Repeating measurements by transient elastography in nonalcoholic fatty liver disease patients with high liver stiffness

bs_bs_banner doi:10.1111/jgh.14311 HEPATOLOGY Repeating measurements by transient elastography in nonalcoholic fatty liver disease patients with high liver stiffness Jeremy Chak-Lun Chow,* Grace Lai-Hung Wong,*, Anthony Wing-Hung Chan, Sally She-Ting Shu,*, Carmen Ka-Man Chan,*, Julie Ka-Yu Leung,*, Paul Cheung-Lung Choi, Angel Mei-Ling Chim,*, Henry Lik-Yuen Chan*, and Vincent Wai-Sun Wong*, *Department of Medicine and Therapeutics, State Key laboratory of Digestive Disease and Department of Anatomical and Cellular Pathology, The Chinese University of Hong Kong, Shatin, Hong Kong Key words FibroScan, liver biopsy, liver fibrosis, liver stiffness measurement, non-alcoholic steatohepatitis. Accepted for publication 1 June 2018. Correspondence Vincent Wong, Department of Medicine and Therapeutics, 9/F, Clinical Sciences Building, Prince of Wales Hospital, 30-32 Ngan Shing Street, Shatin, Hong Kong. Email: wongv@cuhk.edu.hk Declaration of conflict of interest: Grace Wong, Henry Chan, and Vincent Wong have served as speakers for Echosens. The other authors report no potential conflict of interest. Author contributions: Vincent Wong was responsible for the study design. Grace Wong, Anthony Chan, Sally Shu, Carmen Chan, Julie Leung, Paul Choi, and Angel Chim were responsible for the collection of data. Jeremy Chow and Vincent Wong were responsible for data analysis. Henry Chan was responsible for the administrative support. Jeremy Chow and Vincent Wong were responsible for drafting the manuscript. All authors read and approved the final version of the article, including the authorship list. Abstract Background and Aim: The Baveno VI Consensus recommends repeating examination in patients with high liver stiffness measurement (LSM) by transient elastography to reduce false-positive diagnosis of advanced liver disease. We tested whether repeating transient elastography can increase the overall diagnostic accuracy. Methods: Ninety-seven patients with non-alcoholic fatty liver disease who underwent two FibroScan examinations within 6 months prior to liver biopsy were evaluated. An LSM cut-off of 7.9 kpa was used to exclude F3-4 fibrosis. Results: Seventy-eight patients had high LSM at baseline, among whom 27 had low LSM on repeated testing; only four had F3 and none had cirrhosis. In contrast, 31 of 51 patients with high LSM at both examinations had F3-4. Nineteen patients had low LSM at baseline; none of them had F3-4 regardless of the second LSM results. If we took LSM <7.9 kpa at either examination as sufficient to exclude F3-4, the negative predictive value remained high at 91%. The positive predictive value for F3-4 increased from 45% in patients with high LSM at baseline to 61% in those with high LSM at both examinations. Sensitivity analysis using different cut-offs yielded similar results, with 76% of patients with LSM >12 kpa at both examinations having F3-4. Conclusions: Transient elastography is a highly sensitive screening test to exclude F3-4 fibrosis in non-alcoholic fatty liver disease patients. One-third of patients with high LSM may have normal results on repeated examination. By repeating examination in cases with high LSM, one may spare patients from unnecessary liver biopsy. Introduction Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease worldwide. 1,2 Depending on the histological severity, NAFLD may progress to cirrhosis and hepatocellular carcinoma. 3,4 In reality, these only constitute a minority of NAFLD patients, as most patients with simple steatosis have excellent prognosis and are not at risk of hepatic complications. 5 Among the various histological features, the degree of fibrosis is the strongest predictor of overall and liver-related mortality. 6,7 Traditionally, liver biopsy has been the first choice in assessing liver fibrosis, but its invasiveness, low patient acceptance, and risk of bleeding complications limit its broader application. 8 It is simply not feasible nor effective to perform liver biopsy for a disease that affects one-quarter to one-third of the general population. Current guidelines recommend the use of non-invasive tests of liver fibrosis as initial assessment of NAFLD. 