International Journal of Cardiology 175 (2014) 455–463

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Durability of large diameter right ventricular outflow tract conduits in adults with congenital heart disease Jonathan Buber a,⁎, Gabriele Egidy Assenza a, Alice Huang a, Anne Marie Valente a, Sitaram M. Emani b, Kimberlee Gauvreau a, Audrey C. Marshal a, Doff B. McElhinney a, Michael J. Landzberg a a b

Department of Cardiology, Boston Children's Hospital, Boston, MA, United States Department of Cardiac Surgery, Boston Children's Hospital, Boston, MA, United States

a r t i c l e

i n f o

Article history: Received 23 February 2014 Received in revised form 3 May 2014 Accepted 20 June 2014 Available online 28 June 2014 Keywords: Outflow tract conduit Adults Congenital heart disease

a b s t r a c t Background: Subpulmonary ventricular outflow conduits are utilized routinely to repair complex congenital cardiac abnormalities, but are limited by the inevitable degeneration and need for reintervention. Data on conduit durability and propensity to dysfunction in the adult population are limited. Methods: The study included 288 consecutive patients ≥18 years of age who were evaluated between 1991 and 2010 after placement of a ≥18 mm conduit. Freedom from hemodynamic conduit dysfunction served as our primary outcome. Freedom from reintervention, overall mortality and heart transplantation were also evaluated. Results: Median age at conduit implant was 19 years and median follow-up duration was 13 years. Probabilities of survival without conduit dysfunction and reintervention at 5, 10 and 15 years were 87%, 63%, and 49%, and 95%, 81%, and 56%, respectively. Smaller conduit diameter (18–20 mm) was associated with lower probability of survival without dysfunction in the entire study cohort, with prominent effects in patients in both the lowest and the highest age quartiles. Other parameters with similar associations were higher BMI, native anatomy of tetralogy of Fallot or truncus arteriosus, and active smoking. Conclusions: Adult congenital heart disease patients with conduit diameter ≥18 mm had an approximately 50% chance of developing hemodynamic conduit dysfunction and undergoing conduit reintervention by 15 years of post-implant, and a 30% likelihood of undergoing conduit reoperation in the same time frame. The importance of these data is underscored by the increasing number of adults with congenital heart diseases seeking care and the recent advances in transcatheter valve replacement for dysfunctional conduits. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction It has been nearly five decades since an artificial conduit connecting the right ventricle (RV) and the pulmonary artery (PA) was first placed at surgery in a 6-year-old child with pulmonary atresia [1]. As survival among patients with complex congenital heart disease continues to improve and adults have become the fastest growing segment of the congenital heart disease population [2], the performance of RV outflow tract (RVOT) and RV to PA conduits in adults has become increasingly important. A principal shortcoming of the artificial conduits is limited durability, which inevitably leads to the need for reintervention. In data derived mainly from pediatric literature, variables repeatedly shown to be associated with shorter time to reintervention include younger age,

higher RVOT pressure gradient, conduit type (mainly homografts), underlying anatomy [mainly tetralogy of Fallot (TOF)], and multiple surgeries [3–9]. Data on conduit-related adverse outcomes in adult congenital heart disease (ACHD) patients are more limited, and are insufficient to counsel conduit-implanted adults about the expected durability of their conduit, whether it was implanted during adulthood or earlier in life. The aim of this study was therefore to evaluate conduit durability and factors associated with the development of conduit dysfunction or reintervention among ACHD patients, regardless of the age at implantation. The intention of this design was to provide focused data on patients that are either transitioning from pediatric to adult cardiac care, or undergoing a conduit implant procedure in adulthood. 2. Materials and methods 2.1. Study population and data collection

⁎ Corresponding author at: Adult Congenital Heart Disease Service, Department of Cardiology, Boston Children's Hospital, Boston, MA 02115, United States. Tel.: +1 617 355 5072; fax: +1 617 739 8632. E-mail address: [email protected] (J. Buber).

http://dx.doi.org/10.1016/j.ijcard.2014.06.023 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.

This study included consecutive patients with biventricular congenital heart diseases who underwent conduit placement between the RV and the PA at surgery and evaluated at age ≥ 18 years at our institution between 1/1/91 and 12/31/10. For convenience, the

456

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463

Fig. 1. Conduit diameters from a sample of 60 patients aged N18 at the time of conduit replacement surgery.

