Does A1c Have to Change for Hypoglycemia Reduction?

The article by Weiss et al. ¹ published in this issue of Diabetes Technology & Therapeutics points to the complex relationships between A1c levels and hypoglycemia occurrence in patients with type 1 diabetes and, moreover, to the question of their expected respective evolution following insulin regimen changes and use of diabetes technology aiming at improved glucose control. So far, the relationships between A1c level and hypoglycemia occurrence have been deeply influenced at the individual level by two main contributors: the mode of insulin delivery, with its own pharmacokinetic and -dynamic properties, and patient behavior in the management of insulin therapy, deriving from his or her motivation to reach a certain A1c target and his or her fear of hypoglycemia. Automated adjustment of insulin delivery based on glucose levels driven by new coming algorithms introduces a third player in the game.

The A1c level is a summary of the percentage of time spent in hyperglycemia and the amplitude of hyperglycemia above normal range during the preceding 2–3-month period. ² The Diabetes Control and Complications Trial has well demonstrated the tight relationship between long-term A1c levels and the development of diabetes complications related to chronic hyperglycemia. ³ However, another message from the Diabetes Control and Complications Trial was that the reduction of A1c levels was associated with an increased risk of severe hypoglycemia. ⁴ This relationship was observed with an insulin regimen made up of regular and NPH insulin for multiple daily injections and regular insulin in pumps.

Insulin therapy has dramatically changed during the two last decades thanks to insulin analogs and developed diabetes technology. The currently available treatment modes allow a better fulfillment of insulin needs at each time and result in lower A1c levels because glucose peaks due to meals or any stressful condition and glucose levels determined by hepatic glucose production are better controlled.

The availability of fast- and long-acting insulin analogs has improved the efficacy of basal-bolus insulin regimens on A1c levels versus basal-bolus insulin therapy using regular and NPH insulin. ⁵ The quicker action of fast-acting analogs reduces the magnitude and time spent in hyperglycemia following meals, ⁶ while the combined long-acting analogs can prevent blood glucose level elevation in fasting and late postmeal periods. ^7,8 Continuous subcutaneous insulin infusion (CSII) from insulin pumps can further reduce time spent in hyperglycemia by making easier frequent adjustments of insulin delivery in case of hyperglycemia. ⁹ CSII using fast-acting analogs results in even lower A1c levels than CSII using regular insulin because it allows a higher insulin bolus to avoid postmeal hyperglycemia with less risk of later hypoglycemia. ¹⁰ Sensor-augmented pumps can lead to further improvement of A1c levels by promoting insulin correction doses when needed via the availability of continuous glucose monitoring of the glucose level. ¹¹

Because this gradual evolution of insulin treatment modes is also associated with a higher flexibility and more focused adjustment of insulin levels, it also results in a lower occurrence of hypoglycemia. According to this logic, one may conclude that the refinement of insulin delivery modes leads to a parallel reduction of A1c and hypoglycemic events. Intraperitoneal insulin therapy confirms the effectiveness of closer-to-physiology pharmacokinetics and -dynamics of insulin delivery on both A1c and hypoglycemia. ¹² The availability of a system that identifies low glucose levels when the patient may be unaware of them and acts automatically to stop insulin delivery can provide a specific benefit by a further reduction of hypoglycemic events as well as time spent in hypoglycemia as shown in the article by Weiss et al. ¹ Whatever the initial A1c level, the patients who lowered A1c levels also experienced lower occurrence and/or severity of hypoglycemia. The even more sophisticated systems of closed-loop insulin delivery based on control-to-range algorithms currently under investigation also combine improved percentage of time spent in close-to-normal glucose range (i.e., expected lower A1c level on longer term) and less or stable percentage of time in hypoglycemia, but no increase in hypoglycemia. ^13,14

This parallel decrease in A1c level and hypoglycemia allowed by new treatment modes fails in patients who keep high A1c levels and frequent hypoglycemia. When using the low glucose suspend algorithm in this category of patients, only those who decreased their A1c level also significantly reduced hypoglycemia as reported by Weiss et al. ¹ This phenomenon may be related to the behavior of patients in insulin management, which can be changed by the availability of the new safety system. Indeed, patients who keep higher A1c levels in spite of use of a sensor-augmented pump may be less motivated for reaching target A1c or may be prevented from adjusting insulin delivery in order to do so because of a greater intrinsic fear of hypoglycemia. The role of education to insulin management in changing both A1c level and occurrence of hypoglycemia was demonstrated a long time ago and further illustrated by the DAFNE study. ¹⁵ However, behaviors may also be driven by some features of diabetes, such as high sensitivity to insulin or less residual insulin secretion. ¹⁶ Recent studies have shown that higher glucose variability or more frequent hypoglycemia can be influenced by these factors. ^16,17 Many other factors can be considered as having a role in patients' accepted target for A1c, such as psychological trauma related to previous experience of severe hypoglycemia, variable absorption of subcutaneous insulin, or gastroparesis resulting in variable glucose absorption.

Higher variability in glucose levels has been reported as correlated with higher A1c levels in type 1 diabetes. ¹⁸ Although this relationship can be broken by more adjustable mode of insulin delivery (e.g., insulin pumps), the current question is whether the availability of algorithms that modulate insulin delivery without patient decision can further reduce this deleterious combination of high glucose variability leading to defect of insulin adjustment by fear of hypoglycemia and ending in sustained A1c level above target. The article of Weiss et al. ¹ suggests that the availability of the low glucose suspend tool may play a role in promoting better adjustment of insulin doses leading to lower A1c levels with reduction of hypoglycemia.

Coming back to the question in the title of this commentary, improving A1c thanks to the support of innovative technology for diabetes is indeed a path that can be associated with hypoglycemia reduction. When A1c is already close to target, hypoglycemia reduction can be obtained with diabetes technology with no change of A1c. Keeping a high A1c level or increasing the A1c level lowers the probability of reducing hypoglycemia, but severity of hypoglycemia may still be reduced by diabetes technology.