- Email: [email protected]
Fontan-associated protein-losing enteropathy and heart transplant: A Pediatric Heart Transplant Study analysis
http://www.jhltonline.org
ORIGINAL CLINICAL SCIENCE
Fontan-associated protein-losing enteropathy and heart transplant: A Pediatric Heart Transplant Study analysis Kurt R. Schumacher, MD, MS,a Jeffrey Gossett, MD,b Kristine Guleserian, MD,c David C. Naftel, PhD,d Elizabeth Pruitt, MSPH,d Debra Dodd, MD,e Michael Carboni, MD,f Jacqueline Lamour, MD,g Stephen Pophal, MD,h Mary Zamberlan, MS, CPNP,a and Robert J. Gajarski, MDa From the aDepartment of Pediatrics, Mott Children’s Hospital, University of Michigan, Ann Arbor, Michigan; bDepartment of Pediatrics, Ann & Robert H. Lurie Children’s Hospital, Chicago, Illinois; cDepartment of Cardiothoracic Surgery, Children’s Medical Center, University of Texas, Dallas, Texas; dDepartment of Surgery, University of Alabama at Birmingham, Birmingham, Alabama; eDepartment of Pediatrics, Monroe Carell Jr. Children’s Hospital at Vanderbilt, Nashville, Tennessee; f Department of Pediatrics, Duke Children’s Hospital, Durham, North Carolina; gDepartment of Pediatrics, The Children’s Hospital at Montefiore, New York, New York; and the hDepartment of Pediatrics, Phoenix Children’s Hospital, Phoenix, Arizona.
KEYWORDS: Fontan; protein-losing enteropathy; pediatric; heart transplantation; waiting list
BACKGROUND: Post-Fontan protein-losing enteropathy (PLE) is associated with significant morbidity and mortality. Although heart transplantation (HTx) can be curative, PLE may increase the risk of morbidity before and after HTx. This study analyzed the influence of PLE influence on waiting list and post-HTx outcomes in a pediatric cohort. METHODS: Fontan patients listed for HTx and enrolled in the Pediatric Heart Transplant Study from 1999 to 2012 were stratified by a diagnosis of PLE, and the association of PLE with waiting list and post-HTx mortality, rejection, and infection was analyzed. RESULTS: Compared with non-PLE Fontan patients (n ¼ 260), PLE patients listed for HTx (n ¼ 96) were older (11.9 years vs 7.6 years; p ¼ 0.003), had a larger body surface area (1.1 m2 vs 0.9 m2; p ¼ 0.0001), had lower serum bilirubin (0.5 vs 0.9 mg/dl; p ¼ 0.01), lower B-type natriuretic peptide (59 vs 227 pg/ml; p ¼ 0.006), and were less likely to be on a ventilator (3% vs 13%; p ¼ 0.006). PLE patients had lower waiting list mortality than non-PLE Fontan patients (p o 0.0001). There were no intergroup differences for post-HTx survival or times to the first infection or rejection. PLE was not independently associated with increased post-HTx mortality at any time point. CONCLUSIONS: In this multicenter cohort, the diagnosis of PLE alone was not associated with increased waiting list mortality or post-HTx morbidity or mortality. Given the limitations of our data, this analysis suggests that PLE patients in the pediatric age group have outcomes similar to their nonPLE counterparts. Additional multicenter studies of PLE patients with targeted collection of PLEspecific information will be necessary to fully delineate the risks conferred by PLE for HTx. J Heart Lung Transplant ]]]];]:]]]–]]] r 2015 International Society for Heart and Lung Transplantation. All rights reserved.
Reprint requests: Kurt R. Schumacher, MD, MS, Congenital Heart Center, C.S. Mott Children’s Hospital, 1540 E Hospital Dr, Ann Arbor, MI 48109. Telephone: 734-615-2369. Fax: 734-935-9470. E-mail address: [email protected]
Staged palliation to Fontan circulation is the current standard of care for patients with single-ventricle physiology. Unfortunately, severe and life-threatening complications may occur after the Fontan palliation, including protein-losing
1053-2498/$ - see front matter r 2015 International Society for Heart and Lung Transplantation. All rights reserved. http://dx.doi.org/10.1016/j.healun.2015.03.022
2
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
enteropathy (PLE). PLE is characterized by the abnormal loss of protein into the enteral lumen, which results in hypoproteinemia and hypoalbuminemia,1 and although it has been reported in 1% to 11% of Fontan patients,1,2 the true prevalence is unknown. The consequences of enteral protein loss are numerous, including loss of intravascular oncotic pressure leading to diffuse edema, disturbed calcium regulation, enhanced coagulation, quantitative immune abnormalities with probable immune dysregulation, poor bone density, and growth failure1,3; these morbidities cumulatively contribute to marked morbidity and mortality in this patient group.4 PLE is frequently characterized by a waxing and waning course punctuated by flares in protein loss and resulting symptoms. Further, the severity of protein loss and resulting sequelae may vary significantly between patients. This heterogeneity in the disease often makes evaluating the response to therapy challenging. Thus, although multiple pharmacotherapeutic options and anatomic interventions have demonstrated some efficacy in temporizing PLE symptoms, no available therapy has been shown to be consistently curative except heart transplantation (HTx).5–8 The 2006 Pediatric Heart Transplant Study (PHTS) evaluation of Fontan patients undergoing HTx found no difference in survival between PLE and non-PLE patients but only included 25 PLE patients in the cohort.8 This study was limited by a small sample size and limited duration of follow-up. More recently, reports from larger centers, including PLE patients in the current era, found worsened survival in Fontan patients with preserved ventricular systolic function, most of whom had PLE.5,9 This creates a more concerning picture for PLE patients after HTx. Compared with the 2006 PHTS Fontan study, which examined patients who received transplants from 1993 to 2001, the growth of the PHTS database over recent years allows a significantly larger cohort of modern-era patients with at least 1 year of follow-up to be analyzed and allows PLE patients to be analyzed as a separate group entirely. This analysis can now include evaluation of risk factors within the PLE patient group that were not performed in the previous PHTS Fontan study. Thus, to better understand the influence of PLE on waiting list and HTx outcomes, this study used the PHTS database to (1) compare the characteristics and pre-HTx and post-HTx outcomes of Fontan patients with and without PLE, and (2) determine unique pre-HTx risk factors for pre-HTx and post-HTx outcomes in PLE patients. We hypothesized that Fontan patients with PLE possess disease-specific characteristics that increase their risk of poor outcome after HTx.
Methods Patient population and data collection This retrospective cohort study analyzed data from 3,668 patients undergoing HTx between January 1, 1999, and December 31, 2012, at 35 participating PHTS institutions. Demographic, clinical, and event data were collected for all HTx recipients from each participating center. As has been previously described, the data were collected using coded event forms and sent to the data
analysis center at the University of Alabama in Birmingham, where the information was entered into a secure database, verified, and corrected as needed.10 Institutional Review Boards from participating centers approved all studies. The study excluded patients who received multiple solid-organ transplants. Follow-up on all patients was complete through December 31, 2012.
