"We now tell a patient that if we put an ICD in 100 patients with heart disease like yours, over the next five years we would expect 30 patients will die anyway, seven to eight will be saved by the ICD, 10-20 will have a shock when they don't need a shock, five to 15 will have other complications, and the rest of the patients will never experience a shock from their devices."Could this be true? Have we oversold these devices?
The crux of the argument is the difference between “relative risk reduction” and “absolute risk reduction.” “Relative” to patients without an ICD implanted, there were 23% fewer deaths compared to the non-ICD group in the SCD-HeFT trial, on which she based the above data. In “absolute” terms, if 100 patients were implanted, 7.1 fewer deaths occurred in the ICD-treated group than in the non-ICD treated group, all other things being equal.
Now, you’re the one with the crummy heart. A doctor leans forward and gives you the data. Which would you choose?
But to take this a bit further, Dr Stevenson added those other more questionable tidbits of data – I especially like the one that says “10-20 will have a shock when they don't need a shock.” Data regarding appropriate, life saving shocks were not mentioned. Why not? Here’s the REAL data from the New England Journal of Medicine publication:
829 patients received a single-lead ICD in the SCD-HeFT trial. From the NEJM paper:
Of the 829 patients in the ICD group, 259 (31 percent) were known to have received shocks from their device for any cause, with 177 (68 percent of the shocks, or 21 percent of the ICD group) receiving shocks for rapid ventricular tachycardia or fibrillation. During five years of follow-up, the average annual rate of ICD shocks was 7.5 percent. For appropriate shocks only (i.e., shocks for rapid, sustained ventricular tachycardia or fibrillation), the average annual rate of ICD shocks was 5.1 percent.So as I see it, using Dr. Stevenson’s analogy: of 100 patients, 7 to 8 patients will receive a shock each year with 5 or 6 of these patients receiving an appropriate shock, while 2 or 3 will have an inappropriate shock. Said another way, 2-3 patients over 5 years gives us about 10-15 patients getting inappropriate shocks, while 25 to 30 receive appropriate, potentially life-saving shocks over the same time period. The argument for an ICD sounds a bit stronger when you give both sides of the argument, doesn’t it?
The medical device industry is very concerned about the drop in ICD implantation rates recently. Industry is worried the recall fiasco is what caused their ICD implantation rates to plummet. I suggest this might not be the main reason. Instead, I think Dr. Stevenson has struck a cord regarding the real reason ICD implant rates have fallen (remember, Dr. Stephenson is a referring doctor). Referring doctors now know that ICD’s are not infallible, certainly, but more importantly, they realize the number of their patients that would have to be implanted to save one life: 12 to 15.
Is this too many? Too much cost? You decide.
It is what it is. But to avoid the issue in asymptomatic patients with weak heart muscles might result in tragedies, too. Coaches Maggie Dixon and Randy Walker come to mind. They looked good didn’t they? Hell, they were coaching.
This is tough to swallow, but given the effectiveness of ICD’s over drug therapy, we’d be fools to ignore the benefit of ICDs to our patients, all things considered. But certainly, being informed about their upsides and downsides in a balanced manner will go a long way toward helping our patients make this tough, life-long decision.
Now the tough question: if your patient can't decide, what would you recommend?