CDU Washington Editor

The race is on — or at least the participants have entered the starting gate.

Xience, a second-generation drug-eluting stent (DES), this one from Abbott Vascular (Santa Clara, California), received a strong recommendation from the FDA's cardiovascular device panel in late November, putting it on the track next to the Endeavor from Medtronic (Minneapolis), which in October had receive a strong panel recommendation. The panel voted 9-1 for approvable with the conditions of the use of American College of Cardiology/American Heart Association guidelines on DAPT and a post-approval study (PAS) to be designed jointly by the sponsor and FDA.

The industry will now be holding its collective breath to see 1) if and when the devices receives final FDA approval, 2) if both are approved, which is approved first, 3) if both are approved, which receives fastest uptake in the market, and 4) how rapidly and to what extent the new 2.0 device, or devices, elbow aside the first-generation products from Cordis (Miami Lakes, Florida) and Boston Scientific (Natick, Massachusetts).

Abbott's request for a positive recommendation was significantly supported by the presence of an all-star roster of clinical investigators involved with the product's pivotal trial. And, as with Medtronic's submission for the Endeavor, Abbott submitted reams of supporting data to buttress the application at the Nov. 29 panel hearing.

Abbott acquired the Xience when Boston Sci purchased Guidant in 2006 and was required by the Federal Trade Commission to divest the Xience, but Boston Sci was also allowed to offer the exact same product under its own brand, the Promus DES.

Deliverability key

At the panel meeting to review the Xience, Murthy Simhambhatla, MD, Abbott's VP for DES products, said the Xience is based the Multi-Link Vision, a bare-metal stent (BMS), the decision to do this predicated on the fact that the Multi-Link has, he said, "proven to be highly deliverable, a characteristic that was preserved even after the coating."

The Multi-Link is made of cobalt-chromium, which Simhambhatla said "allows for thinner struts without compromise to radio-opacity or flexibility." Another factor was that radiological studies suggest that thicker stents incur progressively greater — hence excessive - endothelial formation. Simhambhatla also said that about 80% of the everolimus is eluted by 30 days and "substantially all is eluted by 120 days."

Gregg Stone, MD, the chairman of CRF (New York), told the panel that three-year data for the SPIRIT I trial — the first-in-man trial for the device — showed no cardiac deaths in either arm and that the data trended toward lower target vessel revascularization (TVR), and lower major adverse cardiac events (MACE), although the study was not powered to evaluate these.

"Most importantly, there were no cases of late-stent thrombosis out to three years" in either the Xience or the bare metal stent (BMS), he said. Stone said that the SPIRIT II trial — which included patients with up to two de novo lesions — randomized 300 patients in a 3:1 ratio to Taxus, with a primary endpoint of angiographic in-stent late loss at 180 days.

Stone said that in the primary endpoint for SPIRIT II, "not only was the Xience not inferior, it was highly superior," exhibiting a 69% reduction in in-stent late loss of lumen diameter. He noted that intravascular ultrasound (IVUS) studies indicated a 64% reduction in percentage volume obstruction. While the study was not powered for clinical endpoints, thrombosis to one year was low for both devices.

"For the first time in a randomized trial, you can see a reduction in MACE," Stone said in reference to numbers that indicated an incidence of MACE of 9.2% for Taxus vs. 2.7% for Xience. For target vessel failure, the numbers indicated an incidence of 9.2% for Taxus and 4.5% for Xience.

Meta-analysis a positive

Stone, the co-principal investigator for SPIRIT III, told the panel that the meta-analysis of the SPIRIT II and III trials was spurred by an interest in conducting an "examination of lower-frequency safety and efficacy endpoints, such as death, myocardial infarction, stent thrombosis and target lesion revascularization [TLR]," which SPIRIT III "was not powered to examine." CRF conducted the meta-analysis, which Stone said was statistically justifiable in part because of the similarity in inclusion/exclusion criteria.

The conclusion from the meta-analysis: that Xience "results in significant reductions in angiographic in-stent and in-segment late loss and binary restenosis," Stone said.

Mitchell Krucoff, MD, a professor of medicine at Duke University Medical Center (Durham, North Carolina) and another prominent voice in the field, provided an overview of safety for DES devices in general, saying, "we have to keep in mind that effectiveness measures and safety measures in coronary implants are closely related."

He said that more than 15 peer-reviewed meta-analyses regarding BMS vs. DES have been decisive. "Out of all that complexity," Krucoff said, the message is that "the regulatory review process worked" because Taxus "is safe and effective relative to BMS."

Krucoff pointed out that enrollment in Spirit III was completed prior to FDA's new requirements for DES trials, and of the 1,002 enrollees, 603 have hit the two-year follow-up. He noted that all-cause death rates were about 30% lower in the Xience arm for the two-year data that Abbott had on hand.

On to SPIRIT IV and

Krucoff also said that Abbott is sponsoring SPIRIT IV, a continued-access prospective randomized trial of almost 3,700 patients that will focus on ischemia-driven MACE at 270 days. SPIRIT V, he said, will collect data from 2,700 patients in a registry, 300 of whom will be diabetics. Both studies were the subject of safety data board meetings, Krucoff said, and no events have surfaced as yet.

