Author + information
- Received January 8, 2016
- Revision received February 12, 2016
- Accepted March 3, 2016
- Published online May 18, 2016.
- Andrés Enriquez, MDa,∗ (, )
- Jim Biagi, MDb,
- Damian Redfearn, MDa,
- Usama Boles, MDa,
- Dalia Kamel, MDb,
- Fariha Sadiq Ali, MDa,
- Wilma M. Hopman, MAa,
- Kevin A. Michael, MDa,
- Christopher Simpson, MDa,
- Hoshiar Abdollah, MDa,
- Debra Campbell, RNa and
- Adrian Baranchuk, MDa
- aHeart Rhythm Service, Queen’s University and Kingston General Hospital, Kingston, Ontario, Canada
- bCancer Center of Southeastern Ontario, Kingston, Ontario, Canada
- ↵∗Reprint requests and correspondence:
Dr. Andres Enriquez, Division of Cardiology, Heart Rhythm Service, Kingston General Hospital, Queen's University, 76 Stuart Street, Kingston, Ontario K7L 2V7, Canada.
Objectives This study evaluated the incidence of ventricular arrhythmia and implantable cardioverter-defibrillators (ICDs) therapies in patients with a diagnosis of cancer.
Background Cardiac disease and cancer are prevalent conditions and share common predisposing factors. No studies have assessed the impact of cancer on the burden of ventricular arrhythmia in patients with cancer and ICDs.
Methods Retrospective study of patients with an ICD and cancer followed from January 2007 to June 2015. Rates of ventricular tachycardia (VT) and ventricular fibrillation (VF) before and after patients’ cancers were diagnosed were evaluated by searching device data collection systems. Rates were adjusted for length of follow-up and compared using the Wilcoxon test, and times to first therapy following diagnosis (stages I to III vs. IV) were compared using Kaplan-Meier curves and log-rank test.
Results Among 1,598 patients with an ICD, 209 patients (13.1%) had a pathological diagnosis of malignancy; and in 102 patients (6.4%), malignancy was diagnosed following device insertion. After the diagnosis of cancer, 32% of patients experienced VT/VF over 23.2 ± 23.6 months, and the frequency of arrhythmic events was significantly increased after the diagnosis (1.19 ± 0.32 vs. 0.12 ± 0.21 episodes per month, respectively; p = 0.03). The incidence of VT/VF was markedly higher in patients with stage IV cancer than in those with earlier stages (p = 0.03). In this group, the incidence of VT/VF was 41.2%, with an average of 7.2 ± 18.5 events per patient, all of whom received ICD shocks. The rate of ICD deactivation in stage IV patients was 35.3%. Inappropriate therapies occurred in 13.7%, and atrial fibrillation was the most frequent cause.
Conclusions One-third of patients who had received ICDs developed ventricular arrhythmia after a diagnosis of cancer. The incidence was significantly higher in those with advanced metastatic disease. Findings underscore the need to discuss ICD management as part of end-of-life care.
Cardiac disease and cancer frequently coexist. These 2 conditions respectively represent the first and second most common causes of death in industrialized countries (1), both are prevalent in advanced-age population and they share common predisposing factors such as tobacco use or an unhealthy lifestyle.
Ventricular arrhythmia and sudden death are responsible for one-half of all deaths in patients with heart disease (2). Insertion of an implantable cardioverter-defibrillator (ICD) has been shown to improve survival in high-risk groups and is a therapy frequently prescribed in the context of advanced heart disease. The impact of cancer on the natural history of heart disease is not completely understood, and no studies have investigated the burden of ventricular arrhythmia in patients with cancer. This information is relevant for patient care and has ethical implications because of the decision to withdraw ICD therapy after a diagnosis of advanced cancer and the time to do so are recurrent questions for patients, their families, and health care providers.
Several factors may contribute to an increased incidence of ventricular arrhythmia in patients with cancer, including electrolytic abnormalities, systemic proinflammatory state, cachexia, or drug toxicity. Antineoplastic chemotherapy can be complicated by cardiotoxicity through different mechanisms, including direct cardiac injury, QT prolongation, or ischemia (3–5). In addition, secondary dissemination to the heart is more common than thought, with some series reporting up to 25% in post-mortem studies (6,7).
The primary aim of this study was to determine the incidence of ventricular arrhythmia and ICD therapies in patients with cardiac disease and a pathological diagnosis of cancer.
Study design and patients
We conducted a retrospective study of patients with an ICD and a concurrent diagnosis of cancer. All patients with an ICD implanted and/or followed at Kingston General Hospital and Peterborough Regional Health Centre from January 2007 to June 2015 were screened for inclusion.
