Patient Preferences for Deactivation of Implantable Cardioverter-Defibrillators

While implantable cardioverter-defibrillators (ICDs) prolong life, painful shocks can occur at the end of life, and physicians rarely discuss the option of device deactivation.^1,2 To prevent shocks at the end of life that do not meaningfully prolong survival, a recent Heart Rhythm Society consensus statement recommended proactive communication with patients regarding deactivation.³ However, most surveys have found that the majority of ICD patients would not consider deactivation even in deleterious future health states including terminal cancer,⁴ constant dyspnea,⁵ or frequent shocks.^5,6

We examined preferences for ICD deactivation in the context of health outcomes such as functional and cognitive disabilities shown to matter most to patients.⁷ Our survey provided an informational description of the potential benefits and burdens of the shocking function of ICDs and subsequently presented scenarios describing broad health outcomes common in patients approaching the end of life.

We recruited ICD patients 50 years and older who were followed in a single academic electrophysiology practice (Yale University). Eligible participants were English speaking and had received a new ICD implant or generator change between January 1, 2006, and December 31, 2009. We excluded patients who were cognitively impaired, had an initial device within the past year, or had a cardiac resynchronization therapy-capable device. Eligible participants were contacted by telephone by 1 investigator (J.A.D.), and those consenting were administered a 20-minute telephone survey. The study was approved by the Yale Human Investigation Committee.

The survey (eAppendix ) included 2 open-ended questions: “What do you feel are the potential benefits of your ICD?” and “What do you fell are the potential harms of your ICD?” Responses were transcribed and later categorized by 3 investigators (J.A.D., T.R.F., and R.L.), with discrepancies resolved through consensus. Participants were then read a standardized script providing the best current evidence regarding benefits and burdens of ICDs. Subsequently, participants were asked whether they would want their ICD deactivated in 5 scenarios representing key domains of health⁸ that strongly influence treatment preferences.⁷ The scenarios were as follows: (1) permanently unable to get out of bed, (2) permanent memory problems, (3) burden to family members, (4) prolonged mechanical ventilation (>1 month), and (5) advanced incurable disease. Participants responded using a scale of 1 (“definitely no”) to 5 (“definitely yes”); they were classified as wanting deactivation if they responded “possibly yes” or “definitely yes” to at least 1 scenario.

Of 136 eligible patients contacted, 95 (70%) agreed to participate. The mean (SD) age was 71.4 (10.6) years, 28% were female, and 19% were nonwhite. The mean (SD) duration of ICD placement was 4.0 (2.4) years prior, and 29% of participants had received a prior ICD shock.

Thirty-one participants reported “unknown” or “no” benefits of their ICD. Other responses were categorized as restarting the heart (n = 31), living longer (n = 27), improving quality of life (n = 7), other (n = 4 [3 patients thought the ICD prevented atrial fibrillation or stroke and 1 thought that it improved heart function]). Sixty participants could not name a potential harm. The remaining categories were receiving a shock (n = 14), device malfunction (n = 8), nuisance (n = 5), complication (infection or bleeding) (n = 2), inappropriate end-of-life shocks (n = 2), and other (n = 4 [1 was related to a phantom shock and 3 were related to unrealistic fears regarding harming the device—by lying down, touching the device, or falling]) (Figure).

Sixty-seven participants (71%) wanted ICD deactivation in 1 or more scenarios. Responses to individual scenarios ranged from 61% wanting deactivation in the setting of advanced incurable disease to 24% wanting deactivation if permanently unable to get out of bed (eFigure). Participant characteristics associated with wanting deactivation included race and disability but not age (Table).

In contrast to prior findings, in our study, following an informational script the majority (71%) of participants wanted ICD deactivation in at least 1 scenario describing health outcomes that are common in patients approaching the end of life.

There are several potential explanations for why our findings differ from prior investigations suggesting that the majority of patients would not want deactivation even in scenarios reflecting advanced illness.^4-6 One may be the age of our study sample, which consisted exclusively of older individuals (age ≥50 years). Another may be the study protocol, which involved reading each participant an informational script explaining the purpose of their ICD. Our qualitative finding that a sizeable number of participants did not have a good understanding of the benefits or potential burdens of their ICD underscores this possibility and highlights the importance of physician communication of information. Finally, we included a broad range of scenarios reflecting outcomes of advanced functional, cognitive, and medical illness, which further highlight the importance of communication in the framework of goal setting around desired health states,⁹ since tolerance for shocks and discomfort will be based on the ultimate perceived outcome. We believe that our findings emphasize the importance of incorporating multiple patient-centered outcomes into advance care planning for ICD patients.

Correspondence: Dr Lampert, Yale University School of Medicine, Section of Cardiology, 789 Howard Ave, Dana 3, Room 319, New Haven, CT 06520 (rachel.lampert@yale.edu).

Published Online: January 28, 2013. 10.1001/jamainternmed.2013.1883. Corrected March 11, 2013.

Author Contributions:Study concept and design: Dodson, Fried, and Lampert. Acquisition of data: Dodson. Analysis and interpretation of data: Dodson, Fried, Van Ness, Goldstein, and Lampert. Drafting of the manuscript: Dodson. Critical revision of the manuscript for important intellectual content: Fried, Van Ness, Goldstein, and Lampert. Statistical analysis: Dodson and Van Ness. Obtained funding: Fried. Study supervision: Fried, Lampert.

Funding/Support: Support for this study was provided by a training grant in Geriatric Clinical Epidemiology from the National Institute on Aging (T32 AG019134) (Dr Dodson) and a mid-career mentorship award from the National Institute of Aging (K24AG028443) (Dr Fried). This work was supported in part by a grant from the Claude D. Pepper Older Americans Independence Center at Yale University School of Medicine (P30AG021342 National Institutes of Health/National Institute of Aging).