記住我
Pain experiences can be highly variable between individuals and a standardized approach may not be sufficient, or even safe, for all patients.12,17 Genetic differences affected not only pain perception and efficacy, but also the likelihood to experience adverse events related to analgesic use.26 Pharmacogenetics can help explain these differences and how to tailor treatment accordingly. Cytochrome enzymes can be highly variable, with numerous single nucleotide polymorphisms (SNPs) known to impact response to medication.14,15 Single nucleotide polymorphism variation can result in a variety of metabolic phenotypes, with corresponding pharmacokinetic differences affecting clinical drug effectiveness and safety for individual patients.
Nonsteroidal anti-inflammatory drugs (NSAIDs) and opioids are 2 of the most commonly used categories of pain medications in the world.30 Nonsteroidal anti-inflammatory drugs such as ibuprofen are primarily metabolized by CYP2C9 to inactive metabolites2,6,25; SNPs affecting these genes can reduce ibuprofen clearance.31 Reduced ibuprofen clearance may result in more active serum drug levels which can increase clinical effectiveness but also may increase the risk of adverse events. Oxycodone, a frequently prescribed oral opioid, is known to be primarily metabolized by CYP3A4 and CYP2D6 to metabolites of varying activity.23 CYP3A4 is involved in metabolizing around 50% of currently prescribed medications and is the primary metabolizer of oxycodone.12,47 CYP2D6 metabolizes approximately 25% of all clinically used medications6 and is responsible for conversion of oxycodone to its most active metabolite, oxymorphone.5,29 Allelic variants of CYP2D6 lead to different opioid-related metabolic phenotypes, ranging from poor metabolism to ultrarapid metabolism,6 which affects analgesic effect and risk for opioid toxicity for individual patients.28,29 Unmetabolized oxycodone is responsible for most of its analgesic activity.
This study's goal was to determine how metabolic function of CYP2C9 or CYP3A4/CYP2D6 would affect the clinical effectiveness or safety profile of ibuprofen or oxycodone, respectively. We hypothesized that reduced function of CYP2C9 would lead to increased effectiveness and more adverse events with ibuprofen, reduced functioning of CYP3A4 would lead to increased analgesic effectiveness and increased adverse events, and reduced function of CYP2D6 would lead to reduced effectiveness and decreased adverse events with oxycodone.
2. Methods 2.1. Study design and settingThis study was a pragmatic prospective, observational cohort study conducted in the Emergency Department (ED) of the Stollery Children's Hospital (Edmonton, Alberta, Canada), a tertiary care facility with 163 beds. Ethics approval was obtained through the University of Alberta's Health Research Ethics Board (Pro00005942).
2.2. ParticipantsChildren were eligible if aged 4 to 16 years and being discharged from the ED with an acute, nonoperative fracture and a recommendation to use either ibuprofen or oxycodone for pain management. To be included in the analyses relating genomic associations with clinical outcomes, participants required a DNA sample to be obtained and they must have taken their prescribed medication on day 1 following ED discharge. Exclusion criteria included baseline daily use of medications for chronic pain, being prescribed both ibuprofen and oxycodone on discharge, cognitive impairment preventing self-reporting of pain, inability of family to communicate in English, and no telephone access to the family. The use of adjuvant acetaminophen at the family's discretion, while not explicitly encouraged, was not prohibited for the ethical reason of avoiding untreated pain in participants.
