Key Points
Question
Is the reduced function of cytochrome P450 2D6 (CYP2D6), via either CYP2D6 genotype or CYP2D6 inhibitor–mediated phenoconversion, associated with pain-related emergency department (ED) visits among patients taking CYP2D6-metabolized opioids?
Findings
In this cohort study using a national dataset of 31 669 patients prescribed hydrocodone, tramadol, codeine, or oxycodone, patients who were phenotypic CYP2D6 intermediate metabolizers or poor metabolizers, considering CYP2D6 genotype and drug interactions, had significantly more pain-related ED visits than patients who were phenotypic ultrarapid metabolizers or phenotypic normal metabolizers.
Meaning
This study suggests that reduced CYP2D6 activity impairs the analgesic response of commonly used opioids metabolized via that pathway.
Abstract
Importance
Cytochrome P450 2D6 (CYP2D6) bioactivates hydrocodone, tramadol, codeine, and oxycodone to active metabolites that primarily provide analgesic activity. Reduced CYP2D6 activity may be associated with poor pain control.
Objective
To evaluate associations of impaired CYP2D6 activity based on genotype or CYP2D6 inhibitors, alone and together, with analgesic activity of CYP2D6-metabolized opioids among patients with pain.
Design, Setting, and Participants
This retrospective national, community-based cohort study used electronic health records and genetics data from the All of Us Research Program. Participants included adults prescribed at least 1 CYP2D6-metabolized opioid for more than 7 days between January 1, 2014, and December 31, 2022, with whole-genome sequencing data available. Analysis groups were defined by CYP2D6 phenotype, which was determined based on CYP2D6 genotype or CYP2D6 inhibitor–mediated phenoconversion. Statistical analysis was performed from July 1, 2023, to January 15, 2025.
Exposures
CYP2D6-metabolized opioids, with or without concomitant CYP2D6 inhibitor exposure, based on prescription records and overlap with opioids.
Main Outcomes and Measures
The primary outcome was occurrence of any pain-related emergency department (ED) visits during opioid treatment, up to 60 days after opioid initiation. The association between ED visits and CYP2D6 phenotype was assessed using inverse probability treatment weighting–adjusted logistic regression. Additional analyses were conducted by drug and isolating CYP2D6 genotype and inhibitors.
Results
Among 31 669 patients (mean [SD] age, 51.2 [15.4] years; 66.5% women) prescribed CYP2D6-metabolized opioids, 15 960 had reduced CYP2D6 activity, and 15 709 had normal or high CYP2D6 activity based on genotype and inhibitors. A higher percentage of patients with reduced CYP2D6 activity (hereafter referred to as phenotypic intermediate metabolizers [pIMs] or phenotypic poor metabolizers [pPMs]) had experienced pain-related ED visits compared with patients with normal or high CYP2D6 activity (phenotypic normal metabolizers [pNMs] and phenotypic ultrarapid metabolizers [pUMs]) (2.1% vs 1.8%; inverse probability–weighted odds ratio, 1.19; 95% CI, 1.06-1.33). There were no significant differences in ED visits among CYP2D6 genotypic IMs or PMs vs NMs or UMs when testing all 4 drugs together. Among genotypic NMs, ED visits were more frequent among the individuals prescribed CYP2D6 inhibitors (inverse probability–weighted odds ratio, 1.49; 95% CI, 1.32-1.68). In analyses by medication, drug interactions were important for all 4 medications, while genotype associations were significant only for hydrocodone, tramadol, and codeine.
Conclusions and Relevance
In this cohort study, reduced CYP2D6 activity was associated with increased ED visits among individuals treated with CYP2D6-metabolized opioids. This finding suggests that incorporating data on CYP2D6 genotype and accounting for drug interactions in opioid prescribing may improve pain management and reduce ED visits.
This cohort study evaluates whether individuals with reduced vs normal or high CYP2D6 activity have more pain-related emergency department visits when pain taking CYP2D6-metabolized opioids.