8 10 Liver stiffness measurement (LSM) by transient elastography is one of the commonly used non-invasive tests. It can be performed as a point-of-care test, can estimate hepatic steatosis and fibrosis simultaneously, and has superior accuracy in diagnosing advanced fibrosis than generic fibrosis prediction scores. 11,12 When applied to patients with NAFLD, transient elastography showed excellent sensitivity and negative predictive value (NPV), allowing the exclusion of advanced fibrosis. 13 19 However, its limitations lie in its inability to confirm the diagnosis of advanced disease due to its low positive predictive value (PPV). To confirm advanced fibrosis in patients with high liver stiffness, one may proceed directly to liver biopsy or perform another non-invasive test of fibrosis based on a different mechanism. 20 Alternatively, the Journal of Gastroenterology and Hepatology 34 (2019) 241 248 241

Baveno VI Consensus suggests the possibility of doing two LSM on different days in fasting conditions to reduce the risk of falsepositive results. 21 Though it makes sense, the approach has not been systematically evaluated. Therefore, we evaluated NAFLD patients who underwent liver biopsy and two separate transient elastography examinations and determined the role of this approach. Materials and methods Study design. This was a post-hoc analysis of a prospectively recruited cohort. In this study, we aimed to evaluate the effectiveness of repeated transient elastography examinations for the diagnosis of fibrosis and cirrhosis in NAFLD patients. At our center, all patients underwent LSM within 1 week before liver biopsy. In some patients, the decision to proceed to liver biopsy was prompted by LSM results; such patients thus had two transient elastography examinations performed on different days. In this study, we selected patients with baseline transient elastography within 6 months before liver biopsy because major changes in fibrosis stage are uncommon in such a short interval. All patients were adults with liver biopsy showing features of NAFLD (steatosis involving at least 5% of hepatocytes). We excluded patients with excessive alcohol consumption (30 g/day in men and 20 g/day in women), positive hepatitis B surface antigen or antihepatitis C virus antibody, histological features of other liver diseases, secondary causes of hepatic steatosis (e.g. use of systemic corticosteroids, methotrexate, or tamoxifen), history of malignancy including hepatocellular carcinoma, and prior liver resection or transplantation. We further excluded patients with less than 10 valid LSM acquisitions or an interquartile range-to-median ratio of LSM of >0.30. All patients gave informed written consent. The study protocol was approved by the Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee. Clinical assessment. Comprehensive clinical assessments were performed within 1 week before liver biopsy. Past medical history, smoking, and alcohol consumption were recorded with a standard questionnaire. We recorded anthropometric measurements including bodyweight, body height, and waist circumference. Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared. Waist circumference was measured at a level midway between the lower rib margin and iliac crest with the tape all around the body in the horizontal position. A fasting venous blood sample was taken for liver biochemistry, fasting glucose, hemoglobin A 1c, lipid profile, complete blood count, and international normalized ratio of the prothrombin time. Liver biopsy. In this study, liver histology served as the reference standard for evaluating the diagnostic accuracy of transient elastography. Percutaneous liver biopsy was performed using 16G Temno needles. Liver histology was assessed by two experienced histopathologists who were blinded to the clinical data (A. W. H. C. and P. C. L. C.). In case of discrepant results, the pathologists reviewed the slides together and came to a consensus. Liver specimens shorter than 15 mm were excluded. We scored liver histology using the non-alcoholic steatohepatitis (NASH) Clinical Research Network system. 