Evaluation of the progression of chronic liver disease (or its regression with treatment) is an important part of the hepatologist's practice. As the final common pathway of liver injury, fibrosis severity and progression are the most important predictors of outcome, irrespective of aetiology. The importance of discerning the stage of disease is manifold. Those with more advanced stages of fibrosis are more likely to go on to develop cirrhosis (1,2) and evidence of significant fibrosis is an indication to commence treatment in certain conditions, for example in hepatitis C (3). Evaluation of treatment efficacy, in autoimmune liver disease (AILD) for example, is also important. In addition, if cirrhosis is present, specific surveillance for gastrooesophageal varices and hepatocellular carcinoma should be initiated. Liver biopsy is presently the accepted standard for evaluating fibrosis but is limited by the need for hospital admission and sedation in children, sampling error (1/50,000 of the liver is sampled with needle biopsy), and serious risks including bleeding (4,5). Development of reliable and accurate noninvasive methods of determining fibrosis stage is important across all chronic liver diseases of childhood.
Several noninvasive methods of detecting liver fibrosis have achieved acceptance in adults during the past decade. Simple markers derived from large cohorts include the APRI index (6) and the Forns index (7). Algorithms that include measures of extracellular matrix turnover have also been used, for example, the fibrometer (8), the fibrotest (9), the hepascore (10), and the enhanced liver fibrosis test (11). Serum markers often fail to discriminate adequately between lesser grades of fibrosis, however. During the last 5 years, the use of transient elastography (TE) has become more and more accepted as a noninvasive marker of liver disease (12). Magnetic resonance elastography has also shown promise but cost may be preclusive (13). TE involves transmission of vibrations through the chest wall into the liver using an ultrasound probe with a vibration transducer. An elastic shear wave is induced in the liver, and using Young modulus (E), the elasticity of the liver is determined from the velocity of the wave where E = 3rVs2 (Vs is the shear velocity and r is the mass density). Liver stiffness is measured in kiloPascals (kPa). The accuracy of TE has been shown to be excellent in a large number of adult studies (14,15). A few small studies have also been undertaken in children (16,17).
Well-designed paediatric studies using biomarkers (in particular TE), to predict fibrosis, are scarce. Nobili et al (17) reported TE in 52 children with nonalcoholic fatty liver disease (NAFLD) and found excellent diagnostic accuracy, although the number of children with significant or severe fibrosis was small, thus possibly overestimating the accuracy of the technique (18).
There is increasing availability of newer therapeutic agents such as antiviral drugs for hepatitis and experimental antifibrogenic agents. This has put the onus on preventative hepatology and resulted in an effort to develop techniques of evaluating liver disease by noninvasive means. The pace of development of these new techniques is much slower in paediatrics because of a lack of adequately powered studies and newer technologies not licensed for use in children.
The aim of the present study was to establish the accuracy of TE in a cohort of children with chronic liver disease with a specific focus on evaluation of disease-specific accuracy.
Study Design and Data Collection
This was a single-centre, prospective cross-sectional study. Children with chronic liver disease undergoing biopsy in a supraregional paediatric liver centre between May 2009 and March 2011 were recruited following informed consent. The study was approved by the National Research Ethics Service via the local research ethics committee. Liver biopsy was undertaken for diagnostic purposes as part of routine clinical care and usually because of abnormal transaminases or suspected chronic liver disease. In 11 children postliver transplant, it was undertaken as routine surveillance for graft dysfunction 10 years after transplantation. In children with hepatitis B and C, biopsy was undertaken as part of protocol before commencing treatment.
Before biopsy, children underwent investigation for chronic liver disease. Screening for hepatitis A, B, and C in addition to Epstein-Barr virus and cytomegalovirus, Wilson disease, AILD, α1-antitrypsin deficiency, and metabolic liver disease was undertaken in the first instance. Ultrasound was done in all of the patients; echogenicity and texture of the liver were noted in addition to spleen size and any anatomical abnormalities.
