Comparative review of drug–drug interactions with epidermal growth factor receptor tyrosine kinase inhibitors for the treatment of non-small-cell lung cancer

Abstract

The development of small-molecule tyrosine kinase inhibitors (TKIs) that target the epidermal growth factor receptor (EGFR) has revolutionized the management of non-small-cell lung cancer (NSCLC). Because these drugs are commonly used in combination with other types of medication, the risk of clinically significant drug–drug interactions (DDIs) is an important consideration, especially for patients using multiple drugs for coexisting medical conditions. Clinicians need to be aware of the potential for clinically important DDIs when considering therapeutic options for individual patients. In this article, we describe the main mechanisms underlying DDIs with the EGFR-TKIs that are currently approved for the treatment of NSCLC, and, specifically, the potential for interactions mediated via effects on gastrointestinal pH, cytochrome P450-dependent metabolism, uridine diphosphate-glucuronosyltransferase, and transporter proteins. We review evidence of such DDIs with the currently approved EGFR-TKIs (gefitinib, erlotinib, afatinib, osimertinib, and icotinib) and discuss several information sources that are available online to aid clinical decision-making. We conclude by summarizing the most clinically relevant DDIs with these EFGR-TKIs and provide recommendations for managing, minimizing, or avoiding DDIs with the different agents.

Introduction

The occurrence of drug–drug interactions (DDIs) is a serious problem for the use of anticancer drugs. DDIs can exacerbate the risk of serious or fatal adverse events, and/or lead to reductions in therapeutic efficacy.¹ In particular, inducers of drug-metabolizing enzymes are known to increase the systemic clearance of many anticancer agents,² for example, long-term anticonvulsant therapy increases the systemic clearance of several antileukemic agents, thereby reducing their clinical efficacy.³

Tyrosine kinase inhibitors (TKIs) are effective for a wide variety of solid and hematologic malignancies, and are now established as standard therapeutic options;¹ more than 20 different TKIs are currently in use.⁴ In particular, the development of small-molecule TKIs that target the epidermal growth factor receptor (EGFR) – EGFR-TKIs – has revolutionized the management of locally advanced/metastatic non-small-cell lung cancer (NSCLC).⁵ This article reviews clinically relevant DDIs with EGFR-TKIs approved for the treatment of NSCLC.

The EGFR (HER1; ErbB1) is a member of the ErbB receptor family, which also includes HER2 (Neu, ErbB2), HER3 (ErbB3), and HER4 (ErbB4).^6,7 The main physiological role of ErbB-linked tyrosine kinases (TKs) is regulation of cellular proliferation.⁶ Somatic EGFR mutations are important oncogenic drivers of NSCLC,^8–10 and occur in 10–15% of the Caucasian patients and around 50% of the Asian patients with metastatic NSCLC and adenocarcinoma histology.¹¹ The most common EGFR mutations are deletions in exon 19 (del19) and L858R point mutations in exon 21.¹² These “activating” mutations lead to activation of intracellular signaling by EGFR in a ligand-independent manner.^13,14 NSCLC patients with activating EGFR mutations become dependent on EGFR activity to stimulate downstream signaling pathways and maintain the malignant phenotype (“oncogene addiction”).^15,16 Consequently, blocking ErbB family pathways with EGFR TKIs can inhibit tumor cell proliferation and initiate apoptosis.

EGFR-TKIs widely available for the treatment of advanced NSCLC include the first-generation reversible TKIs, erlotinib and gefitinib; the second-generation irreversible ErbB family blocker, afatinib; and the third-generation EGFR-wild-type sparing, irreversible EGFR/T790M inhibitor, osimertinib.¹⁷ In China, another first-generation EGFR-TKI, icotinib, is available.¹⁸ EGFR-TKIs are recommended as first-line treatment options for advanced EGFR mutation-positive (EGFRm+) NSCLC,^19,20 having demonstrated robust benefit in terms of progression-free survival (PFS, erlotinib: median 9.7–13.1 months;^21–23 gefitinib: 8.4–10.9 months;^24–28 icotinib 4.0–12.4 months;²⁹ afatinib: 11.0–13.6 months;^25,30,31 osimertinib: 17.7 months).³²

The first-generation EGFR-TKIs, gefitinib and erlotinib, bind reversibly to EGFR TK and thereby inhibit both mutant and (to a lesser extent) wild-type EGFR.³³ In randomized phase III trials, both agents demonstrated improved PFS and response rates, but not overall survival (OS) compared with platinum-doublet chemotherapy.^{21–24,27,28,34} However, patients with EGFRm+ NSCLC inevitably develop resistance to first-generation EGFR-TKIs.³⁵ In the majority (50–60%), resistance develops due to development of a novel T790M missense mutation in exon 20 of the EGFR,^36–38 which impairs binding of the EGFR-TKIs to the kinase domain of the receptor.³⁹

Another first-generation EGFR-TKI, icotinib has a similar chemical structure and physicochemical properties to erlotinib and displays similar clinical efficacy.²⁹ A randomized, double-blind trial in Chinese patients with advanced NSCLC who had failed to respond to chemotherapy concluded that icotinib was non-inferior to gefitinib with regard to PFS; median PFS was 4.6 months for icotinib and 3.4 months for gefitinib.⁴⁰ In a meta-analysis of pooled data for icotinib in patients with NSCLC (11 studies), the pooled mean PFS was 7.34 months (95% confidence interval: 5.60–9.07). EGFRm+ patients had longer PFS (median 11.0 months) than EGFRm- patients (1.97 months).²⁹

Further “generations” of EGFR-TKIs have also been developed, the aim being to improve efficacy and tolerability, and to overcome treatment resistance; data from head-to-head trials suggest improved outcomes with second-/third-generation versus first-generation TKIs.

In contrast to gefitinib and erlotinib, afatinib is an ErbB receptor family inhibitor. Afatinib binds covalently and irreversibly blocks signaling via all hetero- and homodimers formed by ErbB1, but also by HER2, HER3, and HER4.^41,42 It was postulated that this broad irreversible inhibition might delay or avoid the development of resistance.⁴³ In two phase III trials conducted in treatment-naïve NSCLC patients, afatinib demonstrated significant improvements in PFS versus platinum-doublets,^30,31 while pre-specified analyses of both trials identified significant improvements in OS with afatinib versus chemotherapy in patients with del19 mutations.⁴⁴ In a phase IIb study of treatment-naïve patients with advanced EGFRm+ NSCLC, afatinib significantly improved PFS, time to treatment failure (TTF) and objective response rate (ORR) versus gefitinib, with a trend towards improvement in overall survival (OS).^25,45

Osimertinib is an oral, third-generation, irreversible EGFR-TKI inhibitor⁴⁶ that selectively inhibits EGFR harboring both activating mutations and EGFR T790M resistance mutations, with less activity versus wild-type EGFR.⁴⁷ In a double-blind phase III trial in patients with previously untreated, EGFRm+ (del19 or L858R) advanced NSCLC, osimertinib demonstrated efficacy superior to that of standard EGFR-TKIs (gefitinib or erlotinib). Median PFS for osimertinib was significantly longer than for standard EGFR-TKIs (18.9 months versus 10.2 months).³² In a phase I study in patients harboring the EGFR T790M mutation having progressed during therapy with EGFR-TKIs, second-line (or later) osimertinib was highly active, with median PFS in T790M-positive patients (9.6 months) being substantially longer than in T790M-negative patients (2.8 months).⁴⁸ A single-arm study in patients harboring both activating EGFR mutations and the T790M mutation and with progression after prior EGFR-TKI therapy also showed antitumor efficacy of osimertinib (objective response rate 64%; disease control rate 90%) suggesting that osimertinib can overcome T790M-mediated acquired resistance.⁴⁹ Another second-generation EGFR-TKI, dacomitinib, recently became available in the US and EU, but to date there is little published evidence on potentially clinically relevant DDIs with dacomitinib.

