Medical Device Daily Washington Editor
WASHINGTON — Sales of implantable cardioverter defibrillators (ICD) probably topped out in 2006, but a couple of the discussions at the 2007 edition of the Heart Failure Society of America (HFSA; St. Paul, Minnesota) annual meeting suggested that doctors may be growing even more selective about which patients should get these and other electrophysiology devices, in part due to skepticism as to whether clinical trial populations truly mirror the broader patient population.
At least one practitioner previously interviewed earlier this year by Medical Device Daily said that he expects to see as much as a 25% drop in the volume of ICDs he will have implanted in 2007 over the previous year (Medical Device Daily, April 26, 2007).
Justin Ezekowitz, director of the heart function clinic at the University of Alberta (Edmonton, Alberta), discussed his views on the role of devices in heart failure, starting off by reminding the audience that “even on the best medical therapy, left ventricular systolic dysfunction [LVSD] has a high burden of morbidity and mortality.”
On the other hand, he said, he is among those who are of the opinion that while internal defibrillation does reduce mortality and hospitalization, the use of ICDs in ordinary practice does not always parallel their use in trials.
Ezekowitz briefly noted an article he penned for the Annals of Internal Medicine in 2003 in which two reviewers looked over the results of eight trials that collectively included more than 4,900 patients. That analysis pegged the number of deaths at slightly more than 1,150, but “compared with usual care ... ICDs significantly reduced sudden cardiac death ... and all-cause mortality.” Still, even in 2003, Ezekowitz was convinced that the impact of ICDs “on total mortality is sensitive to baseline risk for arrhythmic death.”
Ezekowitz also discussed a meta-analysis he and several other researchers performed for the Agency for Healthcare Research and Quality (AHRQ; Washington) on a series of 11,000 articles appearing in the literature on the use of cardiac resynchronization therapy (CRT, a.k.a., biventricular pacing), which treats arrhythmias in both ventricles, in conjunction with ICDs in treatment of LVSD. He acknowledged that his conclusions might be controversial, perhaps in part a nod to some of the controversies surrounding the uses of meta-analyses in other medical device applications.
According to the report, which AHRQ published in June, “ICD and CRT reduce all-cause mortality in patients with LVSD” that qualified for randomization, but the authors nonetheless said “the incremental benefit of CRT plus ICD over CRT alone in patients with LVSD remains uncertain.”
The analysis is said to have detected no difference in sub-groups in terms of the efficacy of CRT or ICD, but the paper says that “subgroup analyses and meta-regression using aggregate trial data are post-hoc analyses and were underpowered to detect such effects.” Hence, the conclusion that an analysis will have to be done on data from the individual trials to establish which, if any, “clinical subgroups are most likely to benefit from these devices.”
Ezekowitz said that good news tends to be over-reported. “There tends to be publication bias because you want to report something good for your institution,” he said, adding that evaluation of community use “is better than in a randomized clinical trial” in giving a real-world picture of the impact of electrophysiology implants. Still, he saw most follow-ups as “too short to evaluate how someone will live with an ICD.”
As a comparison of the two electrophysiology functions, Ezekowitz said that ICDs reduce mortality in patients with New York Heart Association Class II and III disease with a left ventricular ejection fraction of 35% or less, but that it takes 20 patients a total of 35 months to prevent one death with a moderate safety risk. CRT used in patients with NYHA Class III/IV disease will prevent one death in 23 patients after six months, but the disease set is more severe than that he described for ICDs.
Ezekowitz recommended that patient experiences with CRT be included on a mandatory basis in the national ICD registry, the National Cardiovascular Data Registry, which is maintained by the American College of Cardiology (Washington).
As for safety, Ezekowitz said he had seen data from implantations of ICDs into 6,300 patients in France that demonstrated a risk of infection of about .01%, “which is good,” but he made it clear that experience is a strong predictor. “Those who are putting in less than 30 devices a year” are having more infections and device complications, he said, adding that “if you don’t do it a lot, you’re not going to be able to do it very well.”
Regarding combination therapy with ICDs and CRT units, he cited another paper he authored, this one appearing in the June 13 edition of the Journal of the American Medical Association, that was inconclusive on the question of whether the addition of an ICD to a patient already implanted for CRT provides any meaningful benefit, urging, of course, further studies.
“Inappropriate shocks are still a problem” with ICDs, Ezekowitz said, alluding to a gap in knowledge of which patients with which pathologies should get which devices and combinations of devices and device functions. “We really need to understand where form meets function,” he said.
Ezekowitz did not advocate that the use of electrophysiology equipment be extended to a substantial portion of patients with NYHA Class I disease, but did say that “there is some unmet need” in this population. “We don’t want to become robots blindly following the guidelines,” he said in urging members of the audience to remain open to the idea.
As for the necessary ingredient of risk stratification in determining which patients really will benefit from these devices, Ezekowitz said he did not see microvolt T-wave alternans readings as solid enough evidence without other information. T-wave data, he said,” “is good at risk stratifying, but I wouldn’t use it in isolation.”