Stroke experts debate FDA approval process for devices

Merci Retrieval System was approved to remove blood clots in patients experiencing a stroke.

NEW ORLEANS – When Concentric Medical first introduced the Merci Retrieval System at the American Stroke Association 2004 meeting, scientists called it a “revolutionary” device that would “almost instantly” reverse the damage caused by ischemic stroke.

“How often do we get a chance to reverse a patient’s stroke on the table? We have had patients completely paralyzed on one side of their body, who were made normal almost instantaneously when the clot was retrieved,” said Sidney Starkman, MD, codirector of the UCLA Stroke Center, when presenting data on the first 114 patients in the Merci database at the 2004 meeting.

One year later, the database has grown from 114 to 141 patients, and the device has been approved by the FDA. At this year’s American Stroke Association International Stroke Conference 2005, researchers discussed the approval process.

Corkscrew device

The Merci Retrieval System, sometimes referred to as a “corkscrew” device, consists of a nitinol wire with a helical-shaped tip. The wire is passed through the thrombus within a microcatheter, and once the microcatheter is beyond the occlusion, it is retracted and the wire within reforms the helical shape.

The catheter and the retriever are subsequently pulled back to engage the thrombus, and the entire apparatus is retracted back into the balloon-guided catheter, bringing the thrombus with it. The thrombus and guide catheter can then be removed from the patient.

The FDA approved the device in August 2004 based on a database of 141 patients. Treatment with the Merci device resulted in a 48% rate of TIMI II or III flow, which was significantly higher than the 18% observed in the control group of the PROACT (Prolyse in Acute Cerebral Thromboembolism) II trial, which the researchers had used as a comparator group. There was a 43% mortality rate, a 13% rate of procedure-related serious adverse events, and a 7% rate of serious device-related events.

Rigorous clinical trials needed

Kyra Becker, MD, associate professor of neurology at the University of Washington, was chair of the FDA Neurologic Devices Panel when the committee was asked to consider the device. At the recent stroke conference, Becker criticized the use of the PROACT II database as a historical control to prove safety and efficacy of the Merci system and said stricter standards should be used for device approvals.

In PROACT II, those patients treated with intra-arterial recombinant prourokinase less than six hours after onset of acute stroke had a 66% rate of recanalization and a 40% rate of clinical improvement defined by Rankin score of <2.

“Despite these findings, this agent was not approved by the FDA. Why is there a different standard for approval of devices than there is for drugs?” Becker said. “It’s not completely clear that the Merci retriever is safe, and while it is clear that the device effectively removes clots, it is not at all clear that it improves stroke outcomes.”

Becker said any devices approved for stroke should undergo a rigorous clinical trial process. “We’re treating patients, so the clinical outcome needs to be the primary outcome; you can’t use surrogates. We also need contemporaneous control groups.

“In this day and age it’s probably not going to be a placebo group, but at least a conventional medical therapy. Because medical therapies change over time, there’s a point where historical controls are no longer relevant,” Becker said. “It’s probably necessary to retrieve a clot to improve outcome, but we owe it to our patients to find out for sure.”

Advances are an improvement

Other panel members disputed how realistic a controlled trial would be. Stanley Barnwell, MD, PhD, associate professor of neurointervention at Oregon Health & Science University, said patients and their families would be unlikely to consent to participating in a controlled trial.

“We know that the natural history of this problem is so terribly poor that really any advances are an improvement compared to natural history,” Barnwell said at the stroke conference. “I’m really quite mystified as to why there’s any debate on this issue.

“We’re no longer going to be able to justify a control group of patients with cerebral artery occlusions. They do poorly, and we need to find a way to open them up. Patients and their families will not accept the option of doing nothing.”

Anthony Furlan, MD, section head of stroke and neurologic intensive care at the Cleveland Clinic Foundation, agreed that a randomized controlled trial would be “next to impossible,” but said the FDA needed to develop a pathway to approval for devices that was similar to that for drugs.

“Why was the device approved but the drug was not? If we just look at safe and effective clot removal, the drug and the device look similar, yet one branch of FDA approves the device and another branch doesn’t approve the drug,” Furlan said at the stroke conference.

“The idea that drugs and devices are different in the thrombectomy and reperfusion era becomes an artificial distinction, particularly if you are just talking about safely opening arteries, because drugs like TPA can do this as well. I’m not disagreeing with the drug branch of FDA, I just think the device branch needs to develop a similar pathway.

Cost factors

Randall Higashida, MD, professor of radiology and neurological surgery at the University of California San Francisco, said resources vary for development of devices and drugs.

Higashida estimated that the cost to conduct the PROACT trials exceeded $25 million, and the typical cost for bringing a new stroke drug to market is between $250 million and $300 million.

“To organize and perform a Phase 3 trial of the Merci device as mandated by the FDA would require an additional 450 to 500 patients; involve 80 sites in North America, Europe and Australia; take about three years to complete; and cost $30 to $50 million,” Higashida said. “Such an effort and expense is not reasonable for a startup medical device company.

At Massachusetts General Hospital, Walter Koroshetz, MD, associate professor of neurology, is not waiting for the definitive clinical trial. “Patients indicated for this sort of device are generally those who have ‘nothing to lose’ in the sense that they can expect severe debilitation if we don’t get the clot out,” Koroshetz said.

“The Merci retriever is another tool we can use. It’s not the focus of your stroke program. In terms of the field, the neurology field is way behind and the cardiac field is way ahead. We’ve been forced to use devices made for the heart, and we need to move this field forward so we have devices for the brain.” – by Jeremy Moore

For more information:

• The MERCI device approval. Presented at the American Stroke Association International Stroke Conference 2005. Feb. 2-4, 2005. New Orleans.