Optimized outcome of the Ross procedure in children: single-centre experience^†

Abstract

OBJECTIVES

The Ross procedure with autograft reinforcement has been proposed as a strategy to prevent autograft failure in adults, but outcome data in children during somatic growth remain limited. We investigated long-term outcomes following an individualized autograft reinforcement protocol to evaluate survival and reintervention rates.

METHODS

Between January 1995 and December 2022, 233 patients <18 years [median age: 7 (1–13) years] underwent the Ross procedure, including 60 infants (26%). Most frequently free-root autograft implantation without reinforcement was performed (n = 156, 67%). Autograft reinforcement was applied in 77 patients (33%) using either subcoronary implantation (n = 65, 28%) or external prosthetic support (n = 12, 5%). Kaplan–Meier survival estimates were used for survival and reintervention analyses. Risk factors for reintervention were identified by Cox proportional hazards regression.

RESULTS

Reinforcement was associated with improved survival (5-year survival rates of 97.1% vs 87.0%, 10-year survival rates of 97.1% versus 86.99%, P = 0.017). No differences in autograft reintervention between the groups were found (1-year rates of 100% vs 99.4%, 5-year rates of 100%, P = 0.4852). Right ventricle–pulmonary artery (RV-PA) reintervention-free survival at 5 years was higher for homografts compared to xenografts (96.9% vs 79.4%, P < 0.001).

CONCLUSIONS

The Ross procedure in children demonstrated excellent long-term outcomes with low autograft reintervention rates in both groups. Reinforcement was associated with improved long-term survival while autograft-related reinterventions did not differ significantly between groups. Older age at Ross and homograft use correlated with lower RV-PA reintervention risk. Multicentre evaluation of reinforcement techniques is required to assess the outcome differences observed in this single-centre experience.

Open in new tabDownload slide

INTRODUCTION

The Ross procedure, introduced by Donald Ross in 1967, remains a preferred method for aortic valve replacement, particularly in patients undergoing surgery during somatic growth [1]. By replacing the diseased aortic valve with the patient’s own pulmonary valve (autograft), the Ross procedure offers several advantages, such as excellent haemodynamics, the potential for autograft growth and the avoidance of lifelong anticoagulation [2]. However, the long-term viability of the autograft and the likelihood of reintervention have prompted a critical re-evaluation of the procedure [3–6].

Significant challenges arise from autograft dilation and valvular regurgitation over time, which can compromise the durability of the Ross procedure. To address these issues, strategies such as autograft reinforcement—including subcoronary implantation and modified external autograft support—have been developed [7–9]. These approaches aim to enhance the structural integrity of the autograft and reduce the risk of dilation, thereby improving its long-term performance. Despite encouraging results, the efficacy of autograft reinforcement techniques in reducing reintervention rates and enhancing survival remains a subject of ongoing investigation [10]. In this study, the term “reinforcement” is employed to describe the external supporting techniques in general, including the autologous support in subcoronary technique as well as the prosthetic support by synthetic materials, reflecting widely accepted terminology.

This study aims to evaluate the long-term outcomes of the Ross procedure, focusing on the role of reinforcement techniques in improving survival and reducing reintervention rates. The analysis also examines the impact of right ventricle–pulmonary artery (RV-PA) conduit choice and patient age on postoperative outcomes, which remain under ongoing debate [11, 12]. By leveraging extended follow-up data, the study aims to address gaps in the literature and contribute to the optimization of surgical strategies and patient care for individuals undergoing the Ross procedure.

PATIENTS AND METHODS

Study design and patient selection

This retrospective, single-centre cohort study included patients under 18 years of age who underwent the Ross procedure between January 1995 and December 2022. Inclusion criteria encompassed all patients who underwent autograft implantation, either with reinforcement (i.e. subcoronary implantation or prosthetic support techniques) or without any reinforcement (free-root implantation). Age-based subgroups included infants (<1 year), children (1–12 years) and adolescents (12–17 years).

Ethical considerations

Data collection included comprehensive preoperative, procedural and follow-up information sourced from institutional and referring cardiologist’s medical records. The study was conducted in accordance with the principles of the Declaration of Helsinki and received ethical approval from the institutional review board (Charité—Universitätsmedizin Berlin, EA2/080/20). The need for patient consent was waived, due to the use of retrospective data. In case of follow-up updates by contacting the referring physicians and/or patients, written informed consent from all participants or parents/guardians if participants are underage was acquired.

