SAT-110 Long-Term Impact of Gastric Bypass versus Sleeve Gastrectomy on Nonalcoholic Fatty Liver Disease: Retrospective Observations from Michigan Bariatric Surgery Cohort

Abstract

Non-alcoholic fatty liver disease (NAFLD) is highly prevalent in obese patients. Bariatric surgery is highly effective for treatment of NAFLD. We have previously reported 4 and 5-year weight loss outcomes in patients who underwent bariatric surgery at Michigan Medicine from 2008 to 2013. The aim of this study was to assess the impact of weight loss surgery on persistence of NAFLD up to 5 years post-surgery. Retrospective data on weight, blood pressure, lipid panel, A1c, liver enzymes, NAFLD Fibrosis scores (NFS) and imaging were analyzed with respect to the type of surgery [sleeve gastrectomy (SG) or Roux-en Y gastric bypass (GB)] at baseline and yearly for 5 years. Patients were assessed for presence of underlying NAFLD based on liver biopsy, imaging, or clinical diagnosis (ICD-9 or 10 codes) prior to surgery. Persistent NAFLD was assessed based on presence of hepatic steatosis on abdominal imaging (ultrasound, CT or MRI) over follow-up. Follow up data were available on 221 patients with NAFLD at baseline (GB =128, SG=93); median age 47 years (IQR 40-54) with 73% females and a median BMI of 46.2 kg/m2 (IQR 41.2-52.4). Median BMI was higher in the SG group versus GB owing to the criteria for approval in 2008-2012 (48.1 kg/m2 vs. 45.3 kg/m2, p=0.01). Median NFS was higher in SG vs GB (0.04 vs -0.50, p =0.01). Only 38 patients (17.2%) had a liver biopsy, among which 25 (65.8%) showed non-alcoholic steatohepaitis (NASH) and 3 (7.9%) demonstrated cirrhosis. Both groups had a reduction in NFS over follow up with median NFS significantly lower at all-time points in the GB group, p<0.01); improvement in NFS was more pronounced in the GB group at year 1, p=0.03. Overall, 83 patients (37.5%) had follow-up abdominal imaging with 28 (33.7%) having evidence of persistent NAFLD; 28.7% at 1-2 years, 16.3% at 2-4 years, and 55% at >4 years post-surgery. The nadir of BMI occurred at 1 year (median 31.3 and 37.2, in BG versus SG, respectively), with a peak at year 4 for GB (median 35.6) and year 5 for SG (median 40.4). At all-time points, the median weight loss was highest in the GB group. All metabolic parameters remained improved compared to baseline. The following differences were noted in the GB group versus SG group: LDL was significantly lower at year 1, 2 and 5; cholesterol significantly lower at year 1 and 2; significantly larger reduction in hemoglobin A1c at year 2, 4 and 5; and reduction in TG significantly higher at year 1. In a multivariable model integrating type of surgery, baseline diabetes and baseline hyperlipidemia, baseline hyperlipidemia was independently associated with persistent NAFLD (OR: 3.2, 95% CI 1.04-10.10, p=0.04). Change in weight was not predictive of persistent NAFLD. In conclusion, these observations suggest that bariatric surgery may have a weight-independent effect on NAFLD. Further studies are needed to examine the effect of disease severity and co-morbidities on NAFLD resolution after bariatric surgery.