ICD Decision Making: An Opportunity for Stewardship

John Mandrola


April 14, 2014

Two studies available online from the Journal of the American College of Cardiology shed light on the nuances of ICD decision making[1,2]. The provocative findings urge cardiologists to think hard about our role as stewards in the use of the ICD. One study evaluated the appropriateness of ICD-generator change and the other the utility of the six-minute-walk test in predicting ICD benefit.

These were beautiful studies in that they emphasized the treatment of the person rather than his or her diseases. And this notion—that doctors treat people, not diseases—is everything when it comes to using ICDs.

ICD Generator Change vs De Novo Implant:

Many factors combine to make the decision to replace an ICD different from the original implant decision. First, intuitive thinking, which holds that generator change should be less risky than de novo implants, is wrong. Complications from generator-change surgery are strikingly high [3,4,5] . Second, time changes the patient. People age. They acquire chronic diseases that compete as causes of death and therefore mute ICD benefit. Third, patients change their approach to life and death. Being closer to the inevitable pushes some to consider the notion of a good death vs a bad one. The experience of having an ICD may also alter one's view of the device. Experiencing wide-awake shocks, for instance, might change how one thinks about having another device. Finally, a patient's cardiac condition may have improved such that the relative risk of sudden arrhythmic death is diminished.

These facts transform the decision to replace an ICD generator into a major opportunity for stewardship. Does the patient with ICD-battery depletion still garner greater benefit than risk from the device? All the major guidelines, including the immutable National Coverage Decision (NCD), make no distinction between de novo ICD implant and generator change. The idea that ICD therapy equates to lifelong therapy is not supported by any major consensus document.

University of Pennsylvania Study:

In the first study, researchers performed a retrospective chart review of 231 patients who had undergone replacement of primary prevention ICDs at two VA Medical Centers.

They report four main findings:

  • The prevalence of chronic diseases, such as chronic kidney disease, high blood pressure, neoplasm, and diabetes increased since the original implant.

  • Of 231 patients, 59 (26%) no longer met ICD indications (improved EF and no therapy) while 79 (34%) had unclear indications (LV function had not been reassessed before change).

  • In 3.5 years of follow-up, patients who met indications at the time of generator change were more likely to have received tachycardia therapy than those who did not (10.7% vs 2.8%; p<0.001).

  • The presence of an EF in the range of 30% to 35% at the time of first implant was an independent predictor of not meeting ICD indications at the time of generator change.

In the discussion section, the authors made three major points. First, they emphasized that guideline statements may be correct in not distinguishing a generator change from a new implant. They pointed to the very low shock rates in nonindicated patients and the accumulation of chronic diseases in patients with elective replacement indications (ERI). They acknowledged previous data suggesting persistent arrhythmia risk even when EF improves but noted serious flaws in those studies. Second, the authors made the case to reassess LV function before generator change, especially in those patients who had EFs in 30% to 35% range at original implant. Third, they noted that CRT response associates with lower arrhythmia risk.

They concluded with this bold statement (I recommend reading it slowly):

"[P  rior] studies, as well as our observations, make a case for performing ICD explant instead of generator replacement in patients who experience no appropriate therapies and show significant improvement of LVEF when their devices reach ERI indications. In the cases where improvement of LVEF has occurred with the original device being CRT-D, a CRT-P device could be used instead of CRT-D for replacement."

Functional Capacity and ICD Benefit:

The second study asked whether the ability to walk with a modicum of vigor predicts ICD benefit. Here, a group of researchers analyzed data from the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) to determine if the six-minute-walk distance (6MWD) assists in treatment decisions for patients with heart failure. The role of functional capacity in determining ICD benefit is relevant because of the often-forgotten finding from SCD-HeFT: patients with NYHA class 3 symptoms got no benefit from the ICD. This prespecified analysis was left out of the guidelines because of concerns over the qualitative nature of NYHA functional class. Assessing functional class with the more quantitative six-minute walk overcomes this weakness.

The researchers separated the 2397 patients into tertiles of 6MWDs. Their first finding highlighted the flaws of using NYHA functional class: only half the patients classified as class 3 were in the lowest tertile for 6MWD, while 13% of these supposedly "advanced" heart-failure patients were in the highest tertile.

The main finding, however, was that patients who walked less than 288 m in six minutes (lower tertile) did not benefit from ICD therapy while those with 6MWD greater than 288 m (middle and upper tertiles) benefited substantially.

The authors' conclusion was clear and strong:

"T  hese results confirm and extend the previously reported finding that ICD therapy did not benefit NYHA class 3 patients in the SCD-HeFT trial. These data should be included in shared decision making with patients for whom they apply."

My Thoughts:

It has been said that the best surgeons know when not to operate. The idea, of course, is that judgment is harder to learn than cutting and sewing. The same thinking applies to doctors who implant ICDs. I believe the most difficult aspect of ICD therapy is the (shared) decision to implant the device.

Taken together, these two studies, along with many published risk-stratification algorithms[6,7,8,9,10], focus attention on that narrow notch (sweet spot) of patients who will benefit from the ICD.

Shared decision-making requires caregivers to present information relevant to the patient—the whole person—who sits before us. That person may have an ICD with battery depletion. The study from the University of Pennsylvania urges us to consider the situation now, not then. The SCD-HeFT investigators show us that walking ability, an obvious but overlooked metric of health, predicts ICD benefit. Of course it does.

Real-world doctors recognize the challenge. Like this: I recently encountered a patient with an ICD at ERI, who, in the years since the original primary prevention implant developed significant noncardiac diseases and an improved EF (45%). His goals of care remained life-prolonging—although he continued to smoke. This wasn't going to be easy, I thought. So I paused for a moment, took a breath, mustered my most compassionate tone and began explaining. I tried to communicate that his mortality with a new ICD was the same as without it, and, if he presented in this same situation now, without his ICD, a new one would not be offered. I hadn't even got to the notion of a bad death when he retorted with this zinger: "The government just wants people like me to go off and die."

Ouch. Nobody said doctoring was going to be easy. Yet, in this era of increasingly aggressive low-value/high-cost treatments, it has never been more vital for cardiologists to be stewards. This will require both smarts and courage.

These two studies remind us what we are trying to do with the ICD. The device is not for prolonging death. It is for extending life. And these are not the same—whether it's a new implant or a generator change.



Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.