Major Adverse Cardiac and Cerebrovascular Events After the Ross Procedure
A Report From the German-Dutch Ross Registry
Background—The purpose of the study is to report major cardiac and cerebrovascular events after the Ross procedure in the large adult and pediatric population of the German-Dutch Ross registry. These data could provide an additional basis for discussions among physicians and a source of information for patients.
Methods and Results—One thousand six hundred twenty patients (1420 adults; 1211 male; mean age, 39.2±16.2 years) underwent a Ross procedure between 1988 and 2008. Follow-up was performed on an annual basis (median, 6.2 years; 10 747 patient-years). Early and late mortality were 1.2% (n=19) and 3.6% (n=58; 0.54%/patient-year), respectively. Ninety-three patients underwent 99 reinterventions on the autograft (0.92%/patient-year); 78 reinterventions in 63 patients on the pulmonary conduit were performed (0.73%/patient-year). Freedom from autograft or pulmonary conduit reoperation was 98.2%, 95.1%, and 89% at 1, 5, and 10 years, respectively. Preoperative aortic regurgitation and the root replacement technique without surgical autograft reinforcement were associated with a greater hazard for autograft reoperation. Major internal or external bleeding occurred in 17 (0.15%/patient-year), and a total of 38 patients had composite end point of thrombosis, embolism, or bleeding (0.35%/patient-year). Late endocarditis with medical (n=16) or surgical treatment (n=29) was observed in 38 patients (0.38%/patient-year). Freedom from any valve-related event was 94.9% at 1 year, 90.7% at 5 years, and 82.5% at 10 years.
Conclusions—Although longer follow-up of patients who undergo Ross operation is needed, the present series confirms that the autograft procedure is a valid option to treat aortic valve disease in selected patients. The nonreinforced full root technique and preoperative aortic regurgitation are predictors for autograft failure and warrant further consideration.
Clinical Trial Registration—URL: http://www.clinicaltrials.gov. Unique identifier: NCT00708409.
The Ross operation is an acceptable alternative to conventional aortic valve replacement in selected patients. Advantages of this therapeutic option are the use of the patients’ own valve with favorable hemodynamic characteristics, avoidance of anticoagulant therapy, low thrombogenicity, and the potential to grow in children. Factors contributing to a limited acceptance are the complexity of the operation and the necessity of replacing both the aortic and pulmonary valves. In addition, little clinical long-term information is available regarding the durability of the autograft in the aortic position and the durability of pulmonary conduit substitute. After a renewal of interest in this procedure in the early 1990s, longer-term results are beginning to emerge, focusing mainly on the durability of the procedure and the valve substitutes. However, data on other major cardiac or cerebrovascular events in this patient population remain sparse, coming mainly from small single-center reports with limited follow-up durations. Using data from the large patient population of the German-Dutch Ross Registry, we sought to report on major cardiac or cerebrovascular events observed in 1620 Ross-operated patients over a follow-up of 10 747 patient-years. We believe that the data presented herein could provide a basis for the further judgment of this procedure and could assist physician–patient discussion about the risks, benefits, and expectations after the Ross procedure.
Patients and Methods
Study Population and Operative Data
The German-Dutch Ross Registry includes data from 12 departments of cardiac surgery since 1988 and the systematic prospective registry that was started in January 2002 (Clinical trial ID NCT 00708409). The study database was frozen in November 2009 and, for the purposes of the present report, all events until December 31, 2008 were analyzed. A total of 1620 patients were included in the database and their baseline and follow-up data were analyzed. The responsible surgeon at each center determined the surgical technique (subcoronary; root replacement with or without additional reinforcement procedures). A small subgroup of 30 patients undergoing operation with the root inclusion technique was included in the subcoronary group to create a group with all native root-preserving procedures. Details of the operative techniques and the reasons behind the inclusion of the root inclusion technique patients in the subcoronary group and the separate analysis of the root replacement technique patients with and without reinforcements have been reported previously.1–4 Informed consent was obtained from all patients; the study was approved by the local ethics committee and the authors had full access to and take full responsibility for the integrity of the data and the present article.
