A critical review of low-carbohydrate diets in people with Type 2 diabetes

• Published in: Diabetic medicine 2016-02, Vol.33 (2), p.148-157
• Main Authors: van Wyk, H. J., Davis, R. E., Davies, J. S.
• Format: Article
• Language: English
• Subjects: Biomarkers - blood; Carbohydrates; Diabetes; Diabetes Mellitus, Type 2 - blood; Diabetes Mellitus, Type 2 - complications; Diabetes Mellitus, Type 2 - diet therapy; Diet, Carbohydrate-Restricted - adverse effects; Diet, Diabetic - adverse effects; Diet, Reducing - adverse effects; Energy Intake; Humans; Hyperglycemia - prevention & control; Meta-analysis; Meta-Analysis as Topic; Metabolism; Middle Aged; Obesity - complications; Obesity - diet therapy; Overweight - complications; Overweight - diet therapy; Patient Compliance; Quality; Reproducibility of Results; Studies; Waist Circumference; Weight Loss
• ISSN: 0742-3071; 1464-5491
• DOI: 10.1111/dme.12964

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. Conclusion Total energy intake remains the dietary predictor of body weight. A LCD appears no different from a high‐carbohydrate 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. What's new? This review provides new insights into the effectiveness of low‐carbohydrate diets by demonstrating that limiting inclusion in meta‐analyses to higher quality studies results in more coherent and consistent outcomes. We provide possible explanations for differences in outcomes found when comparing low‐ and high‐carbohydrate diets. This is the first study to assess the actual carbohydrate intake in grams per day (g/day) at the final follow‐up from the available studies included in the meta‐analyses and proposes that the method provides a more useful comparator of actual carbohydrate intake in low‐carbohydrate diets. This study provides recommendations on how to determine a person's recommended carbohydrate intake.