2.3. OutcomesThe primary genetic outcome was the frequency of known CYP2C9, CYP3A4, and CYP2D6 allelic variants relevant to ibuprofen and oxycodone metabolism. The primary clinical outcome measures were postmedication use pain score reduction and occurrence of any adverse event on days 1 to 3 after ED discharge. Pain scores were measured with the Faces Pain Scale-Revised (FPS-R),22 a validated 6-item self-report tool. The following child-reported daily pain score measurements were collected for 3 consecutive days after ED discharge: average pain, minimum pain, maximum pain (that was treated), and postmedication administration pain. For the purposes of this study, pain reduction was defined as the difference between maximal pain and postmedication administration pain on day 1 and was treated as a continuous variable. Adverse event (AE) measurements on days 1 to 3 were categorized as present/absent for occurrence of any adverse event on that day and were coded according to MedDRA (https://www.meddra.org/).24 CYP3A4 and CYP2C9 allelic variants were classified as present or absent. Variants of interest included CYP3A4 *8, *11, *13, and *17; CYP2C9 *1, *2, *3; and *6. CYP2D6 allelic variants were scored, and metabolic phenotype classified based on the previously described and recently updated classification: poor (score 0), intermediate (score 0.25–1), normal (score 1.25–2.25), and ultrarapid (score >2.25) metabolic phenotypes.10
2.4. RecruitmentAll caregivers provided written informed consent and children >6 years, written assent. Trained research assistants confirmed eligibility, collected basic demographic and diagnostic data, completed a medical record review, and procured saliva samples while the patient was in the ED. Oragene (Ottawa, Canada) saliva tube collection kits were used; if the child was unable to produce required saliva volumes spontaneously, saliva sponge swab sticks were used. Saliva samples were transported in preservative and sent by courier to the Canadian Pharmacogenomic Network for Drug Safety (Vancouver, British Columbia) for DNA analyses.9
Participants received a copy of the FPS-R for reference and a logbook on discharge. In the following 3 days, caregivers received daily phone calls during which they were asked to provide their child's self-reported FPS-R pain score for average, maximal, minimal, and postmedication treatment pain. These calls also included questions about home medication use, nonpharmacologic management, adverse events, and functional outcomes. Furthermore, phone calls at 2-week and 6-week postinjury collected details regarding need for medical follow-up.
2.5. Data collection and managementAll data were entered into OpenClinica by a data entry specialist, who validated data on system entry to minimize errors.33 Adverse events data, including side effects, were coded and recorded using MedDRA, a Health Canada recommended terminology for reporting of adverse events (https://www.meddra.org/).24
DNA samples were genotyped for specific gene variants in CYP2C9, CYP3A4, and CYP2D6 with a custom absorption, distribution, metabolism, and excretion (ADME) panel of variants using an Illumina Infinium assay (Illumina, San Diego, CA) and selected CYP2D6 variants by SNaPshot Assay, as previously described.40 The polymorphisms analyzed were CYP2C9 (*1, *2, *3, *5, *6), CYP2D6 (*1, *2, *3, *4, *5, *6, *9, *10, *17, *41), and CYP3A4 (*1B, *2, *3, *4, *7, *8, *11, *13, *14, *16, *17, *18, *19, 11460A>G(K96E), 14313G>A(E174H), 16907T>G(S252A), 20239G>A, 23139T>C (I431T)). Staff-performing DNA analyses were blinded to clinical results.
2.6. Sample size and powerThis paper is an exploratory secondary analysis of data collected for a larger observational cohort study.3 As such, no specific power calculations were conducted for this study.
2.7. Statistical methodsAnalyses were conducted for each medication group separately. Descriptive statistics included mean, SD, counts, and percentages. T tests and χ2 tests assessed mean differences and frequency of AEs between medication groups. Allele-based regression models with only genomic haplotypes as predictors were used to determine if any allele variants within the study population were associated with pain relief or adverse events. For pain reduction, the multiple linear regression model also included variables age, sex, fracture location, fracture reduction, nonpharmacologic management, and ethnicity to provide adjusted estimates. For adverse events, the logistic regression model had variables age, sex, fracture location, sedation in ED, and ethnicity. Linear regression estimates and associated coefficient estimates and standard errors were calculated, as well as odds ratios (ORs) for logistic regression analyses. Statistical analyses were performed in R with missing data removed from tests and regressions; a P-value <0.05 was considered statistically significant.37
3. Results 3.1. Demographic characteristicsOf the 329 children enrolled in the original study,2 a total of 210 (n = 140 ibuprofen and n = 70 oxycodone) had DNA analyses available and were included in this study. Mean age was 11.1 (SD 3.5) years, with 33.8% of the participants being female, and 79.0% self-identifying as Anglo-American (Table 1).
