Background of the FIB-4 Index in Japanese Non-Alcoholic Fatty Liver Disease

ORIGINAL ARTICLE Background of the FIB-4 Index in Japanese Non-Alcoholic Fatty Liver Disease Takashi Wada and Mikio Zeniya Abstract Objective We investigated the distribution and characteristics of the FIB-4 index of liver fibrosis in 1,441 Japanese men (age 50.7±10.2 years) and 304 women (age 53.9±10.3 years) who underwent comprehensive general health checkups and were identified as having non-alcoholic fatty liver disease. With respect to the FIB-4 index, differences according to sex, metabolic indices, and ultrasonic findings were investigated. Methods Among 9,255 individuals who underwent comprehensive general health checkups, 2,750 (29.8%) were found to have mild fatty liver or fatty liver based on ultrasound findings. After excluding patients who consumed 150 g alcohol/week (818 individuals), those testing positive for hepatitis B surface antigens or hepatitis C virus antibody (184 individuals), and those for whom data were insufficient (three individuals), we investigated the FIB-4 indices in the remaining 1,745 subjects. Results There were no sex differences in the FIB-4 index. A total of 1,370 patients (78.5%) exhibited a low cut-off index (COI) (<1.30), 357 (20.5%), exhibited an indeterminate COI (1.30-2.67), and 18 (1.0%) exhibited a high COI (>2.67). There were no associations between the FIB-4 index and the constituent factors of metabolic syndrome. In contrast, there was a significant difference in the ln FIB-4 index between the patients with and without mild fatty liver or fatty liver on ultrasound among men (0.006±0.43 and -0.092±0.39, p<0.001), but not women. Conclusion The FIB-4 index was is significantly lower in men, but not women, with fatty liver. The FIB-4 index must be calculated separately during medical checkups and evaluated in conjunction with ultrasound findings. Key words: liver fibrosis, non-alcoholic fatty liver disease, ultrasound, general health checkup, sex difference () () Introduction Fatty liver disease occurs when fat is deposited in hepatocytes, thereby causing liver damage. This condition is defined as the presence of fat droplets in 30% of hepatocytes in liver tissue, and is clinically diagnosed based on ultrasound findings. Previously, cases of fatty liver were most frequently caused by alcohol consumption; however the disease currently develops as a manifestation of lifestyle-related diseases, such as diabetes mellitus and obesity. Nonalcoholic fatty liver disease (NAFLD) involves fat-induced liver damage resembling that caused by alcohol-related liver damage in the absence of a history of alcohol consumption. According to Eguchi et al. (1), the incidence of NAFLD among non-drinkers is 41% in adult men and 18% in adult women. The ideal modality for diagnosing NAFLD is a liver biopsy, although it is highly invasive (2, 3). The actual incidence of patients who require a liver biopsy is therefore unknown, and clear indication criteria for the procedure are lacking. A number of non-invasive indices have been proposed to identify candidates for liver biopsy, including the NAFLD fibrosis score (NFS) (4), BARD score (5), and FIB- 4 index (6). However, these liver fibrosis indices are not listed in the standard protocol for general medical checkups in the the United States or Europe. Nevertheless, the FIB-4 index is reported to be the most useful parameter (6, 7); it can be calculated from the platelet count, and aspartate Health-Care Center, The Jikei University School of Medicine, Japan Received for publication April 16, 2014; Accepted for publication May 21, 2014 Correspondence to Dr. Takashi Wada, t_wada@jikei.ac.jp 127

