CYP2B6 Genotype

Two people swallow the same 600 mg efavirenz tablet. One is fine by the weekend. The other spends three weeks dizzy, dreaming vividly, unable to think straight, and eventually quits the drug.

A lot of that gap comes down to a single liver enzyme and how much working copy of it you inherited. This test reads the gene behind that enzyme. For one drug it can change the starting dose. For a few others, it changes what to watch.

The Drugs This Enzyme Handles

CYP2B6, or cytochrome P450 2B6, is one of the liver enzymes that process drugs. It handles only a small share of prescriptions. For the drugs that depend on it, inherited slow or fast versions can matter more than they would for a drug with several backup routes.

• Efavirenz: an older HIV drug where this enzyme is the main route out of the body and where CPIC gives dose guidance.
• Nevirapine: another older HIV drug partly handled by this enzyme and CYP3A. Genotype can shift levels, but there is no standard genotype-based dose.
• Methadone: methadone is a mix of two mirror-image forms of the same molecule. This enzyme clears the form tied more closely to heart rhythm effects faster than the form that relieves pain, so a slow genotype can let the first one build up.
• Bupropion: used for depression and for quitting smoking. Here the enzyme is not just removing the drug, it is converting much of it into hydroxybupropion, an active metabolite.
• Ketamine: this enzyme contributes to conversion into norketamine, though CYP3A4 and route of dosing also matter.
• Cyclophosphamide and ifosfamide: chemotherapy prodrugs whose activation involves this enzyme and several others.

Other substrates include artemisinin, propofol, selegiline, and meperidine, but this gene rarely changes a prescribing decision for them. The clear starting-dose rule is for efavirenz.

Efavirenz

This is the CYP2B6 drug with a clear CPIC dose-change recommendation. People with two slow copies reach efavirenz blood levels roughly two to three times higher than normal metabolizers on the same 600 mg dose, with clearance cut by about three quarters. They also carry the highest risk of the dizziness, vivid dreams, insomnia, and trouble concentrating that make people abandon the drug in the first month.

The Clinical Pharmacogenetics Implementation Consortium, the group that writes gene-based prescribing guidelines used in hospitals, recommends considering 400 mg a day for intermediate metabolizers, and 400 mg or even 200 mg a day for poor metabolizers, instead of the standard 600 mg.

That lower dose is not a trade of control for comfort. A large randomized trial in people starting treatment found 400 mg held its own against 600 mg for suppressing the virus through 96 weeks, with fewer drug-related side effects and fewer people stopping because of them.

So if efavirenz is about to be prescribed, the genotype is worth having before the first tablet rather than after three bad weeks. If you are already on it and doing fine, the result mostly tells you whether your levels are likely running high, which matters if side effects show up later or if another drug gets added.

Methadone

Methadone clearance varies a lot between people, and genotype is one measurable reason. Two slow copies can reduce clearance of S-methadone, the form of the molecule tied more closely to QT prolongation on an electrocardiogram.

A post-mortem study found the slow genotype was more common among people who died with methadone in their system than among living people taking it for maintenance. That is an association in a small group, not proof that the gene caused those deaths.

A 2024 CPIC guideline reviewed the methadone evidence and recommends standard dosing, titration, and monitoring for normal, intermediate, and poor metabolizers. The genotype effect on clearance is real and measurable, but not large enough to set a dose by. The result can add context if levels, sedation, or QT concerns appear.

Bupropion, Ketamine, and Chemotherapy

Bupropion is converted into hydroxybupropion, which circulates at roughly ten to seventeen times the level of the parent drug at steady state and carries much of the effect. A slow genotype can mean less of that conversion. A smoking-cessation trial found quit rates differed by genotype, but that finding has never turned into dosing guidance.

Ketamine is partly cleared by the same enzyme, but CYP3A4 and the route of dosing also matter. No routine ketamine dosing guideline uses this genotype.

Cyclophosphamide and ifosfamide depend on several liver enzymes, including this one, to become active. The evidence points in different directions, with some studies reporting faster rather than slower cyclophosphamide clearance in people with two slow copies, so it is too mixed to change chemotherapy from this result alone.

So slow does not mean more drug across the board. For efavirenz and methadone, which arrive active and get taken apart, slow can mean higher exposure. For bupropion, slow can mean less active metabolite. This is a direction-of-risk result, not a good or bad score.

