Department of Pediatric Cardiology, University of Nebraska College of Medicine and Children's Hospital and Medical Center, Omaha, NE
Address reprint requests to: Shelby Kutty, MD, FACC, Division of Pediatric Cardiology, University of Nebraska Medical Center, and Children's Hospital and Medical Center, 8200 Dodge St., Omaha NE 68114. E-mail: firstname.lastname@example.org; fax: 402-955-4356.
Potential conflict of interest: Nothing to report.
Hepatic dysfunction is a recognized complication after Fontan palliation of congenital heart disease. We sought to quantitatively measure hepatic stiffness and vascular Doppler indices using ultrasound (US) and shear wave elastography (SWE) in a Fontan cohort. Subjects were prospectively recruited for echocardiography and real-time hepatic duplex US with SWE for hepatic stiffness (kPa). Doppler peak velocities, velocity time integral, resistive, pulsatility, acceleration indices (RI, PI, AI), and flow volume were measured in celiac artery, superior mesenteric artery, and main portal vein (MPV). A subset underwent cardiac catheterizations with liver biopsy. Correlations were explored between SWE, duplex, hemodynamic, and histopathologic data. In all, 106 subjects were studied including 41 patients with Fontan physiology (age 13.8 ± 6 years, weight 45.4 ± 23 kg) and 65 controls (age 15.0 ± 8.4 years, weight 47.9 ± 22 kg). Patients with Fontan physiology had significantly higher hepatic stiffness (15.6 versus 5.5 kPa, P < 0.0001), higher celiac RI (0.78 versus 0.73, P = 0.04) superior mesenteric artery RI (0.89 versus 0.84, P = 0.005), and celiac PI (1.87 versus 1.6, P = 0.034); while MPV flow volume (287 versus 420 mL/min in controls, P = 0.007) and SMA AI (829 versus 1100, P = 0.002) were lower. Significant correlation was seen for stiffness with ventricular end-diastolic pressure (P = 0.001) and pulmonary artery wedge pressure (P = 0.009). Greater stiffness correlated with greater degrees of histopathologic fibrosis. No significant change was seen in stiffness or other duplex indices with age, gender, time since Fontan, or ventricular morphology. Conclusion: Elevated hepatic afterload in Fontan, manifested by high ventricular end-diastolic pressures and pulmonary arterial wedge pressures, is associated with remarkably increased hepatic stiffness, abnormal vascular flow patterns, and fibrotic histologic changes. The MPV is dilated and carries decreased flow volume, while the celiac and superior mesenteric arterial RI is increased. SWE is feasible in this population and shows promise as a means for predicting disease severity on liver biopsy. (Hepatology 2014;58:251–260)
Univentricular congenital heart disease (CHD) comprises 5% of all CHD admissions in United States. The Fontan procedure has been an effective strategy for the management of univentricular CHD over the past four decades, commonly performed as the final step of a staged surgical palliation. Large numbers of patients with this physiology are surviving into adulthood, and the presence of hepatic abnormalities in these patients is increasingly recognized.[3-6] The pathologic changes in the liver are similar to other forms of cardiac cirrhosis with prominent sinusoidal dilatation and fibrosis, presumably related to high venous pressures, hypoxia, and diminished cardiac output.[4, 7] Progressive hepatic failure is increasingly recognized, and there are isolated reports of hepatocellular carcinoma occurring in these young patients.[3, 8] Assessment of the status of the liver assumes great clinical importance in the follow-up surveillance of these patients; however, currently available methods are inadequate for early identification of these changes.
Liver biopsy is the traditional reference standard for the assessment of hepatic fibrosis or cirrhosis, but is invasive, expensive, and less than ideal method for repeated evaluations.[9, 10] Histologic studies of the liver in Fontan subjects are scarce. Histologic diagnosis has a certain interobserver variability even with standardized scoring, and the acquired sample may not be representative when there is heterogeneous distribution of fibrosis. Moreover, biopsy has a low sensitivity for the early stages of fibrosis. Serum tests of hepatic function including bilirubin, albumin, and coagulation indices are normal in most Fontan patients until the late stages of hepatic dysfunction. Various molecular serum markers (platelet ratio index, Fibrotest, Fibrometer, hyaluronic acid, and matrix metalloproteinases) have been reported as alternative testing for fibrosis; however, low sensitivity remains an issue. The newer biomarkers are not routinely available in most hospitals.