Outcome measures The a priori primary outcome measure was death after HTx. The secondary outcomes studied were time to the first infection and time to the first rejection episode. The PHTS manual defines the outcome “rejection” as a clinical circumstance (biopsy or otherwise driven) leading to an increase in immunotherapy and the outcome “infection” as evidence of an infectious process requiring intravenous therapy or a life-threatening infection requiring oral therapy.11
Statistical analysis We examined data using standard descriptive statistics, including mean and standard deviation, median and interquartile range (IQR), or number and percentage of the total, as appropriate, based on the data’s distribution. Continuous variable distributions were evaluated for equality of variance, and t-tests were used for continuous variable comparisons because they are robust in symmetric distributions. From 3,686 pediatric patients entered into the PHTS undergoing HTx between 1999 and 2012, 356 Fontan patients were stratified into cohorts by the presence or absence of PLE as reported at the time of HTx listing. PLE is reported as a “yes” or “no” check box on the PHTS data-collection forms. The PHTS form manual of operations for form completion does not provide any guidance regarding what constitutes a diagnosis of PLE; centers respond to this item at their own discretion. Similarly, the patient factor “failure to thrive” is also a “yes” or “no” check box without specific criteria set by the PHTS. Two-group comparisons for individual patient factors were performed using appropriate parametric or non-parametric tests as dictated by the statistical characteristics of the data, with t-tests being used for most comparisons of continuous variables. Competing-outcomes analyses examined the temporal influence of PLE on simultaneous risks of death or HTx in patients on the HTx waiting list. Cox proportional hazard analysis examined risk factors for waiting list death censored for HTx in the overall Fontan cohort. Kaplan-Meier time-to-event analyses with log-rank test comparisons evaluated the temporal influence of PLE on post-HTx outcomes. All study patient status was available up to the 10-year study period or until death. No minimum time to follow-up was required to be included in the post-HTx analyses, although every surviving patient had at least 1 year of post-HTx follow-up. Cox proportional hazard analysis assessed PLE as a risk factor for death within the entire study cohort. The PLE cohort was further analyzed using Cox proportional hazard analysis (the variables analyzed are listed in Tables 1 and 2) to determine independent risk factors for postHTx death and Kaplan-Meier time-to-event analyses with right censoring to evaluate the temporal influence of specific factors that could indicated more severe PLE, including intensive care unit hospitalization at HTx, inotrope dependence at HTx, or having a low (o17 kg/m2) or high (421 kg/m2) body mass index (BMI) at HTx, on death within the PLE cohort.
Schumacher et al. Table 1
Influence of Post-Fontan PLE on HTx Outcomes
3
Comparison of Characteristics of Fontan Patients With and Without Protein-Losing Enteropathy at Listing
Characteristicsa
PLE (n ¼ 96)
Non-PLE (n ¼ 260)
p-value
Age at listing, years Body surface area, m2 Panel reactive antibody, % Bilirubin, mg/dl B-type natriuretic peptide, pg/ml C-reactive protein, mg/liter Blood urea nitrogen, mg/dl Protein, g/dl Albumin, g/dl Triglycerides, mg/dl Low-density lipoprotein, mg/dl Male sex White Congenital etiology Status 1 Inotropes (high) Ventilator In intensive care unit In hospital Failure to thrive History of plastic bronchitis NYHA/Ross Class I II III IV Mean right atrial pressure, mm Hg Cardiac index, liters/min/m2 Pulmonary capillary wedge pressure, mm Hg
11.9 1.1 0.0 0.5 58.5 2.9 15.0 4.8 2.7 75.0 87.0 57 82 95 63 37 3 23 41 23 3
7.6 0.9 2.0 0.9 227.0 1.2 16.0 6.5 3.8 73.0 66.0 170 202 259 166 112 34 85 130 79 23
0.0003 0.0084 0.1643 0.0089 0.0063 0.3782 0.8694 0.3097 0.0362 0.6826 0.3578 0.2952 0.1074 0.4617 0.8222 0.4414 0.0063 0.1116 0.2217 0.234 0.042 0.3077
(5.3–14.9) (0.7–1.4) (0.0–15.5) (0.3–0.6) (23.0–163.2) (1.3–7.3) (12.0–24.0) (4.1–5.8) (2.2–3.4) (67.0–104.0) (51.0–91.0) (59.4) (85.4) (99.0) (65.6) (38.5) (3.1) (24.0) (42.7) (23.4) (9.1)
0 0 9 (52.9) 8 (47.1) 14.5 (12.5–17.0) 2.9 (2.5–3.6) 10.5 (8.0–13.0)
(4.0–12.5) (0.6–1.2) (0.0–41.0) (0.5–1.5) (43.5–1,468.0) (0.5–5.6) (13.0–21.0) (5.7–7.2) (3.3–4.3) (62.0–95.8) (48.0–78.0) (65.4) (77.7) (99.6) (64.3) (43.1) (13.1) (32.7) (50.0) (30.4) (26.14)
0 7 (12.3) 28 (49.1) 22 (38.6) 15.0 (10.5–18.0) 3.2 (2.3–3.7) 14.0 (10.0–17.0)
0.6299 0.5741 0.0005
NYHA, New York Heart Association; PLE, protein-losing enteropathy. a Continuous variables are shown as the median (interquartile range) and categoric variables as number (%).
Table 2
Comparison of Characteristics of Fontan Patients With and Without Protein-Losing Enteropathy at Heart Transplantation
Characteristicsa
PLE (n ¼ 70)
Non-PLE (n ¼ 173)
p-value
Age at transplant, years Body surface area, m2 Panel reactive antibody, % Bilirubin, mg/dl B-type natriuretic peptide, pg/ml C-reactive protein, mg/liter Blood urea nitrogen, mg/dl Protein, g/dl Albumin, g/dl Triglycerides, mg/dl Low-density lipoprotein, mg/dl Male White Status 1 Inotropes (high) Ventilator In intensive care unit In hospital Failure to thrive Induction therapy Maintenance steroids
12.2 1.1 0.0 0.5 210.0 0.5 14.0 5.3 3.0 102.5 76.0 40 62 56 37 3 17 36 17 54 18
8.7 0.9 2.0 0.9 751.0 1.6 16.0 6.7 4.0 99.0 69.5 108 131 141 93 18 61 99 54 118 55
0.0041 0.0189 0.7189 0.0682 0.0788 0.4910 0.5755 0.5196 0.2555 0.5479 0.6676 0.4445 0.0248 0.6541 0.8986 0.1242 0.0970 0.4102 0.2822 0.2046 0.9660
(6.1–15.4) (0.8–1.4) (0.0–38.0) (0.3–1.5) (54.0–317.0) (0.5–65.1) (9.0–19.0) (4.4–6.7) (2.4–4.0) (54.0–151.0) (36.0–87.0) (57.1) (88.6) (80.0) (52.9) (4.3) (24.3) (51.4) (24.3) (77.1) (48.7)
(4.9–12.9) (0.7–1.3) (0.0–46.0) (0.5–1.9) (156.0–1,987.0) (0.6–7.7) (10.0–23.0) (5.9–7.5) (3.4–4.4) (62.0–112.0) (54.0–90.0) (62.4) (75.7) (82.5) (53.8) (10.4) (35.3) (57.23) (31.2) (69.0) (48.3)
PLE, protein-losing enteropathy. a Continuous variables are shown as the median (interquartile range) and categoric variables as number (%).
