Abstract
Introduction
Recent real-world studies compared effectiveness and safety of enzalutamide (ENZA) and abiraterone acetate (AA) for metastatic castration-resistant prostate cancer (mCRPC). The growing evidence needs further substantiation with long-term data. This study, the first to use German data, investigated cardiovascular (CV) event risk and overall survival (OS) in patients initiating ENZA or AA. AA (2012) and ENZA (2014) are widely used for mCRPC in Germany.
Methods
This retrospective study used data of chemotherapy-naïve patients with mCRPC on ENZA or AA (2012–2020) from two German claims databases (AOK PLUS and GWQ ServicePlus). The primary endpoint was time to first CV event (CV-related hospitalization) analyzed via a meta-analysis of Cox proportional hazard models of propensity score-matched (PSM) intention-to-treat cohorts. Other endpoints were baseline characteristics, CV event rate, number of CV events per patient, and OS.
Results
Of 2240 patients in the total study population (ENZA, 828; AA, 1412), 796 PSM patients were included in each group. ENZA patients were older and had a higher prevalence of some comorbidities, but without meaningful differences after PSM. Further, 386 patients had ≥ 1 CV event (ENZA, 172; AA, 214). ENZA was associated with a significantly lower risk of CV events (hazard ratio [HR] 0.70, 95% confidence interval [CI] 0.57–0.86, p = 0.001, I2 = 0.0%), CV event rate (0.17 vs 0.23 per person-year; event rate ratio 0.75, 95% CI 0.61–0.92, p = 0.006; I2 = 38.0%), fewer recurrent CV events (HR 0.77, 95% CI 0.61–0.96, p = 0.024; I2 = 0.0%), and prolonged OS (HR 0.79, 95% CI 0.71–0.89, p < 0.001) than AA.
Conclusions
The unmatched ENZA cohort had higher average age and more comorbidities than the AA cohort, but no meaningful differences were noted after PSM. ENZA was associated with a significantly lower risk of CV events and improved OS.
Supplementary Information
The online version contains supplementary material available at 10.1007/s12325-025-03132-8.
Keywords: Abiraterone, Cardiovascular events, Castration-resistant prostate cancer, Enzalutamide, Overall survival, Propensity score matching
Plain Language Summary
This study looked at two treatments for a type of advanced prostate cancer in German patients. The treatments are called enzalutamide (ENZA) and abiraterone acetate (AA). We wanted to see which treatment was safer and more effective, especially in terms of heart-related problems and overall survival. We looked at data from patients who had not received chemotherapy and were treated with either ENZA or AA between 2012 and 2020. We compared the time it took for the first heart-related problem to occur in these patients. We also looked at the number of heart-related problems per each patient and overall survival. The study included 2240 patients (828 treated with ENZA and 1412 treated with AA). An early evaluation found that patients treated with ENZA were older and had a higher prevalence of other health problems. However, after matching the patients in each group based on certain characteristics, we found that these differences were not meaningful. Our results showed that a total of 386 patients had at least one heart-related problem after starting on ENZA or AA. We found that patients treated with ENZA had a lower risk of heart-related problems and lived longer than those treated with AA. In conclusion, even though patients treated with ENZA were older and had more health problems, they had a lower risk of heart-related problems.
Supplementary Information
The online version contains supplementary material available at 10.1007/s12325-025-03132-8.
Key Summary Points
| Why carry out this study? |
| Recently, real-world studies have been conducted to evaluate the effectiveness and safety of enzalutamide (ENZA) vs abiraterone acetate (AA) for metastatic castration-resistant prostate cancer (mCRPC). Our study aims to provide additional support for the existing comparative evidence on long-term outcomes in chemotherapy-naïve mCRPC. This was the first study that used German data. Both medicines have been widely available to treat mCRPC since 2012 (AA) and 2014 (ENZA) in Germany. |
| The aim of the study was to investigate cardiovascular (CV) event risk and overall survival (OS) in patients with mCRPC initiating ENZA or AA. |
| What was learned from the study? |
| The patients in the unmatched ENZA cohort had higher average age and more comorbidities than those in the AA cohort, but no differences in baseline characteristics were observed after propensity-score matching (PSM). |
| Treatment with ENZA was associated with a significantly lower risk of CV events, CV events rate, longer treatment duration, and improved OS than AA. |
| The results need to be confirmed in an earlier disease stage of metastatic hormone sensitive prostate cancer, where patients would have longer exposure to ENZA and AA. |
Introduction
Prostate cancer (PC) is the fourth most prevalent cancer worldwide and the second most common cancer in men globally; in Germany, it ranks as the most frequent cancer in men [1, 2]. Patients with metastatic PC who initially respond to androgen-deprivation therapy (ADT) eventually become castration-resistant as the disease progresses. Abiraterone acetate (AA; plus prednisolone) and enzalutamide (ENZA) are androgen receptor pathway inhibitors (ARPIs) recommended for intensifying ADT in metastatic castration-resistant prostate cancer (mCRPC) [3–5].
While AA and ENZA have been shown to prolong overall survival (OS), the cardiovascular (CV) risk profile of a patient should also be considered in treatment decisions [4]. A recent meta-analysis reported an increased CV risk with ARPIs + ADT compared to ADT alone, although comparisons between the ARPIs were not conducted [6]. Given that CV disease is the most frequent comorbidity and cause of death in patients with PC, guidelines recommend monitoring CV events in patients receiving ARPIs [4, 7, 8]. CV events have been associated with increased morbidity and mortality and reduced quality of life in these patients [9, 10].
