Abstract

Awareness of drug interactions involving opioids is critical for patient treatment as they are common therapeutics used in numerous care settings, including both chronic and disease-related pain. Not only do opioids have narrow therapeutic indexes and are extensively used, but they have the potential to cause severe toxicity. Opioids are the classical pain treatment for patients who suffer from moderate to severe pain. More importantly, opioids are often prescribed in combination with multiple other drugs, especially in patient populations who typically are prescribed a large drug regimen. This review focuses on the current knowledge of common opioid drug–drug interactions (DDIs), focusing specifically on hydrocodone, oxycodone, and morphine DDIs. The DDIs covered in this review include pharmacokinetic DDI arising from enzyme inhibition or induction, primarily due to inhibition of cytochrome p450 enzymes (CYPs). However, opioids such as morphine are metabolized by uridine-5’-diphosphoglucuronosyltransferases (UGTs), principally UGT2B7, and glucuronidation is another important pathway for opioid-drug interactions. This review also covers several pharmacodynamic DDI studies as well as the basics of CYP and UGT metabolism, including detailed opioid metabolism and the potential involvement of metabolizing enzyme gene variation in DDI. Based upon the current literature, further studies are needed to fully investigate and describe the DDI potential with opioids in pain and related disease settings to improve clinical outcomes for patients.

SIGNIFICANCE STATEMENT A review of the literature focusing on drug–drug interactions involving opioids is important because they can be toxic and potentially lethal, occurring through pharmacodynamic interactions as well as pharmacokinetic interactions occurring through inhibition or induction of drug metabolism.

Introduction

Opioid analgesics are widely used in clinical practice for a wide variety of pain management plans for both chronic and acute pain. It is estimated that, in the United States, 50 million adults suffer from chronic pain and 20 million adults have high impact chronic pain (Dahlhamer et al., 2018; Zelaya et al., 2020; Yong et al., 2022), with chronic pain prevalence in older populations higher as compared to younger populations (Johannes et al., 2010; Larsson et al., 2017; Dahlhamer et al., 2018; Zelaya et al., 2020). Although there are numerous treatment options for chronic pain, it is estimated that over 8 million people use opioids for long-term chronic pain management (Reuben et al., 2015; Dowell et al., 2016; Bohnert et al., 2018; Hales et al., 2020; https://www.cdc.gov/drugoverdose/rxrate-maps/state2020.html; Dowell et al., 2022). The cost of chronic pain has been estimated to be $635 billion in annual medical costs, disability, and loss of productivity (Institute of Medicine (US) Committee on Advancing Pain Research, 2011). Hydrocodone, oxycodone, and morphine are among the most widely prescribed or used opioid analgesics (https://nida.nih.gov/publications/drugfacts/prescription-opioids; https://www.cdc.gov/opioids/basics/index.html). All three opioids have been used in the clinic for decades and many studies have been performed focusing on the efficacy of their analgesic properties and side effects, as well as on their metabolism both in vitro and in vivo (Cone and Darwin, 1978; Otton et al., 1993; Coffman et al., 1997; Kaplan et al., 1997; Coffman et al., 1998; Stone et al., 2003; Hutchinson et al., 2004; Lalovic et al., 2004; Adams and Ahdieh, 2005; Lalovic et al., 2006; Kapil et al., 2015). From these studies, pharmacokinetic and pharmacodynamic profiles of these opioids can be gleaned to provide their best efficacy in clinical practice.

A drug’s pharmacokinetic and pharmacodynamic profile can be altered by polypharmacy - i.e., the concomitant use of more than one drug –which can lead to potentially harmful drug–drug interactions (DDI). Such alterations in drug pharmacokinetics or pharmacodynamics can lead to adverse drug events (ADE) that can alter drug efficacy and/or toxicity. Pharmacodynamic DDIs can lead to either enhanced or decreased pharmacological action of the object drug (Overholser and Foster, 2011; Niu et al., 2019). Pharmacokinetic DDIs lead to changes in bioavailability and altered production of active or inactive metabolites (Smith, 2009).

Opioids are central nervous system (CNS) depressants and act upon opioid receptors (mu, delta, and kappa) that are found in the brain, spinal cord, and the periphery (Hersh et al., 2007). Most commonly, opioids have a higher affinity for the mu opioid receptor (MOR) (Theriot et al., 2023). MOR subtypes arise due to splice variants; MOR 1 is associated with analgesia and dependence; MOR 2 is associated with respiratory depression, miosis, and constipation; and MOR 3 is associated with vasodilation (Trescot et al., 2008; Valentino and Volkow, 2018; Dhaliwal and Gupta, 2023). When drugs with similar pharmacological actions are taken together this can lead to pharmacodynamic DDI potentially causing ADE. For example, taking two drugs that both cause depression of the CNS can cause respiratory depression and even death.