person Figure 1. FIB-4 index distribution. Table 1. Subject Attributes Men Women n=1,441 n=304 Age (y) 50.7±10.2 53.9±10.3 BMI(kg/m 2 ) 25.6±3.3 25.5±3.8 Waist circumference (cm) 89.5±8.0 87.9±8.9 Platelet count ( 10 3 /mm 3 ) 232±48 257±58 AST(U/L) 26.1±11.3 23.1±10.5 ALT(U/L) 37.1±24.1 26.7±18.6 FIB-4 index 1.04±0.49 1.05±0.50 Mild fatty liver/ fatty liver 575/866 146/158 BMI: body mass index, AST: aspartate transaminase, ALT: alanine transaminase aminotransferase (AST) and alanine aminotransferase (ALT) levels, variables that are readily obtainable during comprehensive general health checkups (Ningen Dock) and in everyday clinical practice. Therefore, simplicity is a major advantage of this index. To date, there have been no studies of sex differences in the FIB-4 index, at least among Japanese subjects. Although an association between NAFLD and metabolic syndrome has been suggested, no previous studies have investigated the effects of blood pressure, an assessment of which is required for the diagnosis of metabolic syndrome. Nor haves any studies evaluated differences between mild fatty liver and fatty liver, as evidenced on ultrasound. In addition, the FIB- 4 index exhibits a log-normal distribution rather than a normal distribution, although there have been no previous analysis of this issue. We therefore investigated the distribution of the FIB-4 index among patients diagnosed with NAFLD during comprehensive general health checkup, and investigated the association between the FIB-4 index and the constituent factors of metabolic syndrome. Materials and Methods Of the 9,255 individuals who underwent comprehensive general health checkups at the Shimbashi Medical Checkup Office, Jikei University Hospital, during the one-year period between April 2010 and March 2011, 2,750 were identified as having mild fatty liver or fatty liver, based on ultrasound findings. Fatty liver was categorized as either mild fatty liver, in which hepatorenal contrast was equivalent or better, or moderate or worse fatty liver, in which the difference was clear. The diagnosis was made by two experienced ultrasound technicians, and the ultrasound device was a Xario TM XG (Toshiba, Tokyo, Japan) 3.5-MHz convex probe. After excluding 818 individuals with a weekly ethanol intake of >150 g, three individuals with insufficient data, and 184 individuals who tested positive for either hepatitis B surface antigens (HBsAg) or hepatitis C virus antibody (HCVAb), the remaining 1,745 subjects were assessed including 1,441 men (age 50.7±10.2 years) and 304 women (age 53.9±10.3 years). The FIB-4 index was calculated as AST age/(platelet count ALT), and the subjects were categorized into three groups according to the cut-off index (COI) values proposed by Shah et al. (3): low COI (<1.30), indeterminate COI (1.30-2.67), or high COI (>2.67). The study protocol was approved by the Ethics Committee of the Jikei University School of Medicine. Statistical analysis The StatView 4.0 (SAS Institute, Cary, USA) software program was used for the statistical analysis. An analysis of variance (ANOVA) was performed according to Scheffe s method, and a regression analysis was performed using the stepwise technique. A p value of <0.05 was considered to be statistically significant. Results Of a total of 9,255 individuals who underwent comprehensive general health checkups between April 2010 and March 2011, 2,750 (29.8% of the total population) were found to be suffering from mild fatty liver or fatty liver based on the ultrasound findings. In order to identify individuals with NAFLD, who consumed 22 g of ethanol per day, the amount contained in a serving (approximately 180 ml of Japanese sake, i.e., 150 g of alcohol per week; 818 individuals, 8.8%) were excluded. Also excluded were individuals who tested positive for HBsAg or HCVAb (184 individuals, 2.0%), and those with incomplete data (three individuals), leaving a total of 1,745 subjects with NAFLD. This group accounted for 18.9% of all patients who underwent the comprehensive general health checkups, and 63.5% of those diagnosed with mild fatty liver or fatty liver based on ultrasound findings. The mean FIB-4 index among the 1,745 study subjects was 1.04±0.49, with a median value of 0.93. The data were confirmed to have a normal distribution after logarithmic conversion, as shown in Fig. 1. The mean FIB-4 index was 1.04±0.49 among men and 1.05±0.50 among women, with no significant sex differences (Table 1). The COI was low in 1,370 individuals (78.3%), high in 18 individuals (1.0%), and indeterminate in the remaining 128