Inclusion criteria were: 18 years of age or older; ischemic or nonischemic cardiomyopathy or cardiac channelopathy; ICD (single-chamber, dual-chamber, or biventricular) implantation for primary or secondary prevention of ventricular arrhythmia; and confirmed pathologic diagnosis of malignancy. Precancerous conditions, such as actinic keratosis, Barrett's esophagus, cervical dysplasia, or monoclonal gammopathy of undetermined significance (MGUS), were not included. The study was approved by the Health Sciences Research Ethics Board of Queen’s University (HSREB number 6015312).
Ventricular arrhythmias documented by the ICD, defined as any ventricular fast rate falling in the ventricular tachycardia (VT) or ventricular fibrillation (VF) zone, including monitored events sustained for more than 30 seconds and associated with symptoms or requiring cardioversion. Primary outcomes also included appropriate therapies delivered by the ICD, including shocks and antitachycardia pacing (ATP).
Secondary outcomes included time to first appropriate ICD therapy after the diagnosis of cancer; inappropriate ICD therapies (both ATP and shocks) defined as therapies delivered for any rhythm rather than VT/VF; and incidence of electric storm, defined as 2 or more appropriate VT/VF detections in <24 h, treated by ATP or shock or untreated but sustained in a VT-monitoring zone (8–10).
Device data were obtained through analysis of electronic charts from the Cardiac Rhythm Device Clinic at Kingston General Hospital and the satellite Peterborough Cardiac Device Clinic. In both clinics, patient follow-up examinations were conducted at 6-month intervals or more frequently in case of arrhythmic events or technical issues (e.g., lead noise, battery close to replacement indication, suboptimal or rising thresholds, suboptimal sensing). In patients followed through transtelephonic monitoring systems (Carelink [Medtronic Inc., Minneapolis, Minnesota] or Merlin [St. Jude Medical, St. Paul, Minnesota]), an annual visit to the clinic was routinely scheduled for assessment of thresholds. Oncologic data were collected through review of medical registries and electronic files from the Cancer Centre of Southeastern Ontario and the Department of Pathology of Queen’s University.
In patients presenting with VT storm, clinical charts and patient notes were reviewed to identify potential precipitating factors, such as acute ischemic events, electrolyte abnormalities (hypokalemia, hypomagnesemia, or hypocalcemia); QT prolongation; use of cardiotoxic chemotherapy agents; and cardiac, pericardial, or mediastinal involvement by the cancer.
Data were collected in an Excel (Microsoft, Redmond, Washington) file designed for the study and analyzed using both Excel and SPSS version 22.0 software (IBM, Armonk, New York) for Windows (Micrososoft). Initial descriptive analysis included mean ± SD for continuous variables and frequencies and percentages for categorical variables for both the entire sample and the subset with malignancy.
The incidence of appropriate ICD therapies (adjusted for length of follow-up) was compared with that in a control group of ICD patients without diagnosis of cancer followed during the same period. In addition, the frequencies of VT/VF (episodes per month per patient) after the diagnosis of cancer were compared with the frequencies of VT/VF in the same patients before the diagnosis of cancer, using the Wilcoxon signed ranks test, as these data were not normally distributed. Time to first therapy following diagnosis (cancer stage I to III vs. IV) was compared using a Kaplan-Meier curve and the log-rank test. Factors associated with the development of VT/VF were examined using t tests (continuous data) and Pearson chi-square tests (categorical data), followed by multivariate logistic regression.
Characteristics of the population
A total of 1,598 patients with an ICD were followed at the Cardiac Rhythm Device Clinic of Kingston General Hospital and Peterborough Cardiac Device Clinic during the study period. Of 1,598 patients, 209 (13.1%) had a pathologically confirmed diagnosis of malignancy, and in 102 patients (6.4%), the cancer was diagnosed after the device was implanted. Characteristics of the patients are shown in Table 1. Mean age was 70.0 ± 8.4 years of age, and 72.7% were male. The device was single-chamber in 43.5%, dual-chamber in 26.3%, and biventricular (cardiac resynchronization therapy defibrillator [CRTD]) in 30.1%, but 14 patients with an original single- or dual-chamber device underwent subsequent upgrade to a CRTD during the study period. Etiology was ischemic in 67% and indication was primary prevention in 75.1%. Seventy patients (33.5%) had a history of atrial fibrillation or flutter. Only 5 of the 209 patients were lost to follow-up due to relocation away from the catchment area (follow-up: 98%).
