IMPROVED GLYCEMIC CONTROL WITHOUT INCREASING RISK OF HYPOGLYCEMIA: EVERSENSE USE IN A PEDIATRIC POPULATION- THE FEAR NO HYPO STUDY
Abstract
Background and Aims
The “Eversense XL®” sensor is a subcutaneously inserted rtCGM system which lasts up to 180 days. Actual smartphones can be used as a data receiver sent by an on body transmitter with vibration alerts. As of today, it is only approved for use in adults.
Methods
Five children (6-12 years) and 10 adolescents (13-17 years) received the sensor subcutaneously. After a 30-day blinded period, the sensor was used regularly up to 180 days. Primary endpoint was superiority of percent Time Below Range (TBR) <70 mg/dL after 90 days (assessed during 3 weeks before visit) compared to the first 30-day blinded period.
Data sharing with parents was not available during the study.
Results
15 patients (7 male, 12 CSII/3 MDI, diabetes duration 5.5±3.4 years) had the sensor inserted and completed the study. Results are shown in Table 1.No severe hypoglycemic events occurred.
Although TBR<70 mg/dL and TBR <54 mg/dL were reduced to recommended target (Battelino 2019) after 90 days, statistical significance was not reached for either target, glycemic control trended to lower mean values and TIR to enlarge.
Conclusions
In this study population with a low rate of hypoglycemia, Eversense use reduced mean glycaemia without increasing the time in hypoglycemia.Longer duration use of the system should be investigated in an adolescent population to assess long term results, as further improvement of control may take more than 6 months (COMISAIR; Soupal 2019).
IMPROVED GLYCAEMIC VARIABILITY AND HYPOGLYCEMIA THROUGH FLASH GLUCOSE MONITORING IN REAL-WORLD SETTING FROM SPAIN
Abstract
Background and Aims
We investigated glucose patterns in Spain under real-world settings to establish a relationship between glycaemic variability and hypoglycemia risk. The influence of testing frequency in glycaemic parameters was established to demonstrate flash glucose monitoring associations with glycaemic control.
Methods
Patient glucose data obtained in Spain were de-identified and uploaded into a dedicated database. Glycaemic variability (GV) metrics were described and scan rate was determined. The readers were sorted into twenty equally-sized rank-ordered groups, categorised by scan frequency. Glucose parameters were calculated for each group: estimated HbA1c, time below range (TBR) (<70 mg/dL), time in range (TIR) (70-180 mg/dL), and time above range (TAR) (>180 mg/dL). In view of the large sample size, only p <0.001 was considered statistically significant.
Results
Users (n=22,949) performed a mean of 13 scans/day with 82.4 million hours of readings, 37.1 million glucose scans, and 250 million automatically-recorded glucose readings. There was a strong positive correlation of glucose SD and CV with estimated HbA1c, TBR <70 or <54, and TAR >180 mg/dL, and negatively with TIR. GV was significantly lower in the highest compared to lowest scan rate group (39.6 to 3.9 scans/day). A low average use of 0.66 test strips per day for SMBG was observed in this data set.
Conclusions
In real-world conditions, flash glucose monitoring frequent glucose checks are linked to reduced GV and time in hyper- and hypoglycaemia. These results underline the importance of glycemic variability reduction as a central mechanism for glucose control improvement by CGM use in a real-world setting.
PERSONALIZED LINEAR ALGORITHMS FOR REAL-TIME PREDICTION OF HYPOGLYCEMIC EVENTS BASED ON CONTINUOUS GLUCOSE DATA ONLY
Abstract
Background and Aims
In type 1 diabetes management, the availability of continuous glucose monitoring (CGM) data enables the development of predictive models to generate alerts for forthcoming critical events. The aim of this work is performing a comparison of linear data-driven algorithms for the real-time prediction of future blood glucose (BG) concentration fed by CGM data only.
Methods
Auto-Regressive (AR), Auto-Regressive Moving-Average (ARMA) and Auto-Regressive Integrated Moving-Average (ARIMA) models and several identification techniques are used to learn patient specific algorithms to forecast BG concentration. The 20 proposed algorithms have been tested on a dataset composed by 124 subjects monitored for 10 days by the Dexcom G6 CGM Sensor (Dexcom Inc., San Diego, CA). The accuracy of predicted BG has been evaluated using Root Mean Square Error (RMSE), Coefficient of Determination (COD) and prediction delay (delay) while the hypoglycemic event detection capability using Recall, Precision, F1-score, False Positive per day (FP/day) and time gain (TG).