Surgical techniques

Before 2012, the Ross procedure was performed by 4 surgeons almost exclusively using the free-root technique without autograft reinforcement irrespective of age or individual risk factors. After 2012, due to a change in institutional policy as well as in surgical personnel, patients underwent the Ross procedure using either an autograft reinforcement, primarily based on surgeon preference and perceived risk of autograft dilation. Subsequently, an individualized comprehensive approach was employed; infants and children with low-risk factors for autograft dilation up to the age of 2 years received preferably a standard free-root technique without external reinforcement. Older patients and those with a higher risk for autograft failure—such as dilated aortic annulus, predominant aortic regurgitation, or bicuspid aortic valve disease—were screened for subcoronary implantation or external polytetrafluoroethylene support.

Subcoronary implantation provided structural support by securing the autograft within the native aortic root [8]. Alternatively, in selected patients with adult-sized aortic annulus undergoing free-root implantation, external reinforcement was achieved using a thin polytetrafluoroethylene membrane, which was tailored appropriately and sawn into the aortic annulus during the autograft implantation. The membrane was cut out to spare the coronary ostia and was incorporated into the distal aortic anastomosis to mitigate risks of autograft dilation and regurgitation [13].

RV-PA conduit reconstruction was performed preferably using cryopreserved homografts. As an alternative, mainly bovine jugular vein conduits were employed. From 2018 onward, decellularized fresh pulmonary homografts were selectively introduced for adolescent patients to reflect improvements in biocompatibility and reduced immunogenic response compared to cryopreserved homografts. The transition to routinely employing reinforced Ross techniques marked an institutional effort to optimize autograft durability and reduce reintervention rates. Throughout the study period, all surgeries were performed by a consistent team of 3 congenital heart surgeons, ensuring uniformity in technique application and outcome evaluation; however, additional surgeons operated prior to 2012, introducing a potential era effect.

Follow-up and end-points

Follow-up assessments included clinical outcomes, echocardiographic evaluations and survival analyses. The primary end-points were overall survival and freedom from autograft reintervention. Secondary end-points included RV-PA conduit reintervention rates, hospital readmissions and major adverse events such as endocarditis, heart failure and thromboembolism. Echocardiographic data, including aortic valve gradients and pulmonary valve function, were collected preoperatively, immediately postoperatively and during follow-up. Patients with incomplete follow-up data as of 2022 were contacted for additional information, supplemented by consultations with referring cardiologists to ensure comprehensive data acquisition.

Statistical analysis

Descriptive statistics are presented as median with interquartile range (Q1–Q3) for continuous variables and as absolute numbers with percentages for categorical variables. Between-group comparisons were conducted using unadjusted Mann–Whitney U tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables to describe baseline data. We performed a Wilcoxon signed-rank test to compare paired data. Patients were grouped based on the type of autograft implantation technique: reinforced (subcoronary and free-root implantation with external reinforcement) versus non-reinforced techniques. Reintervention-free survival rates and freedom from autograft and RV-PA conduit reinterventions at 1, 5, 10 and 15 years were estimated using Kaplan–Meier survival analysis, and subgroup comparisons were made using the log-rank test. The impact of RV-PA conduit type and patient age on reintervention rates was assessed using a Cox proportional hazards model with the Breslow method for ties. Models including multiple covariates were adjusted for multiple comparisons to mitigate the risk of type I error. A P-value of <0.05 was considered indicative of significance. Statistical analysis was performed using IBM SPSS Statistics version 29.0 (IBM Corp., Armonk, NY, USA) and Stata/MP 18.0 (StataCorp LLC, College Station, TX, USA).

RESULTS

Baseline characteristics

A total of 233 patients who underwent the Ross procedure were included in the study, with a median age of 7 (1–13) years, a median follow-up duration of 5 years (interquartile range 1.25–14.25 years) and a mean follow-up duration of 8.2 years (±7.6). Among these, 77 patients underwent an autograft reinforcement (subcoronary implantation or external prosthetic support), while 156 underwent non-reinforced free-root implantation. Patients in the reinforcement group had a higher incidence of bicuspid or unicuspid valve anatomy (70.1% vs 43.6%; P < 0.001), whereas the non-reinforced group had a significantly higher left ventricular outflow tract maximum gradient [median 30.1 (15.1–91.5) vs 7.9 mmHg (3.9–11.3); P = 0.002]. The baseline characteristics are detailed in Table 1.