4
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
Results During the study period, 96 Fontan patients with PLE and 260 Fontan patients without PLE were listed for HTx. Complete demographic and clinical comparisons between the two groups at the time of listing are detailed in Table 1. At listing, compared with non-PLE Fontan patients, the PLE patients were older, had a higher BSA, lower serum bilirubin, lower B-type natriuretic peptide, lower albumin, had lower pulmonary capillary wedge pressures, were less likely to have plastic bronchitis, and were less likely to require mechanical ventilation (all p o 0.05). Of the patients on the waiting list, 70 PLE patients and 173 non-PLE Fontan patients underwent HTx. Complete patient characteristics and comparisons at the time of HTx for each group are reported in Table 2. Notably, compared with non-PLE Fontan patients at HTx, PLE patients were more likely to be older, white, have a higher BMI, and have lower pulmonary capillary wedge pressures. Also, the serum albumin levels at HTx between the two groups were similar. Panel reactive antibody level was similar between the groups at listing and at HTx. Competing outcomes after listing for the PLE and nonPLE cohort are demonstrated in Figure 1. Compared with the non-PLE group, PLE patients had a similar rate of transplant at 6 months after listing but had a higher survival rate without HTx and lower pre-HTx death rate (p o 0.0001). Among the entire Fontan cohort, PLE did not confer an independent hazard of death. Age at listing (inverse transformed hazard ratio [HR], 1.07; p ¼ 0.001), status 1A at time of listing (HR, 4.90; p o 0.0001), and higher pulmonary artery pressure at listing (HR, 1.22; p o 0.0001) were associated with increased risk of waiting list death. Fewer previous cardiac surgeries were associated with decreased hazard of waiting list death (HR, 0.78; p ¼ 0.02). Serum albumin levels were included in the model as a continuous variable but were not significant. Receiver operator characteristic curves further analyzed albumin level within the PLE group to determine whether a particular serum level conferred increased risk of poor waiting list outcome, but no optimal stratifying value was found.
The median post-HTx duration of follow-up (including patients who died) was 2.4 years (IQR, 0.4, 4.8 years) for non-PLE Fontan patients and 1.1 years (IQR, 0.2, 3.4 years) for PLE patients. Survival after HTx was then compared between the groups (Figure 2). No significant difference in overall postHTx survival was found between PLE and non-PLE Fontan patients; of the 70 PLE Fontan patients undergoing HTx during the study period, 22 died (31%) compared with 40 of 173 non-PLE Fontan patients (23%, p ¼ 0.12). PLE was not a significant risk factor in an overall hazard analysis of the Fontan cohort (p ¼ 0.32). Similarly, no intergroup differences were found for time to the first rejection (p ¼ 0.85; Figure 3) or for time to the first infection (p ¼ 0.47; Figure 4). The patient factors used in the previous comparisons were included in a Cox proportional hazard analysis within the PLE-only cohort to determine PLE-specific risk factors for non-survival after HTx listing censored at transplant and for non-survival after HTx . Only a history of cerebrovascular accident was associated with an increased hazard of death (HR, 3.18; 95% confidence interval, 1.02–9.96; p ¼ 0.0468). Even after stratifying the PLE group by variables that could indicate a more severe presentation of PLE at the time of HTx, no additional PLE-specific risk factors for poor outcome were detected; that is, no association with worse survival was detected for being in an intensive care unit (p ¼ 0.63), on inotropes at HTx (p ¼ 0.72), or having a markedly high or low BMI (compared with normal, p ¼ 0.19). In addition, no difference was found in post-HTx hospital length of stay between PLE and non-PLE patients (39.2 vs 39.9 days, respectively; p ¼ 0.95).
Discussion Despite significant differences in the characteristics between PLE and non-PLE Fontan patients awaiting transplant, a preHTx diagnosis of PLE was not associated with an increased risk of death on the waiting list or after HTx. Furthermore, PLE had no association with the frequency of post-HTx
Figure 1 Competing outcomes of alive, received a transplant, or died for (A) protein-losing enteropathy (PLE) Fontan patients (n ¼ 96) and (B) non-PLE Fontan patients (n ¼ 260).
Schumacher et al.
Influence of Post-Fontan PLE on HTx Outcomes
Figure 2
Time to death stratified by protein-losing enteropathy (PLE) history.
rejection or infection. These results are helpful in evaluating the feasibility of HTx in Fontan patients with PLE that already carry the increased risk of poor outcome after HTx associated with the Fontan procedure.8,12 Both groups had nearly 20% mortality by 1 year post-HTx, which reinforces the high risk assumed by a center that performs a transplant in a Fontan patient. Importantly, based on this analysis, when a patient’s overall risk of mortality after HTx is being evaluated, a diagnosis of PLE in isolation should not be considered factor that increases risk; given that the heterogeneity of PLE as a disease was not reflected in this analysis, whether severely affected individuals are at increased risk remains a question that warrants further exploration.
Figure 3
5
Given the associations between Fontan patients and worsened outcomes after HTx,8 this information is reassuring. These results corroborate the previous PHTS findings that PLE was not a factor for poor outcome in a small cohort of PLE patients.8 However, the current results differ from those of 2 recent single-center reports comparing outcomes among Fontan patients listed for transplant with preserved ejection fraction (PEF), many of whom had PLE. In reporting the St. Louis Children’s experience, Simpson et al5 found significantly higher rates of infection, death from infection, graft failure, and a non-significantly higher death rate in the PEF group (59% of whom had PLE) compared with Fontan patients with impaired ejection
Time to first rejection stratified by protein-losing enteropathy (PLE) history.
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
6
Figure 4
Time to first infection stratified by protein-losing enteropathy (PLE) history.
fraction (IEF). Griffiths et al9 reported the Boston Children’s experience similarly comparing outcomes of Fontan patients referred for HTx with PEF (56% PLE) vs IEF and found increased overall mortality (with or without HTx listing) in the PEF group. Importantly, however, a significant proportion of the PEF group in their cohort was deemed “too sick” for HTx and thus were never placed on the waiting list. Each of these patients subsequently died, whereas only 1 patient with PLE died after HTx. In reporting the experience at Lurie Children’s Hospital, Backer et al12 performed transplants in 22 Fontan patients with PLE; 9 patients had PEF and 13 had IEF. No survival difference was seen between the groups (22% mortality in the PEF vs 23% in the IEF group). These apparent discrepancies warrant further discussion. It is important to draw a distinction between all patients with PLE, as studied in the Griffiths report, and the PLE patients in our study cohort that were deemed suitable for HTx listing. Because some centers participating in the PHTS did not list sick patients with PLE, they were not captured for this study, thus creating an unintended selection bias to the conclusion that PLE is not a risk factor for HTx. Alternatively, if PLE was a significant factor in the decision to not list these patients, based on our data, the use of PLE status in the decision may have been unfounded. Actually, the cohorts in each of these studies are different, and we may not fully understand the hemodynamic and physiologic mechanisms that results in PLE with “normal hemodynamics.” Our study attempted to detect a “sicker” population of PLE patients within the cohort by stratifying the PLE patients by intensive care unit requirement, inotrope use, and BMI, and we were unable to demonstrate increased risk of death in this cohort. Whether a sub-set of higher-risk PLE patients does exist, what their specific characteristics are, and whether they would have poorer HTx outcomes as a result of PLE remain unanswered
because this PLE-specific information is not collected in the PHTS data set. Moreover, a plausible reason explaining why the PHTS analysis did not mirror the Simpson findings that Fontan HTx recipients experience significantly increased incidences of infection and graft failure5 is similarly lacking. Perhaps individual HTx centers may have different acceptable-risk thresholds, and some PLE patients deemed “too sick to list” at one center would be listed at another center and suffer worse outcomes. The Boston and St. Louis experiences both compared PEF vs IEF patients, with most of the PEF patients having concomitant PLE. The Lurie Children’s experience actually stratified the PLE group by ventricular function and found no difference in outcome. Given this, it is also possible that the worse outcomes seen in PEF Fontan-failure patients are driven by poor outcomes in PEF patients without PLE and that PLE patients should not necessarily be included in the same cohort with non-PLE PEF Fontan patients when examining outcomes. Although no statistically significant difference was detected for overall survival between the groups, the p-value of 0.12 should not be completely dismissed. Despite being a large study cohort by typical PLE-study standards, the number of patients studied may have nevertheless resulted in an underpowered study. Finally, patients aged Z 18 years are not included in the PHTS database. This group may consist mainly of patients who have suffered from a longer duration of PLE and who are more severely affected by PLE sequelae that do confer an increase risk at HTx. Additional multicenter studies of PLE patients with targeted collection of PLE-specific information are clearly necessary to fully delineate the spectrum and risk of death conferred by this disease. Data regarding the duration of PLE before listing, the severity of PLE, including need for albumin infusions and hospitalizations for fluid overload and hemodynamic compromise,
Schumacher et al.
