CBD Drug Interactions: List of Drug Groups and Strength of Evidence

Cannabidiol enters the cart as a supplement but acts in the body as a substance altering the metabolism rate of other drugs. In a study involving eighteen healthy adults, a CBD-predominant extract increased exposure to omeprazole by 207% and to losartan by 77%. This is not a test-tube result or assumption but a measurement in humans. This guide shows which liver isoenzymes cannabidiol actually inhibits, with which drugs this matters, and what is still unknown. You will also find a statement repeated by the European Food Safety Authority since 2026: cannabidiol safety cannot be established in people taking medications. It changes the starting point of the entire discussion about combining them.

KEY INFORMATION
• In eighteen healthy adults, a CBD-predominant extract most strongly inhibited CYP2C19, then CYP2C9, followed by CYP3A and CYP1A2 (Bansal et al., Clinical Pharmacology and Therapeutics, 2023).
• The same extract did not change CYP2D6 activity at all, so the often-repeated warning about this pathway is not supported by measurement.
• Exposure to omeprazole increased by 207%, losartan by 77%, midazolam by 56%, caffeine by 39%.
• In children combining cannabidiol with clobazam, norclobazam concentration increased on average by 500% (Geffrey et al., Epilepsia, 2015).
• The EFSA panel states that cannabidiol safety cannot be established in people taking medications (EFSA Panel on Nutrition, 2026).

Which drugs does CBD interact with?

The best documented interactions are with two antiepileptic drugs (clobazam and valproate), alongside tacrolimus and warfarin. In measurements in healthy volunteers, cannabidiol also increased exposure to several drugs, most strongly to omeprazole by 207% (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). Statins and psychotropic drugs appear on the list by mechanism, benzodiazepines by measurement for midazolam, and CYP2D6 activity did not change in this measurement.

The table below organizes what is described in detail in the following sections. The third column indicates the basis of the warning: human measurement, case report, or inference from mechanism. This difference is more important here than the drug name itself, as it determines how strongly the warning can be formulated.

How does CBD affect cytochrome P450?

Cannabidiol is both a substrate and an inhibitor of liver enzymes from the cytochrome P450 family. It is metabolized by CYP2C19 and CYP3A4, while simultaneously slowing several drug-metabolizing pathways (Stout and Cimino, Drug Metabolism Reviews, 2014). The effect is simple: the drug remains in the blood longer than its dosing predicts.

Cytochrome P450 is a family of enzymes responsible for biotransformation of foreign substances. Most oral drugs undergo first-pass metabolism, where these enzymes oxidize or demethylate molecules preparing them for excretion via bile or urine. Blocking one enzyme does not change the number of pills but changes how much substance actually circulates in the body.

A regulatory review describes cannabidiol as involved in interactions both as a perpetrator and as a victim of other substances (Brown and Winterstein, Journal of Clinical Medicine, 2019). The same work points out the importance of P-glycoprotein, a transport protein removing drugs from intestinal and kidney cells. Both mechanisms act simultaneously with tacrolimus.

The review authors note that adverse effects occurred in nearly half of cannabidiol users, with dose dependence. The most common were increased aminotransferase activity, drowsiness, and sleep disturbances. These symptoms are easily attributed to the underlying disease or fatigue, not an over-the-counter product.

Which isoenzymes does CBD inhibit, and which does it not inhibit at all?

The answer was measured in humans, not inferred from test tubes. Eighteen healthy adults received alternately a cannabis extract cookie and a set of probe drugs for five metabolic pathways (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). The CBD-predominant extract inhibited four of them. The fifth, CYP2D6, was not inhibited at all.

The inhibition strength order was: strongest CYP2C19, then CYP2C9, then CYP3A, weakest CYP1A2. An extract containing only tetrahydrocannabinol, without cannabidiol, did not inhibit any of these pathways. This distinction is practical, showing that cannabidiol itself is responsible, not the cannabis extract as such.

The CYP2D6 result is worth remembering separately, as many cannabidiol texts claim the opposite. In this study, CYP2D6 activity remained unchanged, so drugs dependent solely on this pathway are not the main concern. This applies only to pharmacokinetics, as overlapping central nervous system effects remain a separate issue.

