FormalPara Key Points

Polypharmacy for age-related comorbidities is common in patients with nonmetastatic castration-resistant prostate cancer, and this increases the risk of drug–drug interactions (DDIs).

Subanalyses of the phase III ARAMIS trial indicate that darolutamide has low risk of clinically relevant DDIs with comedications commonly used in these patients.

1 Introduction

A prime goal of treatment for patients with nonmetastatic castration-resistant prostate cancer (nmCRPC) is to delay progression to metastatic disease (mCRPC) since increased tumor burden is associated with worse prognosis [1], although the recent introduction of multiple life-extending therapies for mCRPC (radium-223, abiraterone, enzalutamide, sipuleucel-T, and taxanes) has extended post-progression survival [2]. Along with disease progression, adverse events (AEs) associated with anticancer therapies may have a detrimental impact on patients’ quality of life [3], particularly in patients with nmCRPC, given the mostly asymptomatic nature of their disease. As nmCRPC is generally diagnosed in older men who are often prescribed multiple concomitant medications for comorbid conditions [4,5,6], the risk of AEs or loss of efficacy resulting from drug–drug interactions (DDIs) is also a fundamental component in clinical decision making [7].

The second-generation androgen receptor inhibitors enzalutamide and apalutamide have recently been approved for the treatment of nmCRPC after demonstrating significantly prolonged metastasis-free survival in phase III clinical trials [8, 9]. However, this change in the treatment landscape has also brought decision-making challenges. Apalutamide and enzalutamide are associated with higher incidences of certain AEs (e.g., falls, fatigue, hypertension, rashes, and seizures) than are placebo. They also both have the potential for DDIs when coadministered with medications that are substrates for several metabolizing enzymes and drug transporters [8,9,10,11,12,13,14]. For example, coadministration of either drug may reduce the exposure of medications that are substrates for the cytochrome P450 (CYP) enzymes CYP3A4, CYP2C9, and CYP2C19 [15, 16], resulting in loss of activity of common comedications such as anticoagulants, antihypertensives, opioid analgesics, and proton pump inhibitors (PPIs).

Darolutamide is a novel androgen receptor antagonist with a unique molecular structure recently approved by the US FDA for the treatment of nmCRPC [17, 18]. In the phase III ARAMIS trial (NCT02200614), darolutamide demonstrated significantly prolonged metastasis-free survival compared with placebo (40.4 vs. 18.4 months; hazard ratio for metastasis or death in the darolutamide group 0.41; 95% confidence interval [CI] 0.34–0.50; p < 0.001) in patients with nmCRPC receiving androgen-deprivation therapy and delayed time to pain progression (40.3 vs. 25.4 months; hazard ratio 0.65; 95% CI 0.53–0.79) [19]. The DDI potential of darolutamide has been extensively investigated in preclinical and clinical phase I studies with metabolizing enzymes and drug transporters. Darolutamide was sensitive to induction of CYP3A4, the main enzyme mediating its metabolism. Inhibition of CYP3A4, P-glycoprotein (P-gp) and breast cancer resistance protein (BCRP)—both transporters for which darolutamide is a substrate—by comedications was not considered clinically relevant. No relevant inhibition of any CYP enzymes by darolutamide was observed in vitro, and in vivo studies found weak induction of CYP3A4 only, suggesting darolutamide may have no clinically relevant effect on comedications metabolized by CYP enzymes [20]. Therapeutic levels of darolutamide were found to have no effect on P-gp substrates [20], though exposure of rosuvastatin, a substrate for BCRP, organic anion-transporting peptide (OATP)1B1, and OATP1B3, was moderately increased by darolutamide coadministration [21]. This effect was mainly attributed to inhibition of BCRP; however, a contribution of OATP inhibition is considered likely. Overall, the results of these studies indicated that darolutamide has a favorable DDI profile.