2,750 people having fatty liver seen by ultrasound ethanol consumption 150g/week (n=818) incomplete data (n=3) HBsAg and/or HCVAb positive (n=184) 1,745 people diagnosed as having NAFLD FIB-4 index Low (<1.3) Indeterminate (1.3-2.67) High (>2.67) (n=1370, 78.5%) (n=357, 20.5%) (n=18, 1.0%) Male (%) 1135 (82.8%) 293 (82.0%) 13 (72.2%) Age 48.4±8.8 y 61.6±8.1 y 66.6±7.7 y Figure 2. Selection of the subjects. Table 2. Test Results for Men with Different FIB-4 Indices Age (y) BMI (kg/m 2 ) Waist circumference (cm) Platelet count ( 10 3 /mm 3 ) AST (U/L) ALT (U/L) AST/ALT GGT (U/L) Albumin (g/dl) FPG(mg/dL) Neutral lipids (mg/dl) HDL-C (mg/dl) SBP (mmhg) Low (<1.3) Indeterminate (1.3-2.67) High (>2.67) (n=1,135, 78.8%) (n=293, 20.3%) (n=13, 0.9%) 47.8±8.5 61.4±8.4 (p<0.001) 64.8±7.4 (p<0.001) 25.7±3.4 25.3±2.8 24.7±3.8 89.5±8.2 89.3±7.1 87.5±7.1 245±44 188±30 (p<0.001) 138±51 (p<0.001) 25.0±9.3 29.6±15.6 (p<0.001) 47.5±18.0 (p<0.001) 37.5±22.9 34.8±27.6 51.8±33.2 0.75±0.22 0.99±0.30 (p<0.001) 1.06±0.43 (p<0.001) 53.8±41.0 51.8±67.7 76.5±70.9 4.39±0.22 4.28±0.24 (p<0.001) 4.37±0.23 105±25.3 110±26.6 (p<0.001) 103±17 163±109 135±77 131±63 49.9±10.5 50.9±12.0 51.7±18.7 123±13 125±13 118±10 DBP(mmHg) 78±10 79±9 72±11 BMI: body mass index, AST: aspartate transaminase, ALT: alanine transaminase, GGT: gamma-glutamyl transferase, FPG: fasting plasma glucose, HDL-C: high density cholesterol, SBP: systolic blood pressure, DBP: diastolic blood pressure Statistical analysis; versus low group 357 individuals (20.5%) (Fig. 2). Table 2 shows the results for men according to the FIB-4 index, while Table 3 shows the results for women. In terms of associations with metabolic syndrome, the only item for which there was a difference between the groups with different FIB-4 index values in women was fasting plasma glucose, whereas no differences were evident in waist circumference, neutral lipids, high-density lipoprotein (HDL-C), systolic blood pressure, diastolic blood pressure, or fasting plasma glucose in men. There were also no significant differences in albumin or BMI, both of which are included in the prediction formula for the NFS (4), between any of the three FIB-4 index groups. Nor were there any significant differences between the three FIB-4 index groups in terms of plasma glucose, an element of the prediction formula for the BARD score (5). Table 4 shows the clinical findings for the patients diagnosed with mild fatty liver or fatty liver based on an ultrasound diagnosis. There were no significant differences between the two groups in most of the test items associated with metabolic syndrome. However, the FIB-4 index was significantly lower in men, but not women, with fatty liver (p<0.001). When only patients with liver cell injury, as indicated by an ALT level of 31, were investigated, the ln FIB- 4 index was found to be significantly different between men with mild fatty liver versus those with fatty liver (0.006± 0.43 vs. 0.092±0.39, p<0.001). The corresponding values for women were -0.031±0.42 and 0.064±0.22, respectively; this difference was not significant. These findings suggest that it may not be possible to estimate the degree of fibrosis based on the severity of fatty liver on ultrasound. Discussion We herein investigated the distribution and characteristics of the FIB-4 index of liver fibrosis in healthy individuals, using patients with NAFLD as subjects, from the view point of clinical preventive medicine. In Japan, a non-alcoholic status refers to an intake of <20 g of ethanol per day (8). However, the ethanol content in a serving (approximately 180 ml) of Japanese sake is approximately 22-28 g, while that of a medium bottle or long can of beer (500 ml) is 22-129

