Integrating Pharmacogenomic Results and Drug-Induced Phenoconversion Concepts while Conducting a Medication Safety Review Process

Conducting Abstract Opioid appropriateness is primarily focused on insufficient pain control; however, opioid-related adverse drug events (ADE) have also been associated with a negative impact on disability, mood, and quality of life. As progress is made toward determining more appropriate parameters for opioid prescribing, clinicians are beginning to rely on pharmacogenomics (PGx). PGx results can provide valuable patient-specific information about risk for opioid-related ADEs and/or pharmacotherapy failure. The clinical utility of PGx has been associated with limitations; however, considering drug-induced phenoconversion can be a strategy to address limitations in patients with complex drug regimens. This case highlights the importance of considering drug-induced phenoconversion when implementing PGx into clinical practice. This case describes a wheelchair-bound patient who experienced intolerable drowsiness due to oxycodone which was unsuccessful in relieving trigeminal neuralgia-related pain. A clinical pharmacist performed a PGx-informed medication safety review and identified drug interactions causing competitive inhibition of CYP3A4 and drug-drug-gene interactions and drug-induced phenoconversion for CYP2D6 – all of which affected oxycodone metabolism into oxymorphone. This resulted in an increased risk for oxycodone-related ADEs and decreased efficacy. The pharmacist recommended oxycodone discontinuation, optimization of gabapentin, awareness for alternative therapy based on HLA-A PGx results, and patient counseling. Pursuant to therapy changes, the patient experienced better pain control and less drowsiness, which improved her quality of life. This report demonstrates how a PGx-informed medication safety review that assessed drug-drug and drug-drug-gene interactions resulted in better pain management, elimination of opioid-related ADEs, and improved quality of life.


Introduction
The number of opioids prescriptions for pain treatment in the United States in 2011 was 219 million [1,2]. Fortunately, opioid prescriptions have steadily declined to 142 million in 2020 [1]. The belief that opioids are one of the best therapeutic approaches for pain management inhibits further decline. Opioids are not appropriate for all patients and complete pain relief is unattainable for many people. Several factors affect patient response to opioids and should be considered when assessing opioid appropriateness [3]. Pharmacokinetic (PK) and pharmacodynamic (PD) factors contribute to the inter-subject variability observed in opioid response, especially for opioid prodrugs that are bioactivated by the cytochrome (CYP) 2D6 isoenzyme to produce more potent analgesic metabolites (e.g., codeine, tramadol, oxycodone) [4,5]. Administration of other medications with CYP2D6-activated opioids can result in PK or PD interactions, potentially increasing the risk for opioid-related adverse drug events (ADEs) and/or pharmacotherapy failure [4,6]. Given the polymorphic nature of CYP2D6 and the opioid mu receptor, it is assumed that significant variation in PK or PD exists among patients taking opioids. Accordingly, PGx studies have shown correlation with interindividual differences regarding opioid effectiveness [7,8].
Several genetic testing companies have developed PGx panels to help clinicians address inter-subject variability in opioid response; however, the clinical utility of PGx testing has been somewhat controversial in a "real-world setting". Complexities and limitations arise when interpreting PGx results for patients taking multiple medications that can alter the genotypepredicted phenotype this phenomenon is known as drug-induced phenoconversion [5,9]. Access to clinical decision support systems (CDSS) that integrate a patient's medication regimen and PGx results can help pharmacists identify and mitigate medicationrelated problems (e.g., drug-induced phenoconversion) by incorporating these factors into a medication safety review [10][11][12].

Case Report
73-year-old female patient with multiple comorbidities (Table 1) and a partial left leg amputation presented to her provider for a routine visit. She had a history of chronic neuropathic pain treated with gabapentin and oxycodone/acetaminophen. The patient's primary complaint was excessive daytime sedation, which was heavily affecting her quality of life. She recently reported experiencing shock-like pains on the right side of the face often triggered by eating breakfast or brushing her teeth in the morning. The provider identified that the patient's symptoms were due to trigeminal neuralgia (TGN). Pain appeared uncontrolled to a satisfactory level despite already taking pain medication, including oxycodone/acetaminophen 1-2 times daily. The patient reported her pain to be a 5 out of 10 on average using a numeric rating scale (NRS). A PGx test was proposed to help optimize patient's therapy. Results of the PGx test (Table 2) were assessed against the patient's medication regimen (Table 1), aided with an advanced CDSS [(MedWise® (Figure 1)]. The PGx results identified the patient as a CYP2D6 normal metabolizer (NM) suggesting that a normal response to oxycodone was expected based her genotype results (CYP2D6*1*2A). However, the pharmacist noted that there was a drug-drug-gene interaction (DDGI) with duloxetine, resulting in phenoconversion to a CYP2D6 intermediate metabolizer (IM) for oxycodone [5,13]. This DDGI could result in reduced formation of the more potent metabolite, oxymorphone, and could increase the risk of pharmacotherapy failure and may contribute to the patient's inadequate pain management [5]. In addition, several drug-drug interactions (DDIs), particularly with oxycodone, were identified ( Figure 1), including competitive inhibition at CYP3A4 due to the coadministration of amlodipine and atorvastatin [14,15]. When these two drugs are coadministered with the weaker affinity substrate oxycodone, concentrations of oxycodone would increase, thus increasing the risk for oxycodone-related ADEs (e.g., drowsiness, headaches) [16]. CYP2D6 DDGI, taken together, were deemed associated with suboptimal pain management while CYP3A4/5 DDI could explain daytime sedation.  As the patient's pain symptoms were related to TGN, alternative medications were evaluated considering the patient's PGx results, renal function, and potential for drug-gene interactions (DGIs) and DDGIs (Table 3) [17,18,21]. As the patient was prescribed gabapentin 100mg three times a day for neuropathy, the clinical pharmacist recommended optimizing the morning dose to 200mg [17,18,22]. In addition, it was suggested to discontinue oxycodone/acetaminophen. Recommendations and awareness are described in Table 3.
Approximately 2 weeks after implementing these drug regimen modifications, the patient reported feeling significantly less sedated with better pain control (NRS = 0). The patient's quality of life was improved so significantly that she began considering pursuing prosthetic implantation in hopes of walking again.

