Nonalcoholic fatty liver disease (NAFLD) is

Transcription

1 AMERICAN ASSOCIATION FOR THE STUDY OFLIVERD I S E ASES HEPATOLOGY, VOL. 65, NO. 4, 2017 Improved Noninvasive Prediction of Liver Fibrosis by Liver Stiffness Measurement in Patients With Nonalcoholic Fatty Liver Disease Accounting for Controlled Attenuation Parameter Values Salvatore Petta, 1 Vincent Wai-Sun Wong, 2,3 Calogero Camma, 1 Jean-Baptiste Hiriart, 4 Grace Lai-Hung Wong, 2,3 Fabio Marra, 5 Julien Vergniol, 4 Anthony Wing-Hung Chan, 6 Vito Di Marco, 1 Wassil Merrouche, 4 Henry Lik-Yuen Chan, 2,3 Marco Barbara, 1 Brigitte Le-Bail, 7,8 Umberto Arena, 5 Antonio Craxı, 1 and Victor de Ledinghen 4,7 Liver stiffness measurement (LSM) frequently overestimates the severity of liver fibrosis in nonalcoholic fatty liver disease (NAFLD). Controlled attenuation parameter (CAP) is a new parameter provided by the same machine used for LSM and associated with both steatosis and body mass index, the two factors mostlyaffectinglsmperformanceinnafld.weaimedtodetermine whether prediction of liver fibrosis by LSM in NAFLD patients is affected by CAP values. Patients (n 5 324) were assessed by clinical and histological (Kleiner score) features. LSM and CAP were performed using the M probe. CAP values were grouped by tertiles (lower , middle , higher db/m). Among patients with F0-F2 fibrosis, mean LSM values, expressed in kilopascals, increased according to CAP tertiles (6.8 versus 8.6 versus 9.4, P ), and along this line the area under the curve of LSM for the diagnosis of F3-F4 fibrosis was progressively reduced from lower to middle and further to higher CAP tertiles (0.915, ; 0.830, ; 0.806, ). As a consequence, in subjects with F0-F2 fibrosis, the rates of false-positive LSM results for F3-F4 fibrosis increased according to CAP tertiles (7.2% in lower versus 16.6% in middle versus 18.1% in higher). Consistent with this, a decisional flowchart for predicting fibrosis was suggested by combining both LSM and CAP values. Conclusions: In patients with NAFLD, CAP values should always be taken into account in order to avoid overestimations of liver fibrosis assessed by transient elastography. (HEPATOLOGY 2017;65: ). Nonalcoholic fatty liver disease (NAFLD) is regarded as the most relevant liver disease of the twenty-first century. This picture makes NAFLD as the most common cause of chronic liver disease (1) as a dramatically growing risk factor for hepatocellular carcinoma, (2) as a leading indication for liver transplantation, (3) and as a condition leaving individuals at increased risk of extrahepatic mostly cardiovascular morbidity and mortality. (4,5) Due to the high prevalence of NAFLD, it is critical to identify those individuals at higher risk of developing both hepatic and extrahepatic complications. Two prospective cohort studies of the natural history of NAFLD patients have clearly shown that the severity of liver fibrosis is the stronger predictor not only of liver-related complications but also of important extrahepatic diseases, including cardiovascular disease and extrahepatic malignancies. (4,5) Accordingly, the evaluation of liver fibrosis severity has Abbreviations: AUC, area under the curve; BMI, body mass index; CAP, controlled attenuation parameter; CI, confidence interval; IQR, interquartile range; IQR/M, IQR divided by median; LSM, liver stiffness measurement; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; ROC, receiver operating characteristic; TE, transient elastography. Received May 6, 2016; accepted September 8, Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/ /hep.28843/suppinfo. Copyright VC 2016 by the American Association for the Study of Liver Diseases. View this article online at wileyonlinelibrary.com. DOI /hep Potential conflict of interest: Dr. G. Wong, Dr. V. Wong, Dr. H. Chan received lecture fees from Echosens

