Pharmacogenomics: Why the Same Medication Works Differently in Different People

A 52-year-old has a cardiac stent placed after a mild heart attack. The cardiologist prescribes clopidogrel — evidence-based, label-correct, standard of care. Ninety days later he’s back in the hospital with a stent thrombosis. The medication wasn’t wrong. The dose wasn’t wrong. What was wrong is that this patient is a CYP2C19 poor metabolizer — his body cannot convert clopidogrel into its active form. He was taking a drug that was, for his specific biology, functionally inert. And no one checked. This is what pharmacogenomics addresses. The evidence is clear: 9% of all reported drug side effects trace back to medications where a genetic variant increases risk — and three-quarters of those side effects trace to just three genes. The structural problem in modern prescribing is not bad drugs. It’s population-average dosing applied to individual genomes.

What Is Pharmacogenomics?

Pharmacogenomics is the study of how genetic variation affects an individual’s response to drugs. The core concept: you have a set of enzymes in your liver and intestine — primarily the cytochrome P450 (CYP) enzyme family — that metabolize most drugs before they reach their target. The genes encoding these enzymes contain common variants that can dramatically speed up, slow down, or completely eliminate drug metabolism. The result is that two patients receiving the same prescription at the same dose can experience radically different outcomes based entirely on their genomic makeup.

Based on these variants, individuals are classified into four metabolizer categories:

• Poor Metabolizers (PM): Little or no enzyme activity. Drugs build up to toxic levels. Standard doses become overdoses in this population.
• Intermediate Metabolizers (IM): Reduced enzyme activity. May require dose adjustments. Often prescribed standard doses that are subtherapeutic or produce disproportionate side effects.
• Normal Metabolizers (NM): Standard enzyme activity. Standard dosing guidelines apply correctly.
• Rapid / Ultra-Rapid Metabolizers (UM): Excessive enzyme activity. Drugs are cleared too quickly to be effective — or, in the case of prodrugs activated by the enzyme, dangerously high active metabolite levels result.

The key point is that the current prescribing system is built entirely around the Normal Metabolizer. Everyone else — which, depending on the gene, means the majority of the population — is receiving dosing designed for someone else’s biology.

The Evidence — Three Genes, Most of the Risk

A 2025 PLOS Medicine analysis of pharmacogenomic-related adverse drug reactions in a UK database found 9% of all reported ADRs trace to medications where a known genetic variant increases risk. Seventy-five percent of those PGx-mitigatable side effects come from just three genes: CYP2C19, CYP2D6, and SLCO1B1. Forty-seven percent come from psychiatric medications alone. A 2021 meta-analysis covering more than 336,000 subjects across 318 reports found a mean worldwide probability of 36.4% non-normal CYP2D6 phenotype and 61.9% non-normal CYP2C19. That is not a rare variant problem. That is a structural prescribing problem applied to the majority of the population.

CYP2D6 — The Antidepressant Gene

CYP2D6 metabolizes roughly 25% of all commonly prescribed drugs, including most SSRIs, tricyclic antidepressants, many antipsychotics, tamoxifen, codeine, tramadol, and several beta-blockers. CYP2D6 poor metabolizers — approximately 7–10% of European-ancestry adults — accumulate SSRIs to toxic levels at standard doses, get no analgesic effect from codeine (a prodrug requiring CYP2D6 activation to become morphine), and may receive inadequate tamoxifen protection for breast cancer because the active metabolite (endoxifen) is never generated at adequate concentrations.

CYP2D6 ultra-rapid metabolizers — 2–7% of European-ancestry adults, up to 29% in some East African populations — present the mirror-image problem. They convert codeine to morphine so rapidly that levels become dangerous — the basis for the FDA black box warning on codeine in pediatric patients and nursing mothers. For antidepressants, ultra-rapid metabolizers clear SSRIs so quickly that standard doses produce no therapeutic effect, and the clinical response is often interpreted as treatment-resistant depression rather than a metabolism problem.