Discussion
The simultaneous management of pain and multiple comorbidities can present clinicians with challenges. As presented in this case, pain can lead to reduced physical activity, decreased energy level, and depression; and depression can lead to more pain. Furthermore, side effects associated with medicine used to control these conditions may lead to daytime sedation, worsening lethargy, energy, and physical activity. This describes the often-observed prescription cascade: a medication prescribed to treat one condition worsens another condition or causes a new one, leading to the introduction of another drug [23]. This patient case is an example of how incorporating PGx information with druginduced phenoconversion can successfully improve PGx result interpretation in the presence of polypharmacy. This case report illustrates that integrating PGx can help determine appropriate pain management therapy, mitigate opioid use, identify oxycodonerelated side effects, and improve patient quality of life.
In this case study, the patient's pain was inadequately controlled despite being prescribed pain analgesic oxycodone/ acetaminophen. Oxycodone is primarily metabolized by the CYP3A4 enzyme into an inactive metabolite, noroxycodone, while ~10% is metabolized by the CYP2D6 enzyme into a more potent metabolite, oxymorphone [24][25][26]. To exert their effects, both oxycodone and oxymorphone bind and activate muopioid receptors, located in several organs including the brain. Oxymorphone has up to 60 times stronger affinity than oxycodone [27]. Higher concentrations of oxycodone, a weak CYP3A4 affinity substrate, were predicted for this patient due to DDIs and competitive inhibition from the concomitant administration of the CYP3A4 moderate-affinity substrates, atorvastatin and amlodipine ( Figure 1) [5,14,15]. The patient's excessive sedation and headaches were likely worsened by increased oxycodone levels, negatively affecting her quality of life [16,28,29]. The utilization of PGx testing and CYP2D6 results to inform personalized treatment plans has proven to help overcome the inappropriate prescribing of opioids [7]. Although the Clinical Pharmacogenetics Implementation Consortium do not provide CYP2D6-specific therapeutic recommendations for oxycodone, substantial evidence supports the relationship between oxymorphone concentrations and analgesic effects [16,28,30]. A recent review highlighted the reasons for some conflicting results between CYP2D6 polymorphisms and oxycodone response [5]. Among those, the authors highlighted that 1) CYP2D6 is expressed in the liver and in the brain; 2) local formation of oxymorphone in the brain is possibly independent of liver activity; 3) geneticallydetermined poor metabolizers (PM) of CYP2D6 cannot form oxymorphone in the liver and brain; 4) CYP2D6 PMs do not exhibit pain control when administered oxycodone; 5) CYP2D6 inhibition (leading to phenoconversion) could be observed in the liver only or in the liver and in the brain, depending on the capability of the perpetrator drug to cross the blood-brain barrier; 6) thus, under certain conditions, oxymorphone formation could be prevented in the liver leading to low circulating plasma levels while its formation could still occur in the brain [5].
In order for complete phenoconversion to occur (decreased oxycodone efficacy), perpetrating medications must also be able to cross the blood brain barrier [5,29]. Such condition is observed in the current case, as duloxetine, a CYP2D6 moderate-affinity substrate, must accumulate in the brain to exert its effects; therefore, when administered with the CYP2D6 weak-affinity substrate oxycodone, duloxetine can induce a phenoconversion in the brain of CYP2D6 NM to IM [5,13]. This DDGI results in a decreased transformation of the parent drug, oxycodone, into the more potent metabolite, oxymorphone, both in the liver (systemically) and in the brain (locally) [5,27,29].
To address the patient's pain and the expected increased risk for oxycodone-related ADEs and pharmacotherapy failure, alternatives were considered and analyzed for the TGN symptoms. TGN is a rare, chronic pain condition due to nerve dysfunction of the trigeminal nerve [17]. Pharmacological treatment options for improving pain symptoms include several anticonvulsants, with carbamazepine or oxcarbazepine being first-line treatment [17,18,21]. Based on the patient's PGx results (positive for the HLA-A *31:01 allele), the first-line options carbamazepine and oxcarbazepine were excluded due to the increased risk for severe life-threatening cutaneous reactions (e.g., Stevens Johnson syndrome) [31,32]. In addition, these medications possess sedative properties, and drug monitoring would have been necessary since DDIs and DDGIs were expected (Figure 2, Other second-line TGN treatments were evaluated for appropriateness, including phenytoin and topiramate [17,18]. Given the patient's CYP2C9 and HLA-B*15:02 results, no DGIs affecting phenytoin and topiramate were identified (Table 2) [36]. However, if either one of these medications were included in the medication regimen, they would cause several DDIs due to induction of CYP2B6 and CYP2C9 by phenytoin, and of CYP3A4 by phenytoin and topiramate ( Figure 2, Table 3) [37,38]. Of the remaining medications evaluated, gabapentin does not utilize CYP enzymes for metabolism, thus would not contribute to CYP DDIs [22]. It was decided to optimize gabapentin instead of introducing another medication, as the current dose was sub-optimal for TGN and was below the 900mg daily maximum dose appropriate for the patient's creatinine clearance of 45.9ml/min. [18,22]. • DDI: Oxcarbazepine inhibits CYP2C19 metabolism, which will result in phenoconversion to a CYP2C19 PM for pantoprazole and clopidogrel. An alternative antiplatelet therapy will need to be considered if used. Phenytoin (2 nd line) • DGI: Patient is negative for HLA-B*15:02 and is expected to have a typical risk for hypersensitivity reactions with phenytoin. Limited evidence available to suggest phenytoin cannot be used with those HLA-A*31:01 positive.
• DDI: Phenytoin induces CYP2B6, CYP2C9, and CYP3A4 metabolism, which will result in decreased concentrations of amlodipine and atorvastatin (monitoring closely for effectiveness will be recommended) and increased activation of clopidogrel. Topiramate (2 nd line) • No DGIs • DDI: Induces CYP3A4 metabolism, which will result in decreased concentrations of amlodipine and atorvastatin (monitoring closely for effectiveness will be recommended) and increased activation of clopidogrel. Baclofen and lamotrigine (2 nd line) • Competitive inhibition at CYP3A4 (amlodipine and atorvastatin); increasing risk of pharmacotherapy failure.
• Dual antiplatelet therapy has exceeded 1 year of treatment.
Consider re-evaluating benefits/risks of dual antiplatelet therapy, given a higher risk of bleeding with long term use of dual antiplatelet therapy. If deemed clinically appropriate, consider stopping the clopidogrel and continuing aspirin 81mg.
Alternatively, if taking clopidogrel in the morning, consider changing the time of administration of amlodipine and atorvastatin to the evening to avoid competitive inhibition and favor formation of clopidogrel active metabolite.