Most common malignancies were in skin (25% [n = 53]), prostate (12% [n = 26]), breast (12% [n = 25]), lung (9% [n = 18]), and colon (6% [n = 12]). Hematologic cancers together represented 8% of cases. Primary malignancy was unknown in 1 case, and 28 patients had more than one malignancy (Figure 1). Skin cancer was the most prevalent malignancy in total, affecting 65 patients and including 6 melanomas. Prostate cancer was the most common cancer in males, with 37 patients, and breast cancer was the most common cancer in females, with 27 cases.
In those patients in whom cancer was diagnosed after ICD implantation, the distribution of malignancies was 33% in skin (n = 34), 14% in lung (n = 14), 11% in prostate (11%), 7% in colon (n = 7), and 5% in breast (n = 5). In this group, cancer stage at the time of diagnosis was I in 47% (n = 48), II in 22% (n = 22), III in 9% (n = 9), and IV in 17% (n = 17). No staging was possible in 6% of patients (n = 6).
During the 7-year study period, 89 of 204 patients with cancer experienced at least 1 VT/VF episode (43.6%), 83 received appropriate therapies from the ICD (40.6%), and 55 received appropriate ICD shocks (27.0%) (Table 2). Thirty-three patients (16.2%) experienced at least 1 episode of VT storm after a mean follow-up of 40.2 ± 32.3 months, with 5 patients (2.4%) having more than 1 episode. The occurrence of VT/VF was 39.6% (61 of 154) in patients with ICDs for primary prevention and 56.0% (28 of 50) in patients with implants for secondary prevention. Mean time to first VT/VF event after diagnosis of cancer was 16.0 ± 16.8 months, although this was much shorter (5.8 ± 4.9 months) in patients with stage IV cancer.
The occurrence of ventricular arrhythmia was significantly higher than in patients without cancer, followed during the same period, both in those with primary prevention (30.3%; p = 0.021) and secondary prevention indication (40%; p = 0.028), especially considering that mean follow-up in patients without cancer was twice as long (45.05 ± 36.6 months) (Figure 2). In patients without cancer, the mean time to first VT/VF event was 27.6 ± 28.9 months.
In patients with cancer diagnosed after the ICD implantation, the mean time to the diagnosis of cancer was 41.1 ± 28.1 months. Following diagnosis, 32% of the patients experienced ventricular arrhythmia over a period of 23.2 ± 23.6 months. Interestingly, most of these patients (61%) had received the device for primary prevention and had no documented VT/VF prior to cancer diagnosis.
The frequency of VT/VF before the diagnosis of cancer was 0.12 ± 0.21 episodes per month, compared with 1.19 ± 0.32 episodes per month after the diagnosis of cancer, which represents a statistically significant increase in arrhythmia burden by a factor of 10 (p = 0.031, Wilcoxon Signed Ranks test).
Predictors of ventricular arrhythmia in cancer patients
Table 3 compares the clinical characteristics of patients with cancer who developed ventricular arrhythmia compared to those who did not. No significant differences were observed between the 2 groups regarding age, sex, ischemic cause, ejection fraction, or indication (primary versus secondary). Advanced metastatic disease was more prevalent in patients presenting with VT/VF (25.0 vs. 10.0%, respectively; p = 0.047). This group was also more likely to have received chemotherapy (15.4 vs. 2.0%, respectively; p = 0.03). None of the variables remained significant after multivariate analysis.
Ventricular arrhythmia and cancer stage
In the group of patients diagnosed with stage IV cancer (n = 17), the follow-up after the diagnosis of cancer was 8.8 ± 13.4 months and 88% died. In this period, 41.2% of them (n = 7) experienced VT/VF, with an average of 7.2 ± 18.5 events per patient, all of them receiving ICD shocks. Five of these 7 patients (71%) had a primary prevention indication and no prior documentation of VT/VF before the diagnosis of cancer. The incidence of VT/VF was markedly higher than that in patients without systemic dissemination of the disease (stages I to III), as shown in a Kaplan-Meier curve (Figure 3).
Among those patients in whom cancer was diagnosed after ICD implantation (n = 102), 57% underwent surgical excision of the tumor (n = 58), 24% received radiation therapy (n = 24), and 9% received chemotherapeutic drugs (n = 9). Eight patients received potentially cardiotoxic agents, including cyclophosphamide, fluorouracil, cytarabine, paclitaxel, vincristine, vinblastin, cisplatin, etoposide, rituximab, and bortezomib. In patients treated with chemotherapy, the occurrence of VT/VF was particularly high (88.9%; p = 0.0186 compared to patients not treated with chemotherapy).
