Abstract
Background
Pharmacokinetic studies suggest that specific prostate cancer therapies may alter the metabolism of direct oral anticoagulants (DOACs), leading to elevated risks of thrombosis or bleeding.
Methods
To assess the risk of thrombosis or bleeding, population‐based, retrospective, parallel analyses were conducted in Ontario and Alberta, Canada, among adults patients with prostate cancer who were prescribed a DOAC and a potentially interacting androgen receptor pathway inhibitor (including enzalutamide, apalutamide, or abiraterone) compared with non‐DOACs between 2012 and 2023. Analyses were stratified based on a DOAC‐inducer cohort (enzalutamide or apalutamide, which might increase the risk of thrombosis) and a DOAC‐inhibitor cohort (abiraterone, which might increase the risk of bleeding). Overlap weighted Cox proportional hazard models accounting for 37 covariates were performed independently in each jurisdiction and were pooled using random‐effects meta‐analysis.
Results
In total, 2997 individuals were included (2107 from Ontario and 890 from Alberta). In patients who received enzalutamide or apalutamide, there was no increased risk of all thrombosis in the DOAC groups compared with the non‐DOAC groups (pooled hazard ratio, 0.83; 95% confidence interval, 0.36–1.93). In patients who received abiraterone, no significant differences were observed in any bleeding events comparing the DOAC and non‐DOAC groups (pooled hazard ratio, 1.16; 95% confidence interval, 0.10–13.99). The results were consistent in multiple sensitivity analyses.
Conclusions
The concurrent use of enzalutamide, apalutamide, or abiraterone with DOACs did not contribute to clinically meaningful changes in the risk of thrombosis or bleeding in individuals with prostate cancer.
Keywords: androgen receptors, anticoagulants, drug interactions, hemorrhage, prostatic neoplasms, thrombosis
Short abstract
In patients with prostate cancer, direct oral anticoagulants (DOACs), compared with non‐DOACs, combined with enzalutamide or apalutamide were not associated with a higher risk of thrombosis. Similarly, DOACs combined with abiraterone were not associated with a higher risk of bleeding.
INTRODUCTION
Prostate cancer is the most common nonskin cancer in males, accounting for 20% of new cancer cases. 1 Approximately one in eight men will develop prostate cancer in their lifetime. 1 Androgen‐receptor pathway inhibitors (ARPIs), including enzalutamide, apalutamide, and abiraterone, are indicated for nearly all patients with advanced prostate cancer, accounting for 40%–50% of all patients who have prostate cancer. 2 , 3 Thromboembolism is the second leading cause of death in the cancer population. 4 Patients with prostate cancer have a 50% increased risk of thromboembolic diseases compared with the general population. 5 The risks further increase with hormonal therapies like ARPIs. 5 , 6 Anticoagulation is the main treatment for thromboembolism, but it is not without risks, with anticoagulant‐associated hemorrhage as a main complication. 7
Randomized controlled trials have demonstrated the efficacy and safety of direct oral anticoagulants (DOACs) for preventing thromboembolism in patients with atrial fibrillation (AF), 8 , 9 , 10 , 11 and for the prevention and treatment of venous thromboembolism (VTE) in patients with cancer. 12 , 13 , 14 , 15 , 16 , 17 , 18 Major guidelines have since endorsed the use of DOACs for the prevention of AF‐related strokes and for the treatment and prevention of cancer‐associated thrombosis. 19 , 20 , 21 , 22 DOACs are substrates of two predominant pathways: (1) P‐glycoprotein (P‐gp) cell transporters, and (2) cytochrome P450 enzyme (CYP3A4). 23 Apixaban and rivaroxaban are metabolized through both P‐gp and CYP3A4 systems, whereas dabigatran and edoxaban only rely on the P‐gp system. Concomitant medications that significantly induce P‐gp and/or CYP3A4 pathways can lead to a decrease in the concentration of DOACs and a potentially heightened risk of thrombosis. 24 , 25 Conversely, medications that significantly inhibit P‐gp and/or CYP3A4 pathways can lead to an increase in DOAC levels, with a potentially increased risk of bleeding. 24 , 25 Trials to date often exclude patients who are receiving medications that have potential significant interactions, thereby limiting the ability to evaluate the effect of potential drug–drug interactions (DDIs). On the other hand, given the low cost and clinicians' familiarity of warfarin, it continues to be a common oral anticoagulant prescribed worldwide. Studies have indicated that warfarin accounted for approximately 40% of anticoagulants, whereas DOACs accounted for approximately 50% of anticoagulants prescribed for patients with cancer and AF. 26 , 27 The ability to adjust warfarin dosages based on laboratory tests could also be an advantage for situations in which DDIs are of concern. In addition, low‐molecular‐weight heparin (LMWH) remains the preferred anticoagulant in patients with a high risk of bleeding or in the presence of DDI concerns. 19 , 20 , 21
Commonly used ARPIs, such as enzalutamide, apalutamide, and abiraterone, have potential DDIs with DOACs. Enzalutamide and apalutamide are strong inducers of the CYP3A4 and/or P‐gp pathways, 28 , 29 whereas abiraterone is a moderate inhibitor of both pathways. 30 However, the majority of available evidence to date is limited to in‐vitro pharmacokinetic studies, and the clinical implications are unclear. Therefore, we conducted a population‐based, retrospective cohort study to evaluate clinically relevant outcomes in patients receiving concurrent anticoagulants and these potentially interacting ARPIs.
MATERIALS AND METHODS
Data sources
We used linked databases in ICES (previously known as the Institute for Clinical Evaluative Sciences) in Ontario and in Alberta Health Services in Alberta that captured population‐level data for approximately 20 million individuals. ICES is an independent, nonprofit research institute whose legal status under Ontario's health information privacy law allows it to collect and analyze health care and demographic data, without consent, for health system evaluation and improvement. The use of ICES data was authorized under Section 45 of Ontario's Personal Health Information Protection Act (https://www.ices.on.ca/Data‐and‐Privacy/Privacy‐at‐ICES), and additional review by the ethics board was not required. For Alberta, the study was approved by Health Research Ethics Board of Alberta Cancer Committee (HREBA.CC‐24‐0043). Provincial cancer registries were used to identify cancer diagnosis and stage. Hospitalization data were obtained from the Discharge Abstract Database, which contains clinical and demographic data, and the National Ambulatory Care Reporting System, which contains emergency department visits (see Table S1). Health records linked to health administrative databases were used to capture the occurrence of major thrombotic or bleeding events (e.g., through emergency room, hospitalization, and physician claims databases). Medication claims include information on the drug identification number, trade name, therapeutic class, pill strength, quantity dispensed, and days supplied with a low error rate of 0.7%. 31 The data sets were linked using unique encoded identifiers and analyzed at ICES. We used identical cohort builds and common definitions and methods across both provinces, with minor modifications made as required by specific provincial databases.
Study cohort
In Ontario, the study population included older adults (aged 66 years and older) who filled an outpatient prescription for an anticoagulant (DOAC, LMWH, or warfarin) and an ARPI of interest (enzalutamide, apalutamide, or abiraterone) between October 1, 2012, and September 30, 2023. In Ontario, the filling of prescription drugs is captured for individuals older than 65 years under the universal Ontario Drug Benefit for seniors. Therefore, an age cutoff of 66 years was chosen to allow a minimum 1‐year look‐back period for medication prescription. In Alberta, all individuals have drug dispense data; therefore, all adult patients (aged 18 years and older) who were receiving the drugs of interest between March 31, 2012, and September 30, 2023, were included. The index date to enter the cohort was the date of concurrent use of an anticoagulant and the prostate cancer drug of interest. Patients were excluded if they (1) were receiving more than one anticoagulant at index or an anticoagulant other than the drugs of interest, (2) had at least one prescription for any strong CYP3A4 or P‐gp inhibitor (including itraconazole, ketoconazole, voriconazole, posaconazole, tacrolimus, cyclosporine, quinine, and rifampin) in the 90 days before the index date, (3) were on chronic dialysis, and (4) had a history of kidney transplantation before the index date.
