https://orcid.org/0000-0003-0406-5189
Abstract
The explosion of novel anticancer therapies has meant emergence of cardiotoxicity signals including atrial fibrillation (AF). Reliable data concerning the liability of anticancer drugs in inducing AF are scarce. Using the World Health Organization individual case safety report database, VigiBase®, we aimed to determine the association between anticancer drugs and AF.
A disproportionality analysis evaluating the multivariable-adjusted reporting odds ratios for AF with their 99.97% confidence intervals was performed for 176 U.S. Food and Drug Administration (FDA)- or European Medicines Agency (EMA)-labelled anticancer drugs in VigiBase®, followed by a descriptive analysis of AF cases for the anticancer drugs identified in VigiBase®. ClinicalTrial registration number: NCT03530215. A total of 11 757 AF cases associated with at least one anticancer drug were identified in VigiBase® of which 95.8% were deemed serious. Nineteen anticancer drugs were significantly associated with AF of which 14 (74%) are used in haematologic malignancies and 9 (45%) represented new AF associations not previously confirmed in literature including immunomodulating agents (lenalidomide, pomalidomide), several kinase inhibitors (nilotinib, ponatinib, midostaurin), antimetabolites (azacytidine, clofarabine), docetaxel (taxane), and obinutuzumab, an anti-CD20 monoclonal antibody.
Although cancer malignancy itself may generate AF, we identified 19 anticancer drugs significantly associated with a significant increase in AF over-reporting. This pharmacovigilance study provides evidence that anticancer drugs themselves could represent independent risk factors for AF development. Dedicated prospective clinical trials are now required to confirm these 19 associations. This list of suspected anticancer drugs should be known by physicians when confronted to AF in cancer patients, particularly in case of haematologic malignancies.
Introduction
Therapeutic innovations have led new treatment options and improved survival of patients with cancer, but these novel therapies are also associated with short- and long-term adverse effects (AEs). The emergence of associated cardiovascular diseases is increasingly being recognized as an important issue of anticancer drugs.1,2 While older therapies, such as anthracyclines and trastuzumab, were associated with cardiomyopathy (specifically systolic left ventricular dysfunction), there has been a growing appreciation of other cardiac, vascular, and metabolic perturbations as a result of anticancer therapy. In particular, arrhythmias [including atrial fibrillation (AF)] has been a new risk associated with anticancer therapies.3–5
AF is the most common sustained arrhythmia, currently affecting over 33 million individuals worldwide and is predicted to affect 6–12 million people in the USA by 2050 and 17.9 million in Europe by 2060.6,7 Patients with cancer exhibited a 20% higher adjusted risk of AF compared to those without cancer.8 In cancer patients, new onset AF is associated with a two-fold higher risk of thromboembolism complications and a six-fold higher risk of heart failure as well as a 10-fold higher risk of 30-day mortality, even after adjusting for known risk factors.9
The explosion of new anticancer drugs has meant that new therapies may be associated with AF. However, this may not be easy to identify due to competing factors. For example, the diagnosis of cancer alone can predispose patients to AF.9 In surgically managed lung cancer patients, the incidence of post-operative AF after pulmonary resection has been estimated to be as high as 30%.10 The propensity to develop AF is increased by the stresses induced by malignancy and may be related to prior organ injury, immune reaction, systemic inflammation, electrolyte or endocrine abnormalities, impaired oxygenation, modulation of molecular pathways, or metabolic alterations.11,12 Anticancer drugs can influence many of these factors, they have intrinsic toxicity and they may, therefore, contribute to proarrhythmic events.3 Data regarding anticancer drug-associated AF are scarce and are essentially based on non-randomized clinical trials, highlighting the urgent need of ‘real world’ data.4,11 The identification of clinical features of anticancer drugs-related AF is challenging since most clinical trials are underpowered to characterize AF in the cancer population. In general, patients with cardiovascular risk factors or those having pre-existing cardiac diseases (i.e. highest risk of AF) are often excluded from clinical trials. Moreover, monitoring of cardiac arrhythmias is minimal, and only severe AF cases needing immediate medical attention are identified.13 Therefore, defining anticancer drugs associated with AF occurrence remains critical.
The aim of this study was to identify the anticancer drugs associated with AF in Vigibase®, the World Health Organization (WHO) pharmacovigilance database. Other important objectives were to characterize the main clinical features of AF associated with anticancer drugs and to discuss the potential underlying molecular pathways.
Methods
Data source
VigiBase®, the WHO global database of individual case safety reports was used for this study.14,15 VigiBase® was already described elsewhere16,17 and details are shown in the Supplementary material online, Data S1. In this study, we used the Vigibase® extract case level database to allow for multiple adjustments on concomitant medications, reactions, and demographic parameters (details on the case level data structure are shown in the Supplementary material online, Data S1).
Disproportionality analysis in VigiBase®
To identify anticancer drugs associated with AF in VigiBase®, we performed an observational, retrospective, pharmacovigilance study in the WHO database using a disproportionality analysis.16 The study protocol was registered on ClinicalTrials.gov, NCT03530215. Cases were reports with AF and were collected using the ‘preferred term’ (from the System Organ Class MedDRA, v21.1): « atrial fibrillation » from the first report of each anticancer drug until 1 March 2019. Non-cases were all the reports without AF.
The study population was restricted to reports involving at least one liable anticancer drug (hereafter the primary analysis population, see Figure 1). This strategy is common in disproportionality analysis to limit the indication bias because all patients will have received an anticancer drug and hence be likely to have been diagnosed with a cancer.18 The comparability between cases and non-cases is improved with this approach. We examined all anticancer drugs labelled by the US Food and Drug Administration or by the European Medicine Agency prior to January 2018. A total of 176 anticancer drugs were identified (detailed in Supplementary material online, Table S1). We studied the effect of each anticancer drug on the reporting of AF. Another definition of the anticancer drug population using the Anatomical and Therapeutic Classification (ATC) L-class ‘Antineoplastic and Immunomodulating agents’ was used in a sensitivity analysis (L-class population). The full database population, with no restriction on reports, was also used in a sensitivity analysis (Figure 1). Disproportionality estimates, i.e. the reporting odds ratio (ROR) and information component (IC) of AF were computed for each of the 176 anticancer drugs.17,19
Flowchart of cases selection in VigiBase® and algorithm of the MEDLINE search of English review articles published in the last 5 years conducted on 14 January 2019 to identify atrial fibrillation molecular pathways.
Description of the pharmacovigilance cohort in VigiBase®
We described AF cases associated with liable anticancer drugs in Vigibase®. For each report, age, sex, drug indication, time from anticancer drug initiation to AF occurrence (hereafter, time to onset), outcome, concurrent medications, and AEs were collected (see Supplementary material online, Data S1 for details). Concurrent medications of interest were anticoagulants, cardiovascular, or diabetes with a focus on drugs with Class I or III antiarrhythmic properties and digoxin which are more specifically prescribed in patients with an AF history, and systemic corticosteroids. Concurrent diseases or conditions of interest were heart failure, ischaemic stroke, myocardial infarction, haemorrhages, sepsis, colitis, diarrhoea, dehydration, hypokalaemia, and hyperthyroidism.
Identification of the molecular pathways involved in anticancer drug-associated atrial fibrillation
First, a MEDLINE search of review articles published in English in the last 5 years was independently conducted on 14 January 2019 by two reviewers (J.A. and C.D.) according to prespecified selection criteria. The search algorithm is presented in Figure 1. The domains used to perform the MEDLINE search were Domain 1 terms related to AF molecular pathways (CaM kinase II, RYR2/SERCA2a, mitochondrial dysfunction, reactive oxygen species (ROS) production, PI3K/Akt signalling, and MAPK signalling); Domain 2 terms related to anticancer drugs (each anticancer drug with a significant ROR). Finally, the research was performed as follows: (Domain 1) AND (Domain 2). We only retained review articles in English without any temporal restriction. Two authors (J.A. and C.D.) independently screened study titles and abstracts identified by the search against the eligibility criteria. Full reports were obtained for all eligible articles/abstracts. The authors independently extracted data from the selected studies in duplicate using a standardized data extraction form.
