The effect of long-term weight-loss intervention strategies on the dynamics of pancreatic-fat and morphology: An MRI RCT study

Background & aims: The ability to mobilize pancreatic-fat and the meaning of decreased fat in the pancreas remain controversial. We followed the dynamics of pancreatic-fat and its morphology during various long weight-loss induced lifestyle-interventions. Methods: In isolated workplace with monitored/provided lunch, we randomly assigned healthy persons with abdominal obesity or dyslipidemia for one of two 18-month equal-caloric diets: low-fat (LF) or Mediterranean/low-carbohydrate (Med/LC, with provided 1oz walnuts/day), with or without added moderate exercise (supervised gymmembership). We used magnetic-resonance-imaging to quantify pancreatic-fat and morphology. Results: At baseline, 277 eligible participants (mean age = 48 years; 88% men; pancreatic-fat = 17.4 +/- 5.1%) had higher pancreatic-fat in men (17.7 +/- 4.9% vs 14.9 +/- 5.5% in women; p = 0.004). Following 18-month intervention (adherence = 86.3%) and moderate weight-loss (mean = -3.0 +/- 5.5 kg), pancreatic-fat decreased moderately but significantly (-0.26 +/- 2.18% units; p = 0.049). Med/LC diet induced a greater decrease in pancreatic-fat compared to LF (p = 0.043), and the combination of Med/LC diet + exercise exhibited the highest reduction (-0.69% units) as compared to LF diet without exercise (+0.12% units; p = 0.027 between groups). In multivariate regression models, after further adjusted for visceral adiposetissue (DVAT), pancreatic-fat loss associated with both decreases in pancreatic-morphology ratio (perimeter divided by area; beta = 0.361; p < 0.001) and superficial-subcutaneous adipose-tissue loss (beta = 0.242; p = 0.001), but not with changes in intrahepatic-fat (beta = -0.034; p = 0.638). Pancreatic-fat loss associated with increased intake of polyunsaturated-fat (beta = -0.137; p = 0.032), as with improved high-density lipoprotein-cholesterol (HDL; beta = -0.156; p = 0.023) and triglycerides/HDL ratio (beta = 0.162; p = 0.015), independently of DVAT, but not with glycemicecontrol parameters (e.g. HbA1c, HOMA-IR and HOMA-beta; p > 0.2 for all). Conclusions: Pancreatic-fat loss is mainly associated with improved lipid, rather than glycemic profiles. Med/LC diet, mostly with exercise, may benefit pancreatic-fat loss. Pancreatic-morphology could serve as a biomarker of pancreatic-fat state. (c) 2018 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.