Background: Enzalutamide (ENZ) and abiraterone (ABI) are approved for castration-resistant prostate cancer (CRPC). Cardiovascular toxicity was reported for ENZ and ABI in clinical trials. While existing observational studies showed that ENZ would lower risk of cardiovascular disease (CVD) compared to ABI, residual and unmeasured confounding is still concerning.
Objectives: This study aimed to comprehensively compare the CVD risk of ENZ and ABI in metastatic CRPC patients with or without pre-existing CVD.
Methods: We conducted a retrospective cohort study using SEER-Medicare data (2012-2019) on androgen deprivation therapy treated patients (≥65 y.o.) initiating ENZ or ABI. The primary composite outcome was CVD-related hospitalizations (first of ≥1 hospitalization with primary diagnosis of myocardial infarction (MI), ischemic heart disease (IHD), stroke, heart failure (HF), atrial fibrillation (AF), venous thromboembolism (VTE), or hypertension (HTN)). Secondary outcomes were CVD-related outpatient visits (first claim of IHD (≥1), AF (≥2 in 2 years), VTE (≥2 claims ≥7 days apart and ≤ 31 days), and HTN (≥2 in 2 years), and each specific CVD. Patients were followed from the index date (ENZ or ABI initiation) until the earliest of the outcome of interest, treatment discontinuation, death, disenrollment or study period end. We used inverse probability of treatment weighting (IPTW) with baseline confounders (demographics, PC and CVD medication use, and pre-existing CVD) selected based on prior studies and estimated CVD risk using a Fine-Gray subdistribution hazard model with death as a competing risk. Subgroup analysis was conducted in patients without pre-existing CVD.
Results: ENZ and ABI groups included 1557 and 2250 patients, respectively (both groups: 90% White, mean age=78 y.o., 60%-70% had pre-existing CVD and managed with medication). There was non-statistically significant difference between ENZ and ABI groups in CVD-hospitalization (Hazard Ratio (HR) 0.86, 95% confidence interval (CI) 0.65-1.13), CVD outpatient visits (HR 0.95, 95% CI 0.86-1.05), and each of the CVD events (CVD HR, 95% CI: MI 0.55, 0.27-1.14; HF 0.87, 0.70-1.09; Stroke 1.63, 0.82-3.22; IHD 1.00,0.85-1.16; AF 0.88, 0.73-1.06; HTN 0.88, 0.77-1.01; VTE 0.95, 0.86-1.04). ENZ was associated with lower CVD risk compared to ABI in patients without pre-existing CVD (HR 0.76, 95% CI 0.57-1.00).
Conclusions: Our study suggested lower, but not statistically significant difference in CVD risk between ENZ and ABI; ENZ was associated with lower CVD risk in patients without pre-existing CVD. Our study provides additional information to support clinicians’ decision making on ENZ/ABI prescribing in metastatic CRPC patients with different CVD risk.