Statistical analyses
VigiBase® allows for disproportionality analysis (also known as case–non-case analysis),16,17 a method used to assess whether suspected drug-associated AF is reported differentially among anticancer drugs. Disproportionality analysis was already described elsewhere17 and details are shown in the Supplementary material online, Data S1. The ROR estimate has been shown to be most reliable in the presence of at least 3–5 cases for the AE of interest (herein, AF).20 Therefore, we computed the ROR for anticancer drugs including at least five AF cases. In univariable analysis, the time window period was restricted to the period of notification of the anticancer drug group. Then, RORs were adjusted (aRORs) on age (categorized into the following categories: ‘<45 years’, ‘45–64 years’, ‘65–74 years’, and ‘≥75 years’); sex; region (Europa, Asia, America, Oceania, or Africa); the concomitant reporting of cardiovascular or diabetes medications, systemic corticosteroids the concurrent reporting of heart failure, ischaemic stroke, myocardial infarction, sepsis, colitis, diarrhoea, dehydration, hypokalaemia, and hyperthyroidism for each of the 176 anticancer drugs. Because 176 anticancer drugs were simultaneously studied, a conservative Bonferroni adjustment for multiple tests was applied to prevent alpha risk inflation. Hence, a P-value <0.0003 was deemed significant, i.e. comparisons were considered significant when the lower boundary of the 99.97% confidence interval (CI) of the (a)ROR was ≥1. The aRORs (99.97% CI) were computed in the database including only cases with at least one concomitant, interacting, or suspected anticancer drug (primary analysis population; the 176 anticancer drugs are presented in the Supplementary material online, Table S1). In addition, five sensitivity analyses were performed as follows: (i) adjusted ROR excluding cases reporting anticoagulant and/or anti-arrhythmic drugs; (ii) unadjusted ROR in the primary analysis population; (iii) unadjusted ROR in the ATC L-class restricted population; (iv) unadjusted ROR in the full database population; and (v) calculation of the IC (details in Supplementary material online, Data S1).17 Also, disproportionality was used to exploratory assess the association between the underlying cancer and AF, independently of anticancer drugs. χ2 tests (with a P-value <0.05 considered significant) were used to compare the clinical characteristics of AF cases associated with the identified anticancer drugs with those of AF cases that were not associated with anticancer drugs in VigiBase®.
Results
Comparative study in VigiBase®
A total of 2 124 646 reports including at least one of the 176 anticancer drugs were identified (primary analysis population), of which 11 757 (0.55%) were AF cases (Figure 1). In univariate analysis of the primary analysis population, there were 31 anticancer drugs associated with AF. After adjustment on age, sex, region, concurrent cardiovascular or diabetes medications, systemic corticosteroids use, concurrent heart failure, myocardial infarction, stroke, sepsis, hyperthyroidism, colitis, diarrhoea, dehydration, hypokalaemia, and each of the 176 anticancer, 19 anticancer drugs were still associated with AF (10.8% of 176, Figure 2). The strongest association with AF was found for ibrutinib (aROR 8.99; 99.97% CI 7.67–10.52; 1431 AF cases), followed by aldesleukin (aROR 5.01; 99.97% CI 2.49–10.09; 41 AF cases) and nilotinib (aROR 3.92; 99.97% CI 2.85–5.40; 241 AF cases). The detailed results of all the disproportionality analyses, including ROR calculation in several settings (full database population, anticancer drug restricted population, adjusted and unadjusted ROR) and the IC are shown in Supplementary material online, Table S1. Sensitivity analyses are summarized in Supplementary material online, Table S2. Exploratory analysis according to cancer indication (available in 932 234, 43.9%) suggested an association between AF and haematologic malignancies (leukaemia, lymphoma, and myeloma), oesophageal cancer and skin cancer (Supplementary material online, Table S3).
Significant associations between atrial fibrillation cases and anticancer drugs, concurrent medications and reactions, and demographic characteristics of patients in VigiBase® in the primary analysis population (anticancer drug population). Reporting odds ratio were adjusted for age, sex, region, concurrent cardiovascular or diabetes medications, systemic corticosteroids use, concurrent heart failure, myocardial infarction, stroke, sepsis, hyperthyroidism, colitis, diarrhoea, dehydration, hypokalaemia, and each of the 176 anticancer drugs. Data in VigiBase® were extracted on 1 March 2019. A total of 11 757 atrial fibrillation cases, and a total of 2 124 646 anticancer adverse drug reaction reports in the overall database were found at that time. Circle sizes are log-proportional to the number of reports for each factor. AF, atrial fibrillation; CI, confidence interval.
Descriptive cohort in Vigibase®
The 19 anticancer drugs significantly associated with AF in VigiBase® accounted for a total of 6147 AF cases (1.03% of all reported AEs with the 19 anticancer drugs, ranging from 0.38% for docetaxel to 6.4% for ibrutinib). Two-by-two co-prescription rates between these drugs in AF cases are shown in Supplementary material online, Figure S1 and the type of cancer associated with these 19 anticancer drugs are shown in the Supplementary material online, Table S4. In AF cases, the top reported anticancer drugs were lenalidomide (n: 1733), ibrutinib (n: 1431), and rituximab (n: 706). Table 1 shows the clinical characteristics of AF cases associated with each of the 19 anticancer drugs. Time to AF onset from treatment initiation for the 19 anticancer drugs was <3 months in 1173 (66.1% of 1774 available) and <6 months in 1373 (77.4% of 1774 available). Anthracyclines (daunorubicin, idarubicin), midostaurin, and aldesleukin exhibited lower rates of co-reported cardiovascular, diabetes, and anticoagulant medications. Ipilumumab was more frequently co-reported with hyperthyroidism and clofarabine was associated with a high all-cause mortality rate (63.5%). Supplementary material online, Table S5 shows the comparison between AF cases associated with the 19 identified anticancer drugs and AF cases that were not associated with the 19 identified anticancer drugs.
