Secondary Logo

Correlation of Transient Elastography With Severity of Cystic Fibrosis–related Liver Disease

Aqul, Amal*; Jonas, Maureen M.; Harney, Sarah; Raza, Roshan; Sawicki, Gregory S.§; Mitchell, Paul D.; Fawaz, Rima

Journal of Pediatric Gastroenterology and Nutrition: April 2017 - Volume 64 - Issue 4 - p 505–511
doi: 10.1097/MPG.0000000000001448
Original Articles: Hepatology
Free

Objectives: The aim of the study was to evaluate whether liver stiffness measurement (LSM), determined by transient elastography, correlates with presence and severity of liver disease in children and young adults with cystic fibrosis (CF).

Methods: Subjects underwent LSM at routine CF visits. Presence and severity of cystic fibrosis liver disease (CFLD) was determined by clinical parameters. Subjects were classified as no CFLD, CFLD without portal hypertension (PHTN), and CFLD with PHTN. LSM was compared with aspartate aminotransferase/platelet ratio index (APRI) as a correlate to severity of CFLD.

Results: A total of 249 subjects (53% boys; mean age 14 ± 7 years; 7 [3%] <2 years and 74 [30%] 18–25 years) underwent LSM. Subjects were classified as 158 (64%) with no CFLD, 73 (29%) CFLD without PHTN, and 18 (7%) CFLD with PHTN. The median (interquartile range) LSM was different among the 3 groups: 4.4 (3.8–5.4), 5.1 (4.4–6.3), and 14.1 (8.8–24.8) kPa, respectively, with all pairwise comparisons different from one another (P < 0.0001). Similarly, median (interquartile range) APRI was different in groups 1 and 2 compared with CFLD with PHTN: 0.22 (0.17–0.27), 0.24 (0.17–0.33), and 0.53 (0.24–0.84), respectively (P < 0.01). Analysis of receiver operating characteristics for discriminating CFLD with PHTN from the other groups resulted in cut-points at 6.2 kPa (LSM) and 0.35 (APRI). LSM was superior to APRI in discriminating CFLD with PHTN from other groups, with areas under the curve 0.91 (LSM) versus 0.78 (APRI) (P = 0.05).

Conclusions: Liver stiffness, as determined by transient elastography, correlates with the presence and severity of CFLD. Although APRI provided some information regarding severity of liver disease, LSM performed better than APRI in this population.

*Division of Gastroenterology, Hepatology, and Nutrition, University of Texas Southwestern Medical Center, Dallas

Division of Gastroenterology, Hepatology, and Nutrition

Clinical Research Center

§Division of Respiratory Diseases, Cystic Fibrosis Center, Boston Children's Hospital, Boston, MA.

Address correspondence and reprint requests to Rima Fawaz, MD, Boston Children's Hospital, 300 Longwood Ave, Boston, MA 02115 (e-mail: Rima.Fawaz@childrens.harvard.edu).

Received 3 September, 2015

Accepted 19 October, 2016

Echosens (Paris, France) provided the FibroScan machine, technical support, and training of investigators for the purpose of this study. Echosens had no role in study design, collection/analysis/interpretation of data, writing the manuscript, or the decision to submit the manuscript for publication. Intramural funding was provided by the Division of Gastroenterology, Hepatology and Nutrition and the Cystic Fibrosis Center at Boston Children's Hospital.

Portions of this study were presented at The Liver Meeting (the annual meeting of the American Association for the Study of Liver Diseases) on October 31, 2013 in Washington, DC.

The authors report no conflicts of interest.

What Is Known

  • Liver disease is the third leading cause of death in patients with cystic fibrosis.
  • Diagnosis of cystic fibrosis liver disease is often delayed due to the subclinical nature of the disease.
  • There is no reliable tool to identify individuals at risk for developing advanced liver disease and those with precirrhotic liver fibrosis.

What Is New

  • This is the largest study to report on the correlation of transient elastography with the severity of liver disease in children and young adults with cystic fibrosis.
  • Liver stiffness measurement as reported by transient elastography, provides a rapid and noninvasive tool for assessment of presumed liver fibrosis in patients with cystic fibrosis.

