Amitriptyline

Amitriptyline blocks the reuptake of serotonin and norepinephrine into nerve cells, which leaves more of these chemical messengers active. The label states that its mechanism of action in people is not known, and that this reuptake effect is believed by some to explain its antidepressant activity. A sedating effect may appear before the antidepressant effect, which can take as long as 30 days to develop [1].

For depression, the label's usual adult outpatient dose is 75 mg a day, raised to 150 mg a day if needed, with maintenance doses of 50 to 100 mg a day; lower doses are recommended for older adults and adolescents [1]. For nerve pain, adults may start on 25 mg a day [2].

Like other tricyclics, amitriptyline is well absorbed, and on average about half of each dose is metabolized before it reaches the general circulation (first-pass metabolism) [2]. CYP2C19 removes a methyl group to make nortriptyline. Nortriptyline is an antidepressant in its own right (it is also prescribed on its own) and acts more on norepinephrine, while amitriptyline acts more on serotonin. CYP2D6 adds a hydroxyl group to both drugs, producing 10-hydroxy metabolites that are less active. CYP3A4 and CYP1A2 handle smaller shares [2]. Virtually the entire dose is excreted as glucuronide or sulfate conjugates of metabolites, with little unchanged drug in the urine [1].

1. AmitriptylineActive; acts more on serotonin
2. NortriptylineMade by CYP2C19; active; acts more on norepinephrine
3. 10-hydroxy metabolitesMade by CYP2D6 from both drugs; less active

CPIC lists anticholinergic, nervous system and heart effects as the common adverse effects of tricyclics [2]. On the amitriptyline label these include dry mouth, constipation, blurred vision and urinary retention, drowsiness, a drop in blood pressure on standing, and, particularly at high doses, heart rhythm and conduction changes [1]. According to CPIC, tricyclic blood levels are linked to both toxicity and benefit, which is why genes that change those levels matter [2].

CYP2D6

CYP2D6 varies a great deal between people. CPIC's 2016 guideline counted more than 100 versions (alleles), grouped by function: normal (such as *1 and *2), decreased (such as *9, *10, *17, *29 and *41) and no function (such as *3, *4, *5 and *6). CYP2D6*5 is a deletion of the whole gene, and some people carry extra copies, written xN, as in *1x2 [2][3].

To turn two alleles into one result, each allele gets an activity value: 1 for normal function, 0.5 or 0.25 for decreased function and 0 for no function. The two values are added. Under the scale CPIC and the Dutch Pharmacogenetics Working Group (DPWG) agreed in 2019, a total of 0 means poor metabolizer, 0.25 to 1 intermediate, 1.25 to 2.25 normal, and above 2.25 ultrarapid, which is only possible with extra gene copies [4][3]. One result of that change: *1/*4, with one working copy, scores 1.0 and is now an intermediate metabolizer, although the 2016 guideline paper called it normal [4]. A Gene2Rx report shows the CYP2D6 activity score next to the phenotype, and the amitriptyline recommendation is looked up by that score.

Poor metabolizers clear tricyclics slowly. The FDA label says their blood levels are higher than expected at usual doses and that, depending on how much of a tricyclic goes through CYP2D6, the rise can be small or quite large, such as an 8-fold increase in total exposure (AUC) [1]. Ultrarapid metabolizers have the opposite problem: drug levels may stay too low to work. CPIC describes documented cases of ultrarapid metabolizers who needed very large nortriptyline doses to reach target levels and then had high levels of its hydroxy metabolite, which may raise the risk of heart toxicity [2].

CYP2C19

Most CYP2C19 alleles are one of four: the normal *1; *2 and *3, which make no working enzyme; and *17, a change in the gene's promoter that increases how much enzyme is made. Two copies of *17 make an ultrarapid metabolizer and one *17 with a normal allele makes a rapid metabolizer; two no-function alleles make a poor metabolizer. CPIC classes *17 paired with a no-function allele, such as *2/*17, as intermediate, because limited data suggest *17 may not make up for the missing copy [2].

