Differential Impact of Weight Loss on Nonalcoholic Fatty Liver Resolution in a North American Cohort with Obesity

Transcription

1 OBESITY BIOLOGY AND INTEGRATED HYSIOLOGY Differential Impact of Weight Loss on Nonalcoholic Fatty Liver Resolution in a North American Cohort with Vikrant Rachakonda 1, Rachel Wills 2, James. DeLany 2, Erin E. Kershaw 2, and Jaideep Behari 1 Objective: Nonalcoholic fatty liver disease (NAFLD) is closely associated with obesity. In this study, a North American cohort with obesity enrolled in a lifestyle modification program was examined to determine the impact of weight loss on NAFLD resolution and sarcopenia. Methods: Nondiabetic individuals with World Health Organization Class II/III obesity enrolled in a 6- month weight loss intervention were included. Steatosis was measured using computed tomography (CT)-derived liver:spleen attenuation ratio. Body composition was assessed using dual X-ray absorptiometry, air-displacement plethysmography, and CT anthropometry. Results: At baseline, participants with NAFLD had greater visceral adipose tissue (VAT) but similar skeletal muscle area compared to those without NAFLD. After intervention, weight loss was similar in the two groups, but participants with NAFLD lost more VAT than those without NAFLD ( [ to ] cm 2 vs [ to 22.02] cm 2 ; ). In the subset with NAFLD at baseline, participants with NAFLD resolution after intervention lost more VAT than those with persistent NAFLD ( [ to ) cm 2 vs [ to ] cm 2, ). Conclusions: In a Western cohort with obesity, NAFLD was not associated with sarcopenia. After lifestyle modification, there was a differential impact on NAFLD resolution, with twofold greater VAT loss in participants who resolved NAFLD compared with those with persistent NAFLD despite similar weight loss (2017) 00, doi: /oby Introduction With the growing epidemic of obesity and metabolic syndrome, nonalcoholic fatty liver disease (NAFLD) has become the leading etiology of chronic liver disease worldwide (1). Regional distribution of lean and fat mass may influence the development of NAFLD. revious studies have assessed the impact of abdominal adiposity in NAFLD, but both negative and positive associations with visceral and subcutaneous adiposity have been reported (2-7). Decreased muscle mass, termed sarcopenia, was recently identified as a risk factor for NAFLD in Korean populations, and this effect was independent of insulin sensitivity or obesity (8,9). Although studies of US populations have described associations between sarcopenia and obesity-related insulin resistance (10), only one has described decreased muscle mass in participants with NAFLD (11). Lifestyle modification targeting weight loss is currently the most effective therapy for NAFLD. However, the relationship between quantity of weight loss and clinical benefit remains unclear. Early studies have suggested that weight loss of 7% to 10% (12-15) was associated with histologic improvements in steatosis and inflammation, while more recent work has shown that as little as 5% weight loss may result in regression of fibrosis (16). Furthermore, exercisebased interventions in NAFLD reduce intrahepatic triglyceride content independent of weight change (17,18). These disparate findings raise the questions of whether individuals with obesity in Western populations with NAFLD also have sarcopenia and whether weight loss interventions in these patients with NAFLD differentially impact lean and adipose tissue compartments. To address this gap in knowledge, we analyzed a cohort of individuals with severe obesity enrolled in a weight loss interventional trial to assess the relationship between body composition and NAFLD. The aims of our study were to determine whether NAFLD is associated with sarcopenia in individuals with obesity and whether weight loss 1 Division of Gastroenterology, Hepatology, and Nutrition, Department of Medicine, University of ittsburgh, ittsburgh, ennsylvania, USA. Correspondence: Jaideep Behari (beharij@upmc.edu) 2 Division of Endocrinology and Metabolism, Department of Medicine, University of ittsburgh School of Medicine, ittsburgh, ennsylvania, USA. Disclosure: The authors declared no conflicts of interest. Author contributions: VR and JB: study planning and conceptualization; RW, JD, and EEK: data collection; VR and JB: data analysis; VR and JB: drafting manuscript; EEK: critical review of the manuscript. Clinical trial registration: ClinicalTrials.gov identifier NCT Received: 15 January 2017; Accepted: 19 April 2017; ublished online 00 Month doi: /oby VOLUME 00 NUMBER 00 MONTH