At least 1 episode of VT storm was identified in 33 patients (16.2%); 21 of them required hospitalization, and 5 patients died during that admission. In most patients (26 of 33), VT storm occurred after the malignancy was diagnosed, and in 8 patients, the VT storm presented in the context of advanced disseminated disease. As potential contributing factors, electrolytic abnormalities (low potassium and/or magnesium), were detected in 5 of these patients, very frequent ventricular extrasystole in 2 patients, and epinephrine infusion in 1 patient. In another patient, the VT storm occurred in the context of pulmonary infection and QT prolongation secondary to moxifloxacin. Five of the 33 patients were receiving potentially cardiotoxic agents at that time, although none were known to prolong the QT (bortezomib in 2 patients, cyclophosphamide and rituximab in 2 patients, and cytarabine in 1 patient).
The incidence of inappropriate therapies during the study period was 13.7% (n = 28), including 11.3% (n = 23) of inappropriate shocks. Reasons for inappropriate therapies included rapid atrial fibrillation (n = 12), supraventricular tachycardia (n = 9), lead noise (n = 4; including 1 secondary to lead dislodgement and 1 lead fracture), electromagnetic interference (n = 1), T-wave oversensing (n = 1), or both supraventricular tachycardia and T-wave oversensing (n = 1).
Antitachycardia therapies (ATP and shocks) were inactivated in 35.3% of patients whose cancer was diagnosed as stage IV (40% in patients who died during the study period). Time between cancer diagnosis and ICD deactivation in this group was 15.3 ± 20.3 months.
The main finding of this study is that cancer is newly diagnosed in 6% to 7% of patients with an ICD and that almost one-third of these patients will experience ventricular arrhythmia after the malignancy is diagnosed. The incidence of VT/VF is particularly high in patients with advanced disease with systemic dissemination (stage IV cancer), and time to first appropriate therapy is also significantly shorter in this population.
To our knowledge, this is the first study to assess the impact of cancer on the clinical course of patients with an ICD. With an aging population and more than 100,000 ICDs implanted each year in the United States (11), this is not an infrequent clinical scenario, and health providers are expected to be confronted more and more with decision-making in this setting. In a large population-based cohort, Pedersen et al. found that ICD patients have a slightly elevated risk of cancer compared with the general population (standardized incidence rate [SIR] of 1.1; 95% CI: 1.0 to 1.2), mainly driven by tobacco-related cancers (SIR 1.4; 95% CI: 1.2 to 1.6) (12). The incidence rate of cancer in our cohort was 16.8 per 1000 patients per year and tobacco-related cancers represented 34% of all neoplasias.
The reasons behind the high burden of ventricular arrhythmia in cancer patients are not completely understood, but several factors may contribute to this phenomenon. The fact that many patients started experiencing VT/VF after being diagnosed with cancer despite a long arrhythmia-free life before the diagnosis suggests a pathophysiological link between cancer and arrhythmogenesis. The burden of VT/VF measured in events per month increased by a factor of 10 after the diagnosis of cancer. Moreover, 1-fourth of VT storm episodes in our study occurred in the context of terminal disseminated disease. A possible mechanism for this nexus is inflammation, as inflammatory response is up-regulated in the context of cancer (13,14) and evidence supports an association between proinflammatory activity and ventricular arrhythmia (15–17). Inflammation enhances all stages of carcinogenesis, and major inflammatory mediators in cancer include interleukin-1β, interleukin-6, tumor necrosis factor-α, chemokines, and transcription factor NF-κB (13,14).
Direct cardiac involvement by the tumor is another possible mechanism that may explain an increased predisposition to ventricular arrhythmia in advanced stages. VT has been reported to be a manifestation of metastatic cardiac disease and may be the presenting symptom in some cases (18–20). The incidence of cardiac metastasis ranges between 2.3 and 18.3% in post-mortem examinations (6,16) and varies widely depending on the primary tumor. Tumors showing the highest rates of heart metastasis were pleural mesothelioma (48.4%), melanoma (27.8%), lung (21% adenocarcinoma and 18.2% squamous cell carcinoma), and breast (15.5%) (21). Chemotherapy was also associated with a higher incidence of ventricular arrhythmia in our study, a finding that may be attributed to the potentially cardiotoxic effect of the cytostatic agents or the advanced stage of disease in these patients. Other potential factors to consider include emotional stress (22,23), malnutrition and cachexia, electrolyte disturbances secondary to vomiting, diarrhea or decrease in oral intake, radiation-induced cardiac injury, and myocardial ischemia resulting from increased demands, anemia or hypoxia secondary to pulmonary involvement, atelectasis or pleural effusion.
