Results:

From 2008 through 2017, 306 patients from 80 centers underwent TTVR after prior surgical TV replacement (valve-in-valve, n = 284; 93%) or repair (valve-in-ring, n = 22) and were included in this analysis. Fifty-two of these patients (17%) had a prior history of endocarditis, which was the underlying cause of TV dysfunction that ultimately led to surgical replacement or repair. TTVR was performed with a Melody valve in 138 patients (45%) or a Sapien valve in 168 (55%).

In seven patients the implanted valve malpositioned or embolized; four of these patients had a second valve implanted during the same catheterization resulting in elimination of paravalvular regurgitation, two had the valve repositioned to the right atrium or inferior vena cava and underwent surgery emergently (n = 1) or several months later (n = 1), and one with a stable valve but paravalvular leakage underwent surgical TVR later during the same hospitalization.

After TTVR, the mean TV inflow gradient and TR grade measured by echocardiography improved significantly. There was no difference in the early post-intervention gradient according to indication for TTVR, transcatheter valve type, underlying congenital versus acquired TV disease, or prior endocarditis history. Patients with a surgical TV prosthesis size 29 mm or larger had a lower post-TTVR gradient than those with smaller valves (3.6 ± 1.8 mm Hg vs. 4.2 ± 2.3 mm Hg; p = 0.02).

Patients were followed for a median of 15.9 months after implantation. During that time, 36 patients died, eight within 30 days (2.6%). Death was related to procedural complications in two patients, cardiovascular causes unrelated to the procedure in 20, and noncardiovascular causes in 14. Older age (hazard ratio [HR], 1.03 per year) and patient acutely ill and hospitalized before TTVR (HR, 4.4) independently predicted death over time. Prior history of endocarditis and other baseline or procedural factors were not associated with shorter survival and there was no difference in valve-related outcomes according to TTVR valve type.

Thirty-one patients (10%) underwent reintervention on the TV during follow-up: surgical TVR in 18 patients, a second TTVR within the first for malposition or dysfunction in eight patients, re-dilation of the original TTVR in three patients, and device closure of paravalvar leak in two patients. The cumulative incidence of reintervention was 5% at 1 year, 12% at 3 years, and 19% at 5 years. Higher early post-TTVR inflow gradient was associated with shorter freedom from reintervention. Annualized incidence of endocarditis was 1.5% per patient-year.

Eight patients were diagnosed with clinically relevant valve thrombosis (n = 4) or presumed thrombus (immobility or thickening of the leaflets; n = 4), three within several days of TTVR, two within 2 months (early), and three beyond 6 months (late). Higher immediate post-TTVR inflow gradient was associated with increased risk of thrombosis (HR, 1.38 per mm Hg).