Growing pains: HF, pacing specialists ponder pitfalls of CRT's rapid acceptance

April 22, 2004

Thu, 22 Apr 2004 21:00:00

New Orleans, LA and San Diego, CA - Although cardiac resynchronization therapy (CRT) is of undisputed value for some patients with HF, it still has significant limitations that warrant some caution, given the technique's rapid acceptance and currently broad application, write two noted HF specialists in a recent editorial.

Dr Mandeep R Mehra (Source: Ochsner Clinic Foundation)

A handful of randomized trials that support the use of CRT in selected patients with HF, including MIRACLE-ICD and the as-yet unpublished COMPANION trial, form the underpinnings for its current surge in use, observed Dr Mandeep R Mehra (Ochsner Clinic Foundation, New Orleans, LA) and Dr Barry H Greenberg (University of California, San Diego). But they warned of potential dangers from embracing CRT with enthusiasm that may be out of proportion to the conclusiveness of those trials. Their "viewpoint" column appeared in the April 7, 2004 issue of the Journal of the American College of Cardiology.

Limitations of the various CRT clinical trials to date, write Mehra and Greenberg, have included a "substantial" placebo effect in control patients, lack of true intention-to-treat designs, and "a remarkable heterogeneity of response" among trials, including a high rate of nonresponding patients in active-therapy groups.


The purpose of this viewpoint was to distill some of the potential pitfalls in absorbing this evidence too quickly and translating it poorly to the clinical realm.


"Resynchronization therapy is an extraordinarily important advance in heart failure therapeutics, an avenue of therapy that provides seemingly incremental benefit beyond what is provided by the traditional neurohormonal model," Mehra said to heartwire . "The purpose of this viewpoint was to distill some of the potential pitfalls in absorbing this evidence too quickly and translating it poorly to the clinical realm, which could impede our progress in this field."

Those pitfalls, according to Mehra, include:

The premature extension of CRT to patients in NYHA functional class 2, who were generally excluded from the relevant trials.

The use of CRT before "giving drug therapy a chance to work," he said. "One of the ways patients with advanced heart failure reach that stage is that we don't apply appropriate evidence-based pharmacologic therapy."


It's very difficult to tell up front who is going to benefit and who is not.


CRT devices are widely implanted "without the creation of standards for who should implant them and how they should be implanted," Mehra observed. Nor has a prospective system been established for documenting outcomes, as has been done for many other devices. "Whether we are applying them to the right populations and whether we are replicating the results from clinical trials are impossible to ascertain unless we transmit data to a central repository," he said. "I do believe that the burden of developing these repositories should rest to some extent with industry."

Other experts contacted by heartwire were in variable agreement with the Mehra- Greenberg viewpoint.

Dr Marvin A Konstam (Source: New England Medical Center)

According to Dr Marvin A Konstam (New England Medical Center, Boston, MA), CRT "should be considered in significantly symptomatic patients who meet the entry criteria for the trials that have been performed." But, he added, "I think we have to curb our enthusiasm, so to speak, on how much benefit we can anticipate for any one patient."

"I didn't see a single thing in the paper that I didn't agree with," Dr James A Hill (University of Florida, Gainesville) said. With CRT, "you put out a substantial amount of cost with a good bit of faith that it's going to help. In my own practice, I've seen a number of people who've done dramatically well and a number who haven't had much benefit at all. And frankly, it's very difficult to tell up front who is going to benefit and who is not."

Dr James A Hill (Source: University of Florida)

Dr David L Hayes (Mayo Clinic, Rochester, MN) said the issues raised in the viewpoint are already being discussed in the pacing community. "We do need to improve patient selection, and we need proof that whatever new selection techniques we decide are best truly lower the number of nonresponders. There are technical limitations, and we need to overcome those." But he is more bullish on what the current CRT state of the art portends. The Mehra and Greenberg viewpoint, rather, uses a "glass-half-empty" rather than a "glass-half-full" perspective, Hayes said.

Dr David L Hayes (Source: Mayo Clinic)

The CRT clinical trials to date have been convincing, according to Dr Christopher M O'Connor (Duke University, Durham, NC). COMPANION "reflects an almost real-world exposure in that there was a lot of variation in ability to deploy the device," yet it still showed an advantage for CRT, he said. "That suggests the results were probably at the lower end of the benefit. The benefit could actually be larger as we gain more experience." As for some of the concerns raised by the editorialists, O'Connor said, "I think they come down a little hard on the side of caution."

Dr Christopher M O'Connor (Source: Duke University)

Limited supporting data

CRT's clinical-trial database, consisting mostly of patients with severe systolic HF, suggests the technique improves functional capacity, reduces hospitalization, and may reverse ventricular remodeling, the viewpoint acknowledges. But none has shown CRT to reduce mortality.