Cystic fibrosis (CF) is the most common lethal inherited disease in the Caucasian population. Liver disease (LD) is the third leading cause of death, accounting for 2.5% of the overall mortality (1). Cystic fibrosis liver disease (CFLD) includes a wide range of abnormalities with variable degrees of fibrosis, including cirrhosis with portal hypertension (PHTN). Diagnosis of CFLD is often delayed due to the subclinical nature of the disease. In addition, there is no reliable tool to identify individuals with CF at risk for developing advanced LD and those with precirrhotic liver fibrosis (2).

Although liver biopsy is considered the reference standard to determine the severity of fibrosis, it has limited accuracy in CFLD. Cirrhosis is missed in 10% to 30% of adults with different types of LDs (3–5). This is especially pertinent to CFLD given the focal nature of the disease (6). In a small pediatric CFLD study, dual pass biopsies revealed discordance in 38% of the biopsy pairs (7). In addition, liver biopsy is an invasive procedure with unavoidable risks and complications, which limits its repeated use to determine the progression of disease.

Studies have shown the reliability and accuracy of transient elastography (TE) in measuring liver stiffness as a correlate of fibrosis in adult patients (8–13). There have, however, been few studies using this technique in children, particularly with CF.

The aim of the present study was to investigate the correlation between liver stiffness, as determined by TE, and the presence and severity of CFLD in children and young adults.

Back to Top | Article Outline

METHODS

This was a cross-sectional observational study of children and young adults with CF. Patients from 1 month to 25 years of age were recruited during routine visits to the Boston Children's Hospital (BCH) Cystic Fibrosis Center between January 28, 2013 and August 30, 2013. Exclusion criteria were known LD other than CFLD, liver transplant, and pregnancy.

Clinical and biochemical data were obtained from the medical records. This included age, sex, weight, height, presence or absence of splenomegaly as determined by physical examination or ultrasound, pulmonary function test results, and prescription for ursodeoxycholic acid at the time of liver stiffness measurement (LSM). Biochemical parameters obtained within 3 months of the LSM date were recorded, including serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), bilirubin, and platelet count. The AST:platelet ratio index (APRI) was calculated as (AST/upper limit of normal [ULN])/platelets × 100, using, 40 IU/L as the AST ULN at BCH laboratory. Findings from most recent abdominal ultrasound and/or gastrointestinal endoscopy, if performed, were recorded. Body mass index z scores were calculated for all subjects up to the age of 20 years. Scores in subjects younger than 2 years (n = 7) were based on WHO standards, whereas those for older children were based on Centers for Disease Control and Prevention charts (14).

The present study was approved by the BCH institutional review board. Written informed consent was obtained from parents, guardians, and subjects 18 years or older. Assent was obtained from children ages 7 to 17 years.

Back to Top | Article Outline

Liver Stiffness Measurement

TE (FibroScan; Echosens, Paris, France) was performed by 2 trained investigators (S.H., R.R.) who were certified by the manufacturer. Female subjects ages 12 years or older had negative urine pregnancy tests before TE. In each subject, 10 measurements (8 in one patient) were obtained in succession, with results reported in kiloPascals (kPa). The validity of each measurement is assessed by the device. LSM was not recorded if either an insufficient number of valid measurements was obtained, or the interquartile range (IQR) of measurements was ≥1/3 of the median. Following the manufacturer's recommendations, subjects with a thoracic perimeter, measured at the xiphoid process with arms hanging down, <75 cm were examined with the “S” probe, and subjects with a thoracic perimeter >75 cm were examined with the “M” probe. The “S” probe has a probe ultrasound frequency of 5 MHz and measures depths from 15 to 50 mm. The “M” probe works at an ultrasound frequency of 3.4 MHz and measures depths from 25 to 65 mm.

