Abstract
Objective:
The ideal valve substitute for surgical intervention of congenital aortic valve disease in children remains unclear. Data on outcomes beyond 10-15 years after valve replacement are limited but important for evaluating substitute longevity. We aimed to describe up to 25-year death/cardiac transplant by type of valve substitute and assess the potential impact of treatment center. Our hypothesis was that patients with pulmonic valve autograft would have better survival than mechanical prosthetic.
Methods:
This is a retrospective cohort study from the Pediatric Cardiac Care Consortium, a multi-institutional US-based registry of pediatric cardiac interventions, linked with the National Death Index and United Network for Organ Sharing through 2019. Children (0-20 years old) receiving aortic valve replacement (AVR) 1982-2003 were identified. Kaplan-Meier transplant-free survival was calculated, and Cox proportional hazard models estimated hazard ratios for mechanical AVR (M-AVR) versus pulmonic valve autograft.
Results:
Among 911 children, the median age at AVR was 13.4 years (IQR=8.4-16.5) and 73% were male. There were 10 cardiac transplants and 153 deaths, 5 after transplant. The 25-year transplant-free survival post AVR was 86.1% for autograft vs 75.9% for M-AVR and 71.2% for tissue (bioprosthetic or homograft). After adjustment, M-AVR remained related to increased mortality/transplant vs autograft (HR=1.9, 95% CI=1.1-3.4). Surprisingly, survival for patients with M-AVR, but not autograft, was lower for those treated in centers with higher in-hospital mortality.
Conclusions:
Pulmonic valve autograft provides the best long-term outcomes for children with aortic valve disease, but AVR results may depend on a center’s experience or patient selection.
Keywords: congenital heart disease, aortic valve replacement, children, long-term outcomes
INTRODUCTION
Aortic valve replacement (AVR) is less common in children than adults but necessary to treat congenital aortic valve abnormalities when repair is not feasible or has failed.1, 2 Available substitutes for the native valve include pulmonic valve autograft, also known as the Ross procedure (R-AVR), replacement with a mechanical (M-AVR), or tissue-derived valve (T-AVR) [animal bioprosthetic (B-AVR), or homograft (H-AVR)]. As each type has its own advantages and limitations, no substitute is ideal for all children.2-5 Current literature suggests an advantage in R-AVR for up to 10-15 years of follow-up because of its durability, physiologic flow profile, ability to grow with the child, and no need for anticoagulation.4-6 American Heart Association/American College of Cardiology recommendations suggest consideration of R-AVR in young patients when anticoagulation with vitamin K antagonists is contraindicated or undesirable and the operation is performed by an experienced surgeon.7
Multi-institutional comparative data capturing the wider experience after AVR in children and the impact of risk factors beyond 15 years are limited.6, 8, 9 Although recommendations include surgeon’s experience, and R-AVR is a technically challenging procedure, previous research has not addressed this factor when comparing outcomes. We have previously published on patients 1-5 years old,10 but we now aim to describe long-term (up to 25 years) transplant-free survival by type of AVR, focusing on R-AVR and M-AVR, among all children (<21 years old) and assess the impact of patient and treatment center characteristics. We hypothesized that those with R-AVR would have superior outcomes, but these advantages would be dependent on the center’s experience with complex congenital heart surgeries.
METHODS
Data Sources
This study used retrospective cohort data from the Pediatric Cardiac Care Consortium (PCCC), a US-based registry of cardiac surgical and trans-catheter interventions in 47 participating centers between 1982 and 2011.11 Survival or transplant status was ascertained by PCCC records and linkage with the National Death Index (NDI) and Organ Procurement and Transplant Network (OPTN) respectively through the end of 2019 (see eMethods for additional information).12, 13
We included all patients undergoing their first AVR at <21 years old who were US residents, treated in a US center, and had available direct identifiers allowing linkage with the NDI and OPTN. Patients had to receive their first AVR in the PCCC, to avoid immortal person time bias, and before April 15, 2003, for full identifiers to be available due to changes in the Health Insurance Portability and Accountability Act. Patients with acquired aortic valve disease (i.e. rheumatic heart disease, injury, or endocarditis without congenital valve disease), single ventricle physiology, or connective tissue disorders (i.e. Marfan syndrome) were excluded to focus on children with two-ventricle congenital heart disease (CHD) for whom the trajectory of aortic valve disease begins at birth. Other genetic conditions were considered confounders and adjusted for in the model. Finally, patients with missing surgical records, and therefore undetermined type of AVR, were excluded.
Analyses
Patient characteristics were obtained from PCCC records and compared by type of AVR using chi-squared or Fisher’s exact test when an expected cell count was less than five and the Kruskal-Wallis test for continuous variables since normality was not assumed. Children with AVR <1 year of age were analyzed separately because few were treated with M-AVR and this group had very different characteristics. The presence of other left heart obstructive lesions was noted for stenosis of the mitral valve, subaortic area, or aortic arch. Complex heart disease designates patients with additional cardiac anomalies outside of left heart obstructive lesions including double outlet right ventricle (DORV), transposition of the great arteries (TGA), pulmonary atresia, tetralogy of Fallot (TOF), and truncus arteriosus. The year of surgery was split into approximate tertiles.
Treatment center characteristics were total annual surgical volume, proportion of operations classified as “high-risk” (scores 4-6) by the Risk Adjustment for Congenital Heart Surgery version 1 (RACHS-1), and the in-hospital mortality for high-risk procedures. Each was calculated by aggregating all surgeries performed in the center in the six months before and after the date of the AVR, and categorization cut-points were based on associations in the data (see eMethods for more details).