Introduction
In the US, approximately 100 million people experience chronic pain, with an annual cost between $560 billion and $635 billion, exceeding that of heart disease, cancer, or diabetes.1 Opioids metabolized by the enzyme cytochrome P450 2D6 (CYP2D6) include hydrocodone, tramadol, codeine, and oxycodone and represent most opioid prescriptions written in the US.2 CYP2D6 bioactivates hydrocodone, tramadol, and codeine into active metabolites, and the parent compounds have little to no analgesic activity compared with the active metabolites.3,4 Both CYP2D6 and CYP3A4 metabolize oxycodone.4,5 As the parent drug oxycodone and some of the metabolites are active compounds, the data are equivocal on the association of CYP2D6 with oxycodone’s analgesic response.4,5
CYP2D6 is highly polymorphic, and more than 100 alleles have been identified.6 Variant alleles range from loss of function and no enzyme activity (with patients hereafter referred to as poor metabolizers [PMs]) to gene duplication and increased activity (with patients hereafter referred to as ultrarapid metabolizers [UMs]).4,7 Poor metabolizers experience a reduced analgesic response from CYP2D6-metabolized opioids compared with patients who are normal metabolizers (NMs) or intermediate metabolizers (IMs).7 In addition, there are commonly used drugs that the US Food and Drug Administration (FDA) has categorized as strong CYP2D6 inhibitors (bupropion, fluoxetine, paroxetine, terbinafine, and quinidine) (eTable 1 in Supplement 1)8 that lead to the absence of CYP2D6 function and moderate inhibitors (abiraterone, cinacalcet, mirabegron, duloxetine, lorcaserin, and rolapitant) that reduce CYP2D6 activity by about half, with the potential to convert those with the NM phenotype to the PM phenotype or those with the UM phenotype to the IM phenotype.4,9 Many of the CYP2D6 inhibitors are antidepressants, and data indicate that 15% to 25% of patients prescribed opioids are coprescribed a CYP2D6 inhibitor.10
There is, however, little direct evidence of the association of CYP2D6 genotype or inhibitors with analgesic activity of CYP2D6-metabolized opioids in terms of clinical outcomes. A recent clinical study found that patients prescribed a CYP2D6-metabolized opioid alongside a CYP2D6-inhibiting antidepressant had more pain-related emergency department (ED) visits, supporting the hypothesis that CYP2D6 inhibition reduces the analgesic activity of opioids.11 No study from a clinical setting has included CYP2D6 genetics and CYP2D6 drug interactions because genetic data are typically not available within electronic health records (EHRs). In this study, we sought to evaluate the associations of CYP2D6 genotype and CYP2D6 inhibitor–mediated phenoconversion, alone and together, with the analgesic response of hydrocodone, tramadol, codeine, and oxycodone among patients with pain, as assessed by pain-related ED visits using the National Institutes of Health (NIH) All of Us Research Program (All of Us) data. We hypothesize that CYP2D6 PMs and IMs, as defined by genotype, and those patients phenoconverted to these phenotypes by CYP2D6 inhibitors will have worse pain control, as assessed by more frequent pain-related ED visits.
Methods
Data Source and Settings
We conducted a retrospective cohort study using EHR data and short-read whole-genome sequencing (SR-WGS) data from the All of Us Controlled Tier Dataset, version 7, using data from January 1, 2014, to December 31, 2022. This program, led by the NIH, creates one of the largest, most diverse nationwide biomedical data resources in the US.12 The All of Us institutional review board provides approval for all aspects of the All of Us program. All participants in All of Us provided written consent to participate in the research (eMethods in Supplement 1). The details of the All of Us research protocol have been previously published.12 This article complies with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.13
Study Participants and Design
Inclusion criteria were adult patients (aged ≥18 years) who were prescribed at least 1 CYP2D6-metabolized opioid (hydrocodone, tramadol, codeine, or oxycodone) for at least 7 days from January 2014 to December 2022 and had all prescription-related and SR-WGS data available. Exclusion criteria were having cancer-related diagnosis codes in the 6 months prior to the index date and missing or indeterminate CYP2D6 genotypes. Figure 1 outlines the measures used to define the study cohort. Index date (day zero) was 3 days after the date of the opioid prescription start date (Figure 2).
Figure 1. Flowchart of Study Cohort Identification.
Patients were referred to as phenotypic intermediate metabolizers (pUMs), phenotypic normal metabolizers (pNMs), phenotypic poor metabolizers (pPMs), and phenotypic ultrarapid metabolizers (pUMs) based on genotype and concomitant use of CYP2D6 inhibitors.
Figure 2. Graphical Representation of the Study Design.
The index date was set 3 days after the start date of the opioid prescription. Follow-up started from the index date and stopped at the end of 60 days or the end date of the opioid exposure, whichever occurred first. Baseline covariates were assessed for patients 180 days prior to the index date.