In the treatment of NSCLC, EGFR-TKIs are commonly used together with other types of medication. Consequently, the risk of serious DDIs should be taken into consideration when selecting appropriate treatment.⁵ As the increasing molecular stratification of lung cancer has provided more options for targeted intervention and rational combination therapy, a clear understanding of the DDI profiles of different TKIs has become essential. Clinicians need to consider the potential implications of clinically important DDIs when formulating individualized therapeutic strategies for their patients. In this article, we review the key pharmacologic properties of the EGFR-TKIs currently approved for the treatment of NSCLC and the clinically relevant DDIs associated with each agent.

Literature search strategy

We searched the published literature (English language only) relating to established and potential DDIs between the EGFR-TKIs of interest (ie, those currently approved for the treatment of NSCLC) and other prescription drugs, over-the-counter drugs, and herbal medicines. Relevant publications were identified by means of searches of US National Library of Medicine (NLM) PubMed, using the search terms [interaction] OR [drug-drug] AND [Drug name (for each EGFR-TKI)]. Other relevant publications were identified from citations in the key publications identified via NLM PubMed. Further information was obtained from the US and EU prescribing information for each agent (not available for icotinib).

Mechanisms underlying DDIs for EGFR-TKIs used for the treatment of NSCLC

Interactions via effects on gastrointestinal pH

Other than the physicochemical properties of the different TKIs, the most important factor affecting the solubility of/exposure to these agents is gastric pH.¹ For TKIs with a pKa of less than 4–5, concomitant administration of agents that increase stomach pH can reduce TKI solubility, absorption, and bioavailability.^5,50 In particular, clinically relevant DDIs due to changes in gastric pH have been demonstrated between histamine H₂-receptor antagonists such as ranitidine, or proton pump inhibitors (PPIs) such as omeprazole, and a number of TKIs, including crizotinib, dasatinib, erlotinib, gefitinib, lapatinib, and pazopanib.^50,51 As patients with cancer often take acid suppressants for symptoms of gastrointestinal reflux, the potential for such interactions is clinically important.⁵

Interactions via effects on cytochrome P450 (CYP)-dependent metabolism

Phase I metabolism (mostly oxidative) by liver cytochrome P450 (CYP)–dependent enzymes is the most important route of drug metabolism in vivo.¹ Many TKIs are metabolized by this family of enzymes, which makes them prone to metabolic DDIs.^5,50 Indeed, most pharmacokinetic (PK) interactions that affect the EGFR-TKIs involve effects on metabolism, especially via CYP enzymes. Potent enzyme inhibitors and inducers can modify the exposure (the area under the plasma concentration–time curve [AUC] and the maximum plasma concentration [C_max]) of specific EGFR-TKIs, while EGFR-TKIs that are CYP enzyme substrates can affect the PK of other drugs.⁵ Increased or decreased exposure due to alteration of CYP enzyme activity could lead to clinically relevant toxicity or reduced effectiveness of the EGFR-TKIs.¹ Cigarette smoking is also known to induce key CYP enzymes⁵² and has been shown to affect the PK profiles of specific TKIs.⁵

As shown in Table 1, the extent to which different EGFR-TKIs are metabolized by CYP enzymes varies markedly, as do their effects as inducers or inhibitors of CYP enzyme activity. Nevertheless, there is wide potential for interactions with other drugs used in supportive therapy or for the treatment of comorbidities in patients with NSCLC.

Table 1.

Cytochrome P450 (CYP450) enzymes involved in the metabolism of EGFR-TKIs approved for the treatment of NSCLC

	Metabolized by CYP										May inhibit	May induce
	3A4	3A5	2D6	1A1	1A2	1B1	2C8	2C9	2C19	2E1
Erlotinib	+++	+++	+	+	++	+	+	+	–	–	CYP3A4 (m) CYP2C8 (m) CYP1A1 (s)	CYP1A1 CYP1A2
Gefitinib	+++	++	+++	++	+	–	–	–	–	–	CYP2C19 (w) CYP2D6 (w)	–
Afatinib	–	–	–	–	–	–	–	–	–	–	–	–
Osimertinib	+++	+++	–	–	–	–	–	–	–	–	–	CYP3A (w)
Icotinib	+++	++	–	–	+++	–	–	–	–	+++	NR	CYP1A2

Notes: +++, major metabolic route; ++, other significant metabolic route; +, minor metabolic route; –, no interaction.

Abbreviations: w, weak; m, moderate; s, strong; NR, not reported; EGFR-TKI, epidermal growth factor receptor-tyrosine kinase inhibitor.

CYP2D6, in particular, has been reported to be responsible for the metabolism of up to 25% of commonly prescribed drugs.^53,54 Importantly, CYP2D6 is encoded by a highly polymorphic gene, with more than 70 alleles and 130 genetic variations,⁵⁵ which could have a significant influence on up to a half of the drugs metabolized by this enzyme.⁵⁶ There is marked inter-ethnic variation in the frequency of the different alleles, leading to substantial variability in the prevalence of the four main phenotypes (poor, intermediate, extensive, and ultra-rapid metabolizers).⁵⁷ The poor metabolizer phenotype, in particular, has been widely examined in relation to adverse drug reactions.⁵⁸ Poor metabolizers account for 5–10% of the Caucasian populations, but are rarely found in Asia, and are highly variable among people of African ancestry.^59,60 Conversely, about 10–15% of the Caucasians are intermediate metabolizers, compared with up to 50% of the Asians and 30% of the Africans.⁵⁷

Interactions via effects on UDP-glucuronosyltransferase

Uridine-diphosphate (UDP)-glucuronosyltransferase (UGT) catalyzes the conjugation of glucuronic acid to endogenous substances and exogenous compounds.⁴ Most UGT isoforms are expressed in the liver.⁴ Since UGTs are rate-limiting enzymes in the metabolism of various compounds, co-administration of UGT-inhibiting drugs can lead to an increase in the concentration of such compounds in the circulation.⁶¹ In particular, UGT1A1 plays a key role in the metabolism and detoxification of many potentially harmful compounds and drugs, and thus inhibition of UGT1A1 can lead to severe DDIs and other undesirable effects.⁶² UGT1A1 is also involved in the metabolic elimination of endogenous bilirubin, preventing accumulation to toxic levels.⁶² Several TKIs are potent inhibitors of UGT1A1 (eg, erlotinib, lapatinib, nilotinib, pazopanib, regorafenib, and sorafenib),^63–65 and this may underlie some of the adverse events observed with these agents, such as hyperbilirubinemia and hepatotoxicity.^62,64,66