Robert Fiorentino, MD, FDA's lead reviewer for the Xience, said that the agency had some concerns regarding the meta-analysis of SPIRIT II and III data because they used different definitions of adverse events and because SPIRIT II "was not adequately powered to evaluate all the clinical endpoints" made reference to by the sponsor. Fiorentino also noted that SPIRIT II "was not designed to evaluate low-frequency adverse events" and the use of an interim analysis to review the data may have biased the conclusions.

A concern: no subgroup analysis

FDA's complaints about SPIRIT III — despite its meeting its primary and secondary endpoints — were that the trial "was not designed to establish safety and efficacy in patient subgroups" and did not meet FDA's requested follow-up rate of 90%.

Although the 4.0 mm-diameter arm of SPIRIT III was not powered to establish full safety and efficacy, "the Xience 4.0mm stent had results comparable to the RCT" portion of SPIRIT III, according to Fiorentino. However, the 4.0mm arm was not randomized and was not designed to evaluate clinical endpoints, and only 71% of this group of patients had qualifying follow-up angiograms.

He noted that SPIRIT II and III "are the first trials to prospectively analyze" a stent intended for dual-vessel placement, but while there were higher rates of thrombosis in the study arm (four out of 654 against zero in the 319 enrolled in the control arm), all other adverse events were lower in the dual-placement patients on Xience than their counterparts on Taxus.

During the panel discussion, members seemed to revisit the importance of late-stent thrombosis in comparison to other outcomes. Warren Laskey, MD, an interventionist at the University of New Mexico Health Science Center (Albuquerque), asked "are we too hung up on LST? Are we just perseverating and isn't the main thing death and MI?"

Richard Page, MD, the chief of cardiology at the University of Washington School of Medicine (Seattle), said, "I agree that thrombosis is important, but I don't think it's more important than restenosis." He also cited "a real advantage to a stent with real technical advantage" in terms of ease of placement, so "if we see reasonable assurance of effectiveness ... and safety, which we're wrestling with," the device perhaps should be on the market.

The panel initially discussed whether the post-approval study should include controls, but a vote among the members came to a 5-5 tie, and panel chairman Clyde Yancey, MD, professor of cardiology at the University of Texas Southwestern Medical Center (Dallas), cast the deciding vote against the recommendation.

There was no indication from FDA whether it would have acted on such a recommendation, but it would have represented a sharp departure from the norm.

No nays for HeartMate II

On the day following the Xience panel meeting, the circulatory systems panel portaged through another all-day affair, this time over the premarket application for the HeartMate II (HMII) left ventricular assist device (LVAD) from Thoratec (Pleasanton, California).

This meeting, as they all tend to do, featured a set of debates over safety and efficacy as they all do, but also was conducted against a backdrop of two product recalls conducted by Thoratec earlier this year. This too resulted in a unanimous vote to recommend approval.

Thoratec brought along three patients implanted with the HMII to give the panel their perspectives. Whether or not the panelists needed a reminder that heart failure is a terminal condition — and the panel's unanimous vote in favor of approvability indicated that the reminder was unnecessary — that stark reality may have played a role in the decisiveness of the vote.

According to documents filed by FDA prior to the hearing, part of the agency's concern about the HMII trial was that seven of the 133 original patients in the pivotal trial did not meet an inclusion criteria of 1.5 square meters of body surface area, a consideration that the agency indicated was built into the performance goals for the HMII. This also reduced the number of patients who could be included in the primary data analysis from the 133 original enrollees to 126, a difference the FDAers cited at the hearing as another source of concern.

In the company's morning presentation, Don Middlebrook, Thoratec's VP for regulatory affairs, said FDA gave the firm the nod on the pivotal trial in February 2005 and that Thoratec had enrolled all 133 patients by May the following year, "a record pace of enrollment in the history of this technology."

FDA's major deficiency letter was dated this past April, Middlebrook said, but the firm has implanted the HMII in another 280 patients since then, and the registry data from these patients were included in the PMA submission to FDA and in the data presented to the panel. However, neither the agency nor the panel seemed particularly interested in the registry data.

Middlebrook also said that Thoratec's proposed PAS, which he described as a "comprehensive plan," would make use of the fledgling Interagency Registry of Mechanically Assisted Circulatory Support (Intermacs) registry. Intermacs is a joint effort managed by FDA, the Centers for Medicare & Medicaid Services and the National Heart, Lung and Blood Institute of the National Institutes of Health.

The Intermacs web site (intermacs.org) describes the database as "a national registry for patients who are receiving mechanical circulatory support device therapy to treat advanced heart failure."

The University of Alabama at Birmingham and the United Network for Organ Sharing (Richmond, Virginia) also are sponsors of the registry. As of mid-December 2007, they reported that a total of 513 patients were listed at the registry from a total of 82 study sites, with applications from another 25 study sites listed as pending.

