乌鸦传媒

WASHINGTON -- Real-world results with transcatheter aortic valve implantation (TAVI) have been holding up compared with clinical trials, a national registry showed.

The 1-year mortality rate came in at 26.2%, compared with 24.3% in the pivotal PARTNER trial's high surgical risk cohort and 30.7% in its inoperable cohort, , of the Mayo Clinic in Rochester, Minn., and colleagues found.

The 1-year stroke rate was better, at 3.6%, in the Transcatheter Valve Therapy registry than the 8.7% and 11.2% rates, respectively, in the PARTNER trial cohorts, Holmes reported here at the American College of Cardiology meeting.

The stroke finding likely reflected differences in ascertainment, as clinical practice may not include the same kind of rigorous neurologic review used in the trials, Holmes suggested.

But there may also have been an element of "risk creep," according to , of Massachusetts General Hospital in Boston, who was a discussant at the late-breaking clinical trial session.

The registry showed an average Society of Thoracic Surgeons's 30-day predicted risk of surgical mortality (PROM) score of 7.1% in commercial use compared with 11.8% in the PARTNER trial operable cohort.

Fully 57% of the patients in the national registry had a surgical risk score less than 8%, whereas only 12% fit into the with a PROM score over 15%.

The registry included only use of the first-to-market Sapien valve. The period it covered from , through July 2013 predated .

Yet the registry was more on par with the 7.3% average PROM risk score in the pivotal CoreValve trial reported at the conference days earlier.

Clinicians may be selecting slightly lower-risk patients based on factors difficult to capture in the surgical risk score, Holmes suggested.

"It's frailty that determines eligibility" for TAVI, and that's a hard-to-define criterion, agreed , of the St. Vincent Heart Center in Indianapolis, who also was on the discussion panel at the meeting.

But if there is risk creep, the CoreValve findings were at least reassuring that a population at a similar average risk gained a survival advantage from TAVI compared with surgical valve replacement, , of the Detroit Medical Center Vascular Institute, suggested at a press conference she chaired.

"I'm amazed and find it remarkable that we in this country brought almost 200 new centers online and over 20 months had those kind of data," Brown noted.

The registry results were also in the range of what has been reported in registries from countries that approved the technology earlier and from small U.S. registries. Those showed 1-year mortality ranging from 14.3% to 26.0%, and stroke rates of 4.1% to 7.1%.

The joint ACC/STS registry, mandated as part of CMS coverage for TAVI (also called transcatheter aortic valve replacement, or TAVR) pulled in data for nearly all valves implanted across the country during the study period and collected outcomes through Medicare claims for the 5,980 patients ages 65 and older.

The only factor significantly linked to stroke risk in the year after TAVI was female gender (hazard ratio 0.65, 95% CI 0.47-0.91).

Factors linked to mortality were broader and not necessarily surprising, including older age, male sex, greater severity of chronic obstructive lung disease, poorer renal function, and higher surgical risk score, which is seen as a marker of frailty.

"Identification of these associations is essential for developing risk prediction models and will aid in patient selection criteria for TAVR," Holmes said at the late-breaking clinical trial session.

Notably, alternative access routes for the procedure also predicted a higher risk of death by 1 year.

Radial or other access routes carried a relative 42% higher likelihood of mortality compared with transfemoral access, which accounted for 63.7% of cases. The rates at 1 year were 28.8% versus 24.4% (P<0.001).

, of Columbia University Medical Center/New York-Presbyterian Hospital, expressed skepticism at the session as to whether this was simply unmeasured confounding.

Holmes agreed that the majority of the difference may have been largely accounted for by the many factors that go into the determination that a patient may not be a good candidate for transfemoral access, often related to risk of bleeding.

"The higher mortality in that particular cohort cannot be discounted," , of Piedmont Heart Institute in Atlanta, cautioned as a discussant at the press conference.

It will be interesting to watch for trends over time in the database, he added, "particularly as new technologies are introduced into this registry and ... this therapy being diffused into nonclinical sites, nonhigh volume sites as well. We should probably only expect the outcomes in this registry to improve. These are probably the highest event rates we will see."

Comment