Author + information
- Received September 18, 2007
- Revision received January 17, 2008
- Accepted January 26, 2008
- Published online July 8, 2008.
- Page A.W. Anderson, MD⁎,⁎ (, )
- Lynn A. Sleeper, ScD†,
- Lynn Mahony, MD‡,
- Steven D. Colan, MD§,
- Andrew M. Atz, MD∥,
- Roger E. Breitbart, MD§,
- Welton M. Gersony, MD¶,
- Dianne Gallagher, MS†,
- Tal Geva, MD§,
- Renee Margossian, MD§,
- Brian W. McCrindle, MD, MPH#,
- Stephen Paridon, MD⁎⁎,
- Marcy Schwartz, MD§,
- Mario Stylianou, PhD††,
- Richard V. Williams, MD‡‡,
- Bernard J. Clark III, MD⁎⁎,
- Pediatric Heart Network Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Page A. W. Anderson, Duke University Medical Center, Research Park Building 2, Room 113, Box 3218, Durham, North Carolina 27710.
Objectives We characterized a large cohort of children who had a Fontan procedure, with measures of functional health status, ventricular size and function, exercise capacity, heart rhythm, and brain natriuretic peptide (BNP).
Background The characteristics of contemporary Fontan survivors are not well described.
Methods We enrolled 546 children (age 6 to 18 years, mean 11.9 years) and compared them within pre-specified anatomic and procedure subgroups. History and outcome measures were obtained within a 3-month period.
Results Predominant ventricular morphology was 49% left ventricular (LV), 34% right ventricular (RV), and 19% mixed. Ejection fraction (EF) was normal for 73% of subjects; diastolic function grade was normal for 28%. Child Health Questionnaire mean summary scores were lower than for control subjects; however, over 80% of subjects were in the normal range. Brain natriuretic peptide concentration ranged from <4 to 652 pg/ml (median 13 pg/ml). Mean percent predicted peak O2 consumption was 65% and decreased with age. Ejection fraction and EF Z score were lowest, and semilunar and atrioventricular (AV) valve regurgitation were more prevalent in the RV subgroup. Older age at Fontan was associated with more severe AV valve regurgitation. Most outcomes were not associated with a superior cavopulmonary connection before Fontan.
Conclusions Measures of ventricular systolic function and functional health status, although lower on average in the cohort compared with control subjects, were in the majority of subjects within 2 standard deviations of the mean for control subjects. Right ventricular morphology was associated with poorer ventricular and valvular function. Effective strategies to preserve ventricular and valvular function, particularly for patients with RV morphology, are needed.
- brain natriuretic peptide
- cardiac magnetic resonance
- diastolic function
Children who have undergone a Fontan procedure for palliation of a functional single ventricle are at risk for medical complications (1). Current therapy is based on expert opinion, retrospective data collection, and single center small studies. Robust clinical trials are needed to guide care for this population. Trial design requires careful phenotyping and understanding of factors that affect outcomes. To this end, the National Heart, Lung, and Blood Institute–funded Pediatric Heart Network (PHN) conducted the largest observational study to date in children who have undergone a Fontan procedure. The primary aim of this report is to characterize this cohort and specific subgroups, with state-of-the-science techniques to assess functional health status, ventricular size and function, exercise performance, brain natriuretic peptide (BNP) concentration, and heart rhythm.
Study design and components
This cross-sectional study recruited subjects age 6 to 18 years old who had not undergone cardiac surgical intervention in the 6 months before enrollment (2). Anatomic, clinical, and surgical data were collected at enrollment (March 2003 to April 2004) by a detailed medical record review with standardized forms. Each subject's tests were conducted within 3 months. The protocol was approved by each center's institutional review board. Written informed consent and assent were obtained.
A total of 1,078 subjects from 7 centers in the U.S. and Canada were screened; 644 were study eligible, and 546 were enrolled (86% consent rate) (2). Age, time since the Fontan procedure, and functional health status scores, collected for nearly all eligible subjects, were similar for enrolled and eligible but not enrolled subjects.