2 NAFLD Resolution with Weight Loss Rachakonda et al. interventions differentially affect total and regional skeletal muscle and adipose tissue mass in participants with and without NAFLD. We report that in a Western population with severe obesity, the presence of NAFLD was associated with higher visceral adipose tissue (VAT) mass but not with sarcopenia. Furthermore, after the intensive lifestyle intervention program, loss of VAT was greater in individuals with NAFLD than those without NAFLD, and participants who had resolution of NAFLD lost more VAT than those participants in whom NAFLD persisted despite similar total weight loss. Methods Study design and participants This study was performed in a cohort of participants previously enrolled in a clinical trial of weight loss interventions for severe obesity (RENEW, ClinicalTrials.gov trial registration identifier: NCT ) (19). The Institutional Review Board at the University of ittsburgh approved the study, and all participants provided written consent prior to participation. From February 2007 to March 2009, men and women between 30 and 55 years of age were enrolled in a single-blind randomized control trial comparing a dietary weight loss intervention to multimodal weight loss treatment with diet and physical activity. Inclusion criteria included World Health Organization Class II or III obesity (defined as BMI 35 kg/ m 2 ), ability to walk without assistance, and ability to obtain medical clearance for dietary and physical activity interventions. Exclusion criteria included history of coronary artery disease, diagnosis of cancer within 5 years of enrollment, prior bariatric surgery, prior participation in a weight loss program within 1 year of enrollment, history of diabetes mellitus, uncontrolled hypertension, and pregnancy within 6 months of enrollment. articipants with liver enzyme elevations more than 30% above the upper limit of normal laboratory ranges were excluded. The details of the dietary and physical activity interventions have been previously described (19). All clinical, radiographic, laboratory, and anthropometric measurements described below were obtained at baseline in a cohort of 129 participants. A subset (n 5 112) underwent measurements both at baseline and after 6 months of intervention. Determination of NAFLD Hepatic steatosis was evaluated with hepatic and splenic attenuation data from unenhanced abdominal computed tomography (CT) scans as previously described (19,20). Multiple studies have demonstrated that CT-derived indices of liver and spleen attenuation are highly sensitive for identification of moderate steatosis (21-23). In particular, liver:spleen attenuation ratio (L/S ratio) strongly correlates with both histologic quantification of steatosis and intrahepatic triglyceride content (24). As an L/S ratio <1.1 exhibits more than 80% accuracy for identification of 30% or more macrovesicular steatosis, NAFLD was defined as an L/S ratio <1.1. Assessment of body composition Whole-body fat-free mass (FFM) and fat mass (FM) were measured using either dual X-ray absorptiometry or air-displacement plethysmography in participants exceeding the weight capacity of the scanner (136 kg) (19). Unenhanced abdominal CT scans were obtained at baseline and at 6 months to quantify VAT cross-sectional area and abdominal subcutaneous adipose area at the L4/L5 intervertebral level (19,20). Unenhanced CT was also used to measure midthigh skeletal muscle and a subcutaneous cross-sectional area as previously described (19,20). Demographic and clinical evaluation articipant race and ethanol use were self-reported. A detailed assessment of ethanol use was performed as previously described (19,25). articipants underwent measurements of height and weight to calculate BMI. Systolic and diastolic blood pressures were measured using an automated sphygmomanometer. Use of tobacco, antihypertensive agents, and lipid-lowering medications was selfreported. Clinical laboratory parameters were assessed in serum obtained from participants after a 12-hour fast. These parameters included aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase, creatinine, cholesterol (total, very-lowdensity lipoprotein [VLDL], low-density lipoprotein [LDL], highdensity lipoprotein [HDL]), triglycerides, glucose, and insulin. The homeostasis model assessment of insulin resistance was calculated as glucose (mg/dl) 3 insulin (mu/l)/405. Adipokine and cytokine measurement Whole blood specimens were previously collected as part of the RENEW trial. Serum was isolated by centrifugation, aliquoted, and immediately stored at 2808C for future use. Serum levels of interleukin 6 (IL-6), C-reactive protein (CR), adiponectin, and leptin were previously measured as part of other trials (26,27). Statistical analysis Statistical analysis was performed with Stata version 13.0 (StataCorp., Boston, Massachusetts) and Graphad rism version 6.0 (Graphad Software, Inc., La Jolla, California). Categorical variables were reported as absolute frequencies and percentages, and continuous variables were reported as means and 95% confidence intervals. Categorical variables were compared between groups using Fisher s exact test. Baseline values of continuous variables were compared between groups using Welch s t test. Two-way repeated measures analysis of variance (ANOVA) was used to compare postintervention changes between groups. ost hoc paired t tests were performed to assess within-group differences, and unpaired t tests were used to determine differences between group differences. Holm-Sidak methods were applied to correct for multiple comparisons. <0.05 was considered statistically significant. Results Clinical and demographic features of patients Among 129 participants, 11.6% were male, and there were more men with NAFLD than without (Table 1). There were no differences in age, ethnic distribution, antihypertensive medication use, lipidlowering medication use, or tobacco use between individuals with and without NAFLD. The mean BMI of the study cohort was 43.5 ( ), and participants with NAFLD had significantly higher BMI and waist circumference than individuals without NAFLD. There were no differences in systolic or diastolic blood pressures in participants with and without NAFLD. 2 VOLUME 00 NUMBER 00 MONTH