22 Steatosis was graded from 1 to 3 based on the percentage of hepatocytes with steatosis: 1=5 33%, 2 = > 33 66%, 3 = > 66%. Fibrosis was staged from 0 to 4: 0 = absence of fibrosis; 1 = perisinusoidal or portal fibrosis; 2 = perisinusoidal and portal/periportal fibrosis; 3 = septal or bridging fibrosis; and 4 = cirrhosis. NASH was defined as the presence of 5% hepatic steatosis and inflammation with hepatocyte injury (i.e. ballooning), with or without any fibrosis. 9 Transient elastography. Liver stiffness measurement was performed using transient elastography (FibroScan 502 Touch, Echosens, Paris, France) according to the instructions and training provided by the manufacturer. All patients fasted for at least 6 h before LSM. Measurements were performed using the M probe on the right lobe of the liver through intercostal spaces with the patient lying in dorsal decubitus with the right arm in maximal abduction. Ten successful acquisitions were performed on each patient. The median value represented the liver elastic modulus. Only cases with 10 successful acquisitions were evaluated. The liver stiffness was expressed in kilopascal (kpa). The operators were blinded to all clinical data and the diagnosis of the patients. Based on validation in an earlier cohort, we adopted an LSM cut-off of 7.9 kpa to exclude F3 fibrosis with >90% sensitivity and NPV. 14 We also performed sensitivity analysis using a range of other published cut-offs. Statistical analysis. All statistical tests were performed using IBM SPSS Statistics 22.0 (IBM, Armonk, NY, USA). Continuous variables were expressed as mean ± standard deviation or median (interquartile range) as appropriate. The paired t test was performed to evaluate changes in LSM. The area under the receiver-operating characteristics curves was calculated to reflect the overall accuracy of LSM in diagnosing F3-4 fibrosis. Chisquared test was performed to assess differences in fibrosis stage distribution among subgroups. A two-sided P value of <0.05 was taken as statistically significant. Results From July 2006 to July 2017, 534 patients underwent liver biopsy and LSM for suspected NAFLD. Out of this cohort, 117 patients underwent repeated LSM within 6 months. Ninety-seven patients had confirmed NAFLD, valid LSM and satisfactory liver biopsy samples, and were included in the final analysis (Fig. 1). The mean age of this patient cohort was 55 ± 10. Seventy-six (78%) patients had type 2 diabetes, and 61 (63%) patients had hypertension. The mean BMI and waist circumference were 28.3 ± 3.6 kg/m 2 and 97.6 ± 11.4 cm, respectively. Thirty-five (36%) patients had F3-4 fibrosis. Forty-eight (49%) patients had NASH (Table 1). Variability of LSM. Out of the 97 patients included in the final analysis, 19 patients had low LSM results (<7.9 kpa) and 78 patients showed high LSM results (>7.9 kpa) during their first examination. Based on the results of two transient elastography examinations, we classified the patients into four groups; namely, Low-Low (low LSM results at both examinations; n = 13), Low- 242 Journal of Gastroenterology and Hepatology 34 (2019) 241 248

Figure 1 Study flow diagram. LSM, liver stiffness measurement; NAFLD, non-alcoholic fatty liver disease. [Color figure can be viewed at wileyonlinelibrary.com] High (1st examination low LSM results, and 2nd examination high LSM results; n = 6), High-Low (1st examination high LSM results, and 2nd examination low LSM results; n = 27), High-High (high LSM results at both examinations; n = 51) (Fig. 1). BMI was at or above 30 kg/m 2 in 0% of patients in the Low-Low group, 33% in the Low-High group, 15% in the High-Low group, and 43% in the High-High group (P = 0.005). There was no significant weight reduction in the High-Low group between the two visits (bodyweight 72.7 ± 11.5 kg at the first and 72.4 ± 11.0 kg at the second examination; mean change 0.3 ± 1.4 kg; P = 0.23). None of the patients was added on drugs with potential effect on NAFLD/NASH (e.g. vitamin E, pioglitazone, and obeticholic acid). Figure 2 shows the changes in LSM at the two examinations. Among the 19 patients with low LSM at the first examination, the LSM was 6.3 ± 1.2 kpa at the first examination and 7.3 ± 3.2 kpa at the second examination (mean change 1.0 ± 2.7 kpa; 95% confidence interval [CI] 0.2 to 2.3 kpa; P = 0.11). Ten of the 19 (53%) patients had a higher LSM at the second examination, with 6 (32%) exceeding the cut-off of 7.9 kpa. Among 78 patients with high LSM at the first examination, the LSM was 16.0 ± 9.2 kpa at the first examination and 11.3 ± 5.2 kpa at the second examination (mean change 4.7 ± 6.9 kpa; 95% CI 3.2 to 6.3 kpa; P < 0.001). Sixtyseven of 78 (86%) patients had a lower LSM at the second examination, with 27 (35%) falling below the cut-off of 7.9 kpa. Compared with patients with high LSM at both examinations, those with high baseline LSM but low follow-up LSM had lower BMI, waist circumference, serum alanine aminotransferase level, and higher platelet count (Table S1). In keeping with less severe disease, patients in the High-Low group also had less severe steatosis, hepatocyte ballooning and fibrosis, and were thus less likely to have NASH. On the other hand, there was no significant difference in the interquartile range (IQR-to-median) ratios of LSM at both examinations between the two groups suggestive of a difference in the quality of transient elastography examination. Correlation with fibrosis stage. The area under the receiver-operating characteristics curves of LSM in diagnosing F3-4 fibrosis was 0.81 (95% CI 0.72 0.89) at the first assessment and 0.81 (95% CI 0.72 0.91) at the second assessment. There was a gradated increase in the proportion of patients with high fibrosis stages across the four LSM groups (Fig. 3; P < 0.001 for trend). In both the Low-Low and Low-High groups, none of the patients had F3-4 fibrosis. In particular, all but one patient in the Low-Low group had F0-1 fibrosis. In contrast, 31 of 51 (61%) patients in the High-High group had F3-4 fibrosis. Interestingly, among 27 patients in the High-Low group, only 4 (15%) had F3 fibrosis and none had cirrhosis. In the High-Low group, the LSM and its change were similar between patients with F0-2 and F3 fibrosis; the baseline LSM were 12.1 ± 3.2 kpa and 13.2 ± 2.9 kpa (P = 0.53), second LSM were 6.7 ± 1.0 kpa and 6.0 ± 1.3 kpa (P = 0.21), and changes in LSM from baseline to second examination were 5.5 ± 3.1 kpa and 7.3 ± 2.2 kpa (P = 0.28), respectively. Utility of repeated transient elastography examination. Consistent with previous reports, LSM by transient elastography had a high NPV in excluding F3-4 fibrosis. At the first transient elastography examination, 19 patients had LSM <7.9 kpa and none of them had F3-4 fibrosis (NPV 100%; Fig. 4). However, only 35 of 78 patients with LSM 7.9 kpa had F3-4 Journal of Gastroenterology and Hepatology 34 (2019) 241 248 243

Table 1 Clinical characteristics of 97 NAFLD patients Age (years) 55 ± 10 Male gender, n (%) 51 (53) Type 2 diabetes, n (%) 76 (78) Hypertension, n (%) 61 (63) Body mass index (kg/m 2 ) 28.3 ± 3.6 Waist circumference (cm) 97.6 ± 11.4 Alanine aminotransferase (IU/L) 58 ± 37 Aspartate aminotransferase (IU/L) 44 ± 29 Fasting glucose (mmol/l) 7.1 ± 2.2 Hemoglobin A 1c (%) 7.1 ± 1.4 Total cholesterol (mmol/l) 4.4 ± 1.1 HDL-cholesterol (mmol/l) 1.2 ± 0.5 LDL-cholesterol (mmol/l) 2.5 ± 1.0 Triglycerides (mmol/l) 1.9 ± 1.3 Platelet count ( 10 9 /L) 226 ± 61 INR 0.95 ± 0.06 Histological severity Length of liver sample (cm) 2.1 ± 0.6 Steatosis grade (1/2/3) 35/46/16 Lobular inflammation (0/1/2/3) 9/47/40/1 Hepatocyte ballooning (0/1/2) 42/50/5 Fibrosis stage (0/1/2/3/4) 15/31/16/20/15 NAFLD activity score, median (IQR) 4 (3 5) NASH, n (%) 48 (49) Interval between two transient elastography examinations (weeks) 18 ± 8 18 (14 23) Transient elastography 1st transient elastography examination Median LSM (kpa) 14.1 ± 9.1 IQR of LSM (kpa) 2.5 ± 2.3 IQR-to-median ratio of LSM (%) 17.2 ± 7.1 2nd transient elastography examination Median LSM (kpa) 10.5 ± 5.1 IQR of LSM (kpa) 1.6 ± 1.2 IQR-to-median ratio of LSM (%) 14.3 ± 6.8 Continuous variables are expressed as mean ± standard deviation or median (interquartile range). HDL, high density lipoprotein; INR, international normalized ratio; LDL, low density lipoprotein; LSM, liver stiffness measurement; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis. fibrosis, resulting in a PPV of 45%. At the second transient elastography examination, 40 patients had LSM <7.9 kpa and four of them had F3-4 fibrosis. The sensitivity, specificity, PPV, and NPV for the second examination to diagnose F3-4 fibrosis were 89%, 58%, 54%, and 90%, respectively. On the other hand, if we take low LSM <7.9 kpa at either of the two examinations as a negative result, 46 patients would be classified