Characteristics of AF cases with the 19 anticancer drugs associated with AF overreporting in Vigibase®
| Alkylating agents | Androgen deprivation therapy | Antimetabolites | Anthracyclines | Kinase inhibitors | Immune checkpoint inhibitor | Immunomodulating agents | Monoclonal antibodies (anti-CD20) | Protea-some inhibitor | Taxane | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Bruton Tyrosin Kinase inhibitor | BCR-ABL inhibitors | Multi-kinase inhibitor | |||||||||||||||||
| Anticancer drug | Cisplatin | Dacarbazine | Abiraterone | Azacitidine | Clofarabine | Daunorubicin | Idarubicin | Ibrutinib | Nilotinib | Ponatinib | Midostaurin | Ipilimumab | Aldesleukin | Lenalidomide | Pomalidomide | Obinutuzumab | Rituximab | Bortezomib | Docetaxel |
| No | 404 | 29 | 222 | 109 | 52 | 124 | 57 | 1431 | 241 | 72 | 20 | 104 | 41 | 1733 | 308 | 53 | 706 | 547 | 395 |
| Male | 282 (73.8%) [382] | 19 (65.5%) [29] | 216 (100%) [216] | 57 (59.4%) [96] | 32 (65.3%) [49] | 66 (54.1%) [122] | 33 (60%) [55] | 899 (67.4%) [1334] | 131 (58%) [226] | 32 (62.7%) [51] | 12 (60%) [20] | 83 (79.8%) [104] | 29 (78.4%) [37] | 979 (58.2%) [1682] | 176 (58.3%) [302] | 31 (63.3%) [49] | 326 (50.9%) [640] | 318 (61.7%) [515] | 216 (55.2%) [391] |
| Age >65 years | 185 (61.5%) [301] | 20 (76.9%) [26] | 156 (94.5%) [165] | 74 (89.2%) [83] | 32 (72.7%) [44] | 49 (64.5%) [76] | 26 (49.1%) [53] | 885 (84.4%) [1049] | 102 (64.2%) [159] | 29 (69%) [42] | 14 (73.7%) [19] | 63 (66.3%) [95] | 14 (45.2%) [31] | 960 (82.7%) [1161] | 187 (82.7%) [226] | 32 (76.2%) [42] | 366 (74.5%) [491] | 312 (74.3%) [420] | 190 (60.9%) [312] |
| Time to onset, days | 16 (6–49) [126] | 12 (11–28) [5] | 69 (26–176) [72] | 59 (10–218) [75] | 12 (9–26) [8] | 18 (5–31) [18] | 14 (6–35) [13] | 110 (42–294) [335] | 148 (48–440) [102] | 129 (40–227) [23] | 18 (12–114) [13] | 44 (14–60) [15] | 5 (3–13) [15] | 66 (20–304) [375] | 42 (15–90) [76] | 39 (29–56) [9] | 33 (9–102) [157] | 29 (13–74) [137] | 16 (7–42) [200] |
| Concurrent medications [availability is equal to the number of AF cases] | |||||||||||||||||||
| Cardiovascular or diabetes medication | 208 (51.5%) | 16 (55.2%) | 113 (50.9%) | 50 (45.9%) | 24 (46.2%) | 31 (25%) | 32 (56.1%) | 571 (39.9%) | 106 (44%) | 40 (55.6%) | 6 (30%) | 46 (44.2%) | 11 (26.8%) | 1117 (64.5%) | 205 (66.6%) | 15 (28.3%) | 346 (49%) | 393 (71.8%) | 192 (48.6%) |
| Anti-coagulant | 44 (10.9%) | 5 (17.2%) | 33 (14.9%) | 12 (11%) | 5 (9.6%) | 1 (0.8%) | 3 (5.3%) | 239 (16.7%) | 43 (17.8%) | 11 (15.3%) | 3 (15%) | 23 (22.1%) | 1 (2.4%) | 446 (25.7%) | 84 (27.3%) | 10 (18.9%) | 105 (14.9%) | 111 (20.3%) | 58 (14.7%) |
| AAR | 36 (8.9%) | 4 (13.8%) | 21 (9.5%) | 10 (9.2%) | 8 (15.4%) | 5 (4%) | 4 (7%) | 119 (8.3%) | 23 (9.5%) | 8 (11.1%) | 0 (0%) | 10 (9.6%) | 2 (4.9%) | 272 (15.7%) | 41 (13.3%) | 6 (11.3%) | 53 (7.5%) | 64 (11.7%) | 43 (10.9%) |
| Systemic cortico-steroids | 108 (26.7%) | 8 (27.6%) | 129 (58.1%) | 20 (18.3%) | 14 (26.9%) | 25 (20.2%) | 16 (28.1%) | 102 (7.1%) | 8 (3.3%) | 15 (20.8%) | 1 (5%) | 33 (31.7%) | 2 (4.9%) | 825 (47.6%) | 168 (54.5%) | 7 (13.2%) | 353 (50%) | 370 (67.6%) | 128 (32.4%) |
| Concurrent adverse events [availability is equal to the number of AF cases] | |||||||||||||||||||
| Heart failure | 42 (10.4%) | 4 (13.8%) | 38 (17.1%) | 16 (14.7%) | 14 (26.9%) | 28 (22.6%) | 19 (33.3%) | 153 (10.7%) | 30 (12.4%) | 17 (23.6%) | 3 (15%) | 9 (8.7%) | 5 (12.2%) | 201 (11.6%) | 37 (12%) | 4 (7.5%) | 98 (13.9%) | 104 (19%) | 55 (13.9%) |
| Ischaemic stroke | 15 (3.7%) | 0 (0%) | 7 (3.2%) | 6 (5.5%) | 1 (1.9%) | 4 (3.2%) | 0 (0%) | 53 (3.7%) | 22 (9.1%) | 5 (6.9%) | 0 (0%) | 2 (1.9%) | 4 (9.8%) | 70 (4%) | 8 (2.6%) | 1 (1.9%) | 25 (3.5%) | 14 (2.6%) | 18 (4.6%) |
| Myocardial infarction | 13 (3.2%) | 3 (10.3%) | 10 (4.5%) | 5 (4.6%) | 3 (5.8%) | 8 (6.5%) | 3 (5.3%) | 36 (2.5%) | 18 (7.5%) | 4 (5.6%) | 1 (5%) | 3 (2.9%) | 4 (9.8%) | 70 (4%) | 6 (1.9%) | 1 (1.9%) | 26 (3.7%) | 30 (5.5%) | 14 (3.5%) |
| Haemorrhage | 36 (8.9%) | 2 (6.9%) | 16 (7.2%) | 19 (17.4%) | 14 (26.9%) | 22 (17.7%) | 14 (24.6%) | 298 (20.8%) | 23 (9.5%) | 10 (13.9%) | 3 (15%) | 10 (9.6%) | 2 (4.9%) | 107 (6.2%) | 8 (2.6%) | 4 (7.5%) | 84 (11.9%) | 54 (9.9%) | 52 (13.2%) |
| Sepsis | 42 (10.4%) | 6 (20.7%) | 7 (3.2%) | 11 (10.1%) | 16 (30.8%) | 26 (21%) | 8 (14%) | 35 (2.4%) | 2 (0.8%) | 7 (9.7%) | 5 (25%) | 5 (4.8%) | 1 (2.4%) | 111 (6.4%) | 17 (5.5%) | 4 (7.5%) | 59 (8.4%) | 50 (9.1%) | 39 (9.9%) |
| Colitis | 11 (2.7%) | 1 (3.4%) | 1 (0.5%) | 2 (1.8%) | 3 (5.8%) | 3 (2.4%) | 2 (3.5%) | 12 (0.8%) | 1 (0.4%) | 0 (0%) | 1 (5%) | 19 (18.3%) | 0 (0%) | 10 (0.6%) | 1 (0.3%) | 2 (3.8%) | 3 (0.4%) | 7 (1.3%) | 6 (1.5%) |
| Diarrhoea | 42 (10.4%) | 5 (17.2%) | 7 (3.2%) | 9 (8.3%) | 8 (15.4%) | 9 (7.3%) | 15 (26.3%) | 112 (7.8%) | 9 (3.7%) | 7 (9.7%) | 0 (0%) | 20 (19.2%) | 4 (9.8%) | 98 (5.7%) | 11 (3.6%) | 8 (15.1%) | 34 (4.8%) | 51 (9.3%) | 48 (12.2%) |
| Dehydration | 69 (17.1%) | 4 (13.8%) | 6 (2.7%) | 3 (2.8%) | 2 (3.8%) | 4 (3.2%) | 0 (0%) | 20 (1.4%) | 2 (0.8%) | 1 (1.4%) | 0 (0%) | 15 (14.4%) | 0 (0%) | 72 (4.2%) | 6 (1.9%) | 0 (0%) | 20 (2.8%) | 26 (4.8%) | 55 (13.9%) |
| Hypokalaemia | 20 (5%) | 2 (6.9%) | 15 (6.8%) | 2 (1.8%) | 2 (3.8%) | 4 (3.2%) | 3 (5.3%) | 13 (0.9%) | 3 (1.2%) | 0 (0%) | 1 (5%) | 4 (3.8%) | 0 (0%) | 45 (2.6%) | 1 (0.3%) | 0 (0%) | 10 (1.4%) | 26 (4.8%) | 15 (3.8%) |
| Hyperthyroidism | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 10 (4.1%) | 1 (1.4%) | 0 (0%) | 14 (13.5%) | 1 (2.4%) | 8 (0.5%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 2 (0.4%) | 0 (0%) |
| All-cause death | 66 (21.2%) [311] | 4 (17.4%) [23] | 18 (8.7%) [206] | 42 (39.3%) [107] | 33 (63.5%) [52] | 38 (35.8%) [106] | 22 (43.1%) [51] | 130 (9.2%) [1412] | 15 (6.3%) [237] | 16 (22.2%) [72] | 8 (40%) [20] | 23 (22.3%) [103] | 5 (20.8%) [24] | 279 (16.1%) [1730] | 26 (8.4%) [308] | 7 (13.2%) [53] | 114 (16.8%) [679] | 95 (17.6%) [540] | 69 (19.7%) [351] |