CYP2D6 Genotyping and Phenotype Assignment
To determine CYP2D6 star alleles (haplotype patterns at the gene level that have been associated with protein activity levels) and diplotypes from the SR-WGS data, we used a consensus-based approach integrating results from multiple bioinformatic tools, including Aldy,14 Cyrius,15 PyPGx,16 and StellarPGx.17 Each of these tools uses distinct algorithms to address the complexities of CYP2D6 genotyping. By combining the outputs of these tools, we generated more reliable CYP2D6 genotype assignments. The consensus call used a majority agreement between haplotype callers while requiring a call from at least 2 tools. If a consensus call could not be made, an indeterminate genotype was reported. Each star allele was assigned an activity score, which was then used to define the inferred phenotype of the encoded protein based on genotype only or both genotype and CYP2D6 inhibitors (Table 1) following the guidelines of Clinical Pharmacogenetics Implementation Consortium (CPIC).4
Table 1. CYP2D6 Activity Score and Phenotype Adjusted After Consideration of CYP2D6 Inhibitors.
| Allele and phenotypes | Activity scorea |
|---|---|
| Alleleb | |
| Functional (*1, *2 or *35) | 1 |
| Reduced functional (*9,*17,*29 or *41; *10) | 0.5; 0.25 |
| Nonfunctional (*3 through *8, *11 or *15) | 0 |
| Phenotype based on genotype alone or based on genotype and inhibitors both c , d | |
| Ultrarapid metabolizers | >2.25 |
| Normal metabolizers | ≥1.25 to ≤2.25 |
| Intermediate metabolizers | >0 to <1.25 |
| Poor metabolizers | 0 |
The activity score of CYP2D6 is the sum of the values assigned to each allele.
Asterisks indicate star alleles of CYP2D6.
Phenotype based on genotype alone: phenotype is estimated based on the genotype-based activity score (sum of the values assigned to each allele).
Phenotype based on genotype and inhibitors both: here, phenotype is estimated based on the activity score considering the effect of drug interaction. So, CYP2D6 activity score = inhibitor factor × genotype-based activity score; inhibitor factor = 0 for strong CYP2D6 inhibitor, 0.5 for moderate CYP2D6 inhibitor, and 1 for no CYP2D6 inhibitor.
Exposure
We assessed drug exposure based on the dates of prescriptions. In the All of Us EHR data, the medication directions (sig or instructions) were missing, so we used the most common directions of the prescriptions, as observed in a recent project (4 tablets a day),11 to calculate the end dates of the opioids. To accurately evaluate the exposure to concomitant opioid and CYP2D6 inhibitor prescriptions, we implemented a stitching method in which we stitched or combined prescriptions of the same medication with a gap of fewer than 14 days between them and considered the end date of the last prescription as the end date of that opioid exposure. For antidepressant medications, we stitched or combined prescriptions of the same medication only if there was a gap of less than or equal to 3 days between them, and we considered the end date of the last prescription as the end date of that antidepressant medication prescription. Concomitant exposure was then determined based on the start date and calculated end date with a minimum of consecutive 3-day overlap. Patients who did not have any CYP2D6 inhibitor prescription exposure in the last 180 days from the opioid start date and throughout the study period were considered not exposed to inhibitor prescriptions. We considered the first opioid prescriptions for patients who had multiple opioid prescriptions.
Outcome and Follow-Up Periods
Our outcome variable for this study was an ED visit for pain. We identified the first pain-related ED visits based on pain-related diagnostic codes (eTable 2 in Supplement 1) that were captured in the ED during the follow-up period.11 Follow-up started from the index date and stopped at the end of 60 days or the end date of the opioid prescription, whichever occurred first (Figure 2).
Covariates
We assessed 24 covariates (listed in Table 2) associated with patient demographic characteristics, comorbidities, comedications, and other factors that might have been associated with the outcome of ED visits for pain. Participants’ self-reported information about their race and ethnicity and sex from EHR data of All of Us was used to categorize them. These variables help control for confounding variables and understand demographics influences.
Table 2. Characteristics of the Patients.
