Genomics

Pharmacogenomics: Why Your Medications Might Not Be Working

Dr. RP, MD — Board-Certified, Emergency Medicine & Critical Care Medicine — Founder, Analog Precision Medicine

One of the things I took away from medical school in Dublin — and from pharmacology specifically — was a deceptively simple observation that rarely comes up in routine clinical practice: the same drug, at the same dose, does fundamentally different things in different people. Not because of weight or age or kidney function, though those matter too, but because of genetics. The enzymes that metabolize most medications vary substantially between individuals, and that variation produces plasma drug levels that can differ by tenfold or more between patients who appear, on the surface, to be similar.

We learned this. Most of us then practiced medicine for years without doing much with it — prescribing based on population averages, adjusting after something went wrong, rarely testing in advance to find out which metabolic category a particular patient occupied.

That gap is closing. The evidence has crossed a clinical threshold, the cost of testing has dropped substantially, and the case for preemptive pharmacogenomic testing — done once, available for every prescribing decision that follows — is now genuinely strong for a defined set of drug-gene interactions.

The Enzymes That Control Drug Levels

Most medications are metabolized primarily by a family of liver enzymes called the cytochrome P450 (CYP) system. The genes encoding these enzymes are polymorphic — they vary between individuals in ways that produce four broad phenotypic categories: poor metabolizers, intermediate metabolizers, normal (extensive) metabolizers, and ultrarapid metabolizers.

—

CYP2D6 is one of the most clinically important. It metabolizes a wide range of commonly prescribed drugs: antidepressants (fluoxetine, paroxetine, venlafaxine, nortriptyline), antipsychotics, opioids including codeine and tramadol, and beta-blockers including metoprolol. A CYP2D6 poor metabolizer on a standard dose of a CYP2D6-substrate antidepressant may accumulate the drug to toxic levels. An ultrarapid metabolizer may get essentially no therapeutic benefit because the drug clears before it can act.

—

CYP2C19 is equally relevant. It converts clopidogrel — the antiplatelet drug used after coronary stenting — from an inactive prodrug to its active metabolite. CYP2C19 poor metabolizers generate insufficient active drug to inhibit platelet aggregation meaningfully. A patient who just received a coronary stent and is prescribed clopidogrel as a CYP2C19 poor metabolizer is, in pharmacological terms, receiving no effective antiplatelet therapy. The downstream risk — stent thrombosis, recurrent MI — is serious, and it is preventable with a test that could have been done beforehand (Mega et al., N Engl J Med, 2009).[4]

—

SLCO1B1 is a transporter gene affecting statin entry into liver cells. The *5 variant is associated with markedly elevated myopathy risk on simvastatin. Carriers of two copies face roughly a 17-fold increased risk of statin-induced muscle injury at high doses (SEARCH Collaborative Group, N Engl J Med, 2008).[5] In practice, this variant often surfaces only when a patient presents with muscle pain severe enough to prompt a workup — at which point the damage has already occurred.

How Common Is Actionable Variation?

If clinically significant variants were rare, routine testing would be hard to justify. They are not rare.

A 2024 analysis in Molecular Psychiatry examining pharmacogenomic test results from 15,000 psychiatric patients found that 65% had potentially actionable CYP2D6 or CYP2C19 phenotypes — meaning their genotype indicated a prescribing consideration for at least one commonly used medication (Goldfarb et al., Mol Psychiatry, 2024).[6]

“The PREPARE trial found that genotype-guided prescribing reduced clinically relevant adverse drug reactions by 30% compared to standard care.”

The PREPARE trial — the first large-scale, multinational RCT of preemptive pharmacogenomic testing — enrolled 6,944 patients and found that genotype-guided prescribing reduced clinically relevant adverse drug reactions by 30% compared to standard care (Swen et al., Lancet, 2023).[1] This was a rigorous multi-country trial covering 39 medications. The reduction in harm was real and clinically meaningful.

