- Email: [email protected]
Liver Disease in Patients After the Fontan Operation
Liver Disease in Patients After the Fontan Operation Krishna Pundi, BSa, Kavitha N. Pundi, MDb, Patrick S. Kamath, MDc, Frank Cetta, MDb,d, Zhuo Li, BSe, Joseph T. Poterucha, MDb, David J. Driscoll, MDb, and Jonathan N. Johnson, MDb,d,* We reviewed records of all patients with an initial Fontan operation or revision from 1973 to 2012 at our institution (n [ 1,138); 195 patients had postoperative liver data available. Cirrhosis was identified by histopathology or characteristic findings on imaging with an associated diagnosis of cirrhosis by a hepatologist. Of 195 patients with biopsy or imaging, 10-, 20-, and 30-year freedom from cirrhosis was 99%, 94%, and 57%, respectively. There were 40 of 195 patients (21%) diagnosed with cirrhosis (mean age at Fontan 10.7 – 8 years). On multivariate analysis, hypoplastic left heart syndrome was associated with increased risk of cirrhosis (n [ 2 of 16, p [ 0.0133), whereas preoperative sinus rhythm was protective (p [ 0.009). Survival after diagnosis of cirrhosis was 57% and 35%, at 1, and 5 years, respectively. The cause of death was known for 9 patients (5 multiorgan failure, 2 liver failure, and 2 heart failure). In conclusion, there is an incremental occurrence of cirrhosis after the Fontan, which should be considered when designing follow-up protocols for patients after Fontan operation. Ó 2016 Elsevier Inc. All rights reserved. (Am J Cardiol 2016;117:456e460)
Since its introduction in 1968, the Fontan operation has been used for the palliation of patients with a single ventricle.1 The procedure has undergone many modifications.2e4 Numerous late complications after Fontan operation have been reported including arrhythmias, heart failure, protein losing enteropathy, and hepatic dysfunction.5 In patients after Fontan operation, increased central venous pressure and subsequent passive venous congestion have been associated with increased hepatic complications and overall mortality.6e11 However, it is unclear if any specific factors increase the risk of developing hepatic dysfunction, creating a challenge in identifying patients needing close follow-up. The hepatic dysfunction created by vascular congestion has demonstrated a level of reversibility,12,13 providing a possible role for early cardiac transplantation for susceptible patients. In this study, we examined our institution’s experience with hepatic dysfunction and cirrhosis in patients who had a Fontan operation. Methods In this institutional review boardeapproved single-center retrospective study, we reviewed the records of all patients (n ¼ 1,138) who had their initial Fontan operation or Fontan revision at the Mayo Clinic from October 1973 to June 2012. Information regarding demographic, anatomic, preoperative, operative, and postoperative variables was abstracted into a secure electronic patient database. This a Mayo Clinic College of Medicine, bDivision of Pediatric Cardiology, Department of Pediatrics, cDivision of Gastroenterology and Hepatology, Department of Medicine, dDivision of Cardiovascular Diseases, Department of Medicine, and eDivision of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota. Manuscript received July 14, 2015; revised manuscript received and accepted November 3, 2015. See page 459 for disclosure information. *Corresponding author: Tel: (507) 266-0676; fax: (507) 284-3968. E-mail address: [email protected] (J.N. Johnson).
0002-9149/15/$ - see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2015.11.014
database included information regarding preoperative and postoperative echocardiograms, cardiac catheterizations, electrocardiograms, Holter/event monitors, liver imaging, and surgical data. In addition, a medical questionnaire was mailed to all patients not known to be dead at the time of the study (n ¼ 723). Nonresponders received second and third questionnaires. If the subsequent questionnaires were not returned or completed, an attempt was made to contact the patients by telephone. The overall survey response rate was 42% (305 of 723). Data were available in the form of blood tests (aspartate aminotransferase, alanine aminotransferase, or bilirubin), liver ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), magnetic resonance elastography (MRE), or liver biopsy. Cirrhosis was diagnosed based on either (1) histopathologic diagnosis on biopsy or autopsy or (2) correlation of imaging findings (radiologic interpretation of liver cirrhosis on CT, MRI, MRE, or liver ultrasound) with clinical evaluation by a hepatologist with experience caring for Fontan patients (P.S.K). Patients were excluded from this analysis if they had evidence of viral hepatitis, alcoholic liver disease, or had a diagnosis of cirrhosis before their initial Fontan operation. Statistical analysis was performed using the date of the Fontan operation as time ¼ 0. All deaths were considered in the survival analysis, irrespective if they occurred intraoperatively, during immediate postoperative hospitalization, or after hospital discharge. Kaplan-Meier curves were derived to calculate 10-, 20-, and 30-year freedom from cirrhosis. Cox regression models were used to find univariate and multivariate predictors of being diagnosed with cirrhosis. The multivariate model considered univariately significant variables (p <0.05) with model selection using the stepwise method. All statistical tests were 2 sided with the alpha level set at 0.05 for statistical significance. SAS 9.3 was used for the analysis (SAS Institute, Cary, North Carolina). The cutoffs for discrete variables used in univariate and multivariate analyses were defined as follows: preoperative www.ajconline.org
Congenital Heart Disease/Liver Disease After Fontan
457
Table 1 Patient demographics Characteristic
All Fontan Patients (N ¼ 1,138)
Patients with liver data (N ¼ 195)
Patients with Cirrhosis (N ¼ 40)
456 (40%) 10.2 8.3
88 (45%) 9.6 9.2
16 (40%) 10.7 8
619 290 134 95
96 46 44 9
22 9 5 4
Female Age at Operation Type of Fontan operation Atriopulmonary connection Lateral tunnel Extracardiac conduit Other type of Fontan Preoperative anatomy Tricuspid Atresia Double Inlet Left Ventricle Asplenia/Heterotaxy Pulmonary Atresia/Intact Septum Hypoplastic Left Heart Syndrome Mean post-bypass left atrial pressure (mmHg) Mean post-bypass Fontan pressure (mmHg) Post-Fontan Ejection Fraction (%) Most recent Ejection Fraction (%) Most recent Systemic Ventricle End-Diastolic Pressure (mmHg) Average New York Heart Association Functional Class (1-4) Protein-Losing Enteropathy Documented arrhythmia
(54%) (25%) (12%) (9%)
310 (27%) 292 (26%) 78 (7%) 62 (5%) 27 (2%) 9.6 3.2 17.4 3.1 49.5 10.3 47.1 13.4 11 4.3 1.83 0.8 111 (9.7%) 522 (46%)
(49%) (24%) (22%)* (5%)
57 (29%) 40 (21%) 16 (8%) 16 (8%)* 16 (8%) 8.9 2.8 16.9 2.6 51 10 50.6 9.8 10.7 4.1 1.98 0.9 33 (17%)* 141 (72%)*
(55%) (22%) (13%) (10%)
13 (33%) 11 (28%) 4 (10%) 2 (5%) 2 (5%) 9.1 3 17.1 2.9 54 8.1 52 13.5 12.3 4.4 21 7 (18%) 34 (85%)*
* Denotes values that are statistically different from All Fontan Patients (p <0.05).