| Characteristic | Patients, No. (%)a | P value | |
|---|---|---|---|
| pNM and pUM (n = 15 709) | pPM and pIM (n = 15 960) | ||
| Age, mean (SD), y | 50.6 (15.6) | 51.8 (15.1) | <.001 |
| Sex | |||
| Female | 10 227 (65.1) | 10 838 (67.9) | <.001 |
| Male | 5482 (34.9) | 5122 (32.1) | |
| Race and ethnicity | |||
| Black | 2968 (18.9) | 2624 (16.4) | <.001 |
| White | 8747 (55.68) | 10 570 (66.2) | |
| Other or none indicatedb | 3994 (25.4) | 2766 (17.33) | |
| Pain diagnosesc,d | |||
| Back pain | 2031 (12.9) | 2511 (15.7) | <.001 |
| Pain in hand, leg, or joints | 3991 (25.4) | 4479 (28.1) | <.001 |
| Rheumatoid arthritis | 3547 (22.6) | 4183 (26.2) | <.001 |
| Headache (including migraine) | 1416 (9.0) | 1618 (10.1) | <.001 |
| Neuropathic pain | 1810 (11.5) | 2330 (14.6) | <.001 |
| Fibromyalgia | 182 (1.2) | 333 (2.1) | <.001 |
| Injury | 4033 (25.7) | 4220 (26.4) | .12 |
| Comorbiditiesc,d | |||
| Depression | 2036 (13.0) | 3339 (20.9) | <.001 |
| Psychoses | 512 (3.3) | 798 (5.0) | <.001 |
| Anxiety | 1775 (11.3) | 2611 (16.4) | <.001 |
| Opioid use disorder | 168 (1.1) | 216 (1.4) | .02 |
| Diabetes | 2424 (15.4) | 2385 (14.9) | .23 |
| Kidney failure | 927 (5.9) | 867 (5.4) | .07 |
| Liver disease | 846 (5.4) | 821 (5.1) | .34 |
| Medication historyc,d | |||
| Opioid or nonopioid pain medication | 8581 (54.6) | 8637 (54.1) | .37 |
| Benzodiazepines and other sedative-hypnotics | 4723 (30.1) | 5430 (34.0) | <.001 |
| Antipsychotics | 2026 (12.9) | 2829 (17.7) | <.001 |
| CNS medications or stimulants | 168 (1.1) | 414 (2.6) | <.001 |
| Skeletal muscle relaxants | 3170 (20.2) | 3613 (22.6) | <.001 |
| Opioid-related medication use, MME/d | 39.0 (89.6) | 37.6 (83.1) | .13 |
| Health care use (ED visits), mean (SD), No.d | 0.32 (0.97) | 0.31 (0.95) | .63 |
Abbreviations, CNS, central nervous system; ED, emergency department; MME, morphine milligram equivalent; pIM, phenotypic intermediate metabolizer; pNM, phenotypic normal metabolizer; pPM, phenotypic poor metabolizer; pUM, phenotypic ultrarapid metabolizer.
The patients were referred to as pUMs, pNMs, pIMs, and pPMs based on genetic activity and concomitant use of CYP2D6 inhibitors.
Other includes Asian, Middle Eastern, Native Hawaiian, and mixed.
Patients could have more than 1 pain diagnosis, comorbidity, and medication history.
Within 6 months prior to the index date.
Statistical Analysis
Statistical analysis was performed from July 1, 2023, to January 15, 2025. Demographic and baseline characteristics were compared between phenotypically UMs (pUMs) and phenotypically NMs (pNMs) vs phenotypically IMs (pIMs) and phenotypically PMss (pPM) as numbers and percentages (categorical variables) or mean (SD) values (continuous variables). Continuous variables were assessed for normality. Differences were assessed using χ2 tests or 2-sample t tests as appropriate.
We estimated propensity scores using logistic regression to model the probability of receiving the exposure of interest (pNMs or pUMs vs pIMs or pPMs) as a function of baseline patient characteristics. We used the propensity score to generate an inverse probability of treatment weight (IPTW) for each patient. The IPTW was calculated as the inverse of the propensity score for the exposed group and the inverse of 1 minus the propensity score for the comparison group. A candidate list of potential patient characteristics to be included in the propensity score models was determined a priori based on clinical knowledge and published literature of the characteristics likely to be associated with our outcome. Finally, 24 covariates associated with patient demographic characteristics, comorbidities, comedications, and other factors were included in the propensity score model. We assessed covariate balance before and after IPTW using the standardized mean difference, with a value of less than 0.1 considered balanced.18 After IPTW, the patient population was well balanced across all characteristics (eTable 3 in Supplement 1). For the primary analysis, logistic regression using IPTW was conducted to test the association between pNMs or pUMs and pIMs or pPMs with the presence of ED visit (yes or no) as a measure of pain control.
In addition to the primary analysis described, we evaluated the association of CYP2D6 inhibitors with pain control among CYP2D6 NMs as they do not have genetic impairments to the generation of the active metabolite (without confounding by genotype). Within this group, we tested ED visits among those with or without CYP2D6 inhibitor prescriptions. Moreover, we evaluated the associations between CYP2D6 phenotypes based on genotypes (genotypic NMs or UMs vs genotypic IMs or PMs) and ED visits among patients without any inhibitor prescription (not confounded by CYP2D6 inhibitors).
