**4. Clinical Cases**

#### *4.1. The Case of Omeprazole and Clopidogrel*

Proton pump inhibitors (PPIs) are commonly prescribed along with antiplatelet drugs like clopidogrel to reduce the incidence of gastric bleeding during treatment with antiplatelet therapy [25,26]. Omeprazole has long been one of the most widely used PPIs [27]. Omeprazole is a known strong affinity substrate of CYP2C19, leading to the formation of its hydroxy and desmethyl metabolites. The antiplatelet drug clopidogrel is sequentially activated by CYP450 isoforms, including CYP2C19, into its active metabolite (H4) [28]. When omeprazole is co-administered with clopidogrel, omeprazole acts as a strong affinity substrate of CYP2C19 (perpetrator), whereas clopidogrel is a weaker sensitive substrate (victim). (Figure 7) Multiple in vitro studies have reported a potential pharmacokinetic interaction between omeprazole and clopidogrel [29–32]. A clinical study conducted by Angiolillo et al. demonstrated that plasma levels of clopidogrel's active metabolite H4, and consequently the platelet aggregation induced by adenosine diphosphate, were decreased when omeprazole and clopidogrel were administered concomitantly [30]. Various other clinical studies have demonstrated that co-administration of omeprazole and clopidogrel diminishes the antiplatelet activity of clopidogrel [31,32].


**Figure 7.** CYP450 metabolic pathways involved in the metabolism of clopidogrel and omeprazole, and their respective affinities are depicted. Competitive inhibition will be expected at CYP3A4 and mechanism-based inhibition at the CYP2C19 enzymatic level. Clopidogrel is the victim drug and omeprazole acts as the perpetrator drug.

Since omeprazole is a strong affinity substrate for CYP2C19, an "immediate" competitive inhibition is expected between these two drugs. Since competitive inhibition was expected, separating time of administration was considered a logical mitigation strategy to avoid or alleviate the extent of the drug interaction [29]. Others have suggested that increasing the dose of clopidogrel might compensate for the diminished formation of the active metabolite [29]. These recommendations to separate the time of administration of the two drugs or to increase the dose of the victim drug (clopidogrel) come from a sound rationale and have been proven to be efficacious in mitigating drug interactions associated with competitive inhibition. However, it has been shown that following chronic administration, separating the time of administration does not alleviate the reduction in clopidogrel active metabolite (H4) caused by omeprazole [30,33]. This is due to the fact that omeprazole is not only a competitive inhibitor, but also a mechanism-based inhibitor of CYP2C19, which results in a gradual increase in irreversible inhibition of the CYP2C19 enzyme, to a point where clopidogrel activation and its clinical efficacy are significantly impaired. From these observations, the FDA warns that separating the time of administration between these two substrates will not alleviate this interaction [34].

Multiple studies have been conducted to determine the clinical impact of the potential reduced antiplatelet efficacy resulting from this interaction. In two retrospective studies looking at interactions between clopidogrel and PPIs and the effects on clinical outcomes, it was reported either that PPIs were associated with increased cardiac adverse events in acute coronary syndrome patients, or that cardiac adverse events were less common in PPI non-users [35,36]. Short-term mortality odds ratios also favored PPI non-users, but no significant differences were observed in long-term mortality [35]. Though a wide range of PPIs were reviewed, omeprazole and esomeprazole remained the most widely prescribed when all studies were combined. A similar study was conducted by Mahabaleshwarkar et al., which found that PPIs were slightly, but significantly, associated with all-cause mortality [37]. The odds ratio of adverse cardiac events and all-cause mortality for omeprazole in particular was 1.23. In another retrospective cohort study, clopidogrel use post discharge for acute coronary syndrome hospitalizations was studied [38]. Concurrent clopidogrel and PPI use was associated with an increased risk of death or rehospitalization; among patients prescribed a PPI, 60% were on omeprazole [38]. PPI plus clopidogrel use also remained significantly associated with recurrent acute coronary syndrome and revascularization procedures [38]. Another study evaluated the association between various PPIs (all PPIs combined) and individual PPI agents with clopidogrel use and increased risk of hospitalization [39]. There was no significant association between any PPI and increased risk of rehospitalization with clopidogrel, but this association was significant with omeprazole [39].

