Why aren’t my medications working?

You take a medication exactly as directed… and nothing happens. Or it helps a little, then fades. Or the side effects show up before any benefits do.

This is frustrating, and it is also common. Medication response varies from person to person for many reasons, including your diagnosis, dose, other medications, lifestyle, medical conditions, and sometimes your genetics.

Pharmacogenetics (PGx) is the study of how genetic variation influences medication response. It is not the only explanation, but for certain medications it can be a particularly informative one.

What “not working” can look like

People often mean one (or more) of the following:

• Lack of efficacy: the medication does not help the symptom or condition it is meant to treat.
• Too many side effects: the medication causes adverse effects that make it hard to keep taking.
• Unpredictable response: benefits and side effects vary week to week, even with the same dose.
• “Trial and error” cycling: multiple medications in the same category fail in similar ways.

Common reasons medications may not work, genetics included

Before focusing on genetics, it helps to zoom out. Medications can underperform for many non-genetic reasons, such as:

• The diagnosis or treatment target is not the right match for the medication.
• The dose is too low or too high, or it has not been taken long enough to assess.
• Drug-drug interactions change medication levels or block effects.
• Adherence challenges (timing, missed doses, food requirements) reduce effectiveness.
• Differences in absorption, liver function, kidney function, age, and other health conditions change how the body handles medications.
• Lifestyle factors like alcohol use, nicotine, caffeine intake, sleep, and diet can meaningfully alter outcomes.

Pharmacogenetics fits into this bigger picture. It is one factor that can be measured and, for some medications, used to guide safer or more effective prescribing.

Where pharmacogenetics fits

Many medications rely on proteins in your body to do their job, such as:

• Metabolizing enzymes that break down drugs (for example, CYP2D6, CYP2C19, CYP2C9).
• Activation pathways for prodrugs (medications that must be converted into an active form).
• Transporters that move drugs into or out of tissues (for example, SLCO1B1).
• Sensitivity pathways that influence how strongly a drug works (for example, VKORC1 for warfarin).

Genetic variants can change how well these proteins function. That can shift drug levels up or down and influence whether a medication is more likely to be ineffective, overly strong, or more prone to side effects.

Importantly, PGx effects are not “all or nothing.” They are typically probability shifts. A gene result can raise or lower risk, but it does not guarantee a specific outcome.

PGx examples of “it doesn’t work for me” (lack of efficacy)

Below are real-world examples where genetics can contribute to reduced benefit.

Example 1: Clopidogrel (Plavix) that does not provide expected protection

Clopidogrel is a prodrug, meaning it needs to be converted into an active form. A key step in that activation depends on CYP2C19. Some people have CYP2C19 variants that reduce activation. In those cases, clopidogrel can lead to reduced platelet inhibition and a higher risk of major cardiovascular or cerebrovascular events compared with people who activate it normally.

Clinical guidelines (such as CPIC) provide genotype-informed recommendations and often discuss considering an alternative antiplatelet agent in CYP2C19 intermediate or poor metabolizers for specific indications.

Example 2: Codeine or tramadol that barely touches pain

Codeine is also a prodrug. It relies heavily on CYP2D6 to convert codeine into morphine. If someone is a CYP2D6 poor metabolizer, they may convert very little codeine into morphine, leading to reduced analgesic effect even when taking the medication as directed.

CPIC-related guidance recommends considering alternate analgesics for CYP2D6 poor metabolizers (and also for ultrarapid metabolizers, for a different safety reason discussed below). Poor metabolizers especially may not convert the prodrug into its active form at all, so the best they can hope for from codeine might be a placebo effect!

Example 3: Proton pump inhibitors that don’t fully control reflux symptoms

Some common proton pump inhibitors (PPIs) such as omeprazole, lansoprazole, pantoprazole, and dexlansoprazole are significantly influenced by CYP2C19. People with CYP2C19 rapid or ultrarapid metabolism can clear some PPIs more quickly, which can reduce acid suppression and contribute to disappointing symptom control in certain contexts.

CPIC provides dosing guidance for PPIs based on CYP2C19 phenotype, including strategies like increasing starting dose and monitoring efficacy for ultrarapid metabolizers in particular situations.

Example 4: Certain antidepressants that feel ineffective, or never quite “kick in”

For SSRIs such as citalopram and escitalopram, CYP2C19 can influence blood levels. CPIC guidance discusses considering an antidepressant not predominantly metabolized by CYP2C19 for certain CYP2C19 phenotypes, and includes dose-alteration approaches when these medications are still the best clinical choice.

