BACKGROUND: Muscle atrophy and strength loss are common adverse outcomes following bariatric surgery. This randomized, controlled trial investigated the effects of exercise training on bariatric surgery-induced loss of muscle mass and function. Additionally, we investigated the effects of the intervention on molecular and histological mediators of muscle remodelling. METHODS: Eighty women with obesity were randomly assigned to a Roux-en-Y gastric bypass (RYGB: n = 40, age = 42 ± 8 years) or RYGB plus exercise training group (RYGB + ET: n = 40, age = 38 ± 7 years). Clinical and laboratory parameters were assessed at baseline, and 3 (POST3) and 9 months (POST9) after surgery. The 6 month, three-times-a-week, exercise intervention (resistance plus aerobic exercise) was initiated 3 months post-surgery (for RYGB + ET). A healthy, lean, age-matched control group was recruited to provide reference values for selected variables. RESULTS: Surgery resulted in a similar (P = 0.66) reduction in lower-limb muscle strength in RYGB and RYGB+ET (-26% vs. -31%), which was rescued to baseline values in RYGB + ET (P = 0.21 vs. baseline) but not in RYGB (P < 0.01 vs. baseline). Patients in RYGB+ET had greater absolute (214 vs. 120 kg, P < 0.01) and relative (2.4 vs. 1.4 kg/body mass, P < 0.01) muscle strength compared with RYGB alone at POST9. Exercise resulted in better performance in timed-up-and-go (6.3 vs. 7.1 s, P = 0.05) and timed-stand-test (18 vs. 14 repetitions, P < 0.01) compared with RYGB. Fat-free mass was lower (POST9-PRE) after RYBG than RYGB + ET (total: -7.9 vs. -4.9 kg, P < 0.01; lower-limb: -3.8 vs. -2.7 kg, P = 0.02). Surgery reduced Types I (~ - 21%; P = 0.99 between-group comparison) and II fibre cross-sectional areas (~ - 27%; P = 0.88 between-group comparison), which were rescued to baseline values in RYGB+ET (P > 0.05 vs. baseline) but not RYGB (P > 0.01 vs. baseline). RYGB + ET showed greater Type I (5187 vs. 3898 µm2 , P < 0.01) and Type II (5165 vs. 3565 µm2 , P < 0.01) fCSA than RYGB at POST9. RYGB + ET also resulted in increased capillarization (P < 0.01) and satellite cell content (P < 0.01) than RYGB at POST9. Gene-set normalized enrichment scores for the muscle transcriptome revealed that the ubiquitin-mediated proteolysis pathway was suppressed in RYGB + ET at POST9 vs. PRE (NES: -1.7; P < 0.01), but not in RYGB. Atrogin-1 gene expression was lower in RYGB + ET vs. RYGB at POST9 (0.18 vs. 0.71-fold change, P < 0.01). From both genotypic and phenotypic perspectives, the muscle of exercised patients resembled that of healthy lean individuals. CONCLUSIONS: This study provides compelling evidence-from gene to function-that strongly supports the incorporation of exercise into the recovery algorithm for bariatric patients so as to counteract the post-surgical loss of muscle mass and function.