Comparative risk of arrhythmias associated with systemic antifungal agents: a disproportionality analysis of the FAERS database Authors: Xiaohu Yang, Hao Li, Kai Liu, Lingti Kong Published: 2026 ABSTRACT ObjectiveThe real-world cardiac safety profile of systemic antifungal agents has not been thoroughly investigated. Based on the US Food and Drug Administration Adverse Event Reporting System FDA Adverse Event Reporting System database, this study analyzed the arrhythmogenic toxicity of nine systemic antifungal drugs, aiming to provide references for clinical safe medication practices.Research design and methodsAdverse events were described and classified using arrhythmogenic toxicity-related Standardized MedDRA Queries (SMQs) from the MedDRA. To identify the association between systemic antifungal agents and arrhythmogenic toxicity, this study used four algorithms: Reporting odds ratio (ROR), Proportional reporting ratio (PRR), Multi-item gamma-poisson shrinker (MGPS), and Bayesian confidence propagation Neural Network (BCPNN).ResultsA total of 42,393 reports were included. The ranking of the number of positive signals across four types of SMQs was as follows: Itraconazole (4), Fluconazole (3), Posaconazole (3), Voriconazole (2), Caspofungin (1), Amphotericin B (1), Flucytosine (0), Isavuconazole (0), Micafungin (0). Itraconazole demonstrated the strongest ROR value of 2.95 in “Cardiac arrhythmia terms, nonspecific”. The highest ROR values in “Bradyarrhythmias” were 5.53 for both posaconazole and fluconazole. Fluconazole exhibited higher ROR values than other drugs in both “Tachyarrhythmias” and “Torsade de pointes/QT prolongation”, with values of 3.53 and 14.55, respectively.ConclusionThis study employed four disproportionality analysis methods to analyze the association between systemic antifungal agents and arrhythmogenic toxicity signals. Itraconazole, fluconazole, and posaconazole demonstrated stronger arrhythmogenic risks, whereas micafungin, flucytosine, and isavuconazole showed negative signals across all four SMQs. In clinical practice, individual patient risk should be comprehensively assessed to guide personalized drug selection. Introduction Fungal infections impose a substantial and enduring global disease burden, with over 300 million people suffering from severe fungal diseases and approximately 1.4 million deaths attributable to them annually (Denning, 2024; Kriegl et al., 2025; Zhang et al., 2023). The incidence and mortality rates associated with invasive fungal diseases continue to rise, driven by an increasing number of patients with hematological malignancies, iatrogenic or disease-related immunosuppression/immune dysfunction, and the widespread use of indwelling catheters, surgical procedures, and solid organ transplantation. The treatment of these infections remains highly dependent on systemic antifungal agents (Thompson et al., 2024). Current clinical practice has expanded from classical azoles (e.g., fluconazole) to include newer azoles (e.g., isavuconazole) (Dagher et al., 2022; Lewis et al., 2022; Neofytos et al., 2024), echinocandins (e.g., caspofungin), polyenes (e.g., amphotericin B), and flucytosine, constituting a multi-mechanistic therapeutic arsenal. However, this broader range of drug classes has introduced more heterogeneous safety profiles (Lewis and Wiederhold, 2023; Zhang et al., 2023). Of particular concern is cardiotoxicity. Several antifungal agents can prolong the QT interval and induce Torsades de Pointes (TdP), posing a risk of sudden cardiac death. Therefore, it is necessary to clarify the arrhythmogenic toxicity risk profiles of these commonly used drugs (Cao et al., 2026; Li et al., 2023; Ohyama et