Abstract
Background/Aim: In this study, we aimed to evaluate the cardiovascular adverse event (CVAE) signals associated with three proteasome inhibitors (PIs), carfilzomib, bortezomib, and ixazomib, using the Japanese Adverse Drug Event Report (JADER) database, with anticancer agents as a reference.
Patients and Methods: Spontaneous adverse event reports restricted to anticancer agents were extracted from the JADER database between April 2004 and March 2025. A disproportionality analysis was conducted for six predefined CVAE categories based on narrow Standardized MedDRA Queries. Additional analyses were stratified by age (≥70 vs. <70 years) and sex.
Results: In total, 12,181 CVAE reports were identified: 2,303 for carfilzomib, 6,851 for bortezomib, and 3,027 for ixazomib. The overall analysis detected multiple CVAE signals for carfilzomib and bortezomib, including cardiac failure and arrhythmia, whereas no signals were detected for ixazomib. The age-stratified analysis performed revealed new signals that were not observed overall, including venous thromboembolism in older adult patients receiving carfilzomib, cardiomyopathy in those receiving bortezomib, and cardiac failure in those receiving ixazomib. Sex-stratified analysis identified male-specific signals, including thromboembolic and arrhythmic events for carfilzomib, cardiomyopathy for bortezomib, and both cardiac failure and cardiomyopathy for ixazomib. The JADER database analysis identified multiple CVAE signals for carfilzomib and bortezomib, along with new age- and sex-specific signals not observed in the overall population.
Conclusion: These findings highlight the potential influence of patient demographics on CVAE risk and support tailored monitoring strategies for patients receiving PIs, particularly older adult and male patients.
- Proteasome inhibitors
- cardiovascular adverse events
- Japanese Adverse Drug Event Report database
- disproportionality analysis
Introduction
Proteasome inhibitors (PIs), including bortezomib, carfilzomib, and ixazomib, play a central role in the treatment of multiple myeloma by inducing apoptosis of malignant plasma cells via inhibition of the ubiquitin-proteasome system (1-3). While these agents have demonstrated significant therapeutic efficacy, concerns have emerged regarding their potential to cause cardiovascular adverse events (CVAEs), such as heart failure, arrhythmia, and myocardial infarction (4, 5).
Among these three agents, carfilzomib has been shown to be most strongly associated with cardiotoxicity, possibly because of its irreversible inhibitory mechanism and pharmacokinetic properties (4). In our previous analyses using the US Food and Drug Administration Adverse Event Reporting System (FAERS), disproportionality signals for various cardiac events were detected for carfilzomib and bortezomib, whereas no apparent signals were observed for ixazomib (6-9).
The clinical background of patients with multiple myeloma is often complex (10, 11). Most patients are older adults and they may have had prior exposure to cardiotoxic therapies, including anthracyclines and palliative radiation therapy (12, 13). Additionally, the use of corticosteroids and immunomodulatory agents such as dexamethasone and lenalidomide may increase the risk of cardiovascular diseases (14). Differences in formulation and administration route may also influence cardiotoxic profiles; notably, carfilzomib and bortezomib are injectable agents, whereas ixazomib is administered orally (1-3).
We previously conducted comprehensive disproportionality analyses using the Japanese Adverse Drug Event Report (JADER) database for each of the three PIs individually and found consistent signals of cardiac failure across all agents (6-8). Additionally, our previous FAERS-based study, which focused on Food and Drug Administration-approved anticancer agents as the reference population, identified signals for cardiac failure with carfilzomib and bortezomib (9). However, these past studies evaluated each drug independently and did not perform direct comparisons within a unified analytical framework.
In this study, we compared the disproportionality of CVAEs associated with bortezomib, carfilzomib, and ixazomib using the JADER database, employing a common background population consisting of other anticancer agents. We further explored whether patient characteristics influenced the observed reporting patterns by conducting subgroup analyses stratified by age and sex. This analysis was intended to complement our previous findings and provide additional insights into the cardiotoxicity profiles of PIs in the Japanese real-world clinical context.
Patients and Methods
Data source and case selection. The JADER database is publicly available on the Japanese Pharmaceuticals and Medical Devices Agency website (https://www.pmda.go.jp). Since its release on April 1, 2004, JADER has provided open access to adverse event (AE) reports and patient information collected within Japan. The database consisted of four distinct datasets: patient demographics (DEMO), drug information (DRUG), adverse events (REAC), and medical history (HIST). The DEMO file contains basic demographic information such as sex, age, and year of report; the DRUG file provides data on drug name, administration route, and start date; the REAC file records AEs, including their onset date and outcomes; and the HIST file documents the patients’ primary underlying diseases. The AEs in the REAC dataset were coded according to version 28 of the Medical Dictionary for Regulatory Activities (MedDRA) and categorized using Preferred Terms (PTs). In the DRUG file, three classifications were used to describe drug involvement: “suspected drugs,” “concomitant drugs,” and “drug interactions.” We restricted our disproportionality analysis to drugs classified as “suspected drugs” to specifically evaluate drug-associated AEs.