Why is the 2014 risk assessment insufficient today?

A 2014 systematic review concluded that cannabinoid enzyme inhibition and induction generally carry low risk of clinically significant interactions but immediately noted a lack of human data (Stout and Cimino, Drug Metabolism Reviews, 2014). This sentence circulates as evidence of safety, though it contains its own limitation.

Over the next decade, the picture changed due to new measurements, not theories. Registration studies of purified cannabidiol in children with drug-resistant epilepsy, case reports from transplantology, and a probe study of five metabolic pathways in healthy volunteers emerged. Each measured something unavailable in the 2014 review.

By 2017, a cannabidiol safety review directly called for more research on its effects on liver enzymes and drug transporters (Iffland and Grotenhermen, Cannabis and Cannabinoid Research, 2017). The authors noted cannabidiol’s adverse effect profile compares favorably to drugs used for the same indications but left co-administration questions open.

A 2021 review compiled described interactions in one place, listing antiepileptics, antidepressants, opioid analgesics, and unexpected substances like paracetamol and alcohol (Balachandran et al., Journal of General Internal Medicine, 2021). The authors cautiously concluded the list should be known, and its clinical significance studied in controlled trials.

Do other drugs change cannabidiol’s action?

Yes, and this direction is often overlooked. Cannabidiol is involved in interactions not only as a perpetrator but also as a victim of other drugs (Brown and Winterstein, Journal of Clinical Medicine, 2019). Since it is metabolized by CYP3A4 and CYP2C19, anything altering these enzymes also changes its concentration.

A systematic review cites a pharmacokinetic study with ketoconazole, a strong CYP3A4 inhibitor, and cannabis extracts used on oral mucosa (Stout and Cimino, Drug Metabolism Reviews, 2014). The authors consider this confirmation that CYP3A4 is an important metabolic route for both THC and cannabidiol.

The same review notes something opposite with omeprazole. Lack of interaction between cannabidiol extract and omeprazole suggests CYP2C19’s role in cannabidiol metabolism is less significant. The enzyme most strongly inhibited by cannabidiol may not be the one removing it.

Practically, this means two situations. A CYP3A4 inhibitor drug raises cannabidiol concentration, increasing adverse effects risk despite unchanged product labeling. A CYP3A4 inducer, like some antiepileptics and antituberculars, lowers its concentration, so the supplement stops working as intended.

What is a narrow therapeutic window and why does it determine risk?

A narrow therapeutic window means the difference between effective and toxic concentration is small. For such drugs, even moderate exposure changes translate into symptoms, not just lab results. Thus, the same metabolic change may be insignificant for one drug and dangerous for another.

This group includes warfarin, tacrolimus, cyclosporine, clobazam, phenytoin, and lithium. Each has its own monitoring protocol: INR for warfarin, blood drug levels for tacrolimus and cyclosporine, serum levels for phenytoin, and lithium levels for lithium therapy. These protocols exist independently of cannabidiol and must include it.

For comparison, proton pump inhibitors have a wide safety margin, so the 207% exposure increase for omeprazole is large but rarely symptomatic (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). The same mechanism applied to clobazam causes drowsiness and balance issues due to a much narrower margin.

The practical consequence is one: the list of drugs warranting caution before adding cannabidiol depends not on drug popularity but on margin width. Patients cannot assess this themselves, as it requires knowledge of the entire treatment scheme and current test results.

What is known about combining CBD with warfarin?

Only what a single case report and mechanism indicate. Authors describe interaction between warfarin and cannabidiol, discuss metabolism of both, and recommend INR monitoring after adding cannabinoids (Grayson et al., Epilepsy and Behavior Case Reports, 2018). No controlled study has tested this combination.

The mechanism aligns with healthy volunteer measurements. The more potent warfarin enantiomer, S-warfarin, is metabolized by CYP2C9. This pathway showed a 77% increase in losartan exposure with cannabidiol (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). Slowed S-warfarin metabolism prolongs anticoagulant effect and raises INR.