When CYP2C19 is slow, less amitriptyline becomes nortriptyline, and the ratio of amitriptyline to nortriptyline rises. In poor metabolizers, CPIC says the higher amitriptyline levels may increase the chance of side effects. Intermediate metabolizers would be expected to have a modest shift, which CPIC does not consider a reason to change the dose [2]. With *17 the shift goes the other way. In a study of 678 hospitalized Dutch patients taking amitriptyline, citalopram or clomipramine, people with two copies of *17 showed increased metabolism of amitriptyline, although the authors judged the clinical relevance probably limited [5]. CPIC adds that *17 did not change the combined level of amitriptyline plus nortriptyline but was linked to higher nortriptyline levels [2].

How common these results are

The table gives estimates from CPIC's frequency tables, which pool published studies by biogeographic group [6][7].

Share of people with each result, by ancestry (CYP2C19 poor includes likely poor)

Population group	CYP2D6 poor / intermediate / ultrarapid	CYP2C19 poor / rapid or ultrarapid
European	6.5% / 38% / 2.3%	2.4% / 32%
African American or Afro-Caribbean	2.3% / 36% / 4.1%	4.8% / 28%
East Asian	0.8% / 38% / 0.9%	13% / 2.6%
Central or South Asian	2.4% / 28% / 1.5%	8.2% / 21%
Latino	3.1% / 29% / 4.1%	1.2% / 27%

CYP2D6 intermediate results are common in every group in the table, at 28% to 38%. These figures are averages of published studies, many of them small, and CPIC's methods notes warn that the margin of error for estimates like these can be large [6].

Why the two genes are read together

The two enzymes act one after the other, so the combination matters. In a 2005 study, 50 inpatients with depression took a fixed 75 mg of amitriptyline twice a day. Side effects tracked nortriptyline levels, not amitriptyline levels. Patients with two working CYP2D6 copies had side effects far less often than those with only one (12.1% versus 76.5%). None of the 13 patients with two working CYP2D6 copies and only one working CYP2C19 copy reached the study's side-effect threshold, compared with 9 of 11 in the highest-risk group. Genotype was not linked to how well the drug worked [8].

That pattern fits the pathway: with less CYP2C19 activity, less nortriptyline is made, and with less CYP2D6 activity, nortriptyline is cleared more slowly. CPIC's combined table builds on these findings. CPIC classes all of its recommendations for people with unusual results for both genes as optional, because there are few clinical or pharmacokinetic data on such people [2].

The Clinical Pharmacogenetics Implementation Consortium (CPIC) published its tricyclic guideline in 2013 and updated it in 2016. It based the recommendations on amitriptyline and nortriptyline, the two tricyclics with the most pharmacogenetic research, and considers it reasonable to extend them to five other tricyclics, with fewer supporting data [2]. Both genes are CPIC level A for amitriptyline, CPIC's highest level: genetic information should be used to change prescribing [9]. CPIC's online tables apply the 2019 CYP2D6 scoring change, so an activity score of 1.0 now gets the intermediate metabolizer recommendation [10]. Each recommendation is graded strong, moderate or optional, and CPIC uses optional where the data are limited [2].

Three details shape how the advice applies. First, it is written for the higher doses used in depression. For nerve pain treated with lower doses, CPIC recommends no dose change for poor or intermediate metabolizers, with close monitoring for side effects; it suggests considering another drug for CYP2D6 ultrarapid metabolizers, and caution for people with a combination of poor or ultrarapid results. Second, the "starting dose" in the tables means the target steady-state dose: a prescriber may still begin low and build up over several days. Third, many recommendations call for therapeutic drug monitoring, a blood test of drug levels (for amitriptyline, both amitriptyline and nortriptyline), with the dose adjusted until symptoms improve with few or no side effects [2][8].