In randomized controlled trials (RCTs), patients tend to be healthier and have fewer comorbidities, since it is common practice to exclude patients with specific CV conditions. Conversely, patients in real-world evidence (RWE) studies are typically older and have a higher burden of comorbidities. Therefore, generalizing results from the RCTs evaluating AA and ENZA (COU-AA301, COU-AA302, AFFIRM, and PREVAIL), which excluded patients with comorbid cardiac conditions, may be difficult [11–13]. Alternatively, data from real-world practice on the effectiveness and safety of AA and ENZA for chemotherapy-naïve mCRPC remain limited, with existing evidence being primarily from the USA, France, Italy, and Canada. These existing RWE studies consistently demonstrated an increased risk of CV toxicities, such as all-cause, CV-related, and heart failure (HF)-related hospitalizations, and atrial tachyarrhythmia, among AA-treated patients [14–17], and a longer OS in ENZA-treated patients [18–21]. Nevertheless, given the unknown biases and limitations inherent to RWE studies, the generalizability of these findings to the German setting has yet to be validated. This growing evidence also needs further substantiation with long-term data. Therefore, using a state-of-the-art approach, an intention-to-treat (ITT) analysis using propensity score-matched (PSM) cohorts, the current study aimed to estimate the risk of CV events and evaluate OS among patients with chemotherapy-naïve mCRPC initiating ENZA or AA within standard clinical practice in Germany. CV events were defined in terms of CV-related hospitalization, because hospitalization significantly impacts patient care and the healthcare system, encompassing more than just major adverse cardiac events (MACE). This is a common approach in the literature as well [22].
Methods
Study Design and Data Source
The AVENGER study was a retrospective database analysis using anonymized data from two German claims data sources, AOK PLUS and GWQ ServicePlus (referred to as AOK and GWQ hereafter). In the absence of a national/comprehensive dataset, these two databases reflect the German healthcare system, covering approximately 11.0% of the population with statutory health insurance and relying on International Statistical Classification of Diseases and Related Health Problems, German Modification (ICD-10-GM) codes [23]. These datasets encompass the full range of healthcare-related services, where the routinely collected data include demographic data, inpatient and outpatient diagnoses, procedures, admissions, drug usage, and general practitioner and specialist visits. Since this study was based on anonymized and nonidentifiable data, no informed consent or ethics approval from an institutional review board was required. Permission was obtained to access and use patient-level data from AOK PLUS and GWQ, in line with data protection concepts of relevant local regulations [24]. The research followed legal and internal standard processes of database handlers that communicated only aggregated findings with the sponsor.
Study Population
The study included adults with ≥ 2 confirmed outpatient diagnoses and/or ≥ 1 inpatient diagnosis of PC, ≥ 1 prescription or inpatient application of ENZA and/or AA, ≥ 1 inpatient and/or confirmed outpatient diagnosis of metastasis, and castration-resistant prostate cancer (CRPC) defined by medical or surgical castration ≥ 3 months prior to the index date (ENZA/AA initiation) between January 2012 and December 2020. The study period was from January 2010 to June 2021 (Supplementary material, Fig. S1). Patients with another primary malignant neoplasm, prior prescription of chemotherapy and/or radium-223, or ENZA or AA use before or on the index date, were excluded. Eligibility was assessed using ICD-10-GM codes (Supplementary material, Table S1). Of note, AA and ENZA received European Union approval for metastatic hormone-sensitive prostate cancer (mHSPC) in November 2017 and April 2021, respectively.
Study Objectives and Endpoints
The primary objective was the risk of CV event, with time to first CV event, defined as CV-related hospitalization, as the associated primary endpoint. Reasons for the first CV-related hospitalization were defined by the primary diagnosis (ICD-10-GM codes) for the first CV-related hospitalization during follow-up (Supplementary material, Table S2).
Secondary objectives included comparing the ENZA and AA cohorts in terms of the CV event rate (number of CV events per patient-year [PPY], including recurrent events), baseline demographic and clinical patient characteristics, treatment sequences, and treatment durations. Number needed to harm (NNH; the number of patients who need to receive treatment to experience one CV event) was also estimated. Treatment sequence analysis included first-, second-, and third-line life-prolonging PC treatments (including patients not receiving subsequent treatment) and their durations; time to treatment discontinuation or death; time from index date to next treatment; and concomitant therapy.
Exploratory endpoints included OS and adverse events (AEs). OS was defined as the time to all-cause death. AEs observed “very frequently” or “frequently,” based on the summary of product characteristics for ENZA and AA, were identified via documented inpatient or outpatient diagnoses [7, 8]. AE assessment was conducted from the index date until 180 days post-index, treatment discontinuation/switch, death, study end, or end of insurance (whichever occurred first). Severe fractures or falls leading to hospitalization were measured among all patients, while for all other events, only patients without the diagnosis throughout the 24-month covariate assessment window were analyzed (incident events).
Statistical Analysis
All analyses were carried out separately per database as it is not allowed to combine them at patient level.
The primary endpoint was analyzed via a meta-analysis of outcomes from CPH models carried out separately in each database using the ITT approach within the PSM cohort. The PSM methodology was undertaken to render outcome comparisons on different treatments more interpretable by removing significant differences in key baseline risk factors. I2 statistics were reported to present statistical heterogeneity [25]. A PSM method was deployed to obtain two groups of patients (ENZA/AA) with matched baseline characteristics. Covariates were included in the PS calculation, to ensure that a robust comparative analysis was performed (Supplementary material, Table S3). Balance was assessed using the standardized mean difference (SMD) of individual covariates between ENZA/AA patients: 0.1 was taken as the threshold to indicate important imbalance and an SMD of < 0.1 was considered to indicate a negligible difference in the mean or prevalence of a covariate between comparison groups [26]. Sensitivity analyses for primary endpoint included inverse probability of treatment weighting (IPTW) analysis (to adjust for baseline patient characteristics differences), on-treatment analysis (analyzing follow-up data while patients were on treatment only), and an analysis using a cause-specific hazards (CPH) model (considering death as a competing risk rather than a censoring event). Subgroups of patients with or without prior CV disease were also analyzed.
Descriptive analyses of patient characteristics and treatment patterns were performed. Time to next treatment and time to treatment discontinuation were analyzed using Kaplan–Meier (KM) methods. OS was analyzed in a CPH model. AEs were analyzed descriptively, and the difference between treatment groups was evaluated using SMD cutoff of 0.1.
Categorical variables were described by the number and proportion of patients, along with overall sample sizes. Continuous variables were summarized using mean (standard deviation [SD]) and median (interquartile range [IQR]). Results of KM analyses were presented using the median time to event, first and third quartiles, and 95% confidence intervals (CIs). Results of CPH models were presented using hazard ratios (HRs), 95% CIs, and p values. Meta-analyses of CPH results were depicted using forest plots. Statistical significance was assessed at the alpha level of 0.05, corresponding to a 95% CI.