Several studies also report no change in the frequency of cardiovascular adverse events with omeprazole administration during clopidogrel treatment [26,40–42]. Dosing regimens in these studies sugges<sup>t</sup> that the extent of interaction between clopidogrel (75 mg vs. 600 mg) and omeprazole (20 mg vs. 80 mg) may be dose dependent. If an alternative PPI has to be considered, in vitro studies using human liver microsomes have confirmed that PPIs like rabeprazole, lansoprazole, dexlansoprazole, and pantoprazole do not show evidence of mechanism-based inhibition [29]. Clinical studies have also reported that effects of lansoprazole and pantoprazole on clopidogrel antiplatelet activity are not as potent as omeprazole [43–45]. When clopidogrel and PPI coadministration is necessary, switching to a PPI other than omeprazole or esomeprazole may be considered, in light of the evidence presented herein.

#### *4.2. The Case of Paroxetine*

The antidepressant paroxetine is a known substrate of CYP2D6, but also a potent mechanism-based inhibitor of this enzyme. As such, paroxetine is expected to inhibit its own metabolism over time. This has been illustrated in two clinical studies, where the *Cmax* and AUC of paroxetine were increased 5.2- and 7-fold after 2 weeks of paroxetine administration, respectively [46,47]. In the second study, Laine et al. revealed that ultra-rapid metabolizers of CYP2D6 were converted to extensive or poor metabolizers with chronic paroxetine use [47]. The FDA-approved label states that paroxetine takes

nearly 10 days to achieve steady-state concentrations even though the drug has a reported elimination half-life of 21 h; so, within 4–5 days, steady-state levels should be reached under normal circumstances. The label also states that saturation of CYP2D6 contributes to the non-linear pharmacokinetics of paroxetine. Thus, it may be assumed that inhibition of its own mechanism contributes to achieving later than expected steady-state levels and clinical e fficacy [48].

As paroxetine is a strong CYP2D6 a ffinity substrate, it can exhibit acute competitive inhibition when co-administered with sensitive substrates like nortriptyline. (Figure 8) Though separating the time of administration may seem appropriate initially, the e ffects of mechanism-based inhibition over time should be factored into medication risk management; dosage adjustment or substitution of the victim substrate may be necessary. The extent of this interaction also depends on when paroxetine and the victim drug are added to the regimen. If paroxetine is newly added to a victim drug that has already reached steady-state plasma levels, the extent of inhibition and plasma concentrations of the victim drug will increase over time until reaching a new steady state. However, if steady-state levels of paroxetine are already achieved before adding another sensitive substrate, the extent of inhibition will be maximum at initiation and will remain stable, since saturation of enzyme inhibition is already established.


**Figure 8.** CYP450 metabolic pathways involved in the metabolism of nortriptyline and paroxetine and their respective a ffinities for the isoform are depicted. Mechanism-based inhibition at CYP2D6 enzymatic level will be expected. Nortriptyline is the victim drug and paroxetine acts as the perpetrator drug for the CYP2D6 elimination pathway.

#### *4.3. The Case of Erythromycin*

Another example that illustrates the concept of saturability in mechanism-based inhibition is observed with the commonly used macrolide antibiotics erythromycin (Figure 9) or clarithromycin. These drugs are known to cause mechanism-based inhibition of CYP3A4. Clinical studies have demonstrated saturability (i.e., degree of inhibition reaches a maximum value) of enzyme inhibition using CYP3A4-sensitive substrates like alfentanil or midazolam with or without erythromycin pretreatment (treatment or control, respectively). In a study with healthy males after single or multiple oral dose(s) of erythromycin 500 mg, the e ffects on alfentanil pharmacokinetics were measured [49]. A 25% increase of alfentanil half-life was observed following a single erythromycin dose compared to control. After a 7-day pretreatment with erythromycin, the half-life of alfentanil was further increased by 25% (up to 56% compared to control), suggesting an increase in inhibition with time [49]. Similar e ffects were seen on clearance. Although a direct association of this potential drug interaction on clinical outcomes has not been systematically reported, two case reports sugges<sup>t</sup> that erythromycin pretreatment may cause prolonged respiratory depression when alfentanil is administered compared to patients who did not receive erythromycin [50,51].

#### *Pharmaceutics* **2020**, *12*, 846


**Figure 9.** CYP450 metabolic pathways involved in the metabolism of erythromycin and midazolam and their respective affinities for the isoform are depicted. Mechanism-based inhibition at the CYP3A4 enzymatic level will be expected. Midazolam is the victim drug and erythromycin acts as the perpetrator drug.

In another study with 12 healthy volunteers, the effects of erythromycin on midazolam metabolism were studied [52]. It was observed that the *AUC* of midazolam increased 2.3-fold after 2 days of erythromycin pretreatment, compared to control. Following 4 days of pretreatment, midazolam's *AUC* increase was 3.38-fold. After a 7-day pretreatment, a similar increase (3.38-fold) was observed, indicating an increase in inhibition with repeated administration of erythromycin and that a plateau effect had been reached after 4 days of exposure [52].