Noninvasive imaging techniques for detection and staging of hepatic fibrosis have been increasingly evaluated in adult hepatology.[14-19] Standard ultrasound (US) or computed tomography scans are capable of detecting hepatic nodularity, but have limitations to identify fibrosis, especially in the early stages. US techniques for measurement of tissue stiffness (elastography) are newer noninvasive tools that have been used for staging of hepatic fibrosis in adults.[14-21] Shear wave elastography (SWE) is one of the three available US techniques for tissue stiffness measurements; the others are transient elastography (TE, Fibroscan) and acoustic radiation force impulse (ARFI) imaging. SWE has several advantages compared to other methods. SWE allows two-dimensional, real-time, quantitative imaging of tissue stiffness. Being part of a US system, SWE can be performed along with standard US and Doppler examinations of the liver. It has been validated and shown to be fast and accurate in the evaluation of breast, thyroid, prostate, and hepatic stiffness.[15, 19, 20, 22] SWE was more accurate than TE for detection of significant fibrosis (≥F2 METAVIR score) in a recent large prospective study. Moreover, SWE is relatively insensitive to respiratory motion artifacts, and shown to have excellent inter- and intraobserver agreement of measurements. The feasibility of SWE for assessment of pediatric liver has been demonstrated.[23, 24]
The present study was performed under the auspices of the Liver Adult-Pediatric-Congenital-Heart-Disease Dysfunction Study (LADS) Group, a multidisciplinary group formed at our institution in August 2009 with the purpose of having a unified clinical and research approach for the better care of these patients. The study aimed to prospectively (1) measure hepatic stiffness in children and young adults with Fontan physiology using SWE in comparison with healthy controls; (2) correlate hepatic stiffness with echocardiography and catheterization derived indices of cardiac function; and (3) correlate hepatic stiffness with histopathology findings in a subset of patients who underwent liver biopsy.
Patients and Methods
This was a single-center prospective research trial. Patients with Fontan physiology and normal controls were enrolled. The Institutional Review Board approved the study and informed written consent was obtained from all subjects. Healthy volunteer children, adolescents, and young adults were prospectively recruited to serve as normal controls, exclusively for performing research liver ultrasound, Doppler, and SWE as part of this study. Inclusion criteria for controls were (1) age 4-35 years; (2) no previous history of heart disease, liver disease, or any other systemic disease; and (3) the provision of informed consent. The controls were recruited in response to an advertisement placed in the institutional employee newsletter that was approved by the Institutional Review Board. Complete transthoracic echocardiograms were performed in the Fontan subjects on the same day using a Vivid E9 US system (GE Healthcare, Milwaukee, WI).
Hepatic US and SWE Protocol
All subjects were instructed to fast for 4 hours preceding hepatic US. A comprehensive US examination with duplex followed by SWE was performed using the US-SWE system (SuperSonic Imagine Aixplorer, Bothell, WA) and selected broad bandwidth curved transducers (SC6-1 and SL15-4). The real-time B-mode imaging protocol with Doppler assessments of the celiac axis, superior mesenteric artery (SMA), and main portal vein (MPV) are detailed in Table 1. SWE combines ultrafast imaging and acoustic radiation force in a 2-step process: (1) shear wave generation by focusing an US beam, followed by (2) capture of the shear wave progression by very rapid acquisition of US images (up to 20,000 images per second) termed UltraFast Imaging. Quantitative viscoelasticity mapping of the liver was performed from the shear wave propagation, generating a real-time anatomic reference gray scale image and an elastogram color map.
Table 1. Hepatic US and SWE Protocol
B-mode imaging for anatomic assessment
B-mode, color and pulse wave Doppler for display and measurement of velocities, resistive index (RI), pulsatility index (PI), and acceleration index (AI).