2 PETTA ET AL. HEPATOLOGY, April 2017 become the main issue to verify the prognosis of NAFLD patients, and it is now mandatory to identify noninvasive tools to replace liver biopsy, which is invasive, has potential life-threatening complications not applicable to the entire NAFLD population, and is biased by interoberserver or intraobserver agreement, sampling errors, and differences among scores. (6) Presently, noninvasive assessment of fibrosis may be conducted using both specific devices and combined clinical-biochemical markers. (7,8) Liver stiffness measurement (LSM) by transient elastography (TE) has been shown to be sufficiently accurate to predict the fibrosis stage in NAFLD patients, (9-12) but its diagnostic performance is strongly affected by obesity and severity of steatosis, which may increase the rates of false-positive results. (12-15) Notably, in the last few years the FibroScan machine has been equipped with a new software able to read the so-called controlled attenuation parameter (CAP) value, which is a reliable noninvasive marker to quantify hepatic steatosis and has been also independently associated with obesity. (16) Consistent with the link between CAP values and both steatosis and body mass index (BMI) the two factors mostly affecting LSM performance in NAFLD we aimed to determine whether prediction of liver fibrosis by LSM in NAFLD patients is affected by CAP values. Materials and Methods PATIENTS The study assessed consecutive patients with biopsyproven NAFLD, recruited at the University Hospital of Pessac in France (143 patients), the Prince of Wales Hospital in Hong Kong (78 patients), the Gastrointestinal and Liver Unit of the University Hospital in Palermo (69 patients), and the clinical hepatology services of the Azienda Ospedaliera Universitaria Careggi in Florence (34 patients), with available and reliable LSM and CAP values. The diagnosis of NAFLD was based on chronically elevated alanine aminotransferase for at least 6 months, alcohol consumption the last year <20 g/day in females and <30 g/day in males, and steatosis (5% of hepatocytes) at histology with or without necroinflammation and/or fibrosis. Exclusion criteria were as follows: (1) advanced cirrhosis (Child-Turcotte-Pugh B and C); (2) hepatocellular carcinoma; (3) other causes of liver disease or mixed etiologies (alcohol abuse, hepatitis C, hepatitis B, autoimmune liver disease, Wilson disease, hemochromatosis or alpha-1-antitrypsin deficiency); (4) human immunodeficiency virus infection; (5) previous treatment with immunosuppressive drugs and/or regular use of steatosis-inducing drugs, evaluated by a questionnaire (for example, corticosteroid, valproic acid, tamoxifen, amiodarone); or (6) active intravenous drug addiction or use of cannabis. The study was performed in accordance with the principles of the Declaration of Helsinki and its appendices and with local and national laws. Approval was obtained from the hospitals internal review boards and ethics committees, and written informed consent was obtained from all patients. CLINICAL AND LABORATORY ASSESSMENT Demographic, clinical, and anthropometric data were collected at the time of liver biopsy. The diagnosis of type 2 diabetes was based on the revised criteria of the American Diabetes Association, using a value of fasting blood glucose 126 mg/dl on at least two occasions. (17) Current therapy with insulin or oral ARTICLE INFORMATION: From the 1 Sezione di Gastroenterologia, Di.Bi.M.I.S., University of Palermo, Palermo, Italy; 2 Department of Medicine and Therapeutics and 3 State Key Laboratory of Digestive Disease, The Chinese University of Hong Kong, Hong Kong; 4 Centre d Investigation de la Fibrose hepatique, H^opital Haut-Lev^eque, Bordeaux University Hospital, Pessac, France; 5 Dipartimento di Medicina Sperimentale e Clinica, Universita degli Studi di Firenze, Florence, Italy; 6 Department of Anatomical and Cellular Pathology, The Chinese University of Hong Kong, Hong Kong; 7 INSERM U1053 and 8 Service de Pathologie, H^opital Pellegrin, Bordeaux University Hospital, Bordeaux, France. ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO: Salvatore Petta, M.D., Ph.D. Section of Gastroenterology and Hepatology, Di.Bi.M.I.S. Policlinico Universitario Paolo Giaccone Piazza delle Cliniche, Palermo, Italy petsa@inwind.it or salvatore.petta@unipa.it Tel:

3 HEPATOLOGY, Vol. 65, No. 4, 2017 PETTA ET AL. hypoglycemic agents was documented in patients with a previous diagnosis of type 2 diabetes. A 12-hour overnight fasting blood sample was drawn at the time of biopsy to determine serum levels of alanine aminotransferase, platelet count, total cholesterol, triglycerides, and blood glucose. LIVER STIFFNESS MEASUREMENT TE was performed with the FibroScan (Echosens, Paris, France) medical device, using the M probe (also called the standard probe). In each center, LSM and CAP were assessed on the same day as liver biopsy, before the procedure and after an overnight fast, by a trained operator who had previously performed at least 300 determinations in patients with