Abstract
Aim Very low carbohydrate/ketogenic diets (VLC/KDs) are popular but their role in managing pre-diabetes and type 2 diabetes (T2D) is uncertain. This study uses a systematic review and meta-analysis of randomized controlled trials to estimate the effect of these diets in this population. Materials and Methods A systematic review identified randomized controlled trials of at least 6 months duration comparing efficacy and safety of VLC/KDs (<= 50 g carbohydrate or <= 10% total energy from carbohydrate per day) with a control diet (carbohydrate above the VLC/KD threshold) in adults with pre-diabetes or T2D. The primary outcome variable was glycated haemoglobin (HbA1c) after 12 months. The meta-analysis method was inverse variance weighting of mean values for continuous variables. Results Key word searches identified 2290 studies; 2221 were not in scope. A full text review of 69 studies identified eight meeting inclusion criteria; in total, it involved 606 participants. Six studies reported HbA1c (%) at 12 months; four as change from baseline with a fixed effects estimate (95% confidence interval): VLC/KD minus control of 0.01% (-0.22 to 0.25), p = .91; and two as change from baseline: -0.65% (-0.99; -0.31) [-7.1 mmol/mol (-10.8; -3.4)], p < .001. Serum triglycerides were lower with VLC/KD versus control: -0.28 mmol/L (-0.44 to -0.11), p < .001. High-density lipoprotein was higher with an estimate of 0.04 mmol/L (0.01 to 0.08), p = .03, in the five studies reporting 12-month summary data. Conclusions A VLC/KD may cause reductions in HbA1c and triglycerides in those with pre-diabetes or T2D but evidence of an advantage over other strategies is limited. More well-designed studies are required to provide certain evidence.