3 OBESITY BIOLOGY AND INTEGRATED HYSIOLOGY TABLE 1 Baseline demographic, clinical, and metabolic features Total (N 5 129) NAFLD (n 5 58) No NAFLD (n 5 71) Male gender, n (%) 15 (11.6%) 14 (24.1%) 1 (1.4%) <0.001 Age (y) 47.6 ( ) 47.6 ( ) 47.6 ( ) African American, n (%) 48 (37.2%) 17 (29.3%) 31 (43.7%) Tobacco use, n (%) 11 (8.5%) 4 (6.9%) 7 (9.9%) Antihypertensive medication use, n (%) 80 (62.0%) 31 (53.4%) 49 (69.0%) Lipid-lowering medication use, n (%) 12 (9.3%) 3 (5.2%) 9 (12.7%) Weight, kg ( ) ( ) ( ) <0.001 BMI, kg/m ( ) 45.5 ( ) 42.0 ( ) <0.001 Waist circumference, cm ( ) ( ) ( ) <0.001 Systolic blood pressure, mm Hg 135 ( ) 136 ( ) 134 ( ) Diastolic blood pressure, mm Hg 78 (76-79) 78 (76-80) 77 (75-79) Fasting blood glucose, mg/dl 93.5 ( ) 97 ( ) 90.1 ( ) <0.001 Insulin, mu/l 16.6 ( ) 19.0 ( ) 14.6 ( ) HOMA-IR index 3.88 ( ) 4.60 ( ) 3.29 ( ) Total cholesterol, mg/dl ( ) ( ) ( ) HDL, mg/dl 48.0 ( ) 45.2 ( ) 50.2 ( ) LDL, mg/dl ( ) ( ) ( ) VLDL, mg/dl 20.8 ( ) 22.9 ( ) 19.1 ( ) Triglycerides, mg/dl ( ) ( ) ( ) Metabolic syndrome, n (%) 75 (58.1%) 40 (69.0%) 35 (48.6%) AST, U/L 25.4 ( ) 27.1 ( ) 24.0 ( ) ALT, U/L 30.4 ( ) 33.3 ( ) 28.1 ( ) Alkaline phosphatase, U/L 86.5 ( ) 86.8 ( ) 86.3 ( ) Creatinine, mg/dl 0.83 ( ) 0.83 ( ) 0.83 ( ) Weight loss intervention, n (%) Diet only 48.8% 53.4% 49.3% Diet and exercise 51.2% 46.6% 50.7% Adiponectin, mg/ml ( ) ( ) ( ) Leptin, ng/ml ( ) ( ) ( ) IL-6, pg/ml 2.76 ( ) 2.93 ( ) 2.62 ( ) C-reactive protein, mg/ml 8.64 ( ) 9.32 ( ) 8.08 ( ) All values reported as means and 95% confidence intervals unless otherwise specified. ALT, alanine aminotransferase; AST, aspartate aminotransferase; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment of insulin resistance; IL-6, interleukin 6; LDL, low-density lipoprotein; VLDL, very-low-density lipoprotein. articipants with NAFLD exhibited greater metabolic dysfunction, as fasting glucose, insulin, and homeostasis model assessment of insulin resistance were higher in participants with steatosis; HDL was lower, while LDL, VLDL, and triglycerides were higher in participants with NAFLD. AST and ALT levels were mildly elevated in the NAFLD cohort, but fasting levels of CR, IL-6, leptin, and adiponectin did not differ between groups (Table 1). NAFLD is associated with higher VAT mass but not sarcopenia Individuals with NAFLD had higher absolute FM and FFM than those without NAFLD, but percent FM and FFM were similar between groups (Table 2). Cross-sectional midthigh skeletal muscle area was elevated in NAFLD, but both midthigh and abdominal subcutaneous depots were not significantly different in patients with and without NAFLD. Abdominal VAT cross-sectional area was markedly higher in participants with NAFLD. Skeletal muscle mass, as assessed by CT, dual X-ray absorptiometry, and plethysmography, was similar in patients with and without NAFLD. However, participants with NAFLD exhibited a body composition phenotype notable for greater visceral adiposity. Because body composition is dependent on gender and our study cohort was predominately female, we re-analyzed body composition differences after excluding male participants, and our findings on body composition were qualitatively unchanged (data not shown). Weight loss intervention reduces adipose tissue and improves hepatic insulin resistance in individuals with NAFLD A subset of 112 participants underwent a 6-month structured weight loss intervention consisting of either dietary modification alone (DO) or with physical activity (DO1A). Fifty-two participants VOLUME 00 NUMBER 00 MONTH

4 NAFLD Resolution with Weight Loss Rachakonda et al. TABLE 2 Baseline anthropometric measures Total (N 5 129) NAFLD (n 5 58) No NAFLD (n 5 71) Fat-free mass, kg 57.5 ( ) 61.7 ( ) 54.5 ( ) <0.001 Fat-free mass index, kg/m ( ) 22.2 ( ) 20.2 ( ) <0.001 ercent fat-free mass, % 48.7% (47.8%-49.5%) 49.1% (47.7%-50.4%) 48.4% (47.4%-49.4%) Fat mass, kg 59.8 ( ) 62.8 ( ) 57.3 ( ) Fat mass index, kg/m ( ) 22.8 ( ) 21.7 ( ) ercent fat mass, % 50.1% (49.3%-50.9%) 49.8% (48.6%-51.1%) 50.4% (49.3%-51.4%) Midthigh muscle cross-sectional area, cm ( ) ( ) ( ) Midthigh muscle area index, cm 2 /m ( ) 53.3 ( ) 48.8 ( ) Midthigh muscle area/wt, cm 2 /kg 1.2 ( ) 1.2 ( ) 1.2 ( ) Midthigh normal-density muscle, cm ( ) ( ) ( ) <0.001 Midthigh intramuscular fat, cm ( ) 24.1 ( ) 23.0 ( ) Midthigh normal-density muscle, % 82.9% (21.8%-84.0%) 83.7% (82.2%-85.2%) 82.2% (80.7%-83.8%) Midthigh intramuscular fat, % 17.1% (16.0%-18.2%) 16.3% (14.8%-17.8%) 17.8% (16.2%-19.3%) Visceral adipose tissue area, cm ( ) ( ) ( ) Visceral adipose tissue index, cm 2 /m ( ) 78.4 ( ) 63.3 ( ) Subcutaneous adipose area, cm ( ) ( ) ( ) Subcutaneous adipose area index, cm 2 /m ( ) ( ) ( ) Midthigh subcutaneous adipose area, cm ( ) ( ) ( ) All values reported as means and 