Background and Aims

Profusa is developing a soft glucose-sensitive hydrogel, a sterile, small (5.0x0.75x0.65mm when hydrated), intended for subcutaneous injection. The hydrogel is suitable for long term use and could remain in the body permanently. Presented studies assess the platform function, transitioning from animal studies to human studies.

Methods

The glucose sensor is comprised of a hydrogel scaffold with covalently bound fluorescent molecules that produce near infrared light proportional to glucose concentration when interrogated by an optical reader placed on the skin (Figure 1). The v3.0 Reader utilizes a set of LEDs to excite fluorophores within the Hydrogel. The LEDs reside in the Reader which sits on the skin above the Hydrogel. The reader communicates wirelessly to a tablet for control and data transmission (Figure 2). Pre-clincal performance was validated in a swine mode in multiple 8-hour experiments over 3 months.Performance was assessed during in-clinic visits in subjects with diabetes over a period of 3 months. Frequent blood glucose measurements with a Super GL laboratory analyzer were collected as reference.

Results

This study was designed to characterize baseline performance and signal perturbations due to environmental conditions, including motion, temperature, and ambient light. This characterization is used to develop and refine the signal processing algorithm to calculate glucose levels.The following graphs (Figures 3-5) demonstrate the successful identification of a glucose excursion in a single subject during the first two weeks.

Conclusions

Results of using this hydrogel and reader in the swine model are confirmed in this clinical study, tracking glucose during study days over a 3-month period.

Background and Aims

The study aim is to compare sensor accuracy after subcutaneous (s.c.) infusion of glucagon near to versus remote from sensor site.

Methods

Seventeen adults with type 1 diabetes wore two continuous glucose monitors (Dexcom® G6; CGM_glucagon and CGM_control) placed on each side of the abdomen during a 33-h in-clinic visit. During the study visit, participants’ plasma glucose was controlled with DiaCon’s dual-hormone insulin-glucagon closed-loop system, consisting of a smartphone, CGM_control and two s.c. infusion pumps (Dana RS insulin pump, Sooil) for delivery of insulin and glucagon (GlucaGen®, Novo Nordisk). The infusion site for glucagon was placed 2 cm next to the CGM_glucagon. The CGM performance for the 33 hrs were compared with the plasma glucose measured with Yellow Spring Instrument 2900 (YSI).

Results

In total, 1,076 paired YSI–CGM_glucagon and 1,172 paired YSI–CGM_control were collected from the 17 participants. Using YSI as comparator, no difference in the median (interquatile range) absolute relative difference (MARD) for CGM_glucagon and CGM_control was found (16.7 (8.3-22.9) % vs. 11.0 (8.0-17.5) %; P_Wilcoxon= 0.105). Values in zone A + B of Clarke error grid analysis did not differ between CGM_glucagon and CGM_control using YSI as reference measurement (94.1% vs. 94.9%, P_McNemar = 0.103). The precision absolute relative deviation between sensors was 6.9 %.

Conclusions

Sensor accuracy was not significantly affected by s.c. infusion of glucagon near to sensor site.

Background and Aims

We report for the first time normative glucose metrics for %TIR 3.9-10 mmol/L, %TBR <3.9 mmol/L and %TAR >10 mmol/L for adult FreeStyle Libre users in the UK within 4 defined age-groups and the impact of the COVID-19 pandemic on these over time.

Methods

Data was extracted from 8,914 LibreView® de-identified user accounts from adult users aged 18+ with ≥5 days of sensor readings in each month from January to June 2020. Age-group categories were based on self-reported age on LibreView accounts (18-25, 26-49, 50-64, 65+ yrs).

Results

Adult FreeStyle Libre users in this UK study report mean TIR above 51% and median TBR 4.2% or below. In January, prior to the pandemic, the 65+ age group had the highest %TIR (57.9%) while the 18-25 age group had the lowest (51.2%). Consensus targets >70% TIR and <4% TBR were achieved by 11.7% of users aged 65+ compared to 6.0% of those 18-25. Within each age group, TIR change was significant during the analysed months (p<0.001) by repeated measures ANOVA. Comparing January to June 2020, all age groups increased TIR, with those aged 65+ increasing the most (3.1%; p<0.001). The proportion of adults achieving both >70% TIR and <4% TBR targets increased to 15.9% for those 65+ years (p<0.001) and 9.1% for 18-25 (p<0.05).

Conclusions

During lockdown in the UK, the proportion of adults achieving TIR consensus targets increased among FreeStyle Libre users. The glucose metrics suggest an opportunity for even further improvement with continued CGM use.

Background and Aims

COVID-19 patients are often affected by Diabetes Mellitus (DM) and/or hyperglycemia, and the management of their glycemia is challenging.

Aim of this study is to assess clinical utility of intermittent scanning Continuous Glucose Monitoring (is-CGM).

Methods

We assessed many different clinical, bioumoral and Ambulatory Glucose Profile (AGP) derived parameters in 51 consecutive inpatients (male = 