| Alkylating agents | Androgen deprivation therapy | Antimetabolites | Anthracyclines | Kinase inhibitors | Immune checkpoint inhibitor | Immunomodulating agents | Monoclonal antibodies (anti-CD20) | Protea-some inhibitor | Taxane | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Bruton Tyrosin Kinase inhibitor | BCR-ABL inhibitors | Multi-kinase inhibitor | |||||||||||||||||
| Anticancer drug | Cisplatin | Dacarbazine | Abiraterone | Azacitidine | Clofarabine | Daunorubicin | Idarubicin | Ibrutinib | Nilotinib | Ponatinib | Midostaurin | Ipilimumab | Aldesleukin | Lenalidomide | Pomalidomide | Obinutuzumab | Rituximab | Bortezomib | Docetaxel |
| No | 404 | 29 | 222 | 109 | 52 | 124 | 57 | 1431 | 241 | 72 | 20 | 104 | 41 | 1733 | 308 | 53 | 706 | 547 | 395 |
| Male | 282 (73.8%) [382] | 19 (65.5%) [29] | 216 (100%) [216] | 57 (59.4%) [96] | 32 (65.3%) [49] | 66 (54.1%) [122] | 33 (60%) [55] | 899 (67.4%) [1334] | 131 (58%) [226] | 32 (62.7%) [51] | 12 (60%) [20] | 83 (79.8%) [104] | 29 (78.4%) [37] | 979 (58.2%) [1682] | 176 (58.3%) [302] | 31 (63.3%) [49] | 326 (50.9%) [640] | 318 (61.7%) [515] | 216 (55.2%) [391] |
| Age >65 years | 185 (61.5%) [301] | 20 (76.9%) [26] | 156 (94.5%) [165] | 74 (89.2%) [83] | 32 (72.7%) [44] | 49 (64.5%) [76] | 26 (49.1%) [53] | 885 (84.4%) [1049] | 102 (64.2%) [159] | 29 (69%) [42] | 14 (73.7%) [19] | 63 (66.3%) [95] | 14 (45.2%) [31] | 960 (82.7%) [1161] | 187 (82.7%) [226] | 32 (76.2%) [42] | 366 (74.5%) [491] | 312 (74.3%) [420] | 190 (60.9%) [312] |
| Time to onset, days | 16 (6–49) [126] | 12 (11–28) [5] | 69 (26–176) [72] | 59 (10–218) [75] | 12 (9–26) [8] | 18 (5–31) [18] | 14 (6–35) [13] | 110 (42–294) [335] | 148 (48–440) [102] | 129 (40–227) [23] | 18 (12–114) [13] | 44 (14–60) [15] | 5 (3–13) [15] | 66 (20–304) [375] | 42 (15–90) [76] | 39 (29–56) [9] | 33 (9–102) [157] | 29 (13–74) [137] | 16 (7–42) [200] |
| Concurrent medications [availability is equal to the number of AF cases] | |||||||||||||||||||
| Cardiovascular or diabetes medication | 208 (51.5%) | 16 (55.2%) | 113 (50.9%) | 50 (45.9%) | 24 (46.2%) | 31 (25%) | 32 (56.1%) | 571 (39.9%) | 106 (44%) | 40 (55.6%) | 6 (30%) | 46 (44.2%) | 11 (26.8%) | 1117 (64.5%) | 205 (66.6%) | 15 (28.3%) | 346 (49%) | 393 (71.8%) | 192 (48.6%) |
| Anti-coagulant | 44 (10.9%) | 5 (17.2%) | 33 (14.9%) | 12 (11%) | 5 (9.6%) | 1 (0.8%) | 3 (5.3%) | 239 (16.7%) | 43 (17.8%) | 11 (15.3%) | 3 (15%) | 23 (22.1%) | 1 (2.4%) | 446 (25.7%) | 84 (27.3%) | 10 (18.9%) | 105 (14.9%) | 111 (20.3%) | 58 (14.7%) |
| AAR | 36 (8.9%) | 4 (13.8%) | 21 (9.5%) | 10 (9.2%) | 8 (15.4%) | 5 (4%) | 4 (7%) | 119 (8.3%) | 23 (9.5%) | 8 (11.1%) | 0 (0%) | 10 (9.6%) | 2 (4.9%) | 272 (15.7%) | 41 (13.3%) | 6 (11.3%) | 53 (7.5%) | 64 (11.7%) | 43 (10.9%) |
| Systemic cortico-steroids | 108 (26.7%) | 8 (27.6%) | 129 (58.1%) | 20 (18.3%) | 14 (26.9%) | 25 (20.2%) | 16 (28.1%) | 102 (7.1%) | 8 (3.3%) | 15 (20.8%) | 1 (5%) | 33 (31.7%) | 2 (4.9%) | 825 (47.6%) | 168 (54.5%) | 7 (13.2%) | 353 (50%) | 370 (67.6%) | 128 (32.4%) |
| Concurrent adverse events [availability is equal to the number of AF cases] | |||||||||||||||||||
| Heart failure | 42 (10.4%) | 4 (13.8%) | 38 (17.1%) | 16 (14.7%) | 14 (26.9%) | 28 (22.6%) | 19 (33.3%) | 153 (10.7%) | 30 (12.4%) | 17 (23.6%) | 3 (15%) | 9 (8.7%) | 5 (12.2%) | 201 (11.6%) | 37 (12%) | 4 (7.5%) | 98 (13.9%) | 104 (19%) | 55 (13.9%) |
| Ischaemic stroke | 15 (3.7%) | 0 (0%) | 7 (3.2%) | 6 (5.5%) | 1 (1.9%) | 4 (3.2%) | 0 (0%) | 53 (3.7%) | 22 (9.1%) | 5 (6.9%) | 0 (0%) | 2 (1.9%) | 4 (9.8%) | 70 (4%) | 8 (2.6%) | 1 (1.9%) | 25 (3.5%) | 14 (2.6%) | 18 (4.6%) |
| Myocardial infarction | 13 (3.2%) | 3 (10.3%) | 10 (4.5%) | 5 (4.6%) | 3 (5.8%) | 8 (6.5%) | 3 (5.3%) | 36 (2.5%) | 18 (7.5%) | 4 (5.6%) | 1 (5%) | 3 (2.9%) | 4 (9.8%) | 70 (4%) | 6 (1.9%) | 1 (1.9%) | 26 (3.7%) | 30 (5.5%) | 14 (3.5%) |
| Haemorrhage | 36 (8.9%) | 2 (6.9%) | 16 (7.2%) | 19 (17.4%) | 14 (26.9%) | 22 (17.7%) | 14 (24.6%) | 298 (20.8%) | 23 (9.5%) | 10 (13.9%) | 3 (15%) | 10 (9.6%) | 2 (4.9%) | 107 (6.2%) | 8 (2.6%) | 4 (7.5%) | 84 (11.9%) | 54 (9.9%) | 52 (13.2%) |
| Sepsis | 42 (10.4%) | 6 (20.7%) | 7 (3.2%) | 11 (10.1%) | 16 (30.8%) | 26 (21%) | 8 (14%) | 35 (2.4%) | 2 (0.8%) | 7 (9.7%) | 5 (25%) | 5 (4.8%) | 1 (2.4%) | 111 (6.4%) | 17 (5.5%) | 4 (7.5%) | 59 (8.4%) | 50 (9.1%) | 39 (9.9%) |
| Colitis | 11 (2.7%) | 1 (3.4%) | 1 (0.5%) | 2 (1.8%) | 3 (5.8%) | 3 (2.4%) | 2 (3.5%) | 12 (0.8%) | 1 (0.4%) | 0 (0%) | 1 (5%) | 19 (18.3%) | 0 (0%) | 10 (0.6%) | 1 (0.3%) | 2 (3.8%) | 3 (0.4%) | 7 (1.3%) | 6 (1.5%) |
| Diarrhoea | 42 (10.4%) | 5 (17.2%) | 7 (3.2%) | 9 (8.3%) | 8 (15.4%) | 9 (7.3%) | 15 (26.3%) | 112 (7.8%) | 9 (3.7%) | 7 (9.7%) | 0 (0%) | 20 (19.2%) | 4 (9.8%) | 98 (5.7%) | 11 (3.6%) | 8 (15.1%) | 34 (4.8%) | 51 (9.3%) | 48 (12.2%) |
| Dehydration | 69 (17.1%) | 4 (13.8%) | 6 (2.7%) | 3 (2.8%) | 2 (3.8%) | 4 (3.2%) | 0 (0%) | 20 (1.4%) | 2 (0.8%) | 1 (1.4%) | 0 (0%) | 15 (14.4%) | 0 (0%) | 72 (4.2%) | 6 (1.9%) | 0 (0%) | 20 (2.8%) | 26 (4.8%) | 55 (13.9%) |
| Hypokalaemia | 20 (5%) | 2 (6.9%) | 15 (6.8%) | 2 (1.8%) | 2 (3.8%) | 4 (3.2%) | 3 (5.3%) | 13 (0.9%) | 3 (1.2%) | 0 (0%) | 1 (5%) | 4 (3.8%) | 0 (0%) | 45 (2.6%) | 1 (0.3%) | 0 (0%) | 10 (1.4%) | 26 (4.8%) | 15 (3.8%) |
| Hyperthyroidism | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 10 (4.1%) | 1 (1.4%) | 0 (0%) | 14 (13.5%) | 1 (2.4%) | 8 (0.5%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 2 (0.4%) | 0 (0%) |
| All-cause death | 66 (21.2%) [311] | 4 (17.4%) [23] | 18 (8.7%) [206] | 42 (39.3%) [107] | 33 (63.5%) [52] | 38 (35.8%) [106] | 22 (43.1%) [51] | 130 (9.2%) [1412] | 15 (6.3%) [237] | 16 (22.2%) [72] | 8 (40%) [20] | 23 (22.3%) [103] | 5 (20.8%) [24] | 279 (16.1%) [1730] | 26 (8.4%) [308] | 7 (13.2%) [53] | 114 (16.8%) [679] | 95 (17.6%) [540] | 69 (19.7%) [351] |
Data are expressed as n (%) [availability] or the median (interquartile range) [availability].