Table 3. Test Results for Women with Different FIB-4 Indices Low (<1.3) Indeterminate (1.3-2.67) High (>2.67) (n=235, 77.3%) (n=64, 21.1%) (n=5, 1.6%) Age (y) 51.1±9.1 62.5±7.0 (p<0.001) 7 71.2±7.0 (p<0.001) BMI (kg/m 2 ) 25.6±3.7 25.0±4.2 23.8±3.7 Waist circumference (cm) 87.9±8.7 88.2±9.5 84.6±9.1 Platelet count( 10 3 /mm 3 ) 273±53 206±36 (p<0.001) 149± 52 (p<0.001) AST (U/L) 21.6±7.6 26.3±14.1 (p<0.001) 50.2±24.2 (p<0.001) ALT (U/L) 26.0±15.4 26.6±24.6 61.6±38.3 (p=0.001) AST/ALT 0.95±0.29 1.13±0.29 (p<0.001) 0.91±0.27 GGT (U/L) 32.5±26.2 27.5±14.6 44.8±15.1 Albumin (g/dl) 4.20±0.24 4.22±0.21 4.22±0.28 FPG (mg/dl) 101±14.0 106±14.0 (p<0.001) 132±39.3 (p<0.001) Triglyceride (mg/dl) 124±69 110±53 115±56 HDL-C (mg/dl) 60.0±14.0 63.1±15.5 68.4±13.5 SBP (mmhg) 123±15 126±13 125±12 DBP (mmhg) 75±10 76±8 79±12 BMI: body mass index, AST: aspartate transaminase, ALT: alanine transaminase, GGT: gamma-glutamyl transferase, FPG: fasting plasma glucose, HDL-C: high density cholesterol, SBP: systolic blood pressure, DBP: diastolic blood pressure Statistical analysis; versus low group Table 4. Ultrasound Diagnostic Findings for Individuals with Mild Fatty Liver and Fatty Liver Men Women Mild fatty liver Fatty liver Mild fatty liver Fatty liver n=575 n=866 n=146 n=158 Age (y) 52.3±10.6 49.7±9.7*** 53.9±10.4 53.9±10.3 BMI (kg/m 2 ) 24.3±2.3 26.4±3.5*** 24.3±3.3 26.5±4.0*** WC (cm) 86.9±6.3 91.1±8.5*** 85.7±8.4 89.9±8.8*** Platelet count 228±48 234±49* 252±51 262±64* ( 10 3 /mm 3 ) AST (U/L) 23.6±8.7 27.8±8.7*** 20.2±6.1 25.8±12.8*** ALT (U/L) 29.5±17.3 42.1±26.5*** 20.0±9.0 32.9±22.7*** AST/ALT 0.90±0.29 0.74±0.21*** 1.09±0.28 0.89±0.28*** GGT (U/L) 47.5±41.7 57.6±51.3*** 25.3±15.2 37.5±29.1*** Albumin (g/dl) 4.34±0.23 4.38±0.22*** 4.19±0.23 4.23±0.23* FPG (mg/dl) 102±18.3 109±29*** 101±15.0 104±15.5* Triglecreide 143±99 166±105*** 111±57 131±73*** (mg/dl) HDL-C(mg/dL) 51.5±11.5 49.2±10.4*** 62.3±14.9 59.2±13.7 SBP (mmhg) 122±13.0 124±13.1* 120±13.7 126±14.6*** DBP (mmhg) 77±9.7 79±9.8*** 73±9.9 77±9.4*** Ln FIB-4 index 0.006±0.43-0.092±0.39*** -0.031±0.42-0.064±0.22 BMI: body mass index, WC: waist circumference, AST: aspartate transaminase, ALT: alanine transaminase, GGT: gamma-glutamyl transferase, FPG: fasting plasma glucose, HDL-C: high density cholesterol, SBP: systolic blood pressure, DBP: diastolic blood pressure Statistical analysis; *p<0.05, ** p<0.01, *** p<0.001, versus mild fatty liver group 25 g, and some subjects do not drink every day. In the present study, we therefore used the definition of nonalcoholic as an ethanol intake of 150 g per week. Liver fibrosis is more advanced in elderly than in young patients. This time-dependent nature of the degree of fibrosis is believed to underlie the tendency for the FIB-4 index to increase with age (9). NAFLD is also thought to represent the hepatic expression of metabolic syndrome, due to insulin resistance caused by visceral obesity (10). In the present study, however, we found no associations between an elevated FIB-4 index and waist circumference or the values of neutral lipids, HDL-C or, systolic or diastolic blood pressure, all of which are diagnostic constituent factors for metabolic syndrome, with the exception of fasting plasma glucose. In addition, Yatsuji et al. (11) reported that diabetes mellitus has an effect on the incidence of NAFLD in women, consistent with the results of the present study. With respect to performing a comprehensive general health checkup, it is necessary to assess the state of the liver and identify the presence of fatty liver and/or liver dysfunction. It is necessary to identify patients who require liver biopsy from among those who exhibit fibrosis or a prefibrotic condition of NAFLD. Sumida et al. (12) diagnosed the state of fibrosis in liver biopsies of 576 Japanese patients with NAFLD, and investigated whether such staging could be evaluated with the FIB- 4 index and/or six other indices. The researchers regarded stages 3 and 4 disease as being severe, and found that there were 64 such cases (11%). Unlike that used in the present 130