Table 1 - Demographic characteristics (n = 210).* Ibuprofen group (n = 140) Oxycodone group (n = 70) Total cohort (n = 210) Age, mean (SD), y 10.4 (3.7) 12.5 (2.7) 11.1 (3.5) Weight, mean (SD), kg 40.9 (17.2) 48.6 (15.1) 43.5 (16.9) Gender (female), n (%) 56 (40.0) 15 (21.4) 71 (33.8) Origin of family, n (%) Anglo-American 108 (77.9) 58 (82.9) 166 (79.0) Indigenous 6 (4.3) 3 (4.3) 9 (4.3) Asian 6 (4.3) 2 (2.9) 8 (3.8) Hispanic 4 (2.9) 2 (2.9) 6 (2.9) Métis 2 (1.4) 2 (2.9) 4 (1.9) Arab 3 (2.1) 1 (1.4) 4 (1.9) Black 2 (1.4) 1 (1.4) 3 (1.4) East Indian 2 (1.4) 1 (1.4) 3 (1.4) Others† 3 (2.1) 0 (0.0) 3 (1.4) Declined to answer 3 (2.1) 0 (0.0) 3 (1.4) Fracture location, n (%) Upper limb 109 (77.9) 57 (82.6) 166 (79.4) Lower limb 31 (22.1) 12 (17.4) 43 (20.5) Procedural sedation, n (%) 42 (30.0) 24 (34.8) 66 (31.6) Fracture reduction, n (%) 47 (33.6) 25 (36.2) 72 (34.4) Buckle fracture, n (%) 16 (11.4) 2 (2.9) 18 (8.6) Discharge plan, n (%) Follow-up with the orthopedic surgeon 101 (72.1) 56 (80.0) 157 (74.8) Follow-up with the family doctor 12 (8.6) 9 (12.9) 21 (10.0) Referral to the plastic surgeon 4 (2.9) 2 (2.9) 6 (2.9) Return to ED, as needed 2 (1.4) 1 (1.4) 3 (1.4) Return to ED, scheduled 3 (2.1) 0 (0.0) 3 (1.4) Not charted/unable to ask MD 18 (12.9) 2 (2.9) 20 (9.5)*Number of respondents varied per question.
†Includes “African/Anglo-American,” “Anglo-American/Filipino,” and “Anglo-American/Arab.”
ED, emergency department.
Overall, 27.2% (38/140) of children had an allelic variation known to be implicated in clinical outcomes for CYP2C9 and 7.2% (4/57) for CYP3A4, whereas 36.5% (23/63) had “nonnormal” phenotype classification for CYP2D6. CYP2C9 alleles with normal function were found in 72.9% of participants, whereas 26.5% had reduced function, and 0.7% had nonfunctional alleles associated. For CYP3A4, 7.2% of participants' alleles were known to have reduced function; those without variants found were classified as normal functioning wild-type. For CYP2D6, 7.9% of participants were phenotypically classified as poor metabolizers, 28.6% intermediate metabolizers, and 63.5% were normal metabolizers (Table 2).
Table 2 - Distribution of genetic polymorphisms with known functionality. CYP2C9 allele function Allele frequency in the ibuprofen group (n = 140)Based on alleles tested, those without any variants identified were classified to be wild type: *1A. Only alleles with known functionality are presented in this table; thus, proportions are not additive to 100%.