Back to Top | Article Outline

Cystic Fibrosis Liver Disease Categories

Patients were categorized into 3 groups according to clinical, biochemical, sonographic, and endoscopic parameters. These groupings were based on published best-practice guidelines for diagnosis of CFLD (15). Subjects with no LD were defined by most recent ALT < 1.3 × ULN (30 IU/mL is ULN at BCH), not prescribed ursodeoxycholic acid, and normal ultrasound (if done). Subjects categorized as CFLD without PHTN met at least one of the following criteria: most recent ALT > 1.3 × ULN, prescribed ursodeoxycholic acid, or abnormal liver echogenicity without evidence of PHTN. CFLD subjects with PHTN met at least one of the following: splenomegaly on physical examination defined as spleen below costal margin, presence of esophageal varices by traditional or video capsule endoscopy, platelet count <100,000/mm3 or evidence of PHTN on ultrasound (ascites, varices, and/or splenomegaly). Sonographic definition of splenomegaly was spleen length to age >90% of upper confidence limit (16).

Back to Top | Article Outline

Statistical Analyses

Data are described with frequency counts and percentages for categorical variables, median (IQR) for continuous variables that are highly skewed, and mean ± standard deviation otherwise. Unadjusted comparisons were made with Fisher exact test, Wilcoxon rank-sum test, or Student t test, depending upon the nature of the data. The variance of LSM was vastly different across CFLD categories, from 2.6 among subjects without CFLD, to 19.2 among CFLD subjects without PHTN and 475.8 among those with PHTN. Consequently, the analysis of LSM was conducted both parametrically (in kPa units for ease of interpretation) and nonparametrically (in standard normal units to satisfy the analysis of variance assumptions). Data were transformed to a standard normal distribution with mean 0 and variance 1 based on ranked LSM using the method of Blom (17). The omnibus 3-group comparison of LSM (kPa) was assessed with a Kruskal-Wallis test and normalized LSM by analysis of variance. No adjustment was made for pairwise comparisons, since no more than 3 groups were compared (18). Receiver operating characteristic (ROC) curves were constructed for LSM and APRI to evaluate correlation with advanced LD classification. The optimal diagnostic cut-point was based on the Euclidean distance criterion (19). Analysis was performed with SAS v9.3 (Cary, NC).

Back to Top | Article Outline

RESULTS

A total of 348 patients with CF were seen for routine care over a 7-month period in 2013, representing 76% of all patients with CF ages 0 to 25 years in the BCH CF Patient Registry. Of these, 264 subjects were enrolled, and the data from 249 with valid LSM and determinable LD status were analyzed (11/15 not analyzed were <4 years). Derivation of the study population is shown in Figure 1.

FIGURE 1

FIGURE 1

Subjects were 14 years of age, with 3% <2 years and 30% 18 to 25 years. CF was diagnosed by either sweat test or genetic test, with 106 (44%) homozygous and 110 (46%) heterozygous for the ΔF508 CFTR mutation. Among 119 (48%) subjects who had undergone abdominal ultrasound, abnormalities were found in 50 (42%) and signs of PHTN in 16 (13%). ALT and AST were >1.3 times the ULN in 30 (12%) and 13 (5%) of subjects, respectively. One patient had an elevated direct bilirubin levels 1.4 mg/dL. 158 (64%) subjects were classified as no CFLD, 73 (29%) as CFLD without PHTN, and 18 (7%) as CFLD with PHTN (Table 1). The distribution of LSM stratified by suspected CF LD severity is shown in Figure 2. On average, subjects without CFLD have lower LSM than subjects with CFLD in the absence of PHTN (P < 0.0001), and those without PHTN have lower LSM than those with PHTN (P < 0.0001). Defining high LSM as >8.6 kPa (for METAVIR F3 or F4 (20)), 17 of 27 (63.0%) patients had normal platelet levels (as defined by >150,000), compared to 215 of 216 (99.5%) with LSM ≤8.6 kPa (P < 0.0001). Using a platelet count <100,000, 22 of 27 (81.5%) with LSM >8.6 kPa had normal platelet levels compared to 216 of 216 (100%) with LSM ≤8.6 kPa (P < 0.0001). A stricter cut point of >11.6 kPa (for METAVIR F4) results in 8 of 17 (47.1%) with normal platelets compared to 215 of 216 (99.5%) with ≤11.6 kPa (P < 0.0001).