95% confidence intervals unless otherwise specified. (46.0%) met imaging criteria for NAFLD. There were no differences in age and ethnic distribution between individuals with and without NAFLD undergoing weight loss intervention; however, there were more men in the NAFLD cohort (Table 3). There was no difference in prescribed treatment regimens between groups, as similar proportions of participants with and without NAFLD were randomized to the DO and DO1A arms ( ); hence, the effect of weight loss modality would be similarly distributed between groups. TABLE 3 Anthropometric changes after structured lifestyle intervention in subjects with and without NAFLD NAFLD (n 5 52) No NAFLD (n 5 60) (NAFLD) (time 3 NAFLD) Female gender, n (%) 41 (78.8%) 59 (98.3%) <0.001 Age, y 47.6 (41.1 to 52.0) 47.6 (43.0 to 52.9) Caucasian, n (%) 38 (73.1%) 35 (58.3%) Weight loss intervention, n (%) Diet only 25 (48.1%) 28 (46.7%) Diet and exercise 27 (51.9%) 32 (53.3%) Weight change, kg ( to 29.22) < ( to 26.36) <0.001 <0.001 < % or greater weight loss, n (%) 38 (73.1%) 42 (70.0%) Absolute BMI change, kg/m (24.88 to 23.36) < (23.81 to 22.40) < Waist circumference change, cm ( to 24.80) < ( to ) <0.001 < FFM change, kg (23.53 to 21.92) < (22.53 to 21.02) <0.001 < FM change, kg ( to 26.73) < (28.17 to 24.85) < Midthigh muscle area change, cm ( to 22.86) (27.09 to 21.45) < Visceral adipose area change, cm (55.98 to ) < ( to 22.02) Abdominal SC area change, cm ( to ) < ( to ) < Midthigh SC area change, cm ( to ) < ( to ) < All values reported as means and 95% confidence intervals unless otherwise specified. ost hoc paired t tests performed to assess within-group differences, and unpaired t tests used to determine differences between group differences (NAFLD vs. no NAFLD). Holm-Sidak methods applied to correct for multiple comparisons. FFM, fat-free mass; FM, fat mass; SC, subcutaneous. 4 VOLUME 00 NUMBER 00 MONTH

5 OBESITY BIOLOGY AND INTEGRATED HYSIOLOGY TABLE 4 Changes in metabolic measures after weight loss intervention by presence or absence of NAFLD NAFLD (n 5 52) No NAFLD (n 5 60) (NAFLD) (time 3 NAFLD) AST change, IU/L 23.0 (24.9 to 21.2) (22.4 to 1.0) ALT change, IU/L 24.3 (26.5 to 22.0) < (22.2 to 1.9) Alkaline phosphatase change, IU/L 28.9 (213.0 to 24.8) < (212.0 to 24.3) < Fasting glucose change, IU/L 26.5 (29.6 to 23.3) < (23.1 to 22.8) Insulin change, lu/l (27.85 to 22.32) < (26.55 to 21.41) HOMA-IR change, a.u (22.11 to 20.68) < (21.60 to 20.28) < Total cholesterol change, mg/dl 26.0 (212.9 to 1.0) (24.6 to 8.0) HDL change, mg/dl 23.3 (25.6 to 21.1) (24.6 to 20.5) LDL change, mg/dl 24.8 (211.1 to 1.5) (22.8 to 8.8) VLDL change, mg/dl 21.9 (22.9 to 6.6) (24.1 to 4.7) Triglyceride change, mg/dl 3.3 (12.8 to 19.4) (212.1 to 17.5) IL-6 change, pg/ml 1.0 (21.1 to 3.0) (2.1 to 5.9) < C-reactive protein change, lg/ml 24.4 (26.6 to 22.2) < (26.9 to 22.8) < Leptin change, ng/ml (220.2 to 29.2) < (216.7 to 26.5) < Adiponectin change, lg/ml 20.8 (21.9 to 20.3) (21.2 to 0.8) All values reported as means and 95% confidence intervals unless otherwise specified. ost hoc paired t tests used to determine within-group group differences, and unpaired t tests used to determine between-group differences (NAFLD vs. no NAFLD). Holm-Sidak methods applied to correct for multiple comparisons. ALT, alanine aminotransferase; AST, aspartate aminotransferase; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment of insulin resistance; IL-6, interleukin 6; LDL, low-density lipoprotein; VLDL, very-low-density lipoprotein. articipants with NAFLD lost more weight after 6 months of therapy, but there were no statistically significant differences in BMI change between groups (Table 3). Furthermore, a similar proportion with and without NAFLD lost at least 5% body mass (38 [73.1%] participants with NAFLD vs. 42 [70.0%] participants without NAFLD; ). While all participants exhibited significant decreases in FM, FFM, and both abdominal and midthigh subcutaneous adipose tissue areas, midthigh muscle cross-sectional area was not significantly reduced after weight loss intervention. articipants with NAFLD, however, lost significantly more VAT compared to those without NAFLD ( [ to ] cm 2 vs [ to 22.02] cm 2 ; for time 3 group interaction). Together, these findings suggest that weight loss interventions lead to different depot-specific responses depending on whether NAFLD is present. Furthermore, VAT is preferentially affected by weight loss intervention in participants with NAFLD. Changes in metabolic and inflammatory markers after weight loss intervention articipants with NAFLD experienced greater postintervention declines in serum transaminases and fasting serum glucose levels (Table 4). As intrahepatic fat contributes to hepatic insulin resistance as well as hepatocellular injury, improvements in fasting glucose and transaminases may be related to greater reductions in intrahepatic fat after weight loss in NAFLD. Serum leptin, adiponectin, and CR levels decreased similarly in patients with and without NAFLD after weight loss, but lipid profiles were not significantly improved in either cohort. Interestingly, postintervention IL-6 levels increased in patients without NAFLD but remained unchanged in patients