study, the cut-off points for the FIB-4 index was categorized as low at <1.45, indeterminate at 1.45-3.25, and high at >3.25. Compared with other scoring systems, these values were found to be useful for distinguishing between mild and severe fibrosis, with a reported negative predictive value of 98% and positive predictive value of 53%. Of the 395 patients, 91% were classified into the 1.45-3.25 group, indicating that assessing the FIB-4 index may render liver biopsy unnecessary in Japanese patients. Meanwhile, McPherson et al. (13) investigated whether the non-invasive indices of the AST/ALT ratio, AST-toplatelet-ratio index, FIB-4 index, NAFLD fibrosis score (4) and BARD score (5) could be used to evaluate the degree of fibrosis in 145 patients diagnosed with NAFLD based on the findings of a liver biopsy. The authors found that the FIB-4 index outperformed the other indices in terms of its capacity for diagnosing advanced fibrosis. Furthermore, Shah et al. (6) investigated whether the FIB- 4 and seven other indices could be used to assess the presence of fibrosis in 541 patients diagnosed with NAFLD and found a jackknife-validated area under the receiver operating characteristic curve of 0.802 for the FIB-4 index [95% confidence interval (CI), 0.758-0.847], which was superior to that for the other indices. From this viewpoint, Kakuta et al. (14) used the same method applied in the present study to validate the FIB-4 index in individuals undergoing health checkups. In the Kakuta study, the proportion of patients with a low FIB-4 index was 79% (78% in the present study), those with an indeterminate index was 20% (20% in the present study), and those with a high index was 0.9% (1% in the present study). The results of the present study were almost identical to these findings. In the present study, among patients diagnosed with mild fatty liver or fatty liver on an ultrasound diagnosis, the FIB- 4 index was significantly lower in men with fatty liver. However, this trend was not observed in women. This suggests that it is difficult to estimate the degree of fibrosis based on the severity of fatty liver on ultrasound. In this study, we used hepatorenal contrast, which exhibits a better capacity for obtaining an accurate diagnosis of fatty infiltration of the liver than the level of hepatic parenchymal brightness (15). This method was used to categorize fatty liver as either mild or as moderate or worse. The present study showed no differences in the proportions of patients with mild fatty liver and fatty liver among the subjects with an elevated FIB-4 index. The absence of differences in ultrasound findings may have been due to the fact that fatty metamorphosis resolves with the progression of fibrosis in patients with a high COI. This observation suggests that the FIB-4 index is useful for detecting the progression of fibrosis resulting from NAFLD. On comprehensive general health checkups, however, only 1% of patients are categorized as having a high level of fibrosis according to Shah s classification, and it is unclear whether this COI is appropriate for the Japanese population. Future studies are therefore required to verify the COI by reducing the lower limit for the high group, for example from 2.37 to 2.0, and performing liver biopsies in such patients. Ultrasound devices capable of measuring the degree of liver stiffness have recently been developed. In addition, Matsumoto et al. reported an association between liver stiffness and the FIB-4 index (16). That study found a good correlation between ultrasound measurements of liver stiffness and the FIB-4 index (R 2 =0.519, p<0.0001). However, the FIB-4 index, which can be calculated based on blood test results, is easier to apply when evaluating the degree of fibrosis in the liver and should not be replaced. A total of three million or more people a year undergo general medical checkups (Ningen Dock) in Japan. Assessments of liver fibrosis indices, such as the FIB-4 index, are not listed in the standard Ningen Dock program. However, beginning in April 2013, although not evaluated on comprehensive general health checkup records, the FIB-4 index is currently assessed on forms recording the doctor s opinion, and patients classified into the high group are carefully evaluated. One limitation associated with the present study is that the FIB-4 index may be overestimated in patients with a low platelet count compared to the liver function, such as those with pseudo-thrombocytopenia and idiopathic thrombocytopenia. Another limitation is that liver biopsies were not performed. Future studies involving stratified analysis including the presence of concomitant lifestyle-related diseases are required, as well as