chronic liver disease. As recently reported in the literature, (18) we classified all LSM examinations into three reliability categories: very reliable (interquartile range [IQR] divided by median [IQR/M] 0.10), reliable (0.10 < IQR/M 0.30 or IQR/M >0.30 with LSM median <7.1 kpa), and poorly reliable (IQR/M >0.30 with LSM median 7.1 kpa). Only patients with 10 valid measurements were included, and poorly reliable results were excluded from the analysis. HISTOLOGICAL ASSESSMENT In each center, one liver-dedicated expert pathologist, who was unaware of patients identity and history, coded and read histological slides. A minimum 15- mm-long biopsy specimen or the presence of at least 10 complete portal tracts was required. (19) Steatosis was assessed as the percentage of hepatocytes containing fat droplets (minimum 5%) and as a categorical variable. The Kleiner classification (20) was used to stage fibrosis from 0 to 4. Nonalcoholic steatohepatitis (NASH) was considered to be present when steatosis, lobular inflammation, and ballooning were concomitantly present. STATISTICS The results are reported as frequencies and percentage for categorical variables and as mean 6 standard deviation for continuous variables. The Student t test, v 2 test, and analysis of variance were used to compare continuous or categorical variables. Two multiple linear regression models were used to assess the factors independently associated with both CAP and LSM values. The covariates for the multivariate linear analyses were gender, age, BMI 30 kg/m 2, length of biopsy, LSM IQR (only for LSM model), CAP (only for LSM model), CAP IQR (only for CAP model), steatosis, lobular inflammation, ballooning, NASH, and fibrosis. They were chosen as potential confounders, based on their significance in univariate analysis and/or their biological plausibility. In addition, all models were adjusted for enrolling centers. Ballooning, lobular inflammation, and NASH were not entered in the same models due to the issue of collinearity. Receiver operating characteristic (ROC) curves were applied to find the best cutoff values and to identify the area under the ROC curve of LSM able to discriminate the different stages of fibrosis. All analyses were performed using SPSS v statistical package for MacIntosh (SPSS Inc., Chicago, USA). Results PATIENTS The baseline characteristics of the 324 NAFLD patients are shown in Table 1. Mean age was 54.2 years, with a prevalence of females (55%). Forty percent of patients were obese, and diabetes was present in 45% of cases. Mean values for total cholesterol, high-density lipoprotein cholesterol, and triglycerides were within the normal range. At liver biopsy, 35% of patient had fibrosis 3 by Kleiner score. About 1 patient in 3 had steatosis 66%, and features of NASH were recorded in 71% of cases. Supporting Table S1 reports characteristics of patients split according to the enrolling center. FACTORS ASSOCIATED WITH CAP Mean CAP was db/m. Higher CAP values were significantly associated with CAP IQR, BMI 30, steatosis, ballooning, NASH, and severity of fibrosis, even if at multivariate analysis only CAP IQR (P ), BMI 30 (P < 0.001), and steatosis (P < 0.001) were maintained (Table 2). When LSM was included in the model, similar results were reported; and LSM was independently associated with higher CAP values (P ). Finally, when including in the model ballooning instead of NASH, similar results were observed and ballooning was not further associated with CAP (P ). 1147

4 PETTA ET AL. HEPATOLOGY, April 2017 TABLE 1. Demographic, Clinical, Biochemical, and Histological Characteristics of 324 Patients With NAFLD NAFLD Variable (n 5 324) Age, years (mean 6 standard deviation) Male gender 144 (44.4%) BMI 30 kg/m (39.5%) Type 2 diabetes 148 (45.6%) Alanine aminotransferase (IU/L) Platelet count (310 3 /mm 3 ) Cholesterol (mg/dl) Triglycerides (mg/dl) Blood glucose (mg/dl) Liver stiffness (kpa) Stiffness IQR (kpa) CAP (db/m) CAP IQR (db/m) Biopsy specimen length (mm) Histology at biopsy Steatosis (%) Steatosis grade 1 (5%-33%) 86 (26.5%) 2(>33%-66%) 123 (38.0%) 3(>66%) 115 (35.5%) NASH 230 (71%) Stage of fibrosis 0 50 (15.4%) 1 87 (26.9%) 2 72 (22.2%) 3 62 (19.1%) 4 53 (16.4%) Data are given as mean 6 standard deviation, as median and IQR, or as number of cases (%). FACTORS ASSOCIATED WITH LSM AND ACCURACY OF TE FOR FIBROSIS ASSESSMENT Mean LSM was kpa, and 10 patients had IQR/M >0.30 with LSM <7.1 kpa. Liver stiffness