Bariatric surgery consistently induces remission of type 2 diabetes. We tested whether there are diabetes-specific mechanisms in addition to weight loss.

We studied 25 morbidly obese patients (BMI 51.7 ± 1.5 kg/m² [mean ± SEM]), 13 with non-insulin-treated type 2 diabetes (HbA_1c 7.1 ± 0.5% [54 ± 5 mmol/mol]), before and at 2 weeks and 1 year after Roux-en-Y gastric bypass (RYGB). Lean (n = 8, BMI 23.0 ± 0.5 kg/m²) and obese (n = 14) volunteers who were BMI-matched (36.0 ± 1.2) to RYGB patients at 1 year after surgery served as controls. We measured insulin-stimulated glucose disposal (M) and substrate utilisation (euglycaemic clamp/indirect calorimetry), endogenous glucose production (EGP) by 6,6-[²H₂]glucose, lipolysis (rate of appearance of [²H₅]glycerol) and beta cell function (acute insulin response to i.v. glucose [AIR] as determined by C-peptide deconvolution).

At baseline, all obese groups showed typical metabolic abnormalities, with M, glucose oxidation and non-oxidative disposal impaired, and EGP, lipolysis, lipid oxidation and energy expenditure increased. Early after RYGB plasma glucose and insulin levels, and energy expenditure had decreased, while lipid oxidation increased, with M, EGP and AIR unchanged. At 1 year post-RYGB (BMI 34.4 ± 1.1 kg/m²), all diabetic patients were off glucose-lowering treatment and mean HbA_1c was 5.4 ± 0.14% (36 ± 2 mmol/mol) (p = 0.03 vs baseline); AIR also improved significantly. In all RYGB patients, M, substrate oxidation, EGP, energy expenditure and lipolysis improved in proportion to weight loss, and were therefore similar to values in obese controls, but still different from those in lean controls.

In morbidly obese patients, RYGB has metabolic effects on liver, adipose tissue, muscle insulin sensitivity and pattern of substrate utilisation; these effects can be explained by energy intake restriction and weight loss, the former prevailing early after surgery, the latter being dominant in the longer term.