An important mechanism underlying pharmacokinetic DDI includes the interaction of a precipitant drug with the metabolizing enzymes that catalyze the biotransformation of the object drug (Smith, 2009). Metabolic enzymes are divided into phase I and phase II metabolism, with the cytochrome P450 enzymes (CYP) the major superfamily within the phase I metabolic enzymes and the UDP-glucuronosyltransferases (UGT) the major phase II enzyme superfamily. Most opioids are metabolized by both the CYP and UGT family of enzymes (Trescot et al., 2008), with several, including codeine, hydrocodone, oxycodone, tramadol, and morphine, metabolized primarily by CYP3A4, CYP2D6 and UGT2B7 (Cone et al., 1978; Yue et al., 1991a,b; Otton et al., 1993; Pöyhiä et al., 1993; Hagen et al., 1995; Caraco et al., 1996a,b; Coffman et al., 1997; Paar et al., 1997; Coffman et al., 1998; Green et al., 1998; Subrahmanyam et al., 2001; Donnelly et al., 2002; Shapiro and Shear, 2002; Zheng et al., 2002; Stone et al., 2003; Benetton et al., 2004; Hutchinson et al., 2004; Lalovic et al., 2004; Zheng et al., 2004; Baldacci and Thormann, 2006; Lalovic et al., 2006; Madadi and Koren, 2008; Ohno et al., 2008; Jenkins et al., 2009; Nieminen et al., 2009; Cone et al., 2013a,b; Barakat et al., 2014; Elder et al., 2014; Kurogi et al., 2014; DePriest et al., 2016a,b; Romand et al., 2017; Shen et al., 2019). Approximately 25% and 50% of all pharmaceutical drugs are substrates of CYP2D6 and CYP3A4, respectively (Bertz and Granneman, 1997; Evans and Relling, 1999; Ingelman-Sundberg, 2005; Ingelman-Sundberg and Rodriguez-Antona, 2005; Trescot et al., 2008; Zhou, 2009; Zanger and Schwab, 2013), while UGT2B7 is a major UGT isoform that is responsible for catalyzing the biotransformation of numerous xenobiotics and endogenous substances (Bhasker et al., 2000; Tukey and Strassburg, 2000; Stingl et al., 2014; Shen et al., 2019). Since pharmacokinetic DDI can occur by dysregulated drug metabolism, it is important to study potential DDI to characterize the severity of the drug reaction for clinical practice.

This review focuses on three commonly prescribed opioids in the non-cancer chronic pain setting, hydrocodone, oxycodone and morphine, and their respective metabolism and known DDIs. Opioid metabolism is summarized along with the implications of potential DDI due to changes in opioid pharmacokinetics and metabolism. This review includes a description of known associated DDIs and their adverse effects on metabolic pathways and patient response and toxicity. The pharmaceutical drugs identified in this review as altering opioid metabolism can be condensed into general treatment categories. The most common treatments conferred by the drugs that lead to pharmacokinetic DDIs with opioids are antifungals, antibiotics, antivirals, anti-depressants/anti-psychotics, chemotherapeutic/anti-cancer, anticonvulsant, and sedatives.

Furthermore, the classes of drugs identified in this review involved in these DDIs include azoles, protease inhibitors, pharmacokinetic enhancers, non-nucleoside reverse transcriptase inhibitors, quinolones, macrolides, chloramphenicol, ketolides, antimycobacterial, monoamine oxidase inhibitors, selective serotonin reuptake inhibitors, uni- and tricyclic antidepressants, serotonin-norepinephrine reuptake inhibitors, butyrophenones, phenothiazines, atypical second generation antipsychotics, iminostilbenes, benzodiazepines, hydantoins, barbiturates, kinase inhibitors, antiestrogens, poly (ADP-ribose) polymerase inhibitors, isocitrate dehydrogenase-1 inhibitors, among others. Depending on the patient and diagnosis, it is possible that multiples of these precipitant drugs and opioids could be co-prescribed together. Therefore, in addition to patients treated for pain, these DDIs can potentially occur in a wide variety of patient populations outside of typical pain patients.

Methods

Relevant literature was reviewed using PubMed (last searched on December 12, 2022) and the search terms ‘hydrocodone’, ‘morphine’, and ‘oxycodone’. Each search term was also combined with the following additional search terms, ‘metabolism,’ ‘active metabolites,’ ‘drug interaction,’ and ‘drug–drug interaction.’ Searches were filtered for language (English, and text type: free full text), with all studies involving pharmacokinetic DDIs, pharmacodynamic DDIs, and clinical DDIs included in the review. The following studies were excluded from the results: studies showing neutral DDIs, studies in which the opioid was the precipitant drug causing a DDIs, and transporter-related DDIs. These search parameters were used to compile a comprehensive in vitro and in vivo dataset to summarize both pharmacodynamic and pharmacokinetic drug–drug interactions involving the commonly prescribed opioids, hydrocodone, oxycodone, and morphine.