Newer anticoagulants are not exempt. Rivaroxaban and apixaban are substrates of CYP3A4 and P-glycoprotein, both targeted by cannabidiol (Brown and Winterstein, Journal of Clinical Medicine, 2019). The difference is no routine testing reveals changes, so symptom observation remains.

Low molecular weight heparins and fondaparinux are not metabolized by cytochrome P450, so pharmacokinetic interaction with cannabidiol is unlikely. However, patients on any anticoagulant should inform their doctor about supplementation, as bleeding symptoms are always assessed in the treatment context.

How does CBD change clobazam concentration?

The best documented cannabidiol interaction concerns clobazam. In thirteen children with drug-resistant epilepsy taking both drugs, clobazam concentration increased by 60% after four weeks, and norclobazam, its active metabolite, by 500% (Geffrey et al., Epilepsia, 2015). Variability was large but direction clear.

The mechanism explains this asymmetry. Clobazam is demethylated to norclobazam, which is further cleared by CYP2C19. Since cannabidiol most strongly inhibits this pathway, the metabolite accumulates faster than the parent drug. Clinically, this causes drowsiness and impaired coordination, resembling benzodiazepine overdose.

The research team’s practice shows the response. Clobazam dose was reduced in ten of thirteen patients, adverse effects occurred in the same number and resolved after dose reduction. Meanwhile, nine participants had seizure frequency reduced by over half, so the authors do not discourage combination but demand concentration monitoring.

The takeaway for over-the-counter users differs from neurologists. Patients receiving purified cannabidiol as treatment have controlled concentrations and dose adjustment options. Those adding hemp oil to antiepileptic therapy on their own have neither. More on therapy is in the article about CBD in drug-resistant epilepsy.

Does CBD burden the liver with valproate?

The liver signal is real and visible in registration studies. In a Lennox-Gastaut syndrome trial with 225 participants aged 2 to 55, elevated aminotransferase activity occurred in 14 cannabidiol users, i.e., 9% (Devinsky et al., New England Journal of Medicine, 2018). Adverse effects also included drowsiness and appetite loss.

An earlier Dravet syndrome trial showed a similar picture. Abnormal liver tests were more frequent in the cannabidiol group than placebo (Devinsky et al., New England Journal of Medicine, 2017). Both used medicinal doses calculated per patient weight, not supplement content.

A regulatory data review lists increased aminotransferase activity as the most common cannabidiol adverse effect (Brown and Winterstein, Journal of Clinical Medicine, 2019). The EFSA panel adds animal studies show consistent liver toxicity, and human hepatotoxic potential appears especially with concomitant drugs (EFSA Panel on Nutrition, 2026).

Valproate itself burdens the liver, so combining both factors acts in the same direction. The frequency of liver tests is decided by the attending physician. Yellowing of whites, dark urine, or right upper quadrant pain warrant immediate contact, not waiting for scheduled check.

What did the tacrolimus case in a transplant patient show?

It was the first documented clinically significant interaction between purified cannabidiol and tacrolimus. A clinical trial participant with epilepsy taking tacrolimus had about a threefold increase in tacrolimus concentration normalized to dose when receiving 2000-2900 mg cannabidiol daily (Leino et al., American Journal of Transplantation, 2019).

This amount must be emphasized, as it is medicinal, many times higher than cannabidiol content in typical supplements. The case proves the mechanism works and can be dangerous but does not indicate what happens at over-the-counter doses. That has not been studied.

The mechanism is dual. Tacrolimus is metabolized by CYP3A4 and transported by P-glycoprotein, both targeted by cannabidiol (Brown and Winterstein, Journal of Clinical Medicine, 2019). Blocking one raises concentration; blocking both raises it faster than enzyme inhibition alone suggests.

High tacrolimus concentration effects include nephrotoxicity, tremors, headaches, and hypertension. Cyclosporine and sirolimus share similar pathways. Transplant patients take immunosuppressants lifelong, so any CYP3A4-modulating supplement addition requires transplant specialist consultation, not forum opinions.