term RVOT conduits is used throughout this manuscript, although RVOT conduits and RV to PA conduits can be placed at different locations in the RV. (RVOT conduits are defined as conduits anastomosing between the apical part of the RV infundibulum and the PA's, whereas RV–PA conduits are conduits that connect the RV mass to the PA's.) Only patients with a conduit diameter size of ≥18 mm were included, a size chosen based on an exploratory analysis on a random sample who underwent last RVOT conduit implantation at age ≥18 years (Fig. 1). This allowed patients with an “adult sized” conduit implanted at a younger age to be included in the analysis. Patients were included regardless of where the surgery was performed, but for purposes of adequate and complete data acquisition, follow-up visits and conduit-directed interventions (if performed) had to take place at our institution. Any type of conduit was allowed, yet patients with bioprosthetic valves without a conduit were excluded; accordingly, when a conduit is described as a “bioprosthetic valve” in this cohort, that always indicates a conduit that contains a bioprosthetic valve. Patients were also excluded if: 1) the pulmonary ventricle was a morphologic left ventricle; 2) the conduit was deliberately banded; or 3) demographic, clinical and follow-up data were incomplete. Clinical, imaging, and interventional data from the follow-up period and from the time of conduit dysfunction and reintervention (if occurred) were collected from medical records. For outcome analysis, patients born with TOF (with/without pulmonary valve atresia) and those born with truncus arteriosus (TA) were grouped together due to similar surgical consideration and conduit location within the mediastinum. The study protocol, including waiver of consent for retrospective medical record review and angiographic reviews, was approved by the Institutional Committee for Clinical Investigation. 2.2. Definitions The “index surgery”, the surgery in which the first conduit ≥18 mm in diameter was placed, was considered as “time zero” for freedom from event analyses. The internal

conduit diameters were recorded from manufacturer specifications; nominal diameters were used for homograft conduits. The “last” echocardiographic and MRI studies were the latest studies obtained during follow-up, or the last prior to conduit-related intervention (if performed). Mild conduit (pulmonary) regurgitation (PR) was defined as one or both of the following: regurgitant jet width b1/3 of the width of the RVOT diameter and retrograde pressure drop maintained throughout diastole. Moderate PR was defined as one or both of the following: jet width between 1/3 and 2/3 of the RVOT and equilibration between pulmonary artery and RV pressures in late diastole. Severe PR was defined as one or more of the following: jet width N2/3 of RVOT, regurgitation duration/total diastole duration ratio N0.77, pressure half time b100 ms or presence of diastolic flow reversal in branch pulmonary arteries [10–12]. Severe conduit (pulmonary) stenosis (PS) was defined as a mean Doppler gradient N40 mm Hg [13]. The presence of branch pulmonary stenosis (unilateral or bilateral) was determined based on either an echocardiogram or an MRI study, when such was obtained. Severity was determined based on both the cross sectional area and the relative flow (in unilateral stenosis) as demonstrated by either lung perfusion scans or differential flow evaluation in patients who had MRI studies. Hypertension and diabetes were defined in accordance with the corresponding guidelines at the time of the clinic visit. Renal dysfunction was defined as glomerular filtration rate b60 ml/min. Lifestyle habits, including smoking, were recorded during all clinic visits via either verbal or written questionnaires. 2.3. Statistical analysis The primary endpoint of this study was hemodynamically significant conduit dysfunction, defined as severe PS and/or severe PR. Additional outcomes evaluated included conduit-related reintervention (surgical or percutaneous) and death and heart transplantation. For patients who underwent non-valved conduit implantation at the index surgery, time to reintervention was the only primary endpoint, given the free PR that per definition exists after utilization of these conduits at surgery. Continuous demographic, clinical, and procedure-related data are presented as median (minimum–maximum, interquartile range). Categorical data are presented as frequencies and percentages. Time-related outcomes including freedom from hemodynamic conduit dysfunction and freedom from conduit-directed reintervention, as defined above, were depicted with Kaplan–Meier curves. Patients not experiencing the primary endpoint were censored event-free at the time of the last clinical evaluation within the study period, death, or reintervention undertaken before the occurrence of conduit dysfunction. Demographic, historical, procedural and diagnostic features were assessed for association with time-related outcomes using the log-rank test. Age at the time of surgery was evaluated both by quartile and as a continuous variable by 1-year increments. Body mass index (BMI) at the last follow-up visit/visit prior to diagnosis of conduit dysfunction was evaluated as a continuous variable with 1 kg/m2 increments. Interaction-term analysis was carried to assess the probability of survival without conduit dysfunction by conduit diameter group among patients within the 4 age quartile groups by including an interaction term for conduit diameter (18–20, 21–24, and N24 mm). Interactions with age quartile were similarly performed for type of conduit (homografts and bioprosthetic valved conduits), BMI, and native anatomy (TOF + TA, D-loop transposition of the great vessels, conduit after Ross procedure, and other). Multivariable Cox regression analysis was performed with forward stepwise selection of covariates that were significant to p less than 0.10 on univariable analysis or were deemed to have important clinical significance by the

Fig. 2. Flow chart depicting inclusion/exclusion of patients aged N18 years with conduits measuring ≥18 mm in diameter.

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463 investigators. The covariates included in the model were index conduit diameter (as categorical variable only: 18–20, 21–24, and N24 mm), cardiac condition, conduits placed prior to index surgery, BMI, smoking status, maximum instantaneous Doppler gradient N25 mm Hg across the RVOT on the first echocardiogram obtained after surgery, age (as continuous variable), any comorbidity, and conduit type. Hazard ratios are presented with 95% confidence intervals. A two-sided p value b 0.05 was used to denote statistical significance. Analyses were performed with Statistical Analysis System software version 9.3 (SAS Institute Inc., Cary, NC).