Results
Results for the three most performing algorithms are reported in Table 1.
MODEL CLASS | Glucose Prediction Metrics | Hypoglycemic Detection Metrics | ||||||
DELAY (min) | RMSE [mg/dl] | COD (%) | F1-score | Recall | Precision | FP/day | TG (min) | |
ARIMA | 25 | 22.69 | 85.65 | 71.84% | 82.69% | 63.50% | 0.65 | 10 |
ARIMA (day&night) | 25 | 24.72 | 82.55 | 31.35% | 24.59% | 43.24% | 0.44 | 10 |
AR (recursive) | 20 | 28.44 | 79.56 | 69.17% | 82.14% | 59.74% | 0.86 | 15 |
Conclusions
Personalized ARIMA model provides the most encouraging results both in terms of prediction of future glucose level and of hypoglycemic event detection.
BENEFIT OF CONTINUOUS GLUCOSE MONITORING (CGM) IN REDUCING HYPOGLYCEMIA IS SUSTAINED THROUGH 12 MONTHS OF USE AMONG OLDER ADULTS WITH TYPE 1 DIABETES (T1D)
Abstract
Background and Aims
The Wireless Innovation for Seniors with Diabetes Mellitus (WISDM) randomized clinical trial demonstrated reductions in hypoglycemia with CGM compared with blood glucose monitoring over 6 months among older adults with T1D. To determine if this improvement could be sustained over a longer period with less intensive follow-up, participants in the CGM group continued CGM through 12 months.
Methods
Of the adults ≥60 years with T1D (in the RCT CGM group, 100 of 103 (97%) continued on CGM and completed 12 months of follow-up (58% female, median age 68 years, 92% non-Hispanic white, and 58% used insulin pumps). CGM-measured outcomes were calculated at baseline (masked CGM) and 4 weeks prior to each visit over 12 months.
Results
Median percent time <70 mg/dL was reduced from 5.0% at baseline to 2.8% at 12 months (p<0.001). This improvement was evident in the first month and sustained through 12 months (Figure). Similar improvements were observed for time <54 mg/dL (median 1.9% versus 0.4%; p<0.001). In addition, participants spent more time in range 70-180 mg/dL (mean 56% versus. 64%; p<0.001), spent less time in hyperglycemia >180 mg/dL (mean 37% vs. 33%; p<0.001), and had better HbA1c (mean 7.6% versus 7.4%; p=0.01) at 12 months compared with baseline. CGM use remained high with 85% using the sensor ≥6 days per week at both 6 and 12 months.
Conclusions
These results demonstrate CGM may be used effectively in older adults with T1D and provide evidence for further integrating this technology into clinical practice.
EVALUATION OF THE MAGNITUDE OF HYPOGLYCAEMIA IN RELATION TO HBA1C IN MULTIPLE DAILY INSULIN INJECTIONS – AN ANALYSIS WITH RESPECT TO RECENT ADA GUIDELINES
Abstract
Background and Aims
According to recent guidelines, individuals with type 1 diabetes (T1D) should spend less than 4 % per day <3.9 mmol/l (<70 mg/dL) and less than 1 % per day <3.0 mmol/l (<54 mg/dL). The objective of this study was to evaluate time in hypoglycaemia in relation to various HbA1c-levels in individuals with T1D and multiple daily insulin injections (MDI).
Methods
We estimated the association between time in hypoglycaemia and HbA1c during CGM and conventional therapy (self-measurement of blood glucose) in the GOLD randomised trial (n=161), a cross-over study over 69 weeks, using fractional response model.
Results
Time in hypoglycaemia (<3.9 mmol/l and <3.0 mmol/l) increased with lower HbA1c-level estimated by CGM during CGM and conventional therapy, p<0.0001. Time in hypoglycaemia <3.9 mmol/l in relation to HbA1c is shown for CGM and conventional therapy in the figure. During CGM therapy, mean time in hypoglycaemia for individuals with HbA1c 7.0 % (52 mmol/mol) was 5.4 % <3.9 mmol/l and 1.5 % <3.0 mmol/l. The corresponding values for conventional treatment were 9.2 % and 3.5%, respectively. During CGM therapy, 27.3 % of individuals with HbA1c <7.0 % had <1.0 % time in hypoglycaemia <3.0 mmol/l and 27.3 % had <4.0 % time in hypoglycaemia <3.9 mmol/l.