In-hospital mortality at the time of AVR, pacemaker placement at AVR through 30 days after, and reoperations up to 10 years after were determined from PCCC records. Reoperations included subsequent surgeries involving the aortic valve and, for those originally receiving R-AVR, the pulmonary valve. The main outcome of interest, transplant-free survival, was defined as the time from hospital discharge after the AVR until cardiac transplant or death, or the end of the study (December 31, 2019), whichever came first. Kaplan-Meier (KM) transplant-free survival was compared by type of AVR using the log-rank test. The cumulative incidence of reoperations was compared between valve types using the Fine and Gray method considering death and cardiac transplant as competing events.
M- and R-AVR are the most frequently used techniques in children and were the focus of the multivariable analyses. To optimize exchangeability between treatment groups, patients not eligible to receive both types were excluded from this analysis: patients operated before 1991 (first year R-AVR was performed in PCCC) and those with pulmonary valve abnormalities making it unsuitable for use as aortic valve substitute (pulmonary atresia, TGA, TOF, truncus arteriosus communis, and some forms of DORV with TOF physiology). We estimated a propensity score for treatment with M-AVR and then excluded patients with extreme, nonoverlapping values, which would indicate a low probability of receiving the alternative valve type. Within this restricted group, the adjusted association between M-AVR vs R-AVR and other clinical characteristics with long-term survival was assessed with a Cox hazards model accounting for clustering within centers using generalized estimating equations with an exchangeable correlation structure. The proportional hazards assumption was assessed using log-log curves and extended Cox models. Sensitivity analyses to evaluate the robustness of our findings included 1)fine stratification weighting with the propensity score to ensure the balance of clinical factors between the M-AVR and R-AVR groups, 2)exclusion of patients with diagnosed genetic abnormalities, 3)estimation of the hazard ratio for transplant/causes of death other than endocarditis, 4)test of interaction of underlying valve hemodynamics, era of surgery, and center level mortality with high-risk cardiac procedures, and 5)re-dichotomization of percent mortality for high-risk procedures at the median (see eMethods for details).
Analyses were conducted using SAS 9.4 (Cary, NC). The study was approved by the Institutional Review Boards of Emory University School of Medicine, Children’s Mercy Hospital, the University of Minnesota, and University of Georgia and was approved by the NDI and the OPTN.
Patient and Public Involvement
Due to the retrospective nature and secondary linkage of data, no patients or public were directly involved in this study.
RESULTS
There were 1,242 eligible pediatric patients identified with AVR. Of these, 233 did not have adequate identifiers for linkage and 32 did not have complete records, leaving 977 patients from 43 centers eligible for this study (eFigure 1 provides a flow chart of eligibility for each analysis). The distributions of types of AVR through the study period and by age groups are presented in eFigure 2.
AVR < 1 Year of Age
Among eligible patients, 66 were infants. Most (57.6%) received R-AVR, and in-hospital mortality was 56% (eTable 1). Long-term survival among those discharged alive was not statistically significantly lower than other age groups overall (p-value =0.64, Figure 1) or among those receiving R-AVR (25-year transplant-free survival:78.6% vs 92.6%, 81.0% and 87.4% for 1-5, 6-10 and 11-20 years old respectively, p=0.25, 0.81, 0.66) (eFigure 3).
Figure 1.
Long-term transplant-free survival after AVR discharge by age group (infants – 20 years old)a
AVR 1-20 Years of Age
R-AVR was the most frequent method (45.7%), followed by M-AVR (42.9%) and T-AVR (11.4%) (Table 1). Most R-AVR were in the form of full root (92.8%) instead of subcoronary. Median age at AVR was 13.4 years (IQR=8.4-16.5) with those receiving R-AVR being slightly younger. Most patients (51.3%) were treated in a large CHD surgical volume center, and this did not differ by type of AVR. In centers treating those with M-AVR, the proportion of surgeries considered high-risk was lower and their in-hospital mortality for high-risk surgeries was higher, although not statistically significant (Table 2). Since this study includes operations across 20 years, changes in patient and center characteristics across the study eras are shown in eTables 2 and 3.
Table 1.