†Based on: Caudle KE, Sangkuhl K, Whirl-Carrillo M, 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;13(1):116–124. doi:10.1111/CTS.12692.Of note, 11 participants had copy number variations in CYP2D6; it was not possible for our laboratory to analyze which allele was duplicated, and thus, they were removed from regression analyses and Table 2, and are described narratively, instead. Six of these patients were ultrarapid metabolizers with phenotypic activity scores >3.0. Three of these patients had activity scores between 2 and 2.5; classified as normal or ultrarapid metabolizers depending on the allele that was duplicated. Two of these patients had activity scores between 1 and 2, classified as intermediate or normal metabolizers depending on the allele that was duplicated.
3.3. Pain managementUsing an “intention to treat” principle, pain scores, including maximum, minimum, average, postmedication treatment pain, and pain reduction on days 1 to 3, were not clinically different between the ibuprofen and oxycodone groups (Table 3).
Table 3 - Median pain scores, days 1–3. Day 1 Day 2 Day 3 Ibuprofen group (n = 132) Oxycodone group (n = 69) Ibuprofen group (n = 124) Oxycodone group (n = 64) Ibuprofen group (n = 102) Oxycodone group (n = 57) Maximum pain*∆ Pain is defined as maximum pain − posttreatment pain.
Pharmacologic pain management use on day 1 revealed exclusive ibuprofen use in 99.3% of the ibuprofen group. For the oxycodone group, 85.7% used an oxycodone-based therapy, with 18.6% using oxycodone monotherapy and 67.1% using a commercially available oxycodone/acetaminophen combination tablet on day 1 (Supplemental Table 1, https://links.lww.com/PR9/A212). Mean dose of administered ibuprofen was 9.2 mg/kg/dose (SD 3.5) on day 1, 9.0 mg/kg (SD 3.8) on day 2, and 8.5 mg/kg (SD 3.8) on day 3; of note, these are all within the recommended dosing range of 5 to 10 mg/kg/dose. Mean dose of administered oxycodone was 0.11 mg/kg/dose (SD 0.05) on day 1, 0.10 mg/kg (SD 0.04) on day 2, and 0.09 mg/kg (SD 0.04) on day 3; these are all within the recommended dosing range of 0.05 to 0.15 mg/kg/dose.
Nonpharmacologic pain management is presented in Table 4. On day 1, 67.9% of the ibuprofen group and 57.1% of the oxycodone group used casts, whereas 15% and 2.9%, respectively, used splints; these were considered therapeutic treatment for their fractures. Other nonpharmacologic strategies such as ice, tensor, and crutches were used as adjuvants.
Table 4 - Nonpharmacological management used (multiple responses permitted).* Ibuprofen group (n = 140)*Number of respondents varied per day.
ED, emergency department.
Table 5 shows cumulative presence of adverse events throughout days 1 to 3. Cumulatively over days 1 to 3, 69.8% (143/208) of all children reported at least one adverse event; specifically, 60.4% (84/139) in the ibuprofen group and 89.3% (59/66) in the oxycodone group (P < 0.001). On day 1, 61.6% (128/208) of all children reported at least one adverse event; specifically, 52.9% (74/140) in the ibuprofen group and 78.3% (54/69) in the oxycodone group (P < 0.001). The ibuprofen group experienced significantly less appetite loss (P = 0.04), constipation (P = 0.002), dizziness (P < 0.001), drowsiness (P = 0.001), and nausea (P < 0.001) than the oxycodone group. On day 2, 42.7% (88/206) of all children reported adverse events, with 35.7% (50/140) in the ibuprofen group and 57.6% (38/66) in the oxycodone group (P = 0.003). On day 3, 27.0% (55/204) of all children reported adverse events, with 22.1% (31/140) in the ibuprofen group and 32.8% (21/64) in the oxycodone group (P = 0.11) (Supplemental Table 2, https://links.lww.com/PR9/A212).
Table 5 - Cumulative adverse events per patient by the treatment group.* Ibuprofen*Number of respondents varied per day.
AE, adverse events.