Results
Of 52,632 and 48,106 patients with PC diagnosis in the AOK and GWQ databases, respectively, 2240 patients met the eligibility criteria and were included in the study (ENZA, 828; AA, 1412). The PSM cohorts included 796 patients in each group (Supplementary material, Fig. S2).
Baseline Characteristics
Before matching, baseline characteristics differed between the ENZA and AA cohorts: patients in the ENZA cohort were older (77.2 years vs 75.7 years; SMD, AOK 0.19 and GWQ 0.14) and had a numerically higher frequency of diabetes (43.0% vs 35.0%), obesity (25.0% vs 20.3%), arrhythmia (34.3% vs 28.8%), renal impairment (40.7% vs 34.3%), and mean CHA2DS2-VASc score (3.6 vs 3.3). After PSM, no meaningful differences were noted between cohorts (SMD < 0.1) in the overall population (Table 1), or in each database cohort (Supplementary material, Table S3).
Table 1.
Baseline characteristics and comorbidities
| Unmatched cohorts | PSM cohorts | |||
|---|---|---|---|---|
| ENZA (N = 828) | AA (N = 1412) | ENZA (N = 796) | AA (N = 796) | |
| Age, mean (SD)a | 77.2 (8.1) | 75.7 (8.3) | 77.0 (8.1) | 76.8 (8.0) |
| Urologist visits | ||||
| Patients with ≥ 1 urologist visit, n (%) | 817 (98.7) | 1377 (97.5) | 785 (98.6) | 787 (98.9) |
| Number of urologist visits, mean (SD) | 20.9 (10.9) | 20.9 (12.5) | 20.6 (10.5) | 21.4 (12.0) |
| All-cause hospitalizations | ||||
| Patients with ≥ 1 hospitalization, n (%) | 632 (76.3) | 1117 (79.1) | 607 (76.3) | 628 (78.9) |
| Number of hospitalizations, mean (SD) | 2.5 (2.8) | 2.7 (2.9) | 2.5 (2.8) | 2.7 (2.9) |
| Number of hospitalization days, mean (SD) | 15.2 (22.5) | 15.4 (24.1) | 15.2 (22.5) | 15.5 (25.8) |
| Observational time (months), mean (SD) | 24.6 (18.4) | 24.0 (19.0) | 24.9 (18.5) | 22.7 (18.6) |
| CCI, mean (SD) | 10.4 (3.0) | 10.4 (2.9) | 10.3 (3.0) | 10.5 (3.0) |
| CHA2DS2-VASc score, mean (SD) | 3.6 (1.5) | 3.3 (1.5) | 3.6 (1.5) | 3.5 (1.5) |
| Comorbidities, n (%) | ||||
| Anemia | 35 (4.2) | 40 (2.8) | 33 (4.1) | 22 (2.8) |
| Diabetes | 356 (43.0) | 494 (35.0) | 334 (42.0) | 313 (39.3) |
| Obesity | 207 (25.0) | 286 (20.3) | 193 (24.2) | 187 (24.0) |
| Hyperlipidemia | 453 (54.7) | 753 (53.3) | 435 (54.6) | 433 (54.4) |
| Primary hypertension | 693 (83.7) | 1154 (81.7) | 665 (83.5) | 667 (83.8) |
| Angina | 73 (8.8) | 122 (8.6) | 71 (8.9) | 69 (8.7) |
| Cardiomyopathy | 28 (3.4) | 46 (3.3) | 26 (3.3) | 25 (3.1) |
| Stroke | 57 (6.9) | 100 (7.1) | 56 (7.0) | 54 (6.8) |
| Atherosclerosis | 150 (18.1) | 234 (16.6) | 141 (17.7) | 136 (17.1) |
| COPD | 112 (13.5) | 183 (13.0) | 107 (13.4) | 100 (12.6) |
| Liver damage or abnormality | 148 (17.9) | 259 (18.3) | 143 (18.0) | 137 (17.2) |
| Renal impairment | 337 (40.7) | 484 (34.3) | 319 (40.1) | 318 (40.0) |
| Urinary tract infection | 254 (30.7) | 416 (29.5) | 243 (30.5) | 247 (31.0) |
| CV diseases defined as CV events in the study | ||||
| Hypertension | 704 (85.0) | 1165 (82.5) | 674 (84.7) | 670 (84.2) |
| Heart failure | 236 (28.5) | 372 (26.3) | 224 (28.1) | 219 (27.5) |
| Ischemic heart disease | 314 (37.9) | 495 (35.1) | 299 (37.6) | 290 (36.4) |
| Nonrheumatic valve disorders | 129 (15.6) | 259 (18.3) | 124 (15.6) | 156 (19.6) |
| Atrioventricular block | 78 (9.4) | 124 (8.8) | 75 (9.4) | 85 (10.7) |
| Arrhythmia | 284 (34.3) | 407 (28.8) | 265 (33.3) | 259 (32.5) |
| Cerebrovascular disease | 187 (22.6) | 347 (24.6) | 181 (22.7) | 177 (22.2) |
| Peripheral vascular disease | 226 (27.3) | 332 (23.5) | 216 (27.1) | 195 (24.5) |
| Pulmonary circulation disorders | 45 (5.4) | 95 (6.7) | 44 (5.5) | 66 (8.3) |
| Deep vein thrombosis | 60 (7.2) | 87 (6.2) | 56 (7.0) | 46 (5.8) |
| Cardiac arrest | 1 (0.1) | 2 (0.1) | 1 (0.1) | 2 (0.3) |
AA abiraterone acetate, CCI Charlson Comorbidity Index, CHA2DS2-VASc congestive heart failure, hypertension, age, diabetes mellitus, prior stroke or transient ischemic attack or thromboembolism, vascular disease, age, sex category, COPD chronic obstructive pulmonary disease, CV cardiovascular, ENZA enzalutamide, PSM propensity score matching, SD standard deviation
Comorbidities measured during the 2-year covariate assessment window
aSMD ≥ 0.1 before PSM in each cohort; < 0.1 after PSM
Risk of Cardiovascular Events
A total of 386 patients had ≥ 1 CV event (ENZA, 172; AA, 214). The meta-analysis showed a lower risk of CV events in the ENZA cohort than the AA cohort (HR 0.70, 95% CI 0.57–0.86, p = 0.001, I2 = 0.0%; Fig. 1). Median time to first CV event was longer for ENZA than AA (AOK, 58.0 vs 45.0 months; GWQ, 79.6 vs 62.2 months; Fig. 1). Sensitivity analyses confirmed the robustness of the results (Supplementary material, Table S4).
