Death to carbohydrate counting?

Insulin therapy is an effective strategy for achieving glycemic control in patients with type 2 diabetes. Although often neglected, it is important to use an appropriate diet strategy to complement the insulin. Furthermore, a basic tenet of such therapy is that insulin dosage and administration should be appropriate to balance diet and physical activity in order to maintain normoglycemia.

In this issue of Diabetes Care, Bergenstal et al. (1) evaluated two strategies for determining the appropriate dosage of mealtime bolus insulin. A simple algorithm that adjusted bolus insulin dose based on weekly average of premeal glucose was compared with an algorithm based on mealtime carbohydrate counting. The authors demonstrated the equivalence of both the simple strategy and the more elaborate carbohydrate-counting strategy in achieving glycemic control; almost one-half the participants in both groups achieved an A1C <6.5%. Can patients with type 2 diabetes treated with basal:bolus insulin succeed without adding the complexity of carbohydrate counting?

Carbohydrate counting has been around since the 1920s and became integral in managing patients with type 1 diabetes after the landmark findings of the Diabetes Control and Complications Trial (2). However, the efficacy of carbohydrate counting in type 2 diabetes is largely unknown. Potential barriers to carbohydrate counting include the time and effort required for patients to count the carbohydrate content at each meal, patient difficulties in understanding the strategy, and the availability of dietitians or appropriately trained health care providers to teach patients. The authors did not evaluate quality-of-life outcomes, and little is known about how carbohydrate counting affects quality of life. Prior studies have demonstrated that, when given a choice, patients opt to discontinue carbohydrate counting over other strategies (3). Further research is clearly needed on the optimization of carbohydrate counting in type 2 diabetes and also in assessment of the accuracy of patients’ counting in the real world setting.

It would appear that if similar levels of glycemic control can be achieved with few adverse events by using a simple algorithm, then the simple algorithm may be a better strategy for adjusting insulin. Additionally, the simple algorithm may be more feasible to teach patients within underserved settings where a dietitian may not be available. However, are there benefits to counting carbohydrates beyond glycemic control?

In examining the study's secondary outcomes, there is an apparent trend for the carbohydrate-counting group to have less weight gain at the end of the 24-week period. The carbohydrate group had a weight gain of 2.3% compared with a 3.4% increase in the simple algorithm group. The 1% difference in weight gain over 6 months did not reach statistical significance, but the study was not adequately powered to test the significance of such a difference. What would be the weight effects over a longer time period? Could the higher insulin dosage used in the simple algorithm group cause greater weight gain? Weight gain is a well-known side effect of insulin therapy, and previous studies demonstrate weight increases of ≤21% in one year with some insulin regimens (4). Weight management is a critical aspect of type 2 diabetes, and it will be important to evaluate strategies to minimize weight gain while using insulin therapy. Reducing carbohydrates can be an important strategy for improving glycemic control and weight loss. Did the carbohydrate-counting group have a lower carbohydrate or caloric intake than the simple algorithm group? The Look AHEAD (Action for Health in Diabetes) Trial, which is evaluating the potential benefits of weight control in type 2 diabetes, found that the three most common weight control strategies used by participants were increasing fruits and vegetables, cutting out sweets, and eating fewer high-carbohydrate foods (5). Counting carbohydrates may increase dietary awareness of the carbohydrates being consumed and subsequently reduce carbohydrate consumption.

A greater number of patients in the simple algorithm group completed the study in comparison with the carbohydrate-counting groups (91.2 vs. 79.6%), which may suggest greater ease of compliance with the simple algorithm. The overall adverse event rate was similar between the treatment groups, but reported self-blood glucose monitoring <50 mg/dl with symptoms was slightly more common in the carbohydrate-counting than in the simple algorithm group. Although there was no statistically significant difference in the rates of hypoglycemia with using either dosing algorithm, the simple algorithm had 53 episodes of hypoglycemia in 19 patients, whereas the carbohydrate counting group had 37 episodes in 19 patients. Although the reasons for this are unclear, we question whether this reflects more real-time adjustment in bolus insulin dose in the carbohydrate group rather than the weekly adjustment in dose in the simple algorithm.

Insulin management continues to be complex and requires close monitoring both by patients and their physicians. Bergenstal et al. (1) have developed an algorithm to simplify the management of insulin regimens containing basal and mealtime insulin. Patients with type 2 diabetes may achieve glycemic targets with a simple basal:bolus insulin algorithm without the added burden of counting the carbohydrate content of each meal. Glycemic control, however, is one of many aspects of diabetes management, and we need to be mindful of other important aspects including weight control and risks of hypoglycemia. Bergenstal et al. have certainly initiated the impetus for us to examine the relevance of carbohydrate counting for insulin dosing in type 2 diabetes, but carbohydrate counting may have a life for more than just medication adjustment.