Weight loss reduces intermuscular adipose tissue and liver fat compared to exercise: implications for muscle quality and metabolic health

CONTEXT: Weight loss improves insulin sensitivity and liver fat but reduces lean mass. Whether reductions in intermuscular adipose tissue (IMAT) exceed muscle loss and how these effects compare with endurance exercise training without weight loss remains unclear. OBJECTIVE: Test whether weight loss leads to greater reductions in IMAT and hepatic fat than exercise training alone, and whether these changes are reflected in plasma lipidomic and metabolomic profiles in adults with obesity. METHODS: In a 12-week randomized intervention, 46 individuals with obesity were assigned to weight loss (WL-age 39.7 ± 1.3; BMI 35.2 ± 0.9; W/M 8/9), endurance exercise training without weight loss (EX-age 39.7 ± 1.6; BMI 36.0 ± 1.1; W/M 8/8), or delayed-control intervention groups (age 39.8 ± 1.4; BMI 36.8 ± 1.3; W/M 6/7). Changes in IMAT volume, skeletal muscle mass, and liver fat measured by MRI; insulin sensitivity assessed by hyperinsulinemic-euglycemic clamp; and plasma lipidomics and metabolomics by LC-MS based methods. RESULTS: WL reduced body weight (-10.5%; P < .001), liver steatosis (-33.1%; P < .0001), and IMAT volume (-12.7%; P < .0001), while improving insulin sensitivity (42%; P = .004). IMAT decreased significantly more than skeletal muscle mass (-4.2%, P = .17), indicating improved muscle quality. EX increased insulin sensitivity (23%; P = .04) and VO₂peak (7.6%; P < .001) but did not significantly change IMAT or liver fat. DXA overestimated muscle loss compared with MRI. WL decreased plasma sphingolipids and diacylglycerols, whereas EX reduced acylcarnitines. Plasma triacylglycerols and branched-chain amino acids were strongly correlated with liver fat, and triacylglycerols showed the strongest association with IMAT volume. CONCLUSION: Weight loss is more effective than endurance exercise training without weight loss for reducing IMAT and liver steatosis, with IMAT loss exceeding muscle loss. Plasma lipids and metabolites signatures are associated with liver fat and IMAT, supporting their potential utility as noninvasive biomarkers.