values significantly increased according to fibrosis stage: in F , in F , in F , in F , and in F (P < 0.001). Table 3 shows the univariate and multivariate analyses of parameters associated with LSM as continuous variable in the entire cohort and in subgroups with F0-F2 and F3-F4 fibrosis. In the entire cohort higher LSM values were independently linked to higher CAP values (P ), higher LSM IQR values (P < 0.001), and fibrosis (P > 0.001) (Table 3, upper). When both ballooning and lobular inflammation were included in the model instead of NASH, these were not independently associated with LSM (P and P , respectively), while we confirmed the link between LSM and CAP (P ). Notably, when splitting the cohort according to presence or absence of F3-F4 fibrosis, the independent association between LSM and CAP was confirmed in patients with F0-F2 fibrosis (P ) (Table 3, middle) but not in those with F3-F4 fibrosis (P ) (Table 3, lower). ROC curves identified the best cutoffs of LSM able to maximize the accuracy of TE at 8.5 kpa for the diagnosis of F2-F4 (area under the curve [AUC], 0.808; sensitivity, 74.3%; specificity, 73.7%) and at 10.1 kpa for F3-F4 (AUC, 0.861; sensitivity, 77.6%; specificity, 78.4%) fibrosis. Figure 1 summarizes the accuracy of LSM in diagnosing F2-F4 (Fig. 1A) and F3-F4 (Fig. 1B) fibrosis in the entire NAFLD cohort. IMPACT OF CAP ON LSM VALUES AND ON ACCURACY OF TE FOR STAGING FIBROSIS Due to the association of liver stiffness with CAP, we assessed the variations of LSM values and of TE accuracy within the same stages of liver fibrosis and according to CAP tertiles. Supporting Fig. S2 shows Variable TABLE 2. Univariate and Multivariate Analyses of Risk Factors Associated With Controlled Attenuated Parameter as Continuous Variable in 324 Patients With NAFLD by Linear Regression Analysis Univariate Analysis Multivariate Analysis P b Standard error P Age, years Male gender BMI 30 kg/m < <0.001 Biopsy length (mm) CAP IQR (db/m) < Histology at biopsy Steatosis < <0.001 Lobular inflammation Ballooning NASH < Fibrosis

5 HEPATOLOGY, Vol. 65, No. 4, 2017 PETTA ET AL. TABLE 3. Univariate and Multivariate Analyses of Risk Factors Associated With LSM as Continuous Variable in 324 Patients With NAFLD by Linear Regression Analysis Univariate Analysis Multivariate Analysis Variable P b Standard error P Entire NAFLD cohort (n 5 324) Age, years < Male gender BMI 30 (kg/m 2 ) CAP (db/m) Biopsy length (mm) Liver stiffness IQR < <0.001 Histology at biopsy Steatosis Lobular inflammation <0.001 Ballooning <0.001 NASH < Fibrosis < <0.001 F0-F2 NAFLD (n 5 208) Age, years Male gender BMI 30 (kg/m 2 ) CAP (db/m) < Biopsy length (mm) Liver stiffness IQR < <0.001 Histology at biopsy Steatosis Lobular inflammation Ballooning NASH Fibrosis F3-F4 NAFLD (n 5 116) Age, years Male gender BMI 30 (kg/m 2 ) CAP (db/m) Biopsy length (mm) Liver stiffness IQR < <0.001 Histology at biopsy Steatosis Lobular inflammation Ballooning NASH Fibrosis < <0.001 that when patients were considered at each fibrosis stage, for lower grades of fibrosis only, LSM values progressively increased from the lower to the middle and further to the higher CAP tertiles. Along this line, similar results were observed when discriminating patients according to the presence or absence of F2-F4 or F3-F4 fibrosis. Specifically, Fig. 2 shows, in subgroups discriminated on the basis of CAP tertiles, mean LSM values among patients with F0-F1 fibrosis stages (6.5 versus 7.8 versus 8.4, P ; Fig. 2A) and F2-F4 fibrosis stages (13.8 versus 15.2 versus 15.2; P ; Fig. 2B) and among patients with F0-F2 fibrosis stages (6.8 versus 8.6 versus 9.4; P ; Fig. 2C) and F3-F4 fibrosis stages (18.7 versus 18.4 versus 17.0; P ; Fig. 2D). When considering the impact of CAP on TE accuracy, we observed that the AUC of LSM for the diagnosis of F2-F4 (0.822, 95% confidence interval [CI], ; 0.807, 95% CI, ; 0.779, 95% CI, ) and of F3-F4 (0.915, 95% CI, ; 0.830, 95% CI, ; 0.806, 95% CI, ) (Fig. 3) fibrosis was progressively reduced from lower to middle and further to higher CAP tertiles. Sensitivity, specificity, positive predictive values, and negative predictive values of LSM thresholds of 8.5 kpa for F2-F4 fibrosis and of 10.1 kpa for F3-F4 fibrosis, according to CAP tertiles, are reported in Supporting Table S2. Consistent with these data, the rate of false-positive LSM results for the diagnosis of F2-F4 fibrosis 1149