Does CBD affect statins and other cholesterol-lowering drugs?

Atorvastatin and simvastatin are mainly metabolized by CYP3A4, which showed a 56% increase in midazolam exposure with cannabidiol (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). No study directly measured statin concentration changes with cannabidiol; this is inference from mechanism.

Why does this matter? Higher statin concentration raises myopathy risk, i.e., muscle pain and weakness, and in severe cases rhabdomyolysis with kidney damage. Symptoms are nonspecific and easily mistaken for fatigue or joint issues in older adults. This is why the topic recurs in discussions about CBD oil safety in seniors.

Not all statins behave the same. Pravastatin and rosuvastatin depend less on CYP3A4, so theoretically less affected. Switching drugs is a cardiology decision, based on cardiovascular risk profile, not metabolism chemistry alone.

Other cholesterol-lowering drugs have different elimination routes. Ezetimibe undergoes mainly glucuronidation, not cytochrome P450. PCSK9 inhibitors are subcutaneous proteins and do not pass through this system. Data on fibrates are too sparse to conclude, so caution remains standard.

Does CBD enhance benzodiazepine and opioid effects?

Two independent mechanisms act here, easily confused. The first is pharmacokinetic: alprazolam, midazolam, and triazolam are CYP3A4 substrates, and diazepam also CYP2C19, the pathway most strongly inhibited by cannabidiol (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). The second is pharmacodynamic, involving additive central nervous system depression.

A regulatory data review lists drowsiness and sleep disturbances among the most common cannabidiol adverse effects (Brown and Winterstein, Journal of Clinical Medicine, 2019). Adding these to a sedative drug produces effects not visible in lab results but that realistically impair driving ability and increase fall risk in older adults.

Not all benzodiazepines behave the same. Lorazepam and oxazepam are eliminated by glucuronidation, not cytochrome P450, so pharmacokinetic pathway does not apply. Sedation risk remains due to additive effects, not metabolism changes.

For opioids, the picture is similar but less studied. A cannabidiol interaction review lists opioid analgesics among classes with described mutual influence (Balachandran et al., Journal of General Internal Medicine, 2021). Tramadol is a CYP2D6 substrate, and cannabidiol did not inhibit this pathway, so main concern is additive sedation, not concentration change.

What about antidepressants and other psychotropic drugs?

Distinguishing pathways changes the whole assessment. Citalopram and escitalopram are mainly metabolized by CYP2C19, the enzyme most strongly inhibited by cannabidiol (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). Fluoxetine and paroxetine depend on CYP2D6, whose activity remained unchanged in the same study.

This reverses a popular simplification. Many cannabidiol texts claim it inhibits CYP2D6 and thus raises concentrations of antidepressants dependent on this pathway. Healthy volunteer measurements do not confirm this. Attention shifts to CYP2C19-metabolized drugs, not the entire class.

A 2021 interaction review lists antidepressants among classes with described cannabidiol mutual influence but notes most reports require confirmation in controlled studies (Balachandran et al., Journal of General Internal Medicine, 2021). The direction is known, scale unknown. The topic is expanded in the article about cannabis and antidepressants.

Tricyclic antidepressants, now more often used for neuropathic pain than depression, stand apart. They have a narrow therapeutic window and QT interval prolongation risk, with metabolism involving several pathways including CYP2C19. Assessment here is a physician’s task, including ECG monitoring.

Data on antipsychotics are even scarcer. A cannabidiol adverse effect review lists them among CYP3A4-metabolized drugs with increased side effect risk and among drugs associated with liver damage, while antidepressants with cannabidiol may increase drowsiness (Brown and Winterstein, Journal of Clinical Medicine, 2019). Cannabidiol also inhibited CYP1A2, the weakest of four measured pathways: caffeine exposure increased by 39% (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). If you take a psychotropic drug, ask your psychiatrist which pathway metabolizes it.

The works underlying this text do not measure cannabidiol interaction with lithium, so no warning or assurance is given. Lithium has a narrow therapeutic window, so lithium levels are monitored by a doctor anyway.

Can antihypertensive drugs be combined with CBD?

