Magnetic Resonance Imaging Proton Density Fat Fraction Associates With Progression of Fibrosis in Patients With Nonalcoholic Fatty Liver Disease

Markers are needed to predict progression of nonalcoholic fatty liver disease (NAFLD). The proton density fat fraction, measured by magnetic resonance imaging (MRI-PDFF), provides an accurate, validated marker of hepatic steatosis; however, it is not clear whether the PDFF identifies patients at risk for NAFLD progression. We performed a follow-up study of 95 well-characterized patients with biopsy-proven NAFLD and examined the association between liver fat content and fibrosis progression. MRI-PDFF measurements were made at study entry (baseline). Biopsies were collected from patients at baseline and after a mean time period of 1.75 years. Among patients with no fibrosis at baseline, a higher proportion of patients in the higher liver fat group (MRI-PDFF ≥15.7%) had fibrosis progression (38.1%) than in the lower liver fat group (11.8%) (P = .067). In multivariable-adjusted logistic regression models (adjusted for age, sex, ethnicity, and body mass index), patients in the higher liver fat group had a significantly higher risk of fibrosis progression (multivariable-adjusted odds ratio 6.7; 95% confidence interval 1.01–44.1; P = .049). Our findings associate higher liver fat content, measured by MRI-PDFF, with fibrosis progression. Markers are needed to predict progression of nonalcoholic fatty liver disease (NAFLD). The proton density fat fraction, measured by magnetic resonance imaging (MRI-PDFF), provides an accurate, validated marker of hepatic steatosis; however, it is not clear whether the PDFF identifies patients at risk for NAFLD progression. We performed a follow-up study of 95 well-characterized patients with biopsy-proven NAFLD and examined the association between liver fat content and fibrosis progression. MRI-PDFF measurements were made at study entry (baseline). Biopsies were collected from patients at baseline and after a mean time period of 1.75 years. Among patients with no fibrosis at baseline, a higher proportion of patients in the higher liver fat group (MRI-PDFF ≥15.7%) had fibrosis progression (38.1%) than in the lower liver fat group (11.8%) (P = .067). In multivariable-adjusted logistic regression models (adjusted for age, sex, ethnicity, and body mass index), patients in the higher liver fat group had a significantly higher risk of fibrosis progression (multivariable-adjusted odds ratio 6.7; 95% confidence interval 1.01–44.1; P = .049). Our findings associate higher liver fat content, measured by MRI-PDFF, with fibrosis progression. What You Need to KnowBackground and ContextThere are limited data regarding liver fat and its association with disease progression in NAFLD.New FindingsIn NAFLD without fibrosis, higher liver fat content on MRI-PDFF was associated with progression of fibrosis.LimitationsThis was a single center study with lack of validation cohort.ImpactHigher liver fat content may have prognostic value if validated in other independent cohorts. There are limited data regarding liver fat and its association with disease progression in NAFLD. In NAFLD without fibrosis, higher liver fat content on MRI-PDFF was associated with progression of fibrosis. This was a single center study with lack of validation cohort. Higher liver fat content may have prognostic value if validated in other independent cohorts. Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in the United States. It encompasses a spectrum of histologically distinguishable disease, ranging from nonalcoholic fatty liver to nonalcoholic steatohepatitis (NASH), a progressive form characterized by steatosis and hepatocellular injury with ballooning and lobular inflammation, with or without fibrosis.1Kleiner D.E. et al.Hepatology. 2005; 41: 1313-1321Crossref PubMed Scopus (4215) Google Scholar The ability to better predict the risk of disease progression, particularly among patients with early NAFLD, is needed to appropriately focus interventions on the population at higher risk. Magnetic resonance imaging (MRI)-based proton density fat fraction (MRI-PDFF) has emerged as an accurate, reproducible biomarker of hepatic steatosis.2Le T.A. et al.Hepatology. 2012; 56: 922-932Crossref PubMed Scopus (122) Google Scholar, 3Park C.C. et al.Gastroenterology. 