Clinical follow-up was performed at discharge and on a yearly basis. As a result of the different regional provenance of the patients and to support adherence to the program, complete clinical examinations from the referring cardiologists or general practitioners were also accepted. Major adverse cardiac and cerebrovascular events were reported according to the 2008 guidelines.5 All indications for autograft or pulmonary conduit reoperations were in accordance with the American College of Cardiology/American Heart Association guidelines.
The adult and the pediatric population (younger than 16 years) of the Registry are analyzed and reported separately. The cut-off point of 16 years was chosen because at this age the patients are regarded as adults and the technical aspects of the procedure are those of the adult population. Frequencies are given as absolute numbers and percentages. Continuous data are expressed as the mean±SD. Actuarial estimates of survival and freedom from morbid events are made using the Kaplan–Meier method. The survival time of a patient started at the time of surgery and ended at death (event) or at last follow-up (censoring). The long-term survival characteristics of the patient cohort were compared with survival probabilities of the general population obtained from German Life Tables 2005 (Statistisches Bundesamt, Wiesbaden Germany; http://www.destatis.de) and Dutch Life Tables 2008 (http://www.cbs.nl/L-NL/menu/home/default.htm). The contribution of every patient’s follow-up time within each age year is added to obtain the cumulative number of years at risk in each age year for the whole study population. Thereafter, the expected number of deaths (assuming their death rates were the same as the national figure) in each age year is found by multiplying the numbers of years at risk by the age- and gender-matched hazard provided by the life tables. The expected deaths for each age year are added to calculate the total expected deaths in the patient collective. The survival times were simulated based on the patients’ ages and genders but using the life table hazard rates in the simulation. If the simulated time until death exceeds the follow-up time for the corresponding patient, then the follow-up time is recorded and the value is regarded as censored. A Kaplan–Meier curve is fitted to the simulated times. This procedure was repeated enough times and an average curve was calculated and was compared to the Kaplan–Meier survival curve of the actual data. For the various operative technique groups, the instantaneous risk for reoperation was also calculated.
To identify predictive variables for shorter time to reoperation of the autograft or allograft, we performed univariate analyses using the Cox proportional hazard regression model. Multivariate Cox proportional hazard models were used to confirm whether significant (P≤0.10) univariate predictors persisted in the presence of preoperative variables. The presence of interaction and the proportionality of hazards assumption were checked in the final model. The following factors were analyzed as potential risk factors for death or autograft or allograft reoperation: age, year of surgery, gender, presence of comorbidities (diabetes, hypertension, renal failure, coronary artery disease, pulmonary disease, peripheral vascular disease), previous cardiac surgery, preoperative hemodynamics, aortic valve morphology, and homograft donor parameters (diameter, donor recipient age and blood group mismatch, cryopreservation).
Patients, characteristics and operative data are listed in Table 1. Follow-up completeness was 95%, with a mean follow-up of 6.6±4.2 years (range, 0–20.3 years) with 10 747 patient-years.
All-cause early (<30 days) mortality was 1.2% (n=19). All-cause late (>30 days) mortality was 58 (3.6%; 0.54%/patient-year): 34 (0.32%/patient-year) cardiac deaths, 22 noncardiac deaths (0.20%/patient-year), and 2 unknown. Valve-related mortality was 1.2% (20 patients; 0.19%/patient-year). Actuarial cumulative overall survival (including early mortality) for adults was 98.6% at 1 year (95% confidence interval (CI), 98.0%–99.2%), 96.9% at 5 years (95% CI, 95.2%–97.9%), and 94.7 at 10 years (95% CI, 93.1%–93.6%); for children it was 95.0% at 1 year (95% CI, 92.1%–97.9%), 94.4% at 5 years (95% CI, 91.3%–97.5%), and 92.5% at 10 years (95% CI, 88.4%–96.6%).