Measures of Functional Health Status
The Physical and Psychosocial Summary scores of the Child Health Questionnaire (CHQ)-Parent Form (PF) 50 were used (3). The CHQ-PF50 has been validated in healthy and chronically ill children and used as a trial end point in pediatrics (4).
Measures of Ventricular Function and Size
The 2-dimensional echocardiograms and Doppler evaluations of standard short- and long-axis views of the ventricle(s) were centrally interpreted by 1 of 2 readers. When possible, measurements and derived indexes were expressed as Z scores relative to body surface area (BSA) or age in normal children (5). Ventricular anatomic abnormalities (Table 1) were characterized as left ventricular (LV), right ventricular (RV), or mixed (e.g., unbalanced atrioventricular [AV] canal). Subjects were classified as having moderate/severe valve regurgitation if RV, LV, or common AV valve regurgitation was moderate or severe; both right and left AV valve regurgitation grades were mild; native aortic valve or native pulmonary valve regurgitation was moderate or severe; or both native aortic valve and native pulmonary valve regurgitation grades were mild. End-diastolic volume (EDV), end-systolic volume (ESV), and mass were obtained with the biplane-modified Simpson's method. For the mixed morphology group, the volume and mass of each ventricle were measured separately, and the combined values were used for data analysis. Tei index was obtained (6). Ventricular diastolic function was assessed with measures derived from pulsed Doppler interrogation: duration of pulmonary vein flow reversal during atrial systole; tissue Doppler peak early diastolic velocity (E′); tissue Doppler peak late diastolic velocity (A′); AV valve peak early diastolic inflow velocity (E); AV valve peak late diastolic inflow velocity (A); deceleration time of the early AV valve inflow (DT); duration of AV valve late diastolic inflow (AT); and systemic ventricular flow propagation rate (FP). Two grading systems (7) were used (Table 2): 1) restrictive pattern present versus absent, and 2) grades of 0 (no impairment) to 3 (greatest impairment in diastolic filling).
Cardiac magnetic resonance (CMR)
The CMR studies performed with 1.5-T scanners were centrally interpreted by a single reader. Subjects were excluded if they were unable to cooperate; had a pacemaker, defibrillator, permanent pacemaker lead, implanted device considered a contraindication according to institutional guidelines, or in some instances intravascular coils; or were <6 weeks from endovascular device implantation.
The standardized imaging protocol included electrocardiographically triggered gradient echo cine magnetic resonance acquisitions in the vertical and horizontal long-axis planes, followed by contiguous short-axis imaging from the AV junction through the cardiac apex. Outcomes included EDV, ESV, mass indexed to BSA1.3, stroke volume (SV) indexed to BSA, and mass/EDV ratio (5).
A maximal ramp exercise test was performed. Percent predicted of normal for maximum oxygen consumption (VO2) and oxygen consumption at anaerobic threshold (VAT) were calculated (8).
A standard 12-lead electrocardiogram was performed and locally interpreted.
Resting BNP plasma concentration was centrally measured with the Shinogi BNP-32 Human Assay (Mayo Clinical Trial Services, Rochester, Minnesota) (2).
Pre-specified subgroups were defined by ventricular morphology, Fontan procedure type, age at enrollment quartile, age at Fontan procedure quartile, and history of Stage II procedure (superior cavopulmonary connection or hemi-Fontan procedure). Subgroup differences in continuous outcomes were assessed with the t test and analysis of variance or nonparametric testing for highly skewed outcomes. Differences in categorical outcome measures were assessed with the chi-square test and, if ordinal, the Mantel-Haenszel test for linear trend. Modified Bonferroni and exact testing were applied to bootstrapped samples to obtain a p value adjusted for multiple pairwise comparisons (9). Analysis of covariance and multivariate logistic regression were used to determine whether outcomes differed by subgroup after adjustment for age, with log transformation for BNP. Additional multivariate linear, logistic, and multinomial regression modeling was used to analyze outcomes by age at Fontan and history of a Stage II procedure. All analyses were conducted with SAS version 9.1 (SAS Institute, Cary, North Carolina) and S-Plus version 6.2 (Insightful Corp., Seattle, Washington).