Interactions via effects on transporter proteins

In order to reach the portal blood circulation, TKIs need to pass through the gut wall. This involves both passive diffusion and active transport via organic anion-transporting peptide (OATP), organic cation-transporting peptide (OCTP), multidrug-resistance–associated proteins such as ATP-Binding Cassette (ABC) transporter G2 (ABC-G2), efflux transporters such as P-glycoprotein (P-gp, also known as multidrug resistance protein 1 [MDR1], ABC sub-family B, member 1 [ABC-B1] or cluster of differentiation 243 [CD243]), and intestinal metabolic enzymes such as CYP3A4.¹ P-gp and ABC-G2 are expressed in the small intestine, liver, kidneys, and blood–brain barrier (BBB); they appear to regulate the oral absorption, biliary and renal elimination, and also the BBB penetration of several anticancer drugs, including TKIs.^1,2

The role of P-gp in the absorption of TKIs has been widely studied. Some TKIs (eg, crizotinib) are P-gp substrates, so inhibition or induction of P-gp due to concomitant administration of another drug could lead to clinically relevant DDIs.¹ Others such as pazopanib, lapatinib, and gefitinib directly inhibit P-gp activity, so could increase the bioavailability of concomitantly administered P-gp substrates.¹ Other ABC drug transporters such as Breast Cancer Resistance Protein (BCRP) and Multidrug Resistance-associated Protein 2 (MRP2 or ABC-C2) are also recognized for their potential for DDIs.^5,67–69 The role of the uptake solute carrier transporters (eg, OATP, OCTP) in transporter-mediated drug interactions with EGFR-TKIs is less well defined.^5,69,70

Clinically relevant DDIs with selected EGFR-TKIs

Gefitinib

Gefitinib was the first oral quinazoline compound to be referred to as a “selective” EGFR-TKI.⁷¹ In the US, gefitinib is indicated for first-line treatment of patients with metastatic NSCLC who harbor EGFR exon 19 deletions or exon 21 (L858R) substitution mutations, as detected by a test approved by the Food and Drug Administration (FDA).⁷²

Acid-reducing agents

Medications that cause significant sustained elevation in gastric pH, such as PPIs and H₂-receptor antagonists, may reduce the bioavailability, plasma concentration, and efficacy of gefitinib.^72,73 High doses of short-acting antacids may have similar effects if taken regularly close to the time of administration of gefitinib.^73,74 In a recent preclinical study, treatment with omeprazole (10–100 mg/kg orally [p.o.]) and vonoprazan (1–5 mg/kg p.o.) produced significant dose-dependent increases in gastric pH, and the AUC_0–3h of gefitinib (5 mg/kg, p.o.) declined with increasing pH.⁷⁵ In healthy male volunteers, use of the rapid-acting H₂-receptor antagonist ranitidine (at a dose of 450 mg p.o., that increased gastric pH to ≥5 for ≥4 hrs), taken 1 hr before a 250 mg dose of gefitinib, markedly reduced gefitinib exposure (geometric least-squares mean AUC_0–∞ reduced by 47%; C_max by 71%).^74,76 However, inhibitors of gastric secretion had no effect on the efficacy of gefitinib in patients with NSCLC or in those harboring EGFR-activating mutations.⁷⁷

Concomitant use of gefitinib with PPIs should be avoided but, if deemed essential, gefitinib should be taken 12 hrs before or after the PPI.⁷² Similarly, gefitinib should be taken 6 hrs before or after an H₂-receptor antagonist or an antacid.⁷²

CYP450-dependent metabolism

Gefitinib was metabolized at a similar rate when incubated in vitro with recombinant human CYP3A4 or CYP2D6, less efficiently with CYP3A5 and CYP1A1.⁷⁸ The range of metabolites produced by recombinant human CYP3A4⁷⁸ are similar to those generated by human liver microsomes, in which gefitinib was found to be rapidly and extensively metabolized.⁷⁹

O-desmethyl-gefitinib, the major metabolite of gefitinib in human plasma,⁸⁰ was formed mainly by recombinant human CYP2D6,⁷⁸ but in liver microsomes, O-desmethyl-gefitinib was only a minor product;⁷⁹ the metabolism of gefitinib was primarily dependent on CYP3A4 and was not notably reduced in microsomes from CYP2D6 poor metabolizers.⁷⁹ Gefitinib is excreted either unchanged or after metabolism.^73,79,80

Concomitant administration of gefitinib with strong CYP3A4 inhibitors may reduce metabolism and clearance of gefitinib, and may increase its plasma concentration.^72,73 This may be clinically important, as adverse reactions with gefitinib are related to dose and exposure.⁷³ In healthy volunteers, pretreatment with the potent CYP3A4 inhibitor itraconazole (200 mg QD for 12 days) prior to gefitinib (a single dose of 250 mg on day 4) led to an increase in gefitinib exposure (mean AUC) by up to 78%.^74,81 Patients receiving potent CYP3A4 inhibitors should be closely monitored for adverse reactions to gefitinib.⁷³ Notably, the effect of CYP3A4 inhibitors on gefitinib exposure may be greater in CYP2D6 poor metabolizers.⁷³

CYP3A4 inducers may increase gefitinib metabolism and reduce the plasma concentration and efficacy of gefitinib.^72,73 Thus, pretreatment with the strong CYP3A4 inducer rifampicin (600 mg QD for 16 days) prior to gefitinib (a single dose of 500 mg on day 10) led to a reduction in mean gefitinib AUC of up to 83%.⁸¹ Administration of gefitinib with the moderate-to-strong CYP3A4 inducer phenytoin led to a 26% reduction in C_max and a 47% reduction in AUC.⁸² A potential interaction was also reported between gefitinib and herbal medicines including ginseng (a CYP3A4/5 inducer); the patient was a non-responder but became a partial responder after discontinuation of the herbal medicines.⁸³ A similar interaction might be expected with St. John’s Wort (Hypericum perforatum), another CYP3A4 inducer.¹ Concomitant administration of gefitinib with CYP3A4 inducers (phenytoin, carbamazepine, rifampicin, barbiturates, St. John’s Wort, ginseng) should, therefore, be avoided, as treatment efficacy may be reduced.⁷³ If use with a moderate-to-strong CYP3A4 inducer is essential, the dose of gefitinib should be increased to 500 mg/day (provided no severe adverse drug reactions are apparent).⁷² The standard dose (250 mg/day) should be resumed after discontinuation of the CYP3A4 inducer.⁷² Patients taking gefitinib with potent CYP3A4 inhibitors should be carefully monitored, due to the potential for toxicity, while those taking CYP3A4 inducers should be monitored for reduced efficacy.^5,72

As previously noted, gefitinib is mainly metabolized by CYP3A4, and to a lesser extent by CYP2D6,⁷⁹ although the impact of CYP2D6 inhibitors on gefitinib PK has not been evaluated.^5,73 The role of CYP2D6 in the clearance of gefitinib has been evaluated in healthy volunteers genotyped for CYP2D6 status. In CYP2D6 poor metabolizers, O-desmethyl gefitinib was unmeasurable in plasma (confirming that production of this metabolite is mediated by CYP2D6), and gefitinib exposure was twofold higher than in extensive metabolizers.^72,84 The investigators suggested that the absence of metabolite is unlikely to be clinically relevant, as it contributes little to the overall activity of gefitinib, while poor metabolizers have higher exposure to unchanged gefitinib (this is also unlikely to lead to clinically significant changes in the safety and tolerability of gefitinib).⁸⁴ Consequently, prospective screening for CYP2D6 genotype is not warranted before starting gefitinib, and dose adjustments and changes in clinical management strategy are not required for CYP2D6 poor metabolizers.^72,73,84 Giving a potent CYP2D6 inhibitor concomitantly with gefitinib might also increase gefitinib exposure. Consequently, poor CYP2D6 metabolizers and patients who begin taking a CYP2D6 inhibitor together with gefitinib should be closely monitored for adverse reactions to gefitinib.^72,73