as having a low risk of advanced disease, among whom 42 indeed had F0-2 fibrosis (NPV 91%; Fig. 4). Fifty-one patients had LSM 7.9 kpa at both examinations (HIGH-HIGH group), among whom 31 had F3-4 fibrosis (PPV 61%). Thus, repeating transient elastography in patients with high LSM would spare more patients from liver biopsy without jeopardizing the high NPV of the examination. The overall accuracy in diagnosing or excluding F3-4 fibrosis increased from 56% (95% CI 45 66%) for baseline LSM alone to 75% (95% CI 66 84%) for the approach of repeating examination in patients with high LSM at baseline. Sensitivity analysis. Different studies have proposed different LSM cut-offs for the diagnosis of F3-4 fibrosis in NAFLD patients. We therefore repeated the analysis using different cut-offs. Across LSM cut-offs from 7.9 to 12.0 kpa, consistently 35 50% of patients with high LSM at the first examination would be reclassified as low risk on repeated examination (Table 2). As expected, the higher the cut-off, the more likely patients in both the High-Low and High-High groups would have F3-4 fibrosis. However, even at 12.0 kpa, only a minority of patients in the High-Low group had F3-4 fibrosis. Discussion We set out to evaluate the effectiveness of the repeated LSM approach recommended by the Baveno VI Consensus, 21 and whether it can address the issue of low PPV while maintaining its strengths of high NPV and sensitivity. Before one can follow this approach, we need to determine (i) the proportion of patients with high LSM who would have lower measurements on repeated assessment and (ii) the fibrosis stage of such patients. Echoing previous studies, we showed that LSM had excellent sensitivity and NPV, allowing the exclusion of advanced fibrosis in NAFLD patients. 15,18,19 Additionally, the PPV was also low at 45%, limiting its ability to confirm the diagnosis of advanced fibrosis. In this study, none of our patients had obvious causes of false-positive LSM like acute hepatitis, extrahepatic cholestasis, and congestive heart failure. 23 The patients were also fasted to avoid increased LSM with food intake. Nonetheless, high BMI and severe hepatic steatosis are associated with higher LSM, and both factors are common in NAFLD patients. 24 27 The mechanism underlying the association between obesity and increased LSM is unclear, but one possible explanation is interference by the ribs as it is more difficult to locate the intercostal space in obese patients. Because both the degree of hepatic steatosis and the choice of the optimal site for LSM may change on another occasion, this serves as the basis for repeating LSM in patients with high liver stiffness values. In our study, none of the patients in the Low- Low group had BMI 30 kg/m 2. This may explain why they did not have false-positive LSM. Indeed, when we repeated LSM on another date, one-third of the patients with high LSM at first had a lower LSM suggestive of a low likelihood of F3-4 fibrosis. Importantly, 85% of those patients indeed did not have F3-4 fibrosis. On the other hand, if we only perform liver biopsy in patients with high LSM at both examinations, the PPV increased from 45% to 61% for F3-4 disease. Considering the availability and convenience of transient elastography, the strengths of this approach greatly outweigh its limitations. By utilizing different cut-offs, we could further increase the confidence in diagnosing F3-4 fibrosis, with a PPV of 76% when LSM exceeded 12.0 kpa on both occasions (Table 2). One possible explanation of the mild fibrosis and decreased LSM in the High-Low group is that those patients might change their lifestyle and therefore achieve fibrosis regression in between visits. While we cannot exclude this possibility, the mean interval between the two examinations was only 18 weeks, and there was 244 Journal of Gastroenterology and Hepatology 34 (2019) 241 248