The 546 subjects were 11.9 ± 3.4 years old at enrollment; 60% were male. The cohort was short (mean ± SD, 34 ± 30th percentile) and underweight (40 ± 32th percentile). The most common diagnoses were tricuspid atresia and hypoplastic left heart syndrome (Table 1), and 59% had an intracardiac lateral tunnel Fontan procedure. A fenestration was performed in 68% and was found to be patent by echocardiography in 32%. After the Fontan procedure, stroke and/or thrombosis occurred in 8%, seizures in 3%, and protein-losing enteropathy in 4%. The prevalences of developmental and cognitive abnormalities and surgical and catheter-based interventions have been published (10).
The distribution of ventricular morphologic subgroups was: LV, 49%; RV, 34%; and mixed, 18% (Table 2). One-third of subjects had predominant nonsinus rhythm, 10% had a history of atrial tachycardia, and 13% had a pacemaker. Ventricular mass and volume were obtained with echocardiography in 406 and 414 subjects, respectively, and by CMR in 161. Mean ejection fraction (EF) was 59 ± 10% by echocardiography and 57 ± 10% by CMR. The EF was normal (echocardiographic Z score >−2) in 73%, although mean echocardiographic EDV, SV, and EF were lower and mass greater than those of normal subjects. The higher values of mass-to-volume ratio measured by echocardiography as compared with CMR reflect known echocardiography underestimating ventricular volume and overestimating mass (11). Forty-nine percent of subjects had semilunar valve regurgitation, and 74% had AV valve regurgitation. Fifty-eight percent of subjects were taking an angiotensin-converting enzyme inhibitor at enrollment. Diastolic function grade was normal in 28%. The median estimated maximum first derivative of ventricular pressure (dP/dtic) was 1,125 mm Hg/s (normal range 850 to 1,350 mm Hg/s) (12). The Tei index was 0.64 ± 0.19 (normal range 0.29 to 0.41). The median BNP was 13 pg/ml (range <4 to 652 pg/ml, mean 26 ± 48 pg/ml).
The cohort had impaired exercise performance: mean % predicted peak VO2 65 ± 16%; % predicted VAT 78 ± 25%. Peak VO2 and VAT were in the normal range for 28% and 63%, respectively, independent of whether maximal effort was achieved.
Mean CHQ summary scores were lower than those of historically healthy control subjects (3) (45 ± 12 vs. 53 ± 9 for Physical, and 47 ± 10 vs. 51 ± 9 for Psychosocial). Individual scores were in the normal range in 81% and 87%, respectively (Fig. 1).
Male subjects had lower BNP levels than female subjects, even after adjustment for age (p = 0.04). No gender differences were present for % predicted peak VO2 and % VAT. By both echocardiography and CMR, boys had larger EDV than girls (p = 0.04), lower mass/EDV ratio (median by echocardiography 1.1 vs. 1.2, p = 0.002), and higher SV/BSA by CMR imaging (53 ± 14 vs. 48 ± 14, p = 0.03).
Age at enrollment
The type of Fontan procedure differed by age at enrollment (p < 0.001) (Table 2). Older children were more likely to have undergone an atrio-pulmonary connection (36% in ≥15-year-old group vs. 1% to 15% in younger age groups), whereas younger children were more likely to have received a total cavopulmonary connection (TCPC) lateral tunnel (26% and 19% in the 2 younger age groups vs. <5% otherwise). Ventricular morphology differed by age (p = 0.02); in particular, the ≥15-year-old group had a greater proportion with LV morphology than the 9- to <11-year-old group (pairwise adjusted p = 0.03). The BNP increased with age (medians 11 to 14 for the 3 youngest cohorts vs. 17 pg/ml for the ≥15-year-old group, p = 0.020). Most other findings on echocardiography and CMR did not differ by age. Tei index increased with age (p < 0.001). Exercise performance differed among age groups and decreased with age (p < 0.001).