Parents were approached before biopsy and gave informed consent. Inclusion criteria encompassed all those ages 6 to 18 years undergoing biopsy as part of routine clinical care during the study period. Children were excluded if they were too small for TE using the M probe (chest circumference <70 cm), if they had ascites, or if they were younger than 1 year from liver transplant. The variability in liver stiffness in patients posttransplant has been recently described and is more likely to reflect inflammation and vascular/perfusion changes rather than fibrosis (19). Children with known extrahepatic fibrogenic disease were also excluded.
The availability of only the standard M probe for TE at the time of the study limited inclusion criteria to children ages 6 to 18 years. Patients underwent blood sampling and TE on the same day of (or maximum 30 days from) biopsy. Data were collected on age, sex, onset symptoms/signs, indication for biopsy, final diagnosis, height, weight, body mass index (BMI), ultrasound findings, presence of varices, standard blood parameters including alanine transaminase, aspartate aminotransferase, bilirubin, platelet count, γ-glutamyl transpeptidase, alkaline phosphatase, haemoglobin, international normalized ratio, immunoglobulins, and autoantibodies, hepatitis screen, lactate, ceruloplasmin, serum and urinary copper, lipid profile, and oral glucose tolerance test where relevant.
TE was undertaken by a single operator unaware of fibrosis stage and blood biomarker results. TE was measured with the standard M probe using the right lobe of the liver through the intercostal space at the site of biopsy. The exception to this was in post-transplant patients (left lateral segment) in whom a spot was marked for biopsy with ultrasound usually in the epigastric region; this site was also used for probe placement. Ten valid readings were taken in each case aiming for an interquartile range (IQR) of <30% with at least 66% success rate. A 10-mm diameter core of tissue was measured with a depth between 25 and 65 mm.
Liver biopsy was undertaken by a hepatologist via the percutaneous route using the Menghini technique in the majority of cases. In 2 children, the biopsy was undertaken using a transjugular approach by an interventional radiologist in view of thrombocytopenia and thus increased bleeding risk. The formalin-fixed biopsy then underwent paraffin embedding and sectioning with a microtome to 4-μm sections. For routine diagnostic purposes, slides were stained with haematoxylin and eosin, reticulin, Perls, and Orcein. The biopsies were scored for fibrosis by a single hepatohistopathologist blinded to the results of the biomakers. This was done using either the METAVIR or the Kleiner scoring systems according to the diagnosis of the child. Inflammation was scored as none, mild, moderate, and severe. Degrees of steatosis and cholestasis were also recorded.
Descriptive results are expressed as median and IQR (nonparametric data) or number (percentage) of patients with a condition. The χ2 or Fisher exact test was used to compare categorical variables. Comparison of continuous variables between groups was undertaken using the Mann-Whitney U test and the Kruskal-Wallis test. Spearman ρ was used to estimate bivariate correlation between continuous data. Diagnostic performance of noninvasive methods was assessed by area under the receiver operating characteristic (AUROC) curve analysis and expressed as sensitivity, specificity, positive predictive value, negative predictive value, and positive and negative likelihood ratio (LR). Liver biopsy was used as the criterion standard. The receiver operating characteristic (ROC) curve was used to determine the optimal cut-off value for the presence of any/significant/severe fibrosis and cirrhosis. P < 0.05 was considered significant. SPSS version 17.0 (SPSS Inc, Chicago, IL) was used for analysis.
A total of 104 children were recruited to the study between May 2009 and March 2011. Three were excluded because they had an inadequate biopsy specimen (n = 1) or no blood suitable for analysis (n = 2). The remaining 101 children had measurements for at least 2 of the 3 biomarkers under investigation.
