| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2006;82:695-700
© 2006 The Society of Thoracic Surgeons
a The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania
b The Burnham Institute, La Jolla, California
Accepted for publication February 22, 2006.
* Address correspondence to Dr Rychik, Division of Cardiology, Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadelphia, PA 19104 (Email: rychik{at}email.chop.edu).
| Pediatric cardiac surgery:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
METHODS: A cross-sectional analysis of 62 patients after the Fontan operation was performed. Twenty-four hour stool sample was collected for alpha-1-antitrypsin (A1AT) clearance, to determine the presence of abnormal enteric protein loss (AEPL) defined as either an abnormal fecal A1AT clearance of greater than 27 mL/24 hours, or an abnormal fecal A1AT concentration of greater than 54 mg/dL. Subjects underwent ultrasonography of the mesenteric and celiac artery blood flow and blood draw for tumor necrosis factor-alpha (TNF-a), high sensitivity C reactive protein (CRP), brain natriuretic peptide (BNP), angiotensin II, coagulation factors protein S, protein C, and antithrombin III (AT III), and serum transferrin for determination of glycosylation defect.
RESULTS: Age at study was 10.9 ± 3.4 years; 8.6 ± 3.9 years after the Fontan operation. Seven subjects had AEPL. Mesenteric-to-celiac artery flow ratio was lower for the AEPL group, than for the non-AEPL group (p < 0.05). The TNF-a, CRP, BNP, and angiotensin II levels were elevated; however, there was no correlation with AEPL. Abnormalities in coagulation factors were present but did not correlate with AEPL. No glycosylation defects were identified.
CONCLUSIONS: Potential candidate mechanisms for elucidation of the pathophysiology of PLE include abnormal mesenteric vascular resistance and inflammation, conditions uniquely present after the Fontan operation. Targeted investigations of these parameters may provide clues as to the mechanism of onset of PLE after Fontan operation.
Protein-losing enteropathy (PLE) is an enigmatic ailment seen after the Fontan operation. The exact prevalence is unknown; however, studies suggest its presence in 3% to 15% of patients after the Fontan operation [1, 2]. The disease can lead to severe hypoproteinemia with loss of vascular oncotic pressure and development of edema, ascites, and pleural-pericardial effusions. Mortality is high, with 50% survival at five years from onset of diagnosis [1, 2].
The pathophysiologic mechanism of PLE after the Fontan operation is uncertain. A direct association with elevated systemic venous pressures has not been found [2]. Plausible hypotheses have been put forth based upon the successes of various treatment strategies. First, resolution of PLE occurs after interventions that improve cardiac output such as creation of fenestration [3], pacing [4], and heart transplantation [5]. Overall cardiac output is diminished after the Fontan operation [6], and even poorer in patients with PLE [2]. We have hypothesized that chronically diminished cardiac output may alter the systemic distribution of blood flow, shifting blood volume away from the mesenteric circulation toward more vital organs. Our previous work [7] demonstrated that Doppler-derived measures of mesenteric vascular resistance are markedly elevated after the Fontan operation. Alterations in mesenteric arterial flow, in addition to abnormalities of hepatic venous drainage, may lay the groundwork for impaired mesenteric perfusion, resulting in a break in the integrity of the intestinal mucosa and subsequent protein leakage. Second, improvement in PLE can also occur with the administration of systemic corticosteroids [8] or unfractionated heparin [9], agents that act to stabilize cell membranes and inhibit inflammation. It is conceivable that the chronic low cardiac output state in patients after the Fontan operation triggers release of inflammatory markers and neurohormones, which may contribute to the development of PLE. Third, congenital disorders of glycosylation can lead to development of PLE through a mechanism of enterocyte heparan sulfate reduction [10]. Whether such defects in glycosylation exist in patients after the Fontan operation is currently unknown.