al., 2023; Tan et al., 2024; Ünal Yüksekgönül et al., 2021; Yu and Liao, 2022). Due to limitations in enrolled populations and follow-up duration, traditional randomized controlled trials have insufficient power to detect rare or delayed adverse events (AEs) in real-world settings. The US FDA Adverse Event Reporting System (FAERS), a widely used database for post-marketing drug safety surveillance, is now extensively employed for signal detection in pharmacovigilance (Buchanan and Li, 2023; Giunchi et al., 2023; Li et al., 2023; Fernandes et al., 2023; Liabeuf et al., 2025; Ohyama et al., 2023; Fusaroli et al., 2024a; Fusaroli et al., 2024b), providing evidence for label updates and risk communications (Chen et al., 2024; Giunchi et al., 2023; Li et al., 2023). Post-marketing cardiac safety evaluations of systemic antifungals using spontaneous reporting data have also expanded in recent years (Cao et al., 2026; Tan et al., 2024). Consequently, this study utilizes FAERS as the data source to systematically analyze reports of arrhythmogenic toxicity associated with systemic antifungal agents, aiming to provide real-world evidence to inform clinical decision-making regarding their use. Materials and methods Data source The data for this study were sourced from the FAERS database, spanning from the first quarter of 2004 to the second quarter of 2025. The dataset comprises seven data files: patient demographic and administration information (DEMO), drug information (DRUG), adverse events (REAC), patient outcomes (OUTC), reporting source (RPSR), start and end dates for drug treatment (THE), and indications for use or diagnosis (INDI). The following sequential criteria were employed to deduplicate the dataset: (1) For identical CASEIDs, the report with the larger PRIMARYID was retained; (2) Entries with identical PRIMARYIDs were identified as duplicates and excluded. The data processing workflow was illustrated in Figure 1. Flowchart showing data filtering. Data filtering A search was conducted using the generic names (“Isavuconazole”, “Posaconazole”, “Voriconazole”, “Fluconazole”, “Itraconazole”, “Caspofungin”, “Micafungin”, “Amphotericin B”, and “Flucytosine”) and the trade names (“Cresemba”, “Noxafil”, “Vfend”, “Diflucan”, “Sporanox”, “Cancidas”, “Mycamine”, “Fungizone”, and “Ancobon”) listed as the primary suspect drugs. AEs were described and classified using arrhythmogenic toxicity-related Standardized MedDRA Queries (SMQs) from the Medical Dictionary for Regulatory Activities (MedDRA v.27.1). A total of four arrhythmia-related narrow SMQs (“Cardiac arrhythmia terms, nonspecific”, “Bradyarrhythmias”, “Tachyarrhythmias”, and “Torsade de pointes/QT prolongation, QT/TdP”) were included; the specific Preferred Terms (PTs) included in each SMQ are detailed in Supplementary Table S1 (Giunchi et al., 2023; Hung et al., 2023). We also conducted a sensitivity analysis on whether patients received chemotherapy and non-chemotherapy. Data mining To identify the association between systemic antifungal agents and arrhythmogenic toxicity, this study used four algorithms: Reporting odds ratio (ROR), Proportional reporting ratio (PRR), Multi-item gamma-poisson shrinker (MGPS), and Bayesian confidence propagation Neural Network (BCPNN). Table 1 presents the calculation formulas and positive signal thresholds for these four algorithms. To avoid false negatives, only signals that simultaneously met the positive criteria across all four algorithms were considered positive signals (Jiang et al., 2024; Li et al., 2025). TABLE 1Summary of the main algorithms used for signal detection.AlgorithmEquationPositive criteriaROR ROR=ac×bd 95%CI=elnROR±1.961a+1b+1c+1d lower limit