Study design
This was a retrospective, population‐based cohort study. We included patients who were concurrently receiving an ARPI of interest (enzalutamide, apalutamide, or abiraterone) and an anticoagulant. We combined patients who were prescribed LMWH or warfarin into one group (given the small sample size in either group) and compared them with patients who were prescribed a DOAC. Given the distinct different types of potential DDIs, we identified two cohorts of interest: the inducer cohort (concurrent enzalutamide or apalutamide use) and the inhibitor cohort (concurrent abiraterone use). Many confounders were considered, including: age, index year, income, rural residence, history of bleeding, comorbidities within 5 years 32 (such as history of hypertension, diabetes, stroke/transient ischemic attack, AF, myocardial infarction, heart failure, coronary artery disease, angina, coronary artery bypass graft surgery, percutaneous coronary intervention, acute coronary syndrome, peripheral vascular disease, VTE, liver disease, or chronic kidney disease), Charlson comorbidity index, 33 stage of cancer at cancer diagnosis, and other concurrent medications, including angiotensin‐converting enzyme inhibitors or angiotensin II receptor blockers, calcium channel blockers, beta blockers, lipid‐lowering agents, nonsteroidal anti‐inflammatory drugs, proton pump inhibitors, antiplatelet agents, and selective serotonin‐reuptake inhibitors.
Outcomes
The primary efficacy outcome was new arterial or venous thromboses requiring emergency department visit or hospitalization, identified by at least one International Classification of Diseases and Related Health Problems, Tenth Revision (ICD‐10) diagnostic code or billing code in the validated algorithm of the Canadian Institute for Health Information Discharge Abstract Database, National Ambulatory Care Reporting System, Alberta Practitioner Claims, and Ontario Health Insurance Plan databases (see Table S2). Arterial thromboses included the composite events of ischemic stroke, transient ischemic attack, myocardial infarction, percutaneous coronary intervention, or coronary artery bypass graft surgery. VTE outcomes were identified by ICD‐10 diagnostic codes plus at least one procedure code for deep vein thrombosis or pulmonary embolism within 7 days of each other or in the same medical encounter, which had been shown to be a validated approach to identify new VTE cases in Canadian databases. 34 Prior studies have demonstrated high positive predictive values of 97% for ischemic strokes, 35 94% for myocardial infarction, 36 and 82% for pulmonary embolism and a negative predictive value of 99% for pulmonary embolism. 37 , 38 , 39
The primary safety outcome was bleeding event(s) requiring an emergency department visit or hospitalization identified by least one validated ICD‐10 diagnostic code or billing code (see Table S2). Major bleeding events included an upper or lower gastrointestinal hemorrhage or an intracranial hemorrhage. Previous studies demonstrated that these codes have a positive predictive value of 87% and a negative predictive value of 92%. 37 We also evaluated any bleeding events, which included intracranial, gastrointestinal, genitourinary, respiratory tract, joint, ophthalmologic bleeding; or bleeding in unspecified locations; or receipt of a blood transfusion associated with an emergency department visit or hospitalization.
Outcomes were analyzed as‐treated, for which patients were censored when they stopped or switched to another prostate cancer therapy or anticoagulant. Individuals were followed until the primary study outcome, death, loss of provincial health insurance eligibility, or the end of the follow‐up period on December 31, 2023.
Statistical analysis
We calculated the mean ± standard deviation (SD) for continuous variables (or the median and interquartile range [IQR], if skewed) and frequencies and proportions for categorical variables. We compared characteristics between the DOAC and non‐DOAC groups by using standardized differences. These are differences between group means relative to the pooled SD, and >10% was considered a significant difference. 40 We adjusted for differences in all baseline characteristics listed in Table 1 by using an overlap weighting method based on propensity scores. 41 , 42 Missing data were imputed five times and incorporated into the modeling. The first imputation density plot from the inducer cohort is shown in Figure S1 as an example. We used weighted, cause‐specific hazard models to assess the association of anticoagulant groups (DOAC vs. non‐DOAC) with the outcomes of arterial and venous thromboses, or major and all bleeding events. These analyses were conducted using SAS Enterprise Guide, version 7.1 (SAS Institute Inc.). Confidence intervals (CIs) that did not overlap with 1 were considered statistically significant.
TABLE 1.
Baseline characteristics of patients with prostate cancer in the inducer cohort (enzalutamide or apalutamide) comparing direct oral anticoagulants (DOACs) with non‐DOACs.
| Characteristics: Ontario a | No. (%) | Preweighting standardized difference b | Postweighting standardized difference | |
|---|---|---|---|---|
| DOACs, n = 796 | Non‐DOACs, n = 251 | |||
| Age: Mean ± SD, years | 78.0 ± 96.8 | 78.0 ± 7.1 | 0.13 | 0.00 |
| Stage IV at cancer diagnosis, n (%) | 357 (44.8) | 110 (43.8) | 0.02 | 0.00 |
| Nearest census‐based neighborhood income quintile | ||||
| 1: Low | 149 (18.7) | 54 (21.5) | 0.07 | 0.00 |
| 2 | 168 (21.1) | 47 (18.7) | 0.06 | 0.00 |
| 3 | 163 (20.5) | 49 (19.5) | 0.02 | 0.00 |
| 4 | 158 (19.8) | 41 (16.3) | 0.09 | 0.00 |
| 5: High | 158 (19.8) | 60 (23.9) | 0.098 | 0.00 |
| Nonrural residence | 667 (83.8) | 215 (85.7) | 0.05 | 0.00 |
| Comorbidity | ||||
| Major hemorrhage within 1 year prior | 21 (2.6) | 17 (6.8) | 0.20 | 0.00 |
| Hypertension | 591 (74.2) | 192 (76.5) | 0.05 | 0.00 |
| Diabetes | 294 (36.9) | 87 (34.7) | 0.05 | 0.00 |
| Stroke or TIA | 62 (7.8) | 12 (4.8) | 0.12 | 0.00 |
| Atrial fibrillation | 283 (35.6) | 49 (19.5) | 0.37 | 0.00 |
| Myocardial infarction | 50 (6.3) | 10 (4.0) | 0.10 | 0.00 |
| Heart failure | 232 (29.1) | 66 (26.3) | 0.06 | 0.00 |
| CAD | 283 (35.6) | 89 (35.5) | 0.002 | 0.00 |
| Angina | 24 (3.0) | 14 (5.6) | 0.13 | 0.00 |
| CABG | 17 (2.1) | 6 (2.4) | 0.02 | 0.00 |
| PCI | 43 (5.4) | 14 (5.6) | 0.008 | 0.00 |
| ACS | 283 (35.6) | 91 (36.3) | 0.02 | 0.00 |
| PVD | 20 (2.5) | 6 (2.4) | 0.008 | 0.00 |
| History of VTE | 174 (21.9) | 100 (39.8) | 0.40 | 0.00 |
| Liver disease | 47 (5.9) | 25 (10.0) | 0.15 | 0.00 |
| CKD | 240 (30.2) | 80 (31.9) | 0.04 | 0.00 |
| Charlson index: Mean ± SD | 2.82 ± 3.18 | 3.66 ± 3.27 | 0.26 | 0.00 |
| Other concomitant medications | ||||
| ACE or ARB | 301 (37.8) | 108 (43.0) | 0.11 | 0.00 |
| Calcium channel blockers | 253 (31.8) | 76 (30.3) | 0.03 | 0.00 |
| Beta blockers | 264 (33.2) | 91 (36.3) | 0.07 | 0.00 |
| Lipid‐lowering agents | 435 (54.6) | 131 (52.2) | 0.05 | 0.00 |
| NSAIDs | 83 (10.4) | 32 (12.7) | 0.07 | 0.00 |
| Proton pump inhibitors | 230 (28.9) | 106 (42.2) | 0.28 | 0.00 |
| Antiplatelet agents | 33 (4.1) | 11 (4.4) | 0.01 | 0.00 |
| SSRIs | 61 (7.7) | 35 (13.9) | 0.20 | 0.00 |
| DOAC type | ||||
| Dabigatran | 68 (8.5) | NA | NA | |
| Rivaroxaban | 237 (29.8) | |||
| Apixaban | 361 (45.4) | |||
| Edoxaban | 130 (16.3) | |||
| LMWH type | ||||
| Enoxaparin | NA | 48 (19.1) | NA | |
| Dalteparin | 68 (27.1) | |||
| Tinzaparin | 20 (8.0) | |||
| Warfarin | NA | 115 (45.8) | NA | |