TABLE 1

TABLE 1

FIGURE 2

FIGURE 2

LSM in one subject who previously had ascites was 23.1 kPa, LSM in the 5 subjects with varices was 23.1, 24.8, and 45.0 kPa, and 2 subjects with 75.0 kPa (75 being maximum LSM reported by fibroscan). LSM in 1 patient with cholestasis (direct bilirubin 1.4 mg/dL) was 37.4. This patient was categorized as CFLD without PHTN as per our inclusion criteria, although is now known to have advanced CFLD with spontaneous splenorenal shunt and is listed for transplant.

Ten subjects had splenomegaly, either on physical examination or ultrasound, as their only criteria for the CFLD with PHTN category. The LSM of subjects with splenomegaly found either on physical examination (n = 3) or by ultrasound (n = 4) as an only criterion for CFLD with PHTN was lower than LSM of subjects with splenomegaly found on both physical examination and ultrasonography (n = 3). In addition subjects who had splenomegaly and esophageal varices had markedly higher LSM values, with the lowest LSM recoded at 23.1 and highest 75 (n = 5) (Table 2).

TABLE 2

TABLE 2

Unadjusted comparison of LSM and APRI with known correlates of LD is shown in Table 3. Higher LSM and APRI were each associated with ALT >1.3 × ULN, AST >1.3 × ULN, and use of ursodeoxycholic acid. Likewise, each was positively associated with suspected LD severity.

TABLE 3

TABLE 3

The ROC for LSM and APRI with respect to discriminating CFLD with PHTN from either no CFLD or CFLD without PHTN are shown in Figure 3. The area under the ROC curve was greater for LSM than for APRI (P = 0.05), with optimal cut-points for diagnosing CFLD with PHTN at LSM >6.2 and APRI >0.35. Using a combination of LSM and APRI did not provide additional discriminatory ability in identifying patients with CFLD. The area under the curve (AUC) is 0.91 for LSM, 0.78 for APRI, and 0.90 for LSM and APRI together. The AUC for LSM is an improvement over AUC for APRI alone (Δ AUC = 0.91 – 0.78 = 0.13; P = 0.05) as already shown in Figure 3.

FIGURE 3

FIGURE 3

Back to Top | Article Outline

DISCUSSION

Detection and staging of CFLD represents a major clinical challenge due to the lack of sensitive and specific diagnostic tools. Guidelines recommend annual determination of liver enzymes in all patients with CF, with follow-up if abnormal. Ultrasonography and liver biopsy are recommended if a patient has persistently elevated liver enzymes with no explanation (21). Liver biopsy is, however, an imperfect indicator of liver fibrosis in CFLD (8).

In the present study, the ability of LSM to predict the severity of CFLD in a cohort of children and young adults was evaluated. Although few studies have compared TE with different radiological modalities (22,23), to our knowledge, this is the largest cohort of pediatric and young adult patients with CF to be evaluated with LSM for correlation with indicators of severity of CFLD. In our study, platelet count was found to be an insensitive marker of CFLD.

In a review by Flass and Narkewicz (2) it was proposed that liver involvement in CFLD can be classified into 3 groups based on clinical, histological, and imaging data. This was the basis for the classification used in the present study. Since the practice at BCH is to treat patients with CF with persistent elevation of aminotransferases with ursodeoxycholic acid, subjects prescribed ursodeoxycholic acid were categorized as having CFLD even if the most recent ALT was normal. TE was performed only during routine visits to avoid the possible influence of acute illness on LSM.

In the present study, LSM was a simple and reliable tool, with valid measurements in approximately 95% of the subjects. Most of the invalid measurements were seen in patients younger than 3 years, due to lack of patient cooperation. The median LSM in subjects with no CFLD was 4.4 kPa, comparable to the median LSM of 4.7 kPa reported in a recent study of healthy German children (24). The median LSM in subjects with CFLD without PHTN was 5.1 kPa. Although the distribution of LSM in this group was statistically different when compared to both subjects with no LD (P = 0.0001) and subjects with PHTN (P < 0.0001), it falls within the range reported in normal healthy children (24). CFLD that has not progressed to clinically apparent cirrhosis and PHTN is the greatest diagnostic challenge in this population. Identification of an LSM value for early fibrosis could influence management. Monitoring of serial LSM for changes over time may be a more useful strategy for detecting early fibrosis in CFLD.