Influence of Post-Fontan PLE on HTx Outcomes
and the intensity and duration of PLE-specific therapies would be useful in fully characterizing the chronicity, severity, and overall morbidity conferred on an individual by PLE. This necessity is further reinforced when evaluating the overall risk profile of both the non-PLE and PLE Fontan patients in this study. Although not reaching statistical significance, the nonPLE Fontan patients tended to have more traditional postHTx risk factors for non-survival at the time of transplant, including a higher rate of ventilation, higher average B-type natriuretic peptide indicating worse heart failure, and a higher rate of failure to thrive.13,14 Given this trend, the lack of difference in overall survival is somewhat surprising and further suggests a need for analysis including PLE-specific information. Further, serum albumin levels are actually not different between the cohorts at the time of HTx. It is possible that improvement in PLE and its accompanying symptoms explains the equalization of albumin levels. However, it is equally plausible that this represents PLE patients who are being very aggressively managed with albumin replacement therapy, because this frequently occurs at centers to optimize patients before HTx. Without PLE-specific data, these findings are unable to be adequately characterized. The lack of difference in secondary outcome incidence also warrants brief discussion. Reports of apparent immune dysfunction in PLE Fontan patients have described distinct abnormalities in immunoglobulins and lymphocyte populations.15–17 The clinical implications of these findings remain unclear but give cause for concern regarding the potential for rejection and infection in PLE patients after HTx. In the PHTS cohort, there was no difference in the time to the first rejection or infection between PLE and non-PLE patients. Thus, although immunologic abnormalities have been shown in PLE patients, they either resolve after transplant or do not confer clinically relevant and identifiable immune problems (i.e., infection or rejection). Finally, the question remains to what degree HTx cures or ameliorates PLE. Several case series have reported highrates of PLE resolution after HTx.6,7,12 This study cannot directly address the question of PLE resolution because the presence of PLE after HTx is not collected by the PHTS. This study has several limitations that require brief discussion. This is a retrospective, registry-based study of a broad cohort of patients listed for HTx; the nature of retrospective, registry-based research may decrease our ability to discern some specific relationships. Second, as described previously, some specific variables of interest to PLE are not collected in the PHTS database, and PLE itself is only defined as a “yes” or “no” data field without further definition, which may allow significant heterogeneity within the group and potential centerdependent misclassification bias. Third, we are only able to comment on patients listed for HTx entered into the PHTS database and not on all patients with PLE in whom HTx is considered or Fontan patients aged 4 18 years (adults are not included in the PHTS). This potentially causes selection bias in the cohort and excludes from analysis the “sickest” PLE patients who may have been deemed ineligible for HTx.
7
Finally, despite using the entire PHTS database, our patient numbers are still relatively limited due to the overall low incidence Fontan patients with and without PLE requiring transplantation. This is particularly notable regarding late outcomes. Few patients, particularly in the PLE group, have follow-up of 5 years or greater. This lack of patients “at-risk” reduces the power of these analyses to detect differences in late outcomes and limits the study’s ability to adequately characterize the groups long-term after HTx. The potential for type II error and true differences between the groups in late-term outcome is possible and will require future study with larger, late-term cohorts. In conclusion, a diagnosis of PLE was not associated with an increased risk of waiting list mortality or post-HTx morbidity or mortality in this PHTS Fontan cohort. This analysis suggests that PLE patients have outcomes similar to their non-PLE counterparts. However, given the acknowledged lack of PLE-specific information, the potential for selection bias, and the inability to truly stratify PLE by severity within the PHTS database, the question of whether severe PLE conveys risk of post-HTx morbidity remains only partially explored. Further research to determine whether factors unique to this population, but not explored in this analysis, confer differential risk is warranted and will require substantial multicenter collaboration in the future. Given the complexity of care involved with the population, this question remains worth fully answering.
Disclosure statement None of the authors has a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript or other conflicts of interest to disclose.
References 1. Rychik J. Protein-losing enteropathy after Fontan operation. Congenit Heart Dis 2007;2:288-300. 2. Mertens L, Hagler DJ, Sauer U, Somerville J, Gewillig M. Proteinlosing enteropathy after the Fontan operation: an international multicenter study. PLE study group. J Thorac Cardiovasc Surg 1998;115: 1063-73. 3. Meadows J, Jenkins K. Protein-losing enteropathy: integrating a new disease paradigm into recommendations for prevention and treatment. Cardiol Young 2011;21:363-77. 4. Goldberg DJ, Dodds K, Rychik J. Rare problems associated with the Fontan circulation. Cardiol Young 2010;20(Suppl 3):113-9. 5. Simpson KE, Cibulka N, Lee CK, Huddleston CH, Canter CE. Failed Fontan heart transplant candidates with preserved vs impaired ventricular ejection: 2 distinct patient populations. J Heart Lung Transplant 2012;31:545-7. 6. Kanter KR, Mahle WT, Vincent RN, Berg AM, Kogon BE, Kirshbom PM. Heart transplantation in children with a Fontan procedure. Ann Thorac Surg 2011;91:823-9: [discussion 829-30]. 7. Davies RR, Sorabella RA, Yang J, Mosca RS, Chen JM, Quaegebeur JM. Outcomes after transplantation for “failed” Fontan: a singleinstitution experience. J Thorac Cardiovasc Surg 2012;143:1183-92: e4. 8. Bernstein D, Naftel D, Chin C, et al. Outcome of listing for cardiac transplantation for failed Fontan: a multi-institutional study. Circulation 2006;114:273-80. 9. Griffiths ER, Kaza AK, Wyler von Ballmoos MC, et al. Evaluating failing Fontans for heart transplantation: predictors of death. Ann Thorac Surg 2009;88:558-63: [discussion 563-4].
8
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
10. Shaddy RE, Naftel DC, Kirklin JK, et al. Outcome of cardiac transplantation in children. Survival in a contemporary multiinstitutional experience. Pediatric Heart Transplant Study. Circulation 1996;94:II69-73. 11. Pediatric Heart Transplant Study. PHTS manual of operations. 2010. http://www.uab.edu/medicine/phts/images/2010_Forms/PHTS_Manual_ of_Operations_2.18.10.pdf. Accessed February 1, 2014. 12. Backer CL, Russell HM, Pahl E, et al. Heart transplantation for the failing Fontan. Ann Thorac Surg 2013;96:1413-9. 13. Almond CS, Gauvreau K, Canter CE, Rajagopal SK, Piercey GE, Singh TP. A risk-prediction model for in-hospital mortality after heart transplantation in US children. Am J Transplant 2012;12:1240-8.
14. Godown J, Donohue JE, Yu S, Friedland-Little JM, Gajarski RJ, Schumacher KR. Differential effect of body mass index on pediatric heart transplant outcomes based on diagnosis. Pediatr Transplant 2014;18:771-6. 15. Garty BZ. Deficiency of CD4þ lymphocytes due to intestinal loss after Fontan procedure. Eur J Pediatr 2001;160:58-9. 16. Chakrabarti S, Keeton BR, Salmon AP, Vettukattil JJ. Acquired combined immunodeficiency associated with protein losing enteropathy complicating Fontan operation. Heart 2003;89:1130-1. 17. Lenz D, Hambsch J, Schneider P, Tarnok A. Protein-losing enteropathy after Fontan surgery: is assessment of risk patients with immunological data possible? Cytometry B Clin Cytom 2003;53:34-9.