| Characteristics: Alberta a | DOAC, n = 200 | Non‐DOAC, n = 183 | Preweighting standardized difference b | Postweighting standardized difference |
|---|---|---|---|---|
| Age: Mean ± SD, years | 75.4 ± 8.5 | 73.9 ± 8.5 | 0.18 | 0.00 |
| Stage IV at cancer diagnosis | 85 (42.5) | 65 (35.5) | 0.26 | 0.00 |
| Nearest census‐based neighborhood income quintile | ||||
| 1: Low | 40 (20.0) | 34 (18.6) | 0.20 | 0.00 |
| 2 | 56 (28.0) | 39 (21.3) | ||
| 3 | 30 (15.0) | 29 (15.8) | ||
| 4 | 39 (19.5) | 36 (19.7) | ||
| 5: High | 33 (16.5) | 42 (23.0) | ||
| Rural residence | 71 (35.5) | 46 (25.1) | 0.23 | 0.00 |
| Comorbidity | ||||
| Major hemorrhage within 1 year prior | 8 (4.0) | 12 (6.6) | 0.12 | 0.00 |
| Hypertension | 144 (72.0) | 118 (64.5) | 0.16 | 0.00 |
| Diabetes | 74 (37.0) | 48 (26.2) | 0.23 | 0.00 |
| Stroke or TIA | 21 (10.5) | 22 (12.0) | 0.05 | 0.00 |
| Atrial fibrillation | 94 (47.0) | 37 (20.2) | 0.59 | 0.00 |
| Myocardial infarction | 11 (5.5) | 7 (3.8) | 0.08 | 0.00 |
| Heart failure | 59 (29.5) | 44 (24.0) | 0.12 | 0.00 |
| CAD | 30 (15.0) | 20 (10.9) | 0.12 | 0.00 |
| Angina | 5 (2.5) | 0 (0) | 0.23 | 0.00 |
| CABG | 6 (3.0) | 5 (2.7) | 0.02 | 0.00 |
| PCI | 11 (5.5) | 3 (1.6) | 0.21 | 0.00 |
| ACS | 32 (16.0) | 20 (10.9) | 0.15 | 0.00 |
| PVD | 6 (3.0) | 6 (3.3) | 0.02 | 0.00 |
| History of VTE | 9 (4.5) | 13 (7.1) | 0.11 | 0.00 |
| Liver disease | 5 (2.5) | 11 (6.0) | 0.14 | 0.00 |
| CKD | 67 (33.5) | 56 (30.6) | 0.18 | 0.00 |
| Charlson index, mean (SD | 2.90 (3.56) | 4.32 (3.90) | 0.38 | 0.00 |
| Other concomitant medications | ||||
| ACE or ARB | 78 (39.0) | 63 (34.4) | 0.02 | 0.00 |
| Calcium channel blockers | 43 (21.5) | 29 (15.8) | 0.11 | 0.00 |
| Beta blockers | 67 (33.5) | 38 (20.8) | 0.26 | 0.00 |
| Lipid‐lowering agents | 86 (43.0) | 67 (36.6) | 0.05 | 0.00 |
| NSAIDs | 18 (9.0) | 20 (10.9) | 0.41 | 0.00 |
| Proton pump inhibitors | 66 (33.0) | 57 (31.1) | 0.04 | 0.00 |
| Antiplatelet agents | 8 (4.0) | 2 (1.1) | 0.19 | 0.00 |
| SSRIs | 19 (9.5) | 23 (12.6) | 0.15 | 0.00 |
| DOAC type | ||||
| Dabigatran | 22 (11.0) | NA | NA | |
| Rivaroxaban | 75 (37.5) | |||
| Apixaban | 96 (48.0) | |||
| Edoxaban | 7 (3.5) | |||
| LMWH type | ||||
| Enoxaparin | NA | 17 (9.3) | NA | |
| Dalteparin | 23 (12.6) | |||
| Tinzaparin | 91 (49.7) | |||
| Warfarin | NA | 52 (28.4) | NA | |
Abbreviations: ACE, angiotensin‐converting enzyme; ACS, acute coronary syndrome; ARB, angiotensin II receptor blockers; CABG, coronary artery bypass graft surgery; CAD, coronary artery disease; CI, confidence interval; CKD, chronic kidney disease; HR, hazard ratio; NA, not applicable; NSAIDs, nonsteroidal anti‐inflammatory drugs; PCI, percutaneous coronary intervention; PVD, peripheral vascular disease; SD, standard deviation; SSRIs, selective serotonin reuptake inhibitors; TIA, transient ischemic attack; VTE, venous thromboembolism.
Variables with significant missing data: Ontario, stage (15.2%) and estimated glomerular filtration rate (22.9%); Alberta, stage (52.5%) and eGFR glomerular filtration rate (27.2%).
A standardized difference >0.1 indicates a significant difference between the two groups.
To evaluate the consistency of our findings, we performed several additional analysis: (1) comparing DOACs versus LMWH, (2) restricted to patients aged 75 years and older, and (3) excluding patients who had a VTE diagnosis within 6 months before the index date in the Ontario cohort because of its larger sample size. In addition, we conducted meta‐analyses to obtain summary estimates of hazard ratios (HRs) and corresponding 95% CIs from both provinces in the main outcomes of interest, as predicted by the potential DDIs, for all thrombosis events (arterial and venous) in the inducer cohort and all bleeding events in the inhibitor cohort. 43 Random‐effects models were performed using the Hartung–Knapp method to account for potential heterogeneity. 44 Heterogeneity was assessed using I2 and τ2 statistics. Meta‐analyses were conducted using the meta package (version 8.1‐0) in R (version 4.5.0; R Foundation for Statistical Computing).
We report our findings according to the recommended RECORD (Reporting of studies Conducted using Observational Routinely collected health Data) and STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) reporting guidelines. 45
RESULTS
Inducer cohort: Concurrent use with enzalutamide or apalutamide
Cohort characteristics
In Ontario, 1047 male patients aged 66 years and older who were receiving concurrent enzalutamide or apalutamide and an anticoagulant were included, with 796 receiving a DOAC and 251 receiving a non‐DOAC (136 LMWH and 115 warfarin). The mean ± SD age at the index date was 78.7 ± 6.9 years. The majority of men (N = 803; 76.7%) received enzalutamide, and 244 men (23.3%) received apalutamide. DOAC prescriptions included: apixaban (45.4%), rivaroxaban (29.8%), edoxaban (16.3%), and dabigatran (8.5%). The most commonly used LMWH was dalteparin (50% of LMWH use). Table 1 summarizes the baseline characteristics of the Ontario cohort stratified by the DOAC versus non‐DOAC groups. Patients receiving DOAC were slightly older and included more with a history of AF, whereas those in the non‐DOAC group included more with a history of major bleeding within 1 year, a history of VTE, liver disease, and a higher mean Charlson comorbidity index. The median follow‐up durations in the DOAC and non‐DOAC groups were 135 days (IQR, 44–349 days) and 127 days (IQR, 49–244 days), respectively.
In Alberta, 383 adult males were included: with 200 receiving a DOAC and 183 receiving a non‐DOAC (131 LMWH and 52 warfarin; Table 1B). The mean ± SD age of the cohort at the index date was 74.7 ± 8.5 years. In contrast to the Ontario cohort, the majority of men in the Alberta cohort received apalutamide (N = 283; 73.9%), and 100 (26.1%) received enzalutamide. The DOACs prescribed included: apixaban (48.0%), rivaroxaban (37.5%), dabigatran (11.0%), and edoxaban (3.5%). The most commonly used LMWH was tinzaparin (69.5% of LMWH use). Similar to Ontario, patients in Alberta who were prescribed DOAC were older and more had AF and diabetes, whereas men in the non‐DOAC group had a higher percentage of liver disease and a higher Charlson comorbidity index. The median follow‐up durations in the DOAC and non‐DOAC groups were 182 days (IQR, 84–390 days) and 171 days (IQR, 61–383 days), respectively.
All baseline characteristic differences were successfully adjusted after overlap weighting because all standard differences reached 0 (Table 1; see Table S3).
Thrombosis and bleeding outcomes
In the inducer cohort, the expected DDI, if clinically relevant, would lead to an increased risk of thrombosis in patients who were receiving concurrent DOACs (compared with non‐DOACs); therefore, thrombosis outcomes were of the most interest. In Ontario, patients receiving DOACs were not associated with a higher risk of arterial thrombosis (DOAC, 30 of 796 patients [3.8%] vs. non‐DOAC, 11 of 251 patients [4.4%]; weighted HR [after overlap weighting], 0.95; 95% CI, 0.42–2.13; Table 2). Similarly, there were comparable rates in VTE, major or any bleeding events (Table 2). In Alberta, the rates of outcomes appeared largely similar to those in Ontario, except for a numerically higher rate in arterial thrombosis in the DOAC cohort. Similarly, after overlap weighting, there were no significant differences in DOAC vs. non‐DOAC groups in the rates of arterial thrombosis, VTE, and bleeding outcomes (Table 2).
TABLE 2.