The purpose of this study is to investigate the association between PLE after the Fontan operation and possible candidate mechanisms for development of the disease by searching for abnormalities of the following: (1) mesenteric blood flow; (2) systemic inflammation; (3) neurohormonal activation; and (4) protein glycosylation. Our goal is to identify potential candidate parameters for further, more focused investigation, which might better elucidate the pathophysiology of this disorder and ultimately lead to more effective treatment strategies.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Study Protocol
Patients were mailed collection receptacles and instructions for 24 hours of stool to be collected prior to the day of hospital visit. At hospital visit patients (parents) were asked to answer a brief questionnaire on the presence or absence of symptoms relating to PLE, including chronic diarrhea, chronic abdominal pain, or swelling in the abdomen or extremities. A limited physical examination was performed looking for signs of abdominal or pre-tibial edema. Subjects were restricted from food or drink for a minimum of 4 hours prior to ultrasound study; this in order to obtain baseline flow measures unaffected by the stimulus of food. While lying supine, imaging of the celiac artery (CA) and superior mesenteric artery (SMA) was performed using a Phillips Sonos 5500 ultrasonograph (Phillips Medical Systems, Andover, MA) as previously described [7]. Two-dimensional measurements were made of the diameter of the origins of the CA and SMA. Ten subjects had measurements of the CA and SMA diameters repeated by a single observer (JR) while blinded to the original values, with intraobserver variability of less than 10% in all. A pulse-wave Doppler sample was placed at the origin of each vessel and the waveforms recorded. The CA and SMA blood flow was calculated by the following formula: flow = heart rate x velocity time integral x 3.14 x radius of the vessel squared. Flow values were indexed to body surface area. The ratio of SMA-to-CA flow was obtained as an index of relative downstream vascular resistance of the SMA [11]. Resistance index (RI) of Pourcelot was calculated for the SMA by the formula: RI = (peak systolic velocity – end diastolic velocity)/ peak systolic velocity. An RI = 1 reflects no diastolic flow and is the highest level of resistance.
On the same day as ultrasonography, phlebotomy was performed and the following tests obtained: (1) serum alpha-1-antitrypsin (A1AT); (2) serum albumin; (3) liver function; (4) tumor necrosis factor-alpha (TNF-a); (5) high sensitivity C reactive protein (CRP); (6) brain natriuretic peptide (BNP); (7) angiotensin II; (8) serum transferrin for determination of glycosylation defect; and (9) coagulation factors protein S, protein C, and antithrombin III (AT III). The 24-hour stool collection and serum sample for A1AT were used to determine the fecal A1AT stool clearance as well as fecal A1AT concentration.
Patients were divided into two groups based on the presence or absence of abnormal enteric protein loss (AEPL), defined as either an abnormal fecal A1AT clearance of greater than 27 mL/24 hours [12], or an abnormal fecal A1AT concentration of greater than 54 mg/dL [13]. Normal values for TNF-a are based on manufacturer test insert data, with a level greater than 8.2 pg/mL defined as abnormally elevated. Normal values for BNP are age dependant, with a level greater than 30 pg/mL recently defined as abnormal for school age children [14]. Normal values for SMA RI range from 0.75 to 0.9 [15]. As previously published from our laboratory, SMA RI values in normal school age children age 7.2 ± 4.7 years are 0.74 ± 0.08 [7].
Patients with rare congenital disorders of glycosylation (CDG) have defects in N-linked oligosaccharide synthesis, leading to multisystemic pathologies including coagulopathy and psychomotor retardation [16]. Some patients experience bouts of PLE, especially during infections [10]. The CDG patients have abnormally glycosylated transferrin [16], and we used three different methods to determine whether post-Fontan patients have impaired glycosylation, including isoelectric focusing [10], ion-exchange high-performance liquid chromatography (HPLC) analysis [17], and electrospray ionization mass spectrometry (ESI-MS) of immunoaffinity purified material [18]. For ESI-MS, 100 uL serum was mixed with beads containing affinity purified antitransferrin. Bound transferring was eluted with 0.1M glycine buffer at pH 2.8, desalted by C18-ZipTip (Millipore Corp, Bedford, MA), and concentrated in 3 to 5 uL 5% acetic acid in methanol to acetonitrile to water (50:25:25). Analysis was done by nanospray-MS using an API 3000 mass spectrometer (PE-SCIEX, Toronto, Canada), operated in the positive ion mode at 700 V.
