Trying to understand ICDs

Hey all,

A couple of months ago I got a call from Dr. Rich Fogel, an electrophysiologist from St. Vincent’s Hospital in Indiana. He is an old friend from training days; I used to moonlight for his group. Now, Rich is on the leadership track in the Heart Rhythm Society and this year he is in charge of the HRS meeting in Denver. After reading a post I did on ICDs, he asked if I would be willing to give a Hyde Park Lecture at HRS.

I was stunned. These 15 minute talks are TED-like presentations in the convention hall. Just me, my ramblings and a microphone.

In a weak moment I agreed. We settled on the topic of goals of care and end of life issues with ICDs. His thesis is that few patients get talked to about their goals of care before ICD implantation.

Since I have been implanting ICDs for nearly two decades and writing a blog about heart rhythm issues for a few years, it might seem an easy thing to get up there and talk for 15 minutes. Not so much. You see, I have strong feelings about ICDs, and death, and harm, and shared-decision making. My feelings don’t necessarily parallel the majority of Heart Rhythm Society members. And another thing about these talks: they are usually given by distinguished professors who write in medical journals, not regular doctors who scribble graffiti-with-punctuation in blogs. (That’s my son’s take of blogs.)

Ultimately though, the major problem is that using ICDs to prevent sudden cardiac death (SCD) and extend life (not always the same thing) is an extremely complicated topic.

Here are just a couple of teasers on the complexity:

• Heart disease remains the number one killer, and SCD is the number one mode of cardiac death.
• Most patients who suffer SCD have no known heart disease.
• Those with heart disease are at increased risk of SCD, but they are also at risk of dying from other causes. ICDs only reduce death due to arrhythmia, so competing causes of death are a huge problem.
• Risk stratification (aka…determining which patients with heart disease will go on to die suddenly) is woefully inaccurate. More than 13 of 14 patients implanted with an ICD never use it. That means we are wrong many more times than we are right about implanting ICDs.
• We have to expose many patients to the burdens and risks of an ICD to save one life. But make no mistake, all cardiologists have patients alive today because of the ICD. That patient is someone’s dad, mom, brother, sister, wife, husband. You get the picture.
• The clinical trials that established the effectiveness of ICDs were deeply flawed. I’ll go into this in future writings, but the industry-sponsored trials of the early 2000s have been taken as gospel. After a decade, and with further analysis, the magnitude of ICD benefit has probably been greatly over-estimated while the risks have been minimized.
• There are gender-related issues. Women were greatly under-represented in the trials and post-hoc studies have failed to show a convincing benefit in women. For instance, it is well established that women with ICDs get fewer therapies than men, even though they have the same mortality rates.
• Age is also a factor. Our population is aging. As such, patients considered for ICDs present with more co-existing diseases. The clinical trials of ICDs enrolled mostly young and healthy patients. The benefit of an ICD accrues over years, so older patients may not get the same benefit as younger patients. We also know older patients have a higher risk of complications from the device.
• Then there is the matter of talking about death. That’s right, when a patient is referred for an ICD, we must talk about death. This is what the device does after all. But patients who come to the heart doctor for an evaluation are looking for a report. It’s hard for the electrophysiologist to be the first ever doctor to tell them they are actually going to die. “No one said I am going to die?” “I’m doing that bad?” “I feel good?” It’s amazing what patients take back to their primary care doctor or family: “Dr Mandrola said I was too old and sick for an ICD. He said I was going to die.” Remember, doctors have to see more than one patient a morning, so we have to get these conversations done in minutes, not hours.
• Death is a tough issue. Everyone inherently knows they will die, but few like to think about it. And there are two modes of death: sudden and painless (He went to take a nap and never woke up.) and anything but. The biggest risk of the ICD is the shift in mode of death. Yes, it may extend life, but the cost is hospitalizations, sickness, loss of autonomy, pain and many other things that don’t occur when one dies in their sleep suddenly. Try getting proceduralists to talk with patients about this one. I have been surprised at the views of patients and doctors on the matter of life and death. I’ll tell you more in future posts.
• Cost enters the equation. ICDs are expensive. Though this may not be a direct issue in the exam room with a single patient, it is an important overriding consideration. Because of the tremendous impact that ICDs have on the healthcare budget, government agencies have determined implant criteria for which they will pay. These criteria, called the NCD (National Coverage Decision), are inviolate (because nearly no patient can afford the cost of an ICD.) That the NCD was taken directly from the flawed clinical trials, and, that there has been no update in many years, puts patients and doctors in a tough spot. Strict adherence to guidelines and mandates minimizes the component of nuance and judgment in decision-making. That’s a huge negative. The word rationing comes to mind. But a derivative of rationing is rational. If we used ICDs more rationally, perhaps there would be less need for rationing. See. Complicated.

Anyways, I am in severe learning mode. The talk is only 15 minutes, but it is in front of an audience of highly educated heart rhythm specialists.

I’ll keep you posted, but don’t be surprised if we talk about devices, decisions and death in the coming month.

JMM