ResultsMetabolism and Activity of Hydrocodone, Oxycodone, and Morphine and their MetabolitesHydrocodone Metabolism

Greater than 50% of the total hydrocodone dose is metabolized by CYP-mediated phase 1 metabolism (Cone and Darwin, 1978; Cone et al., 1978; Park et al., 1982; Otton et al., 1993; Hutchinson et al., 2004; Barakat et al., 2012; Valtier and Bebarta, 2012). Hydrocodone is a substrate of both CYP2D6 and CYP3A4/5, with it metabolized by CYP2D6 via O-demethylation to its active metabolite, hydromorphone (Otton et al., 1993; Hutchinson et al., 2004), and by CYP3A4 through N-demethylation to a major inactive metabolite, norhydrocodone (see Fig. 1) (Hutchinson et al., 2004). As CYP3A4 and CYP3A5 are highly homologous and metabolize the same substrates (Guengerich, 2005; Daly, 2006; Tseng et al., 2014), the contribution of CYP3A5 to hydrocodone metabolism is still unclear. Hydrocodone, hydromorphone, and norhydrocodone undergo further metabolism by glucuronidation and reduction to minor metabolites (Fig. 1) (Cone and Darwin, 1978; Cone et al., 1978). Hydrocodone has been suggested to be a prodrug in some instances, as the active metabolite, hydromorphone, exhibits a greater analgesic effect than its parent molecule (Trescot et al., 2008). However, previous studies have shown that in the absence of CYP2D6, hydrocodone dosing still elicits analgesic activity, suggesting that hydrocodone has its own analgesic properties (Kaplan et al., 1997; Tomkins et al., 1997). Even though hydromorphone has been shown to have 10- to 33-fold higher affinity to mu-opioid receptors (Hennies et al., 1988; Chen et al., 1991) and greater analgesic potency than morphine when administered subcutaneously (Jaffe JH, 1990), individuals who are poor metabolizers (PM) for CYP2D6 do not exhibit different responses compared to extensive metabolizers (EM) with equivalent hydrocodone dosing (Kaplan et al., 1997; Kapil et al., 2015). This is likely explained by the relatively low abundance of hydromorphone compared to the parent drug in the plasma of individuals taking hydrocodone [with observed plasma levels of hydromorphone at ∼3%–5% of the hydrocodone dose (Cone and Darwin, 1978; Cone et al., 1978; Coller et al., 2009; Hao et al., 2011; Valtier and Bebarta, 2012; Langman et al., 2013; Darwish et al., 2015; Kapil et al., 2015)] and possibly due to the higher rate at which hydrocodone enters the brain as compared to hydromorphone (Kaplan et al., 1997; Schaefer et al., 2017) .

Fig. 1.

Schematic of hydrocodone metabolism.

Hydromorphone is further metabolized through reduction to 6α-and 6β-hydromorphol which undergoes subsequent glucuronidation by UGT2B7 (Coffman et al., 1998) to form hydromorphone-3-glucuronide (Cone et al., 1978; Wright et al., 2001; Trescot et al., 2008) (Fig. 1). Although hydromorphone has low blood brain barrier (BBB) penetration, it is also given as a stand-alone analgesic because of its high opioid effects (Hagen et al., 1995; Wright et al., 2001; Drewes et al., 2013; Landolf et al., 2020).

Oxycodone Metabolism

The metabolism of oxycodone is similar to that of hydrocodone. The primary oxidative pathway for oxycodone is by N-demethylation by CYP3A4/5 to form noroxycodone, an inactive metabolite [see Fig. 2 (Weinstein and Gaylord, 1979; Lalovic et al., 2004, 2006)]. Lalovic et al., found that CYP3A5 is active in oxycodone metabolism (Lalovic et al., 2004), which is consistent with many previous studies demonstrating that most CYP3A4 substrates may also be metabolized by CYP3A5 due to their high sequence homology (Guengerich, 2005; Daly, 2006; Tseng et al., 2014). However, the actual contribution of CYP3A5 (vs. CYP3A4) to oxycodone metabolism has not been firmly established (Naito et al., 2011; Tseng et al., 2014). Oxycodone can also be O-demethylated to oxymorphone, an active metabolite, via CYP2D6 (Otton et al., 1993). Noroxycodone has weak antinociceptive effects and is thus considered inactive in comparison to oxymorphone (Weinstein and Gaylord, 1979; Leow and Smith, 1994; Stamer et al., 2013). Oxymorphone and noroxycodone are further metabolized to noroxymorphone by CYP3A4/5 and CYP2D6, respectively (Lalovic et al., 2004, 2006). Oxycodone also undergoes glucuronidation by UGT2B7 and minimally by UGT2B4 (Moore et al., 2003; Romand et al., 2017). Similar to that observed for hydromorphone, oxymorphone can undergo glucuronidation via UGT2B7 to form oxymorphone-3-glucuronide (Coffman et al., 1998; Adams and Ahdieh, 2005; Lalovic et al., 2006). However, the enzyme(s) involved in the glucuronidation of noroxycodone to noroxycodone-glucuronide have not been characterized (Huddart et al., 2018). Oxycodone, noroxycodone, and oxymorphone are also metabolized by keto-reduction to form α and β- oxycodol, oxymorphol, and noroxycodol, respectively [Fig. 2; (Moore et al., 2003; Baldacci and Thormann, 2005; Lalovic et al., 2006)].