Gefitinib is also a weak inhibitor of CYP2D6 in vitro.^72,81 In patients with solid tumors, co-administration of gefitinib with the CYP2D6 substrate metoprolol led to a 35% increase in metoprolol exposure.⁸¹ This effect may be relevant to CYP2D6 substrates with a narrow therapeutic index⁷³ as it may be necessary to modify the dose of such agents when used concurrently with gefitinib.⁷³

Smoking status is not a relevant consideration for gefitinib.⁵

UDP-glucuronosyltransferases

Gefitinib demonstrated broad inhibition of UGT-mediated glucuronidation in vitro, particularly against UGT1A1, UGT1A7, UGT1A9, and UGT2B7 isotypes.⁶³ The risk of potential DDIs in vivo was predicted by calculating the ratios between the area under the plasma concentration–time curve in the presence and absence of inhibitor (AUCi/AUC). For gefitinib, the AUC ratio at the highest evaluated dose (700 mg/day) was less than 1.3 for the substrates of each inhibited UGT isoform. While the authors acknowledged that in vivo DDIs extrapolated from in vitro data should be interpreted with caution, they concluded that the use of gefitinib is unlikely to lead to clinically significant DDIs via inhibition of glucuronidation.⁶³

Transporter proteins

In vitro evidence indicates that gefitinib is a substrate of P-gp but according to the EMA assessment report for gefitinib, there is no evidence to suggest that this effect has clinical consequences.⁷⁴ Effects of other agents on P-gp are unlikely to influence gefitinib absorption, as P-gp is saturated at higher concentrations.⁷² In one study, gefitinib was reported to directly inhibit the function of P-gp in multidrug-resistant lung and breast cancer cells. The authors suggested that gefitinib may inhibit the excretion of P-gp substrate drugs and that potential DDIs should be evaluated,⁸⁵ although there is no evidence that this effect is clinically relevant. In a preclinical study, simultaneous administration of gefitinib dramatically increased the oral bioavailability of irinotecan,⁸⁶ and in children with refractory solid tumors, use of gefitinib led to a fourfold increase in the bioavailability of oral irinotecan versus historical controls, and significantly reduced the clearance of irinotecan and its active metabolite, SN-38. The authors suggested that these effects may have occurred via inhibition of ABC-G2 by gefitinib.⁸⁷ Gefitinib has been shown to inhibit BCRP in vitro,⁸⁸ but the clinical importance of this effect is unknown.^5,73

Other clinically relevant interactions

Increases in the international normalized ratio (INR) and/or the rate of bleeding events have occurred in patients taking warfarin and gefitinib concomitantly.^72,73,89 Patients taking this combination should be monitored regularly for changes in prothrombin time or INR.^72,73

Concomitant use of sorafenib reduced gefitinib exposure (C_max by 26%, AUC by 38%) via an unknown mechanism, whereas sorafenib exposure was unaffected.⁹⁰ In phase II clinical trials, concomitant use of gefitinib and vinorelbine exacerbated the neutropenic effect of vinorelbine,^73,91 although in a phase I/II trial of gefitinib plus vinorelbine and gemcitabine in patients with metastatic breast cancer the incidence of febrile neutropenia was not a major limiting factor.⁹²

Erlotinib

Erlotinib is a reversible EGFR-TKI that is approved by the FDA as first-line, maintenance, or second-line or subsequent treatment following progression after at least one prior chemotherapy regimen, in patients with metastatic NSCLC who harbor EGFR exon 19 deletions or exon 21 (L858R) substitution mutations, as detected by an FDA-approved test.⁹³

Acid-reducing agents

The solubility of erlotinib is pH-dependent and decreases above pH 5;^1,5,94 therefore, drugs that alter gastrointestinal pH could alter the solubility and absorption of erlotinib, leading to potentially clinically relevant changes in bioavailability.⁹⁴ In a recent preclinical study of concomitant treatment with omeprazole (10–100 mg/kg p.o.) and vonoprazan (1–5 mg/kg p.o.), both of which induced significant dose-dependent increases in gastric pH, the AUC_0–3h of erlotinib (5 mg/kg p.o.) decreased as pH increased.⁷⁵

Combining erlotinib with PPIs should be avoided.^93,94 In healthy volunteers, concomitant use of the PPI omeprazole (40 mg once daily [QD] for 7 days) led to a reduction in erlotinib exposure (46% reduction in AUC; 61% reduction in C_max).⁹⁵ Temporal separation of doses may not eliminate the interaction because PPIs affect the pH of the upper gastrointestinal tract for an extended period.⁹³

Co-administration of H₂-receptor antagonists can also reduce the efficacy of erlotinib. Concomitant use of the H₂-receptor antagonist ranitidine (300 mg QD for 5 days, given 2 hrs before erlotinib) led to a 33% reduction in erlotinib AUC and a 54% reduction in C_max.⁹⁵ Increasing the dose of erlotinib is unlikely to compensate for such reductions in exposure,⁹⁴ but when dosing was staggered (ranitidine was given as a divided dose of 150 mg twice daily [BID] and erlotinib was given 10 hrs after the previous evening dose of ranitidine and 2 hrs before the next morning dose of ranitidine), the reduction in erlotinib exposure was much less marked (15% reduction in AUC; 17% reduction in C_max).⁹⁵ Consequently, if ranitidine co-administration is considered, it should be used in a staggered manner; ie, erlotinib must be taken at least 2 hrs before or 10 hrs after the dose of ranitidine.^93,94

A retrospective analysis of 190 patients with advanced NSCLC indicated that concomitant use of gastric acid suppressants had no significant effect on plasma concentrations of erlotinib, progression-free survival (PFS) or overall survival (OS).⁹⁶ However, in another retrospective review of 544 patients with advanced NSCLC treated with erlotinib, both PFS and OS were significantly reduced in patients who took gastric acid suppressants compared with those who did not (median PFS 1.4 vs 2.3 months, p<0.001; median OS 12.9 vs 16.8 months, p=0.003).⁹⁷ The authors did not speculate about the reasons for the divergent results. According to the product label, if required, antacids should be taken at least 4 hrs before or 2 hrs after erlotinib.⁹⁴

Subsequently, two retrospective studies in patients with EGFR mutations receiving either of the first-generation EGFR-TKIs (erlotinib or gefitinib) also reported that use of acid suppressants had no adverse effects on median PFS and median OS.^98,99 Recently, however, a further retrospective observational study of NSCLC patients taking erlotinib or gefitinib found that median PFS was 84 days in patients taking acid suppressants, compared with 221 days in those not taking acid suppressants (p<0.0001); the type of acid suppressant used did not seem to be important.¹⁰⁰ In the earlier studies, the presence of activating EGFR mutations in a proportion of patients may have conferred increased sensitivity to EGFR TKIs. Consequently, when erlotinib and gefitinib were given concomitantly with an acid suppressant, despite the reduction in bioavailability, the concentrations of erlotinib and gefitinib achieved in plasma may have been sufficient to inhibit mutant EGFR.^98,100 The discrepancy between the outcomes of the earlier studies of erlotinib^96,97 might also be explained by a difference in the proportion of patients with EGFR mutations between the two studies, ie, in the study of Chu et al, the majority of patients did not exhibit EGFR mutations, so the outcomes of erlotinib therapy may have been more sensitive to differences in plasma levels of erlotinib.⁹⁷

