Author + information
- Received November 3, 2011
- Revision received April 25, 2012
- Accepted May 1, 2012
- Published online September 11, 2012.
- Lindsay S. Rogers, MD⁎,†,
- Andrew C. Glatz, MD⁎,†,
- Chitra Ravishankar, MD⁎,†,
- Thomas L. Spray, MD⁎,‡,
- Susan C. Nicolson, MD⁎,§,
- Jack Rychik, MD⁎,†,
- Christina Hayden Rush, BSN⁎,
- J. William Gaynor, MD⁎,‡ and
- David J. Goldberg, MD⁎,†,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. David J. Goldberg, The Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadelphia, Pennsylvania 19104
Objectives The aim of this study was to evaluate Fontan peri-operative outcomes for 771 consecutive patients.
Background Since the initial description by Fontan, mortality associated with the Fontan operation has declined substantially. However, postoperative effusions remain a significant challenge. Effusions are a key determinant of postoperative length-of-stay and have been shown to be associated with the development of protein-losing enteropathy and with decreased survival.
Methods This study was a single-center, retrospective review of 771 patients who underwent Fontan palliation from 1992 to 2009.
Results Patients were divided into 3 eras dictated by shift in clinical practice. Overall mortality was 3.5%, 1% since 1996. Importantly, age at Stage II palliation decreased from Era 1 to Era 3 (7.1 vs. 5.9 months; p = 0.0001), whereas age at Fontan increased (1.7 vs. 2.8 years; p = 0.0001). The proportion of patients with prolonged hospital stay (46.7% vs. 8.2% vs. 19.5%, p < 0.001) decreased substantially after Era 1. A diagnosis of hypoplastic left heart syndrome and longer operative support times were associated with prolonged pleural drainage (odds ratio [OR]: 2.17, p < 0.001; OR: 1.2, p = 0.001) and hospital stay (OR: 1.48, p = 0.05; OR: 1.34, p < 0.001). In patients who underwent invasive assessment, higher pulmonary artery pressure was associated with death (OR: 1.37, p < 0.001) and prolonged hospital stay (OR: 1.09, p = 0.019). Pulmonary arterial pressure ≥15 mm Hg was 90% specific for discriminating unfavorable outcomes.
Conclusions Mortality in the modern era is rare, whereas postoperative pleural drainage remains the dominant morbidity. Elevated pulmonary artery pressure seems to be a marker of unfavorable outcome. Continued investigation is warranted to determine whether medical interventions or alterations to operative strategy can alter peri-operative results and improve long-term outcomes.
The concept of rerouting systemic venous return directly to the pulmonary arteries for the palliation of congenital heart disease was first described by Fontan and Kreutzer in the early 1970s as an alternative approach for management of tricuspid atresia (1,2). Since the original description, the evolution of the “Fontan” has included numerous surgical advances, along with parallel advances in anesthesia management, perfusion strategies, and intra-operative and postoperative care. In sum, these changes have led to substantial reductions in early mortality (3–6) such that several recent reports demonstrate mortality rates of 0% to 5.5% (7–11). However, despite continued improvements, the incidence of postoperative pleural effusions remains high and is a key determinant of postoperative length-of-stay (7,9,11). Prolonged peri-operative course has been associated with the late development of protein-losing enteropathy and with decreased long-term survival, suggesting that improvements in peri-operative course might have both short- and long-term benefits (9). Factors previously shown to influence peri-operative morbidity and mortality include: anatomy, atrioventricular valve regurgitation, ventricular function, pulmonary artery pressure (PAP), support times, modified ultrafiltration (MUF), and Fontan type (12–15).
In this study we review the peri-operative outcome for 771 consecutive patients after the Fontan operation at The Children's Hospital of Philadelphia from January 1, 1992, to December 31, 2009. We identify trends in patient characteristics, describe changes in surgical technique, and evaluate the impact of these variables on early morbidity and mortality with a goal of identifying factors that could be modified to reduce early postoperative morbidity.