ORIGINAL CLINICAL SCIENCE
Fontan-associated protein-losing enteropathy and heart transplant: A Pediatric Heart Transplant Study analysis Kurt R. Schumacher, MD, MS,a Jeffrey Gossett, MD,b Kristine Guleserian, MD,c David C. Naftel, PhD,d Elizabeth Pruitt, MSPH,d Debra Dodd, MD,e Michael Carboni, MD,f Jacqueline Lamour, MD,g Stephen Pophal, MD,h Mary Zamberlan, MS, CPNP,a and Robert J. Gajarski, MDa From the aDepartment of Pediatrics, Mott Children’s Hospital, University of Michigan, Ann Arbor, Michigan; bDepartment of Pediatrics, Ann & Robert H. Lurie Children’s Hospital, Chicago, Illinois; cDepartment of Cardiothoracic Surgery, Children’s Medical Center, University of Texas, Dallas, Texas; dDepartment of Surgery, University of Alabama at Birmingham, Birmingham, Alabama; eDepartment of Pediatrics, Monroe Carell Jr. Children’s Hospital at Vanderbilt, Nashville, Tennessee; f Department of Pediatrics, Duke Children’s Hospital, Durham, North Carolina; gDepartment of Pediatrics, The Children’s Hospital at Montefiore, New York, New York; and the hDepartment of Pediatrics, Phoenix Children’s Hospital, Phoenix, Arizona.
KEYWORDS: Fontan; protein-losing enteropathy; pediatric; heart transplantation; waiting list
BACKGROUND: Post-Fontan protein-losing enteropathy (PLE) is associated with significant morbidity and mortality. Although heart transplantation (HTx) can be curative, PLE may increase the risk of morbidity before and after HTx. This study analyzed the influence of PLE influence on waiting list and post-HTx outcomes in a pediatric cohort. METHODS: Fontan patients listed for HTx and enrolled in the Pediatric Heart Transplant Study from 1999 to 2012 were stratified by a diagnosis of PLE, and the association of PLE with waiting list and post-HTx mortality, rejection, and infection was analyzed. RESULTS: Compared with non-PLE Fontan patients (n ¼ 260), PLE patients listed for HTx (n ¼ 96) were older (11.9 years vs 7.6 years; p ¼ 0.003), had a larger body surface area (1.1 m2 vs 0.9 m2; p ¼ 0.0001), had lower serum bilirubin (0.5 vs 0.9 mg/dl; p ¼ 0.01), lower B-type natriuretic peptide (59 vs 227 pg/ml; p ¼ 0.006), and were less likely to be on a ventilator (3% vs 13%; p ¼ 0.006). PLE patients had lower waiting list mortality than non-PLE Fontan patients (p o 0.0001). There were no intergroup differences for post-HTx survival or times to the first infection or rejection. PLE was not independently associated with increased post-HTx mortality at any time point. CONCLUSIONS: In this multicenter cohort, the diagnosis of PLE alone was not associated with increased waiting list mortality or post-HTx morbidity or mortality. Given the limitations of our data, this analysis suggests that PLE patients in the pediatric age group have outcomes similar to their nonPLE counterparts. Additional multicenter studies of PLE patients with targeted collection of PLEspecific information will be necessary to fully delineate the risks conferred by PLE for HTx. J Heart Lung Transplant ]]]];]:]]]–]]] r 2015 International Society for Heart and Lung Transplantation. All rights reserved.
Reprint requests: Kurt R. Schumacher, MD, MS, Congenital Heart Center, C.S. Mott Children’s Hospital, 1540 E Hospital Dr, Ann Arbor, MI 48109. Telephone: 734-615-2369. Fax: 734-935-9470. E-mail address: [email protected]
Staged palliation to Fontan circulation is the current standard of care for patients with single-ventricle physiology. Unfortunately, severe and life-threatening complications may occur after the Fontan palliation, including protein-losing
1053-2498/$ - see front matter r 2015 International Society for Heart and Lung Transplantation. All rights reserved. http://dx.doi.org/10.1016/j.healun.2015.03.022
2
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
enteropathy (PLE). PLE is characterized by the abnormal loss of protein into the enteral lumen, which results in hypoproteinemia and hypoalbuminemia,1 and although it has been reported in 1% to 11% of Fontan patients,1,2 the true prevalence is unknown. The consequences of enteral protein loss are numerous, including loss of intravascular oncotic pressure leading to diffuse edema, disturbed calcium regulation, enhanced coagulation, quantitative immune abnormalities with probable immune dysregulation, poor bone density, and growth failure1,3; these morbidities cumulatively contribute to marked morbidity and mortality in this patient group.4 PLE is frequently characterized by a waxing and waning course punctuated by flares in protein loss and resulting symptoms. Further, the severity of protein loss and resulting sequelae may vary significantly between patients. This heterogeneity in the disease often makes evaluating the response to therapy challenging. Thus, although multiple pharmacotherapeutic options and anatomic interventions have demonstrated some efficacy in temporizing PLE symptoms, no available therapy has been shown to be consistently curative except heart transplantation (HTx).5–8 The 2006 Pediatric Heart Transplant Study (PHTS) evaluation of Fontan patients undergoing HTx found no difference in survival between PLE and non-PLE patients but only included 25 PLE patients in the cohort.8 This study was limited by a small sample size and limited duration of follow-up. More recently, reports from larger centers, including PLE patients in the current era, found worsened survival in Fontan patients with preserved ventricular systolic function, most of whom had PLE.5,9 This creates a more concerning picture for PLE patients after HTx. Compared with the 2006 PHTS Fontan study, which examined patients who received transplants from 1993 to 2001, the growth of the PHTS database over recent years allows a significantly larger cohort of modern-era patients with at least 1 year of follow-up to be analyzed and allows PLE patients to be analyzed as a separate group entirely. This analysis can now include evaluation of risk factors within the PLE patient group that were not performed in the previous PHTS Fontan study. Thus, to better understand the influence of PLE on waiting list and HTx outcomes, this study used the PHTS database to (1) compare the characteristics and pre-HTx and post-HTx outcomes of Fontan patients with and without PLE, and (2) determine unique pre-HTx risk factors for pre-HTx and post-HTx outcomes in PLE patients. We hypothesized that Fontan patients with PLE possess disease-specific characteristics that increase their risk of poor outcome after HTx.
Methods Patient population and data collection This retrospective cohort study analyzed data from 3,668 patients undergoing HTx between January 1, 1999, and December 31, 2012, at 35 participating PHTS institutions. Demographic, clinical, and event data were collected for all HTx recipients from each participating center. As has been previously described, the data were collected using coded event forms and sent to the data
analysis center at the University of Alabama in Birmingham, where the information was entered into a secure database, verified, and corrected as needed.10 Institutional Review Boards from participating centers approved all studies. The study excluded patients who received multiple solid-organ transplants. Follow-up on all patients was complete through December 31, 2012.