Fig. 1.
Risk of CV events and overall survival in propensity score-matched intention-to-treat cohorts. AA abiraterone acetate, CI confidence interval, CV cardiovascular, ENZA enzalutamide, HR hazard ratio, IPTW inverse probability of treatment weighting, ITT intention-to-treat, PSM propensity score matching, CV event defined as CV-related hospitalization
In an analysis comparing the CV event risk within subgroups of patients with (N = 1434, 90.0%) and without (N = 158, 10.0%) prior CV disease for ENZA vs AA, a significantly lower risk of CV events was seen in the ENZA cohort than the AA cohort (with CV disease, HR 0.72, 95% CI 0.59–0.88, p = 0.002, I2 = 0.0%; without CV disease, HR 0.28, 95% CI 0.09–0.85, p = 0.025; I2 = 0.0%) (Supplementary material, Fig. S3). Specifically, when analyzed by type of prior CV disease within these subgroups, ENZA showed a lower CV event risk than AA across all CV diseases except nonrheumatic valve disease (Supplementary material, Fig. S4). HF, cerebrovascular disease, and ischemic heart disease were common reasons for CV-related hospitalization, occurring similarly in ENZA- and AA-treated patients; HF, 29.0% vs 30.8%; cerebrovascular disease, 18.6% vs 19.6%; ischemic heart disease, 18.0% vs 16.4%, respectively (Table 2). However, arrhythmia occurred at a higher frequency in the AA cohort than the ENZA cohort (13.6% vs 6.4%). Interestingly, hypertension as the reason for CV-related hospitalization was seen at a similar frequency in both groups (ENZA, n = 10, 5.8%; AA, n = 11, 5.1%) (Table 2), with a low event-rate of 1.3% (10/796) to 1.4% (11/796) for each cohort.
Table 2.
Reasons for first CV-related hospitalization post-index
| Variable | AOK & GWQ | AOK | GWQ | |||
|---|---|---|---|---|---|---|
| ENZA | AA | ENZA | AA | ENZA | AA | |
| Total number of patients with ≥ 1 hospitalization | 173 | 214 | 100 | 122 | 72 | 92 |
| Reason for first hospitalization n (%) | ||||||
| Hypertension | 10 (5.8) | 11 (5.1) | 4 (4.0) | 6 (4.9) | 6 (8.2) | 5 (5.4) |
| Heart failure | 50 (29.0) | 66 (30.8) | 35 (35.0) | 41 (33.6) | 15 (20.6) | 25 (27.2) |
| Ischemic heart disease | 31 (18.0) | 35 (16.4) | 15 (15.0) | 19 (15.8) | 16 (21.9) | 16 (17.4) |
| Nonrheumatic valve disorders | 4 (2.3) | 1 (0.5) | 3 (3.0) | 1 (0.8) | 1 (1.4) | 0 (0.0) |
| Atrioventricular block | 6 (3.5) | 3 (1.4) | 6 (6.0) | 2 (1.6) | 0 (0.0) | 1 (1.1) |
| Arrhythmia | 11 (6.4) | 29 (13.6) | 4 (4.0) | 15 (12.3) | 7 (9.6) | 14 (15.2) |
| Cerebrovascular disease | 32 (18.6) | 42 (19.6) | 21 (21.0) | 21 (17.2) | 11 (15.1) | 21 (22.8) |
| Peripheral vascular disease | 7 (4.1) | 11 (5.1) | 2 (2.0) | 9 (7.4) | 5 (6.9) | 2 (2.2) |
| Pulmonary circulation disorders | 10 (5.8) | 10 (4.7) | 4 (4.0) | 5 (4.1) | 6 (8.2) | 5 (5.4) |
| Deep vein thrombosis | 8 (4.7) | 5 (2.3) | 5 (5.0) | 3 (2.5) | 3 (4.1) | 2 (2.2) |
| Cardiac arrest | 3 (1.7) | 1 (0.5) | 1 (1.0) | 0 (0.0) | 2 (2.7) | 1 (1.1) |
AA abiraterone acetate, CV cardiovascular, ENZA enzalutamide
The ENZA cohort had a lower CV event rate (PPY) during follow-up (0.17 vs 0.23; event rate ratio 0.75, 95% CI 0.61–0.92, p = 0.006, I2 = 38.0%; Table 3). The calculated NNH for ENZA versus AA was 18 for pooled data (30 for AOK, 12 for GWQ; Table 3). ENZA was associated with a significantly lower risk of recurrent CV events than AA (HR 0.77, 95% CI 0.61–0.96, p = 0.024, I2 = 38.0%).
Table 3.