2017; 152: 598-607.e2Abstract Full Text Full Text PDF PubMed Scopus (388) Google Scholar, 4Mashhood A. et al.J Magn Reson Imaging. 2013; 37: 1359-1370Crossref PubMed Scopus (60) Google Scholar, 5Loomba R. et al.Hepatology. 2015; 61: 1239-1250Crossref PubMed Scopus (125) Google Scholar Despite validation of this biomarker for accurate assessment of steatosis in NAFLD, there are no data on whether the amount of hepatic fat measured by MRI-PDFF can be used to identify patients at risk for progressive disease. Since recent genetic studies have shown an association between single nucleotide polymorphisms associated with fat metabolism and fibrosis, it is biologically plausible that hepatic steatosis may have prognostic significance. Thus, we performed a longitudinal study of 95 adult patients with well-characterized NAFLD with baseline MRI-PDFF and liver biopsy followed by a second liver biopsy to test the association between higher liver fat (defined as MRI-PDFF ≥15.7%) and progression from no fibrosis (stage 0) on baseline histological assessment to a histological diagnosis of any fibrosis (stage 1–4) (Supplementary Methods). Ninety-five patients with NAFLD with paired liver biopsies and baseline MRI were included (Supplementary Figure 1) and were incorporated in the analysis for the primary and secondary outcome based on baseline histology (Supplementary Figure 2). Patients were stratified into 2 groups by the median MRI-PDFF: higher liver fat group (n = 49) with MRI-PDFF ≥15.7% and lower liver fat group with MRI-PDFF <15.6%. Participants had a mean age of 51 years and were predominantly female (61%), 37% were of Hispanic ethnicity, and the mean body mass index (BMI) was 31.8 kg/m2. At baseline, 40% had no fibrosis, 26% stage 1, 8% stage 2, and 25% stage 3–4 fibrosis (Table 1).Table 1Clinical, Demographic, Histologic, and Imaging Characteristics at Baseline by Dichotomized Liver Fat ContentLower liver fat, n = 49Higher liver fat, n = 46Demographic Age, y, mean (SD)52 (15)50 (11) Male, n (%)18 (37)19 (41) BMI (kg/m2), mean (SD)31.8 (4.7)31.9 (4.7) Race/Ethnicity, n (%)White16 (33)21 (46)African American1 (2)0 (0)Asian9 (19)5 (11)Hispanic19 (37)17 (37)Other5 (10)3 (7) Diabetes, n (%)20 (41)13 (28)Biochemical profile AST (U/L), median (IQR)40 (33)39 (26) ALT (U/L), median (IQR)58 (61)58 (46) Alkaline Phosphatase (U/L), median (IQR)74 (27)74 (28) Total bilirubin (mg/dL), median (IQR)0.5 (0.3)0.4 (0.3) Albumin (g/dL), median (IQR)4.5 (0.3)4.5 (0.2) HOMA-IR, median (IQR)6.3 (7.1)7.0 (9.8) Triglycerides (mg/dL), median (IQR)129 (51)167 (59) Total cholesterol (mg/dL), median (IQR)189 (53)190 (54) HDL (mg/dL), median (IQR)46 (20)41 (14) LDL (mg/dL), median (IQR)109 (42)109 (43) Platelet count (109/L), median (IQR)234 (79)249 (84)Histology Time between biopsies, y, mean (SD)1.75 (1.4)1.74 (1.3) Biopsy length (cm), mean (SD)2.2 (1)2.1 (0.5) Fibrosis stage, n (%)017 (35)21 (46)113 (27)12 (26)23 (6)5 (11)38 (16)8 (17)48 (16)0 (0) NASH Classification, n (%)NAFLD not NASH5 (10)5 (11)Borderline NASH2 (4)2 (4)Definite NASH42 (86)39 (85) Steatosis grade, n (%)00 (0)0 (0)129 (59)2 (4)218 (37)17 (37)32 (4)27 (59) Lobular inflammation grade, n (%)00 (0)1 (2)122 (45)16 (35)226 (53)28 (61)31 (2)1 (2) Ballooning grade, n (%)06 (12)7 (15)126 (53)23 (50)217 (35)16 (25) NAS median (IQR), n (%)4 (1)5 (1)Imaging Baseline MRI-PDFF (%), mean (SD)9.4 (3.3)22.3 (6.3)ALT, alanine aminotransferase; AST, aspartate aminotransferase; HDL, high-density lipoprotein; HOMA-IR = homeostatic model assessment method for insulin resistance (calculated as [fasting insulin (μU/mL)*fasting glucose (mmol/L)]/22.5); LDL, low-density lipoprotein; NAS, NAFLD activity score; NASH, Nonalcoholic steatohepatitis; SD, standard deviation. Open table in a new tab ALT, alanine aminotransferase; AST, aspartate aminotransferase; HDL, high-density lipoprotein; HOMA-IR = homeostatic model assessment method for insulin resistance (calculated as [fasting insulin (μU/mL)*fasting glucose (mmol/L)]/22.5); LDL, low-density lipoprotein; NAS, NAFLD activity score; NASH, Nonalcoholic steatohepatitis; SD, standard deviation. Among the 49 patients with lower liver fat and the 46 patients with higher liver fat, the mean (SD) baseline MRI-PDFF value was 9.4% (± 3.3%) and 22.3% (± 6.3%), respectively. The 2 groups