Outcome measures The a priori primary outcome measure was death after HTx. The secondary outcomes studied were time to the first infection and time to the first rejection episode. The PHTS manual defines the outcome “rejection” as a clinical circumstance (biopsy or otherwise driven) leading to an increase in immunotherapy and the outcome “infection” as evidence of an infectious process requiring intravenous therapy or a life-threatening infection requiring oral therapy.11
Statistical analysis We examined data using standard descriptive statistics, including mean and standard deviation, median and interquartile range (IQR), or number and percentage of the total, as appropriate, based on the data’s distribution. Continuous variable distributions were evaluated for equality of variance, and t-tests were used for continuous variable comparisons because they are robust in symmetric distributions. From 3,686 pediatric patients entered into the PHTS undergoing HTx between 1999 and 2012, 356 Fontan patients were stratified into cohorts by the presence or absence of PLE as reported at the time of HTx listing. PLE is reported as a “yes” or “no” check box on the PHTS data-collection forms. The PHTS form manual of operations for form completion does not provide any guidance regarding what constitutes a diagnosis of PLE; centers respond to this item at their own discretion. Similarly, the patient factor “failure to thrive” is also a “yes” or “no” check box without specific criteria set by the PHTS. Two-group comparisons for individual patient factors were performed using appropriate parametric or non-parametric tests as dictated by the statistical characteristics of the data, with t-tests being used for most comparisons of continuous variables. Competing-outcomes analyses examined the temporal influence of PLE on simultaneous risks of death or HTx in patients on the HTx waiting list. Cox proportional hazard analysis examined risk factors for waiting list death censored for HTx in the overall Fontan cohort. Kaplan-Meier time-to-event analyses with log-rank test comparisons evaluated the temporal influence of PLE on post-HTx outcomes. All study patient status was available up to the 10-year study period or until death. No minimum time to follow-up was required to be included in the post-HTx analyses, although every surviving patient had at least 1 year of post-HTx follow-up. Cox proportional hazard analysis assessed PLE as a risk factor for death within the entire study cohort. The PLE cohort was further analyzed using Cox proportional hazard analysis (the variables analyzed are listed in Tables 1 and 2) to determine independent risk factors for postHTx death and Kaplan-Meier time-to-event analyses with right censoring to evaluate the temporal influence of specific factors that could indicated more severe PLE, including intensive care unit hospitalization at HTx, inotrope dependence at HTx, or having a low (o17 kg/m2) or high (421 kg/m2) body mass index (BMI) at HTx, on death within the PLE cohort.
Schumacher et al. Table 1
Influence of Post-Fontan PLE on HTx Outcomes
3
Comparison of Characteristics of Fontan Patients With and Without Protein-Losing Enteropathy at Listing
Characteristicsa
PLE (n ¼ 96)
Non-PLE (n ¼ 260)
p-value
Age at listing, years Body surface area, m2 Panel reactive antibody, % Bilirubin, mg/dl B-type natriuretic peptide, pg/ml C-reactive protein, mg/liter Blood urea nitrogen, mg/dl Protein, g/dl Albumin, g/dl Triglycerides, mg/dl Low-density lipoprotein, mg/dl Male sex White Congenital etiology Status 1 Inotropes (high) Ventilator In intensive care unit In hospital Failure to thrive History of plastic bronchitis NYHA/Ross Class I II III IV Mean right atrial pressure, mm Hg Cardiac index, liters/min/m2 Pulmonary capillary wedge pressure, mm Hg
11.9 1.1 0.0 0.5 58.5 2.9 15.0 4.8 2.7 75.0 87.0 57 82 95 63 37 3 23 41 23 3
7.6 0.9 2.0 0.9 227.0 1.2 16.0 6.5 3.8 73.0 66.0 170 202 259 166 112 34 85 130 79 23
0.0003 0.0084 0.1643 0.0089 0.0063 0.3782 0.8694 0.3097 0.0362 0.6826 0.3578 0.2952 0.1074 0.4617 0.8222 0.4414 0.0063 0.1116 0.2217 0.234 0.042 0.3077
(5.3–14.9) (0.7–1.4) (0.0–15.5) (0.3–0.6) (23.0–163.2) (1.3–7.3) (12.0–24.0) (4.1–5.8) (2.2–3.4) (67.0–104.0) (51.0–91.0) (59.4) (85.4) (99.0) (65.6) (38.5) (3.1) (24.0) (42.7) (23.4) (9.1)
0 0 9 (52.9) 8 (47.1) 14.5 (12.5–17.0) 2.9 (2.5–3.6) 10.5 (8.0–13.0)
(4.0–12.5) (0.6–1.2) (0.0–41.0) (0.5–1.5) (43.5–1,468.0) (0.5–5.6) (13.0–21.0) (5.7–7.2) (3.3–4.3) (62.0–95.8) (48.0–78.0) (65.4) (77.7) (99.6) (64.3) (43.1) (13.1) (32.7) (50.0) (30.4) (26.14)
0 7 (12.3) 28 (49.1) 22 (38.6) 15.0 (10.5–18.0) 3.2 (2.3–3.7) 14.0 (10.0–17.0)
0.6299 0.5741 0.0005
NYHA, New York Heart Association; PLE, protein-losing enteropathy. a Continuous variables are shown as the median (interquartile range) and categoric variables as number (%).
Table 2
Comparison of Characteristics of Fontan Patients With and Without Protein-Losing Enteropathy at Heart Transplantation
Characteristicsa
PLE (n ¼ 70)
Non-PLE (n ¼ 173)
p-value
Age at transplant, years Body surface area, m2 Panel reactive antibody, % Bilirubin, mg/dl B-type natriuretic peptide, pg/ml C-reactive protein, mg/liter Blood urea nitrogen, mg/dl Protein, g/dl Albumin, g/dl Triglycerides, mg/dl Low-density lipoprotein, mg/dl Male White Status 1 Inotropes (high) Ventilator In intensive care unit In hospital Failure to thrive Induction therapy Maintenance steroids
12.2 1.1 0.0 0.5 210.0 0.5 14.0 5.3 3.0 102.5 76.0 40 62 56 37 3 17 36 17 54 18
8.7 0.9 2.0 0.9 751.0 1.6 16.0 6.7 4.0 99.0 69.5 108 131 141 93 18 61 99 54 118 55
0.0041 0.0189 0.7189 0.0682 0.0788 0.4910 0.5755 0.5196 0.2555 0.5479 0.6676 0.4445 0.0248 0.6541 0.8986 0.1242 0.0970 0.4102 0.2822 0.2046 0.9660
(6.1–15.4) (0.8–1.4) (0.0–38.0) (0.3–1.5) (54.0–317.0) (0.5–65.1) (9.0–19.0) (4.4–6.7) (2.4–4.0) (54.0–151.0) (36.0–87.0) (57.1) (88.6) (80.0) (52.9) (4.3) (24.3) (51.4) (24.3) (77.1) (48.7)
(4.9–12.9) (0.7–1.3) (0.0–46.0) (0.5–1.9) (156.0–1,987.0) (0.6–7.7) (10.0–23.0) (5.9–7.5) (3.4–4.4) (62.0–112.0) (54.0–90.0) (62.4) (75.7) (82.5) (53.8) (10.4) (35.3) (57.23) (31.2) (69.0) (48.3)
PLE, protein-losing enteropathy. a Continuous variables are shown as the median (interquartile range) and categoric variables as number (%).
4
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
Results During the study period, 96 Fontan patients with PLE and 260 Fontan patients without PLE were listed for HTx. Complete demographic and clinical comparisons between the two groups at the time of listing are detailed in Table 1. At listing, compared with non-PLE Fontan patients, the PLE patients were older, had a higher BSA, lower serum bilirubin, lower B-type natriuretic peptide, lower albumin, had lower pulmonary capillary wedge pressures, were less likely to have plastic bronchitis, and were less likely to require mechanical ventilation (all p o 0.05). Of the patients on the waiting list, 70 PLE patients and 173 non-PLE Fontan patients underwent HTx. Complete patient characteristics and comparisons at the time of HTx for each group are reported in Table 2. Notably, compared with non-PLE Fontan patients at HTx, PLE patients were more likely to be older, white, have a higher BMI, and have lower pulmonary capillary wedge pressures. Also, the serum albumin levels at HTx between the two groups were similar. Panel reactive antibody level was similar between the groups at listing and at HTx. Competing outcomes after listing for the PLE and nonPLE cohort are demonstrated in Figure 1. Compared with the non-PLE group, PLE patients had a similar rate of transplant at 6 months after listing but had a higher survival rate without HTx and lower pre-HTx death rate (p o 0.0001). Among the entire Fontan cohort, PLE did not confer an independent hazard of death. Age at listing (inverse transformed hazard ratio [HR], 1.07; p ¼ 0.001), status 1A at time of listing (HR, 4.90; p o 0.0001), and higher pulmonary artery pressure at listing (HR, 1.22; p o 0.0001) were associated with increased risk of waiting list death. Fewer previous cardiac surgeries were associated with decreased hazard of waiting list death (HR, 0.78; p ¼ 0.02). Serum albumin levels were included in the model as a continuous variable but were not significant. Receiver operator characteristic curves further analyzed albumin level within the PLE group to determine whether a particular serum level conferred increased risk of poor waiting list outcome, but no optimal stratifying value was found.
