Restricting carbohydrates and calories in the treatment of type 2 diabetes: a systematic review of the effectiveness of 'low-carbohydrate' interventions with differing energy levels

Abstract

Keywords: Diabetes; HbA1c, glycated haemoglobin; LCD, low-carbohydrate diets; LED, low-energy diets; Low-carbohydrate diet; Low-energy diet; Obesity; T2D, type 2 diabetes; TDR, total diet replacement.

Graphical abstract

Fig. 1.

Interrelationship between energy restriction, weight loss and carbohydrate restriction in improved glycaemic control: carbohydrate and energy restriction are interrelated. (A) In obese individuals with T2D, weight loss is associated with improved glycaemic control⁽¹⁰⁾. This is in accordance with the twin cycle hypothesis, whose central tenet is that excess lipids within the liver and the pancreas drive T2D pathogenesis⁽¹¹⁾. (B) In studies of low-energy feeding, glycaemia improves within days of energy restriction, before significant weight loss has occurred⁽¹²⁾. (C) Carbohydrate restriction improves glycaemia by reducing postprandial glucose rises. While failed repression of gluconeogenesis and glycogenolysis are major causes of hyperglyacemia⁽¹³⁾, dietary carbohydrate intake is the largest driver of postprandial glucose rises. (D) Carbohydrate restriction is also associated with weight loss. This may occur as a function of spontaneous energy restriction or there may be independent effects arising from reduced insulin secretion. Whether or not carbohydrate restriction has independent effects on body weight remains a matter of contentious debate (hence depicted as dashed line)^(14,15). T2D, type 2 diabetes.

Fig. 2.

Study screening and selection.

Fig. 3.

Prescribed daily carbohydrate and energy intakes in included studies. Where a maximum allowance of carbohydrate or energy was prescribed, this value was used; where a range of carbohydrate or energy intakes was prescribed, the mid-point value was taken; where energy intake was unrestricted, a value of 2000 kcal/d was assigned. Squares, no energy restriction (ad libitum feeding); circles, moderate energy restriction (1200–2000 kcal/d); triangles, severe energy restriction (<1200 kcal/d).

Fig. 4.

Risk of bias assessment results: +, low risk of bias; ?, unclear risk of bias; –, high risk of bias.

Fig. 5.

Average improvement in HbA1c and average percentage weight loss at study endpoints. Each point represents the mean value for a single study with the exception of Sato et al.⁽⁴¹⁾ which represents median values. Study endpoints range from 3 to 24 months. Squares, no energy restriction (ad libitum feeding); circles, moderate energy restriction (1200–2000 kcal/d); triangles, severe energy restriction (<1200 kcal/d).

Fig. 6.

Average improvement in HbA1c and average percentage weight loss at 12 months each point represents the mean changes from baseline in HbA1c and weight for a single study, with the exception of Sato et al.⁽⁴¹⁾ which represents median values. Studies were only included if they reported data at 12 months. Squares, no energy restriction (ad libitum feeding); circles, moderate energy restriction (1200–2000 kcal/d); triangles, severe energy restriction (<1200 kcal/d). HbA1c, haemoglobin A1C.