Survival

The overall mortality rate for the cohort was 9.9% (23/233), with 52.2% of deaths (12/23) occurring within 30 days of surgery. Kaplan–Meier survival analysis revealed a higher survival probability in the reinforced group compared to the non-reinforced group (log-rank test, P = 0.017). The estimated 1-year survival rates were 97.1% [95% confidence interval (CI) 88.9–99.3%] for the reinforced group and 92.27% (95% CI 86.79–95.54%) for the non-reinforced group. The 5-year survival rates were 97.1% and 86.99%, respectively (Fig. 1). The median survival was not reached in either group, indicating prolonged survival in both cohorts. At 10 years, survival probabilities were 97.1% for the reinforced group and 86.99% for the non-reinforced group. At 15 years, survival probabilities remained at 97.1% in the reinforced group but declined to 83.38% in the non-reinforced group. The reinforced group experienced 2 deaths, whereas the non-reinforced group had 12 deaths, with a higher proportion of censored cases in the former. Stratification by age revealed a higher hazard in infants (<1 year) compared to older children (1–17 years; log-rank P < 0.001) (Fig. 2).

Reinterventions

Reintervention (including any surgery or heart catheterization) was required in 43.8% of the cohort (102/233), with significantly fewer interventions in the reinforcement group compared to the non-reinforced group (29.9% vs 50.6%; P < 0.001). In the sub-cohort of infants (n = 60), no autograft reinterventions were observed during the follow-up period, regardless of reinforcement status. The median follow-up time for this group was 3.8 years (range: 0.16–10.4). In contrast, patients older than 1 year had a median follow-up time of 6.5 (1.5–14.7) years.

Autograft reinterventions

Kaplan–Meier analysis showed no substantial difference in freedom from autograft reintervention between the reinforced and non-reinforced groups (log-rank test, P = 0.4852). The estimated 1-year rates were 100% for infants and 99.4% (95% CI 95.8–99.9%) for older children and adolescents. The log-rank test for older patients yielded a chi-square value of 1.05 (P = 0.3053), indicating no significant difference between the groups. The 5-year rates were 100% for infants and 94.5% (95% CI: 88.7–97.4%) for older patients. The 10-year rates were 89.5% (95% CI: 81.3–94.2%) and the 15-year rates were 83.2% (95% CI: 74.1–90.1%) for older patients. The reinforced group experienced 3 reinterventions, whereas the non-reinforced group had 15. The log-rank test yielded a chi-square value of 0.49, indicating no statistically significant difference between groups.

Cox proportional hazard analysis for freedom from autograft reintervention including age group, reinforcement, bicuspid aortic valve and sex as covariates was attempted; however, due to the limited number of events, no statistically reasonable results were acquired.

RV-PA reinterventions

Kaplan–Meier analysis confirmed superior freedom from reintervention in the homograft group compared to the xenograft group (log-rank test, P < 0.001), with 1-year rates of 100% vs 99.3%, 5-year rates of 96.9% vs 79.4%, 10-year rates of 76.4% vs 50.1%, and 15-year rates of 70.5% vs 30.2% (Fig. 3).

Cox proportional hazard analysis for freedom from RV-PA conduit reintervention included age group, homograft versus xenograft implantation, and sex as covariates. Homograft implantation was associated with a lower hazard for reintervention [hazard ratio (HR) = 0.216, 95% CI 0.099–0.471, P < 0.001, Fig. 4]. Older age (12–17 years) was also linked to a reduced hazard (HR = 0.344, 95% CI 0.150–0.786, P = 0.011), while sex showed no association (HR = 0.707, 95% CI 0.388–1.288, P = 0.257).

Combined end-point

There was higher autograft-reintervention-free survival in the reinforced group compared to the non-reinforced group (log-rank test, P = 0.0184). The estimated 1-year rates were 98.7% (95% CI 91.1–99.8%) for the reinforced group and 92.3% (95% CI 86.9–95.6%) for the non-reinforced group. The 5-year rates were 86.6% and 67.1%, respectively. The median was not reached in either group, but the 10-year probabilities remained at 86.6% and 43.9%, respectively. The reinforced group experienced 5 events, whereas the non-reinforced group had 36 events. The log-rank test yielded a chi² value of 5.55, supporting a difference in outcomes between groups.

