Connected Insulin Cap Improves Adherence and CGM Metrics
Key Takeaways
- Among adults with insulin-treated diabetes using CGM in Spain, active Insulclock v2.0 use was associated with more on-time and fewer mistimed or missed mealtime insulin doses than masked use.
- Active device use was also associated with improved glucometrics, including higher TIR and favorable movement in other consensus CGM measures.
- In the 47 participants using rapid-acting analog (RI) boluses, median bolus timing shifted earlier relative to CGM-detected excursions during active use, and earlier timing in that subgroup was associated with lower hyperglycemic exposure; the paper reports no significant timing change in the ultrarapid subgroup.
- Questionnaire findings suggested lower treatment burden and better satisfaction across several insulin-therapy domains when the connected cap functions were turned on.
Adults with insulin-treated diabetes (ITD) using MDI and CGM in Spain moved through masked and active device phases so investigators could compare day-to-day mealtime dosing and glycemic patterns with and without those user-facing functions. The study was a pragmatic, multicenter, prospective, open-label observational evaluation with a 1-week run-in, 2 weeks of masked use, and 4 weeks of active use. 82 adults were recruited, and 52 completed analyzable follow-up. The mean age was 54.4 years, 56.6% were women, and 60.4% had type 1 diabetes. Enrollment came from three hospitals and one primary care center in Spain. Masked use withheld app access and notifications, whereas active use turned on reminders for possible missed doses, insulin-stacking alerts within 2 hours, and prompts about injection technique. Meal-related excursions were inferred from a CGM rate-of-change algorithm and paired with rapid-acting insulin data, and the phase comparison tracked adherence, glycemic metrics, and patient-reported outcomes.
Active mode was associated with more on-time mealtime dosing, which increased from 45.5% to 54.4% (p=0.0017), and with higher time in range (TIR), which increased from 56.9% to 61.9% (p=0.0054). Mistimed and missed doses also fell. Glucose management indicator (GMI), time above range greater than 180 mg/dL (TAR180), time below range less than 70 mg/dL (TBR70), and coefficient of variation (CV) also improved significantly, including TAR180 from 40.9% to 36.6% (p=0.0169), while mean glucose and more extreme hypo- and hyperglycemia measures did not clearly differ between phases.
In the paired rapid-acting meal-event analysis, median bolus timing moved from 5.6 minutes after to 5.9 minutes before meal-related excursion onset (p=0.023), and mean bolus dose decreased from 11.2 to 9.1 units (p=0.002). Earlier rapid-acting delivery was also associated with lower TAR180 in regression analysis. Most Insulin Treatment Satisfaction Questionnaire (ITSQ) items improved, including meal planning, prevention of symptomatic hypoglycemia, glucose stability, glycemic control, and ease of insulin dose selection, and the patient-reported satisfaction with diabetes treatment regimens (PRSD) hyperglycemia-risk item also favored active use.
This masked-to-active comparison was open-label, observational, and uncontrolled, so it supports association rather than causation. Follow-up was brief, and meal timing was inferred from interstitial glucose-rate changes rather than directly observed eating, which anchored the insulin-to-meal interval to excursion onset instead of the exact first bite. The authors noted an approximate 10- to 20-minute physiologic and interstitial lag in translating those intervals, and detailed adverse-event counts were not presented. Safety interpretation therefore rests mainly on the reported hypoglycemia metrics and visit-based event collection in this Spanish cohort.
Clinician Questions
How does Insulclock v2.0 classify on-time, mistimed, and missed mealtime insulin doses?
Insulclock v2.0 classified meal excursions with the GRID CGM rate-of-change algorithm, then applied preset timing windows to rapid-acting insulin doses. An on-time bolus was defined as an injection within 45 minutes before excursion onset, and meal-time doses not detected by the rate-of-change method were also categorized as on-time. A mistimed bolus occurred within 60 minutes after excursion onset, and a missed bolus meant no injection from 45 minutes before to 60 minutes after the excursion start.
Which adults with insulin-treated diabetes were represented in the Insulclock v2.0 cohort?
The cohort included adults aged 18 to 65 years with insulin-treated diabetes for at least 1 year who were already using CGM as usual care and could operate the device and complete questionnaires. Participants were enrolled across three hospitals and one primary care center in Spain, and the cohort included both type 1 and type 2 diabetes. The intervention centered on pen-based insulin administration with the connected cap rather than pump therapy.
What changed between masked and active Insulclock v2.0 use?
During masked use, participants had no app access or notifications. Active use turned on reminders for potential missed doses, alerts intended to avoid insulin stacking if a prior injection had occurred within 2 hours, and prompts about injection technique, including short injections.
Why should meal-to-insulin timing with CGM-linked Insulclock data be interpreted cautiously?
Meal timing was inferred from CGM glucose-rate changes rather than directly observed eating, so the insulin-to-meal interval was anchored to excursion onset rather than the exact first bite. The authors noted a combined physiologic and interstitial glucose lag of about 10 to 20 minutes when translating those intervals to meal start, and the short uncontrolled design further limits causal interpretation.