Characteristics of patients receiving AVR at 1-20 year of age for congenital aortic valve disease by valve substitute
| Patient characteristics | Total N = 911 | R-AVR N = 416 (45.7%) | M-AVR N = 391 (42.9%) | T-AVR N = 104 (11.4%) | p-value a |
|---|---|---|---|---|---|
| Males | 669 (73.4) | 294 (70.7) | 300 (76.7) | 75 (72.1) | 0.14 |
| Age at AVR-Median (IQR) | 13.4 (8.4 – 16.5) | 12.6 (7.0-16.1) | 13.9 (9.3-16.9) | 13.9 (8.7-16.5) | <0.01 |
| 1 - 5 years | 143 (15.7) | 84 (20.2) | 40 (10.2) | 19 (18.3) | |
| 6 – 10 years | 199 (21.8) | 95 (22.8) | 85 (21.7) | 19 (18.3) | |
| 11 - 20 years | 569 (62.5) | 237 (57.0) | 266 (68.0) | 66 (63.5) | |
| Era of AVR (~tertiles) | <0.001 | ||||
| 1982 - 1994 | 265 (29.1) | 40 (9.6) | 192 (49.1) | 33 (31.7) | |
| 1995 - 1998 | 348 (38.2) | 194 (46.6) | 119 (30.4) | 35 (33.7) | |
| 1999 - 2003 | 298 (32.7) | 182 (43.8) | 80 (20.5) | 36 (34.6) | |
| Valve hemodynamics at AVR | <0.001 | ||||
| Stenosis | 105 (11.5) | 50 (12.0) | 47 (12.0) | 8 (7.7) | |
| Regurgitation | 279 (30.6) | 59 (14.2) | 174 (44.5) | 46 (44.2) | |
| Both | 527 (57.9) | 307 (73.8) | 170 (43.5) | 50 (48.1) | |
| Other left heart obstruction b | 242 (26.6) | 107 (25.7) | 110 (28.1) | 25 (24.0) | 0.32 |
| Present at 1 level | 169 (18.6) | 70 (16.8) | 84 (21.5) | 15 (14.4) | |
| Present at 2 levels | 65 (7.1) | 33 (7.9) | 22 (5.6) | 10 (9.6) | |
| Present at 3 levels | 8 (0.9) | 4 (1.0) | 4 (1.0) | 0 (0.0) | |
| Genetic conditions (all) | 68 (7.5) | 22 (5.3) | 32 (8.2) | 14 (13.5) | 0.01 |
| Down syndrome | 11 (1.2) | 3 (0.7) | 7 (1.8) | 1 (1.0) | 0.38 |
| Otherc | 57 (6.3) | 19 (4.6) | 25 (6.4) | 13 (12.5) | 0.01 |
| Concurrent procedures | 123 (13.5) | 35 (8.4) | 75 (19.2) | 13 (12.5) | <0.001 |
| Konno procedure | 103 (11.3) | 34 (8.2) | 58 (14.8) | 11 (10.6) | 0.01 |
| Mitral valve repair | 24 (2.6) | 0 (0.0) | 22 (5.6) | 2 (1.9) | <0.001 |
| Arch repair | 7 (0.8) | 2 (0.5) | 4 (1.0) | 1 (1.0) | 0.55 |
| Pacemaker placementd | 44 (4.8) | 17 (4.1) | 24 (6.1) | 3 (2.9) | 0.27 |
| Prior cardiac interventions | 528 (58.0) | 210 (50.5) | 253 (64.7) | 65 (62.5) | <0.001 |
| Balloon valvuloplasty | 142 (15.6) | 99 (23.8) | 30 (7.7) | 13 (12.5) | <0.001 |
| Aortic valve repair | 326 (35.8) | 143 (34.4) | 143 (36.6) | 40 (38.5) | 0.67 |
| Aortic arch repair | 100 (11.0) | 50 (12.0) | 38 (9.7) | 12 (11.5) | 0.69 |
| Pacemaker placement | 20 (2.2) | 2 (0.5) | 17 (4.4) | 1 (1.0) | <0.001 |
| Complex heart disease e | 82 (9.0) | 4 (1.0) | 68 (17.4) | 10 (9.6) | <0.001 |
p-values comparing the distribution of categorical characteristics across valve types were estimated using chi-squared or Fisher’s exact test when at least one expected cell count was less than 5 (left heart obstruction, Down’s syndrome, concurrent mitral valve repair, arch repair or pacemaker placement); Kruskal-Wallis test was used for continuous factors since normality was not assumed
Left heart obstruction was considered as stenosis at the mitral valve, sub-aortic area, or aortic arch
Other syndromes include Digeorge, Goldenhar, Klippel-Feil, Noonan, Osteogenesis Imperfecta, Pierre-Robin, Triple X, Trisomy 15, Turner Syndrome, and Vater Syndrome
During or up to 30 days after AVR procedure
Includes: double outlet right ventricle, dextro- or levo-transposed great arteries, tetralogy of Fallot, truncus arteriosus communis, pulmonary atresia
Table 2.
Characteristics of treatment center treating 1-20 year-old patients in the 12 months around the AVR
| Patients with center data | Total N=823 | R-AVR N=389 (47.3%) | M-AVR N=342 (41.6%) | T-AVR N=92 (11.2%) | p-valuea |
|---|---|---|---|---|---|
| Unique centers - N | 43 | 37 | 40 | 26 | |
| Case volume per 12 mos - Median (IQR) | 204 (145-29) | 204 (145-304) | 206 (145-274) | 196 (136-287) | 0.55 |
| Small (10-99) | 87 (10.6)b | 35 (9.0) | 41 (12.0) | 11 (12.0) | |
| Medium (100-199) | 314 (38.2) | 157 (40.4) | 120 (35.1) | 37 (40.2) | 0.46 |
| Large (≥ 200) | 422 (51.3) | 197 (50.6) | 181 (52.9) | 44 (47.8) | |
| % of high-risk operations – Median (IQR) | 9.8(7.7-12.0) | 10.3 (8.4-12.3) | 9.2 (7.3-11.2) | 10.3 (8.0-12.4) | <0.01 |
| <9.4% | 377 (45.8) | 159 (40.1) | 182 (53.2) | 36 (39.1) | <0.01 |
| ≥9.4% | 446 (54.2) | 230 (59.1) | 160 (46.8) | 56 (60.9) | |
| High risk postoperative mortality - Median (IQR) | 2.4(1.4 – 3.5) | 2.3 (1.3 – 3.5) | 2.5 (1.5 – 3.7) | 2.2 (1.1 – 3.4) | 0.11 |
| Cases in centers <3.8% mortality | 646 (78.5) | 311 (80.0) | 261 (76.3) | 74 (80.4) | 0.44 |
| Cases in centers ≥3.8% mortality | 177 (21.5) | 78 (20.1) | 81 (23.7) | 18 (19.6) |
p-values comparing the distribution of categorical characteristics across valve types were estimated using chi-squared and the Kruskal-Wallis test was used for continuous factors since normality was not assumed
Numbers indicate the count (column percent) unless otherwise specified
In-Hospital Outcomes
A total of 888 patients (97.5%) survived to hospital discharge after their initial AVR. Survival was highest among patients receiving R-AVR (Table 3). After adjustment for the other characteristics, the odds of in-hospital death after M-AVR were 3.6 times that of R-AVR (95%CI=1.0-12.7). Other characteristics statistically significantly related to increased odds of in-hospital mortality were stenotic native aortic valve and AVR at ages 6-10 vs 11-20. (eTable 4).