with NAFLD; as visceral fat is an important source of IL-6, differences in IL-6 change may be influenced by VAT responses to weight loss. These findings demonstrate that individuals with NAFLD experience greater postintervention reductions in hepatic insulin resistance, hepatocellular injury, and IL-6 secretion that those without NAFLD. Reduction of VAT mass is associated with resolution of NAFLD We next assessed the effects of weight loss on body composition in NAFLD subjects to determine whether postintervention changes were associated with resolution of NAFLD, which was defined as a postintervention L/S ratio 1.1. Among 52 subjects with NAFLD, 20 (38.4%) experienced resolution of NAFLD. There were no differences in the assignment of weight loss interventions between subjects with and without NAFLD resolution (10 [50%] resolved NAFLD subjects vs. 15 [46.9%] persistent NAFLD subjects underwent DO [ ]), thus suggesting that weight loss modalities did not differentially influence NAFLD resolution (Table 5). Treatment nonresponders had greater transaminase levels and a lower L/S ratio, but all other measured metabolic markers, inflammatory adipocytokines, and body composition phenotypes were similar between groups. Weight loss intervention was largely successful in patients with NAFLD, and both responders and nonresponders experienced similar degrees of weight loss (Table 6). Furthermore, similar proportions of individuals with and without treatment response lost at least 5% of their body mass; nonresponders, however, experienced smaller improvements in BMI. While postintervention changes in FFM were similar between groups, participants with NAFLD resolution experienced greater declines in FM, which was due to enhanced VAT loss. This finding was not accompanied by significant changes in metabolic and inflammatory markers; only postintervention changes in transaminase levels were significantly different between VOLUME 00 NUMBER 00 MONTH

6 NAFLD Resolution with Weight Loss Rachakonda et al. TABLE 5 Baseline features of subjects with NAFLD undergoing lifestyle intervention NAFLD resolved (n 5 20) NAFLD persistent (n 5 32) Male gender, n (%) 2 (10.0%) 9 (28.1%) Age, y 46.0 ( ) 47.2 ( ) African American, n (%) 5 (25.0%) 9 (28.1%) Weight loss intervention, n (%) Diet only 10 (50.0%) 15 (46.9%) Diet and exercise 10 (50.0%) 17 (53.1%) Weight, kg ( ) ( ) % or greater weight loss, n (%) 18 (90.0%) 20 (62.5%) BMI, kg/m ( ) 45.4 ( ) Waist circumference, cm ( ) ( ) FFM, kg 59.1 ( ) 63.2 ( ) FFM index, kg/m ( ) 22.2 ( ) ercent FFM mass, % 48.2% (45.7%-50.7%) 49.4% (47.7%-51.2%) FM, kg 62.7 ( ) 63.4 ( ) FM, kg/m ( ) 22.6 ( ) ercent FM, % 50.7% (48.3 %-53.2%) 49.5% (47.8%-51.2%) Midthigh muscle cross-sectional area, cm ( ) ( ) Midthigh muscle area index, cm 2 /m ( ) 54.2 ( ) Midthigh muscle area/wt, cm 2 /kg 1.13 ( ) 1.21 ( ) Visceral adipose tissue area, cm ( ) ( ) Visceral adipose tissue index, cm 2 /m ( ) 78.5 ( ) Subcutaneous adipose area, cm ( ) ( ) Subcutaneous adipose area index, cm 2 /m ( ) ( ) Midthigh subcutaneous adipose area, cm ( ) ( ) Midthigh subcutaneous adipose index, cm 2 /m ( ) 71.3 ( ) Fasting blood glucose, mg/dl 99.1 ( ) 96.3 ( ) Insulin, mu/l 17.5 ( ) 19.4 ( ) HOMA-IR index 4.25 ( ) 4.67 ( ) Total cholesterol, mg/dl ( ) ( ) HDL, mg/dl 44.5 ( ) 45.8 ( ) LDL, mg/dl ( ) ( ) VLDL, mg/dl 19.4 ( ) 24.1 ( ) Triglycerides, mg/dl ( ) ( ) AST, U/L 23.7 ( ) 29.3 ( ) ALT, U/L 30.0 ( ) 35.8 ( ) Alkaline phosphatase, U/L 89.6 ( ) 85.2 ( ) Adiponectin, mg/ml ( ) ( ) Leptin, ng/ml 58.3 ( ) 52.7 ( ) IL-6, pg/ml 2.95 ( ) 2.79 ( ) C-reactive protein, mg/ml 8.24 ( ) 9.38 ( ) Liver:spleen ratio, a.u ( ) 0.79 ( ) <0.001 All values reported as means and 95% confidence intervals unless otherwise specified. ALT, alanine aminotransferase; AST, aspartate aminotransferase; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment of insulin resistance; IL-6, interleukin 6; LDL, low-density lipoprotein; VLDL, very-low-density lipoprotein; FM, fat mass; FFM, fat-free mass. responders and nonresponders (Table 7). Of all measured anthropometric, clinical, and metabolic factors, VAT loss alone was associated with NAFLD resolution after weight loss intervention. Finally, we assessed postintervention changes in L/S ratio and found no differences between subjects with and without NAFLD resolution (0.208 [ ] a.u. in responders vs [ ] a.u. in nonresponders; for time 3 group interaction) (Figure 1). Despite no significant differences in baseline weight and body composition, patients with NAFLD resolution appeared to have lower baseline intrahepatic fat content and greater visceral fat loss than treatment nonresponders. 6 VOLUME 00 NUMBER 00 MONTH