Figure 2 LSM change during repeated examinations. (a) Individual changes in patients with baseline LSM <7.9 kpa. (b) Box plot in patients with baseline LSM <7.9 kpa. (c) Individual changes in patients with baseline LSM 7.9 kpa. (d) Box plot in patients with baseline LSM 7.9 kpa. LSM, liver stiffness measurement. [Color figure can be viewed at wileyonlinelibrary.com] Figure 3 Distribution of fibrosis stages by repeated LSM examinations. Low: LSM <7.9 kpa; High: LSM 7.9 kpa. LSM, liver stiffness measurement. Fibrosis stage:, 4;, 3;, 2;, 1;, 0. [Color figure can be viewed at wileyonlinelibrary.com] no significant weight reduction during this period. Besides, only around 10% of NASH patients in drug trials would have spontaneous fibrosis improvement in 1 year. 28 Even with intensive lifestyle intervention, the average decline in LSM was only around 0.5 kpa in a year. 29 We therefore believe that false-positive LSM in obese patients rather than spontaneous fibrosis regression was the main reason for our observation. When performed at the same session, transient elastography is highly reproducible with excellent intraobserver and interobserver agreement. 19,30 However, when the operator places the probe at different positions, a discordance in the predicted fibrosis stage may occur in up to 30% of cases. 31 In another study of 48 inactive carriers of hepatitis B virus, while the median LSM was stable over 10 35 months at three visits, some patients had marked decrease in LSM at the individual level. 32 Because LSM is technically more difficult in obese patients, the proportion of patients with decreased LSM on repeated testing was higher in our NAFLD cohort. Another approach to improve the PPV for diagnosing F3-4 fibrosis is to combine LSM and another non-invasive tests of fibrosis such as the NAFLD fibrosis test or FibroMeter. 20,33,34 When we applied LSM and FibroMeter to our NAFLD patients in another study, the PPV for F3-4 fibrosis increased from 61% for LSM alone to 89% when both tests were positive. 20 Furthermore, magnetic resonance elastography has shown superiority over transient elastography in NAFLD patients, but cost and availability remain obstacles to wider application of the former technique. 35,36 Because aspartate aminotransferase (AST) is not a routine test at our hospital, we could not evaluate the concomitant variability of prediction models of fibrosis such as the NAFLD fibrosis score, Journal of Gastroenterology and Hepatology 34 (2019) 241 248 245

Figure 4 Utility and performance of the repeated LSM approach. Because of high negative predictive value, a low baseline LSM is sufficient to exclude F3-4 fibrosis; it is unnecessary to repeat the examination. In contrast, one-third of patients with high LSM at baseline had normal results on repeated testing. Those patients may be considered as having low risk of F3-4 fibrosis. LSM, liver stiffness measurement. [Color figure can be viewed at wileyonlinelibrary.com] Table 2 Sensitivity analysis of repeated LSM examinations based on different cut-offs Number of patients Cut-off (kpa) Low-Low Low-High High-Low High-High Percentage with F3-4 in High-Low group (%) Percentage with F3-4 in High-High group (%) 7.9 13 6 27 51 15 61 9.0 23 5 26 43 23 65 10.0 25 5 29 38 24 68 11.0 28 4 32 33 31 70 12.0 43 4 25 25 36 76 aspartate aminotransferase-to-platelet ratio index, and the Fibrosis- 4 score. Nonetheless, these prediction models have poor association with each other, 37 are confounded by age, 38 and have inferior diagnostic performance than transient elastography. 14 Changes in these models also correlate poorly with histological changes over time. 3 Our study is unique in its clinicopathological correlation between two transient elastography examinations and liver biopsy. 246 Journal of Gastroenterology and Hepatology 34 (2019) 241 248

Nonetheless, it also has several limitations. First, liver biopsy is an imperfect reference standard with sampling variability. Nonetheless, liver biopsy remains the most acceptable reference standard, and biopsy specimens were examined by two experienced pathologists. We also minimized bias by using good quality samples. Second, patients were recruited from a referral center. However, as the prevalence of advanced fibrosis is much lower in the primary care setting, this repeated LSM approach will be highly applicable in reducing false-positive results and unnecessary referrals. 39 Third, the sample size of our study was relatively small, but the correlation with liver pathology adds weight to the interpretation of results. Furthermore, all patients were Chinese. The study should be replicated in other patient populations, although previous studies in Caucasians have shown similar diagnostic performance. 