Conclusions
Time in hypoglycaemia increases with lower HbA1c levels both during conventional and CGM treatment in persons with T1D treated with MDI. It is difficult for patients to reach targets for time in hypoglycaemia and simultaneously achieving the HbA1c target of <7.0% even during CGM treatment.
PILOT EVALUATION OF CONTINUOUS GLUCOSE MONITORING BY GUARDIAN CONNECT FOR HOSPITALIZED PATIENTS WITH DIABETES
Abstract
Background and Aims
In hospitalization, diabetic patients are exposed to acute situations. Continuous glucose monitoring (CGM) could be a useful tool for the physician to adapt insulin treatment and for therapeutic education of patients. We aimed to describe the glycemic control of hospitalized patients under the CGM device Guardian Connect (GC, Medtronic)
Methods
Guardian Connect GC, composed of a glucose sensor (Enlite-2) connected to an Ipad, was installed in hospitalized diabetic patients under multiple daily injections or external insulin pump. The time in range (TIR) in the target: 70-180mg/dL, in hypoglycemia <70 mg/dL and in hyperglycemia >180 mg/dL was evaluated the last 24 hours of the hospitalization. The satisfaction of patients and medical teams was collected.
Results
61 diabetic patients hospitalized for diabetic foot, acute hypo/hyperglycvemia events.. were equipped with GC (age: 57 years [18-81], T1D: 53%) for 5.7 ± 0.5 days. The TIR in target, hypoglycemia and hyperglycemia were: 63 ±22%, 4 ± 6% and 34 ± 21%. GC was useful for 92% of patients, allowed to anticipate glucose excursions in 85% of cases, facilitated therapeutic education in 69% of cases and was not a burden for 93% of patients. GC remained useful for 72% of nurses but was time consuming for 76 % of nurses due technical problems: loss of signal, calibration
Conclusions
In hospitalization, GC is a useful tool for patient education and to adapt insulin treatment. For the medical team the advantages are minimized due to technical problems. Control studies with larger number of patients are required to evaluate its metabolic impact.
EVALUATION OF CGM USE FEATURES IN ADOLESCENTS WITH TYPE 1 DIABETES (T1D): A REPORT FROM THE CGM INTERVENTION IN TEENS AND YOUNG ADULTS (CITY) STUDY
- Laurel H. Messer, United States of America
- Lauren G. Kanapka, United States of America
- Mark A. Clements, United States of America
- Daniel Desalvo, United States of America
- Kellee Miller, United States of America
- Jennifer Sherr, United States of America
- Ruth S. Weinstock, United States of America
- Lori Laffel, United States of America
Abstract
Background and Aims
To describe use of continuous glucose monitoring (CGM) by adolescents and young adults with T1D enrolled in the CITY trial.
Methods
Participants (N=153) were randomized to CGM (Dexcom G5) or blood glucose (BG) monitoring for 26 weeks, with CGM users receiving educational support. Data on use of CGM are reported for the CGM group (n=74, mean age 18 ± 3 years, 45% female).
Results
At 26-weeks, 86% of participants were using CGM with average use 6.4 (3.5, 7.0) days/week (median, Q1, Q3). Median daily blood glucose (BG) checks in this group was 2.0 (1.8, 3.0) with 98% regularly using CGM readings without a confirmatory BG for insulin dosing (table). Most participants utilized low and high alerts (91% and 84%, respectively), with median low alert setting at 70 mg/dl (70, 80) and median high alert setting at 270 mg/dl (240, 300). The mobile app was used to view CGM data by 81% of participants and 62% of these users shared CGM data via remote monitoring. Most were sharing with one (59%) or two (28%) people: parents/guardians (91%), siblings (13%) and significant others (9%).
Conclusions
In the CITY trial, the majority of adolescents/young adults were using CGM regularly after 6-months, in contrast to previous descriptions of CGM use in this age group. Participants maintained high use of the mobile app, alerts and share features, possibly related to close follow up and education after CGM initiation. Future analyses should assess use of which features predict glycemic improvements.