Data from all patients who underwent a first-time Fontan operation during the study period were included. Demographic, anatomic, and procedural variables were recorded from the medical record. In reporting procedural variables, total support time (TST) represents the time circulation was supported or arrested, equal to cardiopulmonary bypass (CPB) plus deep hypothermic circulatory arrest (DHCA). If multiple periods of CPB and DHCA were used, the times were summed. Duration of pleural drainage was defined as total days from surgery to removal of final chest tube or drain or final thoracentesis, performed during initial admission or during a readmission within 30 days of surgery. Hospital length-of-stay was defined as total days of initial hospital stay plus any additional days during readmission(s) within 30 days of surgery. The Institutional Review Board at The Children's Hospital of Philadelphia approved this study.
The surgical strategies for both the superior cavo-pulmonary connection and the Fontan procedure were not standardized and have evolved considerably over the time period of this study. Because 8 different surgeons were involved in the care of these patients, multiple different techniques have been used with regard to the creation of pathway fenestrations, the use of patch material to augment the branch pulmonary arteries, and the timing and style of cardiopulmonary bypass. Management of cardiopulmonary bypass and myocardial protection were at the discretion of each surgeon, as were the specific surgical techniques used.
All demographic, anatomic, clinical, operative, and postoperative variables were summarized with standard descriptive statistics and expressed as mean ± SD for normally distributed continuous variables, median (with range) for skewed continuous variables, and count (with percentage) for categorical variables. Differences across eras, described in the following text, were assessed with analysis of variance, Kruskal-Wallis (nonparametric analog of analysis of variance), or Pearson chi-square. Duration of hospital stay and chest tube drainage were converted to dichotomous variables on the basis of standards from previous literature and from review of the distribution of our data. Associations between all potential covariates of interest and dichotomous outcome measures were assessed in univariate testing with logistic regression. Multivariate logistic regression was then used to identify factors independently associated with outcomes. Covariates with a p value of <0.2 on univariate analysis and observations in at least 90% of subjects were considered for inclusion in the multivariate analysis. Two clusters of highly collinear covariates identified: TST, CPB, aortic cross clamp, and DHCA; and era, Fontan type, and use of MUF. No more than 1 variable from each collinear cluster was included in the final multivariate model. When more than 1 variable met inclusion criteria in the final model, the variable that produced the model with the lowest Akaike Information Criterion score (post-test goodness-of-fit measure) was selected. Because pre-Fontan cardiac catheterization data were available in <90% of subjects, sub-analysis was performed to generate best-fitting multivariate models, including catheter-derived variables and reported separately. Statistical significance was determined a priori as a 2-tailed p value <0.05. All statistical analyses were performed with STATA software (version 10, StataCorp., College Station, Texas).
A total of 771 patients underwent first-time Fontan operation at The Children's Hospital of Philadelphia between January 1, 1992 and December 31, 2009. The average number of Fontan operations in a given year was 43. On the basis of changes in clinical practice (outlined below), the cohort was divided into 3 eras (Fig. 1): 1992 to 1996 (Era 1, n = 227), 1997 to 2002 (Era 2, n = 195), and 2003 to 2009 (Era 3, n = 349). The eras represent introduction of MUF, change in type of fenestration, and a shift toward the use of the extracardiac conduit (ECC). In Era 1 most Fontan completions were intra-atrial lateral tunnels (IALT). Era 2 was a transition period during which Fontan types were largely balanced between IALT and ECC. In Era 3 most Fontan operations were ECC. Routine use of MUF was introduced in 1996.
Baseline patient characteristics at time of Fontan
Demographic, anatomic, preoperative, and operative characteristics across era are outlined in Table 1. The distribution of ventricular morphology varied significantly across eras as the overall proportion of subjects with hypoplastic left heart syndrome (HLHS) decreased from 57.4% in Era 1 to 44.0% in Era 3. In the most recent era the number of patients with greater than mild atrioventricular valve regurgitation increased significantly compared with the 2 older cohorts. In those who underwent preoperative cardiac catheterization, there were significant differences in ventricular end-diastolic pressure—highest in Era 3, and lowest in Era 2. There was no significant difference in mean PAP across eras. An increase in the use of bidirectional Glenn operation was noted in the 2 most recent cohorts, mirroring the increased use of the ECC. Age at Stage II palliation decreased over the course of the study, leading to longer time between Stage II palliation and the Fontan.