6 PETTA ET AL. HEPATOLOGY, April 2017 FIG. 1. Accuracy of TE for the prediction of F2-F4 (A) and F3-F4 (B) fibrosis in patients with NAFLD. progressively increased from patients in the lower CAP tertile to those in the middle and the higher tertiles (7/111, 6.3%; 14/108, 12.9%; 15/105, 14.2%) (Fig. 4A). Notably, a similar low false-positive rate of about 6% in patients in the middle and the higher CAP tertiles was observed only in those with LSM >10.5 kpa and >11.6 kpa, respectively. Conversely, higher false-positive rates were reported in patients with LSM values ranging from 8.5 (the best threshold for F2-F4 fibrosis in the entire population) to 10.5 kpa for the middle CAP tertile and from 8.5 to 11.6 kpa for the higher CAP tertile. Similar differences in falsepositive LSM results for the diagnosis of F3-F4 fibrosis were observed (8/111, 7.2% in lower CAP tertile; 18/108, 16.6% in middle CAP tertile; 19/105, 18.1% in higher CAP tertile) (Fig. 4B). Again, a similar acceptable false-positive rate of about 7% in patients in the middle and higher CAP tertiles was observed only in those with LSM >12.5 kpa and >13.6 kpa, respectively. Otherwise, higher false-positive rates were reported in patients with LSM values ranging from 10.1 (the best threshold for F3-F4 fibrosis in the entire population) to 12.5 kpa for the middle CAP tertile and from 10.1 to 13.6 kpa for the higher CAP tertile. Based on the above findings, we designed a diagnostic flowchart for the correct interpretation of LSM values according to CAP values for the diagnosis of F2-F4 (Fig. 5A) and F3-F4 (Fig. 5B) fibrosis. Conversely, the impact of CAP on false-negative LSM results was negligible (data not shown). IMPACT OF LSM ON CAP VALUES AND ON ITS ACCURACY FOR THE DIAGNOSIS OF SEVERE STEATOSIS Consistent with the independent direct association of CAP with LSM values and with evidence of the ability of CAP to affect the accuracy of LSM for staging fibrosis, we tested whether LSM in turn affects the diagnostic accuracy of CAP for the diagnosis of severe grade 3 steatosis. In the subgroup of patients with grade 1-2 steatosis mean CAP values were significantly higher in those with LSM >10.1 kpa the threshold for F3-F4 fibrosis in our cohort compared to those without ( versus , P < 0.001) (Supporting Fig. S2A). Conversely, no differences in CAP results were reported among patients with grade 3 steatosis according to the threshold of LSM 10.1 kpa ( versus , P ) (Supporting Fig. S2B). The ROC curve identified the best cutoff of CAP able to maximize its accuracy at 320 db/m for the diagnosis of grade 3 steatosis (AUC, 0.687; 95% CI, ; sensitivity, 63.8%; specificity, 62.5%). Consistent with these data, the rate of false-positive CAP results for the diagnosis of grade 3 steatosis was higher in patients with LSM >10.1 kpa than in those with LSM 10.1 kpa (34/189, 17.9% versus 44/135, 32.5%) (Supporting Fig. S3). Discussion In this study on 324 patients with biopsy-proven NAFLD, we found that CAP, directly related to steatosis and obesity, is independently linked to increased LSM values, especially in patients with lower stages of fibrosis. As a consequence, we reported higher rates of false-positive LSM results for the noninvasive assessment of both F2-F4 and F3-F4 fibrosis by TE in patients in the higher CAP tertiles. Notably, a diagnostic flowchart based on both LSM and CAP was proposed. TE is an easy-to-perform and reliable tool to diagnose liver fibrosis and cirrhosis in a wide spectrum of liver diseases of different etiologies including NAFLD. (10-12) Here, we report that in a large cohort 1150

7 HEPATOLOGY, Vol. 65, No. 4, 2017 PETTA ET AL. FIG. 2. Liver stiffness value distribution in patients with F0-F1 fibrosis stages (A) and F2-F4 fibrosis stages (B) and in patients with F0-F2 fibrosis stages (C) and F3-F4 fibrosis stages (D), according to CAP tertiles. The horizontal bar inside the box represents the mean value. of NAFLD patients the accuracy of TE for the assessment of liver fibrosis was good at diagnosing F2-F4 and F3-F4 fibrosis, as expressed by their areas under the ROCs. These findings are in agreement with published data. (10-12) Besides, we identified LSM values of 8.5 kpa and 10.1 kpa as the best cutoffs for discriminating F2-F4 and F3-F4 fibrosis, respectively. Once again, these LSM cutoff values were comparable to those identified in other studies. (10-12) Some variability may be related to the different prevalence of fibrosis and of several factors affecting liver stiffness independently of the