Background: Animal and human cross-sectional data suggest that bone marrow adipose tissue (MAT) may respond to mechanical loads and exercise. We conducted the first randomised controlled trial of exercise on MAT modulations in humans.
Methods: Forty patients with chronic non-specific low back pain (NSCLBP) were enrolled in a six-month single-blinded randomised controlled trial (ACTRN12615001270505). Twenty patients loaded their spines via progressive upright aerobic and resistance exercises targeting major muscle groups (Exercise). Twenty patients performed non-weightbearing motor control training and manual therapy (Control). Testing occurred at baseline, 3-months (3mo) and 6-months (6mo). Lumbar vertebral fat fraction (VFF) was measured using magnetic resonance imaging axial mDixon sequences.
Results: When compared to baseline (percent change), lumbar vertebral fat fraction (VFF; measured using magnetic resonance imaging axial mDixon sequences) was lower in Exercise at 3mo at L2 (−3.7[6.8]%, p = 0.033) and L4 (−2.6[4.1]%, p = 0.015), but not in Control. There were no between-group effects. The effects of Exercise on VFF were sex-specific, with VFF lower in men at L2, L3, L4 at 3mo and at L1, L2, L3 and L4 at 6mo (p all ≤ 0.05), but not in women. Leg and trunk lean mass were increased at 3mo in Exercise. Changes in VFF correlated significantly with changes in total fat (ρ = 0.40) and lean (ρ = −0.41) masses, but not with lumbar BMD (ρ = −0.10) or visceral adipose tissue volume (ρ = 0.23).
Conclusions: This trial provided first prospective evidence in humans that a moderate exercise intervention may modulate lumbar VFF as a surrogate measure of MAT at 3mo, yet not 6mo. The effect of exercise on MAT may be more prominent in males than females.