of 95%CI>1 a≥3 PRR PRR=aa+b÷cc+d χ2=ad−bc2×a+b+c+da+bc+da+cd+b PRR≥2 χ2≥4,a≥3 BCPNN IC=log2⁡aa+b+c+da+b IC025=elnIC−1.961a+1b+1c+1d IC025>0 MGPS EBGM=aa+b+c+da+ca+b EBGM05=elnEBGM−1.961a+1b+1c+1d EBGM05≥2 N>0 Abbreviations: ROR: reporting odds ratio; PRR: proportional reporting ratio; BCPNN: bayesian confidence propagation neural network; MGPS: multi-item gamma passion shrinker; IC: information component; EBGM: empirical Bayes geometric mean; a: number of reports arising from the suspect adverse event (AE) and suspect drug; b: number of reports arising from the suspect AE, and all other drugs; c: number of reports arising from the suspect drug and other AEs; d: number of reports arising from other drugs and other AEs; CI: confidence interval; χ2: chi-squared; IC025: lower limit of 95% two-sided CI, of the IC; EBGM05: lower limit of 95% one-sided CI, of EBGM. Summary of the main algorithms used for signal detection. Abbreviations: ROR: reporting odds ratio; PRR: proportional reporting ratio; BCPNN: bayesian confidence propagation neural network; MGPS: multi-item gamma passion shrinker; IC: information component; EBGM: empirical Bayes geometric mean; a: number of reports arising from the suspect adverse event (AE) and suspect drug; b: number of reports arising from the suspect AE, and all other drugs; c: number of reports arising from the suspect drug and other AEs; d: number of reports arising from other drugs and other AEs; CI: confidence interval; χ2: chi-squared; IC025: lower limit of 95% two-sided CI, of the IC; EBGM05: lower limit of 95% one-sided CI, of EBGM. Statistical analysis SPSS 26.0 was used for descriptive statistical analysis of the baseline characteristics of reports. R 4.2.2 software (involving the “ggplot2 3.4.2” package and the “pheatmap 1.0.12” package) was employed for disproportionality analysis and data visualization. Results Baseline characteristics Table 2 presents the demographics of the included reports. Among reports with specified age, the age distribution was predominantly in the 18–64.9 years group, followed by the >65 years group. The types of reporters were diverse, including consumers, physicians, pharmacists, and other healthcare professionals. The reporting countries were concentrated in the United States, China, France, Japan, among others, with the United States ranking first in reports for most drugs. TABLE 2Clinical characteristics of the patients.CharacteristicIsavuconazolePosaconazoleVoriconazoleFluconazoleItraconazoleCaspofunginMicafunginAmphotericin BFlucytosine(N = 1,474)(N = 3,809)(N = 12,588)(N = 9,813)(N = 3,899)(N = 2,296)(N = 1950)(N = 6,392)(N = 172)SexFemale570 (38.7%)1,384 (36.3%)4,143 (32.9%)5,071 (51.7%)1,511 (38.8%)824 (35.9%)682 (35.0%)2,212 (34.6%)55 (32.0%)Male780 (52.9%)1904 (50.0%)6,682 (53.1%)3,605 (36.7%)1,615 (41.4%)1,192 (51.9%)1,030 (52.8%)3,493 (54.6%)109 (63.4%)Missing124 (8.4%)521 (13.7%)1763 (14.0%)1,137 (11.6%)773 (19.8%)280 (12.2%)238 (12.2%)687 (10.7%)8 (4.7%)Age(years)<1853 (3.6%)417 (10.9%)1,156 (9.2%)575 (5.9%)286 (7.3%)275 (12.0%)175 (9.0%)931 (14.6%)5 (2.9%)18–64.9265 (18.0%)1,551 (40.7%)4,710 (37.4%)4,439 (45.2%)1,453 (37.3%)1,047 (45.6%)721 (37.0%)3,079 (48.2%)103 (59.9%)>65292 (19.8%)900 (23.7%)3,933 (31.2%)2,404 (24.5%)759 (19.5%)549 (23.9)335 (18.2%)1,259 (19.7%)51 (29.6%)Missing864 (58.6%)941 (24.7%)2,789 (22.2%)2,395 (24.4%)1,401 (35.9%)425 (18.5%)699 (35.8%)1,123 (17.6%)13 (7.6%)Reporter typeConsumer774 (52.5%)695 (18.2%)2,144 (17.0%)2,231 (22.7%)603 (15.5%)165 (7.2%)980 (50.3%)3,016 (47.2%)4 (2.3%)Physician259 (17.6%)1,178 (30.9%)3,914 (31.1%)3,037 (30.9%)1,146 (29.4%)1,119 (48.7%)397 (20.4%)843 (13.2%)42 (24.4%)Pharmacist157 (10.7%)378 (9.9%)1,168 (9.3%)929 (9.5%)507 (13.0%)222 (9.7%)183 (9.4%)471 (7.4%)17 (9.9%)Other health-professional272 (18.4%)1,503 (39.6%)5,018 (39.9%)3,127 (31.9%)1,599 (41%)733 (31.9%)354 (18.1%)1916 (29.9%)105 (61.1%)Missing12 (0.8%)55 (1.4%)344 (2.7%)489 (5.0%)44 (1.1%)57 (2.5%)36 (1.8%)146 (2.3%)4 (2.3%)Reporter country (TOP3)1US 1131 (76.7%)US 2051 (53.9%)US 4698 (37.3%)US 3701 (37.8%)US 1016 (26.1%)US 365 (15.9%)US 868 (44.5%)GB 2697 (42.2%)US 97 (56.4%)2CH 126 (8.5%)FR 385 (10.1%)CN 1078 (8.6%)FR 907 (9.2%)CN 344 (8.8%)FR 306 (13.3%)CN 292 (15.0%)US 1258 (19.7%)JP 15 (8.7%)3JP 60 (4.1%)CN 139 (3.6%)FR 1020 (8.1%)GB 870 (8.9%)JP 294 (7.5%)JP 261 (11.4%)JP 242 (12.4%)JP 252 (3.9%)FR 10 (5.8%)Of country abbreviation: US, united states; CH, switzerland; JP, japan; FR, france; CN, china; GB, united kiongdom. Clinical characteristics of the patients. Of country abbreviation: US, united states; CH, switzerland; JP, japan; FR, france; CN, china; GB, united kiongdom. The annual trend of the number of arrhythmias related reports of each drug was shown in Figure 2. The annual trend of the number of arrhythmias related reports of each systemic antifungal agent. Signal detection The signal strengths of systemic antifungal drugs at the SMQ level were comprehensively evaluated using four algorithms (ROR, PRR, EBGM, and IC), with the results shown in Table 3. The numbers of drug types showing positive signals in the four SMQs—“Cardiac arrhythmia terms, nonspecific”, “Bradyarrhythmias”, “Tachyarrhythmias”, and “QT/TdP”—were 1, 5, 3, and 5, respectively. Among the nine systemic antifungal drugs, only itraconazole showed positive signals in all four SMQs. TABLE 3Signal intensity of arrhythmogenic toxicity caused by systemic antifungal drugs.SMQsDrugFrequencyROR (95% CI)PRR (χ2)EBGM (EBGM05)IC (IC025)Positive signalCardiac arrhythmia terms, nonspecificIsavuconazole10.29 (0.04–2.07)0.29 (1.72)0.29 (0.04)−1.78 (−3.19)NoPosaconazole141.11 (0.66–1.88)1.11 (0.17)1.11 (0.66)0.16 (−0.60)NoVoriconazole450.99 (0.74–1.32)0.99 (0.01)0.99 (0.74)−0.02 (−0.45)NoFluconazole832.13 (1.71–2.64)2.12 (49.36)2.12 (1.71)1.09 (0.75)NoItraconazole432.95 (2.18–3.98)2.94 (55.1)2.94 (2.18)1.56 (1.06)YesCaspofungin101.39 (0.74–2.58)1.38 (1.07)1.38 (0.74)0.47 (−0.45)NoMicafungin81.48 (0.74–2.96)1.48 (1.24)1.48 (0.74)0.56 (−0.47)NoAmphotericin B321.59 (1.12–2.25)1.59 (6.96)1.59 (1.12)0.67 (0.14)NoFlucytosine0NANANANANoBradyarrhythmiasIsavuconazole40.86 (0.32–2.28)0.86 (0.10)0.86 (0.32)−0.22 (−1.47)NoPosaconazole945.53 (4.51–6.77)5.49 (345.15)5.48 (4.47)2.45 (2.09)YesVoriconazole1572.53 (2.16–2.96)2.52 (144)2.52 (2.15)1.33 (1.09)YesFluconazole2925.53 (4.92–6.20)5.48 (1,069.16)5.47 (4.87)2.45 (2.26)YesItraconazole1015.10 (4.19–6.20)5.06 (329.43)5.06 (4.16)2.34 (2.00)YesCaspofungin242.44 (1.63–3.64)2.43 (20.32)2.43 (1.63)1.28 (0.62)YesMicafungin111.49 (0.82–2.69)1.49 (1.77)1.49 (0.82)0.57 (−0.32)NoAmphotericin B451.64 (1.22–2.19)1.64 (11.12)1.63 (1.22)0.71 (0.26)NoFlucytosine0NANANANANoTachyarrhythmiasIsavuconazole90.98 (0.51–1.88)0.98 (0.01)0.98 (0.51)−0.03 (−0.95)NoPosaconazole872.59 (2.10–3.20)2.57 (84.01)2.57 (2.08)1.36 (1.03)YesVoriconazole1831.49 (1.29–1.72)1.49 (29.46)1.49 (1.29)0.57 (0.36)NoFluconazole3683.53 (3.19–3.91)3.50 (658.61)3.50 (3.15)1.81 (1.65)YesItraconazole972.48 (2.03–3.02)2.46 (84.65)2.46 (2.02)1.30 (0.99)YesCaspofungin150.77 (0.46–1.28)0.77 (1.02)0.77 (0.46)−0.37 (−1.08)NoMicafungin201.37 (0.89–2.13)1.37 (2.03)1.37 (0.88)0.46 (−0.2)NoAmphotericin B1001.85 (1.52–2.25)1.84 (38.67)1.84 (1.51)0.88 (0.58)NoFlucytosine32.31 (0.74–7.18)2.30 (2.20)2.30 (0.74)1.20 (−0.66)NoTorsade de pointes/QT prolongationIsavuconazole41.40 (0.53–3.74)1.40 (0.46)1.40 (0.53)0.49 (−0.92)NoPosaconazole1029.85 (8.10–11.98)9.76 (801.68)9.75 (8.02)3.29 (2.88)YesVoriconazole1524.01 (3.42–4.71)4.00 (341.73)3.99 (3.41)2.00 (1.74)YesFluconazole46414.55 (13.27–15.95)14.35 (5,722.33)14.24 (12.99)3.83 (3.66)YesItraconazole907.44 (6.05–9.16)7.39 (497.39)7.38 (6.00)2.88 (2.48)YesCaspofungin183.00 (1.89–4.76)2.99 (23.89)2.99 (1.88)1.58 (0.77)No​Micafungin112.44 (1.35–4.42)2.44 (9.35)2.44 (1.35)1.29 (0.29)NoAmphotericin B613.64 (2.83–4.69)3.63 (116.44)3.63 (2.82)1.86 (1.43)YesFlucytosine24.96 (1.24–19.92)4.94 (6.30)4.94 (1.23)2.31 (−0.58)No Signal intensity of arrhythmogenic toxicity caused by systemic antifungal drugs. Figure 3 presents the forest plot of ROR values. Itraconazole demonstrated a ROR value of 2.95 (95% CI: 2.18–3.98) in “Cardiac arrhythmia terms, nonspecific”. The ROR values in “Bradyarrhythmias” were 5.53 for both posaconazole (95% CI: 4.51–6.77) and fluconazole (95% CI: 4.92–6.20). Fluconazole exhibited positive signal in both “Tachyarrhythmias” and “QT/TdP”, with ROR values of 3.53 (95% CI: 3.19–3.91) and 14.55 (95% CI: 13.27–15.95), respectively. Comparison of ROR at the SMQ level for systemic antifungal agents (forest plot). Figure 4 provides a more intuitive visualization of the differences among various antifungal agents across four algorithms, categorized under four SMQs. Comparison of four algorithms at the SMQ level for systemic antifungal agents (heatmap). This study also conducted sensitivity analysis on whether the patients received chemotherapy and non-chemotherapy, and the results were shown in Supplementary Tables S2, S3 and Figure 5. Comparison of four algorithms at the SMQ level for systemic antifungal agents in patients receiving chemotherapy and non-chemotherapy (heatmap). Abbreviations: chemo, chemotherapy; non-chemo, non-chemotherapy. Clinical outcomes Table 4 presents the association between the frequency of reported AEs for each drug and the clinical outcomes (including Death, Life-Threatening, Disability, Hospitalization, and Other) across different SMQs. For instance, among the adverse reactions related to fluconazole under “Cardiac arrhythmia terms, nonspecific”, the proportions resulting in Death and Life-Threatening outcomes were 16.9% and 18.1%, respectively. TABLE 4Clinical outcomes of arrhythmogenic toxicity caused by systemic antifungal drugs.SMQDrugOutcomeDeathLife-threateningDisabilityHospitalizationOtherMissingCardiac arrhythmia terms, nonspecificIsavuconazole0 (0%)1 (100%)0 (0%)0 (0%)0 (0%)0 (0%)Posaconazole4 (28.6%)5 (35.7%)0 (0%)1 (7.1%)1 (7.1%)3 (21.4%)Voriconazole9 (20.5%)7 (15.9%)0 (0%)2 (4.5%)11 (25.0%)15 (34.1%)Fluconazole14 (16.9%)15 (18.1%)1 (1.2%)3 (3.6%)24 (28.9%)26 (31.3%)Itraconazole3 (7.1%)0 (0%)0 (0%)7 (16.7%)15 (35.7%)17 (40.5%)Caspofungin3 (30.0%)0 (0%)0 (0%)0 (0%)0 (0%)7 (70.0%)Micafungin4 (50.0%)0 (0%)0 (0%)1 (12.5%)3 (37.5%)0 (0%)Amphotericin B8 (25.0%)3 (9.4%)0 (0%)1 (3.1%)1 (3.1%)19 (59.4%)FlucytosineNANANANANANABradyarrhythmiasIsavuconazole2 (50.0%)0 (0%)0 (0%)1 (25.0%)1 (25.0%)0 (0%)Posaconazole8 (10.3%)12 (15.4%)0 (0%)19 (24.4%)26 (33.3%)13 (16.7%)Voriconazole17 (12.7%)26 (19.4%)0 (0%)32 (23.9%)30 (22.4%)29 (21.6%)Fluconazole12 (4.6%)70 (26.6%)0 (0%)47 (17.9%)75 (28.5%)57 (21.7%)Itraconazole6 (6.1%)12 (12.2%)0 (0%)12 (12.2%)41 (41.8%)27 (27.6%)Caspofungin1 (4.8%)3 (14.3%)0 (0%)6 (28.6%)5 (23.8%)6 (28.6%)Micafungin0 (0%)0 (0%)0 (0%)1 (9.1%)9 (81.8%)1 (9.1%)Amphotericin B7 (16.7%)6 (14.3%)0 (0%)7 (16.7%)9 (21.4%)13 (31.0%)FlucytosineNANANANANANATachyarrhythmiasIsavuconazole2 (22.2%)1 (11.1%)0 (0%)2 (22.2%)4 (44.4%)0 (0%)Posaconazole7 (11.1%)21 (33.3%)0 (0%)15 (23.8%)6 (9.5%)14 (22.2%)Voriconazole28 (19.0%)25 (17.0%)0 (0%)25 (17.0%)26 (17.7%)43 (29.3%)Fluconazole11 (4.0%)80 (28.9%)1 (0.4%)41 (14.8%)44 (15.9%)100 (36.1%)Itraconazole3 (3.5%)3 (3.5%)0 (0%)20 (23.3%)27 (31.4%)33 (38.4%)Caspofungin2 (13.3%)0 (0%)0 (0%)1 (6.7%)4 (26.7%)8 (53.3%)Micafungin1 (5.9%)2 (11.8%)0 (0%)1 (5.9%)10 (58.8%)3 (17.6%)Amphotericin B13 (15.9%)9 (11.0%)0 (0%)7 (8.5%)12 (14.6%)41 (50.0%)Flucytosine1 (33.3%)1 (33.3%)0 (0%)0 (0%)0 (0%)1 (33.3%)Torsade de pointes/QT prolongationIsavuconazole2 (50.0%)0 (0%)0 (0%)1 (25.0%)1 (25.0%)0 (0%)Posaconazole9 (11.4%)19 (24.1%)0 (0%)14 (17.7%)27 (34.2%)10 (12.7%)Voriconazole18 (15.0%)23 (19.2%)0 (0%)22 (18.3%)30 (25.0%)27 (22.5%)Fluconazole14 (4.0%)83 (24.0%)0 (0%)63 (18.2%)84 (24.3%)100 (28.9%)Itraconazole6 (7.7%)8 (10.3%)0 (0%)8 (10.3%)35 (44.9%)21 (26.9%)Caspofungin0 (0%)0 (0%)0 (0%)6 (35.3%)4 (23.5%)7 (41.2%)Micafungin0 (0%)1 (14.3%)0 (0%)0 (0%)5 (71.4%)1 (14.3%)Amphotericin B4 (8.7%)7 (15.2%)0 (0%)5 (10.9%)11 (23.9%)19 (41.3%)Flucytosine0 (0%)1 (50.0%)0 (0%)0 (0%)0 (0%)1 (50.0%) Clinical outcomes of arrhythmogenic toxicity caused by systemic antifungal drugs. Discussion Systemic antifungal agents remain indispensable in the treatment and prevention of invasive fungal diseases; however, their real-world cardiac safety profiles have not been fully elucidated (Tisdale et al., 2020). The FAERS provides a valuable supplementary perspective for identifying “disproportionality signals”, which can be used to detect potential safety concerns warranting further epidemiologic and mechanistic validation; this paradigm has been widely adopted in recent pharmacovigilance studies across multiple therapeutic areas (Buchanan and Li, 2023; Khouri et al., 2021; Tan et al., 2024). This study, based on FAERS data, focused on nine systemic antifungal agents: five triazoles (isavuconazole, posaconazole, voriconazole, fluconazole, itraconazole); two echinocandins (caspofungin, micafungin); one polyene (amphotericin B); and one pyrimidine analogue (flucytosine). Four disproportionality analysis methods were employed concurrently for cross-validation to mitigate false-positive or false-negative results associated with any single algorithm, thereby enhancing the robustness of the findings. To reduce study bias, SMQs were utilized instead of single PTs, and four arrhythmia-related SMQs were evaluated. Overall, the ranking of antifungal agents by the number of SMQs showing positive signals was as follows: Itraconazole (4), Fluconazole (3), Posaconazole (3), Voriconazole (2), Caspofungin (1), Amphotericin B (1), Flucytosine (0), Isavuconazole (0), and Micafungin (0). Comparative analysis revealed that the SMQ for “QT/TdP” exhibited the highest heterogeneity and the most prominent overall signal strength, with RORs ranging from 1.40 to 14.55. Among the triazole antifungals, fluconazole, posaconazole, itraconazole, and voriconazole all exhibited positive signal. In contrast, only isavuconazole showed a negative signal (ROR: 1.40, 95% CI: 0.53–3.74). This pattern aligns in direction with previously reported differences in the pharmacological profiles of triazoles and the consensus on the need for enhanced QT monitoring for certain agents (Cao et al., 2026; Ellsworth and Ostrosky-Zeichner, 2020; Khatib et al., 2021; Tisdale et al., 2020; Yu and Liao, 2022). Although the reason may be related to the least number of reports for isavuconazole (Tian et al., 2025), the most likely reason is its potential advantages, including good oral bioavailability (∼100%), lack of QTc interval prolongation, predictable pharmacokinetics, less complex drug interaction characteristics, and improved tolerability (Lewis et al., 2022; Candel et al., 2025). Among the echinocandins, positive signals were observed for caspofungin (ROR: 3.00, 95% CI: 1.89–4.76) and micafungin (ROR: 2.44, 95% CI: 