Over the 18 years included in the review, Fontan type was evenly split between ECC and IALT (Table 1). During the study period, no patient underwent an atrio-pulmonary connection. The median age at Fontan increased from Era 1 to Era 3; as did weight at Fontan. The TST increased over the 3 eras. Although the overall rate of fenestration did not significantly change across era, the use of single-punch or side-to-side fenestrations increased significantly. The number of additional procedures performed at the time of the Fontan was highest in the most recent era.
Outcome comparisons across eras
The variations in outcomes across eras are outlined in Table 2. In the complete cohort there were 27 deaths before hospital discharge, and within 30 days of the Fontan operation, there was a peri-operative mortality rate of 3.5%. However, mortality decreased significantly from Era 1 to Era 3. Overall, there were 8 patients who required Fontan takedown, with no significant variation by era. The largest proportion of patients requiring hospital stay >14 days or with persistent pleural drainage for >14 days occurred in Era 1 and decreased substantially in Era 2 and Era 3.
Risk factors for death, prolonged hospital stay, and prolonged effusion
Significant risk factors for death or Fontan takedown within 30 days of the Fontan operation are outlined in Table 3 and include the presence of a common atrioventricular valve (AVV) and longer DHCA. The use of MUF was associated with decreased odds of death or Fontan takedown within 30 days of Fontan. Risk factors for prolonged hospital stay (>14 days) are outlined in Table 4 and include a diagnosis of HLHS and longer TST. In contrast, older age at the time of Fontan, use of fenestration, and MUF were associated with a lower occurrence of prolonged hospital stay. Risk factors associated with prolonged effusion (>14 days) are described in Table 5 and include a diagnosis of HLHS and longer TST. Creation of a fenestration and use of MUF were associated with a decreased incidence of prolonged effusion. Results of the sub-analysis of patients who received a preoperative cardiac catheterization are shown in Table 6. Higher mean PAP was an independent risk factor for death within 30 days and hospital stay >14 days. In addition, although there was no absolute cut-point, a PAP ≥15 mm Hg was 90% specific in discriminating unfavorable outcomes for prolonged effusion, length-of-stay, and death (Table 7).
Support time and Fontan type
Given the increased length-of-stay and duration of effusion in the most recent era and the identified association with longer TST, we independently evaluated factors associated with increased support times (Table 8). Multivariate analysis demonstrated that ECC, need for additional procedures (such as arch augmentation and atrioventricular valvuloplasty), and larger patient size were independently associated with longer TST.
This study reviews changes in preoperative patient characteristics and surgical techniques and evaluates the impact of those changes on morbidity and mortality for 771 consecutive patients undergoing the Fontan operation over an 18-year period at a single institution. This is the largest contemporary report of peri-operative outcomes after the Fontan operation. In this cohort, overall mortality remained approximately 1% from 1996 to 2009, with duration of pleural tube drainage and length-of-stay that compare favorably with other contemporary reports (Table 9) (7–11). Interestingly, although overall mortality has remained low, duration of pleural drainage and length-of-stay have both increased slightly in the most recent era, likely related to an increase in patient complexity and a change in practice toward greater use of the ECC. Nevertheless, our results demonstrate that in the current era the Fontan operation can be performed with low early postoperative mortality and relatively short hospital stays.
Changes in patient characteristics and surgical technique
Over the 18 years of this study there has been an evolution toward performing Stage II earlier and Fontan completion later; possibly related to concern about inter-stage mortality for children with shunt-dependent physiology (16,17) coupled with the desire to place a sufficiently large ECC. In the most recent era a higher proportion of patients had more than mild AVV regurgitation, ventricular end-diastolic pressure was significantly higher, and a higher proportion of patients required additional procedures at the time of the Fontan operation, suggesting that this might be a higher-risk cohort. These changes in patient characteristics suggest an improvement in the survival of more marginal patients through the first 2 stages of palliation.
