The trusted source for
healthcare information and
Survival After Defibrillator Implantation
Abstract & Commentary
By John P. DiMarco, MD, PhD
Source: Lee DS, et al. Effect of cardiac and noncardiac conditions on survival after defibrillator implantation. J Am Coll Cardiol 2007;49:2408-2415.
Lee and co-workers examined the effects of associated cardiac and noncardiac conditions on survival in ICD recipients. The authors used the Canadian Institute for Health Information discharge abstract and same-day surgery databases and the Province of Ontario's Vital Status database and identified all patients between age 18 and 105 who received a first ICD implant between April, 1997 and March, 2003 in the province. Comorbidities present before ICD insertion were identified from the secondary diagnosis fields of the database and were classified using the Deyo-Charlson comorbidity classification system. Secondary diagnosis data were collected both during the ICD implant admission and from data for hospital admissions in the 3 years before the ICD implant. A cohort of control subjects were matched according to age, prior history of arrhythmia, prior heart failure history and number of significant noncardiac comorbidities.
There were 2,467 patients who received a first ICD implant during the study. The mean age was 62.5 + 13 years and 79% were men. Most of the patients received the ICD for secondary prevention with 82.6% having a history of ventricular tachycardia, ventricular fibrillation or cardiac arrest. Ischemic heart disease was the primary cardiac diagnosis in 59% and 42.5% had a history of congestive heart failure. The most common comorbidities were diabetes (18.2%), diabetes with microvascular complications (2.9%), chronic obstructive pulmonary disease (10.4%), cerebrovascular disease (6.4%), peripheral vascular disease (5.8%), and renal disease (5.8%). By multivariate analysis, age but not gender, was associated with an increased hazard ratio for mortality. Compared to those subjects less than 65 years of age, the hazard ratio (HR) for the age group 65 to 74 was 2.05 and for the group greater than 75, the HR was 3.0. Comorbidities were predictive of an adverse outcome. Heart failure (HR=2.33), diabetes with microvascular complications (HR=2.33), rheumatologic disease (HR=1.89), peripheral vascular disease (HR=1.50), chronic pulmonary disease (HR=1.35), and cancer (HR=1.81) were the comorbidities independently associated with mortality that were in the multivariate analysis. Prior heart failure and, in particular, heart failure with a hospital admission within the 6 months preceding ICD analysis, was also strongly associated with increased mortality. When the hazard ratio for death were adjusted for age, gender and heart failure, noncardiac comorbidities continued to be associated with increased mortality. The hazard ratios for patients with 1, 2, and greater than or equal to 3 noncardiac comorbidities were 1.72, 2.79, and 2.98 respectively. Of note, arrhythmia history was not strongly predictive of an adverse outcome. However, in the subgroup of patients with prior ventricular tachycardia, the ICD implant was associated with a significant increase in survival with an adjusted hazard ratio of 0.80.
The authors conclude that the presence of noncardiac comorbidities and prior clinical heart failure are significant predictors of death in ICD recipients. Increased age and associated comorbidities interact to lead to higher mortalities in older and sicker groups. The authors urge further research to better characterize the role of heart failure and other comorbidities as determinants of outcome in ICD recipients.
This is an important paper that reports outcomes after ICD implant in "a real world" setting. Published indications for ICD insertion are largely based on the entry criteria for the randomized clinical trials that have shown benefits with ICD use for both primary prevention of sudden death and secondary treatment of patients with sustained arrhythmias. Although these randomized clinical trials often excluded patients with significant comorbidities and investigator bias during screening eliminated many patients with complicated conditions during the screening phase, these factors are not mentioned in published indications. In this paper, we see that comorbidities, both cardiac and noncardiac, strongly influence outcome after ICD, and the data emphasize a need to interpret the published indications in light of the individual patient's overall medical condition.
Other studies have reported findings supporting the current analysis but interpretation of the data may be complex. In a prior report from the Canadian Implantable Defibrillator study, Sheldon et al reported that sicker patients were more likely to benefit (Circulation 2000; 101:1660-1664). Other studies have shown that patients with lower ejection fractions were more likely to receive appropriate ICD therapy. However, ICD shocks are also a predictor of mortality and in some comorbidity subgroups in clinical trials (e.g., patients with significant renal disease), no benefits from ICD therapy have been described.
These observations present an interesting problem for clinicians. Sicker patients are indeed more likely to receive ICD shocks than healthier patients, but the individual survival benefit in the sickest patients is likely to be small. Healthier patients with fewer comorbidities are less likely to receive shocks, but the survival benefit in the individual patient who does receive a shock is likely to be great. Enrolling sick patients helps a clinical trial achieve a desired number of events quickly, but societal benefit may be greater per implant if less sick patients receive the device.
The data in this report were from ICD implants between 1997 and 2003. Most of the implants in this series were for secondary prevention. Currently, most ICD implants are for primary prevention with most primary prevention ICDs implanted in patients with very low ejection fractions or advanced heart failure. As shown here, heart failure is a major predictor of mortality. It is, therefore, likely that comorbidities would play an even more important role in the typical patient receiving an ICD today. This only makes it more important for physicians to consider the overall health status of the patient when making a decision about ICD implantation.