Baltimore, MD - Last week, the Centers for Medicare and Medicaid Services (CMS) expanded coverage to thousands of Medicare patients qualifying for implantable cardiac defibrillators (ICDs). However, the CMS decision stopped short of agreeing to reimburse all patients who meet the Multicenter Autonomic Defibrillator Implantation Trial II (MADIT II) criteria, leaving many experts puzzled and disappointed.
Dr Douglas Zipes (Krannert Institute of Cardiology at Indiana University, Indianapolis), past president of the American College of Cardiology, calls the decision "inappropriate."
"What they've done is cherry-pick data that they feel will identify the high-risk population who, more likely than not, will benefit from an ICD, while ignoring the rest of the population," he told heartwire . "They've chosen to ignore the primary objectives of the MADIT II trial."
What they've done is cherry-pick data that they feel will identify the high-risk population who, more likely than not, will benefit from an ICD. . . . They've chosen to ignore the primary objectives of the MADIT II trial.
Zipes would not speculate as to why the CMS chose to limit coverage to a subset of patients meeting MADIT II criteria, instead adding that the MADIT II study results are sound and have already been thoroughly analyzed.
"It is important to note that three of the most respected cardiology organizations in the USthe ACC, the AHA, and NASPEall approved MADIT II indications as a two-way classification, and that factor indicates that these societies have looked at the data and supported the outcome of MADIT II to use it as a guideline," he said. "In addition, the CMS's own advisory council unanimously voted to recommend approval of the MADIT II guidelines. [The CMS has] ignored these, too."
Lead MADIT II investigator Dr Arthur Moss (University of Rochester Medical Center, New York) called the retrospective CMS analysis a "fishing expedition."
"I think the CMS got caught up in slicing and dicing the primary result from the prospective trial to try to identify a subgroup that they thought would have a significantly better benefit," Moss told heartwire . "They looked through all the data until they found something that agreed with what they wanted to do, which was try to reduce the number of patients who would get ICDs."
The decision to expand ICD coverage has raised many difficult policy questions, as the hardware and the surgical implantation frequently places the devices in the ballpark of $25 000, with follow-up care often increasing these costs to around $50 000. It is estimated that about 150 000 people in the US currently have ICDs, but government officials speculate that figure would double if CMS expanded coverage based on the MADIT II inclusion criteria. Last year, Medicare spent approximately $1 billion to pay for ICDs. It is expected the CMS coverage decision will add an additional $400 million to that total.
If an ICD were a buck and a quarter, would we have gone through this entire process of reviewing the evidence at all? Maybe not.
Dr Steve Phurrough, acting director of the CMS Coverage and Analysis Group, defended the CMS decision and denied the government agency sought to limit the number of ICD-eligible patients under Medicare. He also denied cost was a factor in the decision-making process. However, he said the CMS did pay greater attention to the entire approval process with ICDs in MADIT II-type patients due to the enormous costs involved.
"If an ICD were a buck and a quarter, would we have gone through this entire process of reviewing the evidence at all? Maybe not," Phurrough told heartwire . "But because it's a bit more than a buck and a quarter we wanted to make sure the evidence was clear and that it was a benefit. We don't use cost to decide the evidence issue, but we do use cost to decide if this is important enough to address."The relative importance of QRS duration
The contentious CMS decision expands national coverage for the use of ICDs in patients with a previous MI, LVEF of <30%, and QRS duration >120 milliseconds. The decision also expands coverage to include patients with coronary artery disease with prior MI, LVEF of <35%, and inducible, sustained ventricular tachycardia or ventricular fibrillation.
It was expected the CMS would expand ICD coverage to include all MADIT II-type patients, regardless of QRS duration. In the MADIT II study, investigators confirmed a survival benefit from ICD implantation in a vastly enlarged group of patients (prior history of MI and left ventricular dysfunction with an ejection fraction <30%). Published in the March 21, 2002 New England Journal of Medicine, MADIT II demonstrated a 31% decrease in mortality for those patients receiving an ICD compared with those receiving drug therapy alone.
"The part that is unclear, or at least requires more scientific clarification, and in this sense is disappointing, is the decision to limit support for the new expanded indication within the MADIT II population," said Dr Michael E Cain (Barnes-Jewish Hospital, St Louis, MO). "As a physician-scientist, I simply don't have access to data that would help me get to first base on this decision."
Cain, president of the North American Society of Pacing and Electrophysiology (NASPE), told heartwire he is pleased a national coverage decision has been made expanding ICD coverage to additional patients at risk of sudden death but questioned where the data come from to support the decision to extend coverage only to a subset of MADIT II patients.
The only data I go on . . . is the primary MADIT II publication that said in patients with these inclusion criteria the benefit was appreciated across the entire MADIT population.
"The only data I have to go on, where I can read it, see it, and call up people to discuss it, is the primary MADIT II publication that said in patients with these inclusion criteria the benefit was appreciated across the entire MADIT II population," he said. "In the announcement that the CMS has made, they indicated they had access to data that they used to come to this conclusion, but I haven't seen it."Fuzzy math
As previously reported by heartwire , last February the Medicare Coverage Advisory Committee (MCAC) unanimously voted (7-0) to recommend expansion of Medicare coverage for the use of ICDs based on the MADIT II inclusion/exclusion criteria. The MCAC decision was based on the results of the MADIT II trial and did not stipulate QRS duration provided any further risk stratification benefit in patients meeting the MADIT II criteria.