3. Results 3.1. Patients Medical records of 402 patients were reviewed. Complete data were available on 288 consecutive patients with an artificial RVOT conduit sized ≥ 18 mm in diameter, who comprised the study cohort (Fig. 2). The median follow-up duration was 13 years (range 1 to 36.5, interquartile range 6.5–15). Baseline characteristics are presented in Table 1. The distribution of the native anatomy, conduit type, and conduit diameter by age quartile is presented in Fig. 3A, B and C, respectively. The distribution of conduit type by time eras and native anatomy is presented in Fig. 4A and B, respectively. The most common native anatomy was TOF/TA in all age groups, but varied in frequency across the age spectrum; the percentage was highest in the lowest age quartile and lowest in the highest age quartile (p value for between group difference = 0.05). There was a slightly higher percent of patients with TA compared to patients with TOF in the lower age group (73% vs 65%, respectively), while an inverse relationship existed in the highest age group (45% vs. 55%, respectively, data not shown). Conversely, a prior Ross procedure was most common in the highest age quartile and lowest in the lowest age quartile (p = 0.01). Younger patients were also more frequently implanted with homografts, whereas bioprosthetic valves were more frequently utilized in older patients (p value for between lowest and highest age groups difference = 0.07 for both). There was no significant difference in the distribution of conduit type between patients with TA and TOF. 3.2. Outcomes 3.2.1. Survival The 5, 10 and 15 years freedom from death rates were 99% [95% confidence interval (CI): 97–99.5%], 87% (84–93%) and 81% (78–85%), respectively. Fourteen patients died during follow-up, of them 8 developed conduit dysfunction and had undergone conduit-directed reintervention, while 6 were considered free of conduit dysfunction. Of the 8 patients with dysfunction, 3 died in the immediate post-surgical reintervention period. The other 5 died at median of 2.5 years following reintervention; 3 of non-cardiac causes, and 2 of sudden death, presumably arrhythmic. Of the 6 deceased patients free of conduit dysfunction, 4 died of non-cardiac causes and 2 of presumed non-conduit related cardiac causes (1 aortic dissection, 1 sudden death; both had mild-moderate PR without PS on the echocardiogram performed closest to time of death). None of the study patients underwent heart transplantation. 3.2.2. Conduit dysfunction Hemodynamic conduit dysfunction was documented in 158 patients (59% of the patients implanted with valved conduits), and was secondary to severe PS in 90, severe PR in 22, and combined etiology in 48. Of the 138 patients who developed severe PS during follow-up, data concerning the location of the narrowing was available for 131; in 48 (37%) the narrowing was at the site of the distal anastomosis, in 39 (30%) at the level of the valve, in 11 (9%) at the proximal level and in the remaining 33 (25%) there were multiple sites of narrowing. Cardiac MRI was performed during follow-up in 90 patients (31%); of the 76 patients with severe PR, 49 had an MRI performed between conduit dysfunction diagnosis and reintervention. The median calculated

457

indexed right ventricular end diastolic volume (RVEDV) was 152 ml/m2 (range 107–188 ml/m2), with a median Z score of 5.1 (range 2.0–13.4), and a median PR fraction of 43% (range 27–68%). These were significantly higher than the corresponding values recorded in the 41 patients without significant conduit dysfunction at follow-up: median indexed RVEDV 93 ml/m2 (range 45–141 ml/m2), median Z score 1.8 (range 0.8–4.4), and median PR fraction 12% (range 7–36%) (p b 0.001 for between group comparison for all). Median estimated RV function was 49% (range 38%–60%) in the dysfunction group, and 52% in the non-dysfunction group (range 43%–60%, p = 0.10). Similarly, there was no difference in the median estimated left ventricular function between the two groups (56% in the dysfunction group vs. 58% in the non-dysfunction group, p = 0.31).

Table 1 Baseline characteristics of study patients (N = 288). Parameter Male gender Age at first conduit placement (years) Age at index surgery Age N 18 years at index surgery Age at first evaluation Follow-up (years) BMI at last follow-up (kg/m2)

164 (57%) 8 (3.5–33, 0–50) 19.0 (17–22, 9–57) 203 (70%) 20 (18.5–24, 18–72) 13 (6.5–15.5, 1.5–16.5) 23.7 (18–26.5, 16.2–29.8)

Anatomy TOF ± PA Truncus arteriosus D-loop TGA Aortic valve disease, prior Ross procedure Othera Number of surgeries before index surgery Number of conduits placed prior to index surgery

146 (51%) 32 (11%) 44 (15%) 44 (15%) 22 (10%) 1 (0–1, 0–9) 1 (0–1, 0–4)

Conduit-related variables Homograft conduit Aortic Pulmonary Non-homograft Contegra Hancock Carpentier–Edwards Non-valved

187 (65%) 108 (38%) 79 (27%) 77 (26%) 25 (32%)b 32 (41%)b 20 (26%)b 24 (8%)

Index conduit diameter (mm) All 18–20 mm 21–24 mm N24 mm Homografts 18–20 mm 21–24 mm N24 mm Non-homografts 18–20 mm 21–24 mm Branch pulmonary artery stenosis at follow-upe None Mild Moderate-severe Hypertension Active smoking Renal dysfunction