For the ibuprofen group, children who were carriers of the CYP2C9*2 reduced function allelic variant were less likely to experience an adverse event (OR = 0.72, 95% confidence interval [CI] [0.58, 0.89]), compared with children who were carriers of the normal or wildtype CYP2C9*1 allele. In the ibuprofen group, the presence of reduced function CYP2C9*2, CYP2C9*3, and nonfunctional CYP2C9*6 did not influence pain reduction on day 1. For the oxycodone group, no influence was found on pain relief or adverse events, based on CYP3A4 variants and CYP2D6 phenotype classification (Supplemental Tables 3a and 3b, https://links.lww.com/PR9/A212).
3.6. Multivariable analysesFor the ibuprofen group, multivariable logistic regression analyses were conducted for pain reduction on day 1, including the independent clinical variables of age, sex, fracture location, fracture reduction, nonpharmacologic management (ie, ice, elevation, and tensor bandage), ethnicity, and CYP2C9*2, *3, and *6 variants; none demonstrated statistical significance.
For the oxycodone group, linear regression analysis for pain reduction on day 1 included the same clinical variables and CYP2D6 intermediate and extensive metabolizer phenotypes. The use of nonpharmacologic pain management strategies (P = 0.02) was associated with less pain relief for children using oxycodone. Linear regression analysis for pain reduction on day 1 was also performed for CYP3A4, with the same clinical variables; nonpharmacologic pain management strategies (P = 0.006) were associated with less pain relief when using oxycodone.
For the ibuprofen group, multivariable logistic regression analysis for adverse events on day 1 included age, sex, fracture location, procedural sedation in ED, and ethnicity, along with CYP2C9*2, *3, and *6 variants. Children with the CYP2C9*2 allele were less likely to experience adverse events (OR = 0.72, 95% CI [0.58, 0.90]) compared with those with the CYP2C9*1 (wildtype) allele. For the oxycodone group CYP2D6 analysis, the same clinical variables were used, plus intermediate and extensive metabolizer phenotypes were included in the model (Supplemental Tables 4a and 4b, https://links.lww.com/PR9/A212).
3.7. Case studies of CYD2D6 poor metabolismFour individuals who used oxycodone were found to be poor metabolizers, all self-identified as Anglo-American.10,13 One individual had a maximum pain of 4, which reduced to a 0 with oxycodone use; no adverse events were experienced. The next individual had maximum pain of 4, which reduced to a 2; constipation, drowsiness, nausea, and vomiting were reported. The next had maximum pain of 8, which reduced to a 4; drowsiness, nausea, and vomiting were reported. The last selected individual had maximum pain of 6, which reduced to a 2; they reported appetite loss, constipation, drowsiness, dizziness, nausea, and rash.
3.8. Case studies of possible CYD2D6 ultrarapid metabolismThree individuals who used oxycodone were found to have potential ultrarapid metabolism, based on the presence of copy number variations for *1 and *41: 2 self-identified as Anglo-American and 1 as Indigenous. One individual had a maximum pain of 8, which reduced to a 0; adverse events included dizziness and drowsiness. The next individual had a maximum pain of 6, which reduced to 2; appetite loss and nausea were reported. The last had a maximum pain of 10, which reduced to a 6; no adverse events were experienced.
4. DiscussionJust over one quarter of children had an allelic variation known to be implicated in clinical outcomes for CYP2C9, whereas over one-third had “nonnormal” phenotype classification for CYP2D6. Decreased CYP2C9 function alleles (*2 and *3) were found in over one quarter of individuals. No impact on clinical analgesic effectiveness was found with ibuprofen use, but children who were carriers of *2 experienced less adverse events with ibuprofen; this contrasted our original hypothesis that decreased clearance would lead to higher concentrations of ibuprofen, predisposing to adverse events. Decreased CYP3A4 functioning alleles were found in 3.4% of individuals, whereas 36.5% had a decreased functioning phenotype for CYP2D6. We did not find an effect on analgesic effect or occurrence of adverse events with oxycodone use.
