Aims The efficacy of low-carbohydrate diets (LCD) in people with Type 2 diabetes has divided the nutrition community. This review seeks to re-examine the available data to clarify understanding.Methods A comprehensive search of databases was used to identify meta-analyses of LCD in Type 2 diabetes. To improve the quality of the studies analysed, the following inclusion criteria were applied: randomized control trials ≥ 4 weeks in people aged > 18 years with Type 2 diabetes; a carbohydrate intake ≤ 45% of total energy intake per day; and a dietary intake assessment at the end of the study. The resulting studies were subjected to a thematic analysis.Results Nine meta-analyses were identified containing 153 studies. Twelve studies met our amended inclusion criteria. There were no significant differences in metabolic markers, including glycaemic control, between the two diets, although weight loss with a LCD was greater in one study. Carbohydrate intake at 1 year in very LCD (< 50 g of carbohydrates) ranged from 132 to 162 g. In some studies, the difference between diets was as little as 8 g/day of carbohydrates. ConclusionTotal energy intake remains the dietary predictor of body weight. A LCD appears no different from a highcarbohydrate diet in terms of metabolic markers and glycaemic control. Very LCDs may not be sustainable over a medium to longer term as carbohydrate intake in diets within studies often converged toward a more moderate level. The variable quality of studies included in earlier meta-analyses likely explains the previous inconsistent findings between meta-analyses.Diabet. Med. 33: 148-157 (2016)

No abstract