The most notable operative changes over this study period were the introduction of MUF, the shift from the use of an IALT to an ECC, and the shift to a single-punch or side-to-side fenestration creation in the most recent eras. The change in practice toward the use of ECC was primarily based on the expectation that an ECC will provide a better long-term outcome. Longer-term follow-up is needed to determine whether the proposed benefit is indeed valid.
Risk factors for death, prolonged hospital stay, and prolonged effusion
A common AVV and longer DHCA, both markers of surgical complexity, were associated with an increased risk of death or takedown. A diagnosis of HLHS and longer TST were associated with an increased risk of both prolonged effusion and hospital stay. Our data confirm, similar to other reports, the association of MUF and fenestration with a decrease in morbidity and mortality after Fontan (3). Sub-analysis of patients who had a preoperative cardiac catheterization found that higher mean PAP, a potential target for pharmacologic intervention, was also associated with an increased risk of death or takedown and prolonged hospital stay. Our analysis identified a PAP ≥15 mm Hg as a good discriminator of unfavorable outcomes for prolonged effusion, length-of-stay, and death.
Support time and Fontan type
We report an association between longer TST and both duration of hospital stay and duration of pleural drainage. In addition, we found that factors associated with prolonged TST include use of ECC and need for additional procedures at the time of the Fontan operation. Although the need for additional procedures is not easily modifiable, the use of ECC was a change in practice predicated upon the potential long-term benefits of ECC versus IALT. Given the increase in pleural drainage and duration of hospital stay associated with use of ECC, a definitive study evaluating its potential long-term benefit would be useful to justify its continued widespread use. Another potential option proposed by some centers is to perform the Fontan operation without use of CPB. However, reported results of “off-pump” Fontan operations are similar to those reported herein (11,18).
As a retrospective review, our analysis was restricted to data recorded in patient charts and can show only associations and not causality. In our analysis the use of MUF, Fontan type, and era were among the collinear variables, making the individual contributions of each variable difficult to identify. The use of MUF produced the best result on goodness-of-fit testing and therefore was used in this analysis. Although we believe MUF has benefits, there might be other unquantifiable changes across eras, such as specialized nursing care and improvement in postoperative medical management that also contributed significantly to the improved outcomes. Furthermore, because of the relatively small number of outcomes, there is a possibility of over-fitting the final multivariate model for death or takedown.
In this study we describe early postoperative outcomes for 771 consecutive patients after the Fontan operation over an 18-year period at a single institution. Importantly, despite increasing patient complexity, mortality in this cohort was a rare event, and morbidity, primarily in the form of prolonged pleural effusions, was similar to that reported in other contemporary series. Nevertheless, although the use of the ECC might have a theoretical benefit for the prevention of late arrhythmias, in this cohort it was associated with increased pleural drainage. Prolonged pleural drainage has been associated with an increased incidence of protein-losing enteropathy and with decreased survival. Therefore, continued investigation is warranted to determine whether further alterations to operative strategy or targeted interventions aimed at decreasing pulmonary arterial pressure might help to reduce pleural drainage, decrease the frequency of late complications, and improve long-term survival.
This work was supported in part by the Alice Langdon Warner Endowed Chair in Pediatric Cardiothoracic Surgery and the Daniel M. Tabas Endowed Chair in Pediatric Cardiothoracic Surgery. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrioventricular valve
- cardiopulmonary bypass
- deep hypothermic circulatory arrest
- extracardiac conduit
- hypoplastic left heart syndrome
- intra-atrial lateral tunnel
- modified ultrafiltration
- odds ratio
- pulmonary arterial pressure
- total cavopulmonary connection
- total support time
- Received November 3, 2011.
- Revision received April 25, 2012.
- Accepted May 1, 2012.
- American College of Cardiology Foundation
- Fontan F.,
- Baudet E.
- Lemler M.S.,
- Scott W.A.,
- Leonard S.R.,
- Stromberg D.,
- Ramaciotti C.
- Salvin J.W.,
- Scheurer M.A.,
- Laussen P.C.,
- et al.
- Hosein R.B.,
- Clarke A.J.,
- McGuirk S.P.,
- et al.