pulmonary artery pressures (>15 mm Hg), pulmonary arteriolar resistance (>3 U m2), ventricular end-diastolic pressure (>12 mm Hg), post-bypass Fontan pressures (>20 mm Hg), and chest tube duration (>21 days). Bypass times and post-bypass left atrial pressures were analyzed as continuous variables. The complete set of variables evaluated for their association with survival and functional status of survivors are listed in Supplementary Table 1. Results Overall, 1,138 patients had their initial Fontan or Fontan revision operation at our institution, with a mean duration of follow-up of 14.3 9.5 years (median 14.1 years). Clinical follow-up of liver function and associated imaging were available for 195 patients with an average duration of follow-up of 20.6 8.1 years (median 22 years). Demographics of this cohort are described in Table 1. In comparison with all Fontan patients, there was a higher proportion of patients with extracardiac conduits, initial lesion of pulmonary atresia with intact ventricular septum, history of arrhythmia, or history of PLE in patients with liver imaging available (p <0.05). There were no other significant differences between the liver followup and the overall Fontan cohort. In long-term followup, there were no significant differences in cardiac function in the form of ejection fraction or systemic ventricle end-diastolic pressure, or in functional class between the groups. For the 195 patients with available liver data, the mean age at Fontan operation was 9.6 9.2 years with a range of 7 months to 53 years; 40 of 195 patients (21%) had a fenestration performed at the time of the operation, with most of the fenestrations being performed after 1995. At
Figure 1. Kaplan-Meier curve of freedom from developing cirrhosis in 195 patients with liver data after Fontan. There is an incremental incidence of cirrhosis with freedom from cirrhosis of 99% (98.3-100) in 171 patients at 10 years and 94% (90-98) in 115 patients at 20 years.
most recent follow-up, 150 of 195 patients (77%) were alive and 45 (23%) were known to be dead. Liver biochemical tests were available for all 195 patients; there were no significant differences in aspartate aminotransferase, alanine aminotransferase, total bilirubin, or direct bilirubin levels between patients who did or did not develop cirrhosis; 24 of 195 patients (12%) had a liver biopsy (3 of 24 postmortem), of which 23 (96%) were abnormal; 152 of 195 patients (78%) had a liver ultrasound, of which 86 (57%) were noted to be abnormal; 25 of 195 patients (30%) had a liver MRI (72% abnormal), 33 of 195 (17%) had a CT of the liver (91% abnormal), and 33 of 195 (17%) had an MRE of the liver (79% abnormal) with 81 of 195 (42%) having at least
458
The American Journal of Cardiology (www.ajconline.org)
Table 2 Cox regression model for predictors of cirrhosis after the Fontan operation Variable Anatomic Factors Hypoplastic Left Heart Syndrome Pulmonary Artery Stenosis Asplenia Previous Surgical Procedures Bidirectional Glenn Pre-operative Factors Pre-operative Diuretic Use Pre-operative ACE-Inhibitor Use ECG Sinus Rhythm Operative Factors Fontan - Atriopulmonary Connection Fontan - Extracardiac Conduit Fenestration at Initial Operation AV Valve Replacement at Initial Operation Intra-operative Sinus Rhythm
Univariate Hazard Ratio (HR)
p-value
Multivariate Hazard Ratio (HR)
p-value
5.95 2.91 2.56
0.025 0.049 0.081
7.33 NS NS
0.0133 NS NS
4.2
0.021
NS
NS
2.36 4.11 0.33
0.013 0.007 0.014
NS NS 0.31
NS NS 0.009
0.33 4.63 3.9 4.81 0.35
0.002 0.008 0.016 0.01 0.014
NS NS NS NS NS
NS NS NS NS NS
one form of cross-sectional liver imaging. For the 195 patients with available liver data, the overall 10-, 20-, and 30year freedom from cirrhosis was 99%, 94%, and 57% respectively (Figure 1). A total of 40 patients were diagnosed with cirrhosis by pathology or correlation of imaging findings with clinical evaluation by a gastroenterologist. An established clinical diagnosis of cirrhosis by a gastroenterologist was made in 37 of 40 patients on long-term follow-up; the remaining 3 patients had biopsy-proved cirrhosis. Of these 37 patients, the most common clinical characteristics were varices (43%), ascites (35%), and thrombocytopenia (30%); only 1 patient each had hepatorenal syndrome and hepatic encephalopathy. Of the patients with cirrhosis, 10 of 40 had evidence of cirrhosis by pathology, 13 of 40 had evidence of cirrhosis on CT imaging, 8 of 40 on MRI, and 6 of 40 on MRE. Only 6 of 40 patients did not have CT, MRI, or MRE, but all these patients had an abnormal liver ultrasound along with a clinical diagnosis of cirrhosis by the hepatologist. Overall, 31 of 40 patients (78%) with cirrhosis had available ultrasound data and 28 of 31 (90%) were abnormal. MREderived mean liver stiffness data were available for 7 of 40 patients with an average of 6.0 1.3 kPa (>5 kPa interpreted as stage 4 fibrosis or cirrhosis based on institutional grading system derived for adult patients with other etiologies for liver disease). For the 40 patients with cirrhosis, the mean duration from the Fontan operation to diagnosis of cirrhosis was 23.4 6.3 years. There were 22 (55%) who had an atriopulmonary connection, 9 (22%) had a lateral tunnel, 5 (13%) had an extracardiac conduit, and 4 (10%) had other Fontan connections. A total of 6 patients had a fenestration during their initial operation. Only 1 patient had a cardiac transplantation before diagnosis of cirrhosis, 11 months after transplantation. There were 5 patients diagnosed with hepatocellular carcinoma at a mean duration of 20 2.9 years after the Fontan operation (2 died in long-term follow-up), and none had a history of viral hepatitis. The earliest diagnosis of hepatocellular carcinoma occurred in a 20-year old, 17 years after Fontan operation.
