October 25th, 2012

TCT: Two PFO Closure Trials Miss Primary Endpoints

Two trials presented today at the TCT meeting in Miami testing the benefits of PFO closure in patients with cryptogenic stroke have failed to convincingly demonstrate any significant  benefit for the controversial procedure.

The RESPECT (Randomized Evaluation of Recurrent Stroke Comparing PFO Closure to Established Current Standard of Care Treatment) trial randomized 980 patients to PFO closure with the Amplatzer PFO Occluder device or medical therapy. According to lead investigator John Carroll, the rate of recurrent stroke was low in both arms of the trial: 1.6% in the closure group and 3% in the medical-therapy group.

This difference between the groups did not achieve significance in the intention-to-treat (ITT) analyses:

  • ITT raw count: 46.6% risk reduction (p=0.157)
  • ITT Kaplan Meier: 50.8% risk reduction (p=0.083)

However, statistical significance was achieved in the per-protocol and as-treated analyses:

  • Per-protocol Kaplan Meier: 63.4% risk reduction (p=0.032)
  • As-treated Kaplan Meier: 72.7% (p=0.007)

The investigators reported that there were very few device- or procedure-related complications. There was a similar amount of serious adverse events in the two groups (23% with the device and 21.6% with medical therapy).

The investigators concluded that “for carefully selected patients with history of cryptogenic stroke and PFO, the RESPECT Trial provides evidence of benefit in stroke risk reduction from closure with the AMPLATZER PFO Occluder over medical management alone.”

“The optimal secondary prevention strategy following a cryptogenic ischemic stroke in patients who are found to have a PFO has been unknown,” said Carroll, in a TCT press release. “This need to know is particularly intense for young stroke patients who have no or minimal traditional risk factors for ischemic stroke, yet face a risk of recurrent stroke for many decades. RESPECT makes progress in both removing the ‘unknown’ or cryptogenic cause of some strokes and providing high quality data from a large, long-term randomized trial.”

A similar pattern occurred in the smaller PC (Percutaneous Closure of Patent Foramen Ovale versus Medical Treatment in Patients with Cryptogenic Embolism) Trial, in which 414 patients were randomized to PFO closure or medical therapy. The primary endpoint, the composite of death, nonfatal stroke, TIA, and peripheral embolism, occurred in 3.4% of the PFO closure group compared to 5.2% of the control group (relative risk reduction: 37%, p=0.34). The incidence of stroke was 0.5% versus 2.4% (relative risk reduction 80%, p=0.14).

Study investigator Stephan Windecker said that because of a lower-than-expected rate of events after a mean follow-up of 4 years, the trial ended up being underpowered to detect meaningful differences. He concluded that “the observed difference in stroke … may be clinically relevant if confirmed in further studies.”

“We need a definitive trial of this approach if it’s going to be broadly used for PFO closure,” said Deepak Bhatt, in an interview with Bloomberg News. “Anecdotally, there are patients who seem to benefit. It’s unfortunate that none of the trials have been able to absolutely nail that down.”

Earning Respect?

As reported here over the summer, during an earnings call St. Jude CEO Dan Starks told investors that the results of RESPECT were “favorable.” As I wrote then, the danger of this sort of statement  is that the company’s initial evaluation of the results may clash with the eventual judgement of the medical community. Companies are simply in no position to be objective about their own products or trials. Caution in this case was particularly warranted because of the sorry history of negative trials in this area, as highlighted by the failed MIST trial of the STARFlex Septal Closure System. The results of the RESPECT and PC Trials demonstrate that the communication of medical information should not be left in the hands of industry. (St. Jude stock dropped 3.5% on the announcement of the trial results.)

One Response to “TCT: Two PFO Closure Trials Miss Primary Endpoints”

  1. The discussion and debate about the RESPECT results have just begun and additional data must be published to provide a more complete examination of all the results. The complexity of dissecting out an appropriate target population and difficulties of performing a clinical trial when off-label closure is widely available and patients are often unwilling to be randomized are part of the perspective that should be brought to this important discussion. This is more than a device trial but also is about trying to understand why such a large number of relatively young people have cryptogenic strokes. The partership between the neurology-stroke and the cardiology communities will continue to be the only way to make meaningful progress. While the primary endpoint of RESPECT missed its mark (0.08) for being conclusive (0.05)there are many lessons learned in this trial that should not be dismissed but rather used to move the field forward. A secondary prevention trial using a cardiovascular device is truly unique especially when considering the major long-term committment needing from patients willing to participate, trial sites, and the sponsor.