Postoperative outcomes

The incidence of seizures was notably higher in the reinforcement group (9.1% vs 3.2%; P = 0.056), whereas thromboembolism occurred exclusively in the non-reinforced group (3.8%; P = 0.081). Other major complications, including endocarditis (5.2%) and heart failure (3.9%), were comparable between groups. Hospital readmissions occurred in 40.8% of patients, with no significant difference between groups (P = 0.126). Detailed postoperative outcomes are provided in Table 2.

Echocardiographic follow-up

In addition to censoring data available for the entire cohort, echocardiographic follow-up was obtained for 169 of 233 patients (72.5%). Echocardiographic follow-up was not mandatory for study inclusion, as the study focused on clinical end-points. The last available echocardiographic follow-up was performed at a median of 42 months (4–140 months) after the initial surgery (Table 3).

The mean transvalvular pressure gradient decreased from 28.2 mmHg (16.95–40 mmHg) before surgery to 3.1 mmHg (2–4.7 mmHg) at the last follow-up (P < 0.001). Aortic regurgitation severity was improved at the last echo follow-up. Before surgery, 7.4% of patients had no regurgitation, 28.8% had mild, 37.2% had moderate and 26.5% had severe regurgitation. At the last follow-up, 26.5% had no regurgitation, 68.8% had mild, 4.2% had moderate and 0.47% had severe regurgitation (P = 0.774). In the direct comparison between reinforced and non-reinforced autografts, statistical differences were observed in postoperative maximal (P = 0.034) and mean transvalvular gradients (P = 0.039). However, with 4.7 mmHg upper quartile mean gradients remained within clinically acceptable ranges, with no systematic occurrence of higher-grade restenosis. Full freedom from any aortic stenosis was achieved in 97.3% of patients in the non-reinforced group and 91.1% in the reinforcement group at the last follow-up, whereas mild aortic restenosis developed in 7.1% of reinforced vs 2.7% of non-reinforced patients (P = 0.136), moderate aortic restenosis was found in 1 patient (1.8%) of the reinforced group. Conversely, RV-PA conduit gradients increased over time in both groups, leading to the mentioned increase in RV-PA conduit reinterventions over time in xenograft recipients.

DISCUSSION

The long-term outcomes of the Ross procedure in children and young adults, based on our single-centre experience, confirm its feasibility and efficacy, including the recent application of reinforcement techniques [14–16]. This study demonstrates favourable outcomes with low reintervention rates in both groups, while patients who underwent autograft reinforcement—either by subcoronary implantation or autograft wrapping—exhibited superior survival rates. Despite the technically somehow more challenging reinforcement technique, the outcome was comparable to the standard technique of autograft implantation. In our experience, especially in patients with known risk factors for autograft failure, autograft-related morbidity and mortality seem to be successfully mitigated by reinforcement. These findings underscore the importance of tailored surgical strategies in optimizing autograft durability and valve function throughout somatic growth and beyond [17].

In a recent MRI-based long-term analysis assessing the Sinus of Valsalva dilatation following the Ross procedure, reinforcement techniques demonstrated superiority, suggesting long-term benefits with reduced mortality risk and no increase in reintervention rates [18].

Our findings align with previous studies suggesting that structural support can mitigate the risks associated with somatic growth and dilation of the autograft [19–21]. However, the era effect—notably the preferred use of reinforcement techniques since 2012 at our institution—should be acknowledged as a potential confounding factor. Even though during the most recent study period, the preference for subcoronary implantation was apparent, both techniques were still available and employed in a parallel fashion according to the individual risk factors, again with no increase in risks for death, reintervention or major complications.

This study also highlights the critical role of RV-PA conduit choice in determining long-term outcomes. Patients receiving homografts demonstrated better reintervention-free survival compared to those with xenografts. Furthermore, age emerged as a significant determinant, with older patients at surgery exhibiting a lower risk of RV-PA conduit reintervention. These observations suggest that both biological maturity and conduit material play pivotal roles in optimizing outcomes. Our limited experience with the new generation of fresh decellularized pulmonary homografts, especially for adolescent patients showed promising early results, which are out of the scope of this report.