CYP2C19 — The Clopidogrel Problem

Clopidogrel is a prodrug — it must be activated by CYP2C19 before it can exert antiplatelet effects. Approximately 2% of Caucasians and 15–22% of East Asians are CYP2C19 poor metabolizers and get inadequate antiplatelet protection from standard dosing, dramatically increasing stent thrombosis risk. This is not a theoretical concern — this is the clinical scenario that opens this article, repeated tens of thousands of times in hospitals every year. The Clinical Pharmacogenetics Implementation Consortium (CPIC) has issued formal guidelines recommending alternative antiplatelet therapy (prasugrel or ticagrelor) for CYP2C19 poor metabolizers receiving coronary stents. A pharmacogenomic test costs less than $100. The consequence of not performing it can be fatal.

CYP2C19 also governs the metabolism of proton pump inhibitors (PPIs), escitalopram, citalopram, voriconazole, and several other commonly prescribed drugs. A worldwide prevalence of 61.9% non-normal metabolizer status means the majority of patients prescribed these medications are receiving doses calibrated for a metabolizer phenotype that does not describe them.

SLCO1B1 — Statin Myopathy

SLCO1B1 is a transporter gene that moves statins from the bloodstream into liver cells. The 521C variant impairs this transport, leaving statin concentrations elevated in the bloodstream and muscle tissue. Carriers have substantially increased risk of statin-associated myopathy — muscle pain, weakness, and in severe cases rhabdomyolysis — particularly with simvastatin at standard doses. Statins are among the most prescribed drugs in the United States. Millions of patients are on doses that put them at unnecessary risk for a serious and preventable side effect because their SLCO1B1 status was never checked. The test is available. The clinical guidance is established. The prescribing system has simply not made it routine.

Psychiatric Medication Impact

A 2026 Pharmacogenomics Journal analysis of 59,973 hospital admissions found unplanned admissions had significantly higher prevalence of medications with PGx guidelines — 84% — compared to planned admissions at 64%. For patients managing depression, bipolar disorder, anxiety, or ADHD, trial-and-error prescribing is the current standard: 6–8 weeks per medication attempt, often multiple failed trials before finding something that works. Pharmacogenomics does not eliminate all uncertainty from psychiatric prescribing — receptor pharmacology, blood-brain barrier penetration, and other factors matter alongside metabolism — but it removes the preventable failures. Knowing that a patient is a CYP2D6 poor metabolizer before starting an SSRI is the difference between rational prescribing and expensive, harmful guesswork.

The Gap in Standard Care

Despite decades of accumulating evidence — CPIC guidelines for dozens of drug-gene pairs, FDA pharmacogenomic labeling on over 200 drugs, and meta-analyses covering hundreds of thousands of patients — routine pharmacogenomic testing remains non-standard in most clinical practice. Three structural reasons explain the gap:

1. Insurance coverage is inconsistent. Many payers still classify PGx testing as “investigational” for non-cancer applications. Oncology has moved rapidly toward genotype-guided prescribing partly because reimbursement followed. Cardiovascular and psychiatric PGx has lagged. The result is that patients who most need this information are least likely to have it paid for.
2. Medical education has not caught up. Most physicians received minimal genomics training in medical school — a curriculum developed before genome sequencing was clinically accessible at scale. Prescribing habits reflect what was taught, not what the current evidence supports. Physicians are not failing their patients because of malice; they are operating within a system that has not updated its defaults.
3. The “it works for most people” fallacy. Standard dosing guidelines are designed for the average patient. The problem is that “most people” is not a useful category when the consequences of being outside the average are a stent thrombosis, a rhabdomyolysis, or three years of failed antidepressant trials. The patient in front of you is not a population statistic. Their response to medication is determined by their genome, and that genome can be known.