Fig. 2.

Schematic of oxycodone metabolism.

Oxycodone exhibits its own analgesic effects (Lalovic et al., 2006) as inhibition of CYP2D6 did not attenuate its antinociceptive or opioid side effects, and oxycodone readily crosses the BBB (Cleary et al., 1994; Kaiko et al., 1996; Heiskanen et al., 1998; Boström et al., 2005, 2006, 2008; Okura et al., 2008; Lemberg et al., 2010; Drewes et al., 2013). Oxycodone is 1.5 times more potent than morphine when given by oral administration ((WHO), 2018), but they are both considered as medium potency (Eddy and Lee, 1959; Beaver et al., 1977; Thompson et al., 2004; Drewes et al., 2013; Cone et al., 2013a). Oxycodone’s active metabolite, oxymorphone, exhibits a more potent mu-opioid receptor affinity as compared to morphine after parenteral administration and oxymorphone has a 40-fold higher affinity for the mu-opioid receptor as compared to its parent drug, oxycodone and is 10 times more potent than oxycodone when given by intravenous administration (Chen et al., 1991; Lalovic et al., 2006; Babalonis et al., 2021). However, the overall contribution of oxymorphone to the analgesic efficacy of oxycodone is not fully understood (Lalovic et al., 2006; Cone et al., 2013a). Some studies have shown that oxymorphone contributes little analgesic efficacy during oxycodone administration, possibly due to its low relative abundance or its lower BBB permeability (Heiskanen et al., 1998; Lalovic et al., 2006; Zwisler et al., 2009; Lemberg et al., 2010). In contrast, multiple studies have shown that oxymorphone has long lasting analgesic effects with minimal side effects when administered independently (Gimbel and Ahdieh, 2004; Gimbel et al., 2005; Hale et al., 2005; Aqua et al., 2007), and it is prescribed alone as an effective analgesic typically in cancer pain management and obstetrics (Inturrisi, 2002).

Morphine Metabolism

The major metabolic pathway for morphine is via glucuronidation by UGT2B7 to form its major metabolite, morphine-3-glucuronide, and its minor metabolite, morphine-6-glucuronide (Fig. 3) (Coffman et al., 1997, 1998; Donnelly et al., 2002; Stone et al., 2003; Ohno et al., 2008). UGT1A1 and UGT1A8 have also been implicated as playing a role in morphine-6-glucuronide formation (Ohno et al., 2008), while morphine-3-glucuronide formation can also be catalyzed by UGT1A3 and UGT1A8 (Green et al., 1998; Cheng et al., 1999; Stone et al., 2003). Morphine is also N-demethylated by CYP3A4 and CYP2C8 to form a minor metabolite, normorphine (Projean et al., 2003), which can be further metabolized to two glucuronide metabolites (Yeh et al., 1977). Morphine can also be metabolized to minor sulfate metabolites by SULT1A3 [Fig. 3; (Andersson et al., 2014; Kurogi et al., 2014)].

Fig. 3.

Schematic of morphine metabolism.

Morphine-6-glucuronide is considered to be active and may partially contribute to morphine’s analgesic effect (Frances et al., 1992; Portenoy et al., 1992; Klepstad et al., 2000; Kilpatrick and Smith, 2005; Lötsch, 2005; Wittwer and Kern, 2006) given its high affinity for both mu opioid receptors [mu 1 and mu 2; (Pasternak et al., 1987; Frances et al., 1992)]. Several studies have shown that the potency of morphine-6-glucuronide is equal to or more active than morphine depending on the route of administration (Paul et al., 1989; Osborne et al., 1990; Frances et al., 1992; Kilpatrick and Smith, 2005; Wittwer and Kern, 2006; Ohno et al., 2008). However, several studies suggest that the contribution of morphine-6-glucuronide to morphine’s analgesic effects is likely small due to it accounting for only 10% of circulating metabolite in plasma (Hasselström and Säwe, 1993; Lötsch et al., 1996; Andersen et al., 2003; Ing Lorenzini et al., 2012) and because it does not easily cross the BBB as compared to morphine (Frances et al., 1992; Bickel et al., 1996; Wandel et al., 2002; Drewes et al., 2013; Seleman et al., 2014). A systematic review found the weighted mean ratios in serum of morphine and its glucuronide metabolites were 6 (range = 0.2–15) for morphine-3-glucuronide:morphine and 0.9 (range = 0.03–2.6) for morphine-6-glucuronide:morphine in patients with normal renal function given intravenous morphine (Faura et al., 1998). However, some studies suggest that morphine-6-glucuronide may have a major role in morphine analgesia (Hanna et al., 1990; Osborne et al., 1990, 1992), although the analgesic efficacy was found to not be dependent on morphine-6-glucuronide plasma concentration (Osborne et al., 1992). Morphine-6-glucuronide can be administered as its own medication and is well tolerated (Osborne et al., 1992; Penson et al., 2002).