were similar with respect to demographic, biochemical, and histological features at baseline. There were statistically significant differences in median (interquartile range [IQR]) triglycerides between lower (129, IQR 51 mg/dL) and higher liver fat (167, IQR 59 mg/dL) groups (P = .019), and in steatosis grade (P < .001). Among the 38 patients with no fibrosis at baseline, patients in the higher liver fat group (MRI-PDFF ≥15.7%) had higher rate of fibrosis progression than the low-liver-fat group (38.1% vs 11.8%, P = .067). After multivariable-adjustment for baseline factors, chosen a priori, including age, sex, BMI, and race/ethnicity, the multivariable-adjusted odds of fibrosis progression was both clinically and statistically significantly higher in patients in the higher liver fat group than the lower liver fat group (odds ratio 6.67; 95% confidence interval [CI] 1.01–44.1; P = .049) (Figure 1). Sensitivity analyses evaluating the adjusted association between higher liver fat and fibrosis progression of at least 1 stage on follow-up biopsy for all patients with < stage 2, < stage 3, and < stage 4 fibrosis at baseline revealed adjusted odds ratios of 2.94 (95% CI 0.79–11.0, P = .11), 2.21 (95% CI 0.68–7.25, P = .19), and 2.58 (95% CI 0.86–7.73, P = .09), respectively. Among 81 patients with definite NASH histology at baseline, 31% in the higher liver fat group vs 43% in the lower liver fat group (P = .26) had no definite NASH histology at follow-up. In additional analysis restricted to patients with definite NASH and no fibrosis at baseline (n = 28), patients in the higher liver fat group had significantly less improvement to non-NASH histology (40% vs 77%, P = .049) (Supplementary Figure 3). This proof of concept study provides novel data on potential prognostic significance of liver fat content in fibrosis progression in NAFLD. Here, we demonstrate that higher liver fat measured by MRI-PDFF at baseline was associated with early histologic progression to fibrotic NAFLD and less improvement in definite NASH in patients without fibrosis. The findings remained significant despite multivariable-adjustment for age, sex, BMI, and ethnicity. Fibrosis stage has been demonstrated recently to be the strongest predictor of both all-cause and liver-related mortality.6Ekstedt M. et al.Hepatology. 2015; 61: 1547-1554Crossref PubMed Scopus (1318) Google Scholar, 7Angulo P. et al.Gastroenterology. 2015; 149: 389-397.e10Abstract Full Text Full Text PDF PubMed Scopus (1692) Google Scholar, 8Dulai P.S. et al.Hepatology. 2017; 65: 1557-1565Crossref PubMed Scopus (963) Google Scholar Genetic variants associated with NAFLD prevalence and fibrosis severity, including PNPLA3 and TM6SF2, appear to be mechanistically linked to hepatic lipid accumulation, suggesting a pathophysiologic link between steatosis and fibrosis severity.9Romeo S. et al.Nat Genet. 2008; 40: 1461-1465Crossref PubMed Scopus (2229) Google Scholar, 10Rotman Y. et al.Hepatology. 2010; 52: 894-903Crossref PubMed Scopus (368) Google Scholar, 11Kozlitina J. et al.Nat Genet. 2014; 46: 352-356Crossref PubMed Scopus (755) Google Scholar, 12Liu Y.L. et al.Nat Commun. 2014; 5: 4309Crossref PubMed Scopus (409) Google Scholar Our study of patients with well-characterized NAFLD with sequential liver biopsies and MRI-PDFF supports recent epidemiologic studies that have demonstrated a link between baseline severity of steatosis and liver-related mortality, as well as increased liver stiffness at follow-up.13Lallukka S. et al.Sci Rep. 2017; 7: 14561Crossref PubMed Scopus (18) Google Scholar, 14Unalp-Arida A. et al.Hepatology. 2016; 63: 1170-1183Crossref PubMed Scopus (63) Google Scholar Although this study provides a rigorous evaluation of the impact of higher liver fat on histologic progression, we acknowledge the following limitations. Fibrosis progression in NAFLD is typically slow, with a previous systematic review suggesting an average fibrosis progression rate of 1 stage every 7 years in patients with NASH.15Singh S. et al.Clin Gastroenterol Hepatol. 