1.35–4.42). Amphotericin B also exhibited a positive signal (ROR: 3.64, 95% CI: 2.83–4.69). Studies suggest that amphotericin B-associated electrolyte disturbances (e.g., hypokalemia, hypomagnesemia) may increase arrhythmogenic risk by promoting repolarization abnormalities, which is consistent with clinical management recommendations. Although flucytosine, a pyrimidine analog, had a ROR value > 1 (ROR 4.96, 95% CI: 1.24–19.92), it was considered a negative signal due to only two reported cases; this wide confidence interval indicates substantial uncertainty in “signal strength” under sparse counts and warrants cautious interpretation (Khouri et al., 2021). Regarding the “Bradyarrhythmias” SMQ, positive signals were also predominantly concentrated among triazoles. The RORs for posaconazole, fluconazole, itraconazole, and voriconazole were 5.53, 5.53, 5.10, and 2.53, respectively. Isavuconazole again did not show a positive signal (ROR: 0.86, 95% CI: 0.32–2.28). Among non-triazole agents, only caspofungin presented a positive signal (ROR: 2.44, 95% CI: 1.63–3.64). The results are consistent with previously reported adverse reaction cases and database analysis (Uludağ et al., 2013; Panos et al., 2016; Cao et al., 2026). For the SMQ of “Tachyarrhythmias”, the overall signal strength was generally lower than that for QT/TdP. However, several drugs still showed clinically significant disproportionality. The agents presenting positive signals were all triazoles: fluconazole (ROR: 3.53), posaconazole (ROR: 2.59), and itraconazole (ROR: 2.48). Although previous studies have reported that isavuconazole and voriconazole may induce tachycardia, they were not identified as positive signals in the present study may be related to factors such as insufficient report numbers and study bias (Dewan et al., 2017; Keirns et al., 2017). It is noteworthy that the SMQ for “Cardiac arrhythmia terms, nonspecific” demonstrated relatively poor discriminatory power. Most drugs had RORs close to 1 with wide confidence intervals, with only itraconazole showing a positive signal. This result methodologically underscores the importance of SMQ selection; SMQs more closely aligned with specific clinical phenotypes (e.g., QT/TdP or bradyarrhythmias) are more likely to yield interpretable and comparable signals. In contrast, non-specific SMQs may dilute or confound events of different phenotypes, thereby reducing discriminatory power. Figure 4 illustrates the differences among various antifungal agents under four SMQs. It is evident that the ranking and intensity of signals for different SMQs associated with antifungal drugs are not consistent. This suggests that the risk may not stem from a single arrhythmogenic mechanism but rather from the synthesis of multiple factors, including direct electrophysiological effects, drug-drug interactions, inhibition of hERG channel, changes in serum potassium levels, and patient susceptibility, etc (Xie et al., 2025; Syridou et al., 2026; Javandoust et al., 2024). Sensitivity analysis revealed that the arrhythmia-related signals were largely consistent between population receiving and not receiving chemotherapy (Figure 5), and the SMQ for “QT/TdP” was more prominent in the chemotherapy population. This result suggests that in the context of chemotherapy, the increased burden of concomitant medications, electrolyte imbalances, and increased susceptibility to underlying cardiac events may collectively contribute to an increased likelihood of arrhythmia. In addition to signal strength detection, Table 4 summarizes patient outcomes from the reports. For QT/TdP, death and life-threatening outcomes were recurrently reported for multiple drugs and were more