Table 3.
Association of patient characteristics with long-term survival after AVR at 1-20 years of age.
| Total | In-hospital survival after AVR | Long-term transplant-free survival after hospital discharge from AVR | |||||||
|---|---|---|---|---|---|---|---|---|---|
| N | N (%) | p-valuea | 1-year | 5-year | 10-year | 20-year | 25-year | p-valueb | |
| Total | 911 | 888 (97.5) | - | 96.7 | 95.3 | 91.8 | 85.0 | 80.1 | - |
| Valve type | |||||||||
| R-AVR | 416 | 411 (98.8) | 97.8 | 97.1 | 95.6 | 91.2 | 86.1 | ||
| M-AVR | 391 | 377 (96.4) | 0.07 | 94.7 | 92.6 | 88.3 | 80.7 | 75.9 | <0.001 |
| T-AVR | 104 | 100 (96.1) | 100.0 | 98.0 | 89.0 | 76.1 | 71.2 | ||
| Sex | |||||||||
| Male | 669 | 653 (97.6) | 0.67 | 96.5 | 95.0 | 91.9 | 84.5 | 79.8 | 0.73 |
| Female | 242 | 235 (97.1) | 97.5 | 96.2 | 91.5 | 86.4 | 81.0 | ||
| Age at AVR | |||||||||
| 1 – 5 years | 143 | 138 (96.5) | 97.8 | 94.9 | 93.5 | 87.0 | 82.0 | ||
| 6 – 10 years | 199 | 192 (96.5) | 0.27 | 97.4 | 96.4 | 92.2 | 85.8 | 81.8 | 0.90 |
| 11 - 20 years | 569 | 558 (98.1) | 96.2 | 95.0 | 91.2 | 84.5 | 80.0 | ||
| Era of AVR (~tertiles) | |||||||||
| 1982 - 1994 | 265 | 257 (97.0) | 97.3 | 95.7 | 92.6 | 84.4 | 80.9 | ||
| 1995 - 1998 | 348 | 336 (96.6) | 0.12 | 96.7 | 94.9 | 90.2 | 83.9 | - | 0.47 |
| 1999 - 2003 | 298 | 295 (99.0) | 96.3 | 95.3 | 92.9 | - | - | ||
| Pathophysiology at AVR | |||||||||
| Stenosis | 105 | 101 (96.2) | 98.0 | 95.1 | 89.1 | 83.7 | 82.5 | ||
| Regurgitation | 279 | 269 (96.4) | 0.18 | 95.5 | 94.1 | 89.6 | 80.4 | 75.6 | 0.03 |
| Both | 527 | 518 (98.3) | 97.1 | 96.0 | 93.4 | 87.7 | 81.8 | ||
| Other left heart obstruction c | |||||||||
| None | 669 | 661 (98.8) | 97.1 | 96.4 | 92.9 | 86.2 | 81.9 | ||
| Present at 1 level | 169 | 159 (94.1) | <0.01 | 95.0 | 91.8 | 89.3 | 82.4 | 76.2 | 0.22 |
| Present at 2 levels | 65 | 60 (92.3) | 96.7 | 93.3 | 88.3 | 80.0 | 69.8 | ||
| Present at 3 levels | 8 | 8 (100.0) | 100.0 | 87.5 | 75.0 | 75.0 | - | ||
| Genetic conditions d | |||||||||
| None | 843 | 824 (97.8) | 96.8 | 95.3 | 92.5 | 85.9 | 81.0 | - | |
| Down syndrome | 11 | 10 (90.9) | 0.23 | 100.0 | 100.0 | 90.0 | 90.0 | 90.0 | 0.53 |
| Other | 57 | 54 (94.7) | 0.16 | 94.4 | 94.4 | 81.5 | 71.1 | 64.6 | 0.002 |
| Concurrent procedure d | 123 | 116 (94.3) | 0.03 | 95.7 | 92.2 | 88.8 | 77.1 | 66.7 | 0.002 |
| Konno procedure | 103 | 96 (93.2) | 0.01 | 97.9 | 93.8 | 90.6 | 79.5 | 66.1 | 0.04 |
| Mitral valve repair | 24 | 23 (95.8) | 0.46 | 87.0 | 82.6 | 78.3 | 56.2 | 56.2 | <0.001 |
| Arch repair | 7 | 7 (100.0) | 1.00 | 100.0 | 100.0 | 100.0 | 100.0 | - | 0.28 |
| Pacemaker placemente | 44 | 42 (95.5) | 0.31 | 92.9 | 88.1 | 83.3 | 78.3 | 70.3 | 0.09 |
| None | 788 | 772 (98.0) | - | 96.9 | 95.7 | 92.2 | 86.2 | 81.8 | - |
| Prior cardiac interventions d | 528 | 506 (95.8) | <0.01 | 95.5 | 93.1 | 89.1 | 81.8 | 76.5 | 0.003 |
| Balloon valvuloplasty | 142 | 138 (97.2) | 0.77 | 97.8 | 97.1 | 94.9 | 89.6 | 87.9 | 0.09 |
| Aortic valve repair | 326 | 315 (96.6) | 0.22 | 96.5 | 94.9 | 91.1 | 84.1 | 80.0 | 0.99 |
| Aortic arch repair | 100 | 95 (95.0) | 0.10 | 95.8 | 90.5 | 87.4 | 80.6 | 72.2 | 0.22 |
| Pacemaker placement | 20 | 19 (95.0) | 0.40 | 84.2 | 79.0 | 79.0 | 62.4 | 50.0 | 0.002 |
| None | 383 | 382 (99.7) | - | 98.4 | 98.2 | 95.3 | 89.2 | 85.0 | - |
| Complex heart disease | |||||||||
| Yes | 82 | 79 (96.3) | 0.45 | 89.9 | 86.1 | 78.5 | 62.4 | 56.2 | <0.001 |
| No | 829 | 809 (97.6) | 97.4 | 96.2 | 93.1 | 87.2 | 82.4 | ||
| Case volume per 12 months | |||||||||
| Small (10-99) | 87 | 84 (96.6) | 97.6 | 96.4 | 95.2 | 84.5 | 79.9 | ||
| Medium (100-199) | 314 | 305 (97.1) | 0.70 | 97.1 | 95.7 | 90.8 | 83.0 | 77.8 | 0.75 |
| Large (≥ 200) | 422 | 413 (97.9) | 95.9 | 94.4 | 92.3 | 87.2 | 81.3 | ||
| % of high-risk cases | |||||||||
| < 9.4% | 377 | 369 (97.9) | 0.47 | 96.5 | 95.7 | 92.7 | 86.7 | 82.7 | 0.20 |
| ≥ 9.4% | 446 | 433 (97.1) | 96.5 | 94.7 | 91.5 | 84.2 | 76.4 | ||
| High risk postoperative mortality | |||||||||
| < 3.8% | 646 | 635 (98.3) | <0.01 | 97.0 | 95.6 | 93.2 | 87.1 | 81.9 | 0.01 |
| 177 | 167 (94.4) | 94.6 | 93.4 | 87.4 | 78.7 | 72.4 | |||