2015; 13: 643-654.e1–9Abstract Full Text Full Text PDF PubMed Scopus (826) Google Scholar However, a subset of patients progressed more rapidly, and our study provides new evidence that higher degrees of steatosis may be associated with early fibrotic progression. Only a subset of patients had NAFLD without fibrosis and were included in the primary analysis, but the direction of the association between progressive fibrosis and higher liver fat persisted across all groups of noncirrhotic NAFLD at baseline, although it was not statistically significant in sensitivity analysis. More specific measures of visceral adiposity and genetic data on PNPLA3 and TM6SF2 and higher liver fat were not available. Further studies will be needed to validate these findings and assess the role of genotype in fibrosis progression in longitudinal studies. Although larger, multicenter studies confirming these findings are warranted, these results support the hypothesis that the degree of steatosis has prognostic relevance particularly in early NAFLD and may help identify patients at greater risk for progression to fibrotic NASH. Author contributions: Veeral Ajmera: study concept and design, analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript, study supervision, approved final submission. Charlie C. Park: study concept and design, analysis and interpretation of data, critical revision of the manuscript, approved final submission. Cyrielle Caussy: patient visits, data collection, critical revision of the manuscript, approved final submission. Seema Singh: patient visits, data collection, critical revision of the manuscript, approved final submission. Carolyn Hernandez: patient visits, data collection, critical revision of the manuscript, approved final submission. Ricki Bettencourt: statistical analysis, critical revision of the manuscript, approved final submission. Jonathan Hooker: data collection, imaging analysis, critical revision of the manuscript, approved final submission. Ethan Sy: data collection, imaging analysis, critical revision of the manuscript, approved final submission. Cynthia Behling: interpreted biopsies, critical revision of the manuscript, approved final submission. Ronghui Xu: critical revision of the manuscript, approved final submission. Michael Middleton: data collection, imaging analysis, critical revision of the manuscript, approved final submission. Mark A. Valasek: interpreted biopsies, critical revision of the manuscript, approved final submission. Emily Rizo: patient visits, critical revision of the manuscript, approved final submission. Lisa Richards: patient visits, critical revision of the manuscript, approved final submission. Claude B. Sirlin: study concept and design, analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript, obtained funding, study supervision, approved final submission. Rohit Loomba: study concept and design, analysis and interpretation of data, critical revision of the manuscript, obtained funding, study supervision, approved final submission. All authors approved the final version of this article. This is a longitudinal study derived from a well-characterized prospective cohort of patients with biopsy-proven NAFLD. For this study, participants were included if MRI-PDFF was measured contemporaneously with baseline liver biopsy, and if they had subsequent liver biopsy. Between September 2009 and August 2017, 95 adult patients underwent paired clinical liver biopsies for NAFLD assessment and a baseline research MRI-PDFF examination for hepatic steatosis assessment at the University of California San Diego (UCSD) NAFLD Research Center.1Le T.A. Chen J. Changchien C. et al.Effect of colesevelam on liver fat quantified by magnetic resonance in nonalcoholic steatohepatitis: a randomized controlled trial.Hepatology. 2012; 56: 922-932Crossref PubMed Scopus (190) Google Scholar, 2Loomba R. Sirlin C.B. Ang B. et al.Ezetimibe for the treatment of nonalcoholic steatohepatitis: assessment by novel magnetic resonance imaging and magnetic resonance elastography in a randomized trial (MOZART trial).Hepatology. 2015; 61: 1239-1250Crossref PubMed Scopus (247) Google Scholar, 3Loomba R. Schork N. Chen C.H. et al.Heritability of Hepatic Fibrosis and Steatosis Based on a Prospective Twin Study.Gastroenterology. 2015; 149: 1784-1793Abstract Full Text Full Text PDF PubMed Scopus (221) Google Scholar, 4Doycheva I. Cui J. Nguyen P. et al.Non-invasive screening of diabetics in primary care for NAFLD and advanced fibrosis by MRI and MRE.Aliment Pharmacol Ther. 2016; 43: 83-95Crossref PubMed Scopus (137) Google Scholar, 5Loomba R. Cui J. Wolfson T. et al.Novel 3D Magnetic Resonance Elastography for the Noninvasive Diagnosis of Advanced Fibrosis in NAFLD: A Prospective Study.Am J Gastroenterol. 