Lee et al (20) studied TE in 128 patients (ages 1.4 months–27 years) who underwent liver biopsy for various LDs. METAVIR scoring was used to stage fibrosis (25). The median (IQR) LSM in subjects with F3-F4 fibrosis was 14.6 kPa (8.9–22.6 kPa), which is similar to the distribution of LSM in subjects with CFLD and PHTN in the present study, 15.6 kPa (8.8–24.8). APRI has been proposed as a simple noninvasive index of fibrosis and cirrhosis in pediatric patients with viral hepatitis (26). In the present study, the cut-point of APRI to discriminate between subjects with and without PHTN was comparable to what has been reported in CFLD (27,28). APRI, using a cut-point of 0.3 to 0.49, was an accurate predictor of CFLD and PHTN, with a sensitivity ranging between 83% and 87%. When comparing the ROC curves for LSM and APRI using the cut-points of 6.2 kPa and 0.35, respectively, LSM was superior to APRI Interestingly a recent study evaluating the use APRI in predicting degree of fibrosis in pediatric CFLD found a similar cut-off point of >0.264 (29).

There are limitations to the present study. Not all eligible patients seen were enrolled in the study but there was no inherent selection bias since we enrolled patients coming for routine visits on all days on the week and from all providers at BCH. LD categories were broad and inclusive so that no patients with LD would be missed; however, this may have resulted in some misclassifications. Defining splenomegaly as spleen below the costal margin in patients with CF with hyper expanded lungs may have falsely overestimated the presence of PHTN. Patients with abnormal liver echogenicity on ultrasound which is a subjective finding may misclassify patients as having CFLD without PHTN. In addition, no testing was done just for purposes of the present study, so not all data were available for each subject at the time of LSM. The study design, however, provided a large comprehensive subject population, reflecting clinical practice and patient distributions at a large CF center. These limitations may explain the large change in the LSM within the group of CFLD with PHTN; however, it was obvious that subjects with >1 feature of PHTN have higher LSM. In this cohort, only 5 subjects had liver biopsy, most of which were performed >4 years before LSM, so it was not possible to correlate histological findings with LSM. A longitudinal study using serial LSM in this cohort, that is currently underway, is expected to add important information regarding the feasibility of using this technique to detect progression of CFLD.

In conclusion, there has been no established reliable method to detect fibrosis in CFLD. LSM, as measured by TE, provides a rapid, and noninvasive tool for assessment of presumed liver fibrosis in patients with CF. LSM is superior to APRI in detecting PHTN in this population. The use of TE in patients with CF could eventually guide further management regarding the timing of more extensive evaluation for CFLD.