Phurrough said an electrocardiographic substudy from the MADIT II trial, presented last year at the 2002 NASPE Scientific Sessions and previously reported by heartwire , found that QRS duration could single out post-MI patients who benefit most from an ICD and provided the impetus for another look at the MADIT II results.
"We wanted to find out if there was a subgroup within MADIT II that we could identify that had not been previously identified that had a majority of the benefit. The benefit again [in MADIT II] was only six people out of 100," said Phurrough. "Ninety-four people out of 100 didn't benefit from having an ICD. We wanted to better identify those six people."
Phurrough said discrepancies in the data prevented QRS duration from being presented and analyzed at the MCAC meeting. He pointed out that when the CMS asked Guidant (Indianapolis, IN), the sponsor of the MADIT II trial, about the data presented at NASPE, it was told by the company that the presented QRS analysis was incorrect. Guidant then supplied new QRS data that did not show any significant risk-stratification benefit, and as a result, this analysis was not brought forward to the MCAC, the advisory panel that helps the CMS make coverage decisions.
However, as the CMS was finalizing its decision, Phurrough said it double-checked the QRS data supplied by Guidant and realized it, too, was inaccurate. As CMS investigators combed through the QRS data, they realized basic arithmetic mistakes had been made and that data had been transposed improperly into different tables, skewing the risk-stratification benefit of QRS duration >120 milliseconds, said Phurrough.
When we did our own regression analysis on the data that Guidant sent us, we found a very significant difference between the QRS patients greater than 120 milliseconds and QRS patients less than 120 milliseconds.
"I don't know who actually put the data in the table that they sent to us, but there was a mistake made," said Phurrough, adding the CMS corrected the mistake in the table before analyzing the data. "When we did our own regression analysis on the data that Guidant sent us, we found a very significant difference between the QRS patients greater than 120 milliseconds and QRS patients less than 120 milliseconds. That is how we came up with QRS as a discriminator," he said.
For this reason, said Phurrough, QRS duration was not factored into the original MCAC recommendation to approve the ICD indication in MADIT II-type patients but was later used in the CMS decision to expand coverage only to patients meeting MADIT II criteria and with QRS >120 milliseconds.
According to Dr Joseph M Smith, chief medical officer, Guidant, the error appearing in the table was presented during one of the many presentations between Guidant and the CMS. However, he said, the mistake appeared in one summary table only and was immediately corrected.
"The CMS has had access to the full and complete data set since they requested it, for almost well over a year," Smith told heartwire . "We were aware of the mistake, as Steve [Phurrough] brought it to our attention and it was corrected. However, the error was present in a summary table and was not present in the primary data set we originally gave to them."
While Smith contends QRS duration is an important factor to consider, he argues it was not the primary end point of the MADIT II trial. He added the original QRS subset analyses presented at NASPE have not yet gone through the proper peer-review publication channels.
Regarding the retrospective QRS analysis undertaken by the CMS, Moss said it was simply "not good science." He takes exception to the subset analyses and argues that the decision is a fundamental departure from the primary MADIT II results that will deny coverage to large numbers of patients who qualify for an ICD.
They're telling us our overall study is very good, but they think the subset with QRS less than 120 milliseconds doesn't benefit or is behaving differently. It's a very bad scientific precedent.
"Suppose an individual went to the CMS with a negative study and said, 'Look, one subset does benefit a great deal and we want to get approval based on this subset.' They'd be laughed out of there," said Moss. "CMS has done the same thing in reverse. They're telling us our overall study is very good, but they think the subset with QRS less than 120 milliseconds doesn't benefit or is behaving differently. It's a very bad scientific precedent."
Zipes agrees. "QRS is important, but it was not even the focus of this particular study. What a careful scientist would do is note the QRS duration as a possible additional risk factor and use it as a hypothesis to now test in a prospective randomized clinical trial but not base decision-making on a subgroup analysis that was not part of the principal outcomes of the study."Leaving physicians and patients in a pinch
The CMS has stated it would reevaluate the decision when the results of the National Heart, Lung and Blood Institute's Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) are available early next year.
While Cain commends the CMS for the decision and the continuous dialogue between the government agency, NASPE, and independent physicians, he says the decision needs to be clarified further. He admits the decision leaves many electrophysiologists in a difficult position, especially when advising patients.
"When we wear our physician hats, we always must tell our patients the way we believe they need to be treated and be able to defend why we believe a certain treatment is indicated. That is our first priority. The data we have from MADIT II support the use of ICDs in these patients."
You ignored published guidelines from three prominent organizations, and I feel you practiced a level of care below acceptable standards. I'm going to sue you for malpractice.
In September 2002, guidelines written jointly by the ACC, AHA, and NASPE were revised to support the implantation of ICDs based on MADIT II criteria. With the new guidelines already in place, Zipes said the CMS decision leaves physicians in a very difficult position, especially if a patient meets the criteria but does not have a QRS duration >120 milliseconds.
"What happens if that patient leaves and has sudden death for lack of an ICD? Very likely the surviving spouse is going to come back to the physician and say, 'You ignored published guidelines from three prominent organizations, and I feel you practiced a level of care below acceptable standards. I'm going to sue you for malpractice.' The physician is caught between the proverbial rock and a hard place, with CMS saying one thing and published guidelines saying another."
All parties admit the launch of newer, cheaper ICDs will eventually change the landscape of device implantation, and the decision will likely continue to be modified.
Heartwire from Medscape © 2003
Cite this: Reaction to Medicare ICD decision: Experts dismayed full MADIT II criteria not included in coverage - Medscape - Jun 13, 2003.