In general, ventricular function outcomes were worse for the RV subgroup compared with the LV and, to a lesser extent, the mixed subgroups (Table 3). The EF Z score was −0.6 ± 1.8, −1.4 ± 2.3, and −0.5 ± 2.1 for LV, RV, and mixed subgroups, respectively (p < 0.001). Even after age adjustment, E′ was lower and E/E′ was higher in the RV subgroup (p < 0.001) (Fig. 2). The AV valve regurgitation was worst in the RV subgroup, as was semilunar valve regurgitation (present in 65% for RV vs. 42% in LV and mixed subgroups, pairwise adjusted p < 0.001).
Age-adjusted exercise performance was weakly associated with ventricular morphology, with the LV subgroup having higher % predicted peak VO2 (3-group p = 0.03) and % predicted VAT (3-group p = 0.06) than the non-LV groups. Diastolic function grade, BNP, and CHQ summary scores did not differ by ventricular morphology (Table 3).
Type of Fontan procedure
Few differences were found by type of Fontan procedure after adjustment for age. They included higher BNP in subjects with atriopulmonary connection (raw median 18 pg/ml) compared with BNP in subjects with extracardiac conduits and lateral tunnels (raw medians 13 and 10 pg/ml, respectively, pairwise adjusted p = 0.01 and p = 0.03, respectively). Age-adjusted % predicted VAT differed by type of Fontan procedure (p < 0.001); those who received an intracardiac lateral tunnel (adjusted mean ± SE, 73 ± 2%) had lower % predicted VAT than those with an atriopulmonary connection (84 ± 3%) or with an extracardiac conduit (92 ± 4%).
Age at Fontan procedure
Type of Fontan procedure differed by age at Fontan procedure (p < 0.001) (Table 4). Intracardiac lateral tunnel was most commonly used for Fontan procedures performed at <2 years of age (81%) and decreased steadily with age. Conversely, extracardiac tunnel procedures were performed in 6% of subjects who underwent Fontan at <2 years and in 21% of subjects who underwent Fontan at ≥4 years. Age at Fontan was similar for those who did and did not undergo a Stage II procedure (3.5 ± 2.0 years vs. 3.2 ± 2.3 years).
Rhythm status was associated with age at Fontan (p < 0.0001). Normal sinus rhythm was present in 70% to 74% of those with Fontan performed under 3 years of age and 59% to 62% of those with Fontan performed at ≥3 years. This association remained after adjustment for age at enrollment (p = 0.01).
Age-adjusted mean Tei index (0.60, 0.62, 0.66, and 0.68 in the 4 Fontan age groups, p < 0.001) was significantly worse for subjects who had a Fontan at later ages, even after adjustment for age at enrollment and for ventricular morphology. Moderate to severe AV valve regurgitation was more common in children with Fontan performed at ≥3 years (23% to 26%) compared with children with Fontan performed at <2 or at 2 to <3 years (13% to 16%). After adjustment for age at enrollment, greater severity of AV valve regurgitation was associated with older age at Fontan (p = 0.010). Subjects with RV or mixed ventricular morphology who underwent the Fontan at older ages were more likely to have worse E′, and RV subjects who underwent Fontan at older ages had worse E/E′, even after adjustment for age at enrollment (morphology by age at Fontan interactions p ≤ 0.05).
The BNP, CHQ summary scores, and systolic function did not differ by age at Fontan (Table 4), and after adjusting for age, exercise performance was also unrelated to age at Fontan.
Stage II procedure
Subjects who underwent a Stage II procedure (66%) were younger at enrollment (10.9 ± 2.9 years vs. 14.7 ± 3.2 years) and were less likely to have undergone an atrio-pulmonary connection (7% vs. 33%), even after age adjustment (Table 5). The distribution of LV, RV, and mixed subgroup subjects who underwent a Stage II procedure (44%, 40%, and 16%, respectively) was different than in those who did not (62%, 16%, and 22%, respectively; p < 0.001). No differences were found in predominant rhythm by Stage II status. After adjustment for age, Stage II surgery was not associated with ventricular function or exercise performance, except for higher ventricular mass in subjects who underwent Stage II surgery (age-adjusted mean ± SE 1.2 ± 0.1 vs. 0.4 ± 0.3, p = 0.008).