Cardiovascular event rate ratio and risk for recurrent events in propensity score-matched cohorts
| Variable | AOK & GWQ | AOK | GWQ | |||
|---|---|---|---|---|---|---|
| ENZA (N = 796) | AA (N = 796) | ENZA (N = 457) | AA (N = 457) | ENZA (N = 339) | AA (N = 339) | |
| Total number of CV events | 283 | 340 | 167 | 180 | 116 | 160 |
| Total observational time (person-years) | 1648.2 | 1502.3 | 952.9 | 864.0 | 695.3 | 638.3 |
| Event rate: number of CV events per person-year | 0.17 | 0.23 | 0.18 | 0.21 | 0.17 | 0.25 |
| Event rate ratio (ENZA vs AA) (95% CI), p value | 0.75 (0.61–0.92), 0.006a | 0.82 (0.68–1.04), 0.108 | 0.67 (0.52–0.85), 0.001 | |||
| Patients with at least 1 CV event | 172 | 214 | 100 | 122 | 72 | 92 |
| Number of CV events per person, mean (SD) | 0.36 (1.05) | 0.43 (1.09) | 0.37 (0.92) | 0.39 (0.79) | 0.34 (0.93) | 0.47 (1.10) |
| Risk of recurrent CV events, HR (95% CI), p valuea | 0.77 (0.61–0.96), 0.024a | 0.85 (0.67–1.08), 0.175 | 0.67 (0.51–0.89), 0.005 | |||
| Number needed to harm | 18 | 30 | 12 | |||
AA abiraterone acetate, CI confidence interval, CV cardiovascular, ENZA enzalutamide, HR hazard ratio, SD standard deviation
aI2 38%; meta-analysis was performed using a random effects model and restricted maximum likelihood procedure to estimate tau-squared
Overall Survival
In the PSM cohorts, a total of 595 (74.7%) ENZA initiators and 677 (85.1%) AA initiators died during follow-up, with the former showing significantly prolonged OS (HR 0.79, 95% CI 0.71–0.89, p < 0.001, I2 = 0.0%; Fig. 1). Sensitivity analyses confirmed the robustness of the OS results (Supplementary material, Table S4).
Treatment Patterns
Frequencies of concomitant medications and interventions were similar in both groups (Table 4). The median first-line treatment duration was longer for ENZA than AA (AOK, 10.9 vs 8.7 months; GWQ, 9.3 vs 8.5 months; Table 5) in the PSM cohorts. A total of 38.2% of ENZA patients and 38.4% of AA patients did not receive any second-line treatment (13.8% of ENZA patients and 11.4% of AA patients discontinued treatment with no subsequent therapy, and 24.4% of ENZA patients and 27.0% of AA patients died while on first-line treatment) or were censored while on first-line treatment (ENZA, 10.9%; AA, 5.9%). The median time from the index date to first-line treatment (ENZA/AA) discontinuation or death was slightly longer for ENZA than AA in the PSM cohorts (AOK, 11.1 vs 8.7 months; GWQ, 9.5 vs 8.6 months; Table 5).
Table 4.
Treatment received at baseline and in second-line and third-line settings in the propensity score-matched cohorts
| n (%) | ENZA (N = 796) | AA (N = 796) |
|---|---|---|
| At baselinea | ||
| First-generation nonsteroidal antiandrogens | 586 (73.6) | 579 (72.7) |
| LHRH agonists/antagonists | 773 (97.1) | 779 (97.9) |
| LHRH agonists | 752 (94.5) | 749 (94.1) |
| LHRH antagonists | 76 (9.6) | 74 (9.3) |
| Bone targeting agents | 324 (40.7) | 371 (46.6) |
| Surgical castration (orchiectomy) | 16 (2.0) | 15 (1.9) |
| Radiation therapy/brachytherapy | 144 (18.1) | 147 (18.5) |
| Duration of the first-line treatment, months, mean (SD) | 15.1 (14.9) | 12.7 (12.7) |
| Second-line | ||
| Did not receive second-line treatment | ||
| Died while on first-line treatment | 194 (24.4) | 215 (27.0) |
| No second-line treatmentb | 110 (13.8) | 91 (11.4) |
| Censored while on first-line treatment | 87 (10.9) | 47 (5.9) |
| Received second-line treatment | ||
| ENZA | 99 (12.4)c | 234 (29.4) |
| AA | 172 (21.6) | 67 (8.4)c |
| Docetaxel | 100 (12.6) | 110 (13.8) |
| Other | 34 (4.3) | 32 (4.0) |
| Third-line | ||
| Did not receive third-line treatment | ||
| Death/censored/no treatment after first line | 391 (49.1) | 353 (44.4) |
| Died while on second-line treatment | 76 (9.5) | 118 (14.8) |
| No third-line treatmentb | 68 (8.5) | 89 (11.2) |
| Censored while on second-line treatment | ||
| Received third-line treatment | ||
| ENZA | 75 (9.4) | 68 (8.5) |
| AA | 62 (7.8) | 49 (6.2) |
| Docetaxel | 36 (4.5) | 46 (5.8) |
| Other | 30 (3.8) | 18 (2.3) |
| ENZA | 23 (2.9) | 28 (3.5) |
AA abiraterone acetate, CI confidence interval, ENZA enzalutamide, IQR interquartile range, LHRH luteinizing hormone-releasing hormone, TTD time to treatment discontinuation or death, TTNT time to next treatment
aDuring baseline period of 24 months leading up to the index date
bIncludes any patient who discontinued treatment while still observable (thereafter, either censored at end of insurance/study period or died during the study period)
cSecond-line treatment was assumed if there was an observed medication supply gap of ≥ 45 days used in the treatment line identification algorithm. A limitation is patients may still be on first-line treatment in the event of long treatment gaps followed by resumption of the same agent, potential dose reductions leading to extended coverage from a single prescription, or poor therapy adherence delaying the need for prescription refills
Table 5.
Time-to-event analysis in the propensity score-matched cohorts
| Variable | AOK | GWQ | ||
|---|---|---|---|---|
| ENZA (N = 457) | AA (N = 457) | ENZA (N = 339) | AA (N = 339) | |
| Duration of first-line treatment, months | ||||
| n | 457 | 457 | 339 | 339 |
| Median (IQR) | 10.9 (4.3, 23.2) | 8.7 (4.2, 16.5) | 9.3 (3.8, 19.5) | 8.5 (3.7, 16.4) |
| Time to first-line treatment discontinuation or death, months | ||||
| n | 402 | 434 | 307 | 315 |
| Median (95% CI) | 11.1 (9.4–13.3) | 8.7 (8.0–9.6) | 9.5 (7.9–11.0) | 8.6 (7.6–10.0) |
| Time to next treatment, months | ||||
| n | 226 | 241 | 179 | 202 |
| Median (95% CI) | 21.1 (17.9–25.6) | 15.5 (13.0–18.2) | 18.6 (15.3–22.8) | 13.3 (12.1–14.8) |
AA abiraterone acetate, CI confidence interval, ENZA enzalutamide, IQR interquartile range
Adverse Events of Special Interest (AESI)
Severe fractures were reported in 2 (0.3%) patients on AA and none on ENZA. Severe falls were reported in 8 (1.0%) patients on AA and 12 (1.5%) patients on ENZA. The occurrence of other AEs was similar between the treatment cohorts (Supplementary material, Table S5), except for hypokalemia, which was more frequent in patients on AA (12.5% vs 3.2%, SMD = − 0.09).