Table 2 includes factors associated with the development of cirrhosis after Fontan operation by Cox regression analysis. Of note, initial anatomy of hypoplastic left heart syndrome (2 of 16 patients developed cirrhosis) was the only predictor of cirrhosis on multivariate analysis (p ¼ 0.0133); sinus rhythm before the Fontan operation was the only protective factor on multivariate analysis (p ¼ 0.009). After a diagnosis of cirrhosis, survival was 57% at 1 year and 35% at 5 years with a total of 16 patients who died in long-term follow-up. The cause of death was known for 9 patients (5 had multiorgan failure, 2 died of liver failure, and 2 patients died of heart failure.) Discussion Using clinical, radiologic, and pathologic characterization, we found that cirrhosis is primarily diagnosed at least 20 years after Fontan operation as 94% of patients with liver follow-up at 20 years were free of cirrhosis, a number that dropped to 57% by 30 years post-Fontan. We show a lower incidence of cirrhosis compared with the cohort described by Lindsay et al; however, this is likely explained by our use of more stringent criteria for diagnosing cirrhosis, requiring both clinical and imaging evidence.14 We cannot use this study to describe a true incidence of cirrhosis after the Fontan operation because liver follow-up was only available in a portion (17%) of our population, but we begin to characterize those patients who do develop cirrhosis. Hepatic dysfunction is a known long-term complication of the Fontan operation. In this study, we noted that a higher proportion of patients with liver follow-up data had an extracardiac Fontan connection. This is most likely because of more available liver data from patients who had surgery during the last 2 decades, when a transition from atriopulmonary to extracardiac conduit Fontan connection occurred. There were also a higher proportion of patients with a history of arrhythmia or PLE with liver follow-up data, likely explained by more extensive follow-up for patients who developed these complications. Among the 17%
Congenital Heart Disease/Liver Disease After Fontan
of patients who had long-term liver follow-up, 21% (40 of 195) had documented cirrhosis. The cause of hepatic dysfunction after the Fontan procedure is thought to be related to passive venous congestion of the liver. In addition, the long-term compromise of relatively low cardiac output after Fontan operation15 may chronically reduce hepatic blood supply and compromise the ability of the hepatic artery to increase blood supply.16 However, it is not evident how long these changes have to exist for significant hepatic dysfunction to take place. Previous studies on the incidence of hepatic fibrosis and cirrhosis have shown that there is an incremental incidence of hepatic dysfunction with an increase in duration from initial Fontan as manifested by higher levels of portal and sinusoidal fibrosis at autopsy.8,11,17 A retrospective study of hepatic imaging studies and serology in 60 adult patients also indicated that cirrhosis was diagnosed 18.4 5.6 years after the Fontan procedure by hepatic imaging (ultrasound, MRI, and CT), and 22% of these patients had clinical manifestations of cirrhosis.14 One of the most significant challenges in the management of Fontan patients is determining how and when to monitor hepatic function. Recently, Rychik et al18 suggested starting invasive liver testing with liver biopsy 10 years after Fontan operation. Others have suggested that early serial imaging studies can be used to assess the burden of disease in patients, although the timing of such follow-up still poses a challenge.14 Additionally, it has been noted both in our cohort and others that basic liver function testing (aspartate aminotransferase, alanine aminotransferase, and bilirubin) is insensitive in identifying patients at risk for developing cirrhosis.14,19 Considering this, it is important to develop an informed paradigm for screening Fontan patients for liver disease, not only to try to diagnosis cirrhosis early but also balance the risks of invasive liver biopsies and serial radiation. This is additionally challenging as many patients had abnormal findings on liver imaging for years before the manifestation of cirrhosis, and abnormal liver biomarkers do not manifest until late in the disease process. Investigators previously reported autopsy studies, indicating that the presence of heterotaxy increased the severity of hepatic disease in patients after Fontan operation.11 In this study, we found that hypoplastic left heart syndrome was a multivariate predictor of cirrhosis, whereas sinus rhythm before Fontan operation was protective. However, given that only 2 of 16 patients with HLHS developed cirrhosis, it is difficult to determine whether this has clinical significance or attributable to a higher number of patients with HLHS receiving Fontan palliation in the most recent era when vigilance for liver disease has increased. Although the effect of preoperative sinus rhythm may reflect stronger cardiac function or reduced passive congestion, this would need further clarification with long-term sequential analyses of cardiac hemodynamics. Up to 2% to 4% of patients after Fontan operation have required a heart transplant,2,20e22 and it is important to consider the role of cirrhosis in transplant. In young Fontan patients having heart transplantation, the presence of cirrhosis did not affect 1-year mortality or liver function.23 Among patients who have chronic liver disease without cirrhosis, the presence of deteriorating cardiac function
459