Stage II surgery was associated with lower age-adjusted mean log BNP for LV subgroup but not RV and mixed subgroups (Stage II × ventricular morphology interaction p = 0.008). The Psychosocial Summary score was lower in Stage II subjects (age-adjusted mean ± SE 46.4 ± 0.6 vs. 49.7 ± 1.0, p = 0.008) and was not explained by ventricular morphology or other factors. No other age-adjusted study outcomes differed significantly by Stage II surgery status within morphologic subgroup.
This study is the largest to date of children who have undergone the Fontan operation. Systematic data on medical history, demographic variables and quantitative measures of ventricular systolic and diastolic function, exercise performance, neurohormonal response, heart rhythm, and functional health status were obtained. Strengths of the study design are contemporaneous data collection, central interpretation of key measures, and large cohort size from multiple geographically dispersed centers. The novelty of the study follows, in part, from sophisticated measurement of ventricular diastolic function, with tissue Doppler echocardiography, BNP plasma concentration measurement, and large size of the study that permitted for the first time a statistically robust assessment of how outcomes differ by ventricular morphology, age at Fontan, and history of Stage II procedure.
Ejection fraction was normal in the majority (73%) of subjects. Our finding of smaller-than-normal EDV contrasts with a study where EDV was 1.6 times larger than normal (13). Such differences are most likely related to changes in management over time, such as earlier volume-unloading surgery. Smaller EDV, as compared with normal subjects, might be a reflection of aerobic deconditioning and might contribute to the blunted ability to increase stroke volume with exercise (14).
Our finding of abnormal diastolic function in 72% of children who had undergone a Fontan has not been previously reported and is concerning. These indexes are dependent on cardiac loading conditions and are unable to distinguish between enhanced chamber compliance and impaired relaxation (7). However, given our findings, future studies using invasive approaches are needed to ascertain whether these children are at risk for diastolic heart failure.
Functional health status
Over 80% of subjects scored in the normal range on the CHQ. However, on average the parents perceived their children as having lower physical and psychosocial functional status than that of historic healthy control subjects. The lower Physical Summary scores are similar to those for children who have undergone thoracic organ transplantation or cardioverter defibrillator implantation (15,16).
Maximal exercise performance was lower than normal and worse in older subjects, consistent with previous studies (14). The same mechanisms proposed previously to impair exercise performance in single ventricle subjects, including absence of a subpulmonary pumping chamber, abnormal endothelial cell function, increased systemic vascular resistance, decreased muscle mass, and deconditioning, are likely present in our subjects (1,14). The finding that exercise performance (% predicted VAT) was lowest in the intracardiac lateral tunnel group, among Fontan types, was an unexpected and unexplained finding.
The subjects demonstrated a wide range of BNP plasma concentrations. The mean concentration was similar to subjects without congenital heart disease and those with congenital heart disease without ventricular dysfunction and lower than that of single ventricle patients with systemic ventricular failure (17). Our finding of lower age-adjusted BNP concentrations in male versus female subjects is consistent with adult data but not with a report of a gender difference only in healthy post-pubertal children (18).
Ventricular morphologic subgroups
Systolic and Diastolic Function
Our finding of apparently impaired systolic function in the RV subgroup relative to the LV and mixed subgroups is consistent with the general opinion that the structure of the RV is suboptimal for a systemic ventricle (1). The greater prevalence of diastolic dysfunction in the RV subgroup, measured by E′ and E/E′, is not likely a consequence of difference in loading conditions compared with the LV and mixed subgroups (19). Of note, the majority of subjects in each morphologic subgroup had Tei index, E′, and E/E′ values outside a 2-SD range for normal children (20).
The tricuspid valve is thought to be more likely to fail as a systemic AV valve, similar to concerns about the relative inadequacy of the RV as a systemic ventricle. Subjects with RV morphology were most likely and those with LV morphology were least likely to have AV valve regurgitation, consistent with that notion. The higher prevalence of semilunar valve regurgitation in subjects with RV morphology might be related to the aortic reconstruction and intrinsic characteristics of the pulmonary (neo-aortic) valve in patients with hypoplastic left heart syndrome.