Discussion
In this large German claims database study, we found that patients on ENZA had a lower CV event rate, a lower likelihood of CV events, and prolonged OS than those on AA. The results were robust as confirmed across a series of sensitivity and subgroup analyses and align with studies conducted in comparable countries using different datasets [14–21, 27].
Previous studies compared CV risk between AA and ENZA, but methodological differences limit direct comparisons with our study. Administrative claims database studies in the USA [14, 28] and Canada [15] found a higher risk of myocardial infarction (MI) or stroke, major adverse CV events (MI, ischemic stroke, and HF), and CV-related hospitalization risk (HF, ischemic stroke, and acute MI) in the AA than ENZA cohort. Notably, the risk of HF-related hospitalization was significantly higher in the AA cohort than the ENZA cohort [14, 15]. Similarly, a couple of pharmacovigilance studies indicated higher odds of CV events (overall, MI, arrythmia, and HF) for AA, but no signal was found for ENZA [16, 17]. The 2022 European cardio-oncology guideline highlights the higher CV risk associated with AA than ENZA and emphasizes the significance of CV risk assessment and monitoring in patients with PC [4]. Our study further supports these findings by providing additional evidence that, when used for treating mCRPC, ENZA is associated with lower CV risk than AA, particularly in a German context.
Differences in CV risk between AA and ENZA may stem from differences in their mechanisms of action. AA causes deficient glucocorticoid synthesis resulting in hypokalemia, which in turn can lead to arrhythmia [29, 30]. In our study, the higher CV event incidence in the AA group may be due to more frequent hypokalemia (3.2% for ENZA vs 12.5% for AA). Arrhythmia-related hospitalizations were also more common with AA (6.4% for ENZA vs 13.6% for AA), supporting this hypothesis. Hu et al. also reported a higher incidence rate and likelihood of arrhythmia-related hospitalizations for AA versus ENZA (HR 3.01, 95% CI 0.66–13.73) [15]. Additionally, prednisone, administered to manage mineralocorticoid side effects of AA, may contribute to CV risk, as cumulative steroid use increases AEs in patients with CRPC [31]. These findings underscore the need to study CV events in the mHSPC setting because of longer treatment exposures.
Besides CV event risk, a notable finding from the current study was the longer OS in patients receiving ENZA versus AA, which may result from improved efficacy. Previous RWE studies across different regions have consistently shown prolonged OS with ENZA in patients with mCRPC [18–21]. The HR of OS observed in the current study (HR 0.79, 95% CI 0.71–0.89) is consistent with other studies, including Tagawa et al. (HR 0.84, 95% CI 0.76–0.94) [18], Shoen et al. (HR 0.90, 95% CI 0.84–0.96) [19], and George et al. (HR 0.90, 95% CI 0.96–0.86) [20] in the USA, and Scailteux et al. in France (HR 0.90, 95% CI 0.85–0.96) [21].
As a result of its retrospective design, this study has some limitations, including potential missing data for confounding variables and coding imperfections related to outpatient diagnoses. Socio-economic status and ethnicity were not considered in the current analysis; they are a possible area for further work. The severity of baseline CV comorbidities, which impacts the primary outcome, could not be determined or matched between treatment groups. There could be a potential risk of misclassification due to approval time for use of AA and ENZA in mHSPC, which occurred in November 2017 and April 2021, respectively. Despite limitations, our study robustly estimated CV risk by employing both PSM and IPTW, addressing baseline differences. This dual approach is unique and demonstrates methodological rigor, enhancing the reliability and validity of our findings. Further, our findings were supported by numerous sensitivity assessments ensuring robustness across various contexts.
Conclusion
The unmatched ENZA cohort had higher average age and more comorbidities than the AA cohort, but these differences were not meaningful after PSM. Despite the longer treatment duration, ENZA was associated with a significantly lower risk of CV events, lower CV event rates, fewer recurrent CV events, and prolonged OS, compared to AA. These findings need to be substantiated in an earlier disease setting of mHSPC, where patients would have longer exposure to both ARPIs.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgments
Medical Writing/Editorial Assistance
Support for medical writing, editing, and graphic design were provided by Leena Patel (PhD), Olga Klibanov (PharmD), Jessica Warren (BS), and Barbara Aguero (graphic designer) from IQVIA, funded by the study sponsors.
Author Contributions
Axel S. Merseburger had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Substantial contributions to conception and design: Stefan G. Spitzer, Axel S. Merseburger, Sophia Junker, Andrew Chilelli, Matthias Stoelzel, Alexis Serikoff, Eugen Dornstauder, Philipp Hahn. Acquisition of study data: Andrew Chilelli, Alexis Serikoff, Eugen Dornstauder, Philipp Hahn. Analysis of study data: Sophia Junker, Andrew Chilelli, Matthias Stoelzel, Alexis Serikoff, Eugen Dornstauder, Philipp Hahn. Interpretation of study data: Carsten-Henning Ohlmann, Armen Aprikian, Stefan G. Spitzer, Axel S. Merseburger, Andrew Chilelli, Matthias Stoelzel, Alexis Serikoff, Eugen Dornstauder. Data validation and analysis: Sophia Junker, Philipp Hahn. All authors had access to the study results, reviewed and revised the manuscript, and approved the final draft submitted for publication.
Funding
Open Access funding enabled and organized by Projekt DEAL. This study was funded by Astellas Pharma Inc. and Pfizer Inc., the co-developers of enzalutamide. The Rapid Service Fee and Open Access Fee are funded by Astellas Pharma Inc.