2016; 111: 986-994Crossref PubMed Scopus (137) Google Scholar All patients provided written informed consent before enrolling in the study and the study was approved by the UCSD Institutional Review Board. Patients ≥18 years of age with biopsy-proven NAFLD and written informed consent were included. Participants meeting any of the following criteria were excluded from the study: significant alcohol consumption (defined as ≥14 drinks per week for men or ≥7 drinks per week for women) within the previous 2-year period; evidence of active substance use; clinical or laboratory evidence of secondary causes or chronic conditions associated with hepatic steatosis, including nutritional disorders, human immunodeficiency virus infection, and use of steatogenic drugs such as amiodarone, glucocorticoids, methotrexate, l-asparaginase, and valproic acid; underlying liver disease other than NAFLD, including viral hepatitis (assessed with serum hepatitis B surface antigen and hepatitis C RNA assays), hemochromatosis, Wilson disease, alpha-1 antitrypsin deficiency, glycogen storage disease, autoimmune hepatitis, and cholestatic or vascular liver disease; major systemic illnesses; decompensated liver disease (defined as Child-Pugh score >7 points); contraindications to MRI, including metallic implants, claustrophobia, and body circumference exceeding the imaging chamber capacity; pregnancy or attempting to be pregnant; or any other conditions believed by the principal investigator to affect the patient’s competence or compliance to complete the study. Patients with paired liver biopsies and MRI within 1 year of the first liver biopsy and temporally closer to the first liver biopsy than the second were included. At baseline, all patients underwent a standardized clinical evaluation, including detailed history, anthropometric examination, and laboratory testing at the UCSD NAFLD Research Center. A trained clinical investigator documented information including age, sex, height, weight, BMI, ethnicity, and vital signs. Alcohol intake history was obtained in a clinical setting and verified at the research clinic with the Alcohol Use Disorders Identification Test and the Skinner questionnaire. Other causes of liver disease and hepatic steatosis were ruled out systematically based on history and laboratory tests. Participants were instructed to fast for a minimum of 8 hours before collection of laboratory tests. All patients underwent a baseline liver biopsy, followed by a second liver biopsy for assessment by an experienced liver pathologist blinded to patients’ clinical or imaging data. This study used the Nonalcoholic Steatohepatitis Clinical Research Network Histologic Scoring System, in which hepatic fibrosis was scored on a scale from 0 to 4 (0, 1, 2, 3, 4), with stage 4 signifying cirrhosis, hepatic steatosis and lobular inflammation were scored from 0 to 3 (0, 1, 2, 3), and hepatic ballooning was scored from 0 to 2 (0, 1, 2).6Kleiner D.E. Brunt E.M. Van Natta M. et al.Design and validation of a histological scoring system for nonalcoholic fatty liver disease.Hepatology. 2005; 41: 1313-1321Crossref PubMed Scopus (7151) Google Scholar Steatosis, lobular inflammation, and hepatocyte ballooning scores were summed to obtain the NAFLD activity score,7Neuschwander-Tetri B.A. Loomba R. Sanyal A.J. et al.Farnesoid X nuclear receptor ligand obeticholic acid for non-cirrhotic, non-alcoholic steatohepatitis (FLINT): a multicentre, randomised, placebo-controlled trial.Lancet. 2015; 385: 956-965Abstract Full Text Full Text PDF PubMed Scopus (1526) Google Scholar which ranged from 0 to 8. Diagnosis of NASH was classified as definite NASH, NAFLD not NASH, or “borderline” NASH. These categories were assigned before conducting statistical analyses. Abdominal MRI, including an MRI-PDFF sequence, was obtained on a single 3T MR scanner (GE Signa EXCITE HDxt; GE Healthcare, Waukesha, WI) at the UCSD MR3T Research Laboratory using previously described methods.8Tang A. Desai A. Hamilton G. et al.Accuracy of MR imaging-estimated proton density fat fraction for classification of dichotomized histologic steatosis grades in nonalcoholic fatty liver disease.Radiology. 