Successful Primary-Prevention ICD Shocks Present Mortality Paradox

December 13, 2010

December 13, 2010 (Toronto, Ontario) — It might follow that the more severe a case of cardiovascular disease, the greater the risk of potentially fatal ventricular arrhythmias, and so the more life is prolonged by terminating such arrhythmias with implantable cardioverter-defibrillator (ICD) shocks.

Not quite, as shown in a recently published post hoc analysis from a randomized trial that, years ago, made a big impact on device-therapy guidelines [1].

In the Defibrillator in Acute Myocardial Infarction Trial (DINAMIT), which randomized high-risk patients only just recovering from an acute MI to receive or not receive ICDs for primary prevention of sudden death, those with devices suffered fewer arrhythmic deaths but showed significantly greater all-cause mortality than those without devices.

Moreover, their risk of death from any cause climbed more sharply once their devices delivered appropriate therapy, either shocks or antitachycardia pacing.

There's kind of a sweet spot of patients who are just sick enough, but not too sick, in whom the device will give you the maximum bang for your buck.

"In other words, although the device probably saved their lives at the time of therapy, it simply postponed their deaths for a short period of time," write the authors, led by Dr Paul Dorian (St Michael’s Hospital, Toronto, ON).

"The devices are effective in everybody," Dorian observed for heartwire . They can keep serious ventricular arrhythmias from being fatal. It's just that not everyone with indications for a primary-prevention device will ever develop such an arrhythmia that the ICD can terminate.

"There's a catch-22 that we have to confront and cannot get around. If patients are healthy enough to survive for a long time, they're unlikely to need their defibrillator. If they're supersick, they have more need for their defibrillator but are also more likely to die from other reasons," he said.

"There's kind of a sweet spot of patients who are just sick enough, but not too sick, in whom the device will give you the maximum bang for your buck."

Patients in DINAMIT, Dorian said, "in general were a group that was too sick to benefit much from a defibrillator."

The group's analysis was published online December 6, 2010 in Circulation and was presented in part six years ago at the American Heart Association 2004 Scientific Sessions and covered then by heartwire . (Dorian said there's no dramatic explanation for the six-year delay to publication: "We just got overwhelmed with work.")

The main problem is that the disease that gave you the shock can also kill you.

DINAMIT randomized 674 patients with an LVEF <35% and depressed heart-rate variability who were only about one to three weeks past an acute MI and on optimal medications; NYHA class 4 patients were excluded. The trial's outcomes, which were confirmed by the later IRIS randomized study, were instrumental in the omission of patients <40 days post-MI from the guidelines' class-1 primary-prevention ICD recommendation.

In the trial's primary analysis, there was no difference between ICD recipients and nonrecipients in all-cause mortality over a mean follow-up of 2.5 years.

In the current analysis, several markers that significantly increased the risk of both arrhythmic and nonarrhythmic mortality emerged: age >60 years, baseline QRS duration >120 ms, baseline nonsustained VT, hypertension therapy, and a history of MI before the most recent MI.

Adjustment for those predictors yielded an arrhythmic-death hazard ratio of 0.33 (95% CI 0.15–0.71) but a nonarrhythmic-death HR of 1.70 (95% CI 1.00–2.80) for ICD therapy vs no ICD therapy.

The all-cause and nonarrhythmic mortality risks climbed sharply for ICD recipients whose devices delivered appropriate therapy at some point in the trial compared with device patients who never received appropriate therapy.

Relationship Between Delivered ICD Therapy and Mortalitya in DINAMIT, Hazard Ratio (95% CI)

ICD subgroup All-cause mortality, HR (95% CI) Nonarrhythmic death, HR (95% CI)
Appropriate shock vs no appropriate shock 4.9 (2.4–10.2) 4.8 (2.2–10.6)
Any appropriate therapyb vs no appropriate therapy 5.0 (2.6–9.6) 4.3 (2.1–8.9)

a. Adjusted for age >60 years, baseline QRS duration, baseline presence of nonsustained VT, hypertension therapy, and history of MI prior to index MI

b. Any appropriate therapy includes both shocks and antitachycardia pacing

The greatest mortality increases were in patients who received any kind of shock, whether appropriate or inappropriate, vs those who didn't receive either kind of shock: HR 5.6 (95% CI 2.4–10.0) for all-cause mortality and 6.0 (95% CI 2.8–12.7) for nonarrhythmic mortality.

Dorian's preferred explanation for why appropriate therapy--and even inappropriate therapy--appears to identify patients with increased risk of dying from causes other than arrhythmia? Those who get such therapy are sicker in general than those who don't. It's probably not, to any large degree, that the shocks themselves are severely harming the heart, to cite another proposed explanation, he notes.

Based on trial evidence and on experience, Dorian said, "I would infer that shocks are psychologically undesirable and probably cause some degree of minor heart damage, but it's not the main problem. The main problem is that the disease that gave you the shock can also kill you."

In DINAMIT, those who received appropriate shocks had "substantially more episodes of heart failure, MI, and unstable angina during follow-up," the report notes.

To dismiss the link between shocked ventricular tachyarrhythmias and death as an inevitable consequence of a patient doomed to die is to dismiss the possibility of improving survival.

"We don't deny that maybe the shock is doing some additional damage," Dorian said. In some cases, he continued, that damage might be the final blow that kills, but it's not the only thing that caused death. If that were true, it would show in patients with electrical storm who may be hit with dozens of shocks in a day. "Their mortality may be slightly higher than patients who don't have inappropriate shocks, but there's no dramatic effect."

In an accompanying editorial [2], Dr Michael O Sweeney (Brigham and Women’s Hospital, Boston, MA) notes that the approximately fivefold increased risk of death after delivered ICD therapy in DINAMIT is consistent with similar increases seen in the primary-prevention trials MADIT 2 and SCD-HeFT.

Dorian et al seem to favor the most common view of the relationship between delivered ICD shocks and increased mortality in the different trials, according to Sweeney, that "patients with ventricular tachyarrhythmias and shocks are at higher risk for dying, and the former is a marker for, but mechanistically unrelated to, the latter. . . . This default argument persists because it is the easiest to make and always within reach." But, he writes, "It excludes the possibility for understanding alternative and more disruptive explanations, which might lead to improved patient survival."

Sweeney acknowledges that the relationship between shocks and mortality isn't known but describes evidence suggesting it could be causative, at least partially.

"Understanding whether high-voltage shocks delivered into the failing heart preconditioned by ventricular tachyarrhythmia have an adverse effect on mortality is the most important question facing ICD therapy delivery and development. To dismiss the link between shocked ventricular tachyarrhythmias and death as an inevitable consequence of a patient doomed to die is to dismiss the possibility of improving survival."

Dorian lauded Sweeney's editorial and said he "completely agrees" with much of what it says, and "I'm inclined to believe that shocks in and of themselves are a problem, but less of a problem than Dr Sweeney speculated."

The analysis was supported by a grant-in-aid from St Jude Medical. Dorian discloses receiving research grants from Boston Scientific and St Jude Medical and honoraria from and having served on an advisory board for St Jude Medical. Disclosures for the other coauthors are listed in the paper. Sweeney had no disclosures.


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.