Summary of outcomes in the inducer cohort (direct oral anticoagulants [DOACs] vs. non‐DOACs).
| Outcomes: Ontario | No. (%) | Weighted HR [95% CI] | |
|---|---|---|---|
| DOACs, n = 796 | Non‐DOACs, n = 251 | ||
| Arterial thrombosis | 30 (3.8) | 11 (4.4) | 0.95 [0.42–2.13] |
| VTE | 17 (2.1) | 13 (5.2) | 0.49 [0.23–1.04] |
| Major bleeding | 25 (3.1) | 11 (4.4) | 0.67 [0.29–1.55] |
| Any bleeding | 64 (8.0) | 28 (11.2) | 0.86 [0.52–1.24] |
| Outcomes: Alberta | DOACs, n = 200 | Non‐DOACs, n = 183 | Weighted HR [95% CI] |
|---|---|---|---|
| Arterial thrombosis | 18 (9.0) | 7 (3.8) | 1.66 [0.61–4.54] |
| VTE | 6 (3.0) | 4 (2.2) | 0.69 [0.17–2.81] |
| Major bleeding | 5 (2.5) | 7 (3.8) | 0.50 [0.11–2.19] |
| Any bleeding | 17 (8.5) | 16 (8.7) | 0.64 [0.28–1.48] |
| Additional analysis in Ontario | Weighted HR [95% CI] | ||
|---|---|---|---|
| DOACs vs. LMWH: Reference, N = 932 | |||
| Arterial thrombosis | 1.62 [0.41–6.36] | ||
| VTE | 0.37 (0.14–0.98) | ||
| Major bleeding | 1.05 (0.28–3.98) | ||
| Any bleeding | 0.98 (0.46–2.09) | ||
| Age ≥75 years: DOACs vs. non‐DOACs, N = 723 | |||
| Arterial thrombosis | 0.95 [0.37–2.42] | ||
| VTE | 0.39 [0.15–1.01] | ||
| Major bleeding | 1.02 [0.34–3.12] | ||
| Any bleeding | 1.07 [0.56–‐2.06] | ||
| Excluding patients with VTE within 6 months before index: DOACs vs. non‐DOACs, N = 934 | |||
| Arterial thrombosis | 0.78 (0.34‐1.80) | ||
| VTE | 0.44 (0.17‐1.13) | ||
| Major bleeding | 0.80 (0.32‐2.03) | ||
| Any bleeding | 1.01 (0.58‐1.77) | ||
Abbreviations: CI, confidence interval; HR, hazard ratio; LMWH, low‐molecular‐weight heparin; VTE, venous thromboembolism.
Additional analyses
Additional analyses in the Ontario cohort comparing DOAC with LMWH, in patients aged 75 years and older, and after excluding patients who had VTE within 6 months before the index date, demonstrated consistent results (Table 2). The meta‐analysis summarizing results from both provinces indicated no significant differences in the rate of all thrombosis (arterial and venous; [HR, 0.83; 95% CI, 0.36–1.93]), with low heterogeneity (I2 = 22.4%; τ2 = 0.09; p = .28; Figure 1A).
FIGURE 1.
(A) Meta‐analysis of all thrombosis outcomes among patients with prostate cancer in the inducer cohort (enzalutamide or apalutamide) comparing DOACs with non‐DOACs in Ontario and Alberta. (B) Meta‐analysis of all bleeding outcomes among patients with prostate cancer in the inhibitor cohort (abiraterone) comparing DOACs with non‐DOACs in Ontario and Alberta. CI indicates confidence interval; DOACs, direct oral anticoagulants; HR, hazard ratio; SE, standard error.
Inhibitor cohort: Concurrent use with abiraterone
Cohort characteristics
In Ontario, 1060 men who were receiving concurrent abiraterone and an anticoagulant were included: with 735 receiving a DOAC and 325 receiving a non‐DOAC (168 LMWH and 157 warfarin; Table 3). The mean ± SD age of the cohort was 79.8 ± 7.1 years. The patterns of DOAC and LMWH use were similar to those in the inducer cohort; the majority of DOACs were apixaban (50.2%) and rivaroxaban (36.2%), and the most commonly used LMWH was dalteparin (54.8% of LMWH). Patients in the DOAC group were also older and had more cardiac‐related comorbidities (history of AF, coronary artery bypass graft surgery), whereas those in the non‐DOAC group included more with a history of VTE and a higher mean Charlson comorbidity index. The median follow‐up durations in the DOAC and non‐DOAC groups were 186 days (IQR, 99–370 days) and 189 days (IQR, 47–258 days), respectively.
TABLE 3.
Baseline characteristics of patients with prostate cancer in the inhibitor cohort (abiraterone) comparing direct oral anticoagulants (DOACs) versus non‐DOACs.
| Characteristic: Ontario a | No. (%) | |||
|---|---|---|---|---|
| DOACs, n = 735 | Non‐DOACs, n = 325 | Preweighting standardized difference b | Postweighting standardized difference | |
| Age: Mean ± SD, years | 80.6 ± 7.0 | 77.9 ± 7.1 | 0.38 | 0.00 |
| Stage IV at cancer diagnosis | 313 (42.6) | 157 (48.3) | 0.12 | 0.00 |
| Nearest census‐based neighborhood income quintile | ||||
| 1: Low | 145 (19.7) | 69 (21.2) | 0.04 | 0.00 |
| 2 | 134 (18.2) | 23 (7.1) | 0.11 | 0.00 |
| 3 | 155 (21.1) | 57 (17.5) | 0.09 | 0.00 |
| 4 | 140 (19.0) | 70 (21.5) | 0.06 | 0.00 |
| 5: High | 161 (21.9) | 54 (16.6) | 0.13 | 0.00 |
| Nonrural residence | 619 (84.2) | 273 (84.0) | 0.006 | 0.00 |
| Comorbidity | ||||
| Major hemorrhage within 1 year prior | 24 (3.3) | 12 (3.7) | 0.02 | 0.00 |
| Hypertension | 555 (75.5) | 237 (72.9) | 0.06 | 0.00 |
| Diabetes | 248 (33.7) | 117 (36.0) | 0.05 | 0.00 |
| Stroke or TIA | 56 (7.6) | 22 (6.8) | 0.03 | 0.00 |
| Atrial fibrillation | 302 (41.1) | 69 (21.2) | 0.44 | 0.00 |
| Myocardial infarction | 54 (7.3) | 21 (6.5) | 0.04 | 0.00 |
| Heart failure | 229 (31.2) | 84 (25.8) | 0.12 | 0.00 |
| CAD | 265 (36.1) | 100 (30.8) | 0.11 | 0.00 |
| Angina | 27 (3.7) | 19 (5.8) | 0.10 | 0.00 |
| CABG | 9 (1.2) | 15 (4.6) | 0.20 | 0.00 |
| PCI | 31 (4.2) | 11 (3.4) | 0.04 | 0.00 |
| ACS | 273 (37.1) | 103 (31.7) | 0.12 | 0.00 |
| PVD | 23 (3.1) | 8 (2.5) | 0.04 | 0.00 |
| History of VTE | 172 (23.4) | 127 (39.1) | 0.34 | 0.00 |
| Liver disease | 41 (5.6) | 23 (7.1) | 0.06 | 0.00 |
| CKD | 251 (34.1) | 96 (29.5) | 0.10 | 0.00 |
| Charlson index: Mean ± SD | 3.12 ± 3.16 | 3.58 ± 3.23 | 0.15 | 0.00 |
| Other concomitant medications | ||||
| ACE or ARB | 311 (42.3) | 142 (43.7) | 0.03 | 0.00 |
| Calcium channel blockers | 237 (32.2) | 88 (27.1) | 0.11 | 0.00 |
| Beta blockers | 310 (42.2) | 116 (35.7) | 0.13 | 0.00 |
| Lipid‐lowering agents | 422 (57.4) | 167 (51.4) | 0.12 | 0.00 |
| NSAIDs | 100 (13.6) | 50 (15.4) | 0.05 | 0.00 |
| Proton pump inhibitors | 271 (36.9) | 136 (41.8) | 0.10 | 0.00 |
| Antiplatelet agents | 36 (4.9) | 8 (2.5) | 0.13 | 0.00 |
| SSRIs | 62 (8.4) | 27 (8.3) | 0.005 | 0.00 |
| DOAC type | ||||
| Dabigatran | 50 (6.8) | NA | NA | |
| Rivaroxaban | 266 (36.2) | |||
| Apixaban | 369 (50.2) | |||
| Edoxaban | 50 (6.8) | |||
| LMWH type | ||||
| Enoxaparin | NA | 52 (16.0) | NA | |
| Dalteparin | 92 (28.3) | |||
| Tinzaparin | 24 (7.4) | |||
| Warfarin | NA | 157 (48.3) | NA | |
| Characteristic: Alberta a | DOACs, n = 236 | Non‐DOACs, n = 271 | Preweighting standardized difference b | Postweighting standardized difference |