Chi-square or Fisher’s exact p-value when at least one expected cell count was less than 5 (left heart obstruction; Down’s syndrome; other genetic conditions; concurrent Konno, mitral valve repair, arch repair or pacemaker placement; prior balloon valvuloplasty, arch repair, or pacemaker placement; complex heart disease; and high-risk postoperative mortality)
log-rank p-value.
Left heart obstruction was considered as stenosis at the mitral valve, sub-aortic area, or aortic arch.
p-values for these characteristics are calculated separately for each group compared with the reference group of none.
During or up to 30 days after AVR procedure.
Pacemaker placement during or up to 30 days after AVR was somewhat higher in those with M-AVR (6.1%) compared with the other types (R-AVR=4.1%, T-AVR=2.9%) but not statistically significant (p-value=0.46) (Table 1). There was no significant association between pacemaker placement and in-hospital mortality (Table 3).
Long-Term Outcomes
Median follow-up time among survivors was 21.5 years (max=37, IQR=18.5–24.5). During this time, ten patients received a cardiac transplant, five later died, and 148 died without transplant with a median time to event of 11.1 years (IQR=4.6–17.9). Overall, transplant-free survival 25-years after hospital discharge was 80.7% (Table 3 and Figure 2, individual plots for B-AVR and H-AVR are in eFigure 4). Within R-AVR patients there was no difference in those who underwent the subcoronary technique (eFigure 5).
Figure 2.
Long-term transplant-free survival after AVR discharge by valve substitute (ages 1 – 20 years old)a
The majority of the 153 deaths occurring after discharge from AVR were attributed to the CHD (31.4%) or other cardiovascular disease (40.5%). The contributing cause of death did not differ by type of AVR except for endocarditis that contributed to 12.6% of deaths with M-AVR, but only in 2.7% of deaths with R-AVR or T-AVR (p=0.02) (eTable 5).
Valve reoperations occurred in 16.1% of patients by 10 years after AVR. This proportion was highest in those with T-AVR (29.1%), then R-AVR (16.3%), and M-AVR (12.5%) (p<0.01, eFigure 6).
R-AVR vs M-AVR
Of 596 patients identified for this comparison, 14 who received a mitral valve replacement concurrent to the AVR were excluded because they all underwent M-AVR. Another 23 were trimmed with extreme, nonoverlapping propensity scores leaving 372 undergoing R-AVR and 187 with M-AVR. In the adjusted Cox model, M-AVR had a 1.9 times greater rate of death/transplant than R-AVR (95%CI=1.1-3.4). Other factors related to greater death/transplant were male sex, any left heart obstruction, and a genetic syndrome (Figure 3). The covariate balance was improved with the fine stratification weighting so that all standardized differences were less than |0.1| (eTable 6). However, the increased hazard for M-AVR vs R-AVR did not meaningfully change after weighting or the other sensitivity analyses. There was no statistically significant interaction by underlying valve hemodynamics, but the effect of M-AVR vs R-AVR varied by era of surgery and the treatment center’s mortality rate for high-risk procedures. The increased rate of death/transplant for M-AVR persisted, but the hazard ratio was greatest in the most recent surgical era and for centers with higher mortality for high-risk procedures. Dichotomizing in-hospital mortality for high-risk procedures at the median led to a similar trend but attenuated the difference in hazard ratio observed (eTable 7). Exploring this further, those with R-AVR had similar 25-year survival regardless of center characterization (low vs. high: 85.9 vs. 91.0%). However, survival was worse for patients receiving M-AVR in centers with higher in-hospital mortality (low vs. high: 82.2 vs 66.8%) (Figure 4). Compared with 33.6% in the low-risk centers, 51.1% of M-AVR patients in high-risk centers had received prior aortic valve surgery (eTable 8), and among those with previous valve repair, subsequent M-AVR had worse survival (eFigure 7).
Figure 3.
Adjusted hazard ratios and 95% confidence intervals of mortality/transplant after AVR discharge among patients 1-20 years of age eligible for M-AVR or R-AVR
Figure 4.