could warrant earlier consideration for heart transplant to prevent the development of cirrhosis. Even patients with documented cirrhosis may have improvement of hepatic function after transplantation.12 However, it is still unclear if these patients are best served with a heart-only or heart-liver combined transplant strategy. In our cohort, only 1 patient with cirrhosis received a heart transplant. Future studies should assess the long-term cardiac and hepatic function of transplant patients based on the initial presence of cirrhosis. Long-term follow-up for our patients was limited to data from electronic and paper medical records, records from other institutions, and follow-up surveys that were mailed to patients, and thus, a referral bias may be present. Because routine follow-up of liver function was not consistently implemented until the last decade, a significant portion of our cohort did not have this information available and could not be included in the analysis. As a result, our cohort does not have chronological uniformity and likely represents a sicker population than the overall population of Fontan patients. Because of this and our stringent criteria for cirrhosis requiring pathology or abnormal imaging correlated with clinical findings, the overall frequency of liver disease in Fontan patients is likely underestimated by this study. Acknowledgments: The authors would like to recognize Drs Dwight McGoon, Gordon Danielson, Francisco Puga, Joseph Dearani, and Harold Burkhart, who performed many of the operations in this study and Matt Cetta for assistance with data abstraction. Disclosures The authors have no conflicts of interest to disclose. Supplementary Data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. amjcard.2015.11.014. 1. Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax 1971;26:240e248. 2. de Leval MR, Kilner P, Gewillig M, Bull C. Total cavopulmonary connection: a logical alternative to atriopulmonary connection for complex Fontan operations. Experimental studies and early clinical experience. J Thorac Cardiovasc Surg 1988;96:682e695. 3. Jonas RA, Castaneda AR. Modified Fontan procedure: atrial baffle and systemic venous to pulmonary artery anastomotic techniques. J Card Surg 1988;3:91e96. 4. Marcelletti C, Corno A, Giannico S, Marino B. Inferior vena cavapulmonary artery extracardiac conduit. A new form of right heart bypass. J Thorac Cardiovasc Surg 1990;100:228e232. 5. Khairy P, Fernandes SM, Mayer JE Jr, Triedman JK, Walsh EP, Lock JE, Landzberg MJ. Long-term survival, modes of death, and predictors of mortality in patients with Fontan surgery. Circulation 2008;117: 85e92. 6. Asrani SK, Asrani NS, Freese DK, Phillips SD, Warnes CA, Heimbach J, Kamath PS. Congenital heart disease and the liver. Hepatology 2012;56:1160e1169. 7. Cohen AJ, Cleveland DC, Dyck J, Poppe D, Smallhorn J, Freedom RM, Trusler GA, Coles JG, Moes CA, Rebeyka IM, Williams WG. Results of the Fontan procedure for patients with univentricular heart. Ann Thorac Surg 1991;52:1266e1270; discussion 1270e1261. 8. Ghaferi AA, Hutchins GM. Progression of liver pathology in patients undergoing the Fontan procedure: chronic passive congestion, cardiac
460
9.
10. 11.
12.
13. 14. 15. 16.
The American Journal of Cardiology (www.ajconline.org) cirrhosis, hepatic adenoma, and hepatocellular carcinoma. J Thorac Cardiovasc Surg 2005;129:1348e1352. Ginde S, Hohenwalter MD, Foley WD, Sowinski J, Bartz PJ, Venkatapuram S, Weinberg C, Tweddell JS, Earing MG. Noninvasive assessment of liver fibrosis in adult patients following the Fontan procedure. Congenit Heart Dis 2012;7:235e242. Lemmer JH, Coran AG, Behrendt DM, Heidelberger KP, Stern AM. Liver fibrosis (cardiac cirrhosis) five years after modified Fontan operation for tricuspid atresia. J Thorac Cardiovasc Surg 1983;86:757e760. Johnson JA, Cetta F, Graham RP, Smyrk TC, Driscoll DJ, Phillips SD, John AS. Identifying predictors of hepatic disease in patients after the Fontan operation: a postmortem analysis. J Thorac Cardiovasc Surg 2013;146:140e145. Crespo-Leiro MG, Robles O, Paniagua MJ, Marzoa R, Naya C, Flores X, Suarez F, Gomez M, Grille Z, Cuenca JJ, Castro-Beiras A, Arnal F. Reversal of cardiac cirrhosis following orthotopic heart transplantation. Am J Transpl 2008;8:1336e1339. Ellis EL, Mann DA. Clinical evidence for the regression of liver fibrosis. J Hepatol 2012;56:1171e1180. Lindsay I, Johnson J, Everitt MD, Hoffman J, Yetman AT. Impact of liver disease after the fontan operation. Am J Cardiol 2015;115:249e252. Nakamura Y, Yagihara T, Kagisaki K, Hagino I, Kobayashi J. Ventricular performance in long-term survivors after Fontan operation. Ann Thorac Surg 2011;91:172e180. Eipel C, Abshagen K, Vollmar B. Regulation of hepatic blood flow: the hepatic arterial buffer response revisited. World J Gastroenterol 2010;16:6046e6057.
17. Schwartz MC, Sullivan L, Cohen MS, Russo P, John AS, Guo R, Guttenberg M, Rand EB. Hepatic pathology may develop before the Fontan operation in children with functional single ventricle: an autopsy study. J Thorac Cardiovasc Surg 2012;143:904e909. 18. Rychik J, Veldtman G, Rand E, Russo P, Rome JJ, Krok K, Goldberg DJ, Cahill AM, Wells RG. The precarious state of the liver after a Fontan operation: summary of a multidisciplinary symposium. Pediatr Cardiol 2012;33:1001e1012. 19. Kaulitz R, Haber P, Sturm E, Schafer J, Hofbeck M. Serial evaluation of hepatic function profile after Fontan operation. Herz 2014;39: 98e104. 20. Dabal RJ, Kirklin JK, Kukreja M, Brown RN, Cleveland DC, Eddins MC, Lau Y. The modern Fontan operation shows no increase in mortality out to 20 years: a new paradigm. J Thorac Cardiovasc Surg 2014;148:2517e2523.e1. 21. Rossano JW, Shaddy RE. Heart transplant after the Fontan operation. Cardiol Young 2013;23:841e846. 22. d’Udekem Y, Iyengar AJ, Galati JC, Forsdick V, Weintraub RG, Wheaton GR, Bullock A, Justo RN, Grigg LE, Sholler GF, Hope S, Radford DJ, Gentles TL, Celermajer DS, Winlaw DS. Redefining expectations of long-term survival after the Fontan procedure: twenty-five years of follow-up from the entire population of Australia and New Zealand. Circulation 2014;130:S32e38. 23. Simpson KE, Esmaeeli A, Khanna G, White F, Turnmelle Y, Eghtesady P, Boston U, Canter CE. Liver cirrhosis in Fontan patients does not affect 1-year post-heart transplant mortality or markers of liver function. J Heart Lung Transpl 2014;33:170e177.