AAR: Class I or III antiarrhythmics or digoxin.
Characteristics of AF cases with the 19 anticancer drugs associated with AF overreporting in Vigibase®
| Alkylating agents | Androgen deprivation therapy | Antimetabolites | Anthracyclines | Kinase inhibitors | Immune checkpoint inhibitor | Immunomodulating agents | Monoclonal antibodies (anti-CD20) | Protea-some inhibitor | Taxane | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Bruton Tyrosin Kinase inhibitor | BCR-ABL inhibitors | Multi-kinase inhibitor | |||||||||||||||||
| Anticancer drug | Cisplatin | Dacarbazine | Abiraterone | Azacitidine | Clofarabine | Daunorubicin | Idarubicin | Ibrutinib | Nilotinib | Ponatinib | Midostaurin | Ipilimumab | Aldesleukin | Lenalidomide | Pomalidomide | Obinutuzumab | Rituximab | Bortezomib | Docetaxel |
| No | 404 | 29 | 222 | 109 | 52 | 124 | 57 | 1431 | 241 | 72 | 20 | 104 | 41 | 1733 | 308 | 53 | 706 | 547 | 395 |
| Male | 282 (73.8%) [382] | 19 (65.5%) [29] | 216 (100%) [216] | 57 (59.4%) [96] | 32 (65.3%) [49] | 66 (54.1%) [122] | 33 (60%) [55] | 899 (67.4%) [1334] | 131 (58%) [226] | 32 (62.7%) [51] | 12 (60%) [20] | 83 (79.8%) [104] | 29 (78.4%) [37] | 979 (58.2%) [1682] | 176 (58.3%) [302] | 31 (63.3%) [49] | 326 (50.9%) [640] | 318 (61.7%) [515] | 216 (55.2%) [391] |
| Age >65 years | 185 (61.5%) [301] | 20 (76.9%) [26] | 156 (94.5%) [165] | 74 (89.2%) [83] | 32 (72.7%) [44] | 49 (64.5%) [76] | 26 (49.1%) [53] | 885 (84.4%) [1049] | 102 (64.2%) [159] | 29 (69%) [42] | 14 (73.7%) [19] | 63 (66.3%) [95] | 14 (45.2%) [31] | 960 (82.7%) [1161] | 187 (82.7%) [226] | 32 (76.2%) [42] | 366 (74.5%) [491] | 312 (74.3%) [420] | 190 (60.9%) [312] |
| Time to onset, days | 16 (6–49) [126] | 12 (11–28) [5] | 69 (26–176) [72] | 59 (10–218) [75] | 12 (9–26) [8] | 18 (5–31) [18] | 14 (6–35) [13] | 110 (42–294) [335] | 148 (48–440) [102] | 129 (40–227) [23] | 18 (12–114) [13] | 44 (14–60) [15] | 5 (3–13) [15] | 66 (20–304) [375] | 42 (15–90) [76] | 39 (29–56) [9] | 33 (9–102) [157] | 29 (13–74) [137] | 16 (7–42) [200] |
| Concurrent medications [availability is equal to the number of AF cases] | |||||||||||||||||||
| Cardiovascular or diabetes medication | 208 (51.5%) | 16 (55.2%) | 113 (50.9%) | 50 (45.9%) | 24 (46.2%) | 31 (25%) | 32 (56.1%) | 571 (39.9%) | 106 (44%) | 40 (55.6%) | 6 (30%) | 46 (44.2%) | 11 (26.8%) | 1117 (64.5%) | 205 (66.6%) | 15 (28.3%) | 346 (49%) | 393 (71.8%) | 192 (48.6%) |
| Anti-coagulant | 44 (10.9%) | 5 (17.2%) | 33 (14.9%) | 12 (11%) | 5 (9.6%) | 1 (0.8%) | 3 (5.3%) | 239 (16.7%) | 43 (17.8%) | 11 (15.3%) | 3 (15%) | 23 (22.1%) | 1 (2.4%) | 446 (25.7%) | 84 (27.3%) | 10 (18.9%) | 105 (14.9%) | 111 (20.3%) | 58 (14.7%) |
| AAR | 36 (8.9%) | 4 (13.8%) | 21 (9.5%) | 10 (9.2%) | 8 (15.4%) | 5 (4%) | 4 (7%) | 119 (8.3%) | 23 (9.5%) | 8 (11.1%) | 0 (0%) | 10 (9.6%) | 2 (4.9%) | 272 (15.7%) | 41 (13.3%) | 6 (11.3%) | 53 (7.5%) | 64 (11.7%) | 43 (10.9%) |
| Systemic cortico-steroids | 108 (26.7%) | 8 (27.6%) | 129 (58.1%) | 20 (18.3%) | 14 (26.9%) | 25 (20.2%) | 16 (28.1%) | 102 (7.1%) | 8 (3.3%) | 15 (20.8%) | 1 (5%) | 33 (31.7%) | 2 (4.9%) | 825 (47.6%) | 168 (54.5%) | 7 (13.2%) | 353 (50%) | 370 (67.6%) | 128 (32.4%) |
| Concurrent adverse events [availability is equal to the number of AF cases] | |||||||||||||||||||
| Heart failure | 42 (10.4%) | 4 (13.8%) | 38 (17.1%) | 16 (14.7%) | 14 (26.9%) | 28 (22.6%) | 19 (33.3%) | 153 (10.7%) | 30 (12.4%) | 17 (23.6%) | 3 (15%) | 9 (8.7%) | 5 (12.2%) | 201 (11.6%) | 37 (12%) | 4 (7.5%) | 98 (13.9%) | 104 (19%) | 55 (13.9%) |
| Ischaemic stroke | 15 (3.7%) | 0 (0%) | 7 (3.2%) | 6 (5.5%) | 1 (1.9%) | 4 (3.2%) | 0 (0%) | 53 (3.7%) | 22 (9.1%) | 5 (6.9%) | 0 (0%) | 2 (1.9%) | 4 (9.8%) | 70 (4%) | 8 (2.6%) | 1 (1.9%) | 25 (3.5%) | 14 (2.6%) | 18 (4.6%) |
| Myocardial infarction | 13 (3.2%) | 3 (10.3%) | 10 (4.5%) | 5 (4.6%) | 3 (5.8%) | 8 (6.5%) | 3 (5.3%) | 36 (2.5%) | 18 (7.5%) | 4 (5.6%) | 1 (5%) | 3 (2.9%) | 4 (9.8%) | 70 (4%) | 6 (1.9%) | 1 (1.9%) | 26 (3.7%) | 30 (5.5%) | 14 (3.5%) |
| Haemorrhage | 36 (8.9%) | 2 (6.9%) | 16 (7.2%) | 19 (17.4%) | 14 (26.9%) | 22 (17.7%) | 14 (24.6%) | 298 (20.8%) | 23 (9.5%) | 10 (13.9%) | 3 (15%) | 10 (9.6%) | 2 (4.9%) | 107 (6.2%) | 8 (2.6%) | 4 (7.5%) | 84 (11.9%) | 54 (9.9%) | 52 (13.2%) |
| Sepsis | 42 (10.4%) | 6 (20.7%) | 7 (3.2%) | 11 (10.1%) | 16 (30.8%) | 26 (21%) | 8 (14%) | 35 (2.4%) | 2 (0.8%) | 7 (9.7%) | 5 (25%) | 5 (4.8%) | 1 (2.4%) | 111 (6.4%) | 17 (5.5%) | 4 (7.5%) | 59 (8.4%) | 50 (9.1%) | 39 (9.9%) |