Long-Term Survival after discharge among patients (1 - 20 years old) eligible for M-AVR or R-AVR by valve type received and center in-hospital mortality for high-risk cardiac procedures.
DISCUSSION
In this large multi-center study, the 25-year transplant-free survival for children receiving AVR for congenital aortic valve defects was 80% after being discharged alive. Of note, the attrition was not limited to the early postoperative years as seen in other surgically treated CHD where most deaths occur within 5 years after the index surgery. 13, 14 We also found a persistent, long-term survival advantage of R-vs M-AVR. In comparable patients, we estimated a two times greater rate of mortality/transplant in those with M-AVR. Endocarditis was higher in this group compared to the patients receiving any form of tissue valve but did not account for the full association.
In contrast to our hypothesis, the long-term outcomes for R-AVR did not change by center characteristics but were worse for those with M-AVR if done in a center with high in-hospital mortality for complex congenital heart surgery. Beyond simply being a spurious association, this effect modification could reflect center level experience related to patient selection and ineffective prior interventions, which may be evidenced by the high proportion receiving aortic valve repair prior to M-AVR in high-mortality centers. In support of this, among patients with prior surgical valve repair there was worse survival for those then receiving M-AVR instead of R-AVR. Another study found that among patients receiving R-AVR prior surgical repair improved long-term outcomes, potentially due to the ability to delay the valve replacement for a later age, but this is not a fair comparison as it induces a survival bias for those with a prior surgical repair.15 M-AVR was less common in the most recent era potentially indicating a general shift to R-AVR for most centers if the child was eligible for either procedure. Despite our attempt to identify exchangeable groups, those still receiving M-AVR may have more complex conditions. Alternatively, centers with higher surgical mortality and patients receiving M-AVR in the recent era may differ in other aspects of their underlying patient populations such as demographics and socioeconomic status. These factors have been related to short and long-term outcomes in the CHD population, but weren’t available in our cohort.16
In children, few studies have been able to report long-term outcomes after AVR and those that have were single-center studies.2, 4, 17 Meaningful comparison between different forms of aortic valve substitutes is challenging because of the difficulty in obtaining an adequate number of patients with similar key clinical characteristics for whom R-AVR is applicable. The strong potential for such biases is evidenced by the differences in characteristics we noted between those receiving each type of AVR and noted in baseline characteristics in a meta-analysis on this topic.18 Only one randomized trial is available which was limited to adults and compared R-AVR to homograft. Similar to our findings this study found a survival advantage of R-AVR (hazard ratio=4.6).19 Two other studies have attempted to use propensity score matching for comparison of outcomes and both suggested a similar long-term survival advantage for R-AVR in children. The first study was from a single center in Saudi Arabia5 with a case-mix weighted towards rheumatic aortic valve disease and the other was from the U.K.6 with predominant congenital aortic valve disease. Neither assessed potential center effects on outcomes. Importantly, both studies were limited to 10-15 years of follow-up, thus, not able to examine the effects of late developing sequelae post R-AVR. Since patients with R-AVR required higher reoperation for the right or left ventricular outflow tract, as has been found in other studies,3, 18 the extended follow-up time in our study is important to ease concerns over long-term complications expected to develop 15-20 years after the R-AVR.
The survival advantage of R-AVR is likely to be multifactorial including superior hemodynamics, the ability of the autograft valve to grow with the child and decreased risk for endocarditis, pacemaker implantation or complications from anticoagulation.4, 20-23 Despite the apparent advantage of the pulmonic valve autograft, some reports describe a decline in its use in recent years.24, 25 It is unclear whether this simply reflects clinicians’ concern about risks related to this procedure, more careful selection of patients, or more successful use of non-surgical techniques for treatment of aortic valve disease. These questions are beyond the scope of this study, but we note that our results did not show a significant improvement in survival over time.
LIMITATIONS
Despite the strengths of this study, the findings should be interpreted with several limitations. As with other observational studies, multivariable analysis and propensity score estimation/weighting can be performed only for data collected and therefore comparison of outcomes may be subject to residual confounding by important factors not included. For example, sociodemographic characteristics are not collected in the PCCC, although they may impact the treatment a patient receives and their long-term outcomes. Data on reoperations was only available in patients who underwent subsequent surgery at a PCCC center and therefore may be underestimated but is not expected to be differentially so. In fact, 10-year reoperation rates for R-AVR in our registry are within range of other contemporary reports (16.3% vs approximately 5->20% in other reports).9, 26 Some misclassification in other outcomes and risk factors is possible, but this is likely to be non-differential and therefore expected to bias results towards the null. Other meaningful outcomes related to type of AVR and patient quality of life could not be assessed but should be considered. Given the need for anticoagulation, increased risk of complete heart block and endocarditis,4 tickling noise and lower patient-reported quality of life with M-AVR,27 there is reason to suggest that R-AVR would have the advantage there as well. A strength of this study is its inclusion of center-level information, but the recommendations specify the importance of the surgeon which could not be directly evaluated from the available data. The cut-points used for the center level characteristics were data-driven due to the lack of a known clinically relevant breakpoint, and although the trend persisted, the difference in hazard ratios was no longer statistically significant after dichotomization at the median. However, this is what we would expect if there were some sort of threshold effect within the top fifty percent because grouping those with a high hazard above the threshold with those with a lower hazard between the threshold and median value would attenuate the estimated hazard ratio as observed. Finally, despite using a large, multi-center cohort, we were limited by small numbers to fully explore interaction by treatment center characteristics.