61 (21%) 130 (45%) 97 (33%) 50 (27%)c 73 (39%)c 64 (34%)c 34 (34%)d 40 (40%)d

231 (80%) 40 (14%) 17 (6%) 10 (3%) 42 (14%) 5 (2%)

Data presented as number (%) or median (IQR, range). BMI, body mass index; TOF, tetralogy of Fallot; PA, pulmonary atresia; TGA, transposition of the great arteries. a Double-outlet right ventricle, severe pulmonary stenosis and pulmonary atresia with intact ventricular septum. b % of patients with a bioprosthetic valved conduit. c % of patients with a homograft conduit. d % of patients with a non-homograft conduit. e Left and right pulmonary arteries, as evaluated in echocardiogram studies (all patients) and in MRI studies (90 patients, see text).

458

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463

For patients with documented dysfunction, the median time from diagnosis of dysfunction to reintervention was 9.8 months (range 0.5 to 22.3). Reintervention was not undertaken in 16 patients due to high-risk comorbid conditions, consisting of airway disease in 9 and advanced multi-organ dysfunction in 7. Additional 5 patients had surgery that included conduit replacement yet was performed for non-conduit related indications; conduit replacement was undertaken at the time of surgery due to either moderate PS (n = 3) or PR (n = 2). The initial reintervention was surgical in 88 patients, balloon angioplasty/stenting in 44, and transcatheter pulmonary valve replacement with a Melody valve (Medtronic Inc., Minneapolis, MN) in the remaining 20 (this therapy became available at our institution in 2007). Percutaneous interventions were performed mainly in cases of conduit stenosis at the valvular level (initial treatment for 29 of the 39 patients with narrowing at the valvular level) followed by treatment for narrowing at the level of the distal anastomosis (initial treatment for 15 of the 48 patients with narrowing at the distal anastomosis level). Clinical indications included symptom progression (n = 97), recurrent arrhythmias (n = 21), worsening tricuspid regurgitation (n = 19) and endocarditis involving the prosthetic valve (n = 2), while in the remaining 14 patients the indication was worsening PR/PS. None of the patients were operated due to anastomotic dehiscence or pseudoaneurysm. Of the 44 patients who initially underwent balloon angioplasty/stenting, 33 subsequently underwent surgery after a median of 9.5 months, while the remaining 11 had no further intervention by the

Fig. 3. The distribution of native anatomy (A), conduit type (B) and conduit diameter (C) by age quartile in the study cohort.

3.2.3. Conduit reintervention Conduit-directed reintervention was undertaken in 142 of the patients who experienced hemodynamic conduit dysfunction (90%) and in 10 patients with non-valved conduits (41%) during follow-up.

Fig. 4. The distribution of implant era (A) and native anatomy (B) by conduit type in the study cohort.

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463

459

end of follow-up. Seventeen patients were diagnosed with severe branch PA stenosis at follow-up, all of them underwent either balloon dilation of the branch PA's (n = 10, of them 5 also underwent transcatheter-based intervention of the conduit) or surgical augmentation at the time of conduit replacement (n = 7).

Fig. 5. Kaplan–Meier curves for the overall cohort depicting estimated probabilities of survival free from any conduit hemodynamic dysfunction, any conduit reintervention, and surgical conduit reintervention.

3.2.4. Freedom from conduit-related outcomes The Kaplan–Meier survival curves depicting probabilities of freedom from hemodynamic conduit dysfunction, any reintervention, and surgical reintervention are presented in Fig. 5. The cumulative probabilities of freedom from dysfunction at 5, 10 and 15 years were 87% (95% CI: 82–90%), 63% (54–68%) and 49% (39–54%), respectively. The corresponding probabilities for any reintervention were 95% (86–98%), 81% (71–89%) and 56% (47–61%), and those for surgical reintervention only were 97% (93–99%), 90% (81%–98%), and 73% (58–82%). On univariable log rank testing, conduit diameter of 18–20 mm (Table 2 and Fig. 6A), active smoking, native anatomy of TOF/TA, prior conduits placed before the index surgery and the presence of any

Table 2 Associations with hemodynamic conduit dysfunction (N = 264).a Number Gender Male Female Age quartile at index surgery (years) Lowest (7–17) 2nd (17.1–19) 3rd (19.1–22) Highest (22–56) Anatomy TOF ± PA, truncus arteriosus TOF Truncus arteriosus D-loop TGA Aortic valve disease, prior Ross Other Conduits placed prior to index surgery Yes No Conduit type Homograft Aortic Pulmonary Non-homografts Contegra valve Other bioprosthetic valved conduit Index conduit diameter (mm) 18–20 21–24 N24 Maximum instantaneous gradient (mm Hg) at 1st follow-up echocardiogram 15–20 20–25 N25 PR on 1st follow-up echocardiogram None Mild Moderate or severe Branch pulmonary artery stenosis at follow-up None Mild Moderate–severe Smoking Yes No Any co-morbidityc Yes No

Patients with conduit dysfunction at follow-upa

p valueb 0.69

151 113

94 64

59 81 62 62

41 54 35 28

164 134 28 39 41 20

108 82 21 24 16 10

168 96

119 39

187 107 79

104 55 49

25 77

16 38

49 120 95

39 73 46

168 84 12

92 57 9

104 139 21

56 86 16

211 37 16

122 24 12

38 226

30 128

42 222

34 124

0.06

0.048

0.05

0.1

0.03

0.06

0.18

0.07

0.02

0.03

PR, Pulmonic regurgitation. Other abbreviations as in Table 1. a Patients implanted with non-valved conduits were not included in this analysis. b p values are for log-rank test, time from index surgery to conduit failure (time to reintervention for patients with non-valved conduits). c Includes diabetes, hypertension, and renal dysfunction.