Actual vs Expected Death Rate
In this comparison, all patients with follow-up >30 days were included. Observed fatal events were compared with the expected deaths in the age- and gender-matched general German and Dutch populations (Table 2). The Kaplan–Meier actuarial estimates and the estimates of expected survival for the adult and the pediatric populations are displayed in Figures 1 and 2⇓, respectively. In the univariate Cox proportional hazard model, younger age among children (hazard ratio [HR] per year, 0.86; 95% CI, 0.80–0.93; P<0.0001), older age among adults (HR per year, 1.04; 95% CI, 1.02–1.06; P<0.0001), and presence of preoperative risk factors (HR, 1.38; 95% CI, 1.05–1.83; P=0.023) were associated with increased risk for late mortality.
Reoperation Including Endocarditis
One hundred sixty reoperations on Ross-related valves (pulmonary autograft, pulmonary conduit) were required in 137 patients (8.5%; 1.49%/patient-year); the time interval between the initial procedure and the first reoperation was 5.7±4.7 years (range, 0.0–16.3 years; median, 4.9 years). Ninety-three patients underwent 99 reinterventions on the autograft (5.7%; (linearized occurrence rates [LOR], 0.92%/patient-year). Seventy-eight interventions in 63 patients were performed on the pulmonary conduit (3.9%; LOR, 0.72%/patient-year). Of these, 17 procedures with simultaneous autograft and pulmonary conduit interventions were performed in 17 patients.
Table 3 summarizes the indications for 99 autograft reinterventions (including 17 cases of interventions on the autograft and pulmonary conduit). The autograft reoperation procedures (time to first reoperation) were performed 0 to 16.3 years (mean, 6.6±4.9 years; median, 6.8 years) after the initial Ross operation. In 47 patients a mechanical valve was used, in 14 a bioprosthesis was used, and in 6 a homograft was implanted; in 23 patients, an autograft reconstruction was performed. Freedom from reoperation on the autograft and the instantaneous risk of reoperation in children are displayed in Figure 3; the estimates in adults for the different surgical techniques are shown in Figure 4. The univariate Cox proportional hazard model showed evidence that in age group 16 to 40 years (HR, 1.78; 95% CI, 1.12–2.83; P=0.015, in comparison to age group 41 to 60 years), root replacement without surgical reinforcement (HR, 1.82; 95% CI, 1.12–2.95; P=0.015, in comparison to the subcoronary technique) and preoperative aortic regurgitation (HR, 2.07; 95% CI, 1.21–3.55; P=0.0079, in comparison to aortic stenosis as predominant preoperative hemodynamics) were associated with shorter times to reoperation. The multivariate model is displayed in Table 4.
Indications for the 78 reinterventions on the pulmonary conduit (including 17 patients with replacement of the autograft and pulmonary conduit) are presented in Table 3. The pulmonary conduit reoperation procedures (time to first reoperation) were performed from 0 to 16.3 years (mean, 4.7±4.4 years; median, 2.7 years) after the initial Ross operation. No percutaneous procedures were performed. Freedom from reoperation on the pulmonary conduit and the instantaneous risk of reoperation are displayed in Figure 5. The univariate Cox proportional hazard model showed evidence that the younger patient age (HR, 0.97 per year; 95% CI, 0.95–0.98; P<0.0001), younger donor age (HR, 0.96 per year; 95% CI, 0.94–0.98; P=0.0001), absolute age difference between donor and patient (HR per year, 1.06; 95% CI, 1.002–1.048; P=0.032), smaller allograft diameters (HR per mm, 0.87; 95% CI, 0.800–0.95; P=0.0012) were associated with shorter times to allograft reoperation. The multivariable model is displayed in Table 4. Freedom from autograft and pulmonary conduit reoperation and the instantaneous risk of reoperation are displayed in Figure 6. All patients survived the reoperation on Ross-related valves and were alive at the date of the last follow-up inquiry.
Four early endocarditis occurred. Late endocarditis with medical (n=16) or surgical treatment (n=29) was observed in 38 patients (2.3%; LOR, 0.38%/patient-year). Overall, 18 autograft reoperations attributable to endocarditis were performed (1.1%; LOR, 0.18%/patient-year). Pulmonary conduit endocarditis with medical (n=11) or surgical treatment (n=15) occurred in 24 patients (1.5%; LOR, 0.26%/patient-year).