Each table was joined by an identification number and duplicates were removed to create the suspected drug data table. The table contained 2,489,875 reports.
To equalize the background population, only reports of drugs in the WHO ATC codes L01 (ANTINEOPLASTIC AGENTS) and L02 (ENDOCRINE THERAPY) were selected from the reports in the suspected drug data table to create an anticancer drug data table. This table contained 784,153 reports on 202 relevant drugs.
To perform the stratified analysis, we created a data table for males only (n=423,000), females only (n=235,643), those aged ≥70 years (n=299,966), and those aged <70 years (n=422,622) (Figure 1).
Flowchart for constructing a data analysis table. Reports were extracted from the Japanese Adverse Drug Event Report (JADER) database (April 2004-March 2025). Four datasets (DEMO, DRUG, REAC, and HIST) were merged, and records of intravitreal administration of carfilzomib, bortezomib, or ixazomib as suspected drugs were identified. Duplicate reports sharing the same case ID, drug name, or event term were excluded. The resulting dataset was used for the disproportionality analysis. The anticancer drug data table included drugs with WHO ATC codes (L01: ANTINEOPLASTIC AGENTS; L02: ENDOCRINE THERAPY).
Definitions of AEs. CVAEs were defined using 21 PTs from MedDRA version 28.0, selected from the narrow scope of Standardized MedDRA Queries (SMQs) (15). PTs were categorized into six groups: cardiac failure, arrhythmia, ischemic heart disease, cardiomyopathy, thromboembolic events, and hemorrhagic events (Table I).
Cardiovascular adverse event (CVAE) categories defined using the narrow scope of the Standardized MedDRA Queries (SMQs) in MedDRA version 28.0.
Disproportionality analysis. Disproportionality analysis using spontaneous reporting databases is a widely accepted method for detecting potential associations between specific drugs and AEs (16). Reports registered in the JADER database between April 2004 and March 2025 were included. Reporting odds ratios (RORs) were calculated to evaluate whether the reporting frequency of a given AE was disproportionately higher for a specific drug than for all other drugs in the database.
RORs were estimated using 2×2 contingency tables as follows:
a: number of reports of CVAEs associated with PI drugs;
b: number of reports of other AEs associated with PI drugs;
c: number of reports of AEs associated with all other anticancer drugs;
d: number of reports of other AEs associated with all other anticancer drugs.
The ROR was calculated using the following formula:
The 95% confidence interval (CI) was computed as follows:
Disproportionality was considered positive when the lower boundary of the 95% CI exceeded 1.0, indicating a statistically significant increase in the reporting frequency.
Statistical analysis. Statistical analyses were performed using JMP Pro® 18.0 (SAS Institute, Cary, NC, USA).
Ethics approval. The study was designed in compliance with the Ethical Guidelines for Epidemiological Studies established by the Ministry of Health, Labour, and Welfare in Japan and the ethical principles of the Declaration of Helsinki. Given that fully anonymized data extracted from the publicly available JADER database were analyzed without involving direct patient contact or intervention, the Hyogo Medical University Ethics Committee confirmed that formal ethical approval and informed consent were not necessary.
Consent to participate/publish. This study used anonymized data from the JADER database, which is publicly available from the Pharmaceuticals and Medical Devices Agency. As this database contains de-identified patient information that is publicly accessible, institutional review board approval was not required. This study complied with the Agency’s terms of use for the JADER database and was conducted in accordance with the Declaration of Helsinki.
Results
Of the 784,153 anticancer drug AE reports evaluated, 12,181 study-relevant were identified: 2,303 for carfilzomib, 6,851 for bortezomib, and 3,027 for ixazomib. The majority of the patients were aged 70 years or older for carfilzomib and ixazomib, whereas patients in their 60s comprised the largest age group for bortezomib. Males accounted for 49.2%, 52.6%, and 44.6% of reports on carfilzomib, bortezomib, and ixazomib, respectively (Table II).
Characteristics of patients with adverse event (AE) reports associated with carfilzomib, bortezomib and ixazomib.
Overall disproportionality analysis (Table III). Hemodynamic and circulatory disorders: Signals were detected for carfilzomib in cardiac failure (ROR=11.73, 95% CI=10.34-13.31) and pulmonary hypertension (ROR=6.94, 95% CI=4.29-11.24), as well as for bortezomib in cardiac failure (ROR=1.76, 95% CI=1.49-2.07) and shock-associated circulatory or cardiac conditions (excluding torsade de pointes) (ROR=1.95, 95% CI=1.48-2.58).