| Colitis | 11 (2.7%) | 1 (3.4%) | 1 (0.5%) | 2 (1.8%) | 3 (5.8%) | 3 (2.4%) | 2 (3.5%) | 12 (0.8%) | 1 (0.4%) | 0 (0%) | 1 (5%) | 19 (18.3%) | 0 (0%) | 10 (0.6%) | 1 (0.3%) | 2 (3.8%) | 3 (0.4%) | 7 (1.3%) | 6 (1.5%) |
| Diarrhoea | 42 (10.4%) | 5 (17.2%) | 7 (3.2%) | 9 (8.3%) | 8 (15.4%) | 9 (7.3%) | 15 (26.3%) | 112 (7.8%) | 9 (3.7%) | 7 (9.7%) | 0 (0%) | 20 (19.2%) | 4 (9.8%) | 98 (5.7%) | 11 (3.6%) | 8 (15.1%) | 34 (4.8%) | 51 (9.3%) | 48 (12.2%) |
| Dehydration | 69 (17.1%) | 4 (13.8%) | 6 (2.7%) | 3 (2.8%) | 2 (3.8%) | 4 (3.2%) | 0 (0%) | 20 (1.4%) | 2 (0.8%) | 1 (1.4%) | 0 (0%) | 15 (14.4%) | 0 (0%) | 72 (4.2%) | 6 (1.9%) | 0 (0%) | 20 (2.8%) | 26 (4.8%) | 55 (13.9%) |
| Hypokalaemia | 20 (5%) | 2 (6.9%) | 15 (6.8%) | 2 (1.8%) | 2 (3.8%) | 4 (3.2%) | 3 (5.3%) | 13 (0.9%) | 3 (1.2%) | 0 (0%) | 1 (5%) | 4 (3.8%) | 0 (0%) | 45 (2.6%) | 1 (0.3%) | 0 (0%) | 10 (1.4%) | 26 (4.8%) | 15 (3.8%) |
| Hyperthyroidism | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 10 (4.1%) | 1 (1.4%) | 0 (0%) | 14 (13.5%) | 1 (2.4%) | 8 (0.5%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 2 (0.4%) | 0 (0%) |
| All-cause death | 66 (21.2%) [311] | 4 (17.4%) [23] | 18 (8.7%) [206] | 42 (39.3%) [107] | 33 (63.5%) [52] | 38 (35.8%) [106] | 22 (43.1%) [51] | 130 (9.2%) [1412] | 15 (6.3%) [237] | 16 (22.2%) [72] | 8 (40%) [20] | 23 (22.3%) [103] | 5 (20.8%) [24] | 279 (16.1%) [1730] | 26 (8.4%) [308] | 7 (13.2%) [53] | 114 (16.8%) [679] | 95 (17.6%) [540] | 69 (19.7%) [351] |
| Alkylating agents | Androgen deprivation therapy | Antimetabolites | Anthracyclines | Kinase inhibitors | Immune checkpoint inhibitor | Immunomodulating agents | Monoclonal antibodies (anti-CD20) | Protea-some inhibitor | Taxane | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Bruton Tyrosin Kinase inhibitor | BCR-ABL inhibitors | Multi-kinase inhibitor | |||||||||||||||||
| Anticancer drug | Cisplatin | Dacarbazine | Abiraterone | Azacitidine | Clofarabine | Daunorubicin | Idarubicin | Ibrutinib | Nilotinib | Ponatinib | Midostaurin | Ipilimumab | Aldesleukin | Lenalidomide | Pomalidomide | Obinutuzumab | Rituximab | Bortezomib | Docetaxel |
| No | 404 | 29 | 222 | 109 | 52 | 124 | 57 | 1431 | 241 | 72 | 20 | 104 | 41 | 1733 | 308 | 53 | 706 | 547 | 395 |
| Male | 282 (73.8%) [382] | 19 (65.5%) [29] | 216 (100%) [216] | 57 (59.4%) [96] | 32 (65.3%) [49] | 66 (54.1%) [122] | 33 (60%) [55] | 899 (67.4%) [1334] | 131 (58%) [226] | 32 (62.7%) [51] | 12 (60%) [20] | 83 (79.8%) [104] | 29 (78.4%) [37] | 979 (58.2%) [1682] | 176 (58.3%) [302] | 31 (63.3%) [49] | 326 (50.9%) [640] | 318 (61.7%) [515] | 216 (55.2%) [391] |
| Age >65 years | 185 (61.5%) [301] | 20 (76.9%) [26] | 156 (94.5%) [165] | 74 (89.2%) [83] | 32 (72.7%) [44] | 49 (64.5%) [76] | 26 (49.1%) [53] | 885 (84.4%) [1049] | 102 (64.2%) [159] | 29 (69%) [42] | 14 (73.7%) [19] | 63 (66.3%) [95] | 14 (45.2%) [31] | 960 (82.7%) [1161] | 187 (82.7%) [226] | 32 (76.2%) [42] | 366 (74.5%) [491] | 312 (74.3%) [420] | 190 (60.9%) [312] |
| Time to onset, days | 16 (6–49) [126] | 12 (11–28) [5] | 69 (26–176) [72] | 59 (10–218) [75] | 12 (9–26) [8] | 18 (5–31) [18] | 14 (6–35) [13] | 110 (42–294) [335] | 148 (48–440) [102] | 129 (40–227) [23] | 18 (12–114) [13] | 44 (14–60) [15] | 5 (3–13) [15] | 66 (20–304) [375] | 42 (15–90) [76] | 39 (29–56) [9] | 33 (9–102) [157] | 29 (13–74) [137] | 16 (7–42) [200] |
| Concurrent medications [availability is equal to the number of AF cases] | |||||||||||||||||||
| Cardiovascular or diabetes medication | 208 (51.5%) | 16 (55.2%) | 113 (50.9%) | 50 (45.9%) | 24 (46.2%) | 31 (25%) | 32 (56.1%) | 571 (39.9%) | 106 (44%) | 40 (55.6%) | 6 (30%) | 46 (44.2%) | 11 (26.8%) | 1117 (64.5%) | 205 (66.6%) | 15 (28.3%) | 346 (49%) | 393 (71.8%) | 192 (48.6%) |
| Anti-coagulant | 44 (10.9%) | 5 (17.2%) | 33 (14.9%) | 12 (11%) | 5 (9.6%) | 1 (0.8%) | 3 (5.3%) | 239 (16.7%) | 43 (17.8%) | 11 (15.3%) | 3 (15%) | 23 (22.1%) | 1 (2.4%) | 446 (25.7%) | 84 (27.3%) | 10 (18.9%) | 105 (14.9%) | 111 (20.3%) | 58 (14.7%) |
| AAR | 36 (8.9%) | 4 (13.8%) | 21 (9.5%) | 10 (9.2%) | 8 (15.4%) | 5 (4%) | 4 (7%) | 119 (8.3%) | 23 (9.5%) | 8 (11.1%) | 0 (0%) | 10 (9.6%) | 2 (4.9%) | 272 (15.7%) | 41 (13.3%) | 6 (11.3%) | 53 (7.5%) | 64 (11.7%) | 43 (10.9%) |
| Systemic cortico-steroids | 108 (26.7%) | 8 (27.6%) | 129 (58.1%) | 20 (18.3%) | 14 (26.9%) | 25 (20.2%) | 16 (28.1%) | 102 (7.1%) | 8 (3.3%) | 15 (20.8%) | 1 (5%) | 33 (31.7%) | 2 (4.9%) | 825 (47.6%) | 168 (54.5%) | 7 (13.2%) | 353 (50%) | 370 (67.6%) | 128 (32.4%) |
| Concurrent adverse events [availability is equal to the number of AF cases] | |||||||||||||||||||