Thrombosis and Bleeding
Valve-related thrombotic and thromboembolic events occurred in 21 patients (Table 3). Major internal or external bleeding occurred in 17 patients (1.05%; LOR, 0.15%/patient-year). A total of 38 patients experienced the composite end point of thrombosis, embolism, or bleeding (2.3%; LOR, 0.35%/patient-year).
Major Adverse Valve-Related Events
Overall freedom from any major valve-related event (all valve-related mortality; valve-related morbidity: structural and nonstructural valve dysfunction with the need of reoperation, thrombosis, bleeding, embolism, neurological events, endocarditis, and the need for pacemaker implantation within 14 days after operation) was 94.9% at 1 year (95% CI, 95.9%–93.9%), 90.7% at 5 years (95% CI, 92.3%–89.1%), and 82.5% at 10 years (95% CI, 85.0%–80.0%).
The prospective German-Dutch Ross Registry with a considerable number of patients, good midterm follow-up completeness, and follow-up time of >10 000 patient-years offers the opportunity to address key question of major cardiac and cerebrovascular events.
Survival in the Pediatric Population
In most case-series studies addressing the survival of pediatric patients, encouraging results have been published. In the series published by Elkins et al,6 actuarial survival at 8 years was 97%,7 and their updated series showed an actuarial survival of 92% at 12 years.8 Kouchoukos et al9 showed a 10-year actuarial survival of 96%, and in a systematic review the pooled outcome estimate for late mortality in pediatric series was 0.62%/patient-year.10 The reported numbers are similar to the actuarial survival in the registry with 94.4% at 5 years and 92.5% at 13 years. After excluding the early mortality, a comparison to the expected survival based on national hazard rates revealed significant differences, with lower observed patient survival mainly attributable to fatalities occurring within the first year after the procedure. These results are inferior to the results in adult patients. In contrast to adult Ross patients, pediatric patients are more often critically ill, and almost all undergo the procedure after failing previous interventions for congenital aortic stenosis. Furthermore, these patients are more likely to present with complex left ventricular disease and other associated anomalies.
Survival in the Adult Population
In the adult population of the Registry, the observed 94.9% 10-year actuarial survival estimate is comparable to the estimates reported in other large series8 and to the pooled outcome survival estimates of a recently published meta-analysis (0.64%/patient-year10). After the initial survival decrease associated with in-hospital mortality, the observed survival parallels the expected survival calculated from the national hazard rates. Excluding all early fatalities (within 30 days of the initial operation), no significant differences in survival could be observed between the adult Ross patients and the normal population, a fact that underlines the overall good prognosis of the adult Ross-operated patients, after overcoming the early postoperative hazard.
Reoperations on the Autograft Excluding Endocarditis
The estimates of structural and nonstructural autograft deterioration (LOR) with the need for reoperation in pediatric patients is reported to range between 0.24% and 2.82% per patient-year, with a pooled mean of 1.38% per patient-year.10 In contrast, in the present pediatric series reoperations on the autograft are rare within the first decade with an actuarial freedom from reoperation of 100% at 5 years and 97.9% at 10 years, respectively, reflecting a LOR of 0.40% per patient-year. However, graft failure may become apparent thereafter. This is in accordance with an actuarial estimate of 93.8% at 13 years in our series, which is also supported by the data of the series published by Elkins et al.8,11 Aortic regurgitation was the leading cause of reoperation based on nonstructural valve failure with root dilatation; other indications were rare in this pediatric series (no structural valve deterioration, infective endocarditis in 1). The dilatation of the autograft must be prevented to improve the durability of the neoaortic root. In adults and adolescents, this could be effectively achieved by applying reinforcement techniques or by the use of the subcoronary implantation technique.2 The routine use of reinforcement techniques or other implantation techniques failed to improve autograft durability12 or was not applicable, especially in small children. Additionally, the growth potential of the neoaortic root is one of the major advantages of the Ross procedure, which will be unfavorably influenced by all available stabilizing measures.7,13 This limitation and the necessity of reoperation attributable to outgrowth are still an unsolved issue in the pediatric patient undergoing Ross procedure.