CONCLUSIONS
This study suggests that R-AVR has a significant survival benefit compared to M-AVR up to 25 years post AVR, which is important given the concerns for aortic root-related complications and need for right ventricular outflow tract reinterventions after R-AVR8, 28, 29 The disadvantage of the M-AVR was surprisingly greater in centers with less success at high-risk procedures. Our findings may indicate hesitancy for some centers to conduct more complex surgeries resulting in less ideal substitutes and failed repairs when initial replacement may have led to better long-term outcomes.
Supplementary Material
KEY QUESTIONS.
What is already known about the subject?
Previous evaluation of outcomes up to 10-15 years after aortic valve replacement (AVR) in children generally favor the pulmonic valve autograft to mechanical. However, longer-term outcomes and the impact of treatment center characteristics, despite the increased complexity of pulmonic autograft, have not been evaluated.
What does this study add?
This study extends the data on outcomes after AVR up to 25 years, suggesting a persistent long-term survival benefit of pulmonary autograft compared to mechanical, which is important given the concerns for aortic root-related complications and need for right ventricular outflow track reinterventions after autograft. The advantage of the pulmonic valve autograft was surprisingly greater in centers with less success at high-risk procedures.
How might this impact on clinical practice?
These results suggest that more emphasis should be placed on the pulmonic valve autograft technique in pediatric surgical training to maximize its benefit in children with unrepairable aortic valve disease.
ACKNOWLEDGEMENTS
The authors thank all participating PCCC centers and their staffs. We also thank Ms. Amanda Thomas at Emory University for her support.
SOURCES OF FUNDING
This study was supported by National Heart, Lung, and Blood Institute R01 HL122392 and the Department of Defense PR180683. Disclaimer: The data reported here have been supplied by UNOS as the contractor for the Scientific Organ Procurement and Transplantation Network (OPTN). The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the OPTN or the U.S. Government.
Footnotes
COMPETING INTERESTS
None declared.
REFERENCES:
- 1.Karamlou T, Jang K, Williams WG, Caldarone CA, Van Arsdell G, Coles JG and McCrindle BW. Outcomes and associated risk factors for aortic valve replacement in 160 children: a competing-risks analysis. Circulation. 2005;112:3462–9. [DOI] [PubMed] [Google Scholar]
- 2.Brown JW, Ruzmetov M, Vijay P, Rodefeld MD and Turrentine MW. Surgery for aortic stenosis in children: a 40-year experience. Ann Thorac Surg. 2003;76:1398–411. [DOI] [PubMed] [Google Scholar]
- 3.Alsoufi B, Al-Halees Z, Manlhiot C, McCrindle BW, Al-Ahmadi M, Sallehuddin A, Canver CC, Bulbul Z, Joufan M and Fadel B. Mechanical valves versus the Ross procedure for aortic valve replacement in children: propensity-adjusted comparison of long-term outcomes. J Thorac Cardiovasc Surg. 2009;137:362–370 e9. [DOI] [PubMed] [Google Scholar]
- 4.Brown JW, Patel PM, Ivy Lin JH, Habib AS, Rodefeld MD and Turrentine MW. Ross Versus Non-Ross Aortic Valve Replacement in Children: A 22-Year Single Institution Comparison of Outcomes. Ann Thorac Surg. 2016;101:1804–10. [DOI] [PubMed] [Google Scholar]
- 5.Alsoufi B Aortic valve replacement in children: Options and outcomes. J Saudi Heart Assoc. 2014;26:33–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sharabiani MT, Dorobantu DM, Mahani AS, Turner M, Peter Tometzki AJ, Angelini GD, Parry AJ, Caputo M and Stoica SC. Aortic Valve Replacement and the Ross Operation in Children and Young Adults. J Am Coll Cardiol. 2016;67:2858–70. [DOI] [PubMed] [Google Scholar]
- 7.Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP 3rd, Fleisher LA, Jneid H, Mack MJ, McLeod CJ, O’Gara PT, Rigolin VH, Sundt TM 3rd and Thompson A. 2017 AHA/ACC Focused Update of the 2014 AHA/ACC Guideline for the Management of Patients With Valvular Heart Disease: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation. 2017;135:e1159–e1195. [DOI] [PubMed] [Google Scholar]
- 8.Bansal N, Kumar SR, Baker CJ, Lemus R, Wells WJ and Starnes VA. Age-Related Outcomes of the Ross Procedure Over 20 Years. Ann Thorac Surg. 2015;99:2077–83; discussion 2084-5. [DOI] [PubMed] [Google Scholar]
- 9.Nelson JS, Pasquali SK, Pratt CN, Yu S, Donohue JE, Loccoh E, Ohye RG, Bove EL and Hirsch-Romano JC. Long-Term Survival and Reintervention After the Ross Procedure Across the Pediatric Age Spectrum. Ann Thorac Surg. 2015;99:2086–94; discussion 2094-5. [DOI] [PubMed] [Google Scholar]
- 10.Alsoufi B, Knight JH, St Louis J, Raghuveer G and Kochilas L. Are Mechanical Prostheses Valid Alternatives To Ross Procedure In Young Children Under 6 Years Old? Ann Thorac Surg. 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Vinocur JM, Moller JH and Kochilas LK. Putting the Pediatric Cardiac Care Consortium in context: evaluation of scope and case mix compared with other reported surgical datasets. Circ Cardiovasc Qual Outcomes. 2012;5:577–9. [DOI] [PubMed] [Google Scholar]