The result of these three structural failures is that patients are routinely prescribed medications at doses that are, for their specific genome, either dangerous or useless. The information to do better exists. The testing to obtain it is affordable. The clinical framework to interpret it — CPIC guidelines, published drug-gene pair evidence — is established. What is missing is the systematic commitment to making genomic-guided prescribing the standard rather than the exception. A consultation at Pravida Health is a starting point for patients who want prescribing decisions made from their actual biology.

How We Use This at Pravida Health

At Pravida, pharmacogenomic analysis is extracted directly from your whole genome sequencing — no additional test, no additional blood draw. Once your genome is sequenced, your complete pharmacogenomic profile is available permanently and informs every future prescribing decision. We integrate PGx data into clinical care in four specific ways:

1. A personalized medication report. For every CPIC-guideline drug you currently take or might be prescribed, we document your predicted metabolizer phenotype and the dose adjustments or drug substitutions indicated by your genotype. This report travels with you and is updated as CPIC guidelines evolve. It is the foundation of a prescribing relationship built on your biology rather than population averages.
2. A “before you prescribe” reference. Any time you are being considered for a new medication — by us or by another provider — we reference your PGx profile to flag potential issues before the first dose is taken. The goal is to prevent adverse reactions and treatment failures prospectively rather than investigating them after they occur. Book a consultation to discuss how PGx integration works in practice.
3. Statin guidance grounded in SLCO1B1 status. Given the prevalence of SLCO1B1 variants and the ubiquity of statin prescribing, every patient’s lipid management plan is informed by their SLCO1B1 genotype. This does not mean avoiding statins — it means selecting the right statin at the right dose for the specific transporter phenotype of that patient. Rosuvastatin and pravastatin have substantially lower SLCO1B1-mediated myopathy risk than simvastatin; for a 521C carrier, this distinction matters clinically.
4. Psychiatric medication clarity. For patients who have struggled to find effective psychiatric medications — cycling through SSRIs, antipsychotics, mood stabilizers with inconsistent or absent response — PGx data frequently explains years of treatment failures and guides the next step rationally. A CYP2D6 poor metabolizer on a standard SSRI dose is not treatment-resistant. They are over-medicated at the metabolic level. This is a solvable problem. Discuss your medication history and whether PGx analysis changes the clinical picture.

What You Can Do Today

1. Review your current medication list. Look up whether any current medications are CYP2D6, CYP2C19, CYP2C9, or SLCO1B1 substrates — PharmGKB.org is publicly accessible and comprehensive. If they are, and you do not know your genotype for the relevant enzyme, you are prescribing blind. Contact us to begin the conversation about your current medication list and PGx relevance.
2. Ask about PGx testing before starting a new psychiatric or cardiovascular medication. These are the two categories where PGx-related harm is most documented and most preventable. Before a first antidepressant trial or a clopidogrel prescription, the question “have we checked my CYP2C19 or CYP2D6 status?” should be routine. In most clinical settings, it is not asked. Ask it yourself.
3. Understand that “it didn’t work for me” often has a biological explanation. Unusual responses to medications — inadequate effect at standard doses, unexpected side effects, or a pattern of failed trials across a drug class — are signals worth investigating genetically. The alternative explanation — that the patient is simply difficult to treat — is frequently wrong and always less actionable.
4. Request a PGx consultation if you are on five or more medications. Polypharmacy combined with uncharacterized genetic metabolism creates compounding risk. Each drug-gene interaction is not independent — phenoconversion from one drug can alter the effective phenotype for another. Understanding your baseline genomic metabolism is the prerequisite for managing this complexity safely. Schedule a consultation to review your complete medication picture.
5. Get your complete genome sequenced. A comprehensive whole genome sequence includes your full pharmacogenomic profile — permanently available for every future prescription decision, every new diagnosis, every surgical anesthesia plan. This is not a test you repeat. It is a reference that lives with you for the rest of your medical life. At Pravida, WGS is the foundation of our precision medicine program, and PGx is one of the most immediately actionable layers it provides.

Frequently Asked Questions