Number of reports, reporting odds ratios (RORs), and 95% confidence intervals (CIs) for the adverse events (AEs) associated with carfilzomib, bortezomib, and ixazomib, using anticancer drugs as the reference population in the JADER database.
Ischemic and thrombotic events: For carfilzomib, signals were detected for myocardial infarction (ROR=1.90, 95% CI=1.05-3.43) and embolic and thrombotic events, arterial (ROR=4.53, 95% CI=3.55-5.78).
Arrhythmia: Carfilzomib showed signals for conduction defects (ROR=4.44, 95% CI=2.91-6.77) and supraventricular tachyarrhythmias (ROR=6.75, 95% CI=4.57-9.97). Bortezomib showed signals for disorders of the sinus node function (ROR=7.54, 95% CI=4.39-12.96), supraventricular tachyarrhythmias (ROR=2.68, 95% CI=1.88-3.83), and ventricular tachyarrhythmia (ROR=3.36, 95% CI=2.20-5.14).
Cardiomyopathy: A signal was detected for carfilzomib (ROR=1.89, 95% CI=1.12-3.21).
Hemorrhagic and other events: No disproportionality was detected for any of the drugs. No signals were detected for ixazomib in any category during the overall analysis.
In the performed age-stratified analysis (Table IV), new signals not detected in the overall analysis were identified: for carfilzomib, venous thromboembolism in patients aged ≥70 years (ROR=1.95, 95% CI=1.17-3.25); for bortezomib, cardiomyopathy in patients aged ≥70 years (ROR=1.80, 95% CI=1.02-3.19); and for ixazomib, cardiac failure in patients aged ≥70 years (ROR=1.64, 95% CI=1.20-2.24).
Number of reports, reporting odds ratios (RORs), and 95% confidence intervals (CIs) for the selected adverse events (AEs) associated with carfilzomib, bortezomib, and ixazomib, stratified by age (<70 years vs. ≥70 years).
In the sex-stratified analysis (Table V), new signals not detected in the overall analysis were identified: for carfilzomib, venous thromboembolism (ROR=1.67, 95% CI=1.03-2.69) and ventricular tachyarrhythmia (ROR=3.40, 95% CI=1.27-9.11) in males; for bortezomib, cardiomyopathy in males (ROR=2.21, 95% CI=1.34-3.62); and for ixazomib, cardiac failure (ROR=1.80, 95% CI=1.19-2.73) and cardiomyopathy (ROR=2.94, 95% CI=1.46-5.91) in males.
Number of reports, reporting odds ratios (RORs), and 95% confidence intervals (CIs) for the selected adverse events (AEs) associated with carfilzomib, bortezomib, and ixazomib, stratified by sex (male vs. female).
Discussion
In this study, we evaluated the CVAE signals associated with three PIs, carfilzomib, bortezomib, and ixazomib, using the JADER database, restricting the comparator group to anticancer agents. This approach aimed to reduce confounding caused by differences in the underlying disease and treatment context, thereby providing a more clinically relevant comparison.
In the overall analysis, disproportionality signals were detected in multiple CVAE categories for carfilzomib and bortezomib, whereas no signals were detected for ixazomib. For carfilzomib, signals were observed in categories such as cardiac failure, pulmonary hypertension, supraventricular tachyarrhythmias, conduction defects, and cardiomyopathy. For bortezomib, signals were detected in cardiac failure, shock-associated circulatory or cardiac conditions, supraventricular tachyarrhythmia, ventricular tachyarrhythmia, and disorders of sinus node function. These findings are consistent with the reports from clinical trials and other pharmacovigilance studies, including our prior FAERS-based comparative analyses (2, 9, 17-19). The absence of overall signals for ixazomib is consistent with previous JADER reports and may be related to differences in the route of administration and pharmacological properties (3, 20).
The different cardiovascular toxicity profiles observed among the three PIs can be attributed to their distinct inhibition mechanisms and pharmacological properties. Irreversible proteasome inhibition by carfilzomib results in sustained cellular protein dysfunction, whereas reversible inhibition by bortezomib allows for the recovery of proteasomal activity (4). Ixazomib demonstrated the lowest cardiovascular toxicity, likely owing to its oral administration and reduced systemic exposure (21). Cardiomyocytes are particularly vulnerable to proteasome inhibition because of their high protein turnover rates and dependence on protein homeostasis to maintain contractile function (22). Inhibition of the ubiquitin-proteasome system leads to the accumulation of misfolded proteins, triggering endoplasmic reticulum stress (23, 24), mitochondrial dysfunction (25, 26), and the activation of pro-apoptotic pathways (27, 28). These cellular stress pathways may explain the CVAEs observed clinically, with age- and sex-specific signals reflecting the differential baseline cardiovascular reserves in these patient subgroups (29). In line with these mechanistic considerations, recent experimental studies have provided more detailed insights into carfilzomib-associated cardiotoxicity. Experimental evidence indicates that carfilzomib-specific inhibition of the proteasome β2 subunit disrupts protein homeostasis in cardiomyocytes, leading to reductions in contractile force and heart rate (30). In addition, cardiometabolic syndrome, characterized by hypertension, diabetes, and dyslipidemia, has been shown to exacerbate carfilzomib-induced cardiac dysfunction in preclinical models (31). These mechanistic findings may help explain the differences in the reporting frequency and patterns of cardiovascular adverse events among the three proteasome inhibitors analyzed in the present study.
