Short-term dietary reduction of branched-chain amino acids reduces meal-induced insulin secretion and modifies microbiome composition in type 2 diabetes: a randomized controlled crossover trial

Adults with type 2 diabetes in good metabolic control were enrolled in a randomized, double-blind, crossover study in Düsseldorf, Germany. Participants were 40-60 years old, had BMI 28-35 kg/m2, and were treated with lifestyle modification, metformin, or other oral glucose-lowering medication; total analyzed sample size was 12 participants, 8 men and 4 women.

This randomized, double-blind crossover trial tested a 4-week isocaloric dietary reduction in branched-chain amino acids while keeping total protein intake constant at 1 g/kg body weight per day. During the BCAA-reduced periods, about 60% of protein intake was replaced with an amino acid powder lacking BCAAs; the comparison period used a matched amino acid powder containing all amino acids. Both regimens were delivered alongside regular foods.

Short-term BCAA reduction produced favorable postprandial metabolic and microbiome changes, but it did not improve whole-body insulin sensitivity during clamp testing. After the BCAA-reduced diet, meal-induced insulin and C-peptide responses were lower or improved versus the BCAA-containing diet, with incremental insulin release iAUC 29 ± 19 vs 21 ± 11 and incremental C-peptide release 2.8 ± 0.9 vs 2.5 ± 0.8 (both P < 0.05), while postprandial insulin sensitivity indices improved (OGIS 279 ± 94 vs 346 ± 91 mL·min-1·m-2, P < 0.01; PREDIM 2.6 ± 0.9 vs 3.3 ± 1.3 mg·kg-1·min-1, P < 0.01). Fasting BCAAs fell from 507 ± 90 to 422 ± 56 μmol/L with the BCAA-reduced diet (P < 0.001), and fecal microbiota shifted toward less Firmicutes and more Bacteroidetes. The authors concluded that short-term BCAA reduction may improve postprandial insulin handling and adipose-tissue mitochondrial efficiency in type 2 diabetes, but longer-term studies are needed to assess safety and metabolic benefit.

The study was very small (n = 12) and short in duration (4 weeks), which limits precision and long-term inference. Participants were a selected group with relatively well-controlled type 2 diabetes, so generalizability is limited. Findings were largely mechanistic and postprandial; whole-body insulin sensitivity by clamp did not change, and clinical outcomes were not assessed.