In adult patients, the pooled estimates of structural and nonstructural autograft deterioration with the need for reoperation was reported to range between 0.15% and 1.90% per patient-year, with a pooled mean of 0.78% per patient-year.10 These estimates are congruent with the present actuarial freedom from reoperation probability of 97.6% at 5 years and 93.7% at 10 years, reflecting a linearized rate of 0.74% per patient-year. The actuarial data are convincing up to 12 years; beyond this cut-off point, no robust data are available. Because the definition of autograft failure differs in numerous reports in literature, comparisons with other series are limited8 and thus we restricted our estimates on the hard end point “reoperation.” Aortic regurgitation was the leading cause of reoperation in 90% of all reoperations. The mechanisms of graft failure differs between the surgical implantation techniques; autograft procedures as root replacements without reinforcement interventions are prone to nonstructural valve deterioration caused by root dilatation in 41 of 47 (LOR, 1.82%/patient-year), whereas root replacement with reinforcement (LOR, 0.32%/patient-year) or the subcoronary technique (LOR, 0.48%/patient-year) revealed dilated root with the need for reoperation only in 2 of 8 and 0, respectively. The practical benefit of stabilization measures at the annular level of the neoaortic root has been reported previously based on the large database of the registry.2 The leading cause of autograft valve failure with the need for reoperation in the subcoronary group is structural valve deterioration (80% of all reoperations), mainly as cusp prolapse (69% of all structural valve deteriorations). This problem remains a surgical challenge in the subcoronary implantation technique. Additionally, our data suggest that patients with primary aortic regurgitation have an increased risk for autograft failure and might have an increased risk for reoperation. This potential limitation of the procedure is apparent in the pediatric and adult populations and has been reported by various groups.8
Reoperation on the Pulmonary Conduit Excluding Endocarditis
For the combined end point structural and nonstructural degeneration of the pulmonary conduit with the need for reoperation in pediatric patients, a LOR between 0.40% and 4.9% per patient-year (pooled mean, 1.6%/ patient-year) was estimated.10 In accordance with this meta-analysis, we also observed an apparently higher reoperation rate of the pulmonary conduit compared to autograft reoperations. The present series revealed actuarial estimates of freedom from conduit reoperation of 90.6% at 5 years and 87.1% at 10 years, respectively, which correspond to a LOR of 1.32% per patient-year. As in other series, the predominant indication for reoperation of the pulmonary conduit was stenosis (57%). Pure regurgitation is uncommon (12%). Several predictors of conduit failure have been reported;14–19 however, most of them cannot be taken into clinical consideration because of the limited availability of donor grafts.
In the adult population, the LOR estimates of structural and nonstructural pulmonary conduit deterioration with the need for reoperation has been reported to range between 0.12% and 1.27% per patient-year (pooled mean, 0.55%/patient-year).10 The present actuarial freedom from pulmonary conduit reoperation probability of 99.0% at 5 years and 97.0% at 10 years results in a linearized rate of 0.30% per patient-year. In most cases, a pulmonary allograft was the first choice for reconstruction of the right ventricular outflow tract. Similar to the pediatric population, predictors of pulmonary conduit failure in multivariate analyses could hardly be considered in the choice of the graft.
Autograft and Pulmonary Conduit Endocarditis
The observed numbers of autograft endocarditis in the present series are low, with a LOR in the pediatric patients of 0.07% per patient-year (compared to a pooled mean estimate of 0.15%/patient-year10); pediatric pulmonary conduit endocarditis was seen with LOR of 0.40% per patient-year (pooled mean estimate, 0.26%/patient-year). In our adult series, autograft and pulmonary conduit endocarditis are similar to that reported by Takkenberg et al.10 Although the absolute numbers are low, it must be emphasized that 15% of all autograft reoperations and 20% of all pulmonary conduit reoperations were caused by infective endocarditis. All cases of endocarditis during follow-up were unrelated to active aortic valve endocarditis before the procedure supporting the use of the Ross operation for treating patients with infective endocarditis of the aortic valve. In most cases, precipitating factors (hematologic disorders, corticosteroid therapy, diabetes mellitus, drug or alcohol abuse) could be identified.
