Medical Device Daily Washington Editor

GAITHERSBURG, Maryland — FDA has held discussions with industry regarding alternatives to the standard randomized, controlled clinical trial (RCT) for cardiovascular devices to treat atrial fibrillations (AF), but putting alternative designs into play for the devices to treat AF has been a complicated task.

The circulatory systems advisory committee met last week to consider allowing greater flexibility in the trials for various technologies addressing this application, but the panel’s consensus was that an RCT is still the only appropriate means for testing safety and efficacy in these devices.

The panel’s reluctance to accept the idea probably was fed by two issues: the variability in the persistence and the clinical significance of episodes of AF, and differences of opinion among sponsors on the degree of difficulty in recruiting patients who are willing to be randomized into a drug therapy control arm of these trials.

Lesley Ewing, a contract employee at the Office of Device Evaluation, described AF as “an important public health problem” that can be asymptomatic, with medical treatment usually palliative and not effective for maintenance. The gold standard for surgical treatment is the Cox Maze procedure, she said, now usually performed as the modified RF Cox-Maze.

Ewing said there are “many approaches to catheter ablation,” with efficacy rates in the literature ranging from 28% to 100%. Complications of catheter ablation, including death, stroke, tamponade, and pulmonary vein stenosis, are “inconsistently reported,” she said.

Part of the agency’s dilemma, Ewing said, is very basic: “What is an appropriate method to characterize effectiveness?”

Among the proposed measures are absence of AF and absence of symptomatic AF. Another question still up in the air is how to define a reduction of AF burden.

Felipe Aguel, an engineer at FDA’s Office of Science and Engineering Laboratories said that an alternative design — one that allowed a one-arm study with objective performance criteria — would make it “potentially easier to enroll patients” and might pull down the sample size of the study population, but also that “it is unclear whether suitable data are available” to aid in the design of such a study.

Aguel said that a device maker and FDA might be able to agree on a hybrid design, one that would randomize patients into both a control and study arm, but the patients in the control arm would get ablation by another device.

In such a scenario, both primary efficacy and primary safety comparisons would be based on data for the alternative ablation method. However, the data from the study arm would also be stacked up against the OPC standard.

While such an approach was likely to ease enrollment problems at least somewhat, Aguel said, “the use of multiple catheters” in some procedures “makes it difficult to standardize procedures.” The resulting pooled data would likely be confounded by the presence of such mixed data.

During the morning’s open public hearing, Albert Waldo, MD, professor of cardiology at Case Western Reserve University (Cleveland, Ohio), said he was participating in the NaviStar ThermoCool catheter trial, which is “widely used,” including for AF. Biosense Webster (Diamond Bar, California), maker of the NaviStar, “recognizes the importance and value” of examination of safety and efficacy, he said, but the trial for paroxysmal AF is “encountering major challenges.”

Waldo said that the trial, which randomizes to drug therapy, “has taken three-plus years” to fulfill enrollment objectives. He said that of all the prospective participants screened, only 3% were enrolled and that 11% refused to participate due to reluctance to end up in the control arm. However, the majority of the screened prospects, 62%, flunked inclusion/exclusion criteria.

The company’s efforts to recruit were apparently uneven in terms of cost and benefit. In one surge of recruiting, involving ads on television and the web as well as through doctor’s offices, the company spent about a half-million dollars and netted only three patients for the trial.

“We would suggest that greater flexibility is needed in trial design,” and companies need to be able to tailor inclusion/exclusion criteria, he said. He also made a case for allowing clinical investigators some flexibility in deciding on adjunctive catheters to be used in the procedure, saying that “the issue is not what’s best — the issue is what works.”

Helen Barrow, chief medical officer for CryoCor (San Diego, California) – considered in the lead to win FDA approval for an AF-ablation device — said the company had recently completed enrollment in the trial for its cryocatheter.

The company obtained an investigational device exemption from FDA in August 2004, and the first patient was enrolled that November. Barrow said that while recruitment had proved challenging, it was nonetheless do-able.

Burke Barrett, VP for regulatory affairs at CardioFocus (Marlborough, Massachusetts), said that the company is working on a balloon catheter designed to treat pulmonary-based AF and that “our experiences with patient recruitment to date has been very challenging.”

Even offering the possibility of crossover to any patients going into the control arm does not seem to help much with willingness to be randomized, Barrett said, and that enrolling 200 “may mean screening 10,000 patients.”

Addressing the notion of a performance goal, Hugh Calkins, a professor of medicine at Johns Hopkins University Hospital (Baltimore), told the panel that a number of trials for paroxysmal AF will classify asymptomatic AF as a successful outcome, but that this is not a universally held view. He also said that using ablation as a control treatment would allow more effective comparisons and provide a sufficiently large patient base to populate studies.

The panel hashed out a number of questions, including what might serve as a standard outcome for successful treatment of AF as well as what duration of follow-up might be appropriate to track adverse events for drugs used to treat AF.

This latter discussion was prompted by a comment that adverse events for devices tend to show up much more quickly than those for drugs, but panelist James Neaton, PhD, a professor of biostatistics at the University of Minnesota (Minneapolis), suggested that longer follow-up would deal with this.

Another dilemma is that the substantial body of patients whose electrocardiograms display tracks that are typical of AF, but who never experience physical symptoms.

In the end, the panel concluded, with little dissent, that RCTs should remain the standard for trials.