460

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463

diameter 18–20 mm were significantly associated with shorter freedom from dysfunction. Prior Ross procedure conversely showed a reverse association with this outcome (Table 3). In both the highest and lowest age quartiles the probability of freedom from conduit dysfunction was lower for patients with smaller diameter conduits, although with different profiles over time (Fig. 7A and B). Select patients' parameters with conduit dysfunction at follow-up in these two groups are shown in Table 4: important differences included mode of conduit dysfunction (PS in 69% of the patients in the lowest age group vs. PR in 68% of the highest age group patients), conduits placed prior to index surgery (68% in the highest quartile age group vs. 36% in the lowest) and type of conduit utilized at the index surgery (homografts in the lowest age group with dysfunction, bioprosthetic valved conduits, especially of the Contegra type, in the highest age group with dysfunction). In the second and third age quartiles only a trend existed towards shorter freedom from dysfunction with 18–20 mm diameter conduits (log rank p = 0.07 and 0.21 for the comparison between the largest and the smallest conduit groups, respectively). Additional parameters evaluated for interaction with age included conduit type, BMI and native anatomy; none showed a statistically significant association with the study's outcome. To evaluate for a potential referral bias to our institution within the study cohort, we performed a pre-specified analysis by the site in which the index surgery was performed. Briefly, there was a trend towards shorter freedom from conduit dysfunction among patients operated initially at other centers, which became evident ≈ 5 years post index surgery, but this did not reach statistical significance (log rank p = 0.09). 4. Discussion

Fig. 6. Kaplan–Meier curves for the overall cohort depicting the estimated probability of survival free from conduit hemodynamic dysfunction by conduit diameter (A) and native anatomy (B).

comorbidity were all associated with shorter freedom from dysfunction. Conversely, prior Ross procedure was associated with longer freedom from dysfunction (Fig. 6B). Freedom from dysfunction rates was similar at 5, 10 and 15 years after the index surgery between patients with TOF and TA (82% vs. 79%, 54% vs. 50% and 42% vs. 38%, respectively, data not shown, displayed as a single group in Fig. 6B). Conduit type, age quartile at the time of the procedure, and individual comorbidities did not show a significant association with conduit dysfunction (Table 2). On multivariable Cox regression analysis, higher BMI at last followup visit/visit prior to diagnosis of conduit dysfunction and conduit

As the number of individuals with complex congenital heart diseases surviving into adulthood continues to rise, it is important to characterize outcomes particularly relevant to this population [2]. Given the marked anatomical and interventional heterogeneity, however, analyses incorporating such data need to be planned in a fashion that will address a spectrum of interventions performed at different ages, ideally without compromising data integrity [14]. With this in mind, our study design allowed the inclusion of patients who underwent conduit placement at any age, as long as the conduit was in the “adult diameter” range. We thus believe that this report is likely to provide focused data to assist in counseling ACHD patients on the expected longevity of their conduit, whether placed in adulthood or earlier in life. 4.1. Durability of RVOT conduits in adults The main findings of this study were that adult patients with an RVOT conduit ≥18 mm in diameter had an approximately 50% chance of developing severe conduit dysfunction and undergoing conduit

Table 3 Multivariable analysis: associations with shorter freedom from hemodynamic conduit dysfunction.a Covariate

Hazard ratio (95% confidence interval)

p value

Index conduit diameter (mm) 18–20 21–24 N24

2.58 (1.05–3.91) 2.10 (0.96–4.07) 1

0.03 0.08 –

Anatomy Aortic valve disease, prior Ross procedure TOF ± PA, truncus arteriosus, D-loop TGA Conduits placed prior to index surgery BMI at last follow-upb Smoking

0.74 (0.56–0.92) 1 1.92 (0.92–3.14) 3.19 (1.07–4.32) 1.62 (1.00–2.62)

0.04 0.10 0.01 0.05

a Further adjusted for age as a continuous variable, maximum instantaneous Doppler gradient N25 mm Hg across the RVOT conduit at first follow-up echocardiogram, any comorbidity, branch pulmonary artery stenosis and type of conduit placed at index surgery. Patients implanted with non-valved conduits were not included in this analysis. b Per 1 kg/m2 increment.