prospective studies of the association between the FIB-4 index and the long-term prognosis (mortality associated with liver disease). In conclusion, the proportions of patients with fibrosis categorized into three different grades of severity according to the FIB-4 index in the present study are similar to reported at other institutions. Differences in age, as well as the platelet count, and AST level, are evident among patients with a high FIB-4 index (indicating advanced fibrosis); these parameters are also included in the formula for calculating the FIB-4 index. However, in the present study, there was were no differences in the FIB-4 index between the patients with mild fatty liver and fatty liver as detected on ultrasound, and diagnosing fibrosis was problematic. Subjects who exhibit a high COI on general health checkups may need to undergo liver biopsies, although prospective studies are required to clarify this issue. The authors state that they have no Conflict of Interest (COI). References 1. Eguchi Y, Hyogo H, Ono M, et al.; JSG-NAFLD. Prevalence and associated metabolic factors of nonalcoholic fatty liver disease in the general population from 2009 to 2010 in Japan: a multicenter large retrospective study. J Gastroenterol 47: 586-595, 2012. 2. Vuppalanchi R, Chalasani N. Nonalcoholic fatty liver disease and nonalcoholic steatohepatitis: selected practical issues in their 131

evaluation and management. Hepatology 46: 582-589, 2007. 3. Wieckowska A, McCullough AJ, Feldstein AE. Noninvasive diagnosis and monitoring of nonalcoholic steatohepatitis: present and future. Hepatology 49: 306-317, 2009. 4. Angulo P, Hui JM, Marchesini G, et al. The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology 45: 846-854, 2007. 5. Harrison SA, Oliver D, Arnold HL, Gogia S, Neuschwander-Tetri BA. Development and validation of a simple NAFLD clinical scoring system for identifying patients without advanced disease. Gut 57: 1441-1447, 2008. 6. Shah AG, Lydecker A, Murray K, Tetri BN, Contos MJ, Sanyal AJ; Nash Clinical Research Network. Collaborators (114). Comparison of noninvasive markers of fibrosis in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol 7: 1104-1112, 2009. 7. McPherson S, Stewart SF, Henderson E, Burt AD, Day CP. Simple non-invasive fibrosis scoring systems can reliably exclude advanced fibrosis in patients with non-alcoholic fatty liver disease. Gut 59: 1265-1269, 2010. 8. Okanoue T, Saibara T, Nono M, Sumida Y, Hashimoto E, Tamura S, et al. JSH Consensus Kobe 2009; Diagnosis and treatment of NASH. Kanzo 50: 741-747, 2009 (in Japanese). 9. Angulo P, Keach JC, Batts KP, Lindor KD. Independent predictors of liver fibrosis in patients with nonalcoholic steatohepatitis. Hepatology 30: 1356-1362, 1999. 10. Watanabe S, Yaginuma R, Ikejima K, Miyazaki A. Liver diseases and metabolic syndrome. J Gastroenterol 43: 509-518, 2008. 11. Yatsuji S, Hashimoto E, Tobari M, Tokushige K, Shiratori K. Influence of age and gender in Japanese patients with non-alcololic steatohepatitis. Hepatol Res 37: 1034-1043, 2007. 12. Sumida Y, Yoneda M, Hyogo H, et al.; Japan Study Group of Nonalcoholic Fatty Liver Disease. Validation of the FIB4 index in a Japanese nonalcoholic fatty liver disease population. BMC Gastroenterol 12: 2, 2012. 13. McPherson S, Stewart SF, Henderson E, Burt AD, Day CP. Simple non-invasive fibrosis scoring systems can reliably exclude advanced fibrosis in patients with non-alcoholic fatty liver disease. Gut 59: 1265-1269, 2010. 14. Kakuta Y, Sakuragi S, Ohno T, Sakai K, Kanemasa K. The distribution of FIB-4 index in Japanese subjects undergoing a health checkup who are diagnosed as having nonalcoholic fatty liver disease. Kanzo 52: 390-392, 2011 (in Japanese, Abstract in English). 15. Sanyal AJ; American Gastroenterological Association. AGA technical review on nonalcoholic fatty liver disease. Gastroenterology 123: 1705-1725, 2002. 16. Matsumoto T, Sohda T, Nishizawa S, et al. Non-invasive evaluation of liver fibrosis in patients with chronic liver diseases. Med Bull Fukuoka Univ 38: 1-6, 2011. 2015 The Japanese Society of Internal Medicine http://www.naika.or.jp/imonline/index.html 132