Different alleles implicated in drug metabolism can predispose individuals to faster or slower metabolism of various analgesics.5,26 Variants may lead to decreased medication clearance and increase risk of drug toxicity or other adverse events,44 such as gastric bleeding after NSAID use25; conversely, for prodrugs requiring activation, poor metabolism may lead to insufficient pain relief.4 The CYP2D6 normal metabolizer phenotype is known to be most common.6 Poor metabolizer status is more common among those of Anglo-American descent, with about 10% prevalence11; our study demonstrated 7.5% prevalence. Conversely, ultrarapid metabolism is uncommon in those of Anglo-American descent but more prevalent in East Africa, found in up to 29% of those with East African descent.26,41 Our study, representing a mostly Anglo-American population, supports these prior characterizations, with our participants found to be mostly CYP2D6 normal metabolizers but also having some representation of poor and intermediate metabolism. Although some prior studies have shown that oxycodone metabolite contribution to overall analgesic effect is affected by CYP2D6 phenotype,38 other studies have suggested that most analgesic effect comes from the parent compound oxycodone, rather than metabolites that are found in much lower concentrations.27,29 Our findings of no impact of CYP2D6 phenotype classification on clinical effect is congruent with the latter. Thus, our small study would imply that a child's CYP2D6 phenotype classification may not influence their response to oxycodone. This must be confirmed with larger scale studies and randomized allocation of participants.
CYP2C9 is less extensively characterized than CYP2D6, with allelic variants' function categorized as no function, decreased function, normal function, or unknown function based on currently described variants catalogued in genomic databases.8,16 CYP2C9*2 and *3 alleles are associated with poor metabolism and reduced clearance of NSAIDs and are known to be more common in those of Anglo-American descent,12,44 affecting up to 20%21; this slow clearance may put patients at higher risk for adverse events from ibuprofen.31 Our study reaffirmed this distribution, with slightly over 25% of our mostly Anglo-American population having those decreased functioning alleles. Prior study of CYP2C9 in adults has demonstrated that those with decreased functioning alleles are at an increased risk for developing acute gastrointestinal bleeding after taking ibuprofen and other NSAIDs.1,32 Furthermore, both the *2 and *3 have been previously demonstrated in adults to be associated with up to a 10-fold decrease in ibuprofen clearance (and thus increased clinical pain relief).18 By contrast, our study demonstrated no impact of CYP2C9 allelic variations on clinical analgesic effectiveness, and children who were carriers of *2 experienced less adverse events with ibuprofen, instead of the expected increase, based on adult studies. These findings merit further exploration with a larger cohort of participants as it is possible that these results are due to our smaller sample size or, perhaps, that children's metabolism and clinical effects of ibuprofen are influenced by more than pharmacogenetics, alone.
CYP3A4 serves as the primary metabolizer of oxycodone, converting it to noroxycodone.23 Allelic variation has been noted in particular genetic ancestry groups, with variants such as CYP3A4*1B—which may have increased enzyme activity35—found in 2% to 10% of Anglo-Americans, 9% to 11% of Hispanics, and 35% to 67% of African-Americans populations35; increased enzyme activity may lead to less analgesia as more oxycodone is converted to its inactive metabolite, noroxycodone.23 Our population had no participants with increased functioning alleles; rather, 4 participants had a reduced functioning allele. No clinical correlations with clinical effectiveness or adverse events were noted in our small study cohort who received oxycodone, precluding definitive conclusions in this regard.
A recent large translational study of adults has shown that providing adults with a QR code that stored the participants' test results for a 12 gene–drug interaction panel resulted in a 7% lower incidence of adverse events whenever a new medication was prescribed.42 Although personalized medicine based in pharmacogenomic profiles has clear clinical considerations for some cardiac medications (clopidogrel and warfarin)7,39 and chemotherapy (tamoxifen),
留言 (0)