|---|---|---|---|---|
| Age: Mean ± SD | 77.6 ± 8.6 | 75.6 ± 9.0 | 0.22 | 0.00 |
| Stage IV at cancer diagnosis | 100 (42.4) | 104 (38.4) | 0.02 | 0.00 |
| Nearest census‐based neighborhood income quintile | ||||
| 1: Low | 43 (18.2) | 56 (20.7) | 0.19 | 0.00 |
| 2 | 64 (27.1) | 62 (22.9) | ||
| 3 | 40 (16.9) | 55 (20.3) | ||
| 4 | 33 (14.0) | 46 (17.0) | ||
| 5: High | 52 (22.0) | 46 (17.0) | ||
| Rural residence | 75 (31.8) | 62 (22.9) | 0.20 | 0.00 |
| Comorbidity | ||||
| Major hemorrhage within 1 year before | 10 (4.2) | 8 (3.0) | 0.07 | 0.00 |
| Hypertension | 167 (70.8) | 172 (63.5) | 0.16 | 0.00 |
| Diabetes | 62 (26.3) | 68 (25.1) | 0.03 | 0.00 |
| Stroke or TIA | 45 (19.1) | 33 (12.2) | 0.19 | 0.00 |
| Atrial fibrillation | 102 (43.2) | 77 (28.4) | 0.31 | 0.00 |
| Myocardial infarction | 7 (3.0) | 10 (3.7) | 0.04 | 0.00 |
| Heart failure | 46 (19.5) | 59 (21.8) | 0.06 | 0.00 |
| CAD | 25 (10.6) | 24 (8.9) | 0.06 | 0.00 |
| Angina | 5 (2.1) | 4 (1.5) | 0.05 | 0.00 |
| CABG | 2 (0.8) | 3 (1.1) | 0.03 | 0.00 |
| PCI | 5 (2.1) | 5 (1.8) | 0.02 | 0.0.0 |
| ACS | 26 (11.0) | 24 (8.9) | 0.07 | 0.00 |
| PVD | 9 (3.8) | 8 (3.0) | 0.05 | 0.00 |
| History of VTE | 20 (8.5) | 17 (6.3) | 0.08 | 0.00 |
| Liver disease | 10 (4.2) | 10 (3.7) | 0.03 | 0.00 |
| CKD | 76 (32.2) | 78 (28.8) | 0.07 | 0.00 |
| Charlson index: Mean ± SD | 3.35 ± 4.0 | 4.09 ± 4.0 | 0.19 | 0.00 |
| Other concomitant medications | ||||
| ACE or ARB | 88 (37.3) | 100 (36.9) | 0.006 | 0.00 |
| Calcium channel blockers | 54 (22.9) | 55 (20.3) | 0.07 | 0.00 |
| Beta blockers | 80 (33.9) | 71 (26.2) | 0.20 | 0.00 |
| Lipid lowering agents | 102 (43.2) | 96 (35.4) | 0.19 | 0.00 |
| NSAIDs | 27 (11.4) | 37 (13.7) | 0.08 | 0.00 |
| Proton pump inhibitors | 71 (30.1) | 95 (35.1) | 0.13 | 0.00 |
| Antiplatelet agents | 13 (5.5) | 7 (2.6) | 0.17 | 0.00 |
| SSRI | 21 (8.9) | 40 (14.8) | 0.21 | 0.00 |
| DOAC type | ||||
| Dabigatran | 25 (10.6) | NA | NA | |
| Rivaroxaban | 87 (36.9) | |||
| Apixaban | 120 (50.8) | |||
| Edoxaban | 4 (1.7) | |||
| LMWH type | ||||
| Enoxaparin | NA | 14 (5.2) | NA | |
| Dalteparin | 53 (19.6) | |||
| Tinzaparin | 96 (35.4) | |||
| Warfarin | NA | 108 (39.9) | NA | |
Abbreviations: ACE, angiotensin‐converting enzyme; ACS, acute coronary syndrome; ARB, angiotensin II receptor blockers; CABG, coronary artery bypass graft surgery; CAD, coronary artery disease; CI, confidence interval; CKD, chronic kidney disease; HR, hazard ratio; LMWH, low‐molecular‐weight heparin; NA, not applicable; NSAIDs, nonsteroidal anti‐inflammatory drugs; PCI, percutaneous coronary intervention; PVD, peripheral vascular disease; SD, standard deviation; SSRIs, selective serotonin reuptake inhibitors; TIA, transient ischemic attacks; VTE, venous thromboembolism.
Variables with significant missing data: Ontario, stage (9.8%) and estimated glomerular filtration rate (21.7%); Alberta, stage (46.4%) and estimated glomerular filtration rate (26.0%).
A standardized difference >0.1 indicates a significant difference between the two groups.
In Alberta, a total of 507 adult males were included: 236 DOAC and 271 non‐DOAC (163 LMWH and 108 warfarin; Table 3). The mean ± SD age was 76.5 ± 8.8 years. Similar to the inducer cohort, the most commonly used DOACs were apixaban (50.8%) and rivaroxaban (36.9%), and the most commonly used LMWH was tinzaparin (58.9% of LMWH). The comorbidity distribution was also similar to that of the inducer cohort. The median follow‐up durations in the DOAC and non‐DOAC groups were 188 days (IQR, 76–396 days) and 182 days (IQR, 60–387 days), respectively. Again, baseline characteristic differences were adjusted after overlap weighting (Table 3; see Table S4).
Bleeding and thrombosis outcomes
In the inhibitor cohort, bleeding outcomes were of the most interest because the expected DDI (if clinically important) would lead to an increased risk of bleeding in patients receiving concurrent DOACs. In Ontario, patients receiving DOACs were not associated with an increased risk of major or any bleeding episodes (major bleeding: DOAC, 40 of 735 patients [5.4%] vs. non‐DOAC, 13 of 325 patients [4.0%]; weighted HR, 1.00; 95% CI, 0.47–2.15; any bleeding: DOAC vs. non‐DOAC, 14.1% vs. 11.1%; weighted HR, 0.99; 95% CI, 0.64–1.55; Table 4). Similarly, there were comparable rates in arterial thrombosis or VTE events (Table 4). The rates of bleeding outcomes were also comparable between the two provinces. No significant differences were observed between the DOAC groups versus the non‐DOAC groups in all outcomes after overlap weighting in the Alberta cohort (Table 4).
TABLE 4.
Summary of outcomes in the inhibitor cohort (direct oral anticoagulants [DOACs] vs. non‐DOACs).
| Outcomes: Ontario | No. (%) | Weighted HR [95% CI] | |
|---|---|---|---|
| DOAC, n = 735 | Non‐DOAC, n = 325 | ||
| Major bleeding | 40 (5.4) | 13 (4.0%) | 1.00 [0.47–2.15] |
| Any bleeding | 104 (14.1) | 36 (11.1%) | 0.99 [0.64–1.55] |
| Arterial thrombosis | 43 (5.9) | 10 (3.1%) | 1.24 [0.57–2.70] |
| VTE | 27 (3.7) | 20 (6.2%) | 0.54 [0.28–1.07] |
| Outcomes: Alberta | DOAC, n = 236 | Non‐DOAC, n = 271 | Weighted HR [95% CI] |
|---|---|---|---|
| Major bleeding | 9 (3.8) | 13 (4.8) | 0.96 [0.34–2.72] |
| Any bleeding | 33 (14.0) | 32 (11.8) | 1.48 [0.83–2.63] |
| Arterial thrombosis | 20 (8.5) | 21 (7.7) | 0.94 [0.46–1.92] |
| VTE | 4 (1.7) | 12 (4.4) | 0.56 [0.17–1.82] |
| Additional analysis in Ontario | Weighted HR [95% CI] | ||
|---|---|---|---|
| DOAC vs. LMWH: Reference, N = 903 | |||
| Major bleeding | 1.77 [0.40–7.76] | ||
| Any bleeding | 0.89 [0.45–1.78] | ||
| Arterial thrombosis | 0.72 [0.18–2.93] | ||
| VTE | 0.74 [0.31–1.79] | ||
| Age ≥75 years: DOAC vs. non‐DOAC, N = 782 | |||
| Major bleeding | 0.96 [0.39–2.39] | ||
| Any bleeding | 1.29 [0.74–2.25] | ||
| Arterial thrombosis | 1.63 [0.66–4.01] | ||
| VTE | 0.65 [0.26–1.65] | ||
| Excluding patients with VTE within 6 months before to index: DOAC vs. non‐DOAC, N = 940 | |||
| Major bleeding | 0.76 [0.35–1.64] | ||
| Any bleeding | 0.90 [0.56–1.45] | ||
| Arterial thrombosis | 1.13 [0.51–2.51] | ||
| VTE | 0.53 [0.23–1.20] | ||
Abbreviations: CI, confidence interval; HR, hazard ratio; LMWH, low‐molecular‐weight heparin; VTE, venous thromboembolism.