Superior mesenteric artery:
B-mode, color, and pulse wave Doppler for display and measurement of velocities, resistive index (RI), pulsatility index (PI), and acceleration index (AI).
B-mode and color
Left, mid and right hepatic vein Doppler 1 cm distal to junction with inferior vena cava with measurement of systolic and diastolic velocities in each vein
(maximal forward and reverse flow)
Main portal vein:
B-mode, color, and pulse wave Doppler
Volume flow measurement (sample volume as wide as vessel diameter, with angle correction to be parallel to vessel walls)
Assessment for ascites
Right lobe shear wave elastography:
Activate after obtaining adequate B-mode image, adjust depth
Acquire 3 times with the subject holding breath
Scanning of the right liver lobe was performed with the subject supine, with the right arm in full abduction to maximally enlarge the intercostal space and increase access to the right hypochondrium. For SWE acquisition, the subject was asked to hold breath for at least 4 seconds at any time during the respiratory cycle (avoiding deep inspiration), and the largest intercostal interval was obtained by slight tilt of the transducer, using the best brightness on the B-mode image as a guide. Presence of a vessel in the SWE image was avoided with slight angulation of the scanning plane. At least three SWE image acquisitions were made in each study. One of two experienced sonographers performed all acquisitions.
The entire study was transferred by way of a flash drive to a Macintosh computer workstation for SWE quantification using the OsiriX DICOM review software and Q-box plug-in. The Q-box was positioned over an area of relative homogeneous elastogram (avoiding acoustic shadows, vessels, and Glisson's capsule) within a zone of uniform liver parenchyma (Fig. 1). The SWE gain (70%), Q-box diameter (15 mm), elastogram range (0-40 kPa), and depth (4-7 cm) were set by default per the manufacturer's recommendations. The mean elastogram value expressed in terms of Young's Modulus (kPa) in the area delineated by the Q-box was taken. The measurement was repeated three times to obtain three independent SWE maps of the patient in the same scanning plane, and the average of the three independent mean elastogram values was taken as the final hepatic stiffness for the patient. A single investigator (Q.P.) performed all measurements.
Transjugular Liver Biopsy
After obtaining vascular access in the right internal jugular vein using either a low lateral or a high approach with US guidance, a 9 FR sheath was placed. Standard hemodynamic and angiographic assessment of the Fontan circulation was performed. A 7 FR liver access and biopsy needle (Cook, Bloomington, IN) was then used to obtain hepatic vein wedge pressure, followed by hepatic vein and inferior caval/Fontan pressure. An angled tip glide catheter was used to preferentially enter the right lobe of the liver as far posteriorly as possible under biplane fluoroscopy guidance. A small volume hand injection of contrast was performed to confirm hepatic vein location. A 0.035 Amplatz superstiff wire with a 7 cm floppy tip was used to exchange for a 7 FR curved biopsy sheath, followed by repeat hand contrast injection. The sheath was then rotated to point anteriorly so as to position as much liver parenchyma as possible between its tip and the liver capsule. A 19 G core needle biopsy system was then introduced into the sheath while maintaining anterior direction of the posteriorly positioned tip of the sheath. The needle was then extruded just beyond the tip of the sheath before “firing” the spring-loaded needle into the hepatic parenchyma. While maintaining sheath position the needle was removed and technical support removed the biopsy sample from the needle onto a saline-soaked pad within a specimen container. The sheath was slightly repositioned by pull back followed by repeating the biopsy maneuver. Three specimens were routinely obtained in this manner. The biopsy sheath was removed from the coaxial jugular venous sheath and hemostasis was obtained.
Sections of the liver obtained from biopsy were stained with hematoxylin and eosin, trichrome, and reticulin. Two pathologists who were blinded to patient characteristics reviewed the specimens independently. The number of portal areas available for review was noted. Semiquantitative analysis was performed using previously established staging for sinusoidal dilatation (0-3), portal inflammation (0-3), and sinusoidal fibrosis (0-4). The degree of portal fibrosis was scored as: none, portal fibrosis, periportal fibrosis, bridging fibrosis, or cirrhosis. If the two pathologists scored a feature of the specimen differently, then they conferred to arrive at a final score.