The FDA label does not give genotype-specific doses. Its precautions section explains that CYP2D6 poor metabolizers have higher than expected tricyclic levels, that drugs which block CYP2D6 can make normal metabolizers resemble poor metabolizers, and that it is desirable to monitor tricyclic blood levels when a CYP2D6 inhibitor is added [1]. FDA's Table of Pharmacogenetic Associations lists amitriptyline and CYP2D6 in its section for effects on drug levels only: ultrarapid, intermediate and poor metabolizers "may" have altered levels [11]. Neither FDA source addresses CYP2C19 for amitriptyline.

When both genes have a result, a Gene2Rx report shows amitriptyline as one combined entry: each gene's phenotype, with one recommendation from CPIC's combined table, restated below. CYP2D6 groups follow the activity score: poor 0, intermediate 0.25 to 1, normal 1.25 to 2.25, ultrarapid above 2.25.

CYP2D6 and CYP2C19 results and amitriptyline

Result	What it means	CPIC advice to prescribers (strength)
CYP2D6 normal metabolizer CYP2C19 normal or intermediate metabolizer	Typical clearance. CYP2C19 intermediate metabolizers have a slightly higher amitriptyline to nortriptyline ratio that does not call for a dose change.	Start with the recommended starting dose (strong).
CYP2D6 normal metabolizer CYP2C19 poor metabolizer	Less amitriptyline is converted to nortriptyline, so amitriptyline levels are higher, which may make side effects more likely.	Avoid amitriptyline. If it is still warranted, consider a 50% lower starting dose and use blood-level monitoring to guide changes (moderate).
CYP2D6 normal metabolizer CYP2C19 rapid or ultrarapid metabolizer	More amitriptyline is converted to nortriptyline, which may affect response or side effects.	Consider a drug not metabolized by CYP2C19, such as nortriptyline or desipramine. If amitriptyline is used, guide the dose with blood-level monitoring (optional).
CYP2D6 intermediate metabolizer CYP2C19 normal metabolizer	Slower CYP2D6 clearance; higher drug levels make side effects more likely.	Consider a 25% lower starting dose and use blood-level monitoring to guide changes (moderate).
CYP2D6 intermediate metabolizer CYP2C19 intermediate metabolizer	Both pathways somewhat slower than typical.	Consider a 25% lower starting dose; blood-level monitoring is strongly recommended (optional).
CYP2D6 intermediate metabolizer CYP2C19 poor metabolizer	Slower CYP2D6 clearance and much less conversion to nortriptyline.	Avoid amitriptyline. If it is warranted, blood-level monitoring to guide the dose is strongly recommended (optional).
CYP2D6 intermediate metabolizer CYP2C19 rapid or ultrarapid metabolizer	Slower CYP2D6 clearance with faster conversion to nortriptyline.	Consider a drug not metabolized by CYP2C19. If amitriptyline is warranted, blood-level monitoring is strongly recommended (optional).
CYP2D6 poor metabolizer CYP2C19 normal metabolizer	Greatly reduced CYP2D6 clearance; much higher drug levels and a higher chance of side effects.	Avoid amitriptyline. If it is still warranted, consider a 50% lower starting dose with blood-level monitoring (strong).
CYP2D6 poor metabolizer CYP2C19 intermediate metabolizer	Greatly reduced CYP2D6 clearance and somewhat slower CYP2C19.	Avoid amitriptyline. If it is warranted, consider a 50% lower starting dose; blood-level monitoring is strongly recommended (optional).
CYP2D6 poor metabolizer CYP2C19 poor, rapid or ultrarapid metabolizer	Greatly reduced CYP2D6 clearance with an unusual CYP2C19 result as well.	Avoid amitriptyline. If it is warranted, blood-level monitoring to guide the dose is strongly recommended (optional).
CYP2D6 ultrarapid metabolizer CYP2C19 normal metabolizer	Faster CYP2D6 clearance; drug levels may be too low to work.	Avoid amitriptyline. If it is still warranted, consider titrating to a higher target dose than usual, guided by blood-level monitoring (strong).
CYP2D6 ultrarapid metabolizer CYP2C19 intermediate, poor, rapid or ultrarapid metabolizer	Faster CYP2D6 clearance with an unusual CYP2C19 result as well.	Avoid amitriptyline. If it is warranted, blood-level monitoring is strongly recommended (optional).
Indeterminate (either gene)	The alleles found for that gene could not be assigned a function, so no phenotype was given. This is not the same as a normal result.	No combined recommendation. The report gives CPIC's own recommendation for the gene that was called; there is no genotype-based advice for the gene that was not.