Additional analyses
Similar additional analyses in the Ontario inhibitor cohort (comparing DOAC vs. LMWH, in patients aged 75 years and older, and excluding patients with VTE within 6 months before the index date) were conducted, and the results were consistent with the main analysis (Table 4). The meta‐analysis indicated no significant differences in all bleeding events (HR, 1.16; 95% CI, 0.10–13.99) with low heterogeneity (I2 = 15.0%; τ2 = 0.01; p = .28; Figure 1B).
DISCUSSION
To our knowledge, this is the largest population‐based cohort study to date evaluating clinically relevant outcomes in patients with prostate cancer who received concurrent anticoagulants and potentially interacting prostate cancer therapies. In contrast to the DDI concerns based on pharmacokinetic data, we observed no significant differences in arterial or venous thrombosis and bleeding events in either cohort receiving concurrent DOACs and prostate cancer drugs with CYP3A4 and/or P‐gp inducer or inhibitor activities.
With a median age of diagnosis of 67, patients with advanced prostate cancer are generally older, and they have multiple comorbidities and polypharmacy, which can lead to an even higher risk of thrombosis and bleeding. 46 In patients who require anticoagulation and an ARPI for prostate cancer at the same time, clinicians often face a dilemma when choosing the best combination therapy given the concern of potential DDIs. Because of the excellent efficacy of ARPIs in these patients, the survival and duration of treatment continue to be prolonged in recent years (5‐year survival, >90%), 1 making this issue even more relevant in terms of years and lives at risk. In‐vitro studies demonstrate that both enzalutamide and apalutamide are strong inducers of the CYP3A4 pathway, and the co‐administration of enzalutamide or apalutamide reduces the area under the curve of midazolam (a CYP3A4 substrate) by 86% and 92%, respectively. 28 , 29 In addition, apalutamide is a P‐gp inducer. 29 Therefore, the concomitant use of enzalutamide or apalutamide could potentially lower DOAC drug levels and lead to an increased risk of thrombosis. In contrast, abiraterone demonstrates moderate inhibition of the CYP3A4 and P‐gp pathways in vitro 30 , 47 and, in theory, can lead to an increase in DOAC drug levels and an elevated risk of bleeding. Given the concern of DDIs, combinations of enzalutamide, apalutamide, or abiraterone with DOACs are cautioned against by guidance consensus. 48 , 49 , 50 However, whether the in‐vitro pharmacokinetic concerns translate into clinically significant outcomes is unclear, and our current study aims to address this knowledge gap. Understanding the outcomes of concurrent anticoagulants and prostate cancer therapies will have high effects and provide immediate clinical benefit for many patients.
Prior studies demonstrated comparable effectiveness between DOACs and warfarin in patients who had cancer with AF 11 or cancer‐associated VTE, 51 with a higher bleeding risk from warfarin. 52 However, there is little clinical information on the effect of potential drug interactions. In a recent post‐hoc analysis of randomized controlled trials investigating apalutamide in prostate cancer, the rates of thrombosis were comparable in those receiving concurrent oral anticoagulants plus apalutamide compared with anticoagulants plus a placebo, but the numbers of concurrent DOAC users were far too small (N = 15–25). 53 Determining the accurate risks of thrombosis and bleeding is essential to facilitate appropriate monitoring, patient consultation, and potential therapy adjustments. Administrative data sets are useful to study drug interactions that pose serious clinical outcomes (such as thrombosis or hemorrhage) yet are rare to be frequently detected in prospective clinical trials. Our results provide reassuring data indicating that the concurrent use of DOACs with enzalutamide, apalutamide, or abiraterone could be safe and anticoagulant switches, discontinuation, or dose adjustment could be avoided if patients were to start one of these ARPIs.
The strength of our study includes high external validity because all eligible patients in Ontario and Alberta were included. Our findings remained consistent across different provinces and additional analyses. The provincial health care databases in Canada have several strengths, including extremely well recorded prescription drug use 31 and little emigration from the province (<1% per year 54 ).
Limitations
Our study has limitations. Given the retrospective design and nonrandomized attribution of anticoagulants, there were differences in baseline characteristics between the two groups, as expected. We successfully adjusted for the differences in all baseline characteristics by using an overlap weighting method based on propensity scores, but residual confounding remained possible. Over‐the‐counter medications, such as aspirin or nonsteroidal anti‐inflammatory drugs, might not be fully captured in the database, but we expected nondifferential use between the DOAC and non‐DOAC groups. Although health administrative data are not collected for research purposes and are subject to misclassification bias, we used validated and reliable methods to define our outcome variables, with a high positive predictive value. 35 , 36 , 37 , 38 , 39 , 55 We used multiple sources of data for covariates (inpatient, outpatient, emergency department) and acknowledge that, because of the long follow‐up of the study, we may have misclassified some covariates as not being present, leading to an underestimate of comorbid indications of anticoagulation. This may bias our estimates toward the null. It is also reassuring that the outcome rates are consistent between the two provinces and mostly in line with randomized controlled trials among patients with cancer‐associated thrombosis, 56 although direct comparisons of rates are not possible given different study designs and outcome definitions. The rates of arterial thrombosis in Alberta may seem numerically higher, which could be multifactorial, including different anticoagulant prescription patterns, the inclusion of younger patients (an age cutoff of 18 years in Alberta), and the addition of claim codes to identify percutaneous coronary intervention and coronary artery bypass graft surgery in Alberta because they were not captured as well in the National Ambulatory Care Reporting System/Discharge Abstract Database. Despite these variables, results comparing the DOAC and non‐DOAC groups were consistent. Warfarin could have DDIs with prostate cancer agents, and the time‐in‐therapeutic range for patients receiving warfarin was not available in the database, so the quality of warfarin management is unknown. However, this reflected real‐life data, and we conducted a sensitivity analysis in the Ontario cohort in which we removed the patients who were receiving warfarin, and the results were consistent. The sample size remained modest to allow meaningful subgroup analyses by individual anticoagulant that might have different DDI potential and outcomes. However, to our knowledge, our study is the largest of its kind. The use of newer cancer agents like darolutamide were rare during the study period; therefore, such data were not available to be included in the analysis. The causes of death or medication compliance rates were not available in the databases
CONCLUSIONS
In this largest cohort to date evaluating DDIs between DOACs and enzalutamide, apalutamide, or abiraterone, their concurrent use was not associated with an increased risk of arterial or venous thrombosis or bleeding events. Our results can have implications in the prescription of anticoagulants for this population.
AUTHOR CONTRIBUTIONS
Tzu‐Fei Wang: Conceptualization, methodology, investigation, writing–original draft, writing–review and editing, supervision, and funding acquisition. Anna Clarke: Methodology, formal analysis, and data curation. Mitchell Rath: Methodology, formal analysis, and data curation. Samantha Yoo: Methodology, formal analysis, and data curation. Deena Fremont: Data curation, writing–review and editing, and project administration. Cynthia Wu: Resources and writing–review and editing. Pietro Ravani: Methodology, writing–review and editing, and supervision. Dominick Bossé: Investigation and writing–review and editing. Robert Talarico: Methodology, investigation, writing–review and editing, and supervision. Marc Carrier: Conceptualization, methodology, and writing–review and editing. Manish M. Sood: Conceptualization, methodology, investigation, writing–review and editing, funding acquisition, and supervision.
CONFLICT OF INTEREST STATEMENT
Dominick Bossie reports grants/contracts from Ipsen; and personal/consulting fees from Astellas Pharma Canada, AstraZeneca Canada, Bayer, Bristol Myers Squibb Canada, EISAI INC., EMD Serono, Ipsen, Merck & Company Inc., and Pfizer Canada Inc. outside the submitted work. Marc Carrier reports grants/contracts from Bristol Myers Squibb Canada, LEO Pharma Inc. and Pfizer; and personal/consulting fees from Anthos Pharmaceuticals, Bayer, Bristol Myers Squibb Canada, Leo Pharma, Pfizer, Regeneron, Sanofi US Services Inc., and Servier outside the submitted work. Manish M. Sood reports personal/consulting fees from Bayer and Otsuka Pharmaceutical and personal and speaking fees from AstraZeneca Canada outside the submitted work. The remaining authors disclosed no conflicts of interest.
Supporting information
Supplementary Material
ACKNOWLEDGMENTS
The authors sincerely thank Nickolas Beauregard from ICES for his assistance in completing additional sensitivity analyses. We thank IQVIA Solutions Canada Inc. for use of their Drug Information File.