The median post-HTx duration of follow-up (including patients who died) was 2.4 years (IQR, 0.4, 4.8 years) for non-PLE Fontan patients and 1.1 years (IQR, 0.2, 3.4 years) for PLE patients. Survival after HTx was then compared between the groups (Figure 2). No significant difference in overall postHTx survival was found between PLE and non-PLE Fontan patients; of the 70 PLE Fontan patients undergoing HTx during the study period, 22 died (31%) compared with 40 of 173 non-PLE Fontan patients (23%, p ¼ 0.12). PLE was not a significant risk factor in an overall hazard analysis of the Fontan cohort (p ¼ 0.32). Similarly, no intergroup differences were found for time to the first rejection (p ¼ 0.85; Figure 3) or for time to the first infection (p ¼ 0.47; Figure 4). The patient factors used in the previous comparisons were included in a Cox proportional hazard analysis within the PLE-only cohort to determine PLE-specific risk factors for non-survival after HTx listing censored at transplant and for non-survival after HTx . Only a history of cerebrovascular accident was associated with an increased hazard of death (HR, 3.18; 95% confidence interval, 1.02–9.96; p ¼ 0.0468). Even after stratifying the PLE group by variables that could indicate a more severe presentation of PLE at the time of HTx, no additional PLE-specific risk factors for poor outcome were detected; that is, no association with worse survival was detected for being in an intensive care unit (p ¼ 0.63), on inotropes at HTx (p ¼ 0.72), or having a markedly high or low BMI (compared with normal, p ¼ 0.19). In addition, no difference was found in post-HTx hospital length of stay between PLE and non-PLE patients (39.2 vs 39.9 days, respectively; p ¼ 0.95).
Discussion Despite significant differences in the characteristics between PLE and non-PLE Fontan patients awaiting transplant, a preHTx diagnosis of PLE was not associated with an increased risk of death on the waiting list or after HTx. Furthermore, PLE had no association with the frequency of post-HTx
Figure 1 Competing outcomes of alive, received a transplant, or died for (A) protein-losing enteropathy (PLE) Fontan patients (n ¼ 96) and (B) non-PLE Fontan patients (n ¼ 260).
Schumacher et al.
Influence of Post-Fontan PLE on HTx Outcomes
Figure 2
Time to death stratified by protein-losing enteropathy (PLE) history.
rejection or infection. These results are helpful in evaluating the feasibility of HTx in Fontan patients with PLE that already carry the increased risk of poor outcome after HTx associated with the Fontan procedure.8,12 Both groups had nearly 20% mortality by 1 year post-HTx, which reinforces the high risk assumed by a center that performs a transplant in a Fontan patient. Importantly, based on this analysis, when a patient’s overall risk of mortality after HTx is being evaluated, a diagnosis of PLE in isolation should not be considered factor that increases risk; given that the heterogeneity of PLE as a disease was not reflected in this analysis, whether severely affected individuals are at increased risk remains a question that warrants further exploration.
Figure 3
5
Given the associations between Fontan patients and worsened outcomes after HTx,8 this information is reassuring. These results corroborate the previous PHTS findings that PLE was not a factor for poor outcome in a small cohort of PLE patients.8 However, the current results differ from those of 2 recent single-center reports comparing outcomes among Fontan patients listed for transplant with preserved ejection fraction (PEF), many of whom had PLE. In reporting the St. Louis Children’s experience, Simpson et al5 found significantly higher rates of infection, death from infection, graft failure, and a non-significantly higher death rate in the PEF group (59% of whom had PLE) compared with Fontan patients with impaired ejection
Time to first rejection stratified by protein-losing enteropathy (PLE) history.
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
6
Figure 4
Time to first infection stratified by protein-losing enteropathy (PLE) history.
fraction (IEF). Griffiths et al9 reported the Boston Children’s experience similarly comparing outcomes of Fontan patients referred for HTx with PEF (56% PLE) vs IEF and found increased overall mortality (with or without HTx listing) in the PEF group. Importantly, however, a significant proportion of the PEF group in their cohort was deemed “too sick” for HTx and thus were never placed on the waiting list. Each of these patients subsequently died, whereas only 1 patient with PLE died after HTx. In reporting the experience at Lurie Children’s Hospital, Backer et al12 performed transplants in 22 Fontan patients with PLE; 9 patients had PEF and 13 had IEF. No survival difference was seen between the groups (22% mortality in the PEF vs 23% in the IEF group). These apparent discrepancies warrant further discussion. It is important to draw a distinction between all patients with PLE, as studied in the Griffiths report, and the PLE patients in our study cohort that were deemed suitable for HTx listing. Because some centers participating in the PHTS did not list sick patients with PLE, they were not captured for this study, thus creating an unintended selection bias to the conclusion that PLE is not a risk factor for HTx. Alternatively, if PLE was a significant factor in the decision to not list these patients, based on our data, the use of PLE status in the decision may have been unfounded. Actually, the cohorts in each of these studies are different, and we may not fully understand the hemodynamic and physiologic mechanisms that results in PLE with “normal hemodynamics.” Our study attempted to detect a “sicker” population of PLE patients within the cohort by stratifying the PLE patients by intensive care unit requirement, inotrope use, and BMI, and we were unable to demonstrate increased risk of death in this cohort. Whether a sub-set of higher-risk PLE patients does exist, what their specific characteristics are, and whether they would have poorer HTx outcomes as a result of PLE remain unanswered
because this PLE-specific information is not collected in the PHTS data set. Moreover, a plausible reason explaining why the PHTS analysis did not mirror the Simpson findings that Fontan HTx recipients experience significantly increased incidences of infection and graft failure5 is similarly lacking. Perhaps individual HTx centers may have different acceptable-risk thresholds, and some PLE patients deemed “too sick to list” at one center would be listed at another center and suffer worse outcomes. The Boston and St. Louis experiences both compared PEF vs IEF patients, with most of the PEF patients having concomitant PLE. The Lurie Children’s experience actually stratified the PLE group by ventricular function and found no difference in outcome. Given this, it is also possible that the worse outcomes seen in PEF Fontan-failure patients are driven by poor outcomes in PEF patients without PLE and that PLE patients should not necessarily be included in the same cohort with non-PLE PEF Fontan patients when examining outcomes. Although no statistically significant difference was detected for overall survival between the groups, the p-value of 0.12 should not be completely dismissed. Despite being a large study cohort by typical PLE-study standards, the number of patients studied may have nevertheless resulted in an underpowered study. Finally, patients aged Z 18 years are not included in the PHTS database. This group may consist mainly of patients who have suffered from a longer duration of PLE and who are more severely affected by PLE sequelae that do confer an increase risk at HTx. Additional multicenter studies of PLE patients with targeted collection of PLE-specific information are clearly necessary to fully delineate the spectrum and risk of death conferred by this disease. Data regarding the duration of PLE before listing, the severity of PLE, including need for albumin infusions and hospitalizations for fluid overload and hemodynamic compromise,
Schumacher et al.
