Response to Comment on Šoupal et al. Glycemic Outcomes in Adults With T1D Are Impacted More by Continuous Glucose Monitoring Than by Insulin Delivery Method: 3 Years of Follow-up From the COMISAIR Study. Diabetes Care 2020;43:37–43

We thank Telliam and Thivolet (1) for their interest in our study of the use of real-time continuous glucose monitoring (rtCGM) versus self-monitoring of blood glucose with insulin pumps and multiple daily insulin injection (MDI) therapy. However, we have some questions about their study findings.

First, it is impossible to assess the significance of their findings because they have not provided baseline values for comparison. Was there an increase or decrease in time spent with blood glucose <70 mg/dL?

Second, as shown in their Fig. 1, the time below range (TBR) (<70 mg/dL) was notably lower among participants with higher mean glucose (e.g., >200 mg/dL). This suggests that while these MiniMed 640G users may have experienced less TBR, their percentage of time spent above range remained suboptimal. The recent CGM consensus report recommends a composite metric for assessment, combining reductions in TBR with increases in time in range (2). In our study, both the rtCGM+MDI and sensor-augmented pump study groups achieved significant improvements in both metrics (3). Importantly, 14 patients were treated with a Medtronic Paradigm Veo pump with low glucose suspend; there was no difference between Paradigm Veo and Animas Vibe with Dexcom G4 (n = 12) in hypoglycemia. Because data regarding improvements in time in range were not reported in the letter from Telliam and Thivolet (1), it is difficult to assess the full significance of their study findings.

Third, their letter does not offer relevant data, such as study duration, percentage of time MiniMed 640G was used in the suspend mode, and number of participants who discontinued 640G. In our study, the mean percentage of rtCGM use was high in both the rtCGM+MDI group (85.7%) and the rtCGM+CSII (continuous subcutaneous insulin infusion) group (86.7%) under real-world conditions over a 3-year period (3). It is difficult to assess the true value and utility of the 640G system without information about participants’ percentage of use and long-term persistence, particularly within the context of a controlled trial.

Fourth, use of an insulin pump with the FreeStyle Libre CGM system raises serious concern about using this intervention as a comparator. Unlike sensors that automatically alert patients to current or impending hypoglycemia, the FreeStyle Libre requires users to scan their sensor in order to detect current and/or impending acute glycemic events. In essence, their study compared a system that automatically suspends insulin delivery when low glucose is detected with an intervention that relies on participants to scan frequently in order to detect and address current or impending hypoglycemia. This difference between FreeStyle Libre and other CGM systems is particularly relevant if the study included a large number of participants with impaired hypoglycemia awareness. Unfortunately, this information was not reported. Use of a sensor that provides automatic alarms and alerts would have been more appropriate in their comparison.

We agree that the study described by Telliam and Thivolet demonstrates that use of an insulin delivery system that provides preventive low–blood glucose suspend management is an effective therapy for many (but not all) individuals at high risk for severe hypoglycemia. A long-duration, real-world evaluation, using CSII with sensors that provide automatic alarms/alerts, is needed to justify the additional costs associated with 640G therapy. Not all sensors are equivalent.