Data Availability
All data generated or analyzed during this study, which support the findings of this study, are included within this article and its supplementary material files. Researchers may access analyses not present in the manuscript from the corresponding author upon reasonable request. For the Astellas criteria on data sharing, see: https://clinicalstudydatarequest.com/Study-Sponsors/Study-Sponsors-Astellas.aspx.
Declarations
Conflict of Interest
Carsten-Henning Ohlmann: received support for the present publication from Astellas; received payment or honoraria for lectures, presentations, speakers bureaus, publication writing or educational events from Astellas, Janssen-Cilag, and AstraZeneca; received payment for expert testimony from Janssen-Cilag; received support for attending meetings and/or travel from Bayer and Janssen-Cilag; and participated on a data safety monitoring board or advisory board of Bayer, Janssen-Cilag, and AstraZeneca. Axel S. Merseburger: received support for the present publication from Astellas; received payment or honoraria for lectures, presentations, speaker bureaus, publication writing, or educational events from Janssen, Astellas, and Bayer. He is an Editorial Board member of Advances in Therapy. He was not involved in the selection of peer reviewers for the manuscript nor any of the subsequent editorial decisions. Armen Aprikian: received payment or honoraria for lectures, presentations, speaker bureaus, publication writing, or educational events; received support for attending meetings and/or travel from Tolmar and Tersera; and has leadership or fiduciary role in other board, society, committee, or advocacy group, paid or unpaid for Canadian Urology Association and Cedars Cancer Foundation. Alexis Serikoff: received support for the present publication from Astellas. Andrew Chilelli: received support for the present publication and attending meetings and/or travel from Astellas. Eugen Dornstauder: received support for the present publication from Astellas. Andrew Chilelli, Eugen Dornstauder, Matthias Stoelzel, Alexis Serikoff: full-time employees of Astellas Pharma. Sophia Junker: is an employee of Ingress-Health HWM GmbH, a Cytel Company that was contracted by Astellas Pharma Europe Ltd. to conduct this research. Philipp Hahn: is an employee of IPAM e.V. that was contracted by Cytel to conduct this research. Stefan G. Spitzer: has received support from Astellas as a consultant.
Ethical Approval
Since this study was based on anonymized and nonidentifiable data, no informed consent or ethics approval from an institutional review board was required. The research followed legal and internal standard processes of database handlers that communicated only aggregated findings with the sponsor. Permission was obtained to access and use patient-level data from AOK PLUS and GWQ, in line with data protection concepts of relevant local regulations [24].
Footnotes
Prior Presentation: A poster presenting part of this work titled “Cardiovascular events and overall survival in patients treated with Abiraterone acetate Vs ENzalutamide for chemotherapy-naïve metastatic castration-resistant prostate cancer in GERmany: AVENGER study” was presented at the 16th European Multidisciplinary Congress on Urological Cancers (EMUC) 2024, 7–10 November 2024, Lisbon, Portugal.
References
- 1.Robert Koch Institute, Association of Population-Based Cancer Registries in Germany. Cancer in Germany 2017/2018. 13th ed. Robert Koch Institute, Berlin; 2022.
- 2.World Health Organization. Global cancer burden growing, amidst mounting need for services. Geneva: WHO; 2024. [PMC free article] [PubMed] [Google Scholar]
- 3.Deutsche Gesellschaft für Urologie e.V. S3-Leitlinie Prostatakarzinom. Version 6.2. 2021. https://www.leitlinienprogramm-onkologie.de/fileadmin/user_upload/Downloads/Leitlinien/Prostatatkarzinom/Version_6/LL_Prostatakarzinom_Langversion_6.2.pdf. Accessed 13 Aug 2024.
- 4.Lyon AR, López-Fernández T, Couch LS, et al. 2022 ESC Guidelines on cardio-oncology developed in collaboration with the European Hematology Association (EHA), the European Society for Therapeutic Radiology and Oncology (ESTRO) and the International Cardio-Oncology Society (IC-OS). Eur Heart J Cardiovasc Imag. 2022;23:e333–465. 10.1093/ehjci/jeac106. [DOI] [PubMed] [Google Scholar]
- 5.Parker C, Castro E, Fizazi K, et al. Prostate cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2020;31:1119–34. 10.1016/j.annonc.2020.06.011. [DOI] [PubMed] [Google Scholar]
- 6.El-Taji O, Taktak S, Jones C, Brown M, Clarke N, Sachdeva A. Cardiovascular events and androgen receptor signaling inhibitors in advanced prostate cancer: a systematic review and meta-analysis. JAMA Oncol. 2024;10(7):874–84. 10.1001/jamaoncol.2024.1549. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Janssen Biotech. ZYTIGA 250 mg tablets Summary of Product Characteristics. https://www.ema.europa.eu/en/documents/product-information/zytiga-epar-product-information_en.pdf. Accessed 13 Aug 2024.
- 8.Astellas Pharma. Xtandi - 40 mg soft capsules Summary of Product Characteristics. https://www.ema.europa.eu/en/documents/product-information/xtandi-epar-product-information_en.pdf. Accessed 13 Aug 2024.