#### *4.4. The Case of Mirabegron*

A widely prescribed drug in the treatment of overactive bladder, mirabegron, also displays characteristics of mechanism-based inhibition for CYP2D6. (Figure 10) The difference in the degree of metabolism inhibition between a competitive inhibitor and a mechanism-based inhibitor is perceived when they are compared using the same victim drug. One study investigated the victim drug desipramine, with the potential competitive inhibitor duloxetine and the mechanism-based inhibitor mirabegron. In this study, duloxetine (moderate CYP2D6 affinity substrate) 30 mg twice a day was administered for 10 days, after which desipramine (weak CYP2D6 affinity substrate) 50 mg was administered as a single dose. Here, duloxetine would act as a competitive inhibitor and desipramine as a victim drug. Accordingly, a 1.2-fold increase in *AUC* and 0.6-fold increase in *Cmax* of desipramine was observed [53]. In a similar study design, desipramine was administered with or without mirabegron pretreatment. First, desipramine was administered alone and, after a washout period, mirabegron (100 mg) was administered for 13 days. On the fourteenth day, mirabegron 100 mg and desipramine 50 mg were co-administered. A 3.41-fold increase in the desipramine *AUC* was observed with mirabegron pretreatment compared to control [54]. This increase in the *AUC* of desipramine was much larger than the increase observed with duloxetine, a "purely" potential competitive inhibitor. It is important to note that in the short term, mechanism-based inhibitors can act as competitive inhibitors if the other drug has lower affinity for the metabolizing enzyme. Duloxetine and mirabegron are both substrates of CYP2D6 with moderate affinity, exhibiting potential competitive inhibition over desipramine; therefore, following a single dose of each drug, a similar level of CYP2D6 inhibition is expected towards the CYP2D6 victim drug. (Figure 10) The higher inhibition observed with multiple doses of mirabegron versus single dose of duloxetine is explained by the mechanism-based inhibition observed over time. Similar effects of chronic mirabegron administration on metoprolol (a weak CYP2D6 substrate) pharmacokinetics are also reported [54]. A 3.3-fold increase in the metoprolol AUC was observed following a pretreatment of 5 days with mirabegron in CYP2D6 normal (i.e., previously called "extensive") metabolizer subjects [55]. The coadministration of quinidine (a potent CYP2D6 inhibitor) with metoprolol was associated with a similar magnitude of increase in metoprolol's AUC after a single dose [56]. The intensity of drug–drug interaction through "purely" competitive inhibition is expected to be lower between substrates compared to quinidine's inhibition; therefore, the mechanism-based inhibition property of mirabegron can explain why the magnitude of drug–drug interaction observed between mirabegron and metoprolol is similar to that observed between quinidine and metoprolol.


**Figure 10.** CYP450 metabolic pathways involved in the metabolism of desipramine, duloxetine, metoprolol, mirabegron, and their respective affinities for the isoform are depicted. Competitive inhibition will be expected at CYP2D6 between duloxetine (perpetrator; CYP2D6 substrate with higher affinity) and either desipramine or metoprolol (victim drugs; both CYP2D6 substrates with weaker affinity). Mechanism-based inhibition at CYP2D6 will be expected between mirabegron and desipramine, metoprolol, or duloxetine.

In addition to the case examples discussed above, Figure 11 provides a list of CYP450 mechanism-based inhibitors, along with the enzyme inhibited and relevant CYP450 pathways involved in their metabolism. This list is not exhaustive, but provides a quick reference for commonly used medications.

In addition to drug–drug interactions, high variability in terms of CYP450 expression and/or activities can be explained by genetic polymorphisms in genes encoding specific isoforms (such as *CYP2C9*, *CYP2C19*, and *CYP2D6*). This variability on CYP450 expression/activities translates into intersubject variability in drug disposition and drug response. Often, the impact of genetic polymorphisms and drug–drug interactions on CYP450s have been studied separately. However, an interaction exists between these factors. Genetic polymorphisms could also contribute to variability observed in the magnitude of drug–drug interactions observed between two drugs. So, genetic polymorphisms in drug-metabolizing enzymes can affect the occurrence of phenoconversion induced by drug inhibitors. As reported by Storelli et al., differences in CYP2D6 inhibition observed in vitro with paroxetine (mechanism-based inhibitor) or duloxetine (competitive inhibitor) across *CYP2D6* genotypes were not related to their inhibition parameters but likely due to a differential level of functional enzymes as a function of the *CYP2D6* genotype [57,58].

**Figure 11.** List of commonly prescribed medications producing mechanism-based inhibition.
