Redo patients are generally sicker
The propensity matching was designed to eliminate baseline differences for the outcome comparisons, but Dr. Makkar did point out that redo-TAVR patients were sicker than the native TAVR patients. For example, when compared prior to propensity matching, the STS score was higher (8.3 vs. 5.2; P < .01), more patients had atrial fibrillation (47.9% vs. 36.2%; P < .01), and more patients had a prior stroke (15.0% vs. 10.7%; P < .01).
The registry only has follow-up out to 1 year, but participating patients were matched to a claims database to capture outcomes out to 3 years. Mortality rates at long-term follow-up were not significantly different for redo vs. native TAVR (42.2% vs. 40.3% respectively; P = .98); for the entire dataset or when compared in subgroups defined by Edwards valve redo of an Edwards valve (P = .909) or an Edwards valve redo of a non-Edwards device (P = .871).
Whether an early redo, defined as 12 months after the index TAVR procedure, or a late redo, the rate of mortality ranged from approximately 16% to 18% with no significant difference between redo and native TAVR.
A moderator for the late-breaking trials session where these data were presented, Darren Mylotte, MD, a consultant in cardiology for the Galway University Hospitals, Ireland, challenged Dr. Makkar about the potential for selection bias. He said redo patients might be the ones that interventionalists feel confident about helping, making this comparison unrepresentative.
“I think that the selection bias is likely to cut both ways,” Dr. Makkar replied. For many patients with a failed TAVR, he explained that clinicians might think, “There is nothing to be done for this patient except to try a redo.”
Dr. Makkar reports financial relationships with Abbott, Cordis, Edwards Lifesciences, and Medtronic. Dr. Cribier reports a financial relationship with Edwards Lifesciences. Dr. Mylotte reports no potential conflicts of interest.