PGx examples of “the side effects are the problem”

Genetics can also shift the risk of adverse effects, sometimes substantially.

Example 5: Muscle pain or weakness on simvastatin

Statin-associated muscle symptoms can be influenced by variants in SLCO1B1, a transporter that affects statin movement into the liver. Certain SLCO1B1 phenotypes can increase systemic exposure to specific statins, which can raise the likelihood of musculoskeletal side effects in some people.

CPIC has a statin guideline covering SLCO1B1 (and related genes) with therapeutic recommendations to reduce risk of statin-associated musculoskeletal symptoms.

Example 6: Severe blood count suppression on thiopurines (azathioprine, mercaptopurine, thioguanine)

Thiopurines can cause myelosuppression, and this risk is strongly affected by TPMT and NUDT15 function. Decreased-function variants can lead to higher levels of active metabolites and a higher risk of serious toxicity at standard doses.

CPIC provides dosing recommendations based on TPMT and NUDT15 genotype, including substantially reduced starting doses for certain phenotypes to mitigate toxicity risk.

Example 7: Severe toxicity with fluorouracil or capecitabine

Fluoropyrimidines such as 5-fluorouracil (5-FU) and capecitabine can cause severe toxicity in people with reduced DPYD (DPD enzyme) activity.

CPIC recommends genotype-guided dose adjustments, including a 25% to 50% dose reduction from the standard dose for DPYD intermediate metabolizers with a specific captures activity score (noting evidence limitations and the need for careful clinical monitoring).

Example 8: Warfarin that causes bleeding, or does not anticoagulate as expected

Warfarin response can vary widely. Variants in CYP2C9 (metabolism) and VKORC1 (sensitivity) can shift dose needs. If someone is more sensitive than expected, a standard dose can increase bleeding risk. If someone needs a higher dose, underdosing can reduce effectiveness.

CPIC provides pharmacogenetics-guided warfarin dosing recommendations based on genotype.

Pharmacogenetics is only one reason, but it can be a powerful one

Even when PGx is relevant, it usually operates alongside other contributors: drug interactions, dosing strategy, adherence, organ function, and the underlying biology of the condition being treated.

The practical value of PGx is that it can sometimes help explain patterns like:

• “This drug never works for me.”
• “I get side effects at low doses.”
• “I’ve failed multiple medications in the same class.”
• “A prodrug seems ineffective, even when taken correctly.”

When a medication has strong evidence and clinical guidance, a genetic result can help a clinician choose a different medication, adjust dosing, or increase monitoring.

How can you find out if PGx might be relevant for you?

In many cases, PGx information can be determined with a genetic test and used long-term, because your genetics do not change over time. However, test quality and coverage vary:

• Some tests are clinical laboratory tests.
• Some results can come from direct-to-consumer genotyping data, which can be incomplete and may contain errors, so results should be used carefully and often confirmed clinically when decisions are high-stakes.

How Gene2Rx can help

Gene2Rx is designed to help people quickly explore whether known, actionable pharmacogenetic guidance might apply to medications they take, based on supported genetic data sources.

With Gene2Rx:

• You can upload genetic data from supported consumer services and receive a report that is ready in about a minute.
• The platform analyzes genetic variation with potential relevance to 103 medications, using guidance published by the FDA and CPIC where available.
• The report is intended to help you identify medications where you may have a typical or atypical response and to support a more informed discussion with your healthcare professional.

What to do next if you feel like your medications aren’t working

Here is a clinician-friendly, practical approach:

1. Write down what “not working” means for you: no benefit, partial benefit, side effects, or unpredictable response.
2. List all substances you take that can affect response: prescriptions, OTC meds, supplements, nicotine, caffeine, alcohol.
3. Ask your clinician to review interactions and dose strategy, including whether enough time has passed to assess benefit.
4. Consider pharmacogenetics if you see a pattern (nonresponse, side effects at low doses, repeated failures).
5. If you have supported genetic data, use Gene2Rx to identify medications with FDA or CPIC guidance and bring the relevant sections to your appointment for discussion as a starting point for introducing pharmacogenetics to the conversation.

Medication response should not be a guessing game. PGx will not explain every case, but when it applies, it can make the next step clearer, faster, and safer.