Comparisons were made between the values of patients with and those without AEPL by a nonparametric t test. The Fisher exact test was used to compare categoric variables. Laboratory values were tested for associated relationship by Pearson correlation coefficient. Data are reported as mean and SD. A p value of less than 0.05 is significant.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Age at time of study was 10.9 ± 3.4 years (range, 3.3 to 18.5). Average time from Fontan operation was 8.6 ± 3.9 years. There was no difference in age at study, or time from Fontan operation, for patients with or without AEPL (Table 1). Predominant ventricular morphology consisted of right ventricle in 41 (66%), left ventricle in 19 (31%), and indeterminate in 2 (3%). There was no difference in distribution of ventricular morphology between AEPL and non-AEPL groups. All patients had a lateral tunnel type of Fontan operation. Forty-four patients (71%) were taking angiotensin-converting enzyme (ACE) inhibitors and 9 patients (15%) were taking warfarin medication at the time of study.
|
Inflammatory Markers
The TNF-a values ranged from 4 to 46.3 pg/mL, with an average of 8.3 ± 6.4 pg/mL. Twenty-three patients (37%) had abnormally elevated TNF-a levels (>8.2 pg/mL); 4 of 7 (57%) with AEPL and 19 of 55 (35%) without AEPL (p = 0.4). There was no difference in mean TNF-a levels between patients with or without AEPL. The CRP levels ranged from 0.2 to 28.5 mg/L, with an average of 2.2 ± 5 mg/L. The CRP was increased to a level greater than 3 mg/L in 16%, greater than 1 mg/L in 35%, and greater than 0.3 mg/L in 70% of the study population. C reactive protein levels trended higher in the AEPL group; however, this did not reach statistical significance. There was no significant relationship between TNF-a or CRP and the amount of SMA flow, SMA resistance index, or ratio of SMA-to-CA flow.
Neurohormonal Levels
The BNP values ranged from 10 to 94 pg/mL, with an average of 22 ± 19 pg/mL. Thirteen patients (21%) had abnormally elevated BNP values (>30 pg/mL). There was no difference in BNP values between the patients with and without AEPL; however, the highest value of 94 pg/mL was in a patient with active PLE. Angiotensin II levels ranged widely from 1 to 604 pg/mL, with an average of 52 ± 101 pg/mL. There was no difference in angiotensin II levels between patients with and without AEPL. There was no significant relationship found between levels of neurohormonal agents and inflammatory markers or mesenteric blood flow indices. Patients taking ACE inhibitors had significantly lower levels of angiotensin II (15 ± 12 pg/mL vs 64 ± 115 pg/mL; p < 0.001), lower SMA RI (0.94 ± 0.07 vs 0.98 ± 0.04; p = 0.01), and greater SMA-to-CA blood flow ratio (0.35 ± 0.19 vs 0.25 ± 0.09; p < 0.01) than patients not taking ACE inhibitors.
Liver Function Tests and Coagulation Factors
All liver function tests were within normal limits or just mildly elevated for all patients. There was no difference in liver function test values between patients with or without AEPL. Protein S was significantly lower and protein C significantly higher in patients with AEPL than in those without AEPL. There was no difference in antithrombin III levels between the groups.
Glycosylation Defect
Sera from age-matched controls and patients with known types of CDG were analyzed by isoelectric focusing, ion-exchange HPLC, and ESI-MS. None of the post-Fontan patients showed pathologic patterns; all were within the normal range.
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Cardiac Output and Mesenteric Flow After Fontan Operation
We propose a mechanism for the development of AEPL that derives from the principle that patients after the Fontan operation exist in a chronic state of low cardiac output [3, 6, 19]. Similar to what is seen in acute circulatory shock, chronic low-flow states lead to a redistribution of blood flow away from nonvital organs such as the mesenteric circulation. If this is so, then abnormalities of mesenteric vascular resistance are to be expected. In a previous study, we found significant differences in SMA resistance between an age-matched normal population and patients after Fontan, with highest vascular resistance in those with PLE after Fontan [7].