To evaluate the impact of clinically relevant DDIs with darolutamide in the target patient population, we conducted prespecified and post hoc analyses in patients with nmCRPC who participated in the phase III ARAMIS trial [19], including assessments of comorbidities and comedication use, the effects of concomitant medications on the pharmacokinetics of darolutamide, and the impact of coadministering darolutamide with concomitant medications on patient safety. Given that statin use is very common in this population and darolutamide has been observed to increase rosuvastatin exposure [21], we conducted a subanalysis of the impact of using concomitant statins with darolutamide on patient safety.

2 Patients and Methods

2.1 Trial Design

The phase III ARAMIS trial (NCT02200614) protocol and design have been described in detail elsewhere [19]. Briefly, this was a randomized, double-blind, placebo-controlled, phase III clinical trial conducted globally in 409 centers in 36 countries to evaluate the efficacy and safety of darolutamide in patients with nmCRPC. The trial was approved by local institutional review boards and was conducted in compliance with the principles of the Declaration of Helsinki and in accordance with the International Conference on Harmonisation guidelines for Good Clinical Practice. All patients provided written informed consent.

2.2 Treatment

Men were randomized 2:1 to receive darolutamide 600 mg twice daily (given as two 300 mg tablets) or placebo while continuing androgen-deprivation therapy. Patients continued study medications until protocol-defined progression, discontinuation due to AEs, or withdrawal of consent.

2.3 Population Pharmacokinetic Model

An exploratory population pharmacokinetic analysis was performed to evaluate the variability in the pharmacokinetics of darolutamide, its two pharmacologically active diastereomers, (S,R)-darolutamide and (S,S)-darolutamide, and the active metabolite, keto-darolutamide, based on data generated from the phase III ARAMIS study [19]. The population pharmacokinetic model was developed using data from a subset of patients with valid pharmacokinetic samples collected sparsely at several times during the 12-h application interval (detailed in the Electronic Supplementary Material [ESM]—Methods) and information on drug disposition pathways identified preclinically. The bioanalytical methods used for pharmacokinetic sampling are summarized in the ESM—Methods.

All predefined concomitant medications taken by at least one patient during study treatment were included as a covariate in the population pharmacokinetic model (Table 1 in the ESM). The impact of BCRP inhibitors, CYP3A4 inducers/inhibitors, P-gp inducers/inhibitors, PPIs, and/or uridine 5′-diphospho-glucuronosyltransferase (UGT) inducers/inhibitors on the pharmacokinetics of darolutamide was investigated. Comedications with any relevant DDI potential, weak or strong, were included.

2.4 Safety Analyses

At each study visit, data on treatment-emergent AEs (TEAEs), including type, severity, seriousness, and whether they were related to study treatment based on investigator assessment, were recorded. A subgroup analysis was also performed to compare the incidence of TEAEs for patients receiving statins (3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors) versus those not receiving statins. This analysis included all TEAEs reported in concomitant statin users versus nonusers regardless of when the event occurred during the study, i.e., without consideration for when darolutamide and the statin were coadministered. A post hoc analysis evaluating TEAE incidence in patients using statins known to be BCRP substrates versus nonusers of BCRP-substrate statins was also performed.

2.5 Statistical Analysis

Patient demographics, comorbidities, and concomitant medications were summarized descriptively for all randomized patients, i.e., the full analysis set. Comedications were categorized according to potential for relevant DDIs (CYP inducers/inhibitors/substrates, UGT substrates, and drug transporter substrates) based on in vitro and phase I clinical studies of darolutamide [21] and established DDI profiles for the currently available androgen-receptor inhibitors, enzalutamide and apalutamide [11,12,13,14].

Pharmacokinetic model development was conducted via nonlinear mixed-effects modelling using NONMEM (version 7.3, ICON Development Solutions plc, Manchester, UK) and R software [22]. Full details of the population pharmacokinetic analysis are given in the ESM—Methods.

Safety evaluations were conducted in all patients who were randomized and received at least one dose of study medication.