Data are expressed as mean ± standard deviations with ranges in parentheses. The Student t test was used for comparisons between Fontan patients and controls. In the Fontan group, hepatic stiffness was compared between those with lower (<2) and higher (≥2) fibrosis grades on histology. Linear regressions of hepatic stiffness and vascular Doppler indices against patient age, body surface area, gender, and duration of Fontan were explored. Intraobserver agreements for 20 repeated SWE measurements were determined using Bland-Altman analysis to identify possible bias (mean divergence) and the limits of agreement (2 SD of the divergence). Statistical significance was defined as P < 0.05. Statistical analysis was performed with Minitab 16.1 (Minitab, State College, PA).
In all, 106 subjects were studied. The Fontan group consisted of 41 patients (29 male, 12 female, time since Fontan 11 ± 6 years), and there were 65 normal controls (36 male, 29 female). Demographics of subjects are summarized in Table 2. The Fontan types were extracardiac in 19 patients (46%) and lateral tunnel in 22 (54%). Echocardiographic assessment performed on the day of US-SWE showed normal systolic function in 35 (86%) patients, mildly depressed function in 5 (12%), and moderately depressed in 1 (2%). The ventricular morphology was left ventricular in 25 (61%) and right ventricular in 16 (39%). Atrioventricular regurgitation was none or mild in 36 (88%) patients and moderate in 5 (12%). Ten patients (24%) had an open fenestration in the Fontan conduit.
Table 2. Subject Demographics
Mean ± SD (range)
BSA, body surface area.
There was no statistically significant difference in hepatic stiffness between genders in the patient group (males 16.3 ± 5.7, females 13.8 ± 2.8, P = 0.145) or in the control group (males 5.7 ± 0.9, females 5.4 ± 0.84, P = 0.170). Age-wise comparisons were performed subgrouping controls into ≤10, 11-20, and >20 years: the respective stiffness values were 5.6 ± 0.9, 5.6 ± 0.8, and 5.4 ± 0.9, demonstrating no differences related to age (P = 0.99, 0.40, and 0.40 among subgroups).
It was feasible to obtain SWE measurements in all subjects. Hepatic stiffness was markedly increased in Fontan patients; 15.6 ± 5.1 (range 7.4 to 36.8 kPa) in Fontan, and 5.5 ± 0.87 (range 2 to 8.2 kPa) in controls (P < 0.0001). Vascular Doppler indices in Fontan patients compared to controls are summarized in Table 3. Higher celiac artery resistive index (RI) (0.78 versus 0.73, P = 0.04) and SMA RI (0.89 versus 0.84, P = 0.005) were seen in patients with Fontan physiology compared to controls. The celiac artery pulsatility index (PI) (1.87 versus 1.6, P = 0.034) was also higher in Fontan. In the SMA, the acceleration index (AI) (829 ± 330 in patients versus 1100 ± 463 cm/s2 in controls, P = 0.002) and velocity-time integral (VTI) (30.3 ± 15 in patients versus 37 ± 15 cm in controls, P = 0.03) were significantly lower in Fontan patients compared to controls, suggesting decreased SMA flow. The flow volume in the MPV was markedly decreased in Fontan patients (287 versus 420 mL/min in controls, P = 0.007) and the VTI was lower (17.7 ± 16 versus 23 ± 7.2 in controls, P = 0.036).
Table 3. Vascular Doppler Indices in the Celiac Artery, Superior Mesenteric Artery, and Main Portal Vein in All Subjects
Mean ± SD (range)
CA, Celiac artery; SMA, Superior mesenteric artery; MPV, Main portal vein; RI, Resistive index; PI, Pulsatility index; S/D, systolic-diastolic velocity ratio; AI, Acceleration index; VTI, Velocity-time integral. *Statistical significance.