7 OBESITY BIOLOGY AND INTEGRATED HYSIOLOGY TABLE 6 Body composition changes after lifestyle intervention in NAFLD subjects stratified by treatment response NAFLD resolved (n 5 20) NAFLD persist (n 5 32) (group) (time 3 group) Liver:spleen ratio change, a.u (0.124 to 0.292) < (0.109 to 0.243) < Weight change, kg ( to ) < ( to 26.81) <0.001 < % or greater weight loss, n (%) 18 (90.0%) 20 (62.5%) Absolute BMI change, kg/m (26.4 to 23.8) < (24.6 to 22.5) < Waist circumference change, cm 29.6 (217.9 to 21.3) (213.4 to 20.4) FFM change, kg (24.63 to 21.59) (23.69 to 21.28) < FM change, kg ( to 27.57) < (29.61 to 24.84) < Midthigh muscle area change, cm ( to 1.94) ( to 0.01) Visceral adipose area change, cm ( to ) < ( to ) Abdominal SC area change, cm ( to ) < ( to ) < Midthigh SC area change, cm ( to ) < ( to ) < All values reported as means and 95% confidence intervals unless otherwise specified. ost hoc paired t tests performed to assess within-group differences, and unpaired t tests were used to determine differences between group differences (NAFLD resolution vs NAFLD persistence). Holm-Sidak methods applied to correct for multiple comparisons. FFM, fat-free mass; FM, fat mass; SC, subcutaneous. Discussion In this study, we analyzed a North American cohort of individuals without diabetes and with severe obesity enrolled in a lifestyle modification interventional trial to assess the relationship between NAFLD and body composition before and after weight loss. We report three key observations. First, subjects with NAFLD did not exhibit lower muscle mass compared to those without NAFLD but had greater visceral adiposity and higher fasting blood glucose. Second, despite similar degrees of BMI reduction in response to weight loss intervention, subjects with NAFLD experienced greater improvements in VAT mass, waist circumference, hepatic insulin sensitivity, and serum transaminases than those without NAFLD. Third, the subsets of individuals with resolution of NAFLD after lifestyle modification experienced a twofold greater decrease in VAT compared to those with persistent NAFLD. revious work has suggested that NAFLD is associated with sarcopenia (8,9,11). In an analysis of more than 4,000 Korean individuals, TABLE 7 Changes in metabolic measures after weight loss intervention in NAFLD subjects stratified by treatment response NAFLD resolved (n 5 20) NAFLD persists (n 5 32) (group) (time 3 group) AST change, IU/L 21.0 (23.7 to 1.7) (-6.5 to 22.1) < ALT change, IU/L 21.9 (25.4 to 1.7) (-8.7 to 22.9) < Alkaline phosphatase change, IU/L 28.5 (215.3 to 21.7) (-14.7 to 23.7) Fasting glucose change, IU/L 29.5 (215.7 to 23.4) (29.5 to 0.4) Insulin change, lu/l ( to 0.20) (29.05 to 21.14) HOMA-IR change, a.u ( to ) ( to ) Total cholesterol change, mg/dl (235.4 to 14.3) (222.0 to 16.7) HDL change, mg/dl 22.8 (26.5 to 0.9) (26.5 to 20.7) LDL change, mg/dl (218.9 to 21.7) (28.1 to 5.4) VLDL change, mg/dl 1.2 (26.2 to 8.6) (23.3 to 7.9) Triglyceride change, mg/dl 4.2 (216.8 to 25.2) (213.7 to 19.2) IL-6 change, pg/ml ( to 2.599) ( to 2.901) C-reactive protein change, lg/ml ( to ) ( to ) < Leptin change, ng/ml ( to ) < ( to 24.26) Adiponectin change, lg/ml ( to 1.602) ( to 0.384) All values reported as means and 95% confidence intervals unless otherwise specified. ost hoc paired t tests performed to assess within-group differences, and unpaired t tests used to determine differences between group differences (NAFLD resolution vs. NAFLD persistence). ALT, alanine aminotransferase; AST, aspartate aminotransferase; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment of insulin resistance; IL-6, interleukin 6; LDL, low-density lipoprotein; VLDL, very-low-density lipoprotein. VOLUME 00 NUMBER 00 MONTH

8 NAFLD Resolution with Weight Loss Rachakonda et al. ballooning (14). articipants with as little as 5% weight loss had improvements in both insulin resistance and steatosis, while those who lost > 9% body weight experienced significant improvements in all components of the NAFLD activity score (12). One recent study demonstrated a dose-response relationship between weight loss and histologic improvement; subjects with as little as 5% weight loss had nonworsening of liver fibrosis, while weight loss > 10% was associated with high rates of resolution of steatohepatitis (16). On the other hand, a recent trial of bariatric surgery for the treatment of NAFLD demonstrated significant histologic improvements independent of weight loss (34), and multiple exercise-based interventions for NAFLD decreased intrahepatic triglyceride content in the absence of significant weight loss (17,35-37). These findings highlight the need for a better understanding of factors influencing hepatic responses to weight loss interventions. Figure 1 Changes in liver:spleen attenuation ratio (L/S ratio) in individuals with NAFLD persistence and resolution after weight loss intervention. While pre- and post-treatment 1-L/S were lower in subjects with NAFLD resolution ( <0.05 for both), rate of change of L/S did not differ between groups. the prevalence of NAFLD was higher in individuals with and without obesity when sarcopenia was present (9). Similarly, the L4 paraspinal muscle cross-sectional area was lower in a cohort of North American subjects with biopsy-proven NAFLD compared to non- NAFLD controls (11). However, in our study, multiple measures of skeletal muscle mass were similar in subjects with and without NAFLD. These discrepant results may be related, in part, to differences in assessment of steatosis, as clinical laboratory-based prediction models and liver biopsies were used to identify NAFLD in the aforementioned work, while CT-based definitions of steatosis were used in the present analysis. However, as our study included only North American individuals