- 12.Spector LG, Menk JS, Vinocur JM, Oster ME, Harvey BA, St Louis JD, Moller J and Kochilas LK. In-Hospital Vital Status and Heart Transplants After Intervention for Congenital Heart Disease in the Pediatric Cardiac Care Consortium: Completeness of Ascertainment Using the National Death Index and United Network for Organ Sharing Datasets. J Am Heart Assoc. 2016;5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Spector LG, Menk JS, Knight JH, McCracken C, Thomas AS, Vinocur JM, Oster ME, St Louis JD, Moller JH and Kochilas L. Trends in Long-Term Mortality After Congenital Heart Surgery. J Am Coll Cardiol. 2018;71:2434–2446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Smith CA, McCracken C, Thomas AS, Spector LG, St Louis JD, Oster ME, Moller JH and Kochilas L. Long-term Outcomes of Tetralogy of Fallot: A Study From the Pediatric Cardiac Care Consortium. JAMA Cardiol. 2019;4:34–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Buratto E, Wallace FRO, Fricke TA, Brink J, d’Udekem Y, Brizard CP and Konstantinov IE. Ross Procedures in Children With Previous Aortic Valve Surgery. J Am Coll Cardiol. 2020;76:1564–1573. [DOI] [PubMed] [Google Scholar]
- 16.Pace ND, Oster ME, Forestieri NE, Enright D, Knight J and Meyer RE. Sociodemographic Factors and Survival of Infants With Congenital Heart Defects. Pediatrics. 2018;142. [DOI] [PubMed] [Google Scholar]
- 17.Bove T, Bradt N, Martens T, De Wolf D, Francois K, de Beco G, Sluysmans T, Rubay J and Poncelet A. The Pulmonary Autograft After the Ross Operation: Results of 25-Year Follow-Up in a Pediatric Cohort. Ann Thorac Surg. 2021;111:159–167. [DOI] [PubMed] [Google Scholar]
- 18.Etnel JR, Elmont LC, Ertekin E, Mokhles MM, Heuvelman HJ, Roos-Hesselink JW, de Jong PL, Helbing WA, Bogers AJ and Takkenberg JJ. Outcome after aortic valve replacement in children: A systematic review and meta-analysis. J Thorac Cardiovasc Surg. 2016;151:143–52 e1–3. [DOI] [PubMed] [Google Scholar]
- 19.El-Hamamsy I, Eryigit Z, Stevens LM, Sarang Z, George R, Clark L, Melina G, Takkenberg JJ and Yacoub MH. Long-term outcomes after autograft versus homograft aortic root replacement in adults with aortic valve disease: a randomised controlled trial. Lancet. 2010;376:524–31. [DOI] [PubMed] [Google Scholar]
- 20.Rabkin-Aikawa E, Aikawa M, Farber M, Kratz JR, Garcia-Cardena G, Kouchoukos NT, Mitchell MB, Jonas RA and Schoen FJ. Clinical pulmonary autograft valves: pathologic evidence of adaptive remodeling in the aortic site. J Thorac Cardiovasc Surg. 2004;128:552–61. [DOI] [PubMed] [Google Scholar]
- 21.Simon P, Aschauer C, Moidl R, Marx M, Keznickl FP, Eigenbauer E, Wolner E and Wollenek G. Growth of the pulmonary autograft after the Ross operation in childhood. Eur J Cardiothorac Surg. 2001;19:118–21. [DOI] [PubMed] [Google Scholar]
- 22.Khwaja S, Nigro JJ and Starnes VA. The Ross procedure is an ideal aortic valve replacement operation for the teen patient. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2005:173–5. [DOI] [PubMed] [Google Scholar]
- 23.Mazine A, Rocha RV, El-Hamamsy I, Ouzounian M, Yanagawa B, Bhatt DL, Verma S and Friedrich JO. Ross Procedure vs Mechanical Aortic Valve Replacement in Adults: A Systematic Review and Meta-analysis. JAMA Cardiol. 2018;3:978–987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zebele C, Chivasso P, Sedmakov C, Angelini G, Caputo M, Parry A and Stoica S. The Ross Operation in Children and Young Adults: 12-Year Results and Trends From the UK National Database. World J Pediatr Congenit Heart Surg. 2014;5:406–12. [DOI] [PubMed] [Google Scholar]
- 25.Carrel T The Ross procedure is under-used although long-term results show superior results to those obtained following mechanical aortic valve replacement. Ann Transl Med. 2016;4:S21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Charitos EI, Takkenberg JJ, Hanke T, Gorski A, Botha C, Franke U, Dodge-Khatami A, Hoerer J, Lange R, Moritz A, Ferrari-Kuehne K, Hetzer R, Huebler M, Bogers AJ, Stierle U, Sievers HH and Hemmer W. Reoperations on the pulmonary autograft and pulmonary homograft after the Ross procedure: An update on the German Dutch Ross Registry. J Thorac Cardiovasc Surg. 2012;144:813–21; discussion 821-3. [DOI] [PubMed] [Google Scholar]
- 27.Aicher D, Holz A, Feldner S, Kollner V and Schafers HJ. Quality of life after aortic valve surgery: replacement versus reconstruction. J Thorac Cardiovasc Surg. 2011;142:e19–24. [DOI] [PubMed] [Google Scholar]
- 28.Mokhles MM, Rizopoulos D, Andrinopoulou ER, Bekkers JA, Roos-Hesselink JW, Lesaffre E, Bogers AJ and Takkenberg JJ. Autograft and pulmonary allograft performance in the second post-operative decade after the Ross procedure: insights from the Rotterdam Prospective Cohort Study. Eur Heart J. 2012;33:2213–24. [DOI] [PubMed] [Google Scholar]
- 29.Brancaccio G, Polito A, Hoxha S, Gandolfo F, Giannico S, Amodeo A and Carotti A. The Ross procedure in patients aged less than 18 years: the midterm results. J Thorac Cardiovasc Surg. 2014;147:383–8. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.