ICD implantation: Knowing when to hold'em

Ghosh and associates studied 2467 implantable cardioverter-defibrillator (ICD) recipients in Ontario, Canada.¹ The authors examined clinical factors associated with mortality and concluded that advanced age, heart failure, and other comorbidities predicted greater mortality in this population. This study is important because it examines a “real-world” population, not just patients enrolled in clinical trials.

One of the most important factors we encounter in clinical practice is the increasing age of our patients. Grimm and associates evaluated 500 ICD recipients, including those younger and older than 75 years of age.² Both cohorts had equivalent rates of appropriate ICD therapies; thus, the rate of sudden death was low in both groups. Nevertheless, overall mortality was significantly higher in the older patients because of a higher mortality from heart failure. Likewise, in a review of 1866 elderly patients enrolled in secondary prevention trials, ICD placement reduced all-cause mortality in younger patients, but not in older ones, possibly because of a higher incidence of nonarrhythmic death in older patients.³

In contrast, a study of 1232 patients enrolled in MADIT-II (Multicenter Automatic Defibrillator Implantation Trial II), a primary prevention trial, showed that ICD patients over the age of 75 years enjoyed a reduction in mortality.⁴ The authors concluded that ICD therapy should not be withheld based on age alone. A trial of 250 consecutive primary and secondary prevention patients showed similar ICD treatment frequency in younger patients versus those averaging 79 years.⁵ This study supports the view that older patients may benefit from ICD implantation.

Comorbidities play an important role, regardless of age. In a study of 80 secondary prevention patients, predictors of time to death included age, creatinine ≥1.3 mg/dL, and QRS >120 ms.⁶ The presence of more than 1 risk factor was associated with a higher mortality. Goldenberg and associates published a substudy of MADIT-II, in which a risk score was applied.⁷ The risk factors included New York Heart Association functional class II heart failure, age >70 years, elevated blood urea nitrogen (BUN), QRS prolongation, and atrial fibrillation. In addition, a very high-risk group was defined as having a BUN >50 mg/dL or a creatinine >2.5 mg/dL. Mortality was lowest in patients without risk factors and highest in those with renal dysfunction; however, no benefit was seen with ICD placement in patients without risk factors or in those with the highest mortality risk. This U-shaped curve defined patients at both ends of the risk spectrum who may derive less benefit from an ICD. The authors suggest that a simple clinical risk score that includes age, heart failure class, BUN, QRS duration, and the presence of atrial fibrillation can help define patients who may not benefit from ICD placement.

We have seen a dramatic increase in use of ICDs along with expanded indications for these devices. Practicing physicians are charged with the difficult task of applying clinical trial data to “real-world” patients. Studies like that conducted by Ghosh and associates help guide clinical judgment by improving our understanding of when ICD implantation won’t improve survival; knowing when to hold ’em may be the most difficult clinical decision we can make.