Thrombosis, Thromboembolism, and Bleeding
Valve thrombosis, thromboembolism, and bleeding events are uncommon in patients who underwent the Ross procedure. The LOR of 0.18% per year in the pediatric series and 0.35% per patient-year in the adults (normal adult population, 0.13%/patient-year) are identical with the reported pooled data.10 Especially in the adult series, the composite event related more to other cardiac and extracardiac factors than to the valve itself (eg, embolism in atrial fibrillation, anticoagulation-related bleedings).
The German-Dutch Ross Registry is a nonrandomized registry, prospectively recruiting patients since 2002 and all historical Ross patients for the period 1988 to 2002. The operative volume and experience varies across the 13 centers. A small subgroup of patients undergoing operation with the root inclusion technique, as well as patients undergoing operation with a modification of the subcoronary technique (preservation of the non coronary sinus), are included in the subcoronary root population. Analysis showed no difference regarding major outcomes within these techniques. The various operative groups (children, subcoronary technique, root replacement technique with additional reinforcement procedures, root replacement without additional reinforcement procedures) have different follow-up durations. In the reported LOR, reoperations in adult patients who underwent operation with the root replacement technique without autograft reinforcement have been included, although this technique is being abandoned among the participating centers.
The present series confirms that the autograft procedure is a valuable option to treat aortic valve disease in children, adolescents, and young adults. Preoperative aortic regurgitation and the nonreinforced full root technique are predictors for autograft reoperation and require special consideration.
The authors thank Mrs. Katrin Meyer for her excellent data management and secretarial support at the Registry Site in the Department of Cardiac and Thoracic Vascular Surgery, University Clinics Schleswig-Holstein, Campus Lübeck.
R. Lange is a Consultant/Advisory Board member (<$10 000) for Medtronic, Edwards, and MDS. G. Ziemer is an expert witness in medical malpractice case (one-quarter year, <$10 000). H.-H. Sievers received honoraria (<$10 000) from the Ross meeting 2009 (Atlanta, GA) and Cryolife.
Presented at the 2009 American Heart Association meeting in Orlando, Fla, November 14–18, 2009.
Charitos EI, Hanke T, Stierle U, Robinson DR, Bogers AJ, Hemmer W, Bechtel M, Misfeld M, Gorski A, Boehm JO, Rein JG, Botha CA, Lange R, Hoerer J, Moritz A, Wahlers T, Franke UF, Breuer M, Ferrari-Kuehne K, Hetzer R, Huebler M, Ziemer G, Takkenberg JJ, Sievers HH. Autograft reinforcement to preserve autograft function after the ross procedure: a report from the german-dutch ross registry. Circulation. 2009; 120: S146–S154.
Hanke T, Stierle U, Boehm JO, Botha CA, Matthias Bechtel JF, Erasmi A, Misfeld M, Hemmer W, Rein JG, Robinson DR, Lange R, Horer J, Moritz A, Ozaslan F, Wahlers T, Franke UF, Hetzer R, Hubler M, Ziemer G, Graf B, Ross DN, Sievers HH. Autograft regurgitation and aortic root dimensions after the Ross procedure: the German Ross Registry experience. Circulation. 2007; 116: I251–258.
Akins CW, Miller DC, Turina MI, Kouchoukos NT, Blackstone EH, Grunkemeier GL, Takkenberg JJ, David TE, Butchart EG, Adams DH, Shahian DM, Hagl S, Mayer JE, Lytle BW. Guidelines for reporting mortality and morbidity after cardiac valve interventions. J Thorac Cardiovasc Surg. 2008; 135: 732–738.
Takkenberg JJ, Klieverik LM, Schoof PH, van Suylen RJ, van Herwerden LA, Zondervan PE, Roos-Hesselink JW, Eijkemans MJ, Yacoub MH, Bogers AJ. The Ross procedure: a systematic review and meta-analysis. Circulation. 2009; 119: 222–228.
Meyns B, Jashari R, Gewillig M, Mertens L, Komarek A, Lesaffre E, Budts W, Daenen W. Factors influencing the survival of cryopreserved homografts. The second homograft performs as well as the first. Eur J Cardiothorac Surg. 2005; 28: 211–216.