In contrast, morphine-3-glucuronide does not exhibit analgesic effects (Shimomura et al., 1971; Christensen and Jørgensen, 1987; Oguri et al., 1987; Pasternak et al., 1987; Qian-Ling et al., 1992; Wittwer and Kern, 2006) and was shown to antagonize the analgesic effects of both morphine and morphine-6-glucuronide in both mice and rats (Qian-Ling et al., 1992; Christrup, 1997; Faura et al., 1997). Studies have shown that morphine-3-glucuronide causes neuroexcitatory effects such as hyperalgesia, allodynia, and myoclonus (Andersen et al., 2003; Roeckel et al., 2017).

Pharmacodynamic Drug–Drug Interactions

Pharmacodynamic drug interactions involving opioids typically occur when opioids are taken in conjunction with other CNS depressants due to polypharmacy, potentially leading to life-threatening ADE (Prostran et al., 2016; Matos et al., 2020). Pharmacodynamic drug interactions can either be additive, synergistic, or antagonistic (Pérez-Mañá et al., 2018; Niu et al., 2019). An example of an additive or synergistic interaction would be concomitantly taking an opioid with a CNS depressant, such as a benzodiazepine, which could lead to respiratory depression and potentially death (Mirakbari et al., 2003; Dowell et al., 2016; Hwang et al., 2016; Bingham et al., 2020). Since both opioids and CNS drugs have narrow therapeutic indexes, close physician monitoring of the patient should occur under these circumstances, adjusting the dose as necessary to ensure an ADE does not occur. Generally, opioids should not be prescribed or taken together with CNS depressants, such as benzodiazepines, antipsychotics, muscle relaxers, or tranquilizers (FDA, 2017). While there is an overall lack of clinical information investigating the pharmacodynamic interaction and side effects of opioids with these drug classes (Jones et al., 2012; Leonard and Kangas, 2020), there is substantial evidence suggesting an increase in overdose when opioids and benzodiazepines are combined (Park et al., 2015; Bachhuber et al., 2016; Dasgupta et al., 2016; Sun et al., 2017; Dowell et al., 2022; https://nida.nih.gov/research-topics/opioids/benzodiazepines-opioids#). An example of an antagonistic interaction would be the administration of naloxone after an opioid overdose, in which naloxone reverses the effects of the opioid overdose (Boom et al., 2012; Dunne, 2018). Pharmacodynamic DDI can be beneficial if medications are prescribed deliberately and safely (Niu et al., 2019). Furthermore, depending on the pharmacodynamic interaction (additive/synergistic vs antagonistic) smaller amounts of opioids may need to be prescribed as their analgesic effects may still be effective if taken concurrently with other drugs (Pick, 1997; Niu et al., 2019). Specific examples of pharmacodynamic interactions can be found in Table 1.

TABLE 1

In vivo and in vitro studies performed to examine potential drug–drug interactions between precipitant drugs and either hydrocodone, morphine, or oxycodone

Drug–Drug Interactions Where Opioids are the Object Drug

Mu-receptor agonists, specifically opioids, typically have a narrow therapeutic range (Grönlund et al., 2010a,b). Thus, at normal dosages, opioid taken concomitantly with a drug that could potentially inhibit opioid metabolism could potentially lead to adverse drug events. Conversely, taking a concomitant drug that is an inducer of opioid metabolism could potentially lead to subtherapeutic efficacy. Furthermore, prodrugs may exhibit the opposite effect, where if prodrug metabolism is inhibited, there would be subtherapeutic efficacy, and if it is induced there is a higher risk of adverse drug events due to higher plasma concentrations. Altered dosing schedules may be needed to maintain safe therapeutic plasma concentrations of the opioid in the presence of either an enzyme inhibitor or inducer. Therapeutic monitoring may become of increased importance if a known inducer or inhibitor of metabolizing enzyme is prescribed with a drug that is a substrate of an enzyme involved in opioid metabolism to ensure that prescribed opioids are within their therapeutic window.

Genotype Influence on Metabolism

For personalized approaches to patient care and the prescribing of prescription drugs, it is important to consider individual genotypes for key metabolizing enzymes when prescribing concomitant drugs that are known to be inhibitors or inducers of the key metabolizing enzymes. The major metabolizing CYP enzymes 3A4/5, 2C9, 2D6, 2E1, 1A2, and 2C19 (Shapiro and Shear, 2002), as well as UGT2B7, the major enzyme important in morphine metabolism (Bhasker et al., 2000; Shen et al., 2019), all exhibit high-prevalence polymorphisms potentially important in DDI.