amount of liver fibrosis in the various examined populations. A relevant issue from our data is that CAP values not only are independently and directly associated with higher LSM values, especially in patients with low grades of fibrosis, but also are able to interfere with the overall diagnostic performance of TE in patients with NAFLD, leading to overestimation of liver fibrosis, resulting in an increase in false-positive rates for the diagnosis of both F2-F4 and F3-F4 fibrosis. We can speculate that higher CAP values reduce the diagnostic accuracy of TE for fibrosis because CAP represents a noninvasive surrogate of factors influencing liver stiffness other than fibrosis in patients with lower stages of disease. Along this line and consistent with literature data, (16) in our study CAP values were directly linked with the severity of steatosis and with obesity. Notably, the first can potentially affect stiffness influencing the architectural structure of the liver, thus changing the propagation time of the vibratory wave through the liver. (12) Similarly, obesity is associated with a higher skin-to-liver distance of the hepatic capsula, and the presence of nonhepatic tissue in the 1151

8 PETTA ET AL. HEPATOLOGY, April 2017 score (22,23) and/or liver biopsy should be considered. Adoption of the adjustments proposed herein may avoid the use of more accurate techniques, such as magnetic resonance elastography, which are expensive and at this moment not widely available. (24,25) The present study also reports a false-negative rate of about 15% for F3-F4 fibrosis. This potentially implies a lack of screening for hepatocellular carcinoma and for esophageal varices in patients at increased risk. Consistent with this limit we consider necessary in patients with LSM values <10.1 kpa performance of a careful clinical, biochemical (platelet levels), and ultrasonographic assessment that could suggest further FIG. 3. AUC of LSM for F3-F4 fibrosis prediction according to CAP tertiles. The AUC of liver stiffness for F3-F4 fibrosis was higher in patients in the lower compared to those in the middle (P ) and in the higher (P ) CAP tertiles. volume explored by the M probe might increase stiffness values. (15,21) Due to the effect of CAP on the diagnostic accuracy of TE, especially in patients with lower stages of fibrosis, we propose some adjustments for a better interpretation of LSM results. Specifically, we suggest consideration of LSM measurements as accurate in all patients in the lower CAP tertile. In the middle CAP tertile, LSM values ranging from 8.5 to 10.5 kpa for F2-F4 fibrosis and from 10.1 to 12.5 kpa for F3-F4 fibrosis expose patients to a higher risk of false-positive results, while LSM values >10.5 kpa and >12.5 kpa make us confident in diagnosing F2-F4 and F3-F4 fibrosis, respectively. Similarly, in patients in the higher CAP tertile, F2-F4 and F3-F4 fibrosis may be confirmed with LSM values >11.6 and >13.6 kpa, respectively, while LSM values ranging from 8.5 to 10.5 kpa for F2-F4 fibrosis and from 10.1 to 12.5 kpa for F3-F4 fibrosis expose patients to a higher risk of false-positive results. It is important to underscore that LSM and CAP values are simultaneously provided by the same FibroScan machine, which is widely available; and their integration could help us to reduce overestimation of liver fibrosis and false-positive rates in a high proportion of patients. In the remaining patients where FibroScan does not work well, combination with noninvasive scores such as NAFLD fibrosis FIG. 4. Distribution of false-positive LSM results for the diagnosis of F2-F4 (A) and of F3-F4 (B) fibrosis according to CAP tertiles. 1152

9 HEPATOLOGY, Vol. 65, No. 4, 2017 PETTA ET AL. FIG. 5. Proposed diagnostic flowchart for F2-F4 (A) and F3-F4 (B) fibrosis according to LSM values and CAP tertiles. The best thresholds for F2-F4 (8.5 kpa) and F3-F4 (10.1 kpa) fibrosis calculated in the entire population of 324 patients with NAFLD were used. For practical purposes, CAP tertile thresholds were rounded from 299 db/m to 300 db/m and from 338 db/m to 340 db/m without changes in results. Abbreviation: NFS, NAFLD fibrosis score. investigations in some patients. Consistent with this suggestion, in the present study, among patients with false-negative results for F3-F4 fibrosis, 50% were >60 years, 73% had diabetes, and 19% had platelet levels <150,000/mm 3. In our study we also found that LSM values affect the accuracy of CAP