researchers and educators in plant-based medicine, pharmacogenomics is a lens that puts the "variable response" problem into sharper focus. A compound's pharmacology doesn't exist in isolation from the person taking it. The CYP profile, receptor genotype, transporter variants, and Phase II enzyme activity of any given individual shape the pharmacokinetic and pharmacodynamic picture substantially.

This complicates the interpretation of clinical data considerably. A trial that shows no significant effect at the population level might contain a pharmacogenomically-stratified subpopulation where the effect is real and measurable, but if the trial wasn't designed to capture that, the signal disappears into noise. Conversely, a signal in a homogeneous trial population might not generalise to a more diverse one.

I was running along the foreshore near Pyrmont last Tuesday, half-listening to a true-crime podcast and half-thinking about this exact problem, the way forensic toxicology has actually been ahead of clinical pharmacology in taking genetic variation seriously, because the stakes of misinterpretation in court are so visible. There's something instructive in that.

For cannabinoid pharmacology specifically, the interaction between CYP2C9 and CYP3A4 genotype and the metabolism of cannabidiol and tetrahydrocannabinol is an active area of study. A 2020 paper in the British Journal of Clinical Pharmacology identified CYP2C9 as a primary metabolic pathway for THC, with genotype-dependent variation in plasma concentrations. The authors were appropriately cautious about clinical extrapolation from their data; which is the right call at this stage of the field.

The short version

Pharmacogenomics is not a niche or future concern. It's a present reality that shapes how every person encounters every substance they take. The CYP enzyme variants are the most studied, but receptor polymorphisms, transporter variants, and Phase II enzyme genetics all contribute. Population representation in clinical trials has historically been poor, which means a lot of what gets called "standard" pharmacology is actually "pharmacology for a fairly specific demographic." The field is correcting for that, slowly.

Understanding genetic sources of variability doesn't mean personalised genomic testing is a prerequisite for informed pharmacology education, but it does mean the question "why do people respond differently?" has answers that go well beyond "they're more sensitive" or "they took it with food."

Sources

• Pharmacogenomics: The Right Drug for the Right Person, NCBI/NLM
• CYP2C9 genotype and THC metabolism; British Journal of Clinical Pharmacology (via PubMed Central)
• Medicinal Cannabis Guidance Documents, TGA
• Genomics in Healthcare, Australian Commission on Safety and Quality in Health Care

, Priya Nandakumar, Pharmacology writer; peptides & PK