"So far, we know that there is a trend toward decreased mortality in most of the clinical trials, an approximately 23% relative reduction in mortality that is not statistically significant," Mehra told heartwire . "Most of the benefit . . . comes from reduction in hospitalization rates, and the clearest evidence comes from the COMPANION trial." That trial was not blinded and has yet even to be published, he noted.


Most of the benefit . . . comes from reduction in hospitalization rates.


Moreover, the extra nursing and other forms of attention patients receive in clinical trials by itself can dramatically lower hospitalization rates. "There's no way to quantitate what the effect of that closer follow-up will be on hospitalization in an unblinded trial. That's why the only way of showing this therapy is robustly effective is by having a concurrent signal on the mortality end point."

Trial interpretation can be further confused by the substantial placebo effect observed in CRT trials. "Up to half of the patients in some of the trials, in the device-turned-off groups, actually felt better by one NYHA functional class," Mehra observed.

Dr Angelo Auricchio (Source: University Hospital)

It depends on the end points, observed Dr Angelo Auricchio (University Hospital, Magdeburg, Germany). "The authors focused on symptoms and walking distance, which are weak parameters and very sensitive to a placebo effect." He said oxygen consumption at maximal exercise testing is "a much more reliable parameter," for which the placebo effect is minor.

Auricchio, who played a role in the COMPANION trial's design phase, said that study was not among those that may have lacked an intention-to-treat design. Whereas some CRT trials randomized patients only after successful device implantation, Auricchio noted that COMPANION's patients were randomized before the surgery.

O'Connor also took exception to the viewpoint's reservations about the CRT clinical trials, particularly COMPANION and its end points: "Nobody can argue with all-cause mortality/hospitalization. The results were consistent, they were robust, they make sense. It's pretty much a home run as far as trials go."


The results were consistent, they were robust, they make sense. It's pretty much a home run as far as trials go.

Response rates vary widely

Mehra and Greenberg point to the wide heterogeneity of response rates in the CRT trials to date. "What we call the placebo-subtracted response rate, meaning the response rate attributable to the device alone, ranges anywhere from a 15% to 25% benefit, depending on the study," said Mehra when interviewed. "Removing this heterogeneity is critical for the advancement of the field."

Konstam observed that the average treatment effect that emerges from clinical trials casting a wide enrollment net inherently will not show up in every patient who meets the trial's criteria. Hence, under current guidelines based on the trials' entry criteria, it is certain that some patients will not benefit from CRT but will be exposed to the treatments risks. "I don't think we know nearly enough about identifying a priori patients who will benefit and patients who will not benefit," he said.

"The clinical trials to date have, by and large, all met their end points and showed a benefit for CRT," Hayes emphasized. Still, the field is "in a real learning stage" with regard to patient selection.


I don't think we know nearly enough about identifying a priori patients who will benefit and patients who will not benefit.


Requiring CRT recipients to have a QRS duration >120 ms to indicate out-of-sync ventricles is another major stumbling point, the editorialists suggest. "While QRS lengthening probably means a sicker heart, it doesn't necessarily mean ventricular dyssynchrony," Mehra said. "There are studies that have suggested that about 75% of patients with a QRS greater than 120 ms will in fact have dyssynchronous contractility, which means one in four patients will not." Other studies, he added, suggest that about half of patients with a QRS of about 100 to 120 ms also have ventricular dyssynchrony.

"We may actually be missing patients who could benefit," Mehra said. "But by the same token, we may be misusing this treatment in as much as a quarter of patients who have QRS lengthening."

But the CRT trials included the QRS eligibility criteria, so for now, the evidence base is what it is. "At this point, you'd really be on shaky ground to offer the therapy in patients with normal QRS durations," Konstam said.


We may be misusing this treatment in as much as a quarter of patients who have QRS lengthening.


According to Hayes, QRS lengthening seems to predict ventricular dyssynchrony across populations, "so it's reasonable to use it as part of the criteria." What's needed, however, is a more direct method for assessing dyssynchrony, as "the QRS does not tell us the whole story in a large number of patients. We need some other technique, and it's going to be some form of echo Doppler."

Response rates also differ because of the clash between standardized percutaneous lead-insertion techniques and inherent variation in patients' myocardial electrical and mechanical abnormalities, the editorialists write.

For example, Mehra said, the device's third electrode is typically placed near the lateral wall of the left ventricle. "That's considered optimum placement because most dyssynchronous contractility can be abrogated from that location." However, the best lead position is likely to vary depending on the pattern of myocardial scarring or HF etiology. The optimum location for some patients, he said, "might be better reached by epicardial leads placed through a minimally invasive surgical technique."