•
High risk for ADEs in the elderly population.
If efficacy achieved and dual antiplatelet therapy is no longer needed, consider a gradual discontinuation of pantoprazole, monitoring for continued efficacy.
If dual antiplatelet therapy to be continued and/or heart burn symptoms still present, continue to monitor for ADEs associated with long-term use of pantoprazole.

Melatonin 3mg
• Competitive inhibition at CYP1A2 with duloxetine, increasing risk of over-sedation.

•
Due to desensitization of receptors, higher doses do not translate to increased benefit.
A deprescribing strategy was considered to mitigate oversleepiness due to CYP1A2 DDI, as duloxetine can cause sleepiness and act as a competitive inhibitor of melatonin metabolism, a weak CYP1A2 affinity substrate ( Figure 1, Table 3) [39]. Ultimately, the provider increased the morning dose of gabapentin to 200mg (400mg daily) and discontinued melatonin and oxycodone/ acetaminophen; close monitoring was continued for pantoprazole and clopidogrel.
After adjusting the patient's medication regimen, her pain was controlled (NRS = 0) and her quality of life significantly improved, which initiated her pursuit to walk again. Lemke et al. reported that incorporating and discussing PGx findings offers reassurance and psychological benefit to the patients [40]. Given that some patients are cautious and sometimes hesitant to discontinue or change their medication, particularly opioids, presenting them with PGx-driven personalized assessments can help earn the patients' trust.

Conclusions
Employing a personalized, multimodal approach to pain management demonstrates promise in implementing changes regarding the overuse and misuse of opioids. This case demonstrates that in the context of polypharmacy, integration of pharmacogenomic information into a decision-making process is successful when concomitant medications are considered. This resulted in improvement in medication safety, in the patientprovider relationship, and in the patient's overall quality of life as she strives to walk again.