CYP450-dependent metabolism

The metabolism of erlotinib is mediated predominantly by CYP3A4/3A5 in liver and intestine, and to a lesser extent by CYP1A2 and CYP2C8, as well as extra-hepatically by pulmonary CYP1A1 and CYP1B1 in tumor tissue.^78,101 The active metabolite, O-desmethyl erlotinib, subsequently undergoes oxidation and glucuronidation.^5,78,101 Extrahepatic metabolism by CYP3A4 in the intestine, CYP1A1 in the lungs, and CYP1B1 in tumor tissue may also contribute to the clearance of erlotinib.⁹⁴

Potent inhibitors of CYP3A4 reduce erlotinib metabolism, leading to an increase in plasma erlotinib concentrations.⁹⁴ Thus, concomitant use of erlotinib with the potent CYP3A4 inhibitor ketoconazole (200 mg p.o., BID for 5 days) led to an 86% increase in erlotinib AUC.^94,102 In an open-label, crossover study in male and female healthy volunteers, co-administration of erlotinib (100 mg p.o. on days 1 and 15) with the combined CYP3A4 and CYP1A2 inhibitor ciprofloxacin (750 mg BID on days 13–18) led to a 39% increase in erlotinib AUC and a (non-significant) 17% increase in C_max.^93,103 The EU label advises caution when combining erlotinib with ciprofloxacin or potent CYP1A2 inhibitors such as fluvoxamine; if severe adverse events occur, the dose of erlotinib should be reduced.⁹⁴ The US label recommends against concurrent use of erlotinib with strong CYP3A4 inhibitors or CYP3A4/1A2 inhibitors, but if this is unavoidable, and if severe adverse reactions occur, it suggests reducing the dose of erlotinib in 50-mg steps.⁹³

Potent CYP3A4 inducers increase erlotinib metabolism and reduce plasma erlotinib concentrations.¹⁰⁴ Thus, pretreatment with rifampicin (600 mg p.o., QD for 7 days) led to a 69% reduction in the median AUC of erlotinib.⁹⁴ Concomitant use of erlotinib with CYP3A4 inducers should be avoided; alternatively, the use of a higher dose of erlotinib may be considered (300 or 450 mg, compared with the standard dose of 150 mg).^93,94 Patient safety should be closely monitored (including renal and liver function and serum electrolytes).⁹⁴

Other strong and moderate CYP3A4 inducers (eg, enzalutamide, phenytoin, carbamazepine, barbiturates, and herbal preparations containing St. John’s Wort) may also reduce erlotinib exposure.^1,94 Caution is advised when using these agents concomitantly with erlotinib, and alternative treatments should be considered when possible.⁹⁴

Cigarette smoking has been shown to markedly reduce erlotinib exposure via an increase in CYP1A1/1A2 activity. In healthy volunteers, the geometric mean of erlotinib AUC_0-∞ following a single 150 mg dose was 2.8-fold lower in smokers than in non-smokers, and was similar to that in non-smokers after a dose of 300 mg. C_max in smokers was two-thirds of that in non-smokers, and C_24h was 8.3-fold lower than in non-smokers.¹⁰⁵

In patients with solid tumors, clearance of erlotinib was 24% faster in current smokers than former smokers/never-smokers.¹⁰⁶ The increase in clearance seems to be related to induction of CYP1A1/1A2 in smokers.² In current smokers with NSCLC, the response rate to erlotinib was markedly lower than that in never-smokers (3.9% vs 24.7%; p<0.001).¹⁰⁷ In another study, the maximum tolerated dose (MTD) of erlotinib in current smokers with NSCLC was 300 mg/day¹⁰⁸ (ie, double the MTD previously established in unselected patients), although the authors did not speculate on the underlying mechanisms.¹⁰⁹ However, the efficacy and long-term safety of doses greater than the recommended starting doses have yet to be established in patients who continue to smoke.⁹⁴ In current smokers with locally advanced or metastatic NSCLC, a dose of erlotinib of 300 mg/day led to higher plasma concentrations than were achieved by the standard dose of 150 mg/day, but no incremental efficacy benefit was demonstrated.¹¹⁰ Patients should be encouraged to stop smoking as soon as possible before initiating erlotinib.⁹⁴ The US label advises increasing the dose of erlotinib in current smokers to 300 mg (maximum), returning immediately to the recommended dose (150 or 100 mg/day) on cessation of smoking.⁹³ The US label also recommends against the use of erlotinib together with moderate CYP1A2 inducers.⁹³

Erlotinib is itself a potent inhibitor of CYP1A1 and a moderate inhibitor of CYP3A4 and CYP2C8.⁹⁴ The physiologic relevance of CYP1A1 inhibition by erlotinib is unclear, given the limited expression of CYP1A1 in humans.⁹⁴ Pretreatment with, or co-administration of, erlotinib did not alter the clearance of the CYP3A4 substrates midazolam and erythromycin, but reduced the oral bioavailability of midazolam.⁹⁴ In another study, concomitant use of erlotinib did not affect the PK of paclitaxel (a CYP3A4/2C8 substrate).¹¹¹ While the EU label suggests that clinically relevant effects of erlotinib on the PK profiles of other CYP3A4 substrates are unlikely,⁹⁴ case reports suggest a need for caution during use of erlotinib together with CYP3A4 or CYP2C8 substrates. These include a case of rhabdomyolysis due to increased simvastatin exposure in a patient receiving concomitant erlotinib.¹¹² In another patient, toxicities that occurred during use of phenytoin were exacerbated following addition of erlotinib.¹¹³ Clinicians are advised to be aware of these potential interactions when combining these drugs with erlotinib and to proceed with caution.¹

UDP-glucuronosyltransferases

Erlotinib is a selective and potent competitive inhibitor of glucuronidation by UGT1A1 in vitro and exerts potent mixed inhibition of bilirubin glucuronidation in human liver microsomes.⁶³ Based on these findings, co-administration of erlotinib (≥100 mg/day) is predicted to increase the AUC of drugs predominantly cleared by UGT1A1 by ≥30% and to cause clinically significant DDIs when given with such agents.⁶³ Cheng et al (2017) found erlotinib to be a potent noncompetitive inhibitor of UGT1A1.⁶² Patients with low UGT1A1 expression or genetic glucuronidation disorders (eg, Gilbert’s disease) could develop high serum concentrations of bilirubin and must be treated with caution.⁹⁴

Transporter proteins

In vitro transport studies demonstrated that erlotinib is a substrate for, and inhibitor of, both P-gp and BCRP.^114–116 Concomitant administration of P-gp inhibitors such as cyclosporine and verapamil may lead to altered distribution and/or altered elimination of erlotinib.⁹⁴ The clinical relevance of this interaction is unclear,⁵ but clinicians should be aware of the potential for an increase in adverse events when using erlotinib in the presence of P-gp inhibitors.⁹⁴

Erlotinib and its active metabolite OSI-420 are substrates for human organic anion transporter 3 (OAT3) and, to a lesser extent, organic cation transporter 2 (OCT2)¹¹⁶ but the clinical implications of these properties have not been fully elucidated.²