CYP2D6

CYP2D6 has multiple single nucleotide polymorphisms (SNPs) that lead to functional variants of the enzyme and individuals can be categorized into different CP2D6 genotype groups based on the functionality of their CYP2D6 genotypes. Those that have two normal function alleles of CYP2D6 are EM, and these include the *1, *2, and *35 allelic variants (Bradford, 2002; Gaedigk et al., 2018, 2020, 2021). Individuals who have two nonfunctioning, one less functional allele and one nonfunctioning allele, or deleted genes, are PM and include the *3, *4, *5, *6, *7, and *8 variants (Gaedigk et al., 2018, 2020, 2021). Intermediate metabolizers (IM) are those individuals either homozygous for less functional alleles or have one normal function allele and one less function allele or deleted gene. The *9, *10, *17 and *41 alleles are associated with the IM phenotype (Gaedigk et al., 2018, 2020, 2021). Individuals who take a drug that is a CYP2D6 inhibitor or inducer can experience phenoconversion from an EM to PM metabolizer phenotype, potentially altering clinical response by affecting the metabolic clearance of the victim drug and potentially leading to an ADE (Shah and Smith, 2015). Conversely, an individual exhibiting the EM phenotype can exhibit an ultra-rapid metabolizer (UM) phenotype if they are taking a drug that induces transcription of the drug metabolizing enzyme of interest (Shah and Smith, 2015). As CYP2D6 is highly polymorphic, therapeutic monitoring may be necessary, with dosage adjustments based upon an individual’s genotype to ensure optimal plasma drug concentration, especially for those drugs with a narrow therapeutic index, and this may be particularly true for opioid dosing.

CYP3A4

Similar to CYP2D6, the CYP3A4 gene also exhibits numerous SNPs; however, many of these alleles have yet to show variation in CYP3A4 activity in vivo (Westlind-Johnsson et al., 2006). EM phenotype individuals are those that have any of the CYP3A4 *1 allelic sub-variants (Gonzalez et al., 1988; Gaedigk et al., 2018, 2020, 2021). The CYP3A4 PM phenotype is considered when the rare *20 allelic variant is present (Westlind-Johnsson et al., 2006). CYP3A4 alleles considered to have decreased enzymatic function in vivo include the *18A and *22 alleles, and these are linked to individuals with the CYP3A4 IM phenotype (Dai et al., 2001; Kang et al., 2009; Elens et al., 2011a,b; Wang et al., 2011a; Gaedigk et al., 2018, 2020, 2021). Genetic variation in the CYP3A4 enzyme is particularly important when considering potential DDIs since CYP3A4 accounts for approximately 50% of all drug metabolism (Trescot et al., 2008).

UGT2B7

The most prevalent UGT2B7 SNP is at codon 268 (His>Tyr) (approximately 50% prevalence in Caucasians) and it has been associated with varying functionalities depending on the substrate (Lazarska et al., 2018). The UGT2B7*2 variant has been shown to exhibit either increased activity or decreased activity depending on the drug examined (Thibaudeau et al., 2006; Bélanger et al., 2009; Wang et al., 2011b). Other UGT2B7 polymorphisms include less prevalent (i.e., minor allele frequency <3%) synonymous, nonsynonymous, and promoter SNPs (Bhasker et al., 2000; Wang et al., 2018b), but their effect on UGT2B7 activity or expression and overall drug metabolism has been much less studied.

Table 2 lists known inhibitors and inducers of the metabolizing enzymes CYP2D6, CYP3A4, and UGT2B7, which are the major enzymes involved in hydrocodone, oxycodone, and morphine metabolism. Known inhibitors of these metabolizing enzymes could potentially lead to changes in pharmacokinetic disposition, such as increased exposure to a given substrate or, conversely, lead to a decrease in exposure of an active metabolite after a prodrug is given. This increase in exposure could potentially increase efficacy or lead to accumulation of active drug, potentially resulting in toxicity or adverse drug events, or lead to inefficacy due to poor metabolism of a prodrug resulting in little to no formation of the active metabolite. In contrast, inducers of these metabolizing enzymes could potentially lead to sub-therapeutic levels of active drug decreasing the drug’s efficacy. There have been relatively few clinical trials examining the effects of CYP2D6 or CYP3A4 inhibitors in concomitant administration with hydrocodone (Kapil et al., 2015). The majority of inhibition studies have been performed in vitro, primarily in human liver microsomes or human hepatocytes to determine potential DDI. In contrast, there have been extensive studies performed to examine the potential drug-drug interactions with oxycodone and morphine both in vitro and in vivo. These include in vitro mechanistic studies as well as in vivo human clinical trials to identify pharmacokinetic and pharmacodynamic changes. Briefly, 8 DDI were identified involving hydrocodone, including some that were in the same study, 21 involving oxycodone, and 30 involving morphine. A few major DDI studies are described below for each opioid; these include studies that show significant DDI including pharmacokinetic and drug–gene interactions. In vitro DDI studies are not described in depth. Other known DDI that were not described in detail below include other trials as well as in vitro studies. These can be found in Table 1.