In the current study, mesenteric vascular resistance index was markedly higher in both the AEPL and non-AEPL groups than in published values for subjects with normal circulation (0.75 to 0.9) [15]. In addition, we indexed SMA flow to CA flow as a more accurate measure of mesenteric flow maldistribution. In this manner, SMA flow can be indexed within each patient based on their own individual standard, as CA resistance is typically very low in the normal population but may vary in the Fontan population. We found a significantly lower ratio of mesenteric-to-celiac artery flow in the AEPL group, suggesting that diminished mesenteric flow may play a role in the disease. Diminished flow in combination with elevated venous pressures can lead to a decrease in the gut perfusion profile. Impaired mesenteric perfusion may result in modulation of intestinal cell membrane function and promote cellular apoptosis, factors which contribute to increased intestinal permeability and protein leakage [20, 21].
The Role of Inflammation
In adults, the low-flow state of chronic congestive heart failure results in stimulation of the inflammatory system [22]. This phenomenon may similarly be occurring in patients after the Fontan operation. Our study supports this notion as over one-third of our patients exhibited elevation of inflammatory markers TNF-a and CRP. Although elevation in cytokines has been demonstrated early after a Fontan operation [23], we found levels to be elevated years after surgery. Cytokines can induce vasoconstriction and impair endothelial dependant vasodilation in resistance arteries [24], hence it is possible that TNF-a may play a role in increasing mesenteric vascular resistance and contribute to a predisposition to AEPL in patients after the Fontan operation. The inflammation seen in patients with Fontan circulation may additionally result in fundamental changes in endothelial function throughout the vessels in the body. Investigators have found abnormalities in endothelial function after a Fontan operation utilizing brachial artery reactivity measures [25]. Determining if patients with AEPL after a Fontan operation have greater degrees of endothelial dysfunction than those without AEPL would be of great interest.
Tumor necrosis factor-alpha may play a direct role in altering intestinal cell membrane permeability to intravascular protein. Recently, Bode and colleagues [26] created the first in vitro model of PLE by measuring the flux of albumin through a monolayer of intestinal HT29 cells. The TNF-a treatment increased albumin flux across the layer of cells by a factor of fourfold, as TNF-a compromises epithelial barrier function by the disruption of tight junctions [27]. In additional experiments, the loss of cell basolateral heparin sulfate proteoglycan, either by heparanase digestion or inhibition of heparin synthesis, resulted in a 1 
- fold increase in albumin permeability. The combined affects of TNF-a and heparin sulfate loss resulted in a synergistic effect with a sevenfold increase in albumin permeability. These in vitro findings offer an explanation as to why treatment with exogenous heparin may diminish the degree of protein loss in some patients [9].
Steroid treatment has been successfully used to manage AEPL after a Fontan operation, strongly suggesting an inflammatory component to the disease [8]. However, our study failed to demonstrate an association between the presence of inflammation and AEPL; hence the link between inflammation and AEPL is still elusive. This may be due to a number of factors. First, the number of subjects with AEPL enrolled in our study was relatively small. Second, it is conceivable that many of the serum markers for inflammation are lost in the stool in patients with AEPL, which may spuriously lower the values we obtained. Third, a rise in inflammatory markers may occur intermittently and sporadically after a Fontan operation, as the time course and natural history of this phenomenon is not yet well-defined. As our study design was cross-sectional and not longitudinal, it is not known if serial evaluation in the AEPL subjects would reveal abnormalities of inflammation or if those with elevated markers for inflammation would at some future point manifest AEPL.
Congenital Disorders of Glycosylation
Impaired protein glycosylation may contribute to PLE [16, 26, 28, 29]. Patients with a deficiency in phosphomannose isomerase had periodic PLE, which resolved when impaired glycosylation was corrected by daily supplements of mannose that alleviated the metabolic block [29]. Some patients with other types of CDG also have PLE [29]. Abnormal transferrin glycosylation is a useful marker in CDG patients, which we examined in our cohort. In all cases, patterns appeared within the normal range. These findings indicate that none of the patients had major N-glycosylation deficiencies comparable with CDG patients, but it does not eliminate the possibility that organ specific or localized glycosylation abnormalities may exist. Loss of heparan sulfate proteoglycan from the basolateral surface of intestinal epithelial cells correlates with PLE in CDG patients [10], which implicate a role for glycosaminoglycan chains, in agreement with recent in vitro studies [26].
