From a sample of 873 FreeStyle Libre 2 (FSL2, Abbott, USA) users with T1DM treated in the endocrinology department of a university hospital in Spain, described elsewhere [28,29,30], a total of 755 patients (429 men [56.8%]; median age 49 years, range 18–90 years) with sensor usage ≥ 70% [31] were selected for this cross-sectional study. T1DM was defined by the presence of a diagnosis of diabetes [32], evidence of pancreatic islet autoimmunity (positivity for autoantibodies to the glutamic acid decarboxylase, the tyrosine phosphatase-related islet antigen 2, or the zinc transporter 8) [32], and insulin requirement, given that only patients with insulin dependency were included. Therefore, subjects with stage 1 of T1DM (evidence of pancreatic autoimmunity but normoglycemia, thus no insulin requirement) [32] were excluded. Patients younger than 18 years and pregnant women were also excluded.

FreeStyle Libre 2 is an intermittently scanned CGM system commercialized by Abbott in 2018 [33]. The system provides interstitial glucose measurements via a 14-day lifetime sensor [33]. In the Spanish region of Asturias, where this study was conducted, FSL2 devices are financed by the public healthcare system for people with T1DM. Data were recorded through LibreView, a cloud-based platform that enables FSL2 users to share their glucose data with healthcare practitioners. Data were collected from August 18 to September 21, 2023, selecting the available 14-day period closest to the data collection dates, with a maximum of 1 year prior the data collection date (the oldest piece of data was from August 27, 2022). For patients with sensor usage so low that the system was unable to calculate the glucose management indicator (GMI), the available 14-day period closest to the data collection date that permitted GMI calculation was selected.

Patients were divided into three categories according to the presence of different forms of nocturnal hypoglycemia during the 14-day period, as follows: no nocturnal hypoglycemia, only nocturnal hypoglycemia that was not followed by hyperglycemia before 6:00 am, and ≥ 1 episode of nocturnal hypoglycemia that was followed by hyperglycemia before 6:00 am (PHNH). The number of nights (0:00 am to 6:00 am) with nocturnal hypoglycemic episodes during the 14-day period was recorded, differentiating hypoglycemia that was not followed by hyperglycemia before 6:00 am and PHNH. A patient was considered to have had a night with an episode of nocturnal hypoglycemia when the LibreView graph recorded a glucose value of < 70 mg/dL (either when a glucose determination of < 70 mg/dL was recorded, or when the glucose line went below the 70 mg/dL limit) between 0:00 am and 6:00 am, regardless of its duration (Fig. 1, panels A, B and C). Episodes of nocturnal hypoglycemia were considered PHNH when nocturnal hypoglycemia (hypoglycemia between 0:00 am and 6:00 am) was followed by a glucose value of > 180 mg/dL (either when a glucose determination of > 180 mg/dL was recorded or when the glucose line went above the > 180 mg/dL limit) before 6:00 am (Fig. 1, panels B and C). PHNH episodes were further differentiated by whether hyperglycemia was maintained at 6:00 am or whether hyperglycemia was corrected before 6:00 am (Fig. 1, panels B and C).

Examples of how diurnal and nocturnal hypoglycemia were defined. (A) Nocturnal hypoglycemia that was not followed by hyperglycemia before 6:00 am. Episodes like the one marked with an asterisk, where nocturnal hypoglycemia extended beyond 6:00 am, were classified as both nocturnal and diurnal hypoglycemia. (B) PHNH with hyperglycemia at 6:00 am. (C) PHNH without hyperglycemia at 6:00 am. (D) Diurnal (6:01 am to 23:59 pm) hypoglycemia

CGM-related variables. Percentages of the day during which the patient had certain interstitial glucose concentrations (time above range level 2 [> 250 mg/dL], time above range level 1 [181–250 mg/dL], time in range [70–180 mg/dL], time below range level 1 [54–69 mg/dL], and time below range level 2 [< 54 mg/dL]), GMI (this indicates the patient's estimated glycated hemoglobin level based on the mean glucose level), glucose variability (GV, expressed as the coefficient of variation) [31], number of hypoglycemic events (glucose levels < 70 mg/dL during at least 15 min) that occurred during the 14-day period and its mean duration, number of days during the 14-day period with diurnal hypoglycemic episodes (as recorded in the LibreView graph from 6:01 am to 23:59 pm; Fig. 1, panel D), scans per day (number of times per day the patient scanned the FSL2 device to check the glucose levels), and use of hypoglycemia alarms of the FSL2 devices (whether the alarm was set and, if so, the threshold [mg/dL] at which it was set).

Patient characteristics. Age, sex, duration of diabetes (calculated in years from the time of diagnosis of diabetes according to the American Diabetes Association criteria) [32], diagnosis of latent autoimmune diabetes in adults (LADA, a subtype of antibody-positive T1DM defined as adult-onset diabetes [> 30 years of age at diagnosis] with no need for insulin treatment for at least 6 months after diagnosis) [34], presence of advanced chronic complications of diabetes (including microvascular complications such as diabetic retinopathy [proliferative diabetic retinopathy, diabetic retinopathy with macular edema, and diabetic retinopathy that required photocoagulation], diabetic kidney disease that required evaluation by a nephrologist, and diabetic peripheral neuropathy; and macrovascular complications such as peripheral artery disease that required evaluation by a vascular surgeon, cerebrovascular disease, myocardial infarction, and heart failure), body mass index, and smoking status.

Insulin therapy. The total daily dose of insulin (TDDI), expressed in international units, was recorded. A total of 32 (4.2%) patients were treated with an insulin pump (open-loop insulin pumps only, given that closed-loop systems cannot be integrated with the FSL2 system) [35] and 723 (95.8%) received subcutaneous insulin. Among patients on subcutaneous insulin, 695 (96.1%) used basal-bolus therapy, 21 (2.9%) used basal insulin alone, six (0.8%) used premixed insulins, and one (< 0.1%) used only rapid-acting insulin. For patients on basal-bolus therapy, the specific types of basal and rapid-acting insulins were recorded, distinguishing next-generation basal insulins (degludec and U-300 glargine) from earlier-generation formulations (detemir and U-100 glargine), and next-generation rapid-acting insulins (faster-aspart) from earlier-generation formulations (aspart, glulisine, lispro, and human insulin). Additionally, the timing of basal insulin administration was recorded for patients on basal-bolus therapy.

We employed the chi-square test to compare proportions (with Yates’ correction applied for 2 × 2 contingency tables when the expected frequency in any cell was less than 5), the Mann–Whitney test to compare numerical data between two independent groups, and Spearman’s rank test to assess correlation. Logistic regression was used for multivariate analyses of factors associated with PHNH. All tests were two-tailed and P-values lower than 0.05 were considered statistically significant. Data were analyzed using IBM SPSS Statistics version 20 (IBM Corp., USA).