TABLE 2

Known inducers and inhibitors of three major opioid metabolizing enzymes (adapted from Drug Interactions Flockhart TableTM from the Department of Medicine Clinical Pharmacology Division at Indiana University for CYP2D6 and CYP3A4 (Flockhart et al., 2021) and the Food and Drug Administration Drug Development and Drug Interactions Table of Substrates, Inhibitors and Inducers (FDA, 2022).

Hydrocodone

In a randomized controlled trial, the effects of paroxetine (20 mg), a known time-dependent CYP2D6 inhibitor, was investigated when co-administered with a once daily 20 mg extended-release hydrocodone tablet (Kapil et al., 2015). As compared to placebo-treated controls, there was a decrease in the area under the curve (AUC) of the active metabolite hydromorphone in the paroxetine-treated group (0.64 ng·h/L vs 3.8 ng·h/L). However, the maximum concentration (Cmax), time at maximum concentration, and half-life of hydrocodone were similar regardless of the presence of paroxetine. The AUC and Cmax ranges after paroxetine exposure were within their predetermined range of 80%–125% (FDA, 2020), suggesting that paroxetine did not significantly alter hydrocodone exposure. Few adverse effects were reported, including headache, nausea, and diarrhea, and there was no significant difference in reported side effects between the placebo and treatment groups (Kapil et al., 2015). Even though hydromorphone is pharmacologically active, due to its relatively low plasma levels after hydrocodone administration, the inhibition of its formation is not expected to heavily impact hydrocodone efficacy and pharmacodynamic properties (Kapil et al., 2015).This is consistent with the fact that CYP2D6-mediated O-demethylation accounts for 3% of total hydrocodone metabolism (as compared to CYP3A4 mediated N-demethylation which accounts for 40% of its metabolism (Cone and Darwin, 1978; Cone et al., 1978) and is therefore not expected to heavily impact hydrocodone efficacy and pharmacodynamic properties (Kapil et al., 2015).

A phase I clinical trial was performed to determine if DDI were observed for hydrocodone when co-administered with the treatment regimen for hepatitis C virus (termed 3D) consisting of ombitasvir/paritaprevir/ritonavir and dabusavir. This concomitant administration led to a 27% and 90% increase in Cmax and AUC of hydrocodone, respectively, in the 3D group as compared to placebo controls. The increase in hydrocodone exposure is likely due to ritonavir inhibition of CYP3A4. The researchers recommended a 50% dose reduction in hydrocodone to account for the increase in hydrocodone exposure with concomitant administration with the 3D regimen to avoid potential adverse drug events (Polepally et al., 2016). Future studies will be required to determine how the inhibition of this major metabolic pathway will affect the pharmacodynamics of hydrocodone.

A single case study involving a white male with chronic pain participating in a 2-day protocol was performed to determine if cannabis added to a hydrocodone/acetaminophen regimen could detect pharmacodynamic or pharmacokinetic drug–drug interactions. For both days, the male took his prescribed hydrocodone/acetaminophen regimen (½ tablet of 7.5 mg/325 mg combination) with the addition of smoking one pre-rolled cannabis cigarette (0.5 g; 22.17% THC; 0.12% CBD) on day 2. On day 2, hydrocodone plasma levels observed were lower than on day 1, with pharmacokinetic analysis indicating a more rapid absorption of hydrocodone. The participant also reported lower pain, and this may be explained by the more rapid absorption of hydrocodone in the presence of cannabis (Bindler et al., 2022).

Another single case study involved a 5-year-old girl suffering from respiratory tract/ear infections (Madadi et al., 2010). The girl had been prescribed valproic acid since birth to treat her for seizures (250 mg twice per day) and was prescribed hydrocodone 1 mg/ml, one teaspoon three times/day for 5 days) and clarithromycin for her infection (Madadi et al., 2010). After 24 hours of using newly prescribed hydrocodone and clarithromycin, the child was found unresponsive and pronounced dead at the hospital, with postmortem tests revealing high plasma hydrocodone levels (0.14 µg/ml) with undetectable levels of plasma hydromorphone (<0.008 µg/ml). Genetic testing revealed that the patient has one functionally impaired CYP2D6 allele (*41) and one normal function (*2A) allele (Madadi et al., 2010), suggesting that she was an intermediate-to-poor metabolizer of CYP2D6 substrates like hydrocodone, resulting in decreased hydromorphone formation. Furthermore, clarithromycin is a known inhibitor of CYP3A4 (Rodrigues et al., 1997). Thus, both major metabolic pathways of hydrocodone were impaired either due to a drug-gene interaction (CYP2D6) or chemical inhibition (CYP3A4), leading to increased plasma levels of hydrocodone (Madadi et al., 2010). This case highlights the importance of drug-gene interactions and the complex interplay of DDI involving drug–gene interactions. It is becoming increasingly important to be aware of individual patient drug metabolizing enzyme genotypes to avoid harmful side effects when prescribing medications.