Demographics and Clinical Characteristics
A total of 101 children (62 boys) underwent data collection and analysis. Median age was 13.6 years (range 6–18 years). The diagnostic categories were NAFLD in 37 children, AILD in 27, viral hepatitis in 8 (hepatitis B virus: 4, hepatitis C virus:4), Wilson disease in 5, and in 8 miscellaneous conditions such as familial cholestasis or metabolic liver disease. A total of 16 children were postliver transplant. In the majority of children who were posttransplant (n = 11), liver biopsy was performed as part of routine 10-year surveillance and in the presence of normal transaminases. In the remainder (n = 5), biopsy was undertaken for suspected graft dysfunction. Clinical and laboratory data for these patients are shown in Table 1.
Liver biopsy procedure was uncomplicated in all of the children aside from one who complained of shoulder tip pain following the biopsy. This resolved spontanously without evidence of bleeding or pneumothorax. An adequate core of tissue was obtained in all of the cases. Some degree of fibrosis was evident in 95 (94%) biopsies; there was no evidence of fibrosis in 6 samples, 29 were scored as F1, 25 as F2, 34 as F3, and 7 as F4 (cirrhotic). A score of ≥F2 was taken as significant fibrosis and ≥F3 as severe fibrosis. The most clinically relevant distinction was made between no/minimal and significant fibrosis (<F2 and ≥F2) and between cirrhosis and no cirrhosis (F4 and <F4). Inflammation was graded as none, mild, moderate, and severe. Nine biopsies did not show evidence of inflammation; 63 were graded as mild, 23 as moderate, and 6 as severe inflammatory change. Five biopsies were reported as cholestatic; in 3 this was because of AILD, in 1 because of familial intrahepatic cholestasis, and chronic graft dysfunction in 1 child post-liver transplant. Steatosis was present in 44 biopsies (NAFLD, Wilson disease, glycogen hepatopathy, and in 1 child posttransplant). Demographics and clinical details were compared between diagnostic groups (Table 1).
Performance of TE as a Marker of Fibrosis
TE was successful in all but 7 children (7%); in 4 of these cases, this was because of a high BMI (>30). The other 3 children were postliver transplantation (left lateral segment), and the position of the graft made TE unreliable.
TE was a good discriminator of any fibrosis (≥F1) from no fibrosis (F0) (P = 0.013); of significant fibrosis (≥F2) from minimal/no fibrosis (<F2) (P < 0.001); of severe fibrosis (≥F3) from less severe fibrosis (<F3) (P < 0.001); and cirrhosis (F4) from no cirrhosis (<F4) (P = 0.003) (Fig. 1, Table 2) with areas under the ROC curve of 0.81 (P = 0.013) for any fibrosis; 0.78 (P < 0.001) for significant fibrosis; 0.79 (P < 0.001) for severe fibrosis; and 0.96 (P < 0.001) for cirrhosis.
The procedure was quick (<6 minutes in all cases), easily tolerated, and no child had any complications. TE values correlated to stage of fibrosis (Spearman r = 0.58, P < 0.001) in addition to bilirubin (Spearman r = 0.24, P = 0.02), aspartate aminotransferase (Spearman r = 0.287, P = 0.005), and γ-glutamyl transpeptidase (Spearman r = 0.324, P = 0.002). Sex and age did not correlate well with TE values. Severe inflammation on biopsy had a statistically significant relation with TE (P = 0.013). Severe inflammation was present in only 6 cases however, and lesser degrees of inflammation did not appear to affect TE reading. There was also a significant difference when those with no steatosis were compared with those with severe steatosis (P = 0.037).
The optimal cut-off for significant fibrosis in the present study was 6.9 kPa (sensitivity of 68% and specificity of 78%, positive likelihood ratio [LR+] 2.64, and negative likelihood ratio [LR−] 0.59). For severe fibrosis, the cutoff was 7.5 kPa, with a sensitivity of 72% and specificity of 76% and a LR+ of 2.8 and a LR− of 0.4. A cutoff of 14.1 kPa for cirrhosis has a sensitivity of 100% and specificity of 92%, with a positive likelihood ratio of 11.1.