No significant change was seen in hepatic stiffness or other duplex indices with patient age, gender, duration of Fontan palliation, or ventricular morphology. In Fontan patients, the MPV diameter correlated positively with the hepatic stiffness (P = 0.01), patient age (P = 0.009), and duration of palliation (P = 0.009). The MPV PI correlated inversely with patient age and duration of palliation (Supporting Fig. 1). Among patients who had catheterization data (n = 16), significant correlation was seen for SWE with ventricular end-diastolic pressure (P = 0.001) and pulmonary artery wedge pressure (P = 0.009). The Bland-Altman analysis showed good intraobserver agreement for hepatic stiffness measurements (Supporting Fig. 2). Table 4 shows the relationship of hepatic stiffness and duplex indices with age, gender, body surface area, and duration of Fontan palliation. Association of hepatic stiffness with cardiac catheterization data in Fontan patients is shown in Table 5.
Table 4. Relationship of Hepatic Stiffness and Vascular Doppler Indices With Age, Gender, Body Surface Area, and Duration of Fontan Physiology
Age at study
Time from Fontan
SWE value (kPa)
CA, Celiac artery; SMA, Superior mesenteric artery; MPV, Main portal vein; RI, Resistive index; PI, Pulsatility index; S/D, systolic-diastolic velocity ratio; AI, Acceleration index; VTI, Velocity-time integral; BSA, Body surface area. *Statistical significance.
SWE value (kpa)
Table 5. Association of Hepatic Stiffness With Cardiac Catheterization Indices (n = 16) in the Fontan Group
Regression with hepatic stiffness
Systemic venous O2 (%)
Systemic arterial O2 (%)
Pulmonary artery wedge pressure (mmHg)
Ventricular end diastolic pressure (mmHg)
Pulmonary vascular resistance (Wood units.m2)
Cardiac index (l/min/m2)
The mean interval between SWE and liver biopsy was 59 days (0-140). Histologic analysis revealed no sinusoidal dilatation in two patients, dilatation in <1/3 of the sinusoids (score 1) in seven patients, and dilatation in 1/3-2/3 of the sinusoids (score 2) in one patient. Two of the patients had no sinusoidal fibrosis (score 0), two had fibrosis present in <1/3 of sinusoids (score 1), and six had fibrosis present in 1/3-2/3 of the sinusoids (score 2). Examination of the portal areas revealed no portal fibrosis in three patients, portal fibrosis only in one patient, periportal fibrosis in two patients, bridging fibrosis in two patients, and cirrhosis in two patients (Fig. 2). Portal inflammation was minimal (score 1) in two patients and there was no portal inflammation in eight patients. The measured hepatic stiffness was 13.4 ± 1.3 kPa in those patients with fibrosis score of <2 (n = 4) and 19.8 ± 2.6 kPa in patients with fibrosis score ≥2 (n = 6, presence of periportal fibrosis, bridging fibrosis, or cirrhosis, P = 0.002). With small numbers limiting the analysis, the score appears to be similar in patients with portal fibrosis, bridging fibrosis, and cirrhosis (Table 6).
Table 6. Case by Case Analysis of Hepatic Stiffness and Histopathology Grade in 10 Fontan Subjects
Histopathology grade and hepatic stiffness
Sinusoidal dilatation (0-3)
Portal inflammation (0-3)
Sinusoidal fibrosis (0-3)
- Portal fibrosis only
- Periportal fibrosis
- Bridging fibrosis
Mean stiffness value (kPa)
Hepatic Dysfunction in Fontan Physiology
Patients after Fontan palliation are prone to a wide variety of complications that include myocardial dysfunction, systemic arterial and venous hemodynamic abnormalities, diminished exercise capacity, arrhythmias, hepatic dysfunction, protein-losing enteropathy, somatic growth retardation, neo-aortic root dilation, and thromboembolic events. With the Fontan procedure, separation of the systemic venous and pulmonary venous blood is accomplished using a direct connection from the inferior vena cava to the pulmonary arteries without a subpulmonary ventricular pump. As a result, elevated pressure is transmitted to the liver directly by way of the inferior vena cava and the hepatic veins. There are several reports on the devastating effects of this physiology on the liver.[3-5] Many of the Fontan survivors eventually may need cardiac transplantation in young adulthood for hemodynamic reasons. Significant hepatic dysfunction imposes additional risk for heart transplantation, and combined heart and liver transplantation is sometimes considered. It has been shown that indices of hepatic structure and function (serum fibrosis markers and indocyanine green clearance) in Fontan patients were similar to viral cirrhotic patients, but correlations between the indices and fibrosis was poor.