The HeartMate II trial is described as a bridge-to-transplant trial — although with the paucity of donor hearts, patients may end up with LVADs as a destination device by default — employing a dual primary endpoint.

The first of these was a rate of 65%-70% of the patients surviving to transplantation or to 180 days, and the second primary endpoint was two-year survival free of neurological events.

Recoveries produced by device ...

FDA documents released prior to the hearing indicated that of the 126 patients meeting the BSA criterion, 72 had obtained transplants and 13 were still on the transplant list.

Four of the patients experienced sufficient recovery to leave the hospital, and 29 patients passed away during the first 180 days. The success rate was 63.9%, just short of the stated objective in the study design (calculated by adding the 72 patients transplanted to the 13 still eligible divided by the original 133 enrollees).

Speaking on behalf of Thoratec, Leslie Miller, MD, chief of cardiology at Washington Hospital Center (Washington) and principal investigator for the pivotal trial, said that "all patients were [New York Heart Association] Class IV at enrollment," but after six months, "nearly all the patients had achieved a class I or II functional capacity," a finding that he said was confirmed by the continued access of patients, as well as the cohort of patients with small BSA measurements.

The BSA subgroup, which consisted of the seven in the original trial and another eight who joined later during the continued access period, consisted largely of women, a fact that caught the panel's attention given the general lack of women receiving VADs — and none thus far having been implanted with the AbioCor artificial heart — relative to men.

... but a focus on bleeding

Many members of the circulatory systems panel focused on the 59 bleeding events associated with the procedure (it was not clear how many patients this accounted for) — so much so that one panelist proposed that an approvable recommendation be accompanied by a mandatory co-procedure to prevent bleeding.

This motion consumed tens of minutes without gaining any traction, and the panel ended up passing on a specific recommendation regarding the bleeding phenomenon.

In contrast to the problem with patient hemorrhage, Frank Pagani, MD, associate professor of surgery at the University of Michigan Health Systems (Detroit), informed the panel that 30-day mortality for the HM II was one-half that of the pivotal trial for the company's HeartMate XVE. He also said the implant technique "is the same as the XVE, which is the current standard-of-care," but is a shorter and less difficult procedure, probably due principally to the difference in size.

The panel looked at a number of conditions for approvability, voting 11-1 to require that the PAS track outcomes by sex and to require a concurrent comparator arm.

The panel also voted in an 11-1 landslide to require that the label rule out use of the VAD in patients who cannot be on anti-coagulants (the HeartMate II's smaller pump design mandates the use of an anti-coagulant to maintain function). The panel then voted unanimously to recommend approvability.

Thoratec's CEO, Gary Burbach, told Cardiovascular Device Update that he was not surprised at the panel's conclusion that a PAS should include a comparator population, given the discussion of requiring a controlled study.

However, "[w]e do expect there will be some cannibalizing" of the HeartMate XVE, he said. He added that "we expect the XVE will hold" a place in the market because it requires less anti-coagulation therapy, and that the company sees "quite a bit of enthusiasm" for the HMII from doctors.

A key trend developing?

Perhaps one of the more conspicuous trends in the circulatory system advisory committee hearings held by FDA recently is that the votes are overwhelmingly in favor of an "approvable" recommendation, and of the three most recent cardiovascular system panel hearings, two votes were unanimous.

Add to that the fact that the panel's recommendations for post-approval studies (PAS) often closely parallel the follow-up studies already discussed by the sponsor and the agency, and the hearings begin to feel like a substitute FDA (assuming the unlikely event that FDA will deny final approvals in these cases).

Another development of some interest is that some of the nation's more prominent cardiologists are becoming more comfortable with the relative rates of late-stent thrombosis for DES and the BMS variety.

On the other hand, members of the circulatory systems advisory committee have discussed on several occasions requiring sponsors to include either controls or a comparator population for PAS. To date, motions to require a controlled PAS have withered under the heat of discussion, usually when a panelist mentions the financial and logistical burden on the product's sponsor.

But a recent meeting of the committee concluded with a recommendation that the PAS include a comparator population.

In October, the panel heard from FDA and Medtronic (Minneapolis) on the subject of the firm's application for the Endeavor, a stent that elutes zotarolimus.

At that hearing, Medtronic offered a slew of data on more than 2,200 total enrollees. In that data were two-year data on more than 1,300 patients, and the volume of information was probably the primary driver of a unanimous vote for approvability on the part of the panel.

Still, clinical trial data are seemingly always controversial, and the Endeavor data were no different. During the open comment period at the hearing, Bruce Ferguson, MD, speaking for the Society of Thoracic Surgeons (Chicago), told the panel that the use of composite endpoints in clinical trials was of little interest to patients and that many physicians see data from randomized, controlled trials (RCT) as providing "inadequate information to provide guidance for population subsets."

Ferguson also stated that the "short development cycle of new technologies" is inherently problematic because "new information [on the risks of a device] is always coming." Making the case that "the overall process of device evaluation is not patient-centered," he urged FDA and the panel to insist on the use of coronary bypass as the control population for future DES applications.