For depression — where 30–50% of patients fail to respond adequately to the first antidepressant prescribed — pharmacogenomic guidance shows consistent benefit across multiple trials and meta-analyses. A 2023 systematic review of 12 RCTs found that CYP2D6/CYP2C19-guided antidepressant prescribing significantly improved remission rates compared to standard care (Arnone et al., Neurosci Biobehav Rev, 2023).[2]

Where the Evidence Is Strong vs. Still Developing

The Clinical Pharmacogenetics Implementation Consortium (CPIC) provides evidence-graded prescribing recommendations for drug-gene pairs where the clinical evidence is sufficient. The strongest cases include:

—

CYP2C19 and clopidogrel (antiplatelet efficacy after stenting)

—

CYP2D6 and codeine/tramadol (toxicity risk in ultrarapid metabolizers, inadequate analgesia in poor metabolizers)

—

SLCO1B1 and simvastatin (myopathy risk)

—

CYP2D6/CYP2C19 and antidepressant selection

—

TPMT/NUDT15 and thiopurine immunosuppressants

—

HLA-B*57:01 and abacavir (hypersensitivity in HIV therapy)

The FDA's pharmacogenomic biomarker table now covers over 300 drug labels. These are not fringe science positions. They are established enough to have shaped FDA-required label language.

Where the evidence is still developing: broader CYP3A4/3A5 variants, serotonin transporter and receptor variants in psychiatric prescribing, and several other gene-drug interactions are plausible and show preliminary signal but haven't yet reached the level of actionable clinical guidance.

The Case for Preemptive Testing

Reactive pharmacogenomic testing — done after a side effect or treatment failure — has an obvious limitation: the event has already happened. A one-time preemptive panel covering the clinically established variants creates a permanent reference that is available at every future prescribing decision. Because genetics don't change, the test is done once.

The cost of comprehensive pharmacogenomic panels has dropped significantly — currently in the $250–$500 range for panels covering the major clinically actionable variants, depending on the lab and what's included. Some insurance plans cover specific tests in specific clinical contexts, though coverage remains inconsistent.

The strongest indication for preemptive testing: patients on or likely to be prescribed drugs in the high-evidence categories above; patients with prior unexplained medication side effects or treatment failures; and patients on complex polypharmacy, where the probability of having at least one actionable drug-gene interaction already in the regimen is substantial.

The counterargument — that testing isn't necessary until a problem develops — implicitly accepts that the first evidence of a pharmacogenomic problem will be an adverse event. In the case of clopidogrel and stent thrombosis, that logic is dangerous. In the case of years of antidepressant trials in someone who is a CYP2D6 poor metabolizer, it is avoidable harm.

The information exists. The tests are available. The evidence supports their use. The main barriers are reimbursement and clinical implementation infrastructure — not the science.

References

1.Swen JJ et al. A 12-gene pharmacogenetic panel to prevent adverse drug reactions (PREPARE). Lancet. 2023;401(10374):347–356.
2.Arnone D et al. Effectiveness of pharmacogenomic tests for guiding treatment of depressive disorders. Neurosci Biobehav Rev. 2023;144:104965.
3.Brown LC et al. Pharmacogenomic testing and depressive symptom remission. Clin Pharmacol Ther. 2022;112(6):1303–1317.
4.Mega JL et al. Cytochrome P-450 polymorphisms and response to clopidogrel. N Engl J Med. 2009;360(4):354–362.
5.SEARCH Collaborative Group. SLCO1B1 variants and statin-induced myopathy. N Engl J Med. 2008;359(8):789–799.
6.Goldfarb NM et al. Pharmacogenomic insights in psychiatric care. Mol Psychiatry. 2024;29:3495–3502.

Dr. RP, MD is dual board-certified in Emergency Medicine and Critical Care Medicine and is the founder of Analog Precision Medicine, a precision medicine practice in Southern California. This article is for educational purposes only and does not constitute medical advice or establish a physician-patient relationship.