Hepatic Flow Dynamics in Fontan
High resistance to hepatic venous flow drainage (hepatic afterload) is the common denominator of liver disease in Fontan physiology, and likely explains changes in flow patterns, histology, and stiffness. Not surprisingly, our data confirm dilated MPV and markedly decreased absolute MPV flow volume and velocity-time integral in patients with Fontan physiology compared to controls. The few published studies on infradiaphragmatic venous flow in the Fontan circulation have found increased pulsatility of MPV flow in patients with Fontan physiology, consistent with our findings.[27, 28] The expiratory augmentation and inspiratory reduction of MPV flow was absent in patients with Fontan physiology, while there was no difference between Doppler derived inspiratory and expiratory volumetric flow rates.[27, 29] Although absolute MPV flow volume has never been quantified previously, alterations in portal flow has been suggested and attributed to high resistance to venous drainage and low cardiac output.
Inferences about comparative hepatic vascular indices in Fontan physiology remain speculative. The fibrotic liver likely presents high resistance to vascular inflow, whether arterial or venous. When inflow is within a highly pulsatile arterial structure (such as SMA and CA), high resistance might be expected to manifest in vascular indices. Velocity-based indices might be somewhat blunted by the changes in vascular caliber. Indeed, PI and RI in these arteries were both higher in the Fontan group. When inflow was within a much less pulsatile venous structure (such as the MPV), the impact of high resistance was manifest in the velocity ratios.
Histologic Consequences of High Hepatic Afterload
Hepatic histopathology from an autopsy series has demonstrated chronic passive congestion, centrilobular necrosis, and cardiac cirrhosis in patients with Fontan physiology. In that series, the severity of changes correlated with right atrial pressures, and with the time elapsed after surgery. Schwartz et al. have shown that both portal and sinusoidal fibrosis are common, and can progress over time. Our series confirmed that cirrhosis occurs in Fontan (20% of our sample), and hepatic fibrosis is common but not universal (70%). Others have reported somewhat greater proportions with cirrhosis and broad scars (60%), and sinusoidal dilation and fibrosis in all. Cirrhosis was more common in patients with high venous pressure, underscoring the role of venous hypertension in the progression of hepatic injury. The severity of fibrosis correlated with Fontan duration and hepatic venous pressures, suggesting a role for liver biopsy in the follow-up surveillance.
SWE, Fontan Physiology, and Liver Disease
The discovery of reliable noninvasive markers for liver disease in Fontan physiology would be of great value, so we chose to evaluate the feasibility of SWE for hepatic stiffness combined with US and duplex data for assessment of flow dynamics in this population. Measuring hepatic stiffness by SWE was feasible in all subjects. Our main findings were: (1) SWE permitted reproducible measurements of stiffness in all patients with low intraobserver variability; (2) hepatic stiffness was markedly increased and MPV flow volume was decreased in patients with Fontan physiology, while the celiac and mesenteric arterial resistive indices were higher; (3) in the subset of patients who had liver biopsies, greater hepatic stiffness correlated with greater degrees of hepatic fibrotic change; and (4) hepatic stiffness correlated with ventricular end-diastolic pressure and pulmonary artery wedge pressure, whereas stiffness and duplex indices did not show any relation with age, gender, time since Fontan, or ventricular morphology. Taken together, these findings indicate that Fontan physiology can produce sufficient elevation of hepatic afterload to result in changes in hepatic stiffness and histology.
Previous reports of hepatic stiffness assessment using the same technique used in our study included older patients with hepatitis C. Ferraioli et al. reported median stiffness of 7.6, 10.0, and 15.6 kPa corresponding to fibrosis stages 2, 3, and 4 from a hepatitis C cohort with mean age of 44 years. In an older cohort (mean age 55 years) with hepatitis C, a median stiffness of 9.1 was reported for the entire cohort; and 10.6, 14.5, and 27, respectively, for stages F2, F3, and F4. The median stiffness in our relatively younger Fontan cohort was higher (14.5 kPa) compared to reports in older hepatitis C patients. Because a smaller number of our patients had biopsy, stratification based on fibrosis stage was not feasible.