with severe obesity and without diabetes, our results suggest that ethnic, genetic, and environmental differences between populations may influence susceptibility to NAFLD and other metabolic disorders for a given degree of skeletal muscle loss. In our study, visceral but not subcutaneous adipose tissue mass was associated with NAFLD. This finding is consistent with previous reports of enhanced visceral adiposity in hepatic steatosis (4), inflammation with fibrosis (6,28), and transaminase elevation (29). While visceral adiposity may have only a minor influence on global insulin sensitivity (30,31), anatomic and metabolic features may explain its role in NAFLD development. Visceral fat has greater lipolytic activity than subcutaneous fat (32,33). Because subcutaneous adipose tissue depots were similar in both groups, VAT expansion may enhance hepatic FFA delivery in subjects with NAFLD. Subjects with NAFLD exhibited increased fasting glucose levels, which may be explained not only by hepatic resistance to insulinmediated suppression of gluconeogenesis, but also by increased availability of VAT-derived FFA as a substrate for gluconeogenesis. Although VAT elaborates several metabolically deleterious adipocytokines, circulating levels of these factors were similar between groups in our study. Weight loss is the most effective therapy for NAFLD. However, the relationship between quantity of weight loss and clinical benefit remains unclear. Weight loss greater than 7% was associated with histologic improvement in steatosis, lobular inflammation, and To this end, we assessed metabolic factors, adipocytokines, and body composition changes in patients with and without NAFLD after weight loss. Despite similar degrees of BMI change, subjects with NAFLD experienced greater reductions in VAT area. This was accompanied by significant improvements in AST, ALT, and hepatic insulin sensitivity as measured by fasting plasma glucose. All other metabolic parameters and body composition measures were similarly affected regardless of NAFLD status. Despite similar degrees of weight loss in patients with and without resolution of NAFLD, greater improvements in VAT area were observed in subjects with NAFLD resolution. Although responders and nonresponders alike had similar improvements in L/S ratio, responders had less steatosis at baseline. These findings indicate that VAT loss may be paramount to total weight loss for resolution of NAFLD and that there may be biologic determinants of VAT reduction for a given degree of weight loss. Of note, participants with NAFLD resolution had no disproportionate losses of lean mass or skeletal muscle, suggesting that weight loss interventions may be safe for muscle preservation. Our study has a few limitations. First, as this was a retrospective analysis of prospectively collected data, causal relationships could not be established. Second, the study population was limited to a femalepredominant cohort with World Health Organization Class II and III obesity. Third, NAFLD was defined using L/S ratios from CT imaging; while L/S ratio is useful for quantification of steatosis, it cannot identify individuals with steatohepatitis or liver fibrosis. Finally, no serologic evaluation was performed to exclude viral hepatitis, but individuals with liver enzyme levels more than 30% above the upper limit of normal were excluded from the original RENEW study. Conclusion Among a cohort of nondiabetic subjects with severe obesity, NAFLD was not associated with sarcopenia but rather with greater visceral adiposity and metabolic dysfunction. Individuals with NAFLD resolution had twofold higher VAT loss compared to those without NAFLD resolution, despite similar degree of total body weight loss. Our results suggest that there are individual variations in response to weight loss in patients with obesity and NAFLD, which may have implications in the clinical management of this population.o VC 2017 The Society 8 VOLUME 00 NUMBER 00 MONTH

9 OBESITY BIOLOGY AND INTEGRATED HYSIOLOGY References 1. Targher G, Day C, Bonora E. Risk of cardiovascular disease in patients with nonalcoholic fatty liver disease. N Engl J Med 2010;363: Eguchi Y, Eguchi T, Mizuta T, et al. Visceral fat accumulation and insulin resistance are important factors in nonalcoholic fatty liver disease. J Gastroenterol 2006;41: Holt HB, Wild SH, Wood J, et al. Non-esterified fatty acid concentrations are independently associated with hepatic steatosis in obese subjects. Diabetologia 2006;49: ark BJ, Kim YJ, Kim DH, et al. Visceral adipose tissue area is an independent risk factor for hepatic steatosis. J Gastroenterol Hepatol 2008;23: Thomas EL, Hamilton G, atel N, et al. Hepatic triglyceride content and its relation to body adiposity: a magnetic resonance imaging and proton magnetic resonance spectroscopy study. Gut 2005;54: van der oorten D, Milner KL, Hui J, et al. Visceral fat: a key mediator of steatohepatitis in metabolic liver disease. Hepatology 2008;48: Vongsuvanh R, George J, McLeod D, van der oorten D. Visceral adiposity index is not a predictor of liver histology in patients with non-alcoholic fatty liver disease. J Hepatol 2012;57: Hong HC, Hwang SY, Choi HY, et al. Relationship between