Other clinically relevant interactions

Erlotinib can increase the INR in patients taking warfarin.¹¹⁷ Bleeding events have been reported (including peptic ulcer bleeding, hematemesis, hematochezia, melena, and hemorrhage from possible colitis),⁹³ some of which were fatal.⁹⁴ Patients taking erlotinib with coumarin-derived anticoagulants such as warfarin should be monitored regularly for changes in prothrombin time or INR.^93,94 Dose modifications are not recommended for erlotinib.⁹³

Concomitant use of capecitabine may increase plasma erlotinib concentrations. When combined with capecitabine, there was a significant increase in erlotinib AUC and a borderline increase in C_max (compared with concentrations measured in another study of erlotinib monotherapy).⁹⁴

In patients with advanced solid tumors, carboplatin exposure was reported to increase when administered concomitantly with erlotinib.¹¹¹ However, in an intensive PK study of patients with advanced NSCLC who had participated in a phase III trial of first-line erlotinib plus chemotherapy, the use of erlotinib did not alter systemic exposure of paclitaxel and carboplatin compared with that in the placebo group.¹¹⁸

Afatinib

Afatinib is an oral, irreversible inhibitor of the ErbB family of tyrosine kinases. Afatinib downregulates ErbB signaling by covalently binding to the kinase domains of EGFR, HER2, and HER4, leading to irreversible inhibition of tyrosine kinase autophosphorylation; afatinib also inhibits transphosphorylation of HER3.¹¹⁹ Afatinib is approved by the FDA for first-line treatment of patients with metastatic NSCLC who harbor nonresistant EGFR mutations, as detected by an FDA-approved test. It is also approved for locally advanced or metastatic NSCLC of squamous histology progressing after platinum-based chemotherapy.^120,121

Acid-reducing agents

Afatinib is highly soluble throughout the physiologic pH range (1–7.5).¹²² Consequently, interactions with acid-reducing drugs are not expected.⁵

Cytochrome P450-dependent metabolism

Afatinib undergoes minimal biotransformation, and oxidative CYP-mediated metabolism is of negligible importance.^5,123 Metabolism is mainly governed by non-enzyme catalyzed formation of adducts to proteins and nucleophilic small molecules.^5,123 Consequently, DDIs arising from inhibition or induction of CYP450 enzymes by concomitant medications are unlikely to occur.¹²¹ Smoking status has no significant effect on exposure to afatinib.^5,124

Transporter proteins

Afatinib is a substrate and inhibitor of P-gp in vitro,^5,125 and concomitant use of strong P-gp inhibitors can increase exposure to afatinib.^120,121 In healthy subjects, ritonavir (a strong inhibitor of P-gp and BCRP) given simultaneously or 6 hrs after a single 40 mg dose of afatinib led to minimal increases in afatinib AUC_0–∞ and C_max (by 5% and 11% respectively).¹²⁵ However, in a second study, ritonavir given 1 hr before a single 20 mg dose of afatinib led to a 48% increase in afatinib AUC_0–∞ and a 39% increase in C_max.¹²⁵ Conversely, strong P-gp inducers can reduce exposure to afatinib.¹²⁰ In healthy subjects, pretreatment with the potent P-gp inducer rifampicin (600 mg QD for 7 days) before a single 40 mg dose of afatinib led to a reduction in plasma exposure (34% reduction in AUC_0–∞ and 22% reduction in C_max).^121,125

For patients taking afatinib who require treatment with a P-gp inhibitor, the EMA label recommends using staggered dosing to maximize the interval between the doses of afatinib and the P-gp inhibitor (preferably 6 hrs for P-gp inhibitors dosed BID and 12 hrs for those given QD).¹²⁰ According to the US label, if a patient taking a concomitant P-gp inhibitor experiences toxicities while taking afatinib, their clinician may reduce the afatinib dose by 10 mg, and resume the original dose after discontinuation of the P-gp inhibitor, provided tolerability is acceptable.¹²¹ For those taking a P-gp inducer, the afatinib dose may be increased by 10 mg, subject to tolerability, and the original dose may be resumed 2–3 days after the P-gp inducer is discontinued.¹²¹

Afatinib is a moderate inhibitor of P-gp in vitro,^5,125 but clinical data suggest that changes in plasma concentrations of other P-gp substrates are unlikely to occur due to concomitant administration of afatinib.^120,126

Afatinib is both a substrate and an inhibitor of BCRP in vitro^120,121,126 and may increase the bioavailability of BCRP substrates administered orally, such as rosuvastatin and sulfasalazine.¹²⁰

Osimertinib

Osimertinib is a third-generation potent irreversible EGFR-TKI that has efficacy in patients with advanced NSCLC with EGFR mutations (both sensitizing/activating mutations (del19/L858R) and T790M resistance mutations).⁴⁷ In the USA, osimertinib is indicated for the treatment of patients with metastatic EGFR T790M mutation-positive NSCLC, as detected by an FDA-approved test, who have progressed on or after EGFR TKI therapy, and also for first-line treatment of patients with metastatic NSCLC whose tumors have EGFR exon 19 deletions or exon 21 L858R mutations, as detected by an FDA-approved test.¹²⁷ Similarly, in the EU, osimertinib is indicated for the treatment of adults with locally advanced or metastatic EGFR T790M mutation-positive NSCLC; also for first-line treatment of adults with locally advanced or metastatic NSCLC with activating EGFR mutations.¹²⁸

Acid-reducing agents

In a preclinical study, the AUC_0–3h of osimertinib (5 mg/kg p.o.) was not significantly affected by concomitant omeprazole (10–100 mg/kg p.o.) or vonoprazan (1–5 mg/kg p.o.), both of which caused significant dose-dependent increases in gastric pH.⁷⁵ In an open-label study in healthy male volunteers (n=68), co-administration of omeprazole did not significantly alter osimertinib exposure: the geometric least-squares mean ratio [90% CI] for AUC was 107% [100–113%] and for C_max was 102% [95–109%].¹²⁹

In patients whose gastric pH may be altered by concomitant agents or medical conditions, dose modifications are not required for osimertinib.¹²⁹ Gastric pH-modifying agents can be used with osimertinib without restriction.¹²⁸

CYP450-dependent metabolism

In vitro studies indicate that osimertinib is predominantly metabolized by CYP3A4/5 and is a weak inducer of CYP3A.¹³⁰ Hence, modulators of CYP3A could impact osimertinib metabolism, while osimertinib may alter the exposure of other CYP3A substrates.¹³⁰