Functional health status and BNP
Although measures of ventricular performance and exercise capacity varied according to ventricular morphology, CHQ scores and BNP levels did not.
Age at Fontan
Subjects who were older at time of Fontan had worse AV valve function and decreased likelihood of being in sinus rhythm. Poorer valve function and a decrease in sinus rhythm might be related to a longer period of volume overloading (21). These negative associations with older age at Fontan might be used as a rationale to complete the Fontan at an earlier age.
History of stage II procedure
The performance of a Stage II procedure in single ventricle patients follows from the general assumption that this procedure decreases volume loading and its negative effects. Although we postulated that Stage II surgery might be beneficial in some patients, we found that a Stage II procedure was not associated with laboratory measures except for BNP and ventricular mass and a negative association with Psychosocial Summary score. Our findings suggest that further study is needed to assess the impact of performing a Stage II procedure.
This study was limited in several respects. Because only survivors of the Fontan procedure were studied, our findings might not reflect the characteristics of subjects who died in the years after the Fontan. Although the generalizability of our findings is supported by the enrolled subjects being of similar age and functional health status as the eligible but non-consenting subjects (2) and being from geographically dispersed regions, our subjects were recruited exclusively from major medical centers. Functional health status was measured with parental report instruments, which may not match child perception (22).
This largest-to-date multi-center study of children who have undergone a Fontan procedure provides an overview of functional health status, ventricular performance, and exercise performance in current survivors of the Fontan procedure. Ventricular systolic function and functional health status were within normal range in the majority of subjects. Ventricular function and valvular function were negatively associated with RV morphology. Atrioventricular valve function was negatively associated with older age at Fontan completion. Continued follow-up of these subjects will determine whether functional health status is eventually related to measures of ventricular diastolic function. Effective strategies to preserve ventricular and valvular function, particularly for patients with RV morphology, are needed.
The authors thank Minmin Lu for her dedicated assistance with graphics and analysis of the data.
For a complete list of investigators, please see the online version of this article.
Contemporary Outcomes After the Fontan Procedure: A Pediatric Heart Network Multicenter Study
This work was supported by the National Heart, Lung, and Blood Institute, National Institutes of Health/Department of Health and Human Services, grants U01 HL68269 (to Dr. Anderson), HL68270 (to Drs. Sleeper, Colan, Gallagher, Mahony, Geva, Margossian, and Schwartz), HL68292 (to Dr. Williams), HL68290 (to Dr. Gersony), HL68288 (to Dr. McCrindle), HL68285 (Dr. Breitbart), HL68281 (to Dr. Atz), and HL68279 (to Drs. Clark and Paridon). Drs. Anderson and Clark contributed equally to this work.
- Abbreviations and Acronyms
- brain natriuretic peptide
- body surface area
- cardiac magnetic resonance
- atrioventricular valve peak early diastolic inflow velocity
- tissue Doppler peak early diastolic velocity
- end-diastolic volume
- ejection fraction
- end-systolic volume
- systemic ventricular flow propagation rate
- left ventricle/ventricular
- right ventricle/ventricular
- stroke volume
- peak oxygen consumption at anaerobic threshold
- oxygen consumption
- Received September 18, 2007.
- Revision received January 17, 2008.
- Accepted January 26, 2008.
- American College of Cardiology Foundation
- Khairy P.,
- Poirier N.,
- Mercier L.A.
- Landgraf J.M.,
- Abetz L.,
- Ware J.E.
- Perwien A.R.,
- Kratochvil C.J.,
- Faries D.E.,
- Vaughan B.S.,
- Spencer T.,
- Brown R.T.
- Sluysmans T.,
- Colan S.D.
- Nishimura R.A.,
- Tajik A.J.
- Westfall P.H.,
- Young S.S.
- McCrindle B.W.,
- Williams R.V.,
- Mitchell P.D.,
- et al.
- Sluysmans T.,
- Sanders S.P.,
- van der Velde M.,
- et al.
- McCrindle B.W.,
- Williams R.V.,
- Mital S.,
- et al.
- Koch A.,
- Singer H.
- Zile M.R.,
- Brutsaert D.L.