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463

461

Fig. 7. Kaplan–Meier curves depicting estimated probabilities of survival free from conduit hemodynamic dysfunction in the highest (A) and lowest (B) quartile age groups by conduit diameter. In figure B, Log rank p is shown for the difference between the 3 groups. The corresponding value for the difference between the largest and the smallest conduits is 0.04.

reintervention by 15 years of post-implant, and a 30% likelihood of undergoing conduit reoperation in the same time frame. Although there was a steady hazard for development of severe conduit dysfunction after implant, the risk of intervention was relatively low until 10 years postoperatively, after which there were steeper hazards for any reintervention and reoperation. Most available data on RVOT conduit durability are derived from studies in which reintervention rates served as primary outcome, with several incorporating data on adult patients (albeit rarely analyzed in a separate fashion) [7–9,15]. Although reported series differ in a variety of aspects, most large contemporary reports suggested relatively similar rates of freedom from reintervention at 5, 10 and 15 years of ≈70–85%, 50–65% and 45–60%, respectively [6–9]. These data, however, do not necessarily reflect actual conduit durability, as patients are usually able to tolerate RVOT dysfunction for a period before needing reintervention. In light of the growing role of percutaneous solutions to treat RVOT, however, data on freedom from actual conduit dysfunction may be of more interest, as such procedures have the potential to shorten dysfunction-to-intervention delay [16]. Moreover, recent data reported from a large multicenter MRI study of patients with TOF suggests that higher RV mass and mass:volume ratio may be more important risk factors for death/sustained ventricular tachycardia than PR-related RV volume changes, suggesting that prolonged pressure-loading of any degree may be more serious in this population than was realized, and may support earlier intervention for RVOT conduit obstruction [17]. Freedom from conduit reintervention in this cohort was similar to that reported in a recent study that focused on pulmonary valve vs. conduit replacement in adolescents [18] and, as might be expected, longer than in most previous reports from the pediatric literature

(although inherent differences exist between such analyses). Nevertheless, it is disappointing to confirm that adults undergoing surgical conduit replacement have an even chance of developing severe conduit dysfunction within 15 years.

4.2. Factors associated with hemodynamic conduit dysfunction Utilization of smaller diameter conduits (18–20 mm in this series) was associated with shorter freedom from dysfunction, and there was a notable interaction between age quartile and conduit diameter at implant. This association was particularly strong in the youngest and oldest groups, more than in the middle two age quartiles. Also, there was a more pronounced discrepancy between 21 and 24 mm and N24 mm diameters in the oldest quartile than the youngest, and a substantial difference in the distribution of conduit dysfunction due to PS vs. PR between these groups. Somatic-valvar size discordance is a probable contributor to this effect in the lowest age quartile patients [3–6]. The solution to this problem has yet to be determined: although implanting large-diameter conduits was postulated to be a possible solution to the mismatch, more recent data showed that conduit oversizing in children does not increase longevity, and substantial oversizing may result in shorter freedom from conduit regurgitation [19]. In the highest age quartile there are likely multiple contributors to this effect that are somewhat more speculative. Higher BMI, active smoking and utilization of specific conduits (mainly Contegra in our cohort) were more common in this group. Both higher BMI [18] and utilization of the Contegra valves at surgery [5] were previously reported to be associated with higher reintervention rates.

462

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463

Table 4 Select patients' parameters in the highest and lowest age quartile groups who had hemodynamic conduit dysfunction. Parameter

Highest quartile age group (n = 29)

Lowest quartile age group (n = 38)

p value

BMI at last follow-up (kg/m2) Height change from surgery to dysfunction (cm) Mode of dysfunction Stenosis Regurgitation Combined Conduit type Homografts Aortic homograft Pulmonary homograft Non-homografts Contegra Hancock Carpentier–Edwards Index conduit diameter (mm) 18–20 21–24 N24 Maximum instantaneous gradient (mm Hg) at 1st follow-up echocardiogram 15–20 20–25 N25 Anatomy TOF ± PA, truncus arteriosus TOF Truncus arteriosus D-loop TGA Aortic valve disease, prior Ross Other Conduits placed prior to index surgery Yes No Any comorbidity Smoking Follow-up duration (years)

26.3 (21.1–29.8) 0.5 (0–1.5)

21.5 (16.2–26.4) 8.9 (0–19.7)

0.01 b0.001

4 (13) 19 (68) 6 (20)

27 (69) 4 (13) 7 (18)

12 (41) 8 4 16 (55) 8 3 5

25 (66) 14 11 12 (31.5) 1 7 4

0 9 (31) 20 (69)

17 (45) 13 (34) 8 (21)

16 (55) 10 (34.5) 3 (8)

13 (34) 18 (47) 7 (18)

16 (55) 11 (61) 5 (45) 7 (24) 4 (14) 2 (7)

18 (47) 7 (29) 11 (55) 13 (34) 4 (10.5) 3 (8)

19 (68) 10 (32) 12 (41) 21 (72) 11.7 (1.9–21.3)

14 (36) 24 (64) 11 (28) 0 14.6 (2.4–36.5)

0.04

b0.001

0.05

0.18

0.04

0.8 b0.001 0.04

Data presented as number (%) or median (range). Abbreviations as in previous tables.