2015; 274: 416-425Crossref PubMed Scopus (176) Google Scholar, 9Tang A. Tan J. Sun M. et al.Nonalcoholic fatty liver disease: MR imaging of liver proton density fat fraction to assess hepatic steatosis.Radiology. 2013; 267: 422-431Crossref PubMed Scopus (330) Google Scholar, 10Reeder S.B. Robson P.M. Yu H. et al.Quantification of hepatic steatosis with MRI: the effects of accurate fat spectral modeling.J Magn Reson Imaging. 2009; 29: 1332-1339Crossref PubMed Scopus (199) Google Scholar, 11Hines C.D. Frydrychowicz A. Hamilton G. et al.T(1) independent, T(2) (*) corrected chemical shift based fat-water separation with multi-peak fat spectral modeling is an accurate and precise measure of hepatic steatosis.J Magn Reson Imaging. 2011; 33: 873-881Crossref PubMed Scopus (170) Google Scholar, 12Permutt Z. Le T.A. Peterson M.R. et al.Correlation between liver histology and novel magnetic resonance imaging in adult patients with non-alcoholic fatty liver disease - MRI accurately quantifies hepatic steatosis in NAFLD.Aliment Pharmacol Ther. 2012; 36: 22-29Crossref PubMed Scopus (252) Google Scholar The mean PDFF across 9 regions of interest was evaluated, which has been demonstrated to accurately correlate with both liver histology and magnetic resonance spectroscopy.13Yokoo T. Serai S.D. Pirasteh A. et al.Linearity, Bias, and Precision of Hepatic Proton Density Fat Fraction Measurements by Using MR Imaging: A Meta-Analysis.Radiology. 2017; : 170550Google Scholar, 14Middleton M.S. Heba E.R. Hooker C.A. et al.Agreement Between Magnetic Resonance Imaging Proton Density Fat Fraction Measurements and Pathologist-Assigned Steatosis Grades of Liver Biopsies From Adults With Nonalcoholic Steatohepatitis.Gastroenterology. 2017; 153: 753-761Abstract Full Text Full Text PDF PubMed Scopus (167) Google Scholar, 15Middleton M.S. Van Natta M.L. Heba E.R. et al.Diagnostic accuracy of magnetic resonance imaging hepatic proton density fat fraction in pediatric nonalcoholic fatty liver disease.Hepatology. 2017; PubMed Google Scholar, 16Heba E.R. Desai A. Zand K.A. et al.Accuracy and the effect of possible subject-based confounders of magnitude-based MRI for estimating hepatic proton density fat fraction in adults, using MR spectroscopy as reference.J Magn Reson Imaging. 2016; 43: 398-406Crossref PubMed Scopus (36) Google Scholar The median and IQR time intervals between the baseline liver biopsy and MRI were 35 and 42 days, respectively. The primary outcome was progression from no fibrosis (stage 0) on baseline histological assessment to a histological diagnosis of any fibrosis (stage 1–4) on follow-up assessment. The secondary outcome was improvement from definite NASH to not definite NASH histology (not NASH or borderline NASH) on follow-up. Patients were divided into 2 groups: 49 patients at or below the median MRI-PDFF and 46 patients above the median MRI-PDFF. We hypothesized that the risk of fibrosis progression in the higher liver fat group would be 60% compared with 15% in the lower liver fat group, and power analysis showed that a sample size of 35 would provide 80% power with a 2-tailed alpha of 0.05. Therefore, we had adequate power to detect the aforementioned difference in fibrosis progression with our n = 38 patients without fibrosis at baseline. Descriptive statistics of participant demographic, laboratory, histological, and imaging characteristics at baseline were compared by dichotomized MRI-PDFF–determined liver fat. Univariate and multivariate logistic regression analyses to assess for the association between higher compared with lower liver fat and (1) progression to any fibrosis among patients with no fibrosis at baseline and (2) improvement from definite NASH to not definite NASH among patients with NASH at baseline was performed. Multivariate analysis included baseline demographic and anthropometric characteristics chosen a priori for their known association with NAFLD and included BMI, age, race/ethnicity, and sex. Statistical significance was defined as a 2-tailed P ≤ .05. All statistical analyses were performed on STATA (StataCorp LP, College Station, TX). Supplementary Figure 2View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplementary Figure 3View Large Image Figure ViewerDownload Hi-res image Download (PPT)