In vitro studies focusing on hydrocodone DDI found significant inhibition of hydromorphone formation in the presence of quinidine and furafylline through CYP2D6 inhibition (Hutchinson et al., 2004). Hutchinson et al., also found significant inhibition of norhydrocodone in the presence of ketoconazole and troleandomycin through CYP3A4 inhibition (Hutchinson et al., 2004). These data further suggest that the prescription of hydrocodone with either CYP2D6 or CYP3A4 inhibitors should be avoided.

Oxycodone

There have been three randomized controlled trials examining the effects of paroxetine, a CYP2D6 inhibitor, on the pharmacokinetics of oxycodone. In one of these trials, chronic pain patients were administered oxycodone tailored to each patient need (range 20 mg–320 mg/day) along with 20 mg/day of paroxetine (dose-corrected plasma oxycodone concentrations of patients were similar, but patients were also on other comedications for their chronic conditions; (Lemberg et al., 2010). Compared with placebo controls, paroxetine increased the mean dose-adjusted AUC and Cmax of oxycodone by 19% and 23%, respectively. Paroxetine also increased the noroxycodone AUC and Cmax by 100% and 102%, respectively, as compared to the placebo group. The oxymorphone Cmax and AUC were decreased with paroxetine treatment by 57% and 67% (Lemberg et al., 2010) and noroxymorphone plasma concentrations were also decreased by paroxetine treatment. Several adverse effects were reported during concomitant treatment, including headaches, drowsiness, dizziness, and nausea/vomiting. Interestingly, the analgesic effect of oxycodone was not significantly altered with paroxetine treatment, suggesting that CYP2D6 inhibition may not be of major clinical significance and that oxymorphone formation may not be central to oxycodone’s analgesic effect (Lemberg et al., 2010).

Two other randomized control trials investigated the effect of paroxetine alone or in tandem with the CYP3A4 inhibitor, itraconazole, on oxycodone pharmacokinetics. In an initial study by Gronlund et al., only minimal changes in AUC and Cmax for oxycodone was observed after oral oxycodone administration for the paroxetine-treated group as compared to placebo controls (Grönlund et al., 2010a). However, oxymorphone plasma concentrations decreased by 44% while the AUC for noroxycodone increased by 68% in the presence of paroxetine, suggesting that the inhibition of CYP2D6-mediated oxymorphone formation by paroxetine could potentially result in shunting of oxycodone metabolism to the CYP3A4-mediated formation of noroxycodone. In contrast, administration of both paroxetine and itraconazole led to 1.8- to 1.9-fold increases in Cmax and AUC for oxycodone, with corresponding decreases in the AUC of both oxymorphone and noroxycodone. The same researchers performed a similar randomized study but with intravenous oxycodone administration (Grönlund et al., 2011a). Similar to that observed in their earlier oral oxycodone administration study, paroxetine inhibited oxymorphone formation but with little change in oxycodone pharmacokinetics. With the administration of both paroxetine and itraconazole, there was a significant 2-fold increase in oxycodone exposure as well as inhibition of both oxymorphone and noroxycodone formation. Together, these data are consistent with an effective inhibition of both CYP2D6- and CYP3A4-mediated oxycodone metabolism by the paroxetine-itraconazole combination. While the data further suggested that the inhibition of CYP3A4 may be more clinically relevant to potential oxycodone DDI as compared to CYP2D6 inhibition, the studies by Gronlund et al., did not investigate the effect of itraconazole alone on oxycodone pharmacokinetics. Interestingly, the inhibition of both enzymes failed to cause significant changes in the analgesic effects manifested by oxycodone in the two studies after a single dose of oxycodone, suggesting that with repeated oxycodone administration, the decrease in oxycodone clearance observed in this study can result in accumulated levels of oxycodone leading to adverse side effects (Grönlund et al., 2011a). Saari et al., found that the coadministration of itraconazole (oral) and oxycodone (intravenous) led to decreased plasma clearance by 32% and increased AUC of oxycodone by 51%. Similarly, the AUC of orally administered oxycodone increased 1.44-fold in the presence of itraconazole and the Cmax increased by 45% (Saari et al., 2010). The AUC of the CYP3A4 metabolite, noroxycodone was decreased by 49% and oxymorphone was increased by 359% (Saari et al., 2010).

In a non-randomized controlled trial, subjects were administered oxycodone (0.2 mg/kg) and 20 mg of paroxetine (