Since its introduction in 1968, the Fontan operation has been used for the palliation of patients with a single ventricle.1 The procedure has undergone many modifications.2e4 Numerous late complications after Fontan operation have been reported including arrhythmias, heart failure, protein losing enteropathy, and hepatic dysfunction.5 In patients after Fontan operation, increased central venous pressure and subsequent passive venous congestion have been associated with increased hepatic complications and overall mortality.6e11 However, it is unclear if any specific factors increase the risk of developing hepatic dysfunction, creating a challenge in identifying patients needing close follow-up. The hepatic dysfunction created by vascular congestion has demonstrated a level of reversibility,12,13 providing a possible role for early cardiac transplantation for susceptible patients. In this study, we examined our institution’s experience with hepatic dysfunction and cirrhosis in patients who had a Fontan operation. Methods In this institutional review boardeapproved single-center retrospective study, we reviewed the records of all patients (n ¼ 1,138) who had their initial Fontan operation or Fontan revision at the Mayo Clinic from October 1973 to June 2012. Information regarding demographic, anatomic, preoperative, operative, and postoperative variables was abstracted into a secure electronic patient database. This a Mayo Clinic College of Medicine, bDivision of Pediatric Cardiology, Department of Pediatrics, cDivision of Gastroenterology and Hepatology, Department of Medicine, dDivision of Cardiovascular Diseases, Department of Medicine, and eDivision of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota. Manuscript received July 14, 2015; revised manuscript received and accepted November 3, 2015. See page 459 for disclosure information. *Corresponding author: Tel: (507) 266-0676; fax: (507) 284-3968. E-mail address: [email protected] (J.N. Johnson).
0002-9149/15/$ - see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2015.11.014
database included information regarding preoperative and postoperative echocardiograms, cardiac catheterizations, electrocardiograms, Holter/event monitors, liver imaging, and surgical data. In addition, a medical questionnaire was mailed to all patients not known to be dead at the time of the study (n ¼ 723). Nonresponders received second and third questionnaires. If the subsequent questionnaires were not returned or completed, an attempt was made to contact the patients by telephone. The overall survey response rate was 42% (305 of 723). Data were available in the form of blood tests (aspartate aminotransferase, alanine aminotransferase, or bilirubin), liver ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), magnetic resonance elastography (MRE), or liver biopsy. Cirrhosis was diagnosed based on either (1) histopathologic diagnosis on biopsy or autopsy or (2) correlation of imaging findings (radiologic interpretation of liver cirrhosis on CT, MRI, MRE, or liver ultrasound) with clinical evaluation by a hepatologist with experience caring for Fontan patients (P.S.K). Patients were excluded from this analysis if they had evidence of viral hepatitis, alcoholic liver disease, or had a diagnosis of cirrhosis before their initial Fontan operation. Statistical analysis was performed using the date of the Fontan operation as time ¼ 0. All deaths were considered in the survival analysis, irrespective if they occurred intraoperatively, during immediate postoperative hospitalization, or after hospital discharge. Kaplan-Meier curves were derived to calculate 10-, 20-, and 30-year freedom from cirrhosis. Cox regression models were used to find univariate and multivariate predictors of being diagnosed with cirrhosis. The multivariate model considered univariately significant variables (p <0.05) with model selection using the stepwise method. All statistical tests were 2 sided with the alpha level set at 0.05 for statistical significance. SAS 9.3 was used for the analysis (SAS Institute, Cary, North Carolina). The cutoffs for discrete variables used in univariate and multivariate analyses were defined as follows: preoperative www.ajconline.org
Congenital Heart Disease/Liver Disease After Fontan
457
Table 1 Patient demographics Characteristic
All Fontan Patients (N ¼ 1,138)
Patients with liver data (N ¼ 195)
Patients with Cirrhosis (N ¼ 40)
456 (40%) 10.2 8.3
88 (45%) 9.6 9.2
16 (40%) 10.7 8
619 290 134 95
96 46 44 9
22 9 5 4
Female Age at Operation Type of Fontan operation Atriopulmonary connection Lateral tunnel Extracardiac conduit Other type of Fontan Preoperative anatomy Tricuspid Atresia Double Inlet Left Ventricle Asplenia/Heterotaxy Pulmonary Atresia/Intact Septum Hypoplastic Left Heart Syndrome Mean post-bypass left atrial pressure (mmHg) Mean post-bypass Fontan pressure (mmHg) Post-Fontan Ejection Fraction (%) Most recent Ejection Fraction (%) Most recent Systemic Ventricle End-Diastolic Pressure (mmHg) Average New York Heart Association Functional Class (1-4) Protein-Losing Enteropathy Documented arrhythmia
(54%) (25%) (12%) (9%)
310 (27%) 292 (26%) 78 (7%) 62 (5%) 27 (2%) 9.6 3.2 17.4 3.1 49.5 10.3 47.1 13.4 11 4.3 1.83 0.8 111 (9.7%) 522 (46%)
(49%) (24%) (22%)* (5%)
57 (29%) 40 (21%) 16 (8%) 16 (8%)* 16 (8%) 8.9 2.8 16.9 2.6 51 10 50.6 9.8 10.7 4.1 1.98 0.9 33 (17%)* 141 (72%)*
(55%) (22%) (13%) (10%)
13 (33%) 11 (28%) 4 (10%) 2 (5%) 2 (5%) 9.1 3 17.1 2.9 54 8.1 52 13.5 12.3 4.4 21 7 (18%) 34 (85%)*
* Denotes values that are statistically different from All Fontan Patients (p <0.05).
