The use of isCGM leads to marked reduction in severe hypoglycemia requiring emergency medical service or hospital admission and diabetic ketoacidosis in adult type 1 diabetes patients

**Background:** Hypoglycemia and diabetic ketoacidosis (DKA) are acute, life-threatening complications of type 1 diabetes. While continuous glucose monitoring (CGM) improves glycemic control, evidence on whether intermittently scanned CGM (isCGM) reduces severe hypoglycemia is mixed, partly because many studies rely on hospital admission or self-report data, which may miss events managed by emergency medical services (EMS) at home. This study aimed to determine the effect of isCGM on acute diabetes-related complications by combining EMS and hospital admission data from electronic health records.

**Methods:** This retrospective real-world analysis was conducted in the Siun sote region of Eastern Finland (catchment population ~166,400). From 1764 patients with type 1 diabetes (ICD-10 E10.1–E10.9), 642 adults (≥18 years) who started isCGM (FreeStyle Libre 1, no alarms) before the end of 2019 and were domiciled in the region were included. Data were collected from January 2015 to April 2020. Severe hypoglycemic events were identified from EMS records (validated by a diabetologist) and hospital admissions (ICD-10 E10.00); DKA events from hospital admissions (ICD-10 E10.1). Events at diabetes diagnosis were excluded. The primary outcome was the rate of hypoglycemia requiring EMS involvement or hospital admission and DKA events. HbA1c was recorded at isCGM start and at the last measurement during follow-up. Incidence rates (events/1000 person-years) before and after isCGM initiation were compared using Poisson regression with generalized estimating equations (GEE), with each patient as their own control. Subgroup analyses by baseline HbA1c (<7%, 7–9%, >9%) were performed post hoc.

**Key Results:** Mean age was 47.0 years (SD 16.2), mean baseline HbA1c 8.7% (71.6 mmol/mol), and median follow-up before and after isCGM initiation was 2.87 and 2.45 years, respectively. A total of 220 hypoglycemic events were identified (148 before, 72 after isCGM). The incidence rate of hypoglycemic events decreased significantly from 76 to 50 events/1000 person-years (p=0.043). Of the 213 EMS-attended hypoglycemic events, 51 (23.8%) resulted in transport to the emergency department. DKA events totaled 35 (29 before, 6 after), with incidence dropping from 15 to 4 events/1000 person-years (p=0.002). Combined acute complications (hypoglycemia or DKA) decreased from 90 to 54 events/1000 person-years (p=0.009). Mean HbA1c decreased by 0.28% (3.1 mmol/mol) from baseline to last measurement (p<0.001). In subgroup analyses, patients with baseline HbA1c >9% showed a trend toward fewer hypoglycemic events (p=0.064) and a significant reduction in DKA (31 vs. 5 events/1000 person-years, p=0.002). No significant differences were found by age, gender, or insulin delivery method (CSII vs. MDI).

**Clinical Implications:** This study demonstrates that isCGM without alarm functions significantly reduces severe hypoglycemia requiring EMS or hospital admission and DKA in adult type 1 diabetes patients, in addition to modestly lowering HbA1c. By capturing EMS-attended events not resulting in hospital transport (76.2% of hypoglycemic events), the study provides a more complete picture of severe hypoglycemia than hospital-based data alone. The reduction in DKA was most pronounced in patients with the highest baseline HbA1c (>9%). These findings support the use of isCGM as an effective tool for preventing acute diabetes complications, even without real-time alarms. Limitations include the lack of a control group, relatively small sample size for subgroup analyses, and absence of sensor metrics (e.g., time in range). Further research is needed to confirm long-term benefits and to compare isCGM with alarm-enabled CGM systems.