Influence of Post-Fontan PLE on HTx Outcomes
and the intensity and duration of PLE-specific therapies would be useful in fully characterizing the chronicity, severity, and overall morbidity conferred on an individual by PLE. This necessity is further reinforced when evaluating the overall risk profile of both the non-PLE and PLE Fontan patients in this study. Although not reaching statistical significance, the nonPLE Fontan patients tended to have more traditional postHTx risk factors for non-survival at the time of transplant, including a higher rate of ventilation, higher average B-type natriuretic peptide indicating worse heart failure, and a higher rate of failure to thrive.13,14 Given this trend, the lack of difference in overall survival is somewhat surprising and further suggests a need for analysis including PLE-specific information. Further, serum albumin levels are actually not different between the cohorts at the time of HTx. It is possible that improvement in PLE and its accompanying symptoms explains the equalization of albumin levels. However, it is equally plausible that this represents PLE patients who are being very aggressively managed with albumin replacement therapy, because this frequently occurs at centers to optimize patients before HTx. Without PLE-specific data, these findings are unable to be adequately characterized. The lack of difference in secondary outcome incidence also warrants brief discussion. Reports of apparent immune dysfunction in PLE Fontan patients have described distinct abnormalities in immunoglobulins and lymphocyte populations.15–17 The clinical implications of these findings remain unclear but give cause for concern regarding the potential for rejection and infection in PLE patients after HTx. In the PHTS cohort, there was no difference in the time to the first rejection or infection between PLE and non-PLE patients. Thus, although immunologic abnormalities have been shown in PLE patients, they either resolve after transplant or do not confer clinically relevant and identifiable immune problems (i.e., infection or rejection). Finally, the question remains to what degree HTx cures or ameliorates PLE. Several case series have reported highrates of PLE resolution after HTx.6,7,12 This study cannot directly address the question of PLE resolution because the presence of PLE after HTx is not collected by the PHTS. This study has several limitations that require brief discussion. This is a retrospective, registry-based study of a broad cohort of patients listed for HTx; the nature of retrospective, registry-based research may decrease our ability to discern some specific relationships. Second, as described previously, some specific variables of interest to PLE are not collected in the PHTS database, and PLE itself is only defined as a “yes” or “no” data field without further definition, which may allow significant heterogeneity within the group and potential centerdependent misclassification bias. Third, we are only able to comment on patients listed for HTx entered into the PHTS database and not on all patients with PLE in whom HTx is considered or Fontan patients aged 4 18 years (adults are not included in the PHTS). This potentially causes selection bias in the cohort and excludes from analysis the “sickest” PLE patients who may have been deemed ineligible for HTx.
7
Finally, despite using the entire PHTS database, our patient numbers are still relatively limited due to the overall low incidence Fontan patients with and without PLE requiring transplantation. This is particularly notable regarding late outcomes. Few patients, particularly in the PLE group, have follow-up of 5 years or greater. This lack of patients “at-risk” reduces the power of these analyses to detect differences in late outcomes and limits the study’s ability to adequately characterize the groups long-term after HTx. The potential for type II error and true differences between the groups in late-term outcome is possible and will require future study with larger, late-term cohorts. In conclusion, a diagnosis of PLE was not associated with an increased risk of waiting list mortality or post-HTx morbidity or mortality in this PHTS Fontan cohort. This analysis suggests that PLE patients have outcomes similar to their non-PLE counterparts. However, given the acknowledged lack of PLE-specific information, the potential for selection bias, and the inability to truly stratify PLE by severity within the PHTS database, the question of whether severe PLE conveys risk of post-HTx morbidity remains only partially explored. Further research to determine whether factors unique to this population, but not explored in this analysis, confer differential risk is warranted and will require substantial multicenter collaboration in the future. Given the complexity of care involved with the population, this question remains worth fully answering.
Disclosure statement None of the authors has a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript or other conflicts of interest to disclose.
References 1. Rychik J. Protein-losing enteropathy after Fontan operation. Congenit Heart Dis 2007;2:288-300. 2. Mertens L, Hagler DJ, Sauer U, Somerville J, Gewillig M. Proteinlosing enteropathy after the Fontan operation: an international multicenter study. PLE study group. J Thorac Cardiovasc Surg 1998;115: 1063-73. 3. Meadows J, Jenkins K. Protein-losing enteropathy: integrating a new disease paradigm into recommendations for prevention and treatment. Cardiol Young 2011;21:363-77. 4. Goldberg DJ, Dodds K, Rychik J. Rare problems associated with the Fontan circulation. Cardiol Young 2010;20(Suppl 3):113-9. 5. Simpson KE, Cibulka N, Lee CK, Huddleston CH, Canter CE. Failed Fontan heart transplant candidates with preserved vs impaired ventricular ejection: 2 distinct patient populations. J Heart Lung Transplant 2012;31:545-7. 6. Kanter KR, Mahle WT, Vincent RN, Berg AM, Kogon BE, Kirshbom PM. Heart transplantation in children with a Fontan procedure. Ann Thorac Surg 2011;91:823-9: [discussion 829-30]. 7. Davies RR, Sorabella RA, Yang J, Mosca RS, Chen JM, Quaegebeur JM. Outcomes after transplantation for “failed” Fontan: a singleinstitution experience. J Thorac Cardiovasc Surg 2012;143:1183-92: e4. 8. Bernstein D, Naftel D, Chin C, et al. Outcome of listing for cardiac transplantation for failed Fontan: a multi-institutional study. Circulation 2006;114:273-80. 9. Griffiths ER, Kaza AK, Wyler von Ballmoos MC, et al. Evaluating failing Fontans for heart transplantation: predictors of death. Ann Thorac Surg 2009;88:558-63: [discussion 563-4].
8
The Journal of Heart and Lung Transplantation, Vol ], No ], Month ]]]]
10. Shaddy RE, Naftel DC, Kirklin JK, et al. Outcome of cardiac transplantation in children. Survival in a contemporary multiinstitutional experience. Pediatric Heart Transplant Study. Circulation 1996;94:II69-73. 11. Pediatric Heart Transplant Study. PHTS manual of operations. 2010. http://www.uab.edu/medicine/phts/images/2010_Forms/PHTS_Manual_ of_Operations_2.18.10.pdf. Accessed February 1, 2014. 12. Backer CL, Russell HM, Pahl E, et al. Heart transplantation for the failing Fontan. Ann Thorac Surg 2013;96:1413-9. 13. Almond CS, Gauvreau K, Canter CE, Rajagopal SK, Piercey GE, Singh TP. A risk-prediction model for in-hospital mortality after heart transplantation in US children. Am J Transplant 2012;12:1240-8.
14. Godown J, Donohue JE, Yu S, Friedland-Little JM, Gajarski RJ, Schumacher KR. Differential effect of body mass index on pediatric heart transplant outcomes based on diagnosis. Pediatr Transplant 2014;18:771-6. 15. Garty BZ. Deficiency of CD4þ lymphocytes due to intestinal loss after Fontan procedure. Eur J Pediatr 2001;160:58-9. 16. Chakrabarti S, Keeton BR, Salmon AP, Vettukattil JJ. Acquired combined immunodeficiency associated with protein losing enteropathy complicating Fontan operation. Heart 2003;89:1130-1. 17. Lenz D, Hambsch J, Schneider P, Tarnok A. Protein-losing enteropathy after Fontan surgery: is assessment of risk patients with immunological data possible? Cytometry B Clin Cytom 2003;53:34-9.