pulmonary artery pressures (>15 mm Hg), pulmonary arteriolar resistance (>3 U m2), ventricular end-diastolic pressure (>12 mm Hg), post-bypass Fontan pressures (>20 mm Hg), and chest tube duration (>21 days). Bypass times and post-bypass left atrial pressures were analyzed as continuous variables. The complete set of variables evaluated for their association with survival and functional status of survivors are listed in Supplementary Table 1. Results Overall, 1,138 patients had their initial Fontan or Fontan revision operation at our institution, with a mean duration of follow-up of 14.3 9.5 years (median 14.1 years). Clinical follow-up of liver function and associated imaging were available for 195 patients with an average duration of follow-up of 20.6 8.1 years (median 22 years). Demographics of this cohort are described in Table 1. In comparison with all Fontan patients, there was a higher proportion of patients with extracardiac conduits, initial lesion of pulmonary atresia with intact ventricular septum, history of arrhythmia, or history of PLE in patients with liver imaging available (p <0.05). There were no other significant differences between the liver followup and the overall Fontan cohort. In long-term followup, there were no significant differences in cardiac function in the form of ejection fraction or systemic ventricle end-diastolic pressure, or in functional class between the groups. For the 195 patients with available liver data, the mean age at Fontan operation was 9.6 9.2 years with a range of 7 months to 53 years; 40 of 195 patients (21%) had a fenestration performed at the time of the operation, with most of the fenestrations being performed after 1995. At
Figure 1. Kaplan-Meier curve of freedom from developing cirrhosis in 195 patients with liver data after Fontan. There is an incremental incidence of cirrhosis with freedom from cirrhosis of 99% (98.3-100) in 171 patients at 10 years and 94% (90-98) in 115 patients at 20 years.
most recent follow-up, 150 of 195 patients (77%) were alive and 45 (23%) were known to be dead. Liver biochemical tests were available for all 195 patients; there were no significant differences in aspartate aminotransferase, alanine aminotransferase, total bilirubin, or direct bilirubin levels between patients who did or did not develop cirrhosis; 24 of 195 patients (12%) had a liver biopsy (3 of 24 postmortem), of which 23 (96%) were abnormal; 152 of 195 patients (78%) had a liver ultrasound, of which 86 (57%) were noted to be abnormal; 25 of 195 patients (30%) had a liver MRI (72% abnormal), 33 of 195 (17%) had a CT of the liver (91% abnormal), and 33 of 195 (17%) had an MRE of the liver (79% abnormal) with 81 of 195 (42%) having at least
458
The American Journal of Cardiology (www.ajconline.org)
Table 2 Cox regression model for predictors of cirrhosis after the Fontan operation Variable Anatomic Factors Hypoplastic Left Heart Syndrome Pulmonary Artery Stenosis Asplenia Previous Surgical Procedures Bidirectional Glenn Pre-operative Factors Pre-operative Diuretic Use Pre-operative ACE-Inhibitor Use ECG Sinus Rhythm Operative Factors Fontan - Atriopulmonary Connection Fontan - Extracardiac Conduit Fenestration at Initial Operation AV Valve Replacement at Initial Operation Intra-operative Sinus Rhythm
Univariate Hazard Ratio (HR)
p-value
Multivariate Hazard Ratio (HR)
p-value
5.95 2.91 2.56
0.025 0.049 0.081
7.33 NS NS
0.0133 NS NS
4.2
0.021
NS
NS
2.36 4.11 0.33
0.013 0.007 0.014
NS NS 0.31
NS NS 0.009
0.33 4.63 3.9 4.81 0.35
0.002 0.008 0.016 0.01 0.014
NS NS NS NS NS
NS NS NS NS NS
one form of cross-sectional liver imaging. For the 195 patients with available liver data, the overall 10-, 20-, and 30year freedom from cirrhosis was 99%, 94%, and 57% respectively (Figure 1). A total of 40 patients were diagnosed with cirrhosis by pathology or correlation of imaging findings with clinical evaluation by a gastroenterologist. An established clinical diagnosis of cirrhosis by a gastroenterologist was made in 37 of 40 patients on long-term follow-up; the remaining 3 patients had biopsy-proved cirrhosis. Of these 37 patients, the most common clinical characteristics were varices (43%), ascites (35%), and thrombocytopenia (30%); only 1 patient each had hepatorenal syndrome and hepatic encephalopathy. Of the patients with cirrhosis, 10 of 40 had evidence of cirrhosis by pathology, 13 of 40 had evidence of cirrhosis on CT imaging, 8 of 40 on MRI, and 6 of 40 on MRE. Only 6 of 40 patients did not have CT, MRI, or MRE, but all these patients had an abnormal liver ultrasound along with a clinical diagnosis of cirrhosis by the hepatologist. Overall, 31 of 40 patients (78%) with cirrhosis had available ultrasound data and 28 of 31 (90%) were abnormal. MREderived mean liver stiffness data were available for 7 of 40 patients with an average of 6.0 1.3 kPa (>5 kPa interpreted as stage 4 fibrosis or cirrhosis based on institutional grading system derived for adult patients with other etiologies for liver disease). For the 40 patients with cirrhosis, the mean duration from the Fontan operation to diagnosis of cirrhosis was 23.4 6.3 years. There were 22 (55%) who had an atriopulmonary connection, 9 (22%) had a lateral tunnel, 5 (13%) had an extracardiac conduit, and 4 (10%) had other Fontan connections. A total of 6 patients had a fenestration during their initial operation. Only 1 patient had a cardiac transplantation before diagnosis of cirrhosis, 11 months after transplantation. There were 5 patients diagnosed with hepatocellular carcinoma at a mean duration of 20 2.9 years after the Fontan operation (2 died in long-term follow-up), and none had a history of viral hepatitis. The earliest diagnosis of hepatocellular carcinoma occurred in a 20-year old, 17 years after Fontan operation.