Drug interaction studies with inhibitors, inducers, or substrates of CYP enzymes and transporters have not been conducted systematically for osimertinib.¹²⁷ The effect of strong CYP3A4 inhibitors and inducers on the PK of osimertinib in patients with advanced NSCLC was investigated in two open-label studies.¹³¹ In the first study of 36 patients, concomitant use of the strong CYP3A4 inhibitor itraconazole (200 mg BID; days 6–18) together with osimertinib (80 mg/day, days 1 and 10) had no clinically significant effect on osimertinib exposure; AUC increased by 24% and C_max decreased by 20% versus osimertinib given alone^128,131 (the upper bounds of the 90% CIs of the geometric mean least square mean treatment ratios [itraconazole + osimertinib/osimertinib alone] for AUC and C_max were both below the pre-specified “no-effect” limit of 200%). Similarly, there were no clinically relevant changes in exposure parameters for the active metabolite of osimertinib, AZ5104. The authors suggested that preclinical hepatocyte and recombinant CYP studies¹³² may have overestimated the contribution of cytochrome P450 metabolism to the clearance of osimertinib in the clinic, whereas the availability of multiple elimination pathways for osimertinib might also explain the lack of significant effects of itraconazole.¹³¹ The minor reduction in osimertinib C_max and increase in AUC were interpreted as being due to inhibition of CYP3A by itraconazole, leading to changes in elimination of osimertinib and its metabolites. Osimertinib t_max was 2 hrs longer in patients taking osimertinib plus itraconazole than those taking osimertinib alone (p=0.0002), which suggests that concomitant use of itraconazole may alter the absorption of osimertinib.¹³¹

In the second study (n=40), concomitant use of osimertinib (80 mg/day, days 1–77) with the CYP3A4 inducer rifampicin (600 mg/day, days 29–49) led to a 78% reduction in osimertinib AUC as well as an 82% reduction in AUC and 78% reduction in C_max of AZ5104, the metabolite of osimertinib.^128,131 Although the proportion of white patients was slightly higher in the rifampicin study than the itraconazole study, the authors felt that this was unlikely to have affected the results given that osimertinib exposure does not appear to be affected by ethnicity.¹³³

Consistent with the US and EU labels,^127,128 the researchers concluded that osimertinib can be given concurrently with CYP3A4 inhibitors but that strong CYP3A inducers should be avoided if possible.¹³¹ Concomitant use of St. John’s Wort with osimertinib is specifically contraindicated in the EU.¹²⁸ The US label advises that if concurrent use of a strong CYP3A4 inducer is unavoidable, the dose of osimertinib should be increased to 160 mg/day; the standard dose (80 mg/day) may be resumed 3 weeks after discontinuation of the CYP3A4 inducer.¹²⁷ Moderate CYP3A4 inducers may also reduce osimertinib exposure, so should be used with caution or avoided if possible.¹²⁸ No dose adjustments are required when osimertinib is used with moderate and/or weak CYP3A inducers.¹²⁷

In patients with EGFR mutation-positive NSCLC following disease progression on a prior EGFR-TKI, daily administration of osimertinib increased rosuvastatin exposure but had minimal effects on exposure of the sensitive CYP3A4 substrate, simvastatin.¹³⁴ Clinically relevant interactions between osimertinib and CYP3A4 substrates are therefore unlikely.¹²⁸

In a population PK analysis based on data from 780 patients, Brown et al (2017) found that smoking status had no significant effect on osimertinib PK (dose-normalized AUC at steady state),¹³³ which suggests that CYP1A1 induction does not have a major effect on osimertinib metabolism. Only 3% of the patients were current smokers, which limited the strength of the analysis. Nevertheless, no dosage adjustments are required when treating current smokers with osimertinib.¹²⁸

UDP-glucuronosyltransferases

Based on in vitro studies, osimertinib is not an inhibitor of UGT1A1 or UGT2B7 at clinically relevant concentrations. Intestinal inhibition of UGT1A1 is possible, but the clinical impact is unknown.¹²⁸

Transporter proteins

Osimertinib is a substrate of P-gp and BCRP in vitro,¹³⁵ but this is unlikely to lead to clinically significant DDIs at clinically relevant doses.¹²⁸ Osimertinib is a competitive inhibitor of BCRP transporters in vitro,^128,130 and it may, therefore, increase the exposure of BCRP substrates.^128,130 Concomitant administration of osimertinib led to a 35% increase in the AUC and a 72% increase in the C_max of rosuvastatin (a sensitive BCRP substrate).¹³⁴ Patients taking osimertinib with medications that have a narrow therapeutic index and BCRP-dependent disposition should be closely monitored for changes in tolerability due to increases in osimertinib exposure.^127,128 At clinically relevant concentrations, osimertinib is not a substrate or inhibitor of OATP1B1 or OATP1B3 in vitro.¹²⁸

Icotinib

Icotinib is a second-generation reversible EGFR-TKI, approved by the China Food and Drug Administration (CFDA) for the treatment of advanced NSCLC following progression on at least one platinum-based chemotherapy.^{18,29,136,137} To date, relatively few studies have reported on potential DDIs with icotinib.¹³⁸

CYP450-dependent metabolism

A preclinical pharmacokinetic study and a clinical mass balance study showed that more than 90% of icotinib is eliminated by hepatic metabolism, primarily via CYP450 enzymes; four to six main metabolites were identified.¹³⁹ The main enzymes responsible for icotinib metabolism are CYP3A4, CYP2C19, CYP3A5, and CYP1A2.^138,139 According to Shi et al (2013), the involvement of several enzymes in the metabolism of icotinib means that accumulation of the drug is limited, and explains its relatively short half-life (6 h¹⁴⁰), which is one of the main differences between icotinib and the other EGFR-TKIs.⁴⁰

Zhang et al (2018) recently examined the formation of icotinib metabolites by recombinant CYP isozymes in human liver microsomes, to identify the enzymes responsible for icotinib metabolism.¹³⁸ The metabolic pathways identified in vitro predominantly involved CYP3A4 (accounting for 77–87% of icotinib metabolism), CYP3A5 (5–15%), and CYP1A2 (3.7–7.5%). Metabolism of icotinib via CYP450 2C8, 2C9, 2C19, and 2D6 was insignificant. The authors recommended that clinicians should consider the risk of DDIs when prescribing icotinib with strong CYP3A inhibitors or inducers.¹³⁸ Induction of CYP1A2 in lung cancer patients with smoking history may also contribute to the PK and pharmacologic variability of icotinib.¹³⁸

Chen et al (2015) used a physiologically based PK model (validated using data from a phase I trial of icotinib in healthy Chinese subjects) to simulate DDIs with ketoconazole and rifampin (a potent CYP3A4 inhibitor/inducer, respectively). The model-predicted exposure (AUC) for icotinib was higher when given with ketoconazole (400 mg) and rifampin (600 mg) than when given alone; the AUC ratio for icotinib during concomitant use of ketoconazole and rifampin was 3.22 and 0.55, respectively.¹³⁹

UDP-glucuronosyltransferases

Cheng et al (2017)⁶² found both icotinib and erlotinib to be non-competitive inhibitors of UGT1A1, but the effect of icotinib was weaker than that of erlotinib (IC₅₀ for inhibition of UGT1A1-mediated NCHN-O-glucuronidation in human liver microsomes was 5.15 μmol/L for icotinib vs 0.68 μmol/L for erlotinib). The authors concluded that use of icotinib is unlikely to lead to clinically significant DDIs due to inhibition of UGT1A1.⁶²

