Why does the same medication work brilliantly for one patient and fail completely for another? This question haunts clinical practice daily. A dose of codeine that relieves pain in one person may produce dangerous respiratory depression in another—or have no effect at all.

The variability isn't random. Beneath these different responses lies a complex interplay of genetics, physiology, and environment that we're only beginning to understand systematically. Precision medicine promises to unravel this complexity, but the reality is messier than the marketing suggests.

Understanding why treatments work differently across individuals isn't just academically interesting—it's clinically essential. Every prescription carries implicit assumptions about how a patient will metabolize, distribute, and respond to a drug. Those assumptions are often wrong.

Your DNA writes much of the script for how you'll respond to medications. Pharmacogenomics—the study of how genetic variation affects drug response—has identified hundreds of gene variants that predictably alter treatment outcomes.

The cytochrome P450 enzyme family illustrates this powerfully. CYP2D6, which metabolizes roughly 25% of clinically used drugs, exists in multiple genetic variants. Some people carry variants producing ultrarapid metabolism—they clear drugs so quickly that standard doses never reach therapeutic levels. Others are poor metabolizers who accumulate dangerous drug concentrations from normal doses.

Codeine provides a stark example. The drug requires CYP2D6 to convert it to morphine for analgesic effect. Ultrarapid metabolizers produce morphine so efficiently that standard codeine doses can cause overdose symptoms. Poor metabolizers get almost no pain relief at all. The same prescription, radically different outcomes—determined by genetic variants the prescriber may not know about.

Beyond metabolizing enzymes, genetic variation affects drug transporters that move medications across cell membranes, and receptors where drugs exert their effects. Variants in the SLCO1B1 transporter gene dramatically increase statin-induced muscle toxicity risk. HLA gene variants predict severe hypersensitivity reactions to drugs like abacavir and carbamazepine. These aren't subtle effects—they're clinically transformative differences hiding in patients' genomes.

Takeaway

Genetic variation creates predictable, often dramatic differences in drug response—but most prescribing still happens without this information.

Genetics sets the foundation, but physiology builds upon it. Age, organ function, body composition, and disease states continuously modify how drugs behave in the body—sometimes in ways that dwarf genetic effects.

Age-related changes transform pharmacokinetics across the lifespan. Neonates lack mature metabolizing enzymes and have different body water distribution. Elderly patients often have reduced kidney function, decreased liver blood flow, and altered body fat ratios. A standard adult dose of digoxin may be perfectly appropriate for a 40-year-old but toxic for an 80-year-old with identical genetics.

Organ dysfunction creates obvious but frequently underestimated effects. Kidney impairment extends the half-life of renally cleared drugs, sometimes dramatically. Liver disease can either reduce metabolism (impairing drug clearance) or paradoxically increase it (through enzyme induction in early disease). Heart failure reduces organ perfusion, slowing drug distribution and elimination.

Body composition matters more than total weight. Lipophilic drugs like benzodiazepines distribute extensively into fat tissue. An obese patient and a lean patient of identical weight may have vastly different drug concentrations in their blood after the same dose. Disease states add another layer—inflammation alters drug-metabolizing enzyme expression, critical illness changes protein binding, and the medications patients already take create interaction effects that compound unpredictably.

Takeaway

The same patient will respond differently to the same drug at different points in their life—physiology is a moving target.

Precision medicine promises to integrate genetic, physiological, and clinical data to predict individual treatment responses. The reality, while advancing, remains humbling.

Pharmacogenomic testing has achieved clinical implementation for specific high-stakes scenarios. Pre-treatment HLA-B*5701 testing before abacavir prevents potentially fatal hypersensitivity reactions. CYP2C19 testing guides antiplatelet therapy selection after cardiac stenting. These represent genuine precision medicine successes—genetic tests that directly change clinical decisions and improve outcomes.

But most drug response prediction remains aspirational. For most medications, genetic variation explains only a fraction of response variability. A study of warfarin—one of the most intensively studied drugs—found that even sophisticated algorithms incorporating multiple genetic variants and clinical factors explained only about 50% of dose variability. Half the variability remained unexplained.

The complexity is daunting. Gene-gene interactions, gene-environment interactions, epigenetic modifications, microbiome effects, and factors we haven't yet identified all contribute to treatment response. Machine learning approaches show promise for integrating high-dimensional data, but they require massive datasets and often lack interpretability. Clinical judgment—observing how individual patients actually respond and adjusting accordingly—remains irreplaceable. The best prediction is often what happened last time you tried.

Takeaway

Precision medicine has achieved specific successes, but most treatment response prediction still relies on careful observation and iterative adjustment.

Treatment response variability isn't a failure of medicine—it's an inherent feature of human biology. We are not standardized organisms, and our responses to medications reflect that diversity.

Clinical practice must hold two truths simultaneously: population-level evidence guides initial decisions, but individual responses demand attention and adjustment. The patient in front of you is not the average patient from the clinical trial.

As pharmacogenomic testing becomes more accessible and our understanding of physiological modifiers deepens, we'll make better initial predictions. But the fundamental clinical skill—observing, adjusting, and responding to what actually happens—will remain central to good medical care.