We also conducted subgroup analyses for patients who were taking hydrocodone, tramadol, or codeine because these are drugs whose parent compounds mostly lack analgesic activity and are most likely to have an association with CYP2D6 genotype or CYP2D6 inhibitors. As the data are unclear on the association of CYP2D6 with oxycodone response, we performed a separate subgroup analysis for the oxycodone cohort. For all the analyses, we used logistic regression to understand the association of CYP2D6 genotype or inhibitor alone or together with pain-related ED visits. We reported unweighted odds ratios (ORs) and inverse probability (IP)–weighted ORs. Statistical significance was assessed using 95% CIs, where results were considered significant if the 95% CI did not include the null value. All P values were from 2-sided tests, and results were deemed statistically significant at P < .05. Data preprocessing was conducted in Python, version 3.10 (Python Software Foundation), and all analyses were performed using SAS, version 9.4 (SAS Institute Inc) within the All of Us Researcher Workbench. Some graphical presentations were prepared in Prism, version 10 (GraphPad Software).
Results
Study Population
The initial data pull yielded 98 251 adults who were prescribed at least 1 of the 4 opioids (hydrocodone, tramadol, codeine, or oxycodone); of these, 31 669 (mean [SD] age, 51.2 [15.4] years; 66.5% women and 33.5% men; 18.9% Black, 55.7% White, and 25.4% other race or ethnicity or none indicated) had active opioid prescriptions with a supply greater than 7 days and had available CYP2D6 genotype data, representing the analysis cohort (Figure 1 and Table 2). Among them, 15.0% of patients were exposed to any of the FDA-defined CYP2D6 strong or moderate inhibitors, while 85.0% were not. Basic demographic characteristics (eg, age, race and ethnicity, and sex) were complete with no missing data. We relied on EHR-recorded data, recognizing that certain factors—such as external prescriptions or diagnoses—may not be captured, limiting the assessment of missingness.
In the primary analysis, of 31 669 adults, 15 709 were pNMs or pUMs, and 15 960 were pPMs or pIMs. The demographic and clinical characteristics of the primary analysis are provided in Table 2. The distribution of various opioids and CYP2D6 inhibitors among individuals is presented in eTable 4 in Supplement 1.
Pain-Related ED Visits
Analysis of the Association of Both CYP2D6 Inhibitors and Genotype With Pain-Related ED Visits
A higher percentage of patients who were pPMs or pIMs considering both CYP2D6 genotype and inhibitors experienced pain-related ED visits compared with those who were pUMs and pPMs (2.1% vs 1.8%), with an IP-weighted OR of 1.19 (95% CI, 1.06-1.33) (Figure 3A). Among the different opioid subgroups, higher estimates were observed when examining the combination of hydrocodone, tramadol, and codeine only (IP-weighted OR, 1.34; 95% CI, 1.14-1.58). However, for oxycodone alone, there was no significant increase in ED visits between phenotypes (IP-weighted OR, 1.03; 95% CI, 0.87-1.21).
Figure 3. Association of CYP2D6 Genotype and Inhibitors Alone and Together With Pain-Related Emergency Department (ED) Visits.
A, ED visits among different patients who were CYP2D6 metabolizers: phenotypic normal metabolizers (pNMs) and phenotypic ultrarapid metabolizers (pUMs) vs phenotypic intermediate metabolizers (pIMs) and phenotypic poor metabolizers (pPMs), considering both CYP2D6 inhibitors and genotype. B, ED visits among genotypic normal metabolizers (NMs) comparing those exposed to CYP2D6 inhibitors vs those not exposed to CYP2D6 inhibitors. C, Visits among patients not taking CYP2D6 inhibitors comparing different CYP2D6 metabolizers: genotypic NMs and ultrarapid metabolizers (UMs) vs genotypic intermediate metabolizers (IMs) and poor metabolizers (PMs). Patients were referred to as pUMs, pNMs, pPMs, and pIMs based on genotype and concomitant use of CYP2D6 inhibitors; patients were referred to as UMs, NMs, IMs, and PMs based on genotype only. IP indicates inverse probability; OR, odds ratio.
aStatistically significant.
Analysis of the Association of CYP2D6 Inhibitors Alone With Pain-Related ED Visits
To understand the association of CYP2D6 inhibitors alone without confounding by genotype, we included only the NMs in this analysis. As shown in Figure 3B, this analysis yielded a significant association between ED visits and inhibitor prescription (IP-weighted OR, 1.49; 95% CI, 1.32-1.68). Subgroups based on different opioids yielded consistent findings.
Analysis of the Association of CYP2D6 Genotype Alone With Pain-Related ED Visits
To examine the association of genotypes only, without the confounding influence of inhibitors, we included patients with no CYP2D6 inhibitor prescriptions. Both genotypic IM or PM and UM or NM groups had identical pain-related ED visit rates (1.8%) (Figure 3C), and no significant association was observed (IP-weighted OR, 1.10; 95% CI, 0.96-1.36). In the hydrocodone, tramadol, and codeine subgroup, a significantly higher percentage of genotypic IMs and PMs experienced a pain-related ED visit compared with genotypic NMs and UMs (IP-weighted OR, 1.32; 95% CI, 1.10-1.59). Among oxycodone users, there were no differences in ED visits among genotype IMs and PMs compared with UMs and NMs.