This study received funding from the Canadian Institutes of Health Research (CIHR) and was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long‐Term Care. Core funding for ICES Ottawa is provided by the University of Ottawa and the Ottawa Hospital Research Institute. Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information, the Ontario Ministry of Health, Ontario Health, and Cancer Care Ontario. Tzu‐Fei Wang is the recipient of Tier 2 Research Chair from the University of Ottawa in Cancer and Thromboembolism. Marc Carrier is supported by the Clinical Research Chair from the University of Ottawa in Cancer and Thrombosis. Manish M. Sood is supported by the Department of Medicine Chair for Physician Health and Wellness.
This document used data adapted from the Statistics Canada Postal Code^OM Conversion File, which is based on data licensed from Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from ©Canada Post Corporation and Statistics Canada. Data for Alberta were extracted from the Alberta Health Services Enterprise Data Warehouse with support provided by AbSPORU (the Alberta Strategy for Patient‐Oriented Research SUPPORT Uni), which is funded by the CIHR, Alberta Innovates, the University Hospital Foundation, the University of Alberta, the University of Calgary, and Alberta Health Services.
The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred. The funding sources had no role in the design and conduct of the study, data collection, management, analysis, interpretation, or publication of the article. All authors had full access to all data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from ICES. Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the author(s) with the permission of ICES.
REFERENCES
- 1. Canadian Cancer Society . Prostate cancer statistics. Canadian Cancer Society; 2023. Accessed February 14, 2023. https://cancer.ca/en/cancer‐information/cancer‐types/prostate/statistics [Google Scholar]
- 2. National Comprehensive Cancer Network (NCCN) . NCCN Clinical Practice Guidelines in Oncology, Prostate Cancer, version 2.2020. NCCN Guideline Panel; 2020. [Google Scholar]
- 3. Leigh J, Qureshi D, Sucha E, et al. A population‐based study of factors associated with systemic treatment in advanced prostate cancer decedents. Cancer Med. 2022;12(5):5569‐5579. doi: 10.1002/cam4.5401 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Khorana AA, Francis CW, Culakova E, Kuderer NM, Lyman GH. Thromboembolism is a leading cause of death in cancer patients receiving outpatient chemotherapy. J Thromb Haemost. 2007;5(3):632‐634. doi: 10.1111/j.1538-7836.2007.02374.x [DOI] [PubMed] [Google Scholar]
- 5. Balabanova Y, Farahmand B, Garmo H, Stattin P, Brobert G. Risk of venous thromboembolism in men with prostate cancer compared with men in the general population: a nationwide population‐based cohort study in Sweden. BMJ Open. 2022;12(5):e055485. doi: 10.1136/bmjopen-2021-055485 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Van Hemelrijck M, Adolfsson J, Garmo H, et al. Risk of thromboembolic diseases in men with prostate cancer: results from the population‐based PCBaSe Sweden. Lancet Oncol. 2010;11(5):450‐458. doi: 10.1016/s1470-2045(10)70038-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Canadian Institute for Health Information (CIHI) . Drug Use Among Seniors on Public Drug Programs in Canada, 2012. CIHI; 2014. [Google Scholar]
- 8. Chen ST, Hellkamp AS, Becker RC, et al. Efficacy and safety of rivaroxaban vs. warfarin in patients with non‐valvular atrial fibrillation and a history of cancer: observations from ROCKET AF. Eur Heart J Qual Care Clin Outcomes. 2019;5(2):145‐152. doi: 10.1093/ehjqcco/qcy040 [DOI] [PubMed] [Google Scholar]
- 9. Melloni C, Dunning A, Granger CB, et al. Efficacy and safety of apixaban versus warfarin in patients with atrial fibrillation and a history of cancer: insights from the ARISTOTLE trial. Am J Med. 2017;130(12):1440‐1448.e1. doi: 10.1016/j.amjmed.2017.06.026 [DOI] [PubMed] [Google Scholar]
- 10. Fanola CL, Ruff CT, Murphy SA, et al. Efficacy and safety of edoxaban in patients with active malignancy and atrial fibrillation: analysis of the ENGAGE AF‐TIMI 48 trial. J Am Heart Assoc. 2018;7(16):e008987. doi: 10.1161/jaha.118.008987 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Shah S, Norby FL, Datta YH, et al. Comparative effectiveness of direct oral anticoagulants and warfarin in patients with cancer and atrial fibrillation. Blood Adv. 2018;2(3):200‐209. doi: 10.1182/bloodadvances.2017010694 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Raskob GE, van Es N, Verhamme P, et al. Edoxaban for the treatment of cancer‐associated venous thromboembolism. N Engl J Med. 2018;378(7):615‐624. doi: 10.1056/nejmoa1711948 [DOI] [PubMed] [Google Scholar]
- 13. Young AM, Marshall A, Thirlwall J, et al. Comparison of an oral factor Xa inhibitor with low molecular weight heparin in patients with cancer with venous thromboembolism: results of a randomized trial (SELECT‐D). J Clin Oncol. 2018;36(20):2017‐2023. doi: 10.1200/jco.2018.78.8034 [DOI] [PubMed] [Google Scholar]
- 14. McBane RD 2nd, Wysokinski WE, Le‐Rademacher JG, et al. Apixaban and dalteparin in active malignancy‐associated venous thromboembolism: the ADAM VTE trial. J Thromb Haemost. 2020;18(2):411‐421. doi: 10.1111/jth.14662 [DOI] [PubMed] [Google Scholar]
- 15. Agnelli G, Becattini C, Meyer G, et al. Apixaban for the treatment of venous thromboembolism associated with cancer. N Engl J Med. 2020;382(17):1599‐1607. doi: 10.1056/nejmoa1915103 [DOI] [PubMed] [Google Scholar]
- 16. Planquette B, Bertoletti L, Charles‐Nelson A, et al. Rivaroxaban vs dalteparin in cancer‐associated thromboembolism: a randomized trial. Chest. 2021;161(3):781‐790. doi: 10.1016/j.chest.2021.09.037 [DOI] [PubMed] [Google Scholar]
- 17. Carrier M, Abou‐Nassar K, Mallick R, et al. Apixaban to prevent venous thromboembolism in patients with cancer. N Engl J Med. 2019;380(8):711‐719. doi: 10.1056/nejmoa1814468 [DOI] [PubMed] [Google Scholar]
- 18. Khorana AA, Soff GA, Kakkar AK, et al. Rivaroxaban for thromboprophylaxis in high‐risk ambulatory patients with cancer. N Engl J Med. 2019;380(8):720‐728. doi: 10.1056/nejmoa1814630 [DOI] [PubMed] [Google Scholar]
- 19. Key NS, Khorana AA, Kuderer NM, et al. Venous Thromboembolism prophylaxis and treatment in patients with cancer: ASCO clinical practice guideline update. J Clin Oncol. 2020;38(5):496‐520. doi: 10.1200/jco.19.01461 [DOI] [PubMed] [Google Scholar]
- 20. National Comprehensive Cancer Network (NCCN) . NCCN Guidelines in Cancer‐Associated Venous Thromboembolic Disease. Version 2.2025. NCCN Guideline Panel; 2025. Accessed March 1, 2025. https://www.nccn.org/professionals/physician_gls/pdf/vte.pdf [Google Scholar]
- 21. Lyman GH, Carrier M, Ay C, et al. American Society of Hematology 2021 guidelines for management of venous thromboembolism: prevention and treatment in patients with cancer. Blood Adv. 2021;5(4):927‐974. doi: 10.1182/bloodadvances.2020003442 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Delluc A, Wang TF, Yap ES, et al. Anticoagulation of cancer patients with non‐valvular atrial fibrillation receiving chemotherapy: guidance from the SSC of the ISTH. J Thromb Haemost. 2019;17(8):1247‐1252. doi: 10.1111/jth.14478 [DOI] [PubMed] [Google Scholar]
- 23. Voukalis C, Lip GYH, Shantsila E. Drug‐drug interactions of non‐vitamin K oral anticoagulants. Expert Opin Drug Metab Toxicol. 2016;12:1445‐1461. doi: 10.1080/17425255.2016.1225037 [DOI] [PubMed] [Google Scholar]
- 24. Lin JH, Yamazaki M. Role of P‐glycoprotein in pharmacokinetics. Clin Pharmacokinet. 2003;42(1):59‐98. doi: 10.2165/00003088-200342010-00003 [DOI] [PubMed] [Google Scholar]
- 25. Lin JH, Lu AYH. Inhibition and induction of cytochrome P450 and the clinical implications. Clin Pharmacokinet. 1998;35(5):361‐390. doi: 10.2165/00003088-199835050-00003 [DOI] [PubMed] [Google Scholar]