15,18,19 Finally, for historical reasons, LSM was performed by the M probe. Our findings should be confirmed in studies using the XL probe. In conclusion, transient elastography is a highly sensitive screening test to exclude F3-4 fibrosis in NAFLD patients. Onethird of patients with high LSM may have normal results on repeated examination with a low risk of F3-4 fibrosis. By repeating examination in case of high LSM, one may spare patients from unnecessary liver biopsy and improve the overall diagnostic accuracy. Acknowledgment This project was partially supported by a grant from the Research Grants Council of the Hong Kong SAR Government (Project no. CUHK 14108916). References 1 Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, Wymer M. Global epidemiology of nonalcoholic fatty liver disease-meta-analytic assessment of prevalence, incidence, and outcomes. Hepatology 2016; 64: 73 84. 2 Fan JG, Kim SU, Wong VW. New trends on obesity and NAFLD in Asia. J. Hepatol. 2017; 67: 862 73. 3 Wong VW, Wong GL, Choi PC et al. Disease progression of nonalcoholic fatty liver disease: a prospective study with paired liver biopsies at 3 years. Gut 2010; 59: 969 74. 4 Younossi ZM, Otgonsuren M, Henry L et al. Association of nonalcoholic fatty liver disease (NAFLD) with hepatocellular carcinoma (HCC) in the United States from 2004 to 2009. Hepatology 2015; 62: 1723 30. 5 Liu K, Wong VW, Lau K et al. Prognostic value of controlled attenuation parameter by transient elastography. Am. J. Gastroenterol. 2017; 112: 1812 23. 6 Ekstedt M, Hagstrom H, Nasr P et al. Fibrosis stage is the strongest predictor for disease-specific mortality in NAFLD after up to 33 years of follow-up. Hepatology 2015; 61: 1547 54. 7 Dulai PS, Singh S, Patel J et al. Increased risk of mortality by fibrosis stage in nonalcoholic fatty liver disease: systematic review and metaanalysis. Hepatology 2017; 65: 1557 65. 8 Wong VW, Chan WK, Chitturi S et al. Asia-Pacific working party on non-alcoholic fatty liver disease guidelines 2017-part 1: definition, risk factors and assessment. J. Gastroenterol. Hepatol. 2018; 33: 70 85. 9 Chalasani N, Younossi Z, Lavine JE et al. The diagnosis and management of nonalcoholic fatty liver disease: practice guidance from the American Association for the Study of Liver Diseases. Hepatology 2018; 67: 328 57. 10 European Association for the Study of the Liver, European Association for the Study of Diabetes, European Association for the Study of Obesity. EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease. J. Hepatol. 2016; 64: 1388 402. 11 Karlas T, Petroff D, Sasso M et al. Individual patient data metaanalysis of controlled attenuation parameter (CAP) technology for assessing steatosis. J. Hepatol. 2017; 66: 1022 30. 12 Younossi ZM, Loomba R, Anstee QM et al. Diagnostic modalities for non-alcoholic fatty liver disease (NAFLD), non-alcoholic steatohepatitis (NASH) and associated fibrosis. Hepatology 2017. https://doi.org/10.1002/hep.29721. [Epub ahead of print] 13 Yoneda M, Yoneda M, Mawatari H et al. Noninvasive assessment of liver fibrosis by measurement of stiffness in patients with nonalcoholic fatty liver disease (NAFLD). Dig. Liver Dis. 2008; 40: 371 8. 14 Wong VW, Vergniol J, Wong GL et al. Diagnosis of fibrosis and cirrhosis using liver stiffness measurement in nonalcoholic fatty liver disease. Hepatology 2010; 51: 454 62. 15 Petta S, Di Marco V, Camma C, Butera G, Cabibi D, Craxi A. Reliability of liver stiffness measurement in non-alcoholic fatty liver disease: the effects of body mass index. Aliment. Pharmacol. Ther. 2011; 33: 1350 60. 16 Wong VW, Vergniol J, Wong GL et al. Liver stiffness measurement using XL probe in patients with nonalcoholic fatty liver disease. Am. J. Gastroenterol. 2012; 107: 1862 71. 17 Kwok R, Tse YK, Wong GL et al. Systematic review with metaanalysis: non-invasive assessment of non-alcoholic fatty liver disease the role of transient elastography and plasma cytokeratin-18 fragments. Aliment. Pharmacol. Ther. 2014; 39: 254 69. 18 Tapper EB, Challies T, Nasser I, Afdhal NH, Lai M. The performance of vibration controlled transient elastography in a US cohort of patients with nonalcoholic fatty liver disease. Am. J. Gastroenterol. 2016; 111: 677 84. 