One of these drugs was a probe in the study, so the answer is particularly concrete. Losartan was used to measure CYP2C9 activity, and its exposure increased by 77% after a cannabidiol-predominant extract (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). This drug is used chronically by millions, not an exotic hospital preparation.

Calcium channel blockers - amlodipine, diltiazem, and verapamil - are mainly metabolized by CYP3A4. This pathway responded with a 56% increase in midazolam exposure in the same study. Elevated concentration may cause excessive blood pressure drop, ankle swelling, or dizziness on standing.

Beta-blockers behave variably, and here the CYP2D6 result helps. Metoprolol and carvedilol depend largely on this pathway, which cannabidiol did not inhibit, so pharmacokinetic risk is less than older reports suggest. Atenolol is almost entirely renally excreted, so outside this mechanism.

Another aspect is cannabidiol’s direct cardiovascular effects. The EFSA panel notes liver signals and hormonal changes but makes no blood pressure recommendations (EFSA Panel on Nutrition, 2026). Patients starting supplementation with treated hypertension should measure blood pressure regularly and show results to their doctor.

Which drugs are probably unaffected by cannabidiol?

The list is shorter than readers might wish and is based on drug elimination routes. Drugs excreted unchanged by kidneys and those eliminated by glucuronidation lie outside cannabidiol’s main mechanism. This is not proof of safety but lack of known interaction points.

This group includes levetiracetam, largely renally excreted; lorazepam and oxazepam eliminated by glucuronidation; mycophenolate mofetil metabolized by glucuronyltransferases; and azathioprine dependent on different enzymes. Low molecular weight heparins also do not pass through cytochrome P450.

This reasoning’s limitation is important. The probe study covered five metabolic pathways, but the body has many more (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). Lack of measurement is not the same as measurement showing no effect. A 2021 review lists paracetamol among interacting substances, unexpected there (Balachandran et al., Journal of General Internal Medicine, 2021).

Variability among products adds complexity. A cannabidiol pharmacology review notes over-the-counter products often differ in composition and bioavailability, and even rigorous studies used various sources, complicating comparisons (Britch et al., Psychopharmacology, 2021). Labeling does not replace laboratory analysis.

Does the product form change interaction risk?

It does, more than label numbers suggest. A systematic human pharmacokinetic review shows area under the curve and maximum concentration increase with dose but are reached faster by inhalation than oral administration (Millar et al., Frontiers in Pharmacology, 2018). Concentration rise rate determines when enzyme inhibition peaks.

The same review notes something directly relevant to daily life. Maximum cannabidiol concentration is higher after a meal and with fat-based preparations. The same declared dose taken fasting or after a fatty meal yields different exposure, though the user records the same entry.

Half-life variability is equally large. After mucosal administration, it ranged from 1.4 to 10.9 hours; after chronic oral dosing, 2 to 5 days; and after smoking, about 31 hours. The authors emphasize absolute bioavailability was measured only for inhalation, at 31%.

No such measurement exists for other routes, despite intravenous forms. The inconvenient but honest conclusion: oral product bioavailability percentages in descriptions do not come from human studies. Without knowing how much substance reaches blood, enzyme inhibition strength in a person cannot be predicted.

What does EFSA say about CBD safety in people taking medications?

The 2026 position is brief and firm. The EFSA panel states cannabidiol safety cannot be established in three groups: under 25 years old, pregnant and breastfeeding women, and people taking medications (EFSA Panel on Nutrition, 2026). The last group is exactly the reader of this guide.

The panel gives a temporary safe dose of 0.0275 mg per kilogram body weight per day, about 2 mg for a 70 kg person. This value was derived by benchmark dose method with a 400 uncertainty factor. It is a safety ceiling for the general population, not a dosing recommendation.

Reservations are as important as the number. The estimate applies only to supplements with at least 98% cannabidiol purity, no nanoparticles, safe production, and excluded genotoxicity. The panel notes 2022 knowledge gaps remain, and newer studies have methodological limitations.