Table 2 includes factors associated with the development of cirrhosis after Fontan operation by Cox regression analysis. Of note, initial anatomy of hypoplastic left heart syndrome (2 of 16 patients developed cirrhosis) was the only predictor of cirrhosis on multivariate analysis (p ¼ 0.0133); sinus rhythm before the Fontan operation was the only protective factor on multivariate analysis (p ¼ 0.009). After a diagnosis of cirrhosis, survival was 57% at 1 year and 35% at 5 years with a total of 16 patients who died in long-term follow-up. The cause of death was known for 9 patients (5 had multiorgan failure, 2 died of liver failure, and 2 patients died of heart failure.) Discussion Using clinical, radiologic, and pathologic characterization, we found that cirrhosis is primarily diagnosed at least 20 years after Fontan operation as 94% of patients with liver follow-up at 20 years were free of cirrhosis, a number that dropped to 57% by 30 years post-Fontan. We show a lower incidence of cirrhosis compared with the cohort described by Lindsay et al; however, this is likely explained by our use of more stringent criteria for diagnosing cirrhosis, requiring both clinical and imaging evidence.14 We cannot use this study to describe a true incidence of cirrhosis after the Fontan operation because liver follow-up was only available in a portion (17%) of our population, but we begin to characterize those patients who do develop cirrhosis. Hepatic dysfunction is a known long-term complication of the Fontan operation. In this study, we noted that a higher proportion of patients with liver follow-up data had an extracardiac Fontan connection. This is most likely because of more available liver data from patients who had surgery during the last 2 decades, when a transition from atriopulmonary to extracardiac conduit Fontan connection occurred. There were also a higher proportion of patients with a history of arrhythmia or PLE with liver follow-up data, likely explained by more extensive follow-up for patients who developed these complications. Among the 17%
Congenital Heart Disease/Liver Disease After Fontan
of patients who had long-term liver follow-up, 21% (40 of 195) had documented cirrhosis. The cause of hepatic dysfunction after the Fontan procedure is thought to be related to passive venous congestion of the liver. In addition, the long-term compromise of relatively low cardiac output after Fontan operation15 may chronically reduce hepatic blood supply and compromise the ability of the hepatic artery to increase blood supply.16 However, it is not evident how long these changes have to exist for significant hepatic dysfunction to take place. Previous studies on the incidence of hepatic fibrosis and cirrhosis have shown that there is an incremental incidence of hepatic dysfunction with an increase in duration from initial Fontan as manifested by higher levels of portal and sinusoidal fibrosis at autopsy.8,11,17 A retrospective study of hepatic imaging studies and serology in 60 adult patients also indicated that cirrhosis was diagnosed 18.4 5.6 years after the Fontan procedure by hepatic imaging (ultrasound, MRI, and CT), and 22% of these patients had clinical manifestations of cirrhosis.14 One of the most significant challenges in the management of Fontan patients is determining how and when to monitor hepatic function. Recently, Rychik et al18 suggested starting invasive liver testing with liver biopsy 10 years after Fontan operation. Others have suggested that early serial imaging studies can be used to assess the burden of disease in patients, although the timing of such follow-up still poses a challenge.14 Additionally, it has been noted both in our cohort and others that basic liver function testing (aspartate aminotransferase, alanine aminotransferase, and bilirubin) is insensitive in identifying patients at risk for developing cirrhosis.14,19 Considering this, it is important to develop an informed paradigm for screening Fontan patients for liver disease, not only to try to diagnosis cirrhosis early but also balance the risks of invasive liver biopsies and serial radiation. This is additionally challenging as many patients had abnormal findings on liver imaging for years before the manifestation of cirrhosis, and abnormal liver biomarkers do not manifest until late in the disease process. Investigators previously reported autopsy studies, indicating that the presence of heterotaxy increased the severity of hepatic disease in patients after Fontan operation.11 In this study, we found that hypoplastic left heart syndrome was a multivariate predictor of cirrhosis, whereas sinus rhythm before Fontan operation was protective. However, given that only 2 of 16 patients with HLHS developed cirrhosis, it is difficult to determine whether this has clinical significance or attributable to a higher number of patients with HLHS receiving Fontan palliation in the most recent era when vigilance for liver disease has increased. Although the effect of preoperative sinus rhythm may reflect stronger cardiac function or reduced passive congestion, this would need further clarification with long-term sequential analyses of cardiac hemodynamics. Up to 2% to 4% of patients after Fontan operation have required a heart transplant,2,20e22 and it is important to consider the role of cirrhosis in transplant. In young Fontan patients having heart transplantation, the presence of cirrhosis did not affect 1-year mortality or liver function.23 Among patients who have chronic liver disease without cirrhosis, the presence of deteriorating cardiac function
459
