Long-term durability of bioprosthetic aortic valves: implications from 12,569 implants

Abstract

Background: Increased life expectancy and younger patients' desire to avoid lifelong anticoagulation requires a better understanding of bioprosthetic valve failure. This study evaluates risk factors associated with explantation for structural valve deterioration (SVD) in a long-term series of Carpentier-Edwards PERIMOUNT aortic valves (AV).

Methods: From June 1982 to January 2011, 12,569 patients underwent AV replacement with Edwards Lifesciences Carpentier-Edwards PERIMOUNT stented bovine pericardial prostheses, models 2700PM (n = 310) or 2700 (n = 12,259). Mean age was 71 ± 11 years (range, 18 to 98 years). 93% had native AV disease, 48% underwent concomitant coronary artery bypass grafting, and 26% had additional valve surgery. There were 81,706 patient-years of systematic follow-up data available for analysis. Demographics, intraoperative variables, and 27,386 echocardiographic records were used to identify risks for explant for SVD and assess longitudinal changes in transprosthesis gradients using time-varying covariable analyses.

Results: Three hundred fifty-four explants were performed, with 41% related to endocarditis and 44% to SVD. Actuarial estimates of explant for SVD at 10 and 20 years were 1.9% and 15% overall, respectively, and in patients younger than 60 years, 5.6% and 46%, respectively. Younger age (p < 0.0001), lipid-lowering drugs (p = 0.002), prosthesis-patient mismatch (p = 0.001), and higher postoperative peak and mean AV gradients were associated with explant for SVD (p < 0.0001). The effect of gradient on SVD was greatest in patients younger than 60 years.

Conclusions: Durability of the Carpentier-Edwards PERIMOUNT aortic valve is excellent even in younger patients. Explant for SVD is related to gradient at implantation, especially in younger patients. Strategies to reduce early postoperative AV gradients, such as root enlargement or more efficient prostheses, should be considered.

Copyright © 2015 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.

Figures

Fig 1

Instantaneous time-varying risk (hazard function) of prosthetic valve explantation. Solid lines depict parametric estimates of risk for explants enclosed within a 68% confidence band equivalent to ±1 standard error. (A) Overall time-varying risk for explant for any reason. Number of patients remaining at risk is given below the horizontal axis. (B) Cause-specific time-varying risk for explant. (SVD = structural valve deterioration.)

Fig 2

Competing risks of structural valve deterioration (SVD), explantation for other indications, and death before explant of a stented pericardial aortic valve prosthesis. Solid lines represent point estimates enclosed within a 68% confidence band equivalent to ±1 standard error. (A) Instantaneous risk of competing events. (B) Cumulative incidence of each competing risk. Vertical bars represent 68% confidence limits.

Fig 3

Age and probability of explant owing to structural valve deterioration (SVD). (A) Nomogram of age relationship to SVD from multivariable equation based on preoperative variables alone. (B) Patients are grouped according to age range. Each symbol represents an explant, vertical bars are 68% confidence limits, and numbers along the horizontal axis are patients remaining at risk.

Fig 4

Structural valve deterioration (SVD) at 20 years and prosthesis–patient mismatch, represented by the number of standard deviations the geometric size of the aortic prosthesis deviates from normal. Nomogram is based on preoperative variables alone.

Fig 5

Temporal trend of prevalence of aortic regurgitation (AR) grades 3+/4+ after aortic valve replacement stratified by patients with eventual explant for structural valve deterioration (SVD) versus other patients. Solid lines represent parametric estimates of temporal trend of prevalence of AR grade 3+/4+. Symbols represent data grouped without regard to repeated measurements within time frames to provide crude verification of model fit.

Fig 6

Temporal trend of aortic valve (AV) gradients after AV replacement stratified by patients with eventual explant for structural valve deterioration (SVD) versus other patients. Solid lines represent parametric estimates enclosed within 68% bootstrap percentile confidence intervals. Symbols represent data grouped without regard to repeated measurements within time frame to provide crude verification of model fit. (A) AV mean gradient after AV replacement. (B) AV peak gradient after AV replacement.

Fig 7

Explantation for structural valve deterioration (SVD) and postoperative mean transvalvular pressure gradient. (A) Unadjusted relationship between instantaneous risk of explant owing to SVD (left vertical axis) and temporal trend of mean postoperative aortic valve (AV) mean gradient (right vertical axis). Solid lines represent risk of explant for SVD; dashed lines represent 3 patient-specific profiles of postoperative AV mean gradient. Blue lines (top) represent the trend for a patient whose profile is at the 85th percentile. Purple lines (middle) represent the trend for a patient whose profile is at the 50th percentile. Red lines (bottom) represent the trend for a patient whose profile is at the 15th percentile. (B) Explant owing to SVD by 20 years (left vertical axis) according to postoperative AV peak gradient and age at implantation, with dashed lines representing 68% confidence bands. This is a nomogram of the multivariable equation in Table 2.