in diagnosing severe hepatic steatosis, increasing the number of false-positive results, in a similar fashion as CAP affects LSM diagnostic performance. Further studies specifically assessing the role of CAP for grading steatosis in NAFLD should carefully take into account this feature together with other factors affecting CAP values. This study has some limitations. First, interobserver concordance of LSM and CAP examinations as well as of liver histological readings were not assessed, which potentially affects the interpretation of our results. However, regarding FibroScan, all tests were performed by expert operators following the same protocol and fulfilling validity criteria. Moreover, different relevant studies assessing FibroScan in NAFLD were based on multicenter cohorts and/or on multiple operators. (9-11,16,26,27) Finally, many different studies have reported good interobserver concordance for LSM and CAP. (28-30) Regarding histology, data from the literature (20) and from our group (31) clearly show that overall interobserver agreement for staging severe fibrosis in NAFLD is good; and again different relevant studies assessing FibroScan in histologically defined NAFLD were based on multicenter cohorts. (9-11,26,27) Another 1153

10 PETTA ET AL. HEPATOLOGY, April 2017 limitation is that the analysis of LSM was carried out using a conventional M probe only, while an XL probe was not used. In the setting of NAFLD patients, where the proportion of obese individuals is high, like in the present study, the XL probe could not only increase the reliability of FibroScan (26,27) but may also potentially reduce the effect of CAP as a confounder of LSM, which is less affected by obesity and skin-toliver distance when using the XL probe. Consistent with this, further studies are needed to assess the effect of CAP on LSM measured with the XL probe in obese NAFLD. A further methodological issue is the potentially limited external validity of the results for different populations and settings. Our study included a cohort of NAFLD subjects, largely overweight to obese, who were referred to tertiary centers for liver disease, limiting the applicability of the results in different populations, e.g., patients in primary care. Finally, lack of follow-up data may limit the potency of our results. In conclusion, LSM in NAFLD may lead to overestimation of fibrosis in patients with lower stages of disease. Interpreting LSM values based on the simultaneous determination of CAP, a noninvasive parameter provided by the same machine, could help to more accurately use TE for predicting fibrosis in NAFLD. REFERENCES 1) Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, Wymer M. Global epidemiology of non-alcoholic fatty liver disease meta-analytic assessment of prevalence, incidence and outcomes. HEPATOLOGY 2015;64: ) Dyson J, Jaques B, Chattopadyhay D, Lochan R, Graham J, Das D, et al. Hepatocellular cancer: the impact of obesity, type 2 diabetes and a multidisciplinary team. J Hepatol 2014;60: ) Wong RJ, Aguilar M, Cheung R, Perumpail RB, Harrison SA, Younossi ZM, et al. Nonalcoholic steatohepatitis is the second leading etiology of liver disease among adults awaiting liver transplantation in the United States. Gastroenterology 2015;148: ) Ekstedt M, Hagstr om H, Nasr P, Fredrikson M, Stål P, Kechagias S, 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: ) Angulo P, Kleiner DE, Dam-Larsen S, Adams LA, Bjornsson ES, Charatcharoenwitthaya P, et al. Liver fibrosis, but no other histologic features, associates with long-term outcomes of patients with nonalcoholic fatty liver disease. Gastroenterology 2015;149: ) Ratziu V, Charlotte F, Heurtier A, Gombert S, Giral P, Bruckert E, et al; LIDO Study Group. Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology 2005;128: ) Petta S, Handberg A, Craxı A. Non invasive indexes for the assessment of patients with non-alcoholic fatty liver disease. Curr Pharm Des 2013;19: ) 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: ) Wong VW, Vergniol J, Wong GL, Foucher J, Chan AW, Chermak F, et al. Liver stiffness measurement using XL probe in patients with nonalcoholic fatty liver disease. Am J Gastroenterol 2012;107: ) Wong VW, Vergniol J, Wong GL, Foucher J, Chan HL, Le Bail B, et al. Diagnosis of fibrosis and cirrhosis using liver stiffness measurement in nonalcoholic fatty liver disease. HEPATOL- OGY 2010;51: ) Cassinotto C, Boursier J, De Ledinghen V, Lebigot J, Lapuyade