3 Results

3.1 Patients

In total, 1509 men were randomized to darolutamide (n = 955) or placebo (n = 554) in addition to androgen-deprivation therapy. One patient randomized to the darolutamide arm did not receive study treatment and was therefore excluded from the safety analyses [19]. Baseline demographic and clinical characteristics were similar in the two treatment groups (Table 1). Median age was 74 years in both groups, and median follow-up time was 17.9 months. Most patients had at least one comorbid condition (98.4% in both treatment arms); the most common comorbidities were hypertension, obesity, anomalies of lipid metabolism, and diabetes (Table 2).

Table 1 Demographics and baseline characteristics of patients with nonmetastatic castration-resistant prostate cancer in the phase III study
Table 2 Proportion of patients with ongoing comorbid conditions by preferred MedDRA term (> 5% of total population; full analysis set)

3.2 Concomitant Medication Use

The overwhelming majority of men in the trial were receiving at least one concomitant medication, and rates of comedication use were balanced across the two treatment arms: 98.7% with darolutamide and 98.0% with placebo (Table 3). The most common concomitant medications for comorbid conditions in darolutamide- versus placebo-treated patients were antihypertensives, including drugs affecting the renin–angiotensin system (54.7 vs. 49.8%), β-blockers (29.6 vs. 27.6%), and calcium channel blockers (22.7% in both arms), as well as analgesics (53.8 vs. 50.4%), whereas more than one-third of patients received antithrombotics (42.8 vs. 39.7%) or lipid-modifying agents (34.5 vs. 39.4%). Other cardiovascular agents, diuretics, urological agents, PPIs, systemic antibiotics, and anti-inflammatory/antirheumatic agents were used by at least 20% of patients. These comedications span a range of agents that act as inducers, inhibitors, or substrates of the enzymes predominantly involved in drug metabolism or key drug transporters (e.g., P-gp, BCRP, OATP1B1, OATP1B3) (Table 3).

Table 3 Concomitant medication use (> 10% of total patient population; full analysis set)

3.3 Population Pharmacokinetic Analysis

In total, 388 patients (40% of the total who received darolutamide in the phase III trial) were included in the pharmacokinetic subset. Demographics and baseline characteristics were generally similar to those of the overall darolutamide-treated population, although the pharmacokinetic subset contained higher proportions of Japanese patients and patients with mild renal impairment (Table 1).

Evaluation of the population pharmacokinetic analysis showed that none of the comedications included as covariates in the population pharmacokinetic base model met the criteria for a significant covariate effect. These comedications included CYP3A4 inhibitors (21.6%), P-gp inhibitors (20.1%), PPIs (15.2%), BCRP inhibitors (14.9%), UGT1A9 inhibitors (2.8%), and CYP3A4 inducers (0.5%); examples of drugs in these categories are given in Table 3. No patients received P-gp or UGT1A9 inducers. PPIs exerted no effect on darolutamide clearance. It was not possible to specifically assess an effect of CYP3A4 inducers as so few patients received comedications that were CYP3A4 inducers, such as rifampicin or carbamazepine. These data suggest that, although preclinical data indicate darolutamide exposure may be affected by drugs that are strong CYP3A4 and P-gp inducers or inhibitors [23], there was no significant association between the use of concomitant medications and the pharmacokinetic variability of darolutamide in this study.

3.4 Impact on Adverse Event Incidence of Drugs Coadministered with Darolutamide, Including Breast Cancer Resistance Protein Substrates

We also assessed whether the DDIs that have been identified as clinically relevant for darolutamide (notably, sensitivity to strong CYP3A4 inducers and inhibition of BCRP by darolutamide [23]) had a possible impact on the safety of patients in the ARAMIS trial population. As already noted, the common concomitant medications received by patients in ARAMIS were similar between study arms (Table 3). We have previously reported that the incidence of the most common AEs was similar in patients treated with darolutamide and those who received placebo, and permanent discontinuations of the study drug due to AEs occurred in similar proportions of patients in each treatment arm [19].