- 26. Fradley MG, Ellenberg K, Alomar M, et al. Patterns of anticoagulation use in patients with cancer with atrial fibrillation and/or atrial flutter. JACC CardioOncol. 2020;2(5):747‐754. doi: 10.1016/j.jaccao.2020.09.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Chen S, Hellkamp A, Alhanti B, et al. Abstract 12118: Patterns of anticoagulation use in atrial fibrillation and active cancer: observations from Medicare. Circulation. 2021;114(suppl 1):A12118. [Google Scholar]
- 28. Astellas Pharma Canada, Inc . Product Monograph, Including Patient Medication Information: PrXtandi® enzalutamide capsules. Astellas Pharma Canada, Inc.; 2024. Accessed April 30, 2025. https://www.astellas.com/ca/system/files/pdf/Xtandi_PM_EN.pdf [Google Scholar]
- 29. Janssen Inc . Product Monograph, Including Patient Medication Information: PrERLEADA™ apalutamide tables. Janssen Inc.; 2018. Accessed April 30, 2025. https://pdf.hres.ca/dpd_pm/00046215.PDF [Google Scholar]
- 30. Janssen Inc . Product Monograph: PrZYTIGA® abiraterone acetate tables. Janssen Inc.; 2018. Accessed April 30, 2025. https://pdf.hres.ca/dpd_pm/00044223.PDF [Google Scholar]
- 31. Levy AR, O’Brien BJ, Sellors C, Grootendorst P, Willison D. Coding accuracy of administrative drug claims in the Ontario Drug Benefit database. Can J Clin Pharmacol. 2003;10(2):67‐71. [PubMed] [Google Scholar]
- 32. Tonelli M, Wiebe N, Fortin M, et al. Methods for identifying 30 chronic conditions: application to administrative data. BMC Med Inform Decis Mak. 2015;15(1):31. doi: 10.1186/s12911-015-0155-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD‐9‐CM and ICD‐10 administrative data. Med Care. 2005;43(11):1130‐1139. doi: 10.1097/01.mlr.0000182534.19832.83 [DOI] [PubMed] [Google Scholar]
- 34. Alotaibi GS, Wu C, Senthilselvan A, McMurtry MS. The validity of ICD codes coupled with imaging procedure codes for identifying acute venous thromboembolism using administrative data. Vasc Med. 2015;20(4):364‐368. doi: 10.1177/1358863x15573839 [DOI] [PubMed] [Google Scholar]
- 35. Porter J, Mondor L, Kapral MK, Fang J, Hall RE. How reliable are administrative data for capturing stroke patients and their care. Cerebrovasc Dis Extra. 2016;6(3):96‐106. doi: 10.1159/000449288 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Youngson E, Welsh RC, Kaul P, McAlister F, Quan H, Bakal J. Defining and validating comorbidities and procedures in ICD‐10 health data in ST‐elevation myocardial infarction patients. Medicine (Baltimore). 2016;95(32):e4554. doi: 10.1097/md.0000000000004554 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Arnason T, Wells PS, van Walraven C, Forster AJ. Accuracy of coding for possible warfarin complications in hospital discharge abstracts. Thromb Res. 2006;118(2):253‐262. doi: 10.1016/j.thromres.2005.06.015 [DOI] [PubMed] [Google Scholar]
- 38. Burles K, Innes G, Senior K, Lang E, McRae A. Limitations of pulmonary embolism ICD‐10 codes in emergency department administrative data: let the buyer beware. BMC Med Res Methodol. 2017;17(1):89. doi: 10.1186/s12874-017-0361-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Al‐Ani F, Shariff S, Siqueira L, Seyam A, Lazo‐Langner A. Identifying venous thromboembolism and major bleeding in emergency room discharges using administrative data. Thromb Res. 2015;136(6):1195‐1198. doi: 10.1016/j.thromres.2015.10.035 [DOI] [PubMed] [Google Scholar]
- 40. Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simulat Comput. 2009;38(6):1228‐1234. doi: 10.1080/03610910902859574 [DOI] [Google Scholar]
- 41. Li F, Morgan KL, Zaslavsky AM. Balancing covariates via propensity score weighting. J Am Stat Assoc. 2018;113(521):390‐400. doi: 10.1080/01621459.2016.1260466 [DOI] [Google Scholar]
- 42. Thomas LE, Li F, Pencina MJ. Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. JAMA. 2020;323(23):2417‐2418. doi: 10.1001/jama.2020.7819 [DOI] [PubMed] [Google Scholar]
- 43. Borenstein M, Hedges LV, Higgins JP, Rothstein HR. A basic introduction to fixed‐effect and random‐effects models for meta‐analysis. Res Synth Methods. 2010;1(2):97‐111. doi: 10.1002/jrsm.12 [DOI] [PubMed] [Google Scholar]
- 44. Partlett C, Riley RD. Random effects meta‐analysis: coverage performance of 95% confidence and prediction intervals following REML estimation. Stat Med. 2017;36(2):301‐317. doi: 10.1002/sim.7140 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Benchimol EI, Smeeth L, Guttmann A, et al. The REporting of studies Conducted using Observational Routinely‐collected health Data (RECORD) statement. PLoS Med. 2015;12(10):e1001885. doi: 10.1371/journal.pmed.1001885 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Canadian Cancer Society . Cancer statistics at a glance. Canadian Cancer Society; 2025. Accessed March 1, 2025. https://www.cancer.ca/en/cancer‐information/cancer‐101/cancer‐statistics‐at‐a‐glance/?region=on#:~:text=Cancer%20is%20the%20leading%20cause,women%20will%20die%20from%20cancer. [Google Scholar]
- 47. Short NJ, Connors JM. New oral anticoagulants and the cancer patient. Oncologist. 2014;19(1):82‐93. doi: 10.1634/theoncologist.2013-0239 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Steffel J, Collins R, Antz M, et al. 2021 European Heart Rhythm Association practical guide on the use of non‐vitamin K antagonist oral anticoagulants in patients with atrial fibrillation. Europace. 2021;12(10):1612‐1676. doi: 10.1093/europace/euab065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Musgrave KM, Power K, Laffan M, O'Donnell JS, Thachil J, Maraveyas A. Practical treatment guidance for cancer‐associated thrombosis—managing the challenging patient: a consensus statement. Crit Rev Oncol Hematol. 2022;171:103599. doi: 10.1016/j.critrevonc.2022.103599 [DOI] [PubMed] [Google Scholar]
- 50. Leblanc K, Edwards SJ, Dranitsaris G, et al. Drug interactions between androgen receptor axis‐targeted therapies and antithrombotic therapies in prostate cancer: Delphi consensus. Cancers (Basel). 2024;16(19):3336. doi: 10.3390/cancers16193336 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Carrier M, Cameron C, Delluc A, Castellucci L, Khorana AA, Lee AYY. Efficacy and safety of anticoagulant therapy for the treatment of acute cancer‐associated thrombosis: a systematic review and meta‐analysis. Thromb Res. 2014;134(6):1214‐1219. doi: 10.1016/j.thromres.2014.09.039 [DOI] [PubMed] [Google Scholar]
- 52. Durand M, Schnitzer ME, Pang M, et al. Comparative effectiveness and safety of direct oral anticoagulants versus vitamin K antagonists in nonvalvular atrial fibrillation: a Canadian multicentre observational cohort study. CMAJ Open. 2020;8(4):E877‐E886. doi: 10.9778/cmajo.20200055 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Potdar R, Gartrell BA, Given R, et al. Concomitant use of oral anticoagulants in patients with advanced prostate cancer receiving apalutamide: a post‐hoc analysis of TITAN and SPARTAN studies. Am J Cancer Res. 2022;12(1):445‐450. [PMC free article] [PubMed] [Google Scholar]
- 54. Jha P, Deboer D, Sykora K, Naylor CD. Characteristics and mortality outcomes of thrombolysis trial participants and nonparticipants: a population‐based comparison. J Am Coll Cardiol. 1996;27(6):1335‐1342. doi: 10.1016/0735-1097(96)00018-6 [DOI] [PubMed] [Google Scholar]
- 55. Manuel DG, Rosella LC, Stukel TA. Importance of accurately identifying disease in studies using electronic health records. BMJ. 2010;341:c4226. doi: 10.1136/bmj.c4226, 2010. [DOI] [PubMed] [Google Scholar]
- 56. Frere C, Farge D, Schrag D, Prata PH, Connors JM. Direct oral anticoagulant versus low molecular weight heparin for the treatment of cancer‐associated venous thromboembolism: 2022 updated systematic review and meta‐analysis of randomized controlled trials. J Hematol Oncol. 2022;15(1):69. doi: 10.1186/s13045-022-01289-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material
Data Availability Statement
The data that support the findings of this study are available from ICES. Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the author(s) with the permission of ICES.