B, Cales P, et al. Liver stiffness in nonalcoholic fatty liver disease: a comparison of supersonic shear imaging, FibroScan and ARFI with liver biopsy. HEPATOLOGY 2015;63: ) Petta S, Maida M, Macaluso FS, Di Marco V, Camma C, Cabibi D, et al. The severity of steatosis influences liver stiffness measurement in patients with nonalcoholic fatty liver disease. HEPATOLOGY 2015;62: ) Wong GL, Chan HL, Choi PC, Chan AW, Lo AO, Chim AM, et al. Association between anthropometric parameters and measurements of liver stiffness by transient elastography. Clin Gastroenterol Hepatol 2013;11: ) Das K, Sarkar R, Ahmed SM, Mridha AR, Mukherjee PS, Das K, 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: ) Petta S, Di Marco V, Camma C, Butera G, Cabibi D, Craxı A. Reliability of liver stiffness measurement in non-alcoholic fatty liver disease: the effects of body mass index. Aliment Pharmacol Ther 2011;33: ) De Ledinghen V, Wong GL, Vergniol J, Chan HL, Hiriart JB, Chan AW, et al. Controlled attenuation parameter for the diagnosis of steatosis in non-alcoholic fatty liver disease. J Gastroenterol Hepatol 2016;31: ) American Diabetes Association. Clinical practice recommendations Diabetes Care 2000;23(Suppl.):S4-S19. 18) Boursier J, Zarski JP, de Ledinghen V, Rousselet MC, Sturm N, Lebail B, et al.; Multicentric Group from ANRS/HC/EP23 FIBROSTAR Studies. Determination of reliability criteria for liver stiffness evaluation by transient elastography. HEPATOLOGY 2013;57: ) Colloredo G, Guido M, Sonzogni A, Leandro G. Impact of liver biopsy size on histological evaluation of chronic viral hepatitis: the smaller the sample, the milder the disease. J Hepatol 2003; 39: ) Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. HEPATOLOGY 2005;411: ) Shen F, Zheng RD, Shi JP, Mi YQ, Chen GF, Hu X, et al. Impact of skin capsular distance on the performance of controlled attenuation parameter in patients with chronic liver disease. Liver Int 2015;35: ) Angulo P, Hui JM, Marchesini G, Bugianesi E, George J, Farrell GC, et al. The NAFLD fibrosis score: a noninvasive 1154

11 HEPATOLOGY, Vol. 65, No. 4, 2017 PETTA ET AL. system that identifies liver fibrosis in patients with NAFLD. HEPATOLOGY 2007;45: ) Petta S, Vanni E, Bugianesi E, Di Marco V, Camma C, Cabibi D, et al. The combination of liver stiffness measurement and NAFLD fibrosis score improves the noninvasive diagnostic accuracy for severe liver fibrosis in patients with nonalcoholic fatty liver disease. Liver Int 2015;35: ) Loomba R, Wolfson T, Ang B, Hooker J, Behling C, Peterson M, et al. Magnetic resonance elastography predicts advanced fibrosis in patients with nonalcoholic fatty liver disease: a prospective study. HEPATOLOGY 2014;60: ) Imajo K, Kessoku T, Honda Y, Tomeno W, Ogawa Y, Mawatari H, et al. Magnetic resonance imaging more accurately classifies steatosis and fibrosis in patients with nonalcoholic fatty liver disease than transient elastography. Gastroenterology 2016; 150: ) Myers RP, Pomier-Layrargues G, Kirsch R, Pollett A, Duarte- Rojo A, Wong D, et al. Feasibility and diagnostic performance of the FibroScan XL probe for liver stiffness measurement in overweight and obese patients. HEPATOLOGY 2012;55: ) de Ledinghen V, Wong VW, Vergniol J, Wong GL, Foucher J, Chu SH, et al. Diagnosis of liver fibrosis and cirrhosis using liver stiffness measurement: comparison between M and XL probe of FibroScan VR. J Hepatol 2012;56: ) Afdhal NH, Bacon BR, Patel K, Lawitz EJ, Gordon SC, Nelson DR, et al. Accuracy of FibroScan, compared with histology, in analysis of liver fibrosis in patients with hepatitis B or C: a United States multicenter study. Clin Gastroenterol Hepatol 2015;13: e ) Boursier J, Konate A, Guilluy M, Gorea G, Sawadogo A, Quemener E, et al. Learning curve and interobserver reproducibility evaluation of liver stiffness measurement by transient elastography. Eur J Gastroenterol Hepatol 2008;20: ) Recio E, Cifuentes C, Macıas J, Mira JA, Parra-Sanchez M, Rivero-Juarez A, et al. Interobserver concordance in controlled attenuation parameter measurement, a novel tool for the assessment of hepatic steatosis on the basis of transient elastography. Eur J Gastroenterol Hepatol 2013;25: ) Petta S, Valenti L, Marra F, Grimaudo S, Tripodo C, Bugianesi E, et al. MERTK rs polymorphism affects the severity of fibrosis in non-alcoholic fatty liver disease. J Hepatol 2016;64: Supporting Information Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/ /hep.28843/suppinfo. 1155