could warrant earlier consideration for heart transplant to prevent the development of cirrhosis. Even patients with documented cirrhosis may have improvement of hepatic function after transplantation.12 However, it is still unclear if these patients are best served with a heart-only or heart-liver combined transplant strategy. In our cohort, only 1 patient with cirrhosis received a heart transplant. Future studies should assess the long-term cardiac and hepatic function of transplant patients based on the initial presence of cirrhosis. Long-term follow-up for our patients was limited to data from electronic and paper medical records, records from other institutions, and follow-up surveys that were mailed to patients, and thus, a referral bias may be present. Because routine follow-up of liver function was not consistently implemented until the last decade, a significant portion of our cohort did not have this information available and could not be included in the analysis. As a result, our cohort does not have chronological uniformity and likely represents a sicker population than the overall population of Fontan patients. Because of this and our stringent criteria for cirrhosis requiring pathology or abnormal imaging correlated with clinical findings, the overall frequency of liver disease in Fontan patients is likely underestimated by this study. Acknowledgments: The authors would like to recognize Drs Dwight McGoon, Gordon Danielson, Francisco Puga, Joseph Dearani, and Harold Burkhart, who performed many of the operations in this study and Matt Cetta for assistance with data abstraction. Disclosures The authors have no conflicts of interest to disclose. Supplementary Data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. amjcard.2015.11.014. 1. Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax 1971;26:240e248. 2. de Leval MR, Kilner P, Gewillig M, Bull C. Total cavopulmonary connection: a logical alternative to atriopulmonary connection for complex Fontan operations. Experimental studies and early clinical experience. J Thorac Cardiovasc Surg 1988;96:682e695. 3. Jonas RA, Castaneda AR. Modified Fontan procedure: atrial baffle and systemic venous to pulmonary artery anastomotic techniques. J Card Surg 1988;3:91e96. 4. Marcelletti C, Corno A, Giannico S, Marino B. Inferior vena cavapulmonary artery extracardiac conduit. A new form of right heart bypass. J Thorac Cardiovasc Surg 1990;100:228e232. 5. Khairy P, Fernandes SM, Mayer JE Jr, Triedman JK, Walsh EP, Lock JE, Landzberg MJ. Long-term survival, modes of death, and predictors of mortality in patients with Fontan surgery. Circulation 2008;117: 85e92. 6. Asrani SK, Asrani NS, Freese DK, Phillips SD, Warnes CA, Heimbach J, Kamath PS. Congenital heart disease and the liver. Hepatology 2012;56:1160e1169. 7. Cohen AJ, Cleveland DC, Dyck J, Poppe D, Smallhorn J, Freedom RM, Trusler GA, Coles JG, Moes CA, Rebeyka IM, Williams WG. Results of the Fontan procedure for patients with univentricular heart. Ann Thorac Surg 1991;52:1266e1270; discussion 1270e1261. 8. Ghaferi AA, Hutchins GM. Progression of liver pathology in patients undergoing the Fontan procedure: chronic passive congestion, cardiac
460
9.
10. 11.
12.
13. 14. 15. 16.
The American Journal of Cardiology (www.ajconline.org) cirrhosis, hepatic adenoma, and hepatocellular carcinoma. J Thorac Cardiovasc Surg 2005;129:1348e1352. Ginde S, Hohenwalter MD, Foley WD, Sowinski J, Bartz PJ, Venkatapuram S, Weinberg C, Tweddell JS, Earing MG. Noninvasive assessment of liver fibrosis in adult patients following the Fontan procedure. Congenit Heart Dis 2012;7:235e242. Lemmer JH, Coran AG, Behrendt DM, Heidelberger KP, Stern AM. Liver fibrosis (cardiac cirrhosis) five years after modified Fontan operation for tricuspid atresia. J Thorac Cardiovasc Surg 1983;86:757e760. Johnson JA, Cetta F, Graham RP, Smyrk TC, Driscoll DJ, Phillips SD, John AS. Identifying predictors of hepatic disease in patients after the Fontan operation: a postmortem analysis. J Thorac Cardiovasc Surg 2013;146:140e145. Crespo-Leiro MG, Robles O, Paniagua MJ, Marzoa R, Naya C, Flores X, Suarez F, Gomez M, Grille Z, Cuenca JJ, Castro-Beiras A, Arnal F. Reversal of cardiac cirrhosis following orthotopic heart transplantation. Am J Transpl 2008;8:1336e1339. Ellis EL, Mann DA. Clinical evidence for the regression of liver fibrosis. J Hepatol 2012;56:1171e1180. Lindsay I, Johnson J, Everitt MD, Hoffman J, Yetman AT. Impact of liver disease after the fontan operation. Am J Cardiol 2015;115:249e252. Nakamura Y, Yagihara T, Kagisaki K, Hagino I, Kobayashi J. Ventricular performance in long-term survivors after Fontan operation. Ann Thorac Surg 2011;91:172e180. Eipel C, Abshagen K, Vollmar B. Regulation of hepatic blood flow: the hepatic arterial buffer response revisited. World J Gastroenterol 2010;16:6046e6057.
17. Schwartz MC, Sullivan L, Cohen MS, Russo P, John AS, Guo R, Guttenberg M, Rand EB. Hepatic pathology may develop before the Fontan operation in children with functional single ventricle: an autopsy study. J Thorac Cardiovasc Surg 2012;143:904e909. 18. Rychik J, Veldtman G, Rand E, Russo P, Rome JJ, Krok K, Goldberg DJ, Cahill AM, Wells RG. The precarious state of the liver after a Fontan operation: summary of a multidisciplinary symposium. Pediatr Cardiol 2012;33:1001e1012. 19. Kaulitz R, Haber P, Sturm E, Schafer J, Hofbeck M. Serial evaluation of hepatic function profile after Fontan operation. Herz 2014;39: 98e104. 20. Dabal RJ, Kirklin JK, Kukreja M, Brown RN, Cleveland DC, Eddins MC, Lau Y. The modern Fontan operation shows no increase in mortality out to 20 years: a new paradigm. J Thorac Cardiovasc Surg 2014;148:2517e2523.e1. 21. Rossano JW, Shaddy RE. Heart transplant after the Fontan operation. Cardiol Young 2013;23:841e846. 22. d’Udekem Y, Iyengar AJ, Galati JC, Forsdick V, Weintraub RG, Wheaton GR, Bullock A, Justo RN, Grigg LE, Sholler GF, Hope S, Radford DJ, Gentles TL, Celermajer DS, Winlaw DS. Redefining expectations of long-term survival after the Fontan procedure: twenty-five years of follow-up from the entire population of Australia and New Zealand. Circulation 2014;130:S32e38. 23. Simpson KE, Esmaeeli A, Khanna G, White F, Turnmelle Y, Eghtesady P, Boston U, Canter CE. Liver cirrhosis in Fontan patients does not affect 1-year post-heart transplant mortality or markers of liver function. J Heart Lung Transpl 2014;33:170e177.