| Heart failure | 42 (10.4%) | 4 (13.8%) | 38 (17.1%) | 16 (14.7%) | 14 (26.9%) | 28 (22.6%) | 19 (33.3%) | 153 (10.7%) | 30 (12.4%) | 17 (23.6%) | 3 (15%) | 9 (8.7%) | 5 (12.2%) | 201 (11.6%) | 37 (12%) | 4 (7.5%) | 98 (13.9%) | 104 (19%) | 55 (13.9%) |
| Ischaemic stroke | 15 (3.7%) | 0 (0%) | 7 (3.2%) | 6 (5.5%) | 1 (1.9%) | 4 (3.2%) | 0 (0%) | 53 (3.7%) | 22 (9.1%) | 5 (6.9%) | 0 (0%) | 2 (1.9%) | 4 (9.8%) | 70 (4%) | 8 (2.6%) | 1 (1.9%) | 25 (3.5%) | 14 (2.6%) | 18 (4.6%) |
| Myocardial infarction | 13 (3.2%) | 3 (10.3%) | 10 (4.5%) | 5 (4.6%) | 3 (5.8%) | 8 (6.5%) | 3 (5.3%) | 36 (2.5%) | 18 (7.5%) | 4 (5.6%) | 1 (5%) | 3 (2.9%) | 4 (9.8%) | 70 (4%) | 6 (1.9%) | 1 (1.9%) | 26 (3.7%) | 30 (5.5%) | 14 (3.5%) |
| Haemorrhage | 36 (8.9%) | 2 (6.9%) | 16 (7.2%) | 19 (17.4%) | 14 (26.9%) | 22 (17.7%) | 14 (24.6%) | 298 (20.8%) | 23 (9.5%) | 10 (13.9%) | 3 (15%) | 10 (9.6%) | 2 (4.9%) | 107 (6.2%) | 8 (2.6%) | 4 (7.5%) | 84 (11.9%) | 54 (9.9%) | 52 (13.2%) |
| Sepsis | 42 (10.4%) | 6 (20.7%) | 7 (3.2%) | 11 (10.1%) | 16 (30.8%) | 26 (21%) | 8 (14%) | 35 (2.4%) | 2 (0.8%) | 7 (9.7%) | 5 (25%) | 5 (4.8%) | 1 (2.4%) | 111 (6.4%) | 17 (5.5%) | 4 (7.5%) | 59 (8.4%) | 50 (9.1%) | 39 (9.9%) |
| Colitis | 11 (2.7%) | 1 (3.4%) | 1 (0.5%) | 2 (1.8%) | 3 (5.8%) | 3 (2.4%) | 2 (3.5%) | 12 (0.8%) | 1 (0.4%) | 0 (0%) | 1 (5%) | 19 (18.3%) | 0 (0%) | 10 (0.6%) | 1 (0.3%) | 2 (3.8%) | 3 (0.4%) | 7 (1.3%) | 6 (1.5%) |
| Diarrhoea | 42 (10.4%) | 5 (17.2%) | 7 (3.2%) | 9 (8.3%) | 8 (15.4%) | 9 (7.3%) | 15 (26.3%) | 112 (7.8%) | 9 (3.7%) | 7 (9.7%) | 0 (0%) | 20 (19.2%) | 4 (9.8%) | 98 (5.7%) | 11 (3.6%) | 8 (15.1%) | 34 (4.8%) | 51 (9.3%) | 48 (12.2%) |
| Dehydration | 69 (17.1%) | 4 (13.8%) | 6 (2.7%) | 3 (2.8%) | 2 (3.8%) | 4 (3.2%) | 0 (0%) | 20 (1.4%) | 2 (0.8%) | 1 (1.4%) | 0 (0%) | 15 (14.4%) | 0 (0%) | 72 (4.2%) | 6 (1.9%) | 0 (0%) | 20 (2.8%) | 26 (4.8%) | 55 (13.9%) |
| Hypokalaemia | 20 (5%) | 2 (6.9%) | 15 (6.8%) | 2 (1.8%) | 2 (3.8%) | 4 (3.2%) | 3 (5.3%) | 13 (0.9%) | 3 (1.2%) | 0 (0%) | 1 (5%) | 4 (3.8%) | 0 (0%) | 45 (2.6%) | 1 (0.3%) | 0 (0%) | 10 (1.4%) | 26 (4.8%) | 15 (3.8%) |
| Hyperthyroidism | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 10 (4.1%) | 1 (1.4%) | 0 (0%) | 14 (13.5%) | 1 (2.4%) | 8 (0.5%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 2 (0.4%) | 0 (0%) |
| All-cause death | 66 (21.2%) [311] | 4 (17.4%) [23] | 18 (8.7%) [206] | 42 (39.3%) [107] | 33 (63.5%) [52] | 38 (35.8%) [106] | 22 (43.1%) [51] | 130 (9.2%) [1412] | 15 (6.3%) [237] | 16 (22.2%) [72] | 8 (40%) [20] | 23 (22.3%) [103] | 5 (20.8%) [24] | 279 (16.1%) [1730] | 26 (8.4%) [308] | 7 (13.2%) [53] | 114 (16.8%) [679] | 95 (17.6%) [540] | 69 (19.7%) [351] |
Data are expressed as n (%) [availability] or the median (interquartile range) [availability].
AAR: Class I or III antiarrhythmics or digoxin.
Identification of the molecular pathways involved in anticancer drug-associated atrial fibrillation
After the review process, 26 review articles were selected (Figure 1, Supplementary material online, Data S2). The main molecular pathways underlying AF development implied Ca2+ handling abnormalities, the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) and mitogen-activated protein kinases (MAPK)/extracellular-signal-regulated kinase (ERK) pathways, ROS production/mitochondrial dysfunction and pro-inflammatory pathways leading to atrial fibrosis and atrial action potential changes (Figure 3). Sixteen of the 19 identified anticancer drugs were able to modulate at least one of the molecular pathways involved in AF development (references supporting the modulation of a molecular pathway and the corresponding anticancer drugs are shown in Supplementary material online, Table S6). The main molecular pathways underlying anticancer drug-associated AF implied the PI3K/AKT and MAPK/ERK pathways and ROS production.
(A) Representation of molecular pathways involved in atrial fibrillation initiation, maintenance and progression. (B) Venn diagram of the molecular pathways involved in atrial fibrillation initiation, maintenance, and progression. Number of anticancer drugs (among the 19 anticancer drugs significantly associated with atrial fibrillation overreporting in VigiBase®) modulating each molecular pathway. AKT, protein kinase B; CaM kinase II, Ca2+/calmodulin-dependent protein kinase II; ERK; extracellular-signal-regulated kinase; MAPK, mitogen-activated protein kinases; PI3K, phosphatidylinositol 3-kinase; ROS; reactive oxygen species; RYR2, ryanodine receptor 2; SERCA2a, sarco/endoplasmic reticulum Ca2+-ATPase 2a.