Reading Your Result

Results come back as a pair of star alleles, one inherited from each parent, written like 1/6. The 1 is the reference version with normal function. Common numbered variants each carry a known consequence: 6 produces less working enzyme and 18 produces none, while 4 and *22 produce more.

Your two copies together generate a phenotype call. That call is the lab's shorthand for expected enzyme activity.

• Ultrarapid metabolizer: two increased-function copies, such as 4/4 or 22/22. Uncommon.
• Rapid metabolizer: one normal copy and one increased-function copy, such as 1/4.
• Normal metabolizer: 1/1.
• Intermediate metabolizer: one normal copy and one reduced-function copy, most often 1/6. This is the most common result that is not normal.
• Poor metabolizer: two reduced or non-function copies, such as 6/6 or 6/18. This is the group that reaches the highest efavirenz levels.

Rapid and ultrarapid are real categories here, unlike some drug-metabolism genes where only the slow end has been worked out. For efavirenz, being at the fast end mostly means the standard dose is the right one for you.

Why Ancestry Shifts the Odds

The 6 variant is common in many populations. It accounts for roughly 15 to 30% of gene copies in many European and East Asian groups, and it is considerably more common in people of African ancestry, where reported frequencies often exceed half of gene copies. In some sub-Saharan African cohorts, reduced-function alleles taken together (6 plus *18) account for close to half of all copies.

*18 behaves differently. It is found mostly in people of African ancestry and makes no working protein at all, so a panel that does not test for it can report a reassuring result in someone who is in fact a poor metabolizer. If you have African ancestry, which specific variants your lab covers matters more than it does for many pharmacogenetic tests.

When the Result Can Mislead You

• The panel only sees what it looks for. A 1/1 result means no tested variant was found, not that your gene is certainly normal. Rare variants exist, and they are more likely if your ancestry is thinly represented in the reference data.
• Ancestry changes the miss risk. A panel built around European variants can miss African-ancestry no-function alleles such as *18 unless it names them in the report.
• Other drugs can make the genotype less predictive. Clopidogrel, ticlopidine, and thiotepa can inhibit this enzyme, so a normal metabolizer can temporarily behave more slowly while taking them. Rifampin, phenytoin, and phenobarbital push it the other way. Efavirenz can speed up its own clearance over the first couple of weeks.
• A consumer genetic report is not the same product. Consumer reports often test one or two positions and infer the rest. If the result is going to change a prescription, confirm it with a clinical pharmacogenomic test that reports the variants used for the star-allele call.
• An unexpected variant may come back with no known meaning. Labs call this a variant of uncertain significance. It is an unknown, not a slow result, and it should not be treated as one.
• Stem-cell transplant can change what blood DNA shows. If you have had a bone-marrow or stem-cell transplant, a blood sample may reflect donor DNA. Tell the lab before testing.

One Result, Stored for Later

Your genotype does not change. There is no trend to follow and no retest interval. The only reason to run it again is if the first result came from a method you would not want to base a prescription on.

The value accumulates instead of moving. Put the result somewhere it will surface when someone is writing a prescription: your pharmacy record, your chart, a note on your phone. Most of the drugs this gene touches are ones you might meet once, under pressure, years from now.

What does need repeating is the ordinary monitoring around whichever drug you end up taking. Liver and kidney function still shape drug handling, and a normal genotype will not protect you from a liver that is struggling.

What to Do With a Slow Result

Nothing, until a relevant drug is on the table. This is a result you file, not one you act on the day it arrives.

When one of those drugs does come up, the path is short. Bring the result to whoever is prescribing, and ask a pharmacist if you want someone who reads these guidelines routinely. If the result came from a consumer report, confirm it with a clinical test before changing a prescription.

For efavirenz there is a specific lower starting dose to ask about. For methadone, CPIC recommends standard dosing, titration, and monitoring regardless of genotype, so the result is context rather than a dose rule. For bupropion, ketamine, cyclophosphamide, or ifosfamide, the result has to be interpreted drug by drug.

If you have already had an unexplained bad reaction to one of these drugs, the result can help explain it and warn you before the next one. A poor metabolizer who could not tolerate efavirenz has a reason to flag the genotype before anesthesia involving ketamine, before chemotherapy, and before starting bupropion for depression or smoking cessation.

If your history is a general pattern of odd drug reactions rather than one specific drug, one gene will only explain one slice of it. A wider pharmacogenomic panel tends to answer more of the question.