The incidence of aortic restenosis was predominantly mild and did not require reoperation or raise autograft reintervention rates in either group. This finding may reflect altered root geometry or residual leaflet stress but does not detract from the protective benefits of reinforcement against autograft dilation. Differences in seizure incidence and thromboembolism were not statistically significant (with raw P-values reported). However, these findings may reflect multifactorial perioperative influences, warranting further investigation into neuroprotective strategies. Notably, thromboembolism occurred exclusively in the non-reinforced group (3.8%), possibly indicating haemodynamic or anticoagulation disparities.

In summary, this retrospective analysis of a large single-centre cohort demonstrated that autograft reinforcement by subcoronary implantation or autograft wrapping in children and adolescents is feasible and effective. Moreover, its association with improved survival and comparable reintervention rates, without significant differences in postoperative outcomes is promising. Due to the shorter follow-up period after the recently implemented reinforcement strategy, the possible long-term benefits of our approach are to be investigated. The use of homografts for RV-PA conduit reconstruction was linked to superior long-term outcomes compared to xenografts. Patients aged 12–17 years exhibited a lower risk for reintervention. These results support the individualized use of autograft reinforcement and homografts to optimize the long-term efficacy of the Ross procedure during somatic growth.

Limitations

Limitations of this study include its retrospective, single-centre design and the intrinsic selection bias of non-randomized analyses, which may have assigned patients with particular risk profiles to reinforced or non-reinforced strategies. Because reinforced techniques were adopted predominantly after 2012, improvements in perioperative care or surgeon expertise during this period could partly account for the observed benefits in the reinforced group. Although mild-moderate aortic stenosis and a higher incidence of seizures were noted among reinforced patients, these findings did not necessitate reoperation and warrant further exploration to clarify underlying mechanisms.

Moreover, the lack of multicentric data further limits the generalizability of these findings. The heterogeneity of practices across centres, with differing surgical techniques, patient management protocols, and expertise levels, poses certain challenges to standardization and data pooling. Additionally, only a limited number of centres worldwide have substantial experience with both reinforced and non-reinforced techniques, making multicentric comparisons difficult. Centre-based effects, such as variations in surgeon experience and institutional protocols, may further confound the interpretation of outcomes. Moreover, subgroup analyses for homograft types and more detailed stratifications by age were not feasible, leaving gaps in the understanding of specific outcomes.

Future investigations should focus on prospective, multicentric studies with uniform follow-up protocols to validate these findings. Detailed mechanistic studies are also essential to unravel the biological underpinnings of the observed outcomes, particularly the interplay between reinforcement techniques, age, and valve durability.

CONCLUSION

The Ross procedure provides durable aortic valve replacement in paediatric population, with autograft reinforcement improving long-term outcomes without significantly altering reintervention rates. Regarding RV-PA conduit performance patient-specific factors, conduit selection and age at surgery seem to be pivotal in optimizing outcomes, as homografts outperformed xenografts, and older age correlated with fewer reinterventions. These findings contribute to the evolving landscape of paediatric and adolescent valve surgery, indicating room for improvement of both autograft and RV-PA conduit outcomes. Additional data from multicentre registries or randomized comparisons will help refine patient-centred approaches for this complex procedure.

FUNDING

None declared.

Conflict of interest: None declared.

DATA AVAILABILITY

Derived data supporting the findings of the study are available upon reasonable request from the corresponding author.

Author contributions

Peter Murin: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Supervision; Writing—original draft. Julia Gaal: Data curation; Investigation; Validation; Visualization; Writing—review & editing. Robin Stenzel: Data curation; Investigation; Validation. Viktoria Weixler: Data curation; Investigation; Validation. Olga Romanchenko: Data curation; Investigation; Validation. Raphael Seiler: Data curation; Investigation; Validation. Stanislav Ovroutski: Data curation; Supervision; Validation; Writing—review & editing. Felix Berger: Supervision; Writing—review & editing. Mi-Young Cho: Supervision; Writing—review & editing. Joachim Photiadis: Supervision; Writing—review & editing. Marcus Kelm: Conceptualization; Data curation; Formal analysis; Methodology; Software; Writing—original draft

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks John Santosh Murala, Martin M. Kostolny and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• CI

  Confidence interval

 
• HR

  Hazard ratio

 
• RV-PA

  Right ventricle–pulmonary artery

Author notes