The associations we identified between higher BMI and smoking and shorter freedom from conduit dysfunction are intriguing and may provide important hints as to the potential harmful effects of certain lifestyle habits in ACHD patients with artificial conduits. The difference in the number of smoking patients with dysfunction between the highest (n = 21, 72%) and lowest (n = 0, 0%) age quartiles was striking. Older patients with conduit dysfunction also had significantly higher BMI's than those in the lowest age group. It is possible that the systemic inflammatory state that accompanies both smoking [20] and elevated BMI [21] plays a role in these relationships, and further studies are needed to confirm this assumption. The relationships between a native anatomy of TOF and shorter freedom from conduit dysfunction, and between prior Ross procedure and longer freedom from dysfunction that existed in our cohort come in accordance with reports made by other groups [18,22,23]. Although it is possible that this finding reflects differences in the geometry of the implant site and conduit, which were shown to be important factors in conduit performance [24,25], this explanation is only hypothetical, as data on RVOT geometry prior to or at the time of the index surgery was largely unavailable for collection, and thus not reported in this study. The question of what type of conduit/valve to use for RVOT reconstruction in ACHD patients is unanswered. At many institutions, including ours, there has been a trend towards the use of stented bioprosthetic valves in both adults and children [18,26,27]. In the present study we did not identify differences in durability of different conduit types. Some prior studies that included various valved conduits and patients of all ages found homografts to have shorter longevity [7], while others reported the opposite [28], and a recent review highlighted the variability of findings in this regard [29]. Importantly, our findings may be limited by the fact that we did not evaluate patients with non-conduit

bioprosthetic valves. A prior report from our center found no difference in the longevity of homografts and bioprosthetic valves in adolescents [18], yet it is unknown whether the same is true for adult individuals. Nevertheless, this study should be a useful benchmark for future evaluations along these lines. 5. Limitations Data for this study were collected retrospectively. As our institution is a large tertiary center managing many patients with complex congenital heart disease, selection and referral biases are possible. We attempted to overcome the referral bias by comparing the study's outcome in patients who underwent the index surgery at our center or an outside hospital, which revealed no significant difference. We did not include patients who had 18 mm or larger conduits placed and reintervened upon at b18 years of age, so there may be a “survival” bias inherent in the data from patients with conduits placed before 18 years of age. The decision to undertake conduit reintervention was not standardized, and patients may have undergone intervention with different degrees of PS and PR for variable lengths of time. This limitation was likely of minimal significance, as dysfunction rather than reintervention was the primary outcome, and we found that essentially all patients had severe PR or conduit stenosis at the echocardiogram prior to reintervention. MRI studies were obtained in a minority of study patients, and so we were unable to incorporate this data. As many of the patients included in this analysis were outside referrals after undergoing the index surgery, the precise anatomic data and techniques used at previous procedures, as well data on the PA geometry at the time of the index surgery was not available to us for the majority of these patients and were therefore not incorporated into the analysis. Time to conduit dysfunction and/or

J. Buber et al. / International Journal of Cardiology 175 (2014) 455–463

reintervention can potentially be affected by flow dynamics of RVOT/RVPA conduits. Transcatheter pulmonary valve replacement may lead to shifting patterns of referral for RVOT reintervention, particularly treatment of less severe conduit regurgitation or stenosis. This therapy was available only for the most recent 4 years, and so our findings may not reflect RVOT reintervention practices in the current era. We included patients with 18 mm or larger conduits, which we determined empirically to be indicative “adult size”, but in reality the normal pulmonary valve annulus diameter in adults is larger [30]. Moreover, it is well established that homograft conduits undergo remodeling in situ that frequently results in contraction or shrinkage, but this can vary substantially between patients. However, like most other studies on conduit performance, we only had data on the implanted conduit diameter, not actual conduit lumen diameter over time, which likely confounds our analysis to some degree. References [1] Rastelli GC, Ongley PA, Davis GD. Surgical repair for pulmonary valve atresia with coronary–pulmonary artery fistula: report of a case. Mayo Clin Proc 1965;40:521–7. [2] Care of the adult with congenital heart disease: 32nd Bethesda Conference. J Am Coll Cardiol 2001;37:1161–89. [3] Mokhles MM, van de Woestijne PC, de Jong PL, et al. Clinical outcome and healthrelated quality of life after right-ventricular-outflow-tract reconstruction with an allograft conduit. Eur J Cardiothorac Surg 2011;40:571–8. [4] Poynter JA, Eghtesady P, McCrindle BW, et al. Congenital Heart Surgeons' Society. Association of pulmonary conduit type and size with durability in infants and young children. Ann Thorac Surg 2013;96:1695–702. [5] Urso S, Rega F, Meuris B, et al. The Contegra conduit in the right ventricular outflow tract is an independent risk factor for graft replacement. Eur J Cardiothorac Surg 2011;40:603–9. [6] Mohammadi S, Belli E, Martinovic I, et al. Surgery for right ventricle to pulmonary artery conduit obstruction: risk factors for further reoperation. Eur J Cardiothorac Surg 2005;28:217–22. [7] Dearani JA, Danielson GK, Puga FJ, et al. Late follow-up of 1095 patients undergoing operation for complex congenital heart disease utilizing pulmonary ventricle to pulmonary artery conduits. Ann Thorac Surg 2003;75:399–410. [8] Askovich B, Hawkins JA, Sower CT, et al. Right ventricle-to-pulmonary artery conduit longevity: is it related to allograft size? Ann Thorac Surg 2007;84:907–11. [9] Caldarone CA, McCrindle BW, Van Arsdell GS, et al. Independent factors associated with longevity of prosthetic pulmonary valves and valved conduits. J Thorac Cardiovasc Surg 2000;120:1022–30. [10] Li W, Davlouros PA, Kilner PJ, et al. Doppler-echocardiographic assessment of pulmonary regurgitation in adults with repaired tetralogy of Fallot: comparison with cardiovascular magnetic resonance imaging. Am Heart J 2004;147:165–72. [11] Silversides CK, Veldtman GR, Crossin J, et al. Pressure half-time predicts hemodynamically significant pulmonary regurgitation in adult patients with repaired tetralogy of Fallot. J Am Soc Echocardiogr 2003;16:1057–62.