Back to Top | Article Outline

References

1. Efrati O, Barak A, Modan-Moses D, et al. Liver cirrhosis and portal hypertension in cystic fibrosis. Eur J Gastroenterol Hepatol 2003; 15:1073–1078.
2. Flass T, Narkewicz MR. Cirrhosis and other liver disease in cystic fibrosis. J Cyst Fibros 2013; 12:116–124.
3. Regev A, Berho M, Jeffers LJ, et al. Sampling error and intraobserver variation in liver biopsy in patients with chronic HCV infection. Am J Gastroenterol 2002; 97:2614–2618.
4. Bedossa P, Dargere D, Paradis V. Sampling variability of liver fibrosis in chronic hepatitis C. Hepatology 2003; 38:1449–1457.
5. Westin J, Lagging LM, Wejstal R, et al. Interobserver study of liver histopathology using the Ishak score in patients with chronic hepatitis C virus infection. Liver 1999; 19:183–187.
6. Lindblad A, Glaumann H, Strandvik B. Natural history of liver disease in cystic fibrosis. Hepatology 1999; 30:1151–1158.
7. Lewindon PJ, Shepherd RW, Walsh MJ, et al. Importance of hepatic fibrosis in cystic fibrosis and the predictive value of liver biopsy. Hepatology 2011; 53:193–201.
8. Sandrin L, Fourquet B, Hasquenoph JM, et al. Transient elastography: a new noninvasive method for assessment of hepatic fibrosis. Ultrasound Med Biol 2003; 29:1705–1713.
9. Laharie D, Zerbib F, Adhoute X, et al. Diagnosis of liver fibrosis by transient elastography (FibroScan) and non-invasive methods in Crohn's disease patients treated with methotrexate. Aliment Pharmacol Ther 2006; 23:1621–1628.
10. 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.
11. Castera L, Forns X, Alberti A. Non-invasive evaluation of liver fibrosis using transient elastography. J Hepatol 2008; 48:835–847.
12. Coco B, Oliveri F, Maina AM, et al. Transient elastography: a new surrogate marker of liver fibrosis influenced by major changes of transaminases. J Viral Hepat 2007; 14:360–369.
13. 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.
14. Grummer-Strawn LM, Reinold C, Krebs NF, et al. Use of World Health Organization and CDC growth charts for children aged 0-59 months in the United States. MMWR Recomm Rep 2010; 59:1–15.
15. Debray D, Kelly D, Houwen R, et al. Best practice guidance for the diagnosis and management of cystic fibrosis-associated liver disease. J Cyst Fibros 2011; 10 (suppl 2):S29–S36.
16. Megremis SD, Vlachonikolis IG, Tsilimigaki AM. Spleen length in childhood with US: normal values based on age, sex, and somatometric parameters. Radiology 2004; 231:129–134.
17. Blom G. Statistical Estimates and Transformed Beta Variables. New York, NY:John Wiley & Sons, Inc; 1958.
18. Bauer P. Multiple testing in clinical trials. Stat Med 1991; 10:871–889. discussion 89–90.
19. Zou KH, Yu CR, Liu K, et al. Optimal thresholds by maximizing or minimizing various metrics via ROC-type analysis. Acad Radiol 2013; 20:807–815.
20. Lee CK, Perez-Atayde AR, Mitchell PD, et al. Serum biomarkers and transient elastography as predictors of advanced liver fibrosis in a United States cohort: the Boston children's hospital experience. J Pediatr 2013; 163:1058.e2–1064.e2.
21. Sokol RJ, Durie PR. Recommendations for management of liver and biliary tract disease in cystic fibrosis. Cystic Fibrosis Foundation Hepatobiliary Disease Consensus Group. J Pediatr Gastroenterol Nutr 1999; 28 (suppl 1):S1–S13.
22. Menten R, Leonard A, Clapuyt P, et al. Transient elastography in patients with cystic fibrosis. Pediatr Radiol 2010; 40:1231–1235.
23. Friedrich-Rust M, Schlueter N, Smaczny C, et al. Non-invasive measurement of liver and pancreas fibrosis in patients with cystic fibrosis. J Cyst Fibros 2013; 12:431–439.
24. Engelmann G, Gebhardt C, Wenning D, et al. Feasibility study and control values of transient elastography in healthy children. Eur J Pediatr 2012; 171:353–360.
25. Bedossa P, Poynard T. An algorithm for the grading of activity in chronic hepatitis C. The METAVIR Cooperative Study Group. Hepatology 1996; 24:289–293.
26. McGoogan KE, Smith PB, Choi SS, et al. Performance of the AST-to-platelet ratio index as a noninvasive marker of fibrosis in pediatric patients with chronic viral hepatitis. J Pediatr Gastroenterol Nutr 2010; 50:344–346.
27. Kitson MT, Kemp WW, Iser DM, et al. Utility of transient elastography in the non-invasive evaluation of cystic fibrosis liver disease. Liver Int 2013; 33:698–705.
28. Karlas T, Neuschulz M, Oltmanns A, et al. Non-invasive evaluation of cystic fibrosis related liver disease in adults with ARFI, transient elastography and different fibrosis scores. PLoS One 2012; 7:e42139.
29. Leung DH, Khan M, Minard CG, et al. Aspartate aminotransferase to platelet ratio and fibrosis-4 as biomarkers in biopsy-validated pediatric cystic fibrosis liver disease. Hepatology 2015; 62:1576–1583.
Keywords:

liver disease; liver fibrosis; pediatric

© 2017 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,