That may be the sole reason we are seeing such dramatic heterogeneity in the response rate.


Also, inappropriate definitions for treatment success can cause inconsistent CRT success rates, Mehra observed. "These devices are inserted without the [reliable] detection of mechanical dyssynchrony and without the ascertainment of resynchronization at the end of the procedure." Usually, he said, "successful" therapy amounts to the patient "leaving the electrophysiology laboratory safe and sound and alive after insertion of these triple-lead pacemakers. . . . That may be the sole reason we are seeing such dramatic heterogeneity in the response rate." As alternative gauges of success, he proposes functional measures such as the six-minute walk or cardiopulmonary exercise tests, plus assessment of the degree of dyssynchrony correction and confirmation that lead placement correlates with the site of dyssynchrony.

Auricchio observed that some treatment success, by some measures, may not be evident by hospital discharge. For example, any benefits of reverse remodeling would not show up for many months.

Resist "indication creep"

"The urge to extrapolate potential for benefit to less morbidly ill populations exists but must be resisted," cautioned the editorial writers. Lack of a CRT registry database makes it impossible to determine whether CRT is being used consistently in class-3-4 patients, Mehra said to heartwire .

"There may in fact be a therapeutic window," Mehra added. "True class 4 might be too late for application of this therapy, and class 2 may be too early. The best application may be in the class-3 patient, in whom the risk/benefit ratio would be the most justifiable."


True class 4 might be too late for application of this therapy, and class 2 may be too early.


CRT should be limited to class 3-4 HF, O'Connor said. "On the other hand, it puts a lot on our ability to assess NYHA class." Among class-2-3 HF patients there are those at the margin for which the classification scheme is especially subjective. And, he noted, an HF patient may appear to fit class 3 one week and class 2 the next, and vice versa as the syndrome fluctuates. He would therefore put more weight on such quantifiable criteria as QRS duration and LVEF when selecting patients for CRT, leaving room for exceptions that might allow CRT in some class-2 patients.

A case could be made, O'Connor said, for using CRT in a class-2 patient with an LVEF of 20%, a QRS duration of 140 ms, and an HF of ischemic etiologyan ideal candidate for an implantable cardioverter defibrillator. "Are you going to put in an ICD or a biventricular pacer-ICD, realizing that the likelihood of their progressing to class 3 in the next year or so is pretty reasonable?"

Auricchio said the data showing a reverse-remodeling benefit from CRT can be interpreted as favoring its use in class-2 patients, as it might actually prevent progression to more severe disease.

Hayes doubts CRT is deliberately used currently in patients who are unquestionably in class 2. "It's not part of the guidelines, it's not part of the reimbursement." But he doesn't rule it out for the future. "It's fair to question whether we have really pinpointed the right subsets of the HF population and whether there might be a role for it in patients who are functionally less ill."

Give meds a chance

Currently more problematic, Mehra noted, is the occasional tendency to put some patients on CRT before an adequate attempt at drug therapy. "The technology has been adopted too fast in [some] patients, who could have been better managed with medical therapy, before rushing to judgment with device therapy," Mehra said. Before resorting to CRT, the patient should have demonstrated resistance to "good evidenced-based therapy" such as ACE inhibitors, beta blockers, and aldosterone antagonists. "If they remain symptomatic three or six months down the road, then certainly I would consider device-based therapy."


When I talk about guidelines for CRT labeling, and where I'm willing and not willing to make exceptions, there's one place I draw the line.


Hayes agrees. "When I talk about guidelines for CRT labeling, and where I'm willing and not willing to make exceptions, there's one place I draw the line," he said. "First, patients must be optimized on standard drug therapy that has already been proven to work."

Auricchio, however, suggested that drug therapy typically considered optimal, which would encompass beta blockers, may not be crucial before going to CRT. "Recent data from the MIRACLE-ICD study showed that CRT even in patients not tolerating beta blockers demonstrated reverse remodeling. This is a strong argument against the mandatory use of optimal drug therapy before CRT."

Related links

1. [HeartWire > Heart failure; Mar 26, 2004]

2. [HeartWire > Heart failure; Mar 23, 2004]

3. [HeartWire > News; Dec 18, 2003]

4. [HeartWire > News; Sep 26, 2003]

5. [HeartWire > News; Mar 31, 2003]

6. [HeartWire > News; Feb 11, 2003]

7. [HeartWire > News; Sep 10, 2001]

8. [HeartWire > News; May 07, 2001]

9. [HeartWire > News; Jan 29, 2001]

10. [HeartWire > News; May 17, 2000]


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.