Discrimination of Fibrosis Within Different Disease Groups
Subgroup analysis was undertaken to investigate the performance of the individual markers in the different aetiological groups (Table 3). Only the largest 3 groups were included in the subanalysis: NAFLD, AILD, and children posttransplant. In the remainder (diagnosis hepatitis B, hepatitis C, Wilson disease, and miscellaneous), TE was successful in all of the patients but numbers were too small to calculate reliable sensitivity, specificity, and AUROCs within these subgroups.
Performance of TE in NAFLD
In this subgroup, TE was a good predictor of significant fibrosis (≥F2, P = 0.02) and of severe fibrosis (≥F3, P = 0.002). There was no child with NAFLD-associated cirrhosis for comparison. The AUROC for significant fibrosis was 0.73 (P = 0.04); a cutoff of 6.1 kPa gave a 60% sensitivity and 78% specificity. For severe fibrosis, the AUROC was 0.8 (P = 0.003); a cutoff of 6.9 kPa had a 72% sensitivity and a 85% specificity. TE was not possible in 4 of these children (11%) because of high BMI. The median BMI z score of this group was 2.23 (IQR 1.87–2.41).
Performance of TE in AILD
Twelve children with AILD had type 1 disease, 1 child had type 2 AILD, and the remainder had autoimmune sclerosing cholangitis (8) or autoantibody negative AILD with high immunoglobulin G at presentation and characteristic AILD on liver biopsy. Median length of treatment before biopsy/TE was 1 week (IQR 0–24 weeks). Nine children had been commenced on prednisolone before biopsy, and 5 on both prednisolone and azathioprine. TE performed consistently well in children with AILD. For those with ≥F1 (any fibrosis), the AUROC curve was 0.92 (P = 0.05) and a cutoff of 6.3 kPa gave a 84% sensitivity and a 100% specificity for the presence of fibrosis. For children with significant fibrosis, the AUROC for TE was 0.87 (P = 0.004) and a cutoff of 7.4 kPa gave a 85% sensitivity and 85% specificity. In severe fibrosis, the AUROC was 0.89 (P = 0.001). A cutoff of 9.1 kPa gave a 86% sensitivity and an 85% specificity. Finally, in those with cirrhosis, the AUROC was 0.97 (P = 0.001) with a cutoff of 14.1 kPa, giving 100% sensitivity and 95% specificity.
Performance of TE in Children Post-liver Transplant
This third subgroup was the smallest one (16 children). The median interval from transplant to biopsy/TE was 10.2 years (IQR 5.1–11.4 years). Four children had whole liver allografts, the remainder (12), left lateral segments. Of these, 3 had previously had evidence of acute cellular rejection and 4 had nonspecific hepatitis in the first 2 years post transplant. Two children had previous evidence of de novo autoimmune hepatitis; however, neither was autoantibody positive at the time of the study. Other complications included posttransplant lymphoproliferative disease in 2, hepatic artery thrombosis (late) in 1, biliary stricture in 1, and a hepatic artery—portal vein fistula in 1 child. Two children were off immunosuppression at the time of biopsy (this was because of noncompliance in one case and cessation of treatment because of posttransplant lymphoproliferative disease in the other). The remainder were on a combination of prednisolone, tacrolimus, and mycophenolate or in the case of 3: cyclosporine and prednisolone.
TE failed in 3 (19%), mainly because of the position of the graft (left lateral segment). There was no child in this group without fibrosis. For those with significant fibrosis (≥F2), the AUROC for TE was 0.9 (P = 0.019) with a cutoff of 6.9 kPa, giving 100% sensitivity and 80% specificity. For severe fibrosis (≥F3), the AUROC was 0.83 (P = 0.057); a cutoff of 10.5 kPa had a 60% sensitivity and 75% specificity. Insufficient numbers were available for the analysis of cirrhosis.