Discussion
Our primary analysis of 31 669 individuals taking hydrocodone, tramadol, codeine, or oxycodone showed that a significantly higher percentage of pIMs and pPMs, considering both CYP2D6 genotype and inhibitors, experienced a pain-related ED visit compared with pUMs and pNMs. Among the combined subgroup of patients taking hydrocodone, tramadol, and codeine, we observed consistent results, but in the oxycodone-only analysis, we did not observe a significant increase in ED visits among pIMs or pPMs. This finding is consistent with existing literature that suggests CYP2D6 may play a less critical role in the analgesic activity of oxycodone.
Previous studies indicate that the active metabolites biotransformed from hydrocodone, tramadol, and codeine by CYP2D6 having higher affinity for the μ-opioid receptors are fully or largely responsible for the analgesic response those opioids provide.3,19,20 Our results suggest a clinically important association of reduced CYP2D6 activity with patients who are phenotypically IMs or PMs while taking these drugs having increased pain-related ED visits.
Conversely, oxycodone is metabolized to oxymorphone by CYP2D6 and noroxycodone by CYP3A4, which are further metabolized to noroxymorphone by the activity of CYP2D6 and CYP3A4, respectively.19,21 Oxycodone, oxymorphone, and noroxymorphone are all active compounds; however, the exact association of the parent and different metabolites with the analgesic response is unknown.22 Data have generally suggested that CYP2D6 activity may have smaller associations with the analgesic activity of oxycodone than that of other CYP2D6-metabolized opioids,5,19 and our results support that finding.
In 2 additional analyses, we aimed to separately examine the association of genotype and CYP2D6 inhibitors alone with pain control, ensuring the absence of interference from the other. Among genotypic NMs, pain-related ED visits were more frequent among those prescribed CYP2D6 inhibitors compared with patients not taking CYP2D6 inhibitors, which is consistent with findings of a previous study11 and suggests that the coadministration of CYP2D6 inhibitors with CYP2D6-metabolized opioids may impair opioid response, leading to increased pain-related ED visits. For patients not taking CYP2D6 inhibitors, no overall association was observed across 4 opioids; however, hydrocodone, tramadol, or codeine users with IM and PM genotypes experienced more pain-related ED visits, whereas oxycodone users did not, indicating that oxycodone data diluted the overall analysis.
Unlike CYP2D6 genotype effects, CYP2D6 inhibitors might show a greater association with oxycodone response and pain-related ED visits by broadly disrupting drug metabolism and distribution. Although genetic variations are associated with CYP2D6 activity—often offset by CYP3A4’s role—inhibitors block CYP2D6 and may also inhibit CYP3A4 (eg, by norfluoxetine, a main metabolite of fluoxetine), altering drug-drug interactions23 and/or changing transporter activities.24 These combined associations could lead to pronounced disruptions in oxycodone pharmacokinetics, amplifying their clinical association beyond that of genotype alone.
Given that ED visits are a major factor associated with health care costs, incorporating preemptive CYP2D6 genotyping could proactively identify patients who might be at increased risk of poor pain control by the CYP2D6-metabolized opioids, improving pain management and reducing health care costs associated with ED visits. Clinical consideration of CYP2D6 drug interactions could have a similar impact. Expanding clinical decision support tools to integrate both pharmacogenetics and drug-drug interactions would optimize pain management and potentially reduce opioid-related ED visits for pain.
Our data suggest that patients who are CYP2D6 pIMs or pPMs and are prescribed hydrocodone, tramadol, or codeine have more ED visits than those with normal CYP2D6 function. This finding implies they are experiencing inadequate pain control—which can be estimated based on the known pharmacokinetics of these drugs. These findings are consistent with current CPIC guidelines4 and add strength to the evidence suggesting that use of these data in the clinical setting may lead to improved patient outcomes. CYP2D6 genotype data are readily available from multiple national references and other laboratories, and results are routinely reported as a phenotype (eg, IM). Use of pharmacogenetic data are optimized through pharmacogenetic clinical decision support tools, which are available in some EHR systems. Consideration of the drug interaction along with the genotype has not been optimized with tools in the EHR, but a calculator tool25 is available at PharmGKB26 that can calculate the CYP2D6 phenotype based on genotype and drug interactions. Although genotype is immutable, the inhibitor drug prescriptions can change over time, so genotype must be reassessed with each new opioid prescription.4,11,27 Although our data with oxycodone did not reveal an association with genotype alone, when considering CYP2D6 inhibitors, we did observe increased ED visits. Thus, consideration of drug interactions, and perhaps genotype, may also be reasonable with oxycodone.