prominent among triazoles with high signal strength. Concurrently, a high proportion of “unknown” outcomes indicates that missing outcome documentation remains common in spontaneous reporting systems. For “Tachyarrhythmias”, hospitalization/prolonged hospitalization was more frequently reported than death or life-threatening outcomes. This distribution of outcomes aligns with clinical consensus that drug-related rhythm abnormalities, particularly QT/TdP, can deteriorate rapidly and necessitate monitoring and risk management (Zeppenfeld et al., 2022). Overall, the findings of this study offer several implications for clinical practice. First, fluconazole, posaconazole, and itraconazole showed strong positive signals for both QT/TdP and bradyarrhythmias, suggesting that these agents should be used cautiously when a suitable alternative exists, particularly in populations with pre-existing QT prolongation, electrolyte disturbances, cardiac disease, or a high risk of drug interactions (Lester et al., 2024; Liabeuf et al., 2025; Mehanna and Graaf, 2024; Nakano et al., 2025). Second, isavuconazole consistently showed negative signals across multiple cardiotoxicity-related SMQs in this analysis, indicating it may be a preferable alternative in specific clinical scenarios. This observation aligns with the direction of recent mechanistic and clinical evidence (Ellsworth and Ostrosky-Zeichner, 2020; Ergün et al., 2024; Neofytos et al., 2024). However, it may also be related to the relatively low number of adverse reaction reports, and more real-world evidence is needed to confirm this. Third, while non-triazole agents like amphotericin B and caspofungin exhibited generally milder signals, they are not “zero-risk”. Clinical vigilance remains necessary regarding electrolyte management (especially potassium and magnesium), infusion-related reactions, and concomitant medication risks to avoid the accumulation of factors predisposing to cardiotoxicity (Abdel-Hafez et al., 2022). In addition, FAERS-based analyses have suggested potential sex-specific reporting patterns for voriconazole AEs, supporting demographic-aware surveillance (Xu et al., 2025). This study has several limitations. First, data from FAERS are derived from spontaneous reports, which are susceptible to under-reporting, missing information, and confounding factors such as indication and concomitant medications. Due to the limited number of complete reports, subgroup analyses of relevant influencing factors could not be performed, potentially introducing bias. Second, this study can only demonstrate an association between specific drugs and adverse reactions and cannot directly establish a causal relationship. Therefore, future confirmatory studies and mechanistic studies are warranted to clarify the impact of specific antifungal agents on cardiotoxicity (Boodman and Gupta, 2023). In addition, future work may incorporate standardized QT assessment frameworks (including regulatory guidance and model-informed approaches) to better judge the clinical significance of signals. Conclusion Based on FAERS database, this study employed four disproportionality analysis methods (ROR, PRR, BCPNN, and MGPS) to analyze the association between nine systemic antifungal agents and cardiotoxicity-related signals. Itraconazole, fluconazole, and posaconazole demonstrated stronger cardiotoxicity risks, whereas micafungin, flucytosine, and isavuconazole showed negative signals across all four SMQs. In clinical practice, individual patient risk should be comprehensively assessed to guide personalized drug selection, with reinforcement of electrolyte management, drug interaction review, and electrocardiographic monitoring.