The present study was designed to evaluate TE in a paediatric cohort with liver disease. Across all of the diagnostic groups, TE was particularly good at distinguishing significant fibrosis from no fibrosis and cirrhosis from less severe forms of fibrosis. TE is sometimes not possible in those with a high BMI (although a probe suited to this purpose has recently come on the market). Although not proposed as an alternative to biopsy for diagnostic purposes, TE could serve as a valuable tool in the long-term follow-up of paediatric liver disease. Liver biopsies are rarely repeated following initial diagnostic biopsy in paediatric cohorts and there is a real need for a noninvasive tool that could characterise fibrosis progression or regression. The close correlation of TE readings to extent of histological fibrosis suggests that TE is the ideal tool to follow liver fibrosis over years.
The performance of TE in the NAFLD group exposed certain challenges and limitations. AUROC did not meet the 0.97 to 1.0 levels expected from Nobili et al (17); however, it did reach a realistic AUROC of 0.73 and 0.8 for significant and severe fibrosis, respectively. Cutoff values were similar to those reported in the article. It is possible that the difference in the severity of liver disease in the 2 cohorts had an influence on the results. In the Italian cohort, there were fewer children in the significant or severe fibrosis categories and so the balance was weighted toward those with F1. In addition, difficulty acquiring measurement of TE because of the high BMI of 4 children possibly skewed results in a relatively small population. The availability of an XL probe that is more suitable for those with a higher BMI has improved the diagnostic accuracy of TE in NAFLD. The performance of TE in AILD was impressive with an excellent AUROC in distinguishing any, significant, severe fibrosis and cirrhosis.
It is now widely recognised that graft fibrosis is present in two-thirds of children 5 years after transplantation (20,21). Graft fibrosis usually develops silently in the presence of normal liver function tests and may lead to eventual graft loss. Modification of immunosuppresion may slow the development of fibrosis once identified, and thus a noninvasive means of assessing fibrosis in this cohort is much needed. TE performed well in those posttransplant with an AUROC of 0.9 for significant fibrosis and 0.83 for severe fibrosis (there were insufficient numbers to assess cirrhosis). Thus, the results obtained using TE in this group of patients are particularly encouraging. Previous studies have evaluated the use of TE in the recurrence of hepatitis C posttransplant (22,23), but there has not been any study evaluating posttransplant graft fibrosis in a paediatric cohort. In the present study, some difficulties with measurement of stiffness in the left lateral segment were encountered. The use of the S probe, which requires <4 cm of liver tissue, may improve the performance of TE in this context. In addition, ultrasound may be used to position the probe in an area that is unaffected by large vascular structures or bile ducts.
One of the main limitations of the present study was sample size with relatively small numbers in each of the diagnostic groups. The paucity of children with no fibrosis (F0) also made it difficult to differentiate between no and minimal fibrosis. Thus, the use of TE in primary care is questionable and is probably a more appropriate tool in secondary and tertiary referral centres.
This is, however, the largest paediatric study reported to date using TE in a comparison to liver biopsy. The combination of TE perhaps followed by confirmatory blood biomarkers could be useful both in screening (obese children for nonalcoholic steatohepatitis or children with cystic fibrosis for cystic fibrosis related liver disease) and for longitudinal monitoring of those with established fibrosis in chronic liver disease.
In conclusion, TE was a reliable tool in distinguishing different stages of liver fibrosis in children. Routine use of this technique may significantly decrease the number of biopsies performed and provide a reliable method of noninvasive monitoring of liver disease progression in children.
We thank the Starfish Appeal Charity and EchoSens for support.
1. Matteoni CA, Younossi ZM, Gramlich T, et al. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999; 116:1413–1419.
2. Blasco A, Forns X, Carrion JA, et al. Hepatic venous pressure gradient identifies patients at risk of severe hepatitis C recurrence after liver transplantation. Hepatology 2006; 43:492–499.
3. Strader DB, Wright T, Thomas DL, et al. Diagnosis, management, and treatment of hepatitis C. Hepatology 2004; 39:1147–1171.