SWE in Comparison With Similar Techniques
Transient elastography, an alternative technique, has been used for noninvasive staging of hepatic fibrosis in adults,[16, 31, 32] but requires a dedicated system and measurements are one-dimensional. The accuracy and reproducibility of hepatic stiffness measurements by TE has been demonstrated in children with steatohepatitis. TE enabled accurate identification of patients without fibrosis or significant fibrosis and those with advanced fibrosis. Others have shown that TE is feasible even in very young children, but the type of US probe used, patient sedation, or food intake may influence the results. Other limitations of TE are related to (1) the relatively low volume of hepatic parenchyma explored, (2) absence of US imaging to guide the measurement, and (3) difficulties with obesity or ascites. TE has good sensitivity for cirrhosis (and less for degrees of fibrosis), but the stiffness cutoffs for the various stages have not been validated. SWE, on the other hand, is two-dimensional, and maps a greater hepatic area with its large frequency bandwidth, allowing improved diagnostic accuracy compared to TE. In patients with hepatitis C, SWE was superior to TE in the detection of mild and intermediate grades of fibrosis. A previous TE study found correlation between hepatic stiffness and the time since Fontan. This was not seen in our study, and the divergent observations remain currently unexplained. ARFI, the third US technique, is also one-dimensional, and measures from a small predetermined region of interest. Other limitations of ARFI are insensitivity to early stages of fibrosis, inability to generate an elastogram map, and the limited pediatric experience. Inaccuracies with ARFI result from signal attenuation in obese children and from excessive patient movement in uncooperative small children.
No statistically significant age- or gender-related differences in hepatic stiffness were demonstrated in this study. In normal adults, Sirli et al. found no significant differences in hepatic stiffness across a wide age range (18-70). Similarly, Popescu et al. showed no significant age- or gender-related differences in stiffness in adults spanning a similar age range. The impact of age and gender has been studied in the pediatric age group (1-17 years) as well, demonstrating that while age did not influence hepatic stiffness, an effect of gender was seen with slightly lower values for stiffness in females.[39, 40]
The relatively small sample size was a limitation of this study. The controls were not age- and gender-matched. Several factors including portal hypertension and inflammation may influence hepatic stiffness apart from fibrosis. However, histology data in Fontan patients including data from the present study have shown sinusoidal dilatation, distorted architecture, and fibrosis primarily without evidence of significant inflammation.[26, 41] All patients were studied in the supine position, and the effects of gravity or other variables that may influence venous Doppler patterns in the Fontan circulation were not evaluated. Intraobserver variability in stiffness measurements could be important clinically, so single stiffness determinations will need to be interpreted with caution. Replicated averaged determinations may provide more clinical value than a single measurement.
In conclusion, high ventricular end-diastolic pressure and pulmonary arterial wedge pressures are a manifestation of elevated hepatic afterload in the Fontan circulation. This physiologic derangement is reliably associated with changes in portal venous dimensions and absolute flow volumes, hepatic stiffness, and fibrotic histologic changes in the liver. Hepatic stiffness is markedly increased in Fontan physiology compared to controls. SWE is feasible in this young population and shows promise as a means for predicting severity of disease on liver biopsy. Further studies are needed to better understand the course of hepatic dysfunction, and for better stratification of risk in these patients.
The authors thank the patients who participated in this study. The authors appreciate the assistance of Sandeep Mukherjee, M.D., Marion Meckelburg, B.A., RDMS, Frank Galeoti, Nicole Hardin, M.S., R.T., Susan Walsh, R.N., MHSA, Swetha Natarajan, M.D., Rallyn Renner, RDMS, Kris Houston R.N., BSN, Ling Li, M.D., Ph.D., and Carolyn Chamberlain, R.N., BSN, MPH. Supersonic Imagine, Inc., generously provided equipment and technical support. S.K. receives support from the American College of Cardiology Foundation, the American Heart Association and the Children's Hospital and Medical Center Foundation.