DDIs: aids for clinical decision-making

The growing awareness of the importance of DDIs in the treatment of cancer patients has been reflected by increases in the number and scope of the sources of information and guidance available to aid clinical decision-making.¹⁴¹ Some useful online resources are now available. “Oncology in Practice” (http://oncologypro.esmo.org/Oncology-in-Practice) from the European Society of Medical Oncology includes an overview of the main DDIs for the most frequently used TKIs, prophylaxis and treatment of these DDIs, and information for patients. “Drugs.com” includes a drug interaction checker at https://www.drugs.com/drug_interactions.html, which allows users to specify the agents in a prescription and to obtain information on potential interactions. “SiteGPR” (http://sitegpr.com/fr/) provides evidence-based advice (in French) on dose adjustments for patients with renal impairment, including dose adjustments due to DDIs.¹⁴¹ “Cancer Drug Interactions” (https://cancer-druginteractions.org/), which is endorsed by the British Oncology Pharmacy Association, allows users to select from a list of anticancer drugs and commonly prescribed concomitant medications and obtain information on whether a DDI is likely, together with the rationale and quality of evidence. Watson for Oncology (https://www.ibm.com/watson/health/oncology-and-genomics/oncology/) is an artificial intelligence system that extracts data from medical records and provides evidence-based treatment options tailored to the individual patient.^142,143 Lexicomp Online (http://www.lexi.com) and Micromedex 2.0 (http://micromedex.com) both include interactive tools for evaluation of drug interactions.¹⁴⁴ Interestingly, Muhič et al (2017) found that different DDI screening systems may differ significantly in their ability to detect clinically relevant DDI-related adverse drug reactions.¹⁴⁵ Notably, pharmacy information experts recommend that to address such questions, multiple sources of information should be consulted.¹⁴⁴

Summary and recommendations for clinical practice

The development of the EGFR-TKIs has changed the therapeutic landscape of NSCLC and raised expectations among both patients and physicians. The introduction of these drugs into clinical practice presents challenges for physicians, however, not least due to the risk of DDIs with some agents. When formulating individualized therapeutic strategies for their patients, physicians should be aware of how differences between the PK properties of the different EGFR-TKIs may affect the potential for DDIs and, consequently, the efficacy, optimum dose, and tolerability of the treatment regimen. Moreover, given that cancer patients are often highly polymedicated, physicians must always bear in mind the potential impact of concomitant medications when selecting treatment and addressing the management of side effects. The most important interactions for physicians to be aware of, in terms of their significance to the treatment of patients with NSCLC, are described in Table 2; recommended approaches to managing such interactions are summarized below.

Table 2.

Potentially clinically significant interactions^a for EGFR-TKIs approved for the treatment of NSCLC

Interaction with		Gefitinib	Erlotinib	Afatinib	Osimertinib	Icotinibb
Acid-reducing agents		+	+	–	–	NR
CYP3A4	Inducers	+	+	–	+	+
	Inhibitors	+	+	–	–	+
Smoking status		–	+	–	–	+
UDP-glucuronosyltransferase		–	–	–	–	–
Transporter proteins	P-gp	–	–	+	–	NR
BCRP	–	–	–	–	NR

Notes: ^aPotentially significant impact on clinical efficacy of the EGFR-TKI. ^bLimited published information. +, potentially clinically significant interaction; –, no evidence of a clinically significant interaction.

Abbreviations: CYP3A4, cytochrome P450 3A4; NR, not reported; EGFR-TKI, epidermal growth factor receptor-tyrosine kinase inhibitor; UDP, uridine diphosphate.

Clinically significant interactions with acid-suppressive drugs (PPIs, H₂-receptor antagonists, and antacids) have been demonstrated for EGFR-TKIs that exhibit pH-dependent solubility (ie, gefitinib and erlotinib). When used concomitantly, bioavailability may be reduced, to such an extent that clinical efficacy may be significantly impaired. As patients taking EGFR-TKIs often experience gastrointestinal side effects and routinely use acid-reducing agents for palliation of gastro-esophageal reflux, dyspepsia, gastritis, and mucositis, these DDIs are clinically relevant. If concomitant use is unavoidable, then staggering the dose of the EGFR-TKI and the acid suppressant by several hours may help to reduce the extent of the interaction. Another option would be to prescribe afatinib, which is not subject to this type of interaction (no information is available for icotinib). To manage a clinically significant DDI, a twice-daily PPI could be replaced by a once-daily regimen. Giving the EGFR-TKI 2 hrs before the PPI (and using an enteric-coated formulation of the PPI) should optimize absorption of the EGFR-TKI.

Physicians should exercise caution when prescribing a known CYP enzyme inhibitor or inducer, as concomitant drugs may need to be substituted or the doses adjusted to account for potential reductions or increases in CYP enzyme-mediated metabolism. In addition to the agents mentioned earlier, commonly used CYP450 inhibitors include: amiodarone, cimetidine (1A2); clarithromycin, diltiazem, grapefruit juice, telithromycin (3A4/3A5); amiodarone, fluconazole, fluoxetine, metronidazole, trimethoprim, sulfamethoxazole (2C9); isoniazid (2C19); and amiodarone, cimetidine, diphenhydramine, fluoxetine, paroxetine, quinidine, terbinafine (2D6). Furthermore, phenobarbital is an inducer of CYP1A2, 3A4/5, and 2C9. A table of important substrates, inhibitors, and inducers (with direct links to PubMed citations) is continually updated by Indiana University School of Medicine and can be accessed at https://drug-interactions.medicine.iu.edu/Home.aspx.¹⁴⁶

Gefitinib, erlotinib, osimertinib, and icotinib are predominantly metabolized by CYP3A4; consequently, concomitant administration with a potent CYP3A4 inhibitor may substantially increase plasma concentrations of the EGFR-TKIs. Conversely, co-administration with a strong CYP3A4 inducer may increase EGFR-TKI metabolism, reduce plasma concentrations, and consequently reduce efficacy. Clinicians should take care when treating patients with CYP3A4 inducers, and if an interaction is anticipated, concomitant administration should be avoided if possible.

Plasma concentrations of erlotinib are markedly reduced in smokers; therefore, while receiving erlotinib, current smokers should be advised to stop smoking. For those who continue to smoke, it may be necessary to increase the dose of erlotinib (to a maximum of 300 mg/day). Induction of CYP1A2 in smokers may also influence the metabolism of icotinib.

Gefitinib, erlotinib, osimertinib, and afatinib are substrates for the drug transporter P-gp in vitro (no information is available for icotinib), but clinical findings indicate that clinically relevant DDIs may occur with afatinib only (not with gefitinib, erlotinib or osimertinib). Clinicians should consider staggering or adjusting the dose of afatinib when used in combination with a P-gp inhibitor or inducer.

Abbreviation list

ABC, ATP-binding cassette; AUC, area under the plasma concentration–time curve; BBB, blood–brain barrier; BID, twice daily; C_max, maximum plasma concentration; CYP, cytochrome P450; DDI, drug–drug interaction; EGFR, epidermal growth factor receptor; EMA, European Medicines Agency; FDA, Food and Drug Administration; INR, international normalized ratio; MTD, maximum tolerated dose; NR, not reported; NSCLC, non-small-cell lung cancer; OAT3, organic anion transporter 3; OATP, organic anion-transporting peptide; OCT2, organic cation transporter 2; OCTP, organic cation-transporting peptide; OS, overall survival; PFS, progression-free survival; P-gp, P-glycoprotein; PK, pharmacokinetic[s]; p.o., orally; PPI, proton pump inhibitor; QD, once daily; TKI, tyrosine kinase inhibitor; UDP, uridine diphosphate; UGT, UDP-glucuronosyltransferase.

Disclosure

Dr Junling Li reports grants from Boehringer Ingelheim for medical writing assistance during the preparation of this article. Dr Ziyi Xu reports grants from Boehringer Ingelheim for medical writing assistance during the preparation of this article. The authors report no other conflicts of interest in this work.