RESUMO
BACKGROUND AND AIMS: Diabetes mellitus (DM) is one of the most prevalent chronic noncommunicable diseases globally, and the only way to reduce its complications is good glycaemic control. Insulin remains the only approved treatment for type 1 DM (T1DM) and is used by many with type 2 DM (T2DM). Carbohydrate counting is considered the ideal way to calculate meal-related insulin doses as it allows greater flexibility in diet and could, in some people, reduce the burden of the disease. The primary objective of this systematic review was to assess carbohydrate counting efficacy in reducing glycated haemoglobin (HbA1c ) levels and safety by not increasing hypoglycaemia risk, inducing an increase in body weight or blood lipids, or reducing the quality of life of people with T1DM. METHODS: We included randomised controlled clinical trials with a parallel-group design comparing any carbohydrate counting forms with standard care or other forms of dietary advice or insulin dose calculation in people with T1DM with a follow up period of at least 3 months and with no restrictions in language, age or settings. As a primary outcome, we consider the change of HbA1c levels within at least 3 months. Secondary outcomes were hypoglycaemia events, body weight changes, blood lipids levels, and the total daily insulin dose. We also evaluated health-related quality-of-life changes and questionnaires on satisfaction with treatment of diabetes. RESULTS: Data from 11 studies with 899 patients were retrieved with a mean follow-up of 52 ± 35.5 weeks. Carbohydrate counting is not better in reducing HbA1c levels (SMD-0.24%, 95% CI -0.68 to 0.21) than all dietary advice forms. However, this finding was highly heterogeneous. We identified three studies that account for most of the heterogeneity using clustering algorithms. A second analysis excluding these studies shows a meaningful reduction in HbA1c levels (SMD-0.52%, 95% CI -0.82 to -0.23) with low heterogeneity. In the subgroup analysis, carbohydrate counting significantly reduces HbA1c levels compared with usual diabetes education. Carbohydrate counting does not relate to any substantial change in blood lipids, body weight, hypoglycaemia risk or daily insulin dose. Finally, we analysed the effect of trial duration on reduction in HbA1c levels and found no significant change related to time. CONCLUSIONS: Carbohydrate counting is an efficacious technique to safely reduce the levels of HbA1c in adults and children compared with standard diabetes education, and its effect does not appear to change with prolonged time. Standardisation in reporting important outcomes such as hypoglycaemia and quality of life are vital to produce comparable evidence in carbohydrate counting clinical trials. This systematic review was registered in PROSPERO under code: CRD42020218499.