November 20th, 2012

PFO Occluder Devices Don’t Get No RESPECT

In the RESPECT (Randomized Evaluation of Recurrent Stroke Comparing PFO Closure to Established Current Standard of Care Treatment) trial, 980 patients were randomly assigned to (1) PFO closure with the AMPLATZER PFO occluder device (St. Jude Medical, which sponsored the trial) or (2) medical therapy. The trial’s primary endpoint was a composite of recurrence of nonfatal stroke, fatal ischemic stroke, or early (within 45 days after randomization) all-cause mortality.

Patients assigned to received the AMPLATZER device had a 46.6% relative reduction in event risk compared with those assigned to medical management (9 vs. 16 events).  However, in the intent-to-treat (ITT) analysis, this difference did not reach statistical significance (p = 0.157).

Does what followed call to mind a straight man, a song and dance, or a comedy routine? 

The statistician noted a differential length of follow-up in the two arms, which was attributed to “late trial fatigue”: More of the patients in the medical group than in the device group left the study early (90 vs. 48) — some because they wanted to get an off-label device outside the study — resulting in more years of observation in the device group for events to occur (1,375 vs. 1,184 patient-years; P=0.009).  As a result, the ITT raw-count analysis was deemed invalid and a prespecified Kaplan Meier (i.e., time to event) analysis was performed, which showed a strong trend toward benefit (P value, 0.08–0.09).  Not statistically significant. . .but close. 

The investigators next performed an analysis that excluded events that occurred in the device arm before a closure procedure was attempted. The resulting per protocol analysis (which eliminated 3 of the 9 strokes in the device arm) showed a 63% reduction in the primary endpoint with the device as compared with medical therapy (P=0.032).

The punch line: The RESPECT investigators concluded that their study” provides evidence of benefit in stroke risk reduction from closure with the AMPLATZER PFO Occluder over medical management alone.”

Rodney Dangerfield was famous for claiming, “I don’t get no respect.”  After all the slicing and dicing of these data, should the PFO Occluder device get any respect?  

2 Responses to “PFO Occluder Devices Don’t Get No RESPECT”

  1. Steven Greer, MD says:

    If you believe the empirical and medical imaging evidence that materials pass through the PFO to the aorta, and the device is safe, and the studies are woefully underpowered and impossible to conduct, then why not use it? Would you risk your own brain?

    Mind you, I am bearish and skeptical of stents, ICDs and TAVI, and am no medical device fan, in general.

  2. Joel Wolkowicz, MDCM says:

    Data torture – if you twist and manipulate the data enough, it will say anything you want it to say.