Concomitant use of statins was similar in the darolutamide and placebo arms (32.1 vs. 36.5%; Table 4). In both treatment arms, the incidence of TEAEs was comparable between patients receiving concomitant statins and those not receiving statins (88.2 vs. 80.9% for darolutamide and 82.2 vs. 73.9% for placebo; Table 4). With respect to the type of AEs reported (based on system organ class and preferred term), no differences among statin users in either treatment arm was observed. No meaningful differences in alanine aminotransferase (ALT) values occurred between concomitant statin users and nonusers: in both treatment arms, mean ALT values were slightly higher among statin users than nonusers at baseline and throughout the treatment period. In a subset of patients using statins that are BCRP substrates (approximately 30% of the overall population; Table 2 in the ESM), predefined AEs known to be associated with statins occurred more frequently with darolutamide than placebo (8.6 vs. 3.5%; Table 5). This difference was due to greater incidences of abnormalities in laboratory parameters. A similar difference was observed in the overall study population, indicating that the difference occurred independently of statin exposure and excluding any relevant impact of darolutamide in statin users. The abnormalities in laboratory parameters were notably increased aspartate aminotransferase levels (1.4% with darolutamide vs. 0.2% with placebo, respectively) and increased blood bilirubin levels (1.3% with darolutamide vs. 0% with placebo).

Table 4 Overview of treatment-emergent adverse events by concomitant statin use (safety population)
Table 5 Incidence of pre-defined treatment-emergent adverse events with concomitant use of statins that are breast cancer resistance protein substrates (safety population)

4 Discussion

The potential for DDIs is an important consideration when selecting an androgen receptor-targeting therapy for patients with nmCRPC. Loss of efficacy with treatments for patients’ comorbidities or sustaining an increased risk of AEs are potential outcomes of DDIs [10, 24]. Extensive investigation in preclinical/phase I studies indicated that darolutamide has a favorable DDI profile. Overall, few DDIs were identified at clinically relevant levels: darolutamide was sensitive to CYP3A4 induction, but no clinically relevant effects on darolutamide were expected from drugs that act as inhibitors of CYP enzymes and drug transporters.

Most importantly, darolutamide had few effects on other drugs used to treat comorbid conditions; increased exposure of rosuvastatin was the only interaction of note [23]. This interaction was mainly attributed to BCRP inhibition, although darolutamide-mediated inhibition of OATP transporters may have contributed to the observed effect. By contrast, enzalutamide is a strong CYP3A4 inducer and a moderate CYP2C9 and CYP2C19 inducer, and apalutamide is a strong inducer of CYP3A4 and CYP2C19 and a weak inducer of CYP2C9, UGT, P-gp, BCRP, and OATP1B1 [11, 14]. Additional monitoring may be required as both drugs may lead to suboptimal bioavailability of comedications with potential loss of efficacy.

The present analyses of the phase III ARAMIS trial demonstrated that almost all patients with nmCRPC receiving androgen-deprivation therapy with androgen receptor inhibitor therapy have comorbid conditions that require the use of concomitant medications. Individual patient data in ARAMIS indicated that many patients were receiving multiple comedications, as others have also reported [10]. As expected in this population (median age 74 years), the most common comorbidities included hypertension and other cardiovascular disorders, obesity, hyperlipidemia, osteoarthritis, and diabetes. Consistent with the comorbidities reported, the most frequently prescribed comedications were antihypertensives and agents for other cardiovascular disorders, along with a need for analgesia, treatments for urological and acid-related gastrointestinal disorders, antidepressants, anxiolytics, and dementia treatments. Our population pharmacokinetic analysis demonstrated that these frequently used concomitant drugs had no significant effect on the pharmacokinetics of darolutamide. Comedications that are potent CYP3A4 inducers were rarely used by the overall ARAMIS population, so the risk of DDIs may also be low in clinical practice.