Separate this from legal status. Cannabidiol is not on Polish controlled substances lists, so its trade is not prohibited. A sanitary notification filed by the producer is not a novel food authorization and does not guarantee ingredient safety. These two things, often conflated in advertising, are distinct.

What else affects cytochrome P450 besides drugs?

Liver enzymes respond not only to drugs. Caffeine was a CYP1A2 activity measure in the probe study, and its exposure increased by 39% after a cannabidiol-predominant extract (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). This is the weakest of four measured effects but shows a common beverage is part of the system.

Grapefruit juice contains furanocoumarins inhibiting intestinal CYP3A4. Its effect aligns with cannabidiol, so both factors may sum effects on drugs dependent on this pathway. No study measured this combination, so no specific multiplier is supported.

St. John’s wort acts oppositely, inducing CYP3A4. The net effect with simultaneous cannabidiol use is unpredictable and changes over time, as induction builds over days and inhibition appears faster. From a safety perspective, unpredictability is worse than a known direction.

Alcohol is listed among substances with described mutual influence with cannabidiol (Balachandran et al., Journal of General Internal Medicine, 2021). Additive central nervous system depression occurs, the same mechanism as with benzodiazepines. An evening oil dose followed by a glass of wine is a combination needing no enzyme to cause harm.

How to prepare for a conversation with a doctor or pharmacist?

Start with a list, as assessment is impossible without it. Write down every prescription drug, over-the-counter product, supplement, and herb. For each, note dose, administration time, and start date. Occasional painkillers or antacid products also go on the list, as they use the same enzymes.

Ask specific questions instead of general ones. Asking if cannabidiol is safe has no good answer. Asking if adding cannabidiol changes anything with your drug set and recent tests has an answer and can be recorded. More tips are in the article on how to prepare for a doctor visit.

Bring current test results. For anticoagulant therapy, INR; for epilepsy, drug concentrations; for immunosuppression, tacrolimus or cyclosporine levels; plus liver tests and creatinine. Results older than six months have limited value, as the reference point for later comparisons is the state just before change.

Accept a negative answer if given. Situations where a reasonable recommendation is “do not combine” include transplant patients on tacrolimus, warfarin users post-thrombosis, and children on clobazam. The EFSA panel states more generally that cannabidiol safety cannot be established in people taking medications (EFSA Panel on Nutrition, 2026).

Which symptoms require immediate action?

Four symptom groups do not wait for scheduled visits. First, bleeding signs in anticoagulated patients: gum bleeding, blood in urine, black stools, bruises without injury. These require urgent INR testing and doctor evaluation, as intracranial bleeding is life-threatening.

Second, excessive sedation symptoms: drowsiness preventing normal function, slowed or slurred speech, balance disturbances. With clobazam or benzodiazepine treatment, these may indicate increased drug concentration. In the clobazam combination study, adverse effects resolved after dose reduction (Geffrey et al., Epilepsia, 2015).

Third, muscle and liver symptoms. Muscle pain with dark urine in a statin user requires creatine kinase testing. Jaundice or right upper quadrant pain indicate liver issues, which are closely monitored with cannabidiol (Brown and Winterstein, Journal of Clinical Medicine, 2019).

Fourth, transplant patients. Tremors, severe headache, or blood pressure spikes may indicate tacrolimus concentration increase and require urgent testing. None of these alone prove interaction but all require evaluation, not self-discontinuation.

What does this guide not resolve?

The main limitation concerns quantity. Studies underlying this text used purified cannabidiol in medicinal doses: weight-based in epilepsy trials, 640 mg in enzyme probe study, 2000-2900 mg in tacrolimus case. Typical supplements contain fractions of these amounts.

This leads to honest uncertainty both ways. It is wrong to claim a few oil drops produce the same effects as a medicinal dose, as this was not measured. It is also wrong to claim small amounts mean no interaction, as exposure depends on product form, meal, and individual metabolism (Millar et al., Frontiers in Pharmacology, 2018).

The second limitation concerns products themselves. A cannabidiol pharmacology review notes over-the-counter products often have unknown composition and bioavailability, and even rigorous studies used various sources, complicating comparisons (Britch et al., Psychopharmacology, 2021). Labeling does not replace laboratory analysis.