Stratified analyses revealed signals not observed in the overall population, suggesting potential age- and sex-related differences in CVAE reporting. In patients aged ≥70 years, new signals for venous thromboembolism with carfilzomib, cardiomyopathy with bortezomib, and cardiac failure with ixazomib were identified. These findings may reflect an age-related susceptibility to cardiovascular dysfunction, possibly due to physiological aging or comorbidities (32). Differences in baseline age and sex distributions among the patients treated with the three PIs may partly explain these subgroup-specific findings. The predominance of older adult patients in the carfilzomib-treated and ixazomib-treated groups may have contributed to the emergence of age-specific signs such as venous thromboembolism and cardiac failure. Similarly, the higher proportion of male patients in the carfilzomib-treated and bortezomib-treated groups may be relevant to the detection of male-specific thromboembolic or arrhythmic events (33, 34). In the sex-stratified analysis performed, male-specific signals included venous thromboembolism and ventricular tachyarrhythmia with carfilzomib, cardiomyopathy with bortezomib, and both cardiac failure and cardiomyopathy with ixazomib. These patterns have highlighted the potential influence of patient demographics on CVAE reporting and the value of subgroup analyses in post-marketing safety assessments.
When compared with our previous FAERS analysis, which detected signals for cardiac failure and arrhythmia for carfilzomib and bortezomib, the use of an anticancer-only comparator in the present study may have influenced signal detection, particularly in the subgroup analyses (9). This methodological refinement allows for a more focused evaluation within a homogeneous treatment population and may reduce the potential for overestimating disproportionality owing to background risk differences.
Study limitations. Disproportionality analysis using a spontaneous reporting database cannot establish causality or provide accurate incidence estimates. Detailed information on patient background, concomitant medications, and comorbidities was lacking. A particularly important limitation is the potential impact of reduced statistical power in stratified analyses. For some AEs with significant signals in the overall analysis, the signals disappeared after stratification, likely reflecting a decrease in statistical power owing to the smaller number of reports. CIs based on small counts tended to be wide, increasing the risk of false-negatives even when a true risk was present; however, statistical significance was not reached. Conversely, some signals detected in the overall analysis may represent false-positives owing to confounding factors or reporting bias.
In conclusion, our analysis of the JADER database has identified multiple CVAE signals for carfilzomib and bortezomib, as well as new age- and sex-specific signals that were not observed in the overall population. The integration of these findings with mechanistic and clinical evidence suggests that both cardiomyocyte injury and vascular toxicity may underlie the observed patterns. These results highlight the need for proactive cardiovascular monitoring in patients receiving PIs, with particular attention needed for older adult and male patients, to mitigate risks and optimize treatment outcomes.
Acknowledgements
The Authors would like to thank Editage (www.editage.com) for the English language editing.
Footnotes
Authors’ Contributions
Masaki Fujiwara: Conceptualization, investigation, visualization, writing - original draft, writing - review and editing. Shuji Nagano: Writing - review and editing. Yoshihiro Uesawa: Writing - review and editing. Mayako Uchida: Writing - review and editing. Nobuyuki Muroi: Writing - supervision, review and editing. Tadashi Shimizu: Conceptualization, visualization, supervision, writing - original draft, Writing - review and editing.
Data Availability
The data supporting the findings of this study are available at https://www.info.pmda.go.jp/fukusayoudb/CsvDownload.jsp. All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding authors.
Conflicts of Interest
All Authors declare no conflicts of interest.
Artificial Intelligence (AI) Disclosure
During the preparation of this manuscript, the authors used ChatGPT based on GPT-4o and 5 models (OpenAI, San Francisco, CA, USA) to assist with English language editing, including improving the clarity, tone, and readability. All content generated using this tool was critically reviewed and revised by the Authors, who take full responsibility for the integrity and accuracy of the final manuscript.
- Received November 22, 2025.
- Revision received December 20, 2025.
- Accepted January 9, 2026.
- Copyright © 2026 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
Jump to section
Related Articles
Cited By...
- No citing articles found.