- 9.Wilk M, Waśko-Grabowska A, Szmit S. Cardiovascular complications of prostate cancer treatment. Front Pharmacol. 2020;11:555475. 10.3389/fphar.2020.555475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.van de Poll-Franse LV, Sadetsky N, Kwan L, Litwin MS. Severity of cardiovascular disease and health-related quality of life in men with prostate cancer: a longitudinal analysis from CaPSURE. Qual Life Res. 2008;17:845–55. 10.1007/s11136-008-9356-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Fizazi K, Scher HI, Molina A, et al. Abiraterone acetate for treatment of metastatic castration-resistant prostate cancer: final overall survival analysis of the COU-AA-301 randomised, double-blind, placebo-controlled phase 3 study. Lancet Oncol. 2012;13:983–92. 10.1016/s1470-2045(12)70379-0. [DOI] [PubMed] [Google Scholar]
- 12.Miller K, Carles J, Gschwend JE, Van Poppel H, Diels J, Brookman-May SD. The phase 3 COU-AA-302 study of abiraterone acetate plus prednisone in men with chemotherapy-naïve metastatic castration-resistant prostate cancer: stratified analysis based on pain, prostate-specific antigen, and Gleason score. Eur Urol. 2018;74:17–23. 10.1016/j.eururo.2017.08.035. [DOI] [PubMed] [Google Scholar]
- 13.Armstrong AJ, Al-Adhami M, Lin P, et al. Association between new unconfirmed bone lesions and outcomes in men with metastatic castration-resistant prostate cancer treated with enzalutamide: secondary analysis of the PREVAIL and AFFIRM randomized clinical trials. JAMA Oncol. 2020;6:217–25. 10.1001/jamaoncol.2019.4636. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Conover MM, Weaver J, Fan B, et al. Cardiovascular outcomes among patients with castration-resistant prostate cancer: a comparative safety study using US administrative claims data. Prostate. 2023;83:729–39. 10.1002/pros.24510. [DOI] [PubMed] [Google Scholar]
- 15.Hu J, Aprikian AG, Vanhuyse M, Dragomir A. Comparative cardiovascular safety of novel hormonal agents in metastatic castration-resistant prostate cancer using real-world data. Clin Genitourin Cancer. 2022;20:17–24. 10.1016/j.clgc.2021.08.009. [DOI] [PubMed] [Google Scholar]
- 16.Bretagne M, Lebrun-Vignes B, Pariente A, et al. Heart failure and atrial tachyarrhythmia on abiraterone: a pharmacovigilance study. Arch Cardiovasc Dis. 2020;113:9–21. 10.1016/j.acvd.2019.09.006. [DOI] [PubMed] [Google Scholar]
- 17.Cone EB, Reese S, Marchese M, et al. Cardiovascular toxicities associated with abiraterone compared to enzalutamide—a pharmacovigilance study. EClinicalMedicine. 2021;36:100887. 10.1016/j.eclinm.2021.100887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Tagawa ST, Ramaswamy K, Huang A, et al. Survival outcomes in patients with chemotherapy-naive metastatic castration-resistant prostate cancer treated with enzalutamide or abiraterone acetate. Prostate Cancer Prostatic Dis. 2021;24:1032–40. 10.1038/s41391-021-00318-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Schoen MW, Carson KR, Eisen SA, et al. Survival of veterans treated with enzalutamide and abiraterone for metastatic castrate resistant prostate cancer based on comorbid diseases. Prostate Cancer Prostatic Dis. 2023;26:743–50. 10.1038/s41391-022-00588-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.George DJ, Ramaswamy K, Yang H, et al. Real-world overall survival with abiraterone acetate versus enzalutamide in chemotherapy-naïve patients with metastatic castration-resistant prostate cancer. Prostate Cancer Prostatic Dis. 2024. 10.1038/s41391-024-00816-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Scailteux L-M, Campillo-Gimenez B, Kerbrat S, et al. Overall survival among chemotherapy-naive patients with castration-resistant prostate cancer under abiraterone versus enzalutamide: a direct comparison based on a 2014–2018 French population study (the SPEAR cohort). Am J Epidemiol. 2021;190:413–22. 10.1093/aje/kwaa190. [DOI] [PubMed] [Google Scholar]
- 22.Hu J, Aprikian AG, Vanhuyse M, Dragomir A. Comparative cardiovascular safety of novel hormonal agents in metastatic castration-resistant prostate cancer using real-world data. Clin Genitourin Cancer. 2022;20(1):17–24. [DOI] [PubMed] [Google Scholar]
- 23.Niegisch G, Grimm M-O, Hardtstock F, et al. Treatment patterns and clinical outcomes in metastatic urothelial carcinoma: a German retrospective real-world analysis. Future Oncol. 2024;20(19):1351–66. 10.2217/fon-2023-1065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Verbraucherschutz BdJuf. Zehntes Buch Sozialgesetzbuch - Sozialverwaltungsverfahren und ozialdatenschutz - (SGB X) § 75 Übermittlung von Sozialdaten für die Forschung und Planung. 2022. https://dejure.org/gesetze/SGB_X/75.html. Accessed 23 Jan 2024.
- 25.Deeks JJ, Higgins JPT, Altman DG, eds. Chapter 10: Analysing data and undertaking meta-analyses. In: Higgins JPT, Thomas J, Chandler J, et al, editors. Cochrane Handbook for Systematic Reviews of Interventions. Version 6.4 (updated August 2023). Cochrane, 2023. www.training.cochrane.org/handbook. Accessed 13 Aug 2024.
- 26.Zhao QY, Luo JC, Su Y, Zhang YJ, Tu GW, Luo Z. Propensity score matching with R: conventional methods and new features. Ann Transl Med. 2021;9:812. 10.21037/atm-20-3998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Iacovelli R, Ciccarese C, Bria E, et al. The cardiovascular toxicity of abiraterone and enzalutamide in prostate cancer. Clin Genitourin Cancer. 2018;16:e645–53. 10.1016/j.clgc.2017.12.007. [DOI] [PubMed] [Google Scholar]
- 28.Kulkarni A, Rubin N, Tholkes A, et al. Risk for stroke and myocardial infarction with abiraterone versus enzalutamide in metastatic prostate cancer patients. ESMO Open. 2021;6:100261. 10.1016/j.esmoop.2021.100261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nkwocha BI, Singh M. Abiraterone-induced hypokalemia: a case report. Cureus. 2023;15:e42533. 10.7759/cureus.42533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Thu Kyaw M, Maung ZM. Hypokalemia-induced arrhythmia: a case series and literature review. Cureus. 2022;14:e22940. 10.7759/cureus.22940. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Schultz NM, Penson DF, Wilson S, et al. Adverse events associated with cumulative corticosteroid use in patients with castration-resistant prostate cancer: an administrative claims analysis. Drug Saf. 2020;43:23–33. 10.1007/s40264-019-00867-6. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data generated or analyzed during this study, which support the findings of this study, are included within this article and its supplementary material files. Researchers may access analyses not present in the manuscript from the corresponding author upon reasonable request. For the Astellas criteria on data sharing, see: https://clinicalstudydatarequest.com/Study-Sponsors/Study-Sponsors-Astellas.aspx.