463

[12] Grothoff M, Spors B, Abdul-Khaliq H, Gutberlet M. Evaluation of postoperative pulmonary regurgitation after surgical repair of tetralogy of Fallot: comparison between Doppler echocardiography and MR velocity mapping. Pediatr Radiol 2008;38:186–91. [13] Baumgartner H, Hung J, Bermejo J, et al. Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice. Eur J Echocardiogr 2009;10:1–25. [14] Warnes CA, Liberthson R, Danielson GK, et al. Task force 1: the changing profile of congenital heart disease in adult life. J Am Coll Cardiol 2001;37:1170–5. [15] Oosterhof T, Meijboom FJ, Vliegen HW, et al. Long-term follow-up of homograft function after pulmonary valve replacement in patients with tetralogy of Fallot. Eur Heart J 2006;27:1478–84. [16] McElhinney DB, Hellenbrand WE, Zahn EM, et al. Short- and medium-term outcomes after transcatheter pulmonary valve placement in the expanded multicenter US melody valve trial. Circulation 2010;122:507–16. [17] Valente AM, Gauvreau K, Egidy Assenza G, et al. Contemporary predictors of death and sustained ventricular tachycardia in patients with repaired tetralogy of Fallot enrolled in the INDICATOR cohort. Heart 2014;100:247–53. [18] Batlivala SP, Emani S, Mayer JE, McElhinney DB. Pulmonary valve replacement function in adolescents: a comparison of bioprosthetic valves and homograft conduits. Ann Thorac Surg 2012;93:2007–16. [19] Karamlou T, Ungerleider R, Alsoufi B, et al. Oversizing pulmonary homograft conduits does not significantly decrease allograft failure in children. Eur J Cardiothorac Surg 2005;27:548–53. [20] Tracy RP, Psaty BM, Macy E, et al. Lifetime smoking exposure affects the association of C-reactive protein with cardiovascular disease risk factors and subclinical disease in healthy elderly subjects. Arterioscler Thromb Vasc Biol 1997;17:2167–76. [21] Trayhurn P, Wood IS. Adipokines: inflammation and the pleiotropic role of white adipose tissue. Br J Nutr 2004;92:347–55. [22] Boethig D, Goerler H, Westhoff-Bleck M, et al. Evaluation of 188 consecutive homografts implanted in pulmonary position after 20 years. Eur J Cardiothorac Surg 2007;32:133–42. [23] Selamet Tierney ES, Gersony WM, Altmann K, et al. Pulmonary position cryopreserved homografts: durability in pediatric Ross and non-Ross patients. J Thorac Cardiovasc Surg 2005;130:282–6. [24] Nordmeyer J, Tsang V, Gaudin R, et al. Quantitative assessment of homograft function 1 year after insertion into the pulmonary position: impact of in situ homograft geometry on valve competence. Eur Heart J 2009;30:2147–54. [25] Babu-Narayan SV, Diller GP, Gheta RR, et al. Clinical outcomes of surgical pulmonary valve replacement after repair of tetralogy of Fallot and potential prognostic value of preoperative cardiopulmonary exercise testing. Circulation 2014;129:18–27. [26] Chen XJ, Smith PB, Jaggers J, Lodge AJ. Bioprosthetic pulmonary valve replacement: contemporary analysis of a large, single-center series of 170 cases. J Thorac Cardiovasc Surg 2013;146:1461–6. [27] Giamberti A, Chessa M, Reali M, et al. Porcine bioprosthetic valve in the pulmonary position: mid-term results in the right ventricular outflow tract reconstruction. Pediatr Cardiol 2013;34:1190–3. [28] Homann M, Haehnel JC, Mendler N, et al. Reconstruction of the RVOT with valved biological conduits: 25 years experience with allografts and xenografts. Eur J Cardiothorac Surg 2000;17:624–30. [29] Yuan SM, Mishaly D, Shinfeld A, Raanani E. Right ventricular outflow tract reconstruction: valved conduit of choice and clinical outcomes. J Cardiovasc Med 2008;9:327–37. [30] Capps SB, Elkins RC, Fronk DM. Body surface area as a predictor of aortic and pulmonary valve diameter. J Thorac Cardiovasc Surg 2000;119:975–82.