Type 2 diabetes (T2D) necessitates continuous self-management, emphasizing evidence-based dietary education. Various dietary approaches have shown promise in improving glycemic control, including carbohydrate counting (CC). Advanced CC has proven beneficial for type 1 diabetes (T1D) and T2D patients using basal-bolus insulin, but evidence for basic carbohydrate counting (BCC) in T2D is limited. Accuracy in carbohydrate estimation is crucial; some studies (mostly in T1D) show a link between accurate estimation and better glycemic control, but others do not. Multiple BCC educational approaches exist (gram counting, exchanges, experience-based estimation), yet lack thorough clinical trial evaluation. This study aimed to examine the efficacy of a structured BCC education program, added to standard dietary care, on glycemic control in T2D patients. The hypothesis was that BCC would improve glycemia (HbA1c and mean amplitude of glycaemic excursions [MAGE]) compared to standard care.

Existing literature on dietary interventions for T2D reveals a range of effective approaches, including low-fat, vegetarian, Mediterranean, high-protein, and low-carbohydrate diets. A network meta-analysis demonstrated that several dietary strategies improve glycemic control in T2D. While the total carbohydrate intake significantly impacts postprandial glucose, the effectiveness of BCC specifically for enhancing glycemic control in T2D remains under-researched. Studies on the relationship between carbohydrate estimation accuracy and glycemic control have yielded mixed results, particularly in T2D populations. There is also a lack of consistent educational approaches for BCC in clinical trials, creating a gap in the research on effective strategies for improving carbohydrate estimation skills and self-monitoring in T2D.

This single-center, parallel-group, randomized, controlled, open-label, superiority trial was conducted at Steno Diabetes Center Copenhagen. Participants (18-75 years old) with T2D (diabetes duration ≥12 months), HbA1c 53-97 mmol/mol (7.0-11.0%), and treated with diet or any glucose-lowering medication were included. Exclusion criteria included current CC practice, recent BCC program participation, use of an automated bolus calculator, low daily carbohydrate intake, gastroparesis, uncontrolled medical issues affecting dietary intake, pregnancy, breastfeeding, participation in other clinical trials, or inability to understand information/give informed consent. Participants were randomized 1:1 to either BCC or standard dietary care. The BCC group received a structured group-based educational program (8 hours) focused on practical carbohydrate estimation and management. The standard care group received three individual dietary counseling sessions (2 hours total). The primary outcomes were changes in HbA1c and MAGE from baseline to six months. Secondary outcomes included various metabolic parameters, dietary habits, and quality of life measures. Intention-to-treat analysis was performed using linear mixed-effects models with baseline correction. A sample size of 226 participants was planned, but the COVID-19 pandemic led to study delays and a smaller final sample size of 48 participants (23 BCC, 25 standard care). Statistical significance was set at p < 0.05, with adjustment for false discovery rate for multiple testing.

A total of 48 participants were enrolled (23 in the BCC group, 25 in the standard dietary care group). Seven participants did not receive the allocated intervention. Baseline characteristics were similar between groups. There were no statistically significant differences between groups in changes in HbA1c (-2 mmol/mol [-7 to 4]; p=0.554) or MAGE (-14% [-36 to 16]; p=0.319) from baseline to the end of treatment (6 months). Both groups showed comparable reductions in HbA1c and MAGE. However, the BCC group showed significantly improved median carbohydrate estimation error compared to the standard care group (-55% [-70 to -32]; p<0.001 after multiple testing adjustment). This improvement in carbohydrate estimation skills remained significant after adjustment for multiplicity. No other significant differences were found in secondary outcomes like mean plasma glucose, time-in-range, diabetes diet-related quality of life, or perceived dietitian-related autonomy support. Both groups demonstrated a decline in estimation skills at the 6-month follow-up.

This study's findings do not support the hypothesis that BCC, as an add-on to standard care, improves glycemic control in individuals with long-standing T2D. The lack of significant glycemic improvements despite improved carbohydrate estimation skills underscores the complexity of T2D management. Factors beyond carbohydrate counting, such as medication use and other lifestyle modifications, likely play a more significant role in glycemic outcomes. The observed improvement in carbohydrate estimation in the BCC group, however, highlights the potential benefit of this specific aspect of dietary education. The decline in estimation skills observed at follow-up in both groups underscores the necessity of ongoing support and reinforcement. The study's limitations, particularly the smaller than planned sample size due to COVID-19 disruptions, may have reduced the power to detect clinically significant differences. Medication adjustments during the study may also have influenced outcomes.

This study did not find that BCC improves HbA1c or MAGE compared to standard care in individuals with long-standing T2D. While BCC improved carbohydrate estimation skills, this did not translate into significant glycemic benefits. The results emphasize the need for a more holistic approach to T2D management, perhaps incorporating real-time CGM for personalized feedback, and larger-scale trials investigating the combined effects of BCC and CGM are warranted. Future studies should consider more easily adaptable study designs with improved participant retention strategies and ongoing behavioral intervention support.

The study's main limitations were the smaller than planned sample size and delays in study visits due to the COVID-19 pandemic. This reduced the statistical power, potentially leading to a failure to detect clinically relevant differences. Changes in medication during the study could also have influenced the results. The use of MAGE as a primary outcome, which is now considered less accurate than other measures of glycemic variability, is another limitation. Under- and misreporting of dietary intake, common in obese individuals, may have also influenced the findings.