If only one gene has a result, the single-gene recommendations match the rows above in which the other gene is normal. For example, a CYP2D6 intermediate metabolizer with no CYP2C19 result gets the 25% dose-reduction advice. CPIC applies the same recommendations to "likely intermediate" and "likely poor" CYP2C19 results, which involve decreased-function alleles, as to intermediate and poor results [10].

Bring the report to your prescriber or pharmacist along with a list of everything you take, including over-the-counter products. The two phenotypes and the CYP2D6 activity score are what a clinician needs to find the matching CPIC row. CPIC considers genotype-based dosing most useful when a tricyclic is started. After months on the drug, a genetic test may help less, because the dose may already have been adjusted to response, side effects or blood levels [2]. Do not start, stop or change amitriptyline because of a genetic result without talking to your prescriber. Stopping suddenly after long use can cause nausea, headache and malaise [1].

Medicines that change the picture

Other drugs can make an enzyme behave differently from what the genes predict, which is called phenoconversion. FDA lists bupropion, fluoxetine, paroxetine, quinidine and terbinafine as strong CYP2D6 inhibitors, and abiraterone, cinacalcet, duloxetine and mirabegron among the moderate ones [12]. The amitriptyline label adds that all SSRIs inhibit CYP2D6 to varying degrees, as do cimetidine and the heart rhythm drugs propafenone and flecainide, and that a person stable on a tricyclic "may become abruptly toxic" when an inhibitor is added [1]. CPIC reports that adding a potent CYP2D6 inhibitor such as fluoxetine may raise amitriptyline levels dramatically, and that it has been suggested to treat people on strong CYP2D6 inhibitors like poor metabolizers [2]. Fluoxetine's effect lasts after it is stopped: the label says at least 5 weeks may be needed between stopping fluoxetine and starting a tricyclic [1].

For CYP2C19, FDA lists fluconazole, fluoxetine, fluvoxamine and ticlopidine as strong inhibitors, and rifampin as a strong inducer, a drug that speeds the enzyme up [12]. Fluoxetine is on both strong-inhibitor lists. The label also reports large increases in amitriptyline levels in some people taking topiramate [1]. Phenoconversion is common: in a study of 900 adults treated for depression with venlafaxine, 4% were CYP2D6 poor metabolizers by genotype, but 27% had blood-level ratios typical of poor metabolizers [13].

Other factors

Older age, liver disease and reduced kidney function may call for lower tricyclic doses whatever the genotype [2]. Older adults are more sensitive to anticholinergic effects, may be at higher risk of falls, and generally have higher blood levels for a given dose [1]. The label also warns about serotonin syndrome when amitriptyline is combined with other serotonergic drugs, including SSRIs, SNRIs, triptans and tramadol [1].

A normal result for both genes points to standard dosing. It does not promise that the drug will be well tolerated, since anyone can have the side effects listed on the label, and it is a weak guide to whether the drug will help: CPIC points out that there are fewer data linking genotype to response than linking genotype to drug levels [2].

Consumer arrays from 23andMe, AncestryDNA and MyHeritage are not clinical-grade tests, and their raw data is labeled as not for medical use. They read fixed positions and can miss rare variants, and CPIC's guideline says most genotyping tests do not detect rare or new variants [2]. Array processing uses imputation to infer some variants that were not measured directly, but a variant at a position the data does not cover cannot be seen. A normal result therefore means that none of the assessed variants was found among the positions your data covers; it cannot rule out a variant the data did not cover. "Indeterminate" means the alleles found could not be assigned a function.