Strengths and Limitations
This study has several strengths. First, to our knowledge, this is the first population-based clinical cohort study investigating the association of reduced CYP2D6 activity, considering CYP2D6 genotypes and inhibitor alone (isolating the associations of inhibitors and genetics) and together, with pain control by the CYP2D6-metabolized opioids. Second, use of data from All of Us allowed for a diverse, nationwide sample, enhancing generalizability. Third, we included a large cohort with CYP2D6 genotype data obtained from SR-WGS data available in All of Us.
This study also has several limitations. We used EHR data from All of Us, which may include misclassifications and incomplete records from various health care systems. Some participants may have received care outside the shared EHR network, limiting the assessment of missing data. Emergency department visits may be underreported, as people are likely to go to an ED close to where they live. However, this would be expected to bias toward the null. Prescription directions were unavailable, requiring assumptions based on prior studies.11 Again, actual consumption of the medications could not be verified. Differences between comparison groups may arise from combining individuals with different pain problems,11 but propensity scoring ensured balance, and all ORs were IP weighted.
Conclusions
In this cohort study, reduced CYP2D6 activity was observed to be associated with increased pain-related ED visits among those treated CYP2D6-metabolized opioids. Incorporating CYP2D6 genotyping and accounting for drug interactions in opioid prescribing may improve pain management and reduce ED visits.
eMethods. Patient Consent and Data From All of Us Research Program
eTable 1. List of CYP2D6 Strong and Moderate Inhibitors
eTable 2. List of Emergency Department (ED) Visit SNOMED Standard Concept Name and Codes
eTable 3. Characteristics of the Patients Before and After Inverse Probability of Treatment Weighted (IPTW) Balancing
eTable 4. Distribution of the Different Medications in the Study Population
Data Sharing Statement
References
- 1.Smith TJ, Hillner BE. The cost of pain. JAMA Netw Open. 2019;2(4):e191532. doi: 10.1001/jamanetworkopen.2019.1532 [DOI] [PubMed] [Google Scholar]
- 2.The top 300 of 2022. ClinCalc.com. Accessed January 28, 2024. https://clincalc.com/DrugStats/Top300Drugs.aspx
- 3.Foster A, Mobley E, Wang Z. Complicated pain management in a CYP450 2D6 poor metabolizer. Pain Pract. 2007;7(4):352-356. doi: 10.1111/j.1533-2500.2007.00153.x [DOI] [PubMed] [Google Scholar]
- 4.Crews KR, Monte AA, Huddart R, et al. Clinical Pharmacogenetics Implementation Consortium guideline for CYP2D6, OPRM1, and COMT genotypes and select opioid therapy. Clin Pharmacol Ther. 2021;110(4):888-896. doi: 10.1002/cpt.2149 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Smith DM, Weitzel KW, Elsey AR, et al. CYP2D6-guided opioid therapy improves pain control in CYP2D6 intermediate and poor metabolizers: a pragmatic clinical trial. Genet Med. 2019;21(8):1842-1850. doi: 10.1038/s41436-018-0431-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.PharmVar: Pharmacogene Variation Consortium. PharmVar. Accessed August 6, 2022. http://www.PharmVar.org
- 7.Lötsch J, Skarke C, Schmidt H, et al. Evidence for morphine-independent central nervous opioid effects after administration of codeine: contribution of other codeine metabolites. Clin Pharmacol Ther. 2006;79(1):35-48. doi: 10.1016/j.clpt.2005.09.005 [DOI] [PubMed] [Google Scholar]
- 8.Drug development and drug interactions: table of substrates, inhibitors and inducers. U.S. Food and Drug Administration. Accessed August 6, 2022. https://www.fda.gov/drugs/developmentapprovalprocess/developmentresources/druginteractionslabeling/ucm093664.htm
- 9.Frost DA, Soric MM, Kaiser R, Neugebauer RE. Efficacy of tramadol for pain management in patients receiving strong cytochrome P450 2D6 inhibitors. Pharmacotherapy. 2019;39(6):724-729. doi: 10.1002/phar.2269 [DOI] [PubMed] [Google Scholar]
- 10.Tirkkonen T, Laine K. Drug interactions with the potential to prevent prodrug activation as a common source of irrational prescribing in hospital inpatients. Clin Pharmacol Ther. 2004;76(6):639-647. doi: 10.1016/j.clpt.2004.08.017 [DOI] [PubMed] [Google Scholar]