Hormonal contraception and antidiabetic drugs remain outside the table. No first-level human interaction measurements with cannabidiol were found, so they are not listed as safe or risky.

Levothyroxine, used for hypothyroidism, is different. In 22 children with drug-resistant epilepsy, cannabidiol concentration before the next dose was 109.6% higher in those taking levothyroxine (Brstilo et al., Pharmaceutics, 2023), and the EFSA panel notes this as a possible cannabidiol effect on thyroid hormone metabolism (EFSA Panel on Nutrition, 2026). This observation is from a small drug-treated group, so cannabidiol decisions with levothyroxine are physician’s responsibility.

The practical conclusion remains unchanged despite these gaps. If you take chronic medications, the decision to add cannabidiol belongs to a doctor or pharmacist, not product descriptions. The EFSA panel did not say cannabidiol harms this group, only that safety cannot be established, which is a different statement requiring caution.

Frequently Asked Questions

Does CBD interact with all medications?

No. In a study involving eighteen healthy adults, a CBD-predominant extract inhibited CYP2C19, CYP2C9, CYP3A, and CYP1A2, but did not change CYP2D6 activity at all (Bansal et al., Clinical Pharmacology and Therapeutics, 2023). The risk therefore concerns drugs metabolized by the first four pathways, not every prescription drug.

Can CBD oil be combined with psychotropic drugs?

Only after consulting a psychiatrist. Cannabidiol most strongly inhibits CYP2C19 (Bansal et al., 2023), and this enzyme metabolizes, among others, citalopram and escitalopram. A review of adverse effects lists antipsychotic drugs among CYP3A4 substrates and liver-burdening drugs (Brown and Winterstein, 2019). Antidepressants combined with cannabidiol may increase drowsiness.

What happens to clobazam concentration after adding CBD?

It increases, especially on the active metabolite side. In thirteen children with drug-resistant epilepsy, the average concentration of norclobazam increased by 500% after four weeks, and clobazam itself by 60% (Geffrey et al., Epilepsia, 2015). The clobazam dose was reduced in ten of the thirteen patients.

Can CBD harm the liver when combined with valproate?

An increase in aminotransferase activity was noted in 14 participants taking cannabidiol, i.e., 9% of them, in a study on Lennox-Gastaut syndrome involving 225 people (Devinsky et al., New England Journal of Medicine, 2018). Liver tests with such a combination are ordered and interpreted by the attending physician, not the patient.

Can CBD be taken together with warfarin?

A case was described where the introduction of cannabidiol forced a change in warfarin management, and the authors conclude that INR should be monitored after adding cannabinoids (Grayson et al., Epilepsy and Behavior Case Reports, 2018). Warfarin is a substrate of CYP2C9, a pathway inhibited by CBD. The decision is made by the doctor.

Does CBD increase tacrolimus concentration?

In a described case of a clinical trial participant, tacrolimus concentration normalized to dose increased about threefold when taking 2000-2900 mg of cannabidiol daily (Leino et al., American Journal of Transplantation, 2019). This is a medicinal amount, many times higher than the content of a typical supplement.

How long after stopping CBD should one be cautious about interactions?

Unknown. A pharmacokinetic review in humans shows that cannabidiol half-life varied between administration routes from about 1.4 hours to several days after chronic oral administration (Millar et al., Frontiers in Pharmacology, 2018). The timing for returning to previous drug doses is determined by the doctor.

Is a CBD supplement safer than a cannabidiol drug?

Not in the sense the question implies. The EFSA panel states that cannabidiol safety cannot be established in people simultaneously taking medications (EFSA Panel on Nutrition, 2026). The composition and actual content of over-the-counter products are sometimes unknown (Britch et al., Psychopharmacology, 2021).

This article is for informational and educational purposes and does not constitute medical advice. Before starting cannabis or CBD for therapeutic purposes, consult a doctor, especially if you take other medications, are pregnant, or breastfeeding.

Author: Michał Waluk · Published: 2026-05-11 · Updated: 2026-09-23