News & Views | ICD use in primary prevention: Do we really know enough?
 |
A study of more than 111,000 patients by Al-Khatib et al found that the recipients of non-evidence-based implants had an increased risk of post-procedure complications and in-hospital death, were older, had more comorbidities, more likely to belong to a racial minority group other than black and more likely to receive a dual-chamber ICD (JAMA 2011;305[1]:43-49). The reasons the implants were deemed non-evidence-based were: newly diagnosed heart failure (62.1 percent), within 40 days of an MI (36.8 percent), NYHA class IV symptoms (12 percent) and within three months of CABG surgery (3.2 percent).
“While many of the 22.5 percent of patients shouldn’t have received defibrillators because they were not indicated, that figure might be on the high side for various reasons,” says Alan H. Kadish, MD, of Northwestern University Feinberg School of Medicine in Chicago. For example, the large group of patients indicated as newly diagnosed with heart failure could be inaccurate, as registry data are typically retrospectively abstracted and the precise onset of congestive heart failure might require more history.
Also, patients often fall into a "gray area" where some data indicate an ICD would be helpful, but not enough to make a firm recommendation. "A reasonable allowable percentage of non-evidence-based implants could be 6 percent," he suggests, which would take into consideration physician judgment.
DINAMIT dynamite
Adding to the controversy, results from the DINAMIT study showed that a reduction in sudden death in ICD patients was completely offset by an increase in non-arrhythmic deaths, which were greatest in patients receiving ICD shock therapy (Circulation 2010;122:2645-2652). Dorian et al randomized outpatients with recent (six to 40 days) acute MI, left ventricular dysfunction (ejection fraction less than 35 percent) and low heart rate variability to ICD (311 patients) or standard medical therapy (342 patients).Michael O. Sweeney, MD, of Brigham and Women's Hospital in Boston, says that more than 35 years of investigation indicate that high-voltage shocks cause cardiac tissue damage, though the translation of these effects to adverse clinical outcomes is uncertain. “Collectively, we need to face the possibility that mortality risk could be changed immediately following a mortality dividend from a high-voltage shock. This is a disruptive concept for physicians because it is difficult to confront the possibility that a powerful therapy that is perceived as highly effective—especially one with no alternative—is potentially consequential to patient outcomes.”
Sweeney acknowledges that there are many patients who could benefit from an ICD, even peri-infarct patients, but providers simply need to have a better understanding of the patients who won’t benefit. He adds that the results from DINAMIT and IRIS could have been widely different, if the investigators had applied contemporary shock-reduction strategies. He is calling for new clinical trials to investigate these issues, which also may contribute to understanding differences in outcomes among guideline-compliant and out-of-guideline ICD patients.
While the potential adverse outcomes with utilizing ICDs in primary prevention will undoubtedly continue to be debated, there are recommendations to improve current implantation techniques. In fact, Kadish suggests that physician education is integral to knowing and following the guidelines (variability between hospitals ranged from zero to 40 percent), and patient management is best undertaken by an electrophysiologist (implantations by electrophysiologists resulted in fewer complications and deaths compared with nonelectrophysiologist cardiologists and other surgeons).