4. Cadranel JF, Rufat P, Degos F. Practices of liver biopsy in France: results of a prospective nationwide survey. For the Group of Epidemiology of the French Association for the Study of the Liver (AFEF). Hepatology 2000; 32:477–481.
5. Poniachik J, Bernstein DE, Reddy KR, et al. The role of laparoscopy in the diagnosis of cirrhosis. Gastrointest Endosc 1996; 43:568–571.
6. Wai CT, Greenson JK, Fontana RJ, et al. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology 2003; 38:518–526.
7. Forns X, Ampurdanes S, Llovet JM, et al. Identification of chronic hepatitis C patients without hepatic fibrosis by a simple predictive model. Hepatology 2002; 36:986–992.
8. Cales P, Oberti F, Michalak S, et al. A novel panel of blood markers to assess the degree of liver fibrosis. Hepatology 2005; 42:1373–1381.
9. Imbert-Bismut F, Ratziu V, Pieroni L, et al. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study. Lancet 2001; 357:1069–1075.
10. Adams LA, Bulsara M, Rossi E, et al. Hepascore: an accurate validated predictor of liver fibrosis in chronic hepatitis C infection. Clin Chem 2005;51:1867–73.
11. Guha IN, Parkes J, Roderick P, et al. Noninvasive markers of fibrosis in nonalcoholic fatty liver disease: validating the European Liver Fibrosis Panel and exploring simple markers. Hepatology 2008; 47:455–460.
12. Friedrich-Rust M, Ong MF, Martens S, et al. Performance of transient elastography for the staging of liver fibrosis: a meta-analysis. Gastroenterology 2008; 134:960–974.
13. Binkovitz LA, El-Youssef M, Glaser KJ, et al. Pediatric MR elastography of hepatic fibrosis: principles, technique and early clinical experience. Pediatr Radiol 2012; 42:402–409.
14. Fraquelli M, Rigamonti C, Casazza G, et al. Reproducibility of transient elastography in the evaluation of liver fibrosis in patients with chronic liver disease. Gut 2007; 56:968–973.
15. Arena U, Vizzutti F, Abraldes JG, et al. Reliability of transient elastography for the diagnosis of advanced fibrosis in chronic hepatitis C. Gut 2008; 57:1288–1293.
16. de Ledinghen V, Le Bail B, Rebouissoux L, et al. Liver stiffness measurement in children using FibroScan: feasibility study and comparison with Fibrotest, aspartate transaminase to platelets ratio index, and liver biopsy. J Pediatr Gastroenterol Nutr 2007; 45:443–450.
17. Nobili V, Vizzutti F, Arena U, et al. Accuracy and reproducibility of transient elastography for the diagnosis of fibrosis in pediatric nonalcoholic steatohepatitis. Hepatology 2008; 48:442–448.
18. Poynard T, Muntenau M, Morra R, et al. Methodological aspects of the interpretation of non-invasive biomarkers of liver fibrosis: a 2008 update. Gastroenterol Clin Biol 2008; 32:8–21.
19. Inoue Y, Sugawara Y, Tamura S, et al. Validity and feasibility of transient elastography for the transplanted liver in the peritransplantation period. Transplantation 2009; 88:103–109.
20. Evans HM, Kelly DA, McKiernan PJ, et al. Progressive histological damage in liver allografts following pediatric liver transplantation. Hepatology 2006; 43:1109–1117.
21. Scheenstra R, Peeters PM, Verkade HJ, et al. Graft fibrosis after pediatric liver transplantation: ten years of follow-up. Hepatology 2009; 49:880–886.
22. Donato MF, Rigamonti C, Colombo M. Can protocol liver biopsy be avoided to evaluate post-transplant hepatitis C recurrence? Transient elastography makes it possible. Dig Liver Dis 2010;42:307.
23. Carrion JA, Navasa M, Bosch J, et al. Transient elastography for diagnosis of advanced fibrosis and portal hypertension in patients with hepatitis C recurrence after liver transplantation. Liver Transpl 2006; 12:1791–1798.