This systematic review and meta-analysis determined the impact of structured exercise training, and the influence of associated weight loss, on intrahepatic triglyceride (IHTG) in individuals with non-alcoholic fatty liver disease (NAFLD). It also examined its effect on hepatic insulin sensitivity in individuals with or at increased risk of NAFLD. Analyses were restricted to studies using magnetic resonance spectroscopy or liver biopsy for the measurement of IHTG and isotope-labelled glucose tracer for assessment of hepatic insulin sensitivity. Pooling data from 17 studies (373 exercising participants), exercise training for one to 24 weeks (mode: 12 weeks) elicits an absolute reduction in IHTG of 3.31% (95% CI: -4.41 to -2.22%). Exercise reduces IHTG independent of significant weight change (-2.16 [-2.87 to -1.44]%), but benefits are substantially greater when weight loss occurs (-4.87 [-6.64 to -3.11]%). Furthermore, meta-regression identified a positive association between percentage weight loss and absolute reduction in IHTG (β = 0.99 [0.62 to 1.36], P < 0.001). Pooling of six studies (94 participants) suggests that exercise training also improves basal hepatic insulin sensitivity (mean change in hepatic insulin sensitivity index: 0.13 [0.05 to 0.21] mg m min per μU mL ), but available evidence is limited, and the impact of exercise on insulin-stimulated hepatic insulin sensitivity remains unclear.

Fibroblast growth factor 21 (FGF21), follistatin and leukocyte cell-derived chemotaxin 2 (LECT2) are novel hepatokines that are modulated by metabolic stresses. This study investigated whether exercise intensity modulates the hepatokine response to acute exercise. Ten young, healthy men undertook three 8-h experimental trials: moderate-intensity exercise (MOD; 55% peak oxygen uptake), high-intensity exercise (HIGH; 75% peak oxygen uptake), and control (CON; rest), in a randomised, counterbalanced order. Exercise trials commenced with a treadmill run of varied duration to match gross exercise energy expenditure between trials (MOD vs HIGH; 2475 ± 70 vs 2488 ± 58 kJ). Circulating FGF21, follistatin, LECT2, glucagon, insulin, glucose and nonesterified fatty acids (NEFA) were measured before exercise and at 0, 1, 2, 4, and 7 h postexercise. Plasma FGF21 concentrations were increased up to 4 h postexercise compared with CON (P ≤ 0.022) with greater increases observed at 1, 2, and 4 h postexercise during HIGH versus MOD (P ≤ 0.025). Irrespective of intensity (P ≥ 0.606), plasma follistatin concentrations were elevated at 4 and 7 h postexercise (P ≤ 0.053). Plasma LECT2 concentrations were increased immediately postexercise (P ≤ 0.046) but were not significant after correcting for plasma volume shifts. Plasma glucagon (1 h; P = 0.032) and NEFA (4 and 7 h; P ≤ 0.029) responses to exercise were accentuated in HIGH versus MOD. These findings demonstrate that acute exercise augments circulating FGF21 and follistatin. Exercise-induced changes in FGF21 are intensity-dependent and may support the greater metabolic benefit of high-intensity exercise.