Only the interaction between darolutamide and the combined BCRP and OATP substrate rosuvastatin has been deemed clinically relevant [23], though our safety subgroup analysis in statin users and nonusers could not clearly attribute any imbalance between the treatment arms to the effect of darolutamide on statins. However, the small sample size may preclude observation of rare AEs such as myopathy or rhabdomyolysis, although these are well-characterized and can readily be monitored. Overall, the incidence of AEs did not appear to be increased significantly by DDIs, as similar AE rates are reported with darolutamide and placebo despite the equally high use of concomitant medications in both treatment arms in the ARAMIS study [19]. This observation includes AEs previously reported more frequently with enzalutamide and apalutamide versus placebo in phase III trials (PROSPER and SPARTAN, respectively) such as falls, hypertension, and central nervous system (CNS)-related AEs [8, 9].

The risk of AEs resulting from DDIs is an important consideration for physicians. Enzalutamide and apalutamide have been identified as substrates, inducers, and/or inhibitors of metabolizing enzymes and inhibitors or inducers of drug transporters [15, 16, 25, 26]. Given that the majority of drugs are metabolized by at least one CYP enzyme [4], enzalutamide and apalutamide have the potential for DDIs with a range of medications that are commonly used by patients with prostate cancer, e.g., gastrointestinal, cardiovascular, and analgesic agents (Table 6) [10, 15, 16, 24,25,26]. Consequently, the FDA labels for enzalutamide and apalutamide suggest avoiding the use of susceptible comedications and warn of the potential loss of efficacy of these comedications, respectively [11, 14]. In a retrospective review of pharmacy records for patients with mCRPC, enzalutamide was associated with a high prevalence of potential DDIs with CNS drugs (e.g., opioid analgesics), which could increase the risk of cognitive adverse effects, falls, and fractures [10]. In addition, enzalutamide reduces plasma concentrations of PPIs, which increases the risk of bleeding events in patients receiving PPIs in combination with aspirin or nonsteroidal anti-inflammatory drug prophylaxis. Overall, 85% of patients with mCRPC who received enzalutamide were at risk for DDIs that would require treatment modification of at least one comedication [10]. In a post hoc analysis of the phase III SPARTAN trial, the risk of falls was increased in patients receiving apalutamide and concomitant α-blockers or antidepressants [27]. Hence, the safety of prescribing androgen receptor inhibitors in the setting of polypharmacy warrants consideration by the treating physician.

Table 6 Examples of potential drug–drug interactions between enzalutamide, apalutamide, or darolutamide and concomitant medications

Limitations of our study include the small sample size of the safety subanalyses and the population pharmacokinetic analysis and that the population pharmacokinetic analysis did not investigate the effect of darolutamide on the pharmacokinetics of comedications. Strong CYP3A4 and P-gp inducers were identified as having a relevant effect on darolutamide, but such comedications were rarely used in ARAMIS. For example, the strong CYP3A4-inducing anti-epileptic drugs, carbamazepine, phenytoin, and phenobarbital, were used by few patients. However, as patients with severe infections (which may be treated with systemic antibiotics such as rifampicin) and other serious comorbidities (e.g., recent stroke or myocardial infarction, active viral hepatitis, or human immunodeficiency virus) were excluded, the ARAMIS population did not fully represent the real-world population likely to receive darolutamide treatment. Multiple interactions between drugs were possible since many patients received several comedications, but it was not possible to assess the impact of these on patient safety in our analyses. Our identification of potential DDIs is not exhaustive—DDIs may occur with drugs not investigated in preclinical and phase I studies to date, and DDIs may be specific to individual drugs within a drug class. The safety profile of darolutamide will continue to be monitored in post-marketing studies.

5 Conclusions

These analyses of ARAMIS data suggest a limited potential for clinically relevant DDIs between darolutamide and comedications frequently used to treat age-related comorbidities in patients with nmCRPC (Table 6). Interactions noted with the combined BCRP and OATP substrate rosuvastatin did not seem to translate into increased AEs in the ARAMIS trial to date. Determining the risk–benefit balance of androgen receptor-targeted therapies for patients with nmCRPC who receive multiple comedications is important for optimal management. The findings presented here provide additional information for clinical decision making.