Discussion
The explosion of novel oncology therapies has resulted in better prognosis for many cancer types over the last decade but can also result in cardiovascular risks.2 While much of the focus of the ‘cardio-oncology’ community has been on clinical heart failure associated with the therapies, less is known about the arrhythmogenic risk of anticancer drugs, especially AF. Using the largest medical post-marketing surveillance database, our paper systematically assessed AF association with anticancer drugs in general and we report the largest and most extensive clinical characterization of such cases. We found a significant overreporting association between AF and 19 anticancer drugs. Our study confirmed the association between AF and 10 anticancer drugs including ibrutinib, abiraterone, several cytotoxic agents (anthracyclines, dacarbazine, and cisplatin), rituximab, interleukin-2, bortezomib, and ipilimumab.11,21 Some of these associations were extensively previously described. In a meta-analysis of four randomized trials of ibrutinib, the pooled relative risk (95% CI) of AF associated with ibrutinib as compared with a comparator was 3.9 (2.0–7.5, P < 0.0001) and the pooled rate (95% CI) of AF among ibrutinib recipients in 20 randomized and non-randomized studies was 3.3 (2.5–4.1) per 100 person-years over a median follow-ups of 26 months.22 Aldesleukin (interleukin 2) has been associated with AF in cohort study with rates ranging from 4.3% to 8%.23,24 Interestingly, we reported new associations between AF and nine anticancer drugs, corresponding to several pharmacological classes including immunomodulating agents (lenalidomide, pomalidomide), several kinase inhibitors (nilotinib, ponatinib, midostaurin), antimetabolites (azacytidine, clofarabine), docetaxel (taxane), and obinutuzumab, an anti-CD20 monoclonal antibody (Figure 2). Although AF occurrence was described in case reports and/or phase 1/2 uncontrolled trials in patients treated with azacytidine, clofarabine, docetaxel, nilotinib, ponatinib, or lenalidomide, no evidence could sustain a specific and definite association. A greater than 19 million reports pharmacovigilance database is extremely valuable in this context to evaluate the specific association between anticancer drugs and AF. Among these 19 anticancer drugs, it is interesting to note that 14 (74%) are commonly prescribed in haematologic malignancies, particularly in chronic lymphocytic leukaemia and in multiple myeloma. Vigibase® does not provide the percentage of patients with pre-existing AF. Indirectly, we approached it using the presence of anti-arrhythmic drugs or anticoagulants, as potential ‘surrogates’ of pre-existing AF. We found that 1439 (23.4%) of 6147 cases included at least one anti-arrhythmic drugs or anticoagulant. Excluding these 1439 cases of the analysis did not change the conclusions regarding the 19 anticancer drugs significantly associated with AF.
In our study, we cannot exclude a confounding role of the underlying malignancy, irrespective of the anticancer drugs delivered. To date, there is no robust data showing that one type of malignancy would be mostly associated with AF compared with other type of malignancies. A recent nationwide cohort study highlighted that all major malignancy types were associated with an increased incidence of AF compared to the general population (17.4 per 1000 person-years vs. 3.7 per 1000 person-years, respectively) with a trend for an increase incidence associated with lung cancer.25 In our study, skin, oesophageal, and haematologic malignancies appeared to be significantly associated with AF reporting, irrespective of the anticancer drugs. However, these results must be considered with caution as the type of malignancy was reported in only 43.9% of the reports.
Although we adjusted the disproportionality analysis on an extensive number of confounders for AF reporting, it is impossible to formally exclude other nondrug aetiologies. However, the fact that 16 of the 19 identified anticancer drugs are able to modulate molecular pathways that are already known to be involved in AF development makes the causal link plausible. Moreover, the fact that these molecular pathways involved in AF largely overlap those of left ventricular and vascular toxicities, and oncologic signalling pathways, increases the probability of a causal association. Many pathophysiological processes contribute to the initiation, maintenance, and progression of AF.26 In ‘classical’ AF (e.g., non-anticancer drug-associated AF), abnormalities of Ca2+-handling play a central role in both focal ectopic activity and AF substrate progression. The slightest alteration of any of its components is likely to lead to AF initiation or progression.27 Recently, other pathways have been described as key components of AF development, such as cardiac mitochondrial dysfunction that emerged as a major player in postoperative AF occurrence.28 An interesting finding of our study is that anticancer drug-associated AF does not seem to predominantly implicate the ‘classical’ molecular pathway represented by Ca2+-handling abnormalities. In our analysis, the agents which had the strongest associations (AF-ROR of 3.0 or greater) all affect the PI3K/AKT signalling and MAPK signalling pathways, suggesting that these pathways may be mostly implicated in AF associated with anticancer drugs (Figure 3).
AF is generally manageable without anticancer drug discontinuation and anticancer drug interruption must be weighed against the risk of cancer progression.27 Patients with chronic lymphocytic leukaemia who had ibrutinib interrupted at the onset of AF had a decreased progression-free survival (median 19 months) compared with those who continued ibrutinib or had dose reductions (median 27 months, P = 0.023).30 There is no randomized clinical trial data addressing rate vs. rhythm control strategies in cancer patients experiencing AF associated with anticancer drugs but lessons learned from the management of AF associated with ibrutinib or other drugs can be helpful in this context.27,31 In haemodynamically stable cancer patients with AF related to anticancer drugs, rate control may be preferable to rhythm control because the ability to maintain sinus rhythm after cardioversion may be limited by the continuation of the imputed anticancer drug. Finally, the decision to use anticoagulants in cancer patients with AF should be individualized and remained challenging. The CHA2DS2-VASc and HAS-BLED scores, although not validated in cancer patients, are usually applied to determine individual thrombotic and bleeding risks, respectively.30 The use of anticoagulant agents during cancer therapy poses a particular challenge because of the risks drug–drug interactions and bleeding. Bleeding risk appeared not to be equivalent among the type of malignancy, with gastrointestinal cancer more at risk.1,32
Study limitations
Pharmacovigilance databases suffer from some bias such as underreporting, notoriety, halo bias or lack of information on the exposed population, and sales volumes of drugs. Lack of information includes several clinical observations or conditions (such as previous radiotherapy and/or surgery) that represent potential confounding factors for AF development. This lack of clinical data does not allow to determine if the 19 anticancer drugs associated with AF could directly trigger AF or rather be a precipitating factor in predisposing patients (i.e. in with advanced underlying cancer and/or cardiac disease) where cancer itself could represent a major confounding factor. Pharmacovigilance database facilitate the study (and the early emergence) of rare AE(s) and such associations between drug(s) and AE(s) are not systematically associated with an unfavourable benefit-risk ratio (which is dependent at least on the AE rate and its severity). Although our primary analysis is adjusted on the presence of all the 176 anticancer drugs studied, synergistic effects on top of additive effects may influence AF development in patients treated with concurrent or sequential use of several anticancer drug. However, interaction analyses in VigiBase® have not yet been validated. The case/non-case design of our analyses cannot provide evidence of a definite proof of a causal relationship between the 19 anticancer drugs identified and AF development. Since our analyses were performed on overreporting and not overrisk, we were not able to determine which drug is most likely to cause AF but could only list drugs for which AF was overreported, which is an indirect approach to determine AF risk. Unfortunately, the incidence of anticancer drug-associated AF cannot be estimated in VigiBase®. The proportion of AF cases (1.03%) compared to all reported AEs must be interpret with caution since we chose other anticancer drugs as controls, the overreporting of some adverse events may artificially mask an authentic overrisk related to another adverse event that is less often reported.
Conclusions
Using the largest post-marketing surveillance database, our work identified 19 anticancer drugs that are significantly associated with AF. Of these 19 anticancer drugs, 14 are mainly used in haematologic malignancies and 9 represent new AF associations not previously confirmed in the literature. Oncologists and cardiologists should be aware that these anticancer drugs are suspected to induce AF. These findings may prove useful to physicians when confronted with AF in cancer patients. Dedicated prospective clinical trials are now required to confirm these 19 associations highlighted using the VigiBase® pharmacovigilance database.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available as data were extracted from VigiBase®, the WHO global database of individual case safety reports, made available by the Uppsala Monitoring Centre (Uppsala, Sweden). Restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Requests can be made to the corresponding author who will submit it to the Uppsala Monitoring Centre (Uppsala, Sweden) to obtain its permission.
Supplementary material
Supplementary material is available at European Heart Journal – Cardiovascular Pharmacotherapy online.
Acknowledgements
The information presented in this study does not represent the opinion of the UMC or the World Health Organization. The authors thank the custom searches team at the Uppsala Monitoring Centre (Uppsala, Sweden) research section for providing the Vigibase extract case level data (VigiBase®, the WHO global database of individual case safety reports), without whom this study would not have been possible.
Funding
The study was supported by Caen Normandy University Hospital (CHU Caen Normandie, France) and Normandy University (Université de Caen Normandie, France).
Conflict of interest: none declared.