The high incidence of ventricular arrhythmia in the context of advanced cancer underlines the need to discuss the opportune moment of ICD deactivation when dealing with patients nearing end-of-life. ICDs may painfully prolong the process of dying, and ICD shocks have a negative impact on quality of life (24). There is consensus that withdrawal of ICD therapy in terminally ill patients is ethically justified if consistent with the patient's goals of care (25). In addition, shock therapy can be disabled while ATP function and antibradycardia pacing are maintained. Despite this, we found a low rate of ICD deactivation in patients with advanced cancer (35.3% in stage IV patients), a fact that has also been noted by other studies (26). The reasons for this are diverse. Many patients and relatives are not aware that deactivation is an option (27) and few include mention of ICDs in their advance directives (28). ICD deactivation is sometimes perceived as a form of assisted suicide (29). In addition, studies show that many physicians report a lack of confidence when discussing cessation of ICD therapy (29,30).
This study has certain strengths. First, selection bias was avoided by the inclusion of all ICD patients followed at Kingston General Hospital and Peterborough satellite clinic. The concentration of patient care in only 2 centers explains the minimal loss to follow-up (2%). Both of the device clinics were provided with computer-based systems with which to store and organize data, such as Paceart (Medtronic Inc.) and OneView (ScottCare Cardiovascular Solutions, Cleveland, Ohio), granting thorough information regarding programming, events, and delivered therapies.
Some limitations should also be acknowledged. First, the study was retrospective, the number of patients was relatively small, and the population was heterogeneous. Electrograms of the events were not available for review in all cases and differentiating between VF and fast monomorphic VT was not always possible. In addition, device programming was not uniform for all patients and, therefore, the criteria for VT detection were not the same. The mechanisms of the association between cancer and ventricular arrhythmia are speculative and deserve further investigation.
Almost one-third of patients with an ICD will experience ventricular arrhythmia after a diagnosis of cancer, and the incidence of VT/VF is significantly higher in patients with advanced metastatic disease (stage IV). These findings suggest a possible pathogenic association between both conditions and underscore the need to discuss patient’s views regarding ICD management as part of the end-of-life care.
COMPETENCY IN MEDICAL KNOWLEDGE: The burden of ventricular arrhythmia is high in patients with ICD after a diagnosis of cancer. ICD management should be an important aspect of end-of-life care.
TRANSLATIONAL OUTLOOK: There is a link between ventricular arrhythmia and cancer, particularly in advanced stages of cancer. Mechanisms of this association are speculative and deserve further investigation.
Dr. Baranchuk has received speaker’s honoraria from Medtronic and St Jude Medical; and grants from Medtronic and Bayer. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- antitachycardia pacing
- cardiac resynchronization therapy defibrillator
- implantable cardioverter-defibrillator
- monoclonal gammopathy of undetermined significance
- standardized incidence rate
- ventricular fibrillation
- ventricular tachycardia
- Received January 8, 2016.
- Revision received February 12, 2016.
- Accepted March 3, 2016.
- American College of Cardiology Foundation
- Go A.S.,
- Mozaffarian D.,
- Roger V.L.,
- et al.,
- American Heart Association Statistics Committee and Stroke Statistics Subcommittee
- Tomaselli G.F.,
- Zipes D.P.
- Yeh E.T.,
- Bickford C.L.
- Barbey J.T.,
- Pezzullo J.C.,
- Soignet S.L.
- Kantarjian H.M.,
- Giles F.,
- Gattermann N.,
- et al.
- Reynen K.,
- Köckeritz U.,
- Strasser R.H.
- Kowey P.R.,
- Levine J.H.,
- Herre J.M.,
- et al.
- Pedersen S.B.,
- Nielsen J.C.,
- Bøtker H.E.,
- Farkas D.K.,
- Schmidt M.,
- Sørensen H.T.
- Albert C.M.,
- Ma J.,
- Rifai N.,
- Stampfer M.J.,
- Ridker P.M.
- Sharma J.,
- Brunson J.M.,
- Memon N.,
- Khakoo A.Y.
- Bussani R.,
- De-Giorgio F.,
- Abbate A.,
- Silvestri F.
- Steinberg J.S.,
- Arshad A.,
- Kowalski M.,
- et al.
- Lampert R.,
- Jain D.,
- Burg M.M.,
- et al.
- Schron E.B.,
- Exner D.V.,
- Yao Q.,
- et al.
- Lampert R.,
- Hayes D.L.,
- Annas G.J.,
- et al.,
- American College of Cardiology,
- American Geriatrics Society,
- American Academy of Hospice and Palliative Medicine,
- American Heart Association,
- European Heart Rhythm Association,
- Hospice and Palliative Nurses Association
- Goldstein N.,
- Bradley E.,
- Zeidman J.,
- Mehta D.,
- Morrison R.S.