- 11.Nahid NA, McDonough CW, Wei YJ, et al. Use of CYP2D6 Inhibitors with CYP2D6 Opioids: Association with Emergency Department Visits for Pain. Clin Pharmacol Ther. 2024;116(4):1005-1012. doi: 10.1002/cpt.3314 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Denny JC, Rutter JL, Goldstein DB, et al. ; All of Us Research Program Investigators . The “All of Us” research program. N Engl J Med. 2019;381(7):668-676. doi: 10.1056/NEJMsr1809937 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61(4):344-349. doi: 10.1016/j.jclinepi.2007.11.008 [DOI] [PubMed] [Google Scholar]
- 14.Hari A, Zhou Q, Gonzaludo N, et al. An efficient genotyper and star-allele caller for pharmacogenomics. Genome Res. 2023;33(1):61-70. doi: 10.1101/gr.277075.122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen X, Shen F, Gonzaludo N, et al. Cyrius: accurate CYP2D6 genotyping using whole-genome sequencing data. Pharmacogenomics J. 2021;21(2):251-261. doi: 10.1038/s41397-020-00205-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lee SB, Shin JY, Kwon NJ, Kim C, Seo JS. ClinPharmSeq: a targeted sequencing panel for clinical pharmacogenetics implementation. PLoS One. 2022;17(7):e0272129. doi: 10.1371/journal.pone.0272129 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Twesigomwe D, Drögemöller BI, Wright GEB, et al. StellarPGx: a Nextflow pipeline for calling star alleles in cytochrome P450 genes. Clin Pharmacol Ther. 2021;110(3):741-749. doi: 10.1002/cpt.2173 [DOI] [PubMed] [Google Scholar]
- 18.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083-3107. doi: 10.1002/sim.3697 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Crews KR, Gaedigk A, Dunnenberger HM, et al. ; Clinical Pharmacogenetics Implementation Consortium . Clinical Pharmacogenetics Implementation Consortium guidelines for cytochrome P450 2D6 genotype and codeine therapy: 2014 update. Clin Pharmacol Ther. 2014;95(4):376-382. doi: 10.1038/clpt.2013.254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Volpe DA, McMahon Tobin GA, Mellon RD, et al. Uniform assessment and ranking of opioid μ receptor binding constants for selected opioid drugs. Regul Toxicol Pharmacol. 2011;59(3):385-390. doi: 10.1016/j.yrtph.2010.12.007 [DOI] [PubMed] [Google Scholar]
- 21.Agema BC, Oosten AW, Sassen SDT, et al. Population pharmacokinetics of oxycodone and metabolites in patients with cancer-related pain. Cancers (Basel). 2021;13(11):2768. doi: 10.3390/cancers13112768 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lalovic B, Kharasch E, Hoffer C, Risler L, Liu-Chen LY, Shen DD. Pharmacokinetics and pharmacodynamics of oral oxycodone in healthy human subjects: role of circulating active metabolites. Clin Pharmacol Ther. 2006;79(5):461-479. doi: 10.1016/j.clpt.2006.01.009 [DOI] [PubMed] [Google Scholar]
- 23.Hemeryck A, Belpaire FM. Selective serotonin reuptake inhibitors and cytochrome P-450 mediated drug-drug interactions: an update. Curr Drug Metab. 2002;3(1):13-37. doi: 10.2174/1389200023338017 [DOI] [PubMed] [Google Scholar]
- 24.O’Brien FE, Dinan TG, Griffin BT, Cryan JF. Interactions between antidepressants and P-glycoprotein at the blood-brain barrier: clinical significance of in vitro and in vivo findings. Br J Pharmacol. 2012;165(2):289-312. doi: 10.1111/j.1476-5381.2011.01557.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Cicali EJ, Elchynski AL, Cook KJ, et al. How to integrate CYP2D6 phenoconversion into clinical pharmacogenetics: a tutorial. Clin Pharmacol Ther. 2021;110(3):677-687. doi: 10.1002/cpt.2354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.GSI: Genotype selection interface. PharmGKB. Accessed May 19, 2025. https://pharmgkb.org/
- 27.Nahid NA, Johnson JA. CYP2D6 pharmacogenetics and phenoconversion in personalized medicine. Expert Opin Drug Metab Toxicol. 2022;18(11):769-785. doi: 10.1080/17425255.2022.2160317 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods. Patient Consent and Data From All of Us Research Program
eTable 1. List of CYP2D6 Strong and Moderate Inhibitors
eTable 2. List of Emergency Department (ED) Visit SNOMED Standard Concept Name and Codes
eTable 3. Characteristics of the Patients Before and After Inverse Probability of Treatment Weighted (IPTW) Balancing
eTable 4. Distribution of the Different Medications in the Study Population
Data Sharing Statement