A genotype is one input among several. Side effects or a poor response can also come from the dose, other medicines, age, or liver or kidney function [2]. The label notes that blood levels can help identify people with excessively high levels and those with poor absorption or missed doses [1].

Blood-level monitoring shows how much drug is actually in the body, and CPIC recommends it for many of the results above.

I take a low dose for nerve pain. Does my result still matter?

Less than at depression doses. CPIC's separate low-dose advice, described under What the guidelines recommend, calls for no dose change in poor or intermediate metabolizers but close monitoring for side effects, and the standard tables apply if the dose is raised [2]. That advice is written for nerve pain; CPIC does not address migraine prevention separately.

Is nortriptyline a better fit for my result?

It depends on which gene is unusual. CPIC names nortriptyline and desipramine as tricyclics without major CYP2C19 metabolism, so they sidestep a CYP2C19 poor, rapid or ultrarapid result. Both still depend on CYP2D6, so they do not get around a CYP2D6 poor or ultrarapid result; for those, CPIC suggests a drug not metabolized by CYP2D6 [2].

Why does my report use two genes for amitriptyline but one for nortriptyline?

Nortriptyline is already the product of the CYP2C19 step. CPIC lists nortriptyline and desipramine as affected by CYP2D6 only, while amitriptyline, clomipramine, doxepin, imipramine and trimipramine are affected by both [2]. More on the drug class is on the tricyclic antidepressants page.

1. Amitriptyline hydrochloride tablets, USP (Sandoz): prescribing information. DailyMed, US National Library of Medicine. Indication, dosing, CYP2D6 precautions, drug interactions and side effects.
2. Hicks JK et al. Clinical Pharmacogenetics Implementation Consortium guideline (CPIC) for CYP2D6 and CYP2C19 genotypes and dosing of tricyclic antidepressants: 2016 update. Clin Pharmacol Ther (2017). The CPIC recommendations, how amitriptyline is metabolized, which doses the advice covers, and the evidence behind it.
3. CPIC. CYP2D6 allele functionality table (spreadsheet). Function and activity value of each CYP2D6 allele.
4. Caudle KE et al. Standardizing CYP2D6 genotype to phenotype translation: consensus recommendations from the Clinical Pharmacogenetics Implementation Consortium and Dutch Pharmacogenetics Working Group. Clin Transl Sci (2020). The activity score ranges used for CYP2D6 phenotypes since 2019.
5. de Vos A et al. Association between CYP2C19*17 and metabolism of amitriptyline, citalopram and clomipramine in Dutch hospitalized patients. Pharmacogenomics J (2011). CYP2C19*17 and amitriptyline metabolism.
6. CPIC. CYP2D6 frequency table (spreadsheet). Allele and phenotype frequencies by biogeographic group, with methods and caveats.
7. CPIC. CYP2C19 frequency table (spreadsheet). Phenotype frequencies by biogeographic group.
8. Steimer W et al. Amitriptyline or not, that is the question: pharmacogenetic testing of CYP2D6 and CYP2C19 identifies patients with low or high risk for side effects in amitriptyline therapy. Clin Chem (2005). Combined CYP2D6 and CYP2C19 genotype and side effects.
9. CPIC. Assignment of CPIC levels for genes/drugs (ClinPGx). What CPIC level A means.
10. CPIC guideline for CYP2D6, CYP2C19 and tricyclic antidepressants (ClinPGx guideline page). Current recommendation tables and updates since publication.
11. US Food and Drug Administration. Table of Pharmacogenetic Associations. Amitriptyline and CYP2D6, listed in section 3.
12. US Food and Drug Administration. For healthcare professionals: FDA's examples of drugs that interact with CYP enzymes and transporter systems. Strong and moderate CYP2D6 and CYP2C19 inhibitors and inducers.
13. Preskorn SH et al. Cytochrome P450 2D6 phenoconversion is common in patients being treated for depression: implications for personalized medicine. J Clin Psychiatry (2013). How often measured CYP2D6 activity differs from genotype in practice.