19 Vuppalanchi R, Siddiqui MS, Van Natta ML et al. Performance characteristics of vibration-controlled transient elastography for evaluation of nonalcoholic fatty liver disease. Hepatology 2018; 67: 134 44. 20 Loong TC, Wei JL, Leung JC et al. Application of the combined FibroMeter vibration-controlled transient elastography algorithm in Chinese patients with non-alcoholic fatty liver disease. J. Gastroenterol. Hepatol. 2017; 32: 1363 9. 21 de Franchis R, Baveno VI Faculty. Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: stratifying risk and individualizing care for portal hypertension. J. Hepatol. 2015; 63: 743 52. 22 Kleiner DE, Brunt EM, Van Natta M et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005; 41: 1313 21. 23 Wong VW, Chan HL. Transient elastography. J. Gastroenterol. Hepatol. 2010; 25: 1726 31. 24 Das K, Sarkar R, Ahmed SM et al. "Normal" liver stiffness measure (LSM) values are higher in both lean and obese individuals: a population-based study from a developing country. Hepatology 2012; 55: 584 93. 25 Wong GL, Chan HL, Choi PC et al. Association between anthropometric parameters and measurements of liver stiffness by transient elastography. Clin. Gastroenterol. Hepatol. 2013; 11: 295 302. e1 3. 26 Petta S, Maida M, Macaluso FS et al. The severity of steatosis influences liver stiffness measurement in patients with nonalcoholic fatty liver disease. Hepatology 2015; 62: 1101 10. Journal of Gastroenterology and Hepatology 34 (2019) 241 248 247

27 Petta S, Wong VW, Camma C et al. Improved noninvasive prediction of liver fibrosis by liver stiffness measurement in patients with nonalcoholic fatty liver disease accounting for controlled attenuation parameter values. Hepatology 2017; 65: 1145 55. 28 Friedman SL, Ratziu V, Harrison SA et al. A randomized, placebocontrolled trial of cenicriviroc for treatment of nonalcoholic steatohepatitis with fibrosis. Hepatology 2018; 67: 1754 67. 29 Wong VW, Chan RS, Wong GL et al. Community-based lifestyle modification programme for non-alcoholic fatty liver disease: a randomized controlled trial. J. Hepatol. 2013; 59: 536 42. 30 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 73. 31 Zelber-Sagi S, Yeshua H, Shlomai A et al. Sampling variability of transient elastography according to probe location. Eur. J. Gastroenterol. Hepatol. 2011; 23: 507 14. 32 Castera L, Bernard PH, Le Bail B et al. Transient elastography and biomarkers for liver fibrosis assessment and follow-up of inactive hepatitis B carriers. Aliment. Pharmacol. Ther. 2011; 33: 455 65. 33 Chan WK, Nik Mustapha NR, Mahadeva S. A novel 2-step approach combining the NAFLD fibrosis score and liver stiffness measurement for predicting advanced fibrosis. Hepatol. Int. 2015; 9: 594 602. 34 Boursier J, Vergniol J, Guillet A et al. Diagnostic accuracy and prognostic significance of blood fibrosis tests and liver stiffness measurement by FibroScan in non-alcoholic fatty liver disease. J. Hepatol. 2016; 65: 570 8. 35 Imajo K, Kessoku T, Honda Y et al. Magnetic resonance imaging more accurately classifies steatosis and fibrosis in patients with nonalcoholic fatty liver disease than transient elastography. Gastroenterology 2016; 150: 626626 NaN. 36 Park CC, Nguyen P, Hernandez C et al. Magnetic resonance elastography vs transient elastography in detection of fibrosis and noninvasive measurement of steatosis in patients with biopsy-proven nonalcoholic fatty liver disease. Gastroenterology 2017; 152: 598598 NaN. 37 Morling JR, Fallowfield JA, Guha IN et al. Using non-invasive biomarkers to identify hepatic fibrosis in people with type 2 diabetes mellitus: the Edinburgh type 2 diabetes study. J. Hepatol. 2014; 60: 384 91. 38 McPherson S, Hardy T, Dufour JF et al. Age as a confounding factor for the accurate non-invasive diagnosis of advanced NAFLD fibrosis. Am. J. Gastroenterol. 2017; 112: 740 51. 39 Mahady SE, Macaskill P, Craig JC et al. Diagnostic accuracy of noninvasive fibrosis scores in a population of individuals with a low prevalence of fibrosis. Clin. Gastroenterol. Hepatol. 2017; 15: 14531453 NaN. Supporting information Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. Clinical characteristics of patients in the High-Low (LSM >7.9 kpa at baseline and 7.9 kpa on follow-up) and High-High (LSM >7.9 kpa at both examinations) groups. 248 Journal of Gastroenterology and Hepatology 34 (2019) 241 248