sarcopenia and nonalcoholic fatty liver disease: the Korean Sarcopenic Study. Hepatology 2014;59: Lee YH, Jung KS, Kim SU, Yoon HJ, Yun YJ, Lee BW et al. Sarcopaenia is associated with NAFLD independently of obesity and insulin resistance: Nationwide surveys (KNHANES ). J Hepatol 2015;63: Srikanthan, Hevener AL, Karlamangla AS. Sarcopenia exacerbates obesityassociated insulin resistance and dysglycemia: findings from the National Health and Nutrition Examination Survey III. LoS One 2010;5:e doi: / journal.pone Dasarathy J, eriyalwar, Allampati S, et al. Hypovitaminosis D is associated with increased whole body fat mass and greater severity of non-alcoholic fatty liver disease. Liver Int 2014;34:e118-e Harrison SA, Fecht W, Brunt EM, Neuschwander-Tetri BA. Orlistat for overweight subjects with nonalcoholic steatohepatitis: A randomized, prospective trial. Hepatology 2009;49: etersen KF, Dufour S, Befroy D, Lehrke M, Hendler RE, Shulman GI. Reversal of nonalcoholic hepatic steatosis, hepatic insulin resistance, and hyperglycemia by moderate weight reduction in patients with type 2 diabetes. Diabetes 2005;54: romrat K, Kleiner DE, Niemeier HM, et al. Randomized controlled trial testing the effects of weight loss on nonalcoholic steatohepatitis. Hepatology 2010;51: Vilar Gomez E, Rodriguez De Miranda A, Gra Oramas B, et al. Clinical trial: a nutritional supplement Viusid, in combination with diet and exercise, in patients with nonalcoholic fatty liver disease. Aliment harmacol Ther 2009;30: Vilar-Gomez E, Martinez-erez Y, Calzadilla-Bertot L, et al. Weight loss through lifestyle modification significantly reduces features of nonalcoholic steatohepatitis. Gastroenterology 2015;149: e5; quiz e Hallsworth K, Fattakhova G, Hollingsworth KG, et al. Resistance exercise reduces liver fat and its mediators in non-alcoholic fatty liver disease independent of weight loss. Gut 2011;60: Keating SE, Hackett DA, George J, Johnson NA. Exercise and non-alcoholic fatty liver disease: a systematic review and meta-analysis. J Hepatol 2012;57: Goodpaster BH, Delany J, Otto AD, et al. Effects of diet and physical activity interventions on weight loss and cardiometabolic risk factors in severely obese adults: a randomized trial. JAMA 2010;304: Goodpaster BH, Kelley DE, Wing RR, Meier A, Thaete FL. Effects of weight loss on regional fat distribution and insulin sensitivity in obesity. Diabetes 1999;48: Dunn MA, Behari J, Rogal SS, et al. Hepatic steatosis in diabetic patients does not predict adverse liver-related or cardiovascular outcomes. Liver Int 2013;33: Iwasaki M, Takada Y, Hayashi M, et al. Noninvasive evaluation of graft steatosis in living donor liver transplantation. Transplantation 2004;78: ark SH, Kim N, Kim KW, et al. Macrovesicular hepatic steatosis in living liver donors: use of CT for quantitative and qualitative assessment. Radiology 2006;239: Shores NJ, Link K, Fernandez A, et al. Non-contrasted computed tomography for the accurate measurement of liver steatosis in obese patients. Dig Dis Sci 2011;56: Rachakonda V, Reeves VL, Aljammal J, et al. Serum autotaxin is independently associated with hepatic steatosis in women with severe obesity. (Silver Spring) 2015;23: DeLany J, Kelley DE, Hames KC, Jakicic JM, Goodpaster BH. Effect of physical activity on weight loss, energy expenditure, and energy intake during diet induced weight loss. (Silver Spring) 2014;22: Linkov F, Maxwell GL, Felix AS, et al. Longitudinal evaluation of cancerassociated biomarkers before and after weight loss in RENEW study participants: implications for cancer risk reduction. Gynecol Oncol 2012;125: Marchesini G, Bugianesi E, Forlani G, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003;37: Verrijken A, Francque S, Mertens I, Talloen M, eiffer F, Van Gaal L. Visceral adipose tissue and inflammation correlate with elevated liver tests in a cohort of overweight and obese patients. Int J Obes (Lond) 2010;34: Fabbrini E, Magkos F, Mohammed BS, et al. Intrahepatic fat, not visceral fat, is linked with metabolic complications of obesity. roc Natl Acad Sci U S A 2009; 106: Kantartzis K, Machann J, Schick F, Fritsche A, H aring HU, Stefan N. The impact of liver fat vs visceral fat in determining categories of prediabetes. Diabetologia 2010;53: Martin ML, Jensen MD. Effects of body fat distribution on regional lipolysis in obesity. J Clin Invest 1991;88: Montague CT, rins JB, Sanders L, et al. Depot-related gene expression in human subcutaneous and omental adipocytes. Diabetes 1998;47: Lassailly G, Caiazzo R, Buob D, et al. Bariatric surgery reduces features of nonalcoholic steatohepatitis in morbidly obese patients. Gastroenterology 2015;149: ; quiz e Bacchi E, Negri C, Targher G, et al. Both resistance training and aerobic training reduce hepatic fat content in type 2 diabetic subjects with nonalcoholic fatty liver disease (the RAED2 Randomized Trial). Hepatology 2013;58: Johnson NA, Sachinwalla T, Walton DW, et al. Aerobic exercise training reduces hepatic and visceral lipids in obese individuals without weight loss. Hepatology 2009;50: Keating SE, Hackett DA, arker HM, et al. Effect of aerobic exercise training dose on liver fat and visceral adiposity. J Hepatol 2015;63: VOLUME 00 NUMBER 00 MONTH