Selective Nature of AEPL and the "Response-to-Injury" Model
Why AEPL affects a select group of patients and not others remains puzzling. The AEPL can be seen in patients with "failing Fontan" and abnormal hemodynamics, namely pulmonary artery pressures greater than 20 mm Hg; however, it can also commonly be seen in patients with acceptably low pulmonary artery pressures of less than 15 mm Hg [2]. Although our study is limited by the absence of temporally related cardiac catheterization data such as pulmonary artery pressure, all of our subjects were outpatients and functional status NYHA class II or better. Our study criteria specifically excluded the most severely dysfunctional patients in order to eliminate confounding hemodynamic factors that might influence the parameters we intended to measure. It is safe to conclude that elevated pulmonary artery pressure alone is not the sole determinant of AEPL after a Fontan operation.
It is conceivable that AEPL follows a "response-to-injury" model in that subjects exhibit varying thresholds in their response to the stressor injuries of low cardiac output, impaired mesenteric flow, and inflammation imposed by the Fontan circulation. In many, the "response" of AEPL occurs at markedly abnormal post-Fontan hemodynamics (pulmonary artery pressures > 20 mm Hg), while in some others it occurs at relatively typical post-Fontan hemodynamics (pulmonary artery pressures < 15 mm Hg). An inherent, host predisposition may determine the presence of the enteric protein-loss response to the injury of post-Fontan physiology.
Such host specific variability may be evident in other ways in the single ventricle population. Data suggest that abnormalities of coagulation may be variably inherent in single ventricle patients and is not necessarily acquired in relation to protein loss. Coagulation abnormalities, in fact increased or decreased factors, can be noted prior to a Fontan operation [30, 31]. In our study, abnormalities were most significant in the AEPL patients, but not necessarily that of a diminished factor, as our subjects with AEPL had a significantly increased protein C level compared with those without AEPL. These findings further support the notion that coagulation abnormalities are not related to loss at the enteric level. Nonetheless, patients with AEPL after a Fontan operation are at higher risk for thrombosis than patients without AEPL, and should be managed accordingly with vigilant surveillance [32].
Conclusions
We have demonstrated a number of physiologic abnormalities after a Fontan operation that may either be the cause of, or occur as a consequence of, the development of AEPL. Our study does not conclusively determine the mechanism of the disease but provides important insight into further possible variables for study in what portends to be a complex pathophysiology. In our recruitment for this study, two subjects were identified as having AEPL who were previously unaware of having the disease. This suggests that there is a clinically silent component to the disease, with a much larger prevalence than previously appreciated. As the number of patients with Fontan circulation growing into adulthood continues to expand, physicians may be faced with a growing morbidity and mortality related to AEPL. Multicenter efforts will be necessary in order to recruit adequate numbers of patients for further detailed investigations. The current study points toward the concepts of impaired mesenteric flow and inflammation as possible targets for development of effective treatment strategies for AEPL after a Fontan operation.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Nayak and P.D. Booker The Fontan circulation CEACCP, February 1, 2008; 8(1): 26 - 30. [Full Text] [PDF]
|
|
|
|
 |
|
 T. Nakano, H. Kado, T. Tachibana, K. Hinokiyama, A. Shiose, M. Kajimoto, and Y. Ando Excellent Midterm Outcome of Extracardiac Conduit Total Cavopulmonary Connection: Results of 126 Cases Ann. Thorac. Surg., November 1, 2007; 84(5): 1619 - 1626. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Uzun, J. K. Wong, V. Bhole, and O. Stumper Resolution of Protein-Losing Enteropathy and Normalization of Mesenteric Doppler Flow With Sildenafil After Fontan Ann. Thorac. Surg., December 1, 2006; 82(6): e39 - e40. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. K. Heinemann Invited commentary Ann. Thorac. Surg., August 1, 2006; 82(2): 700 - 701. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
