|| Checking for direct PDF access through Ovid
There have been many reviews on this question with generally favourable outcomes in terms of all-day glucose levels and glycosylated hemoglobin A1 concentration (HbA1c) but the most recent meta-analysis  was negative although only four of a possible 12 studies in type 2 diabetes longer in duration than 12 weeks were included.The 6-month data from a 2-year low-carbohydrate weight-loss study in type 2 diabetes has shown favourable effects on HbA1c and medication use . In this study, the low carbohydrate diet was restricted to 14% of energy as carbohydrate (a target of <50 g/day was set) and compared with a high carbohydrate diet of 53% carbohydrate with an emphasis on low glycemic index foods. Importantly, saturated fat was restricted to less than 10% in both diets so the difference in calories was made up with both protein and unsaturated fat. The diets were isocaloric; 115 participants commenced the study and 93 completed 6 months. The weight loss was excellent at 11.5–12 kg. Although overall HbA1c changes were not different, medication reductions were twice as great in the low carbohydrate group and reductions in glycemic variability were also twice as great. On post-hoc analysis, those individuals with a baseline HbA1c of more than 7.8% (nine low carbohydrate and 14 high carbohydrate individuals) had a lower HbA1c of about 0.6% on the low carbohydrate diet (P < 0.05). A further report from the same study at 12 months showed a fall in HbA1c of 1% in both the groups with no differences between the two diets although medication reductions occurred more frequently in the low carbohydrate group [3▪▪].Similar benefits of a low-carbohydrate weight-loss diet in type 2 diabetes were seen in a smaller cohort in which weight loss was the same in both the low carbohydrate and high carbohydrate group but HbA1c was lower by 0.8% in the low carbohydrate diet. A more than 50% reduction in medication occurred in 70% of the low carbohydrate group versus 30% of the high carbohydrate group . Gower and Goss  found that in 30 women with polycystic ovary syndrome (a risk factor for type 2 diabetes), the low carbohydrate diet lowered fasting insulin and glucose and improved insulin sensitivity and the β-cell response to a test meal. In the low carbohydrate diet, women lost intra-abdominal fat whereas on the high carbohydrate diet lean mass was lost but the two groups were not contrasted statistically. A carbohydrate-free diet in a noncalorie-restricted diet lowers 24-h glucose and insulin levels in people with type 2 diabetes by 39 and 48%, respectively, whereas complete fasting lowers them by 49 and 69%, respectively. Thus, the carbohydrate restriction accounts for 71% of the effect of fasting in type 2 diabetes [6▪]. A position statement on the benefits of low carbohydrate diets for type 2 diabetes has been recently published [7▪] while the mechanisms of the effect of a ketogenic diet on appetite and food intake have been reviewed by Paoli et al.[8▪]. A diet with less than 100 g of carbohydrate per day induces ketones which are acetyl-CoA derivatives formed when there is insufficient oxaloacetate available in the TCA cycle for acetyl-COA to enter. Ketones can be used for energy by muscle and after a delay the brain can use them as well.Fibroblast growth factor 21 (FGF21) is a novel liver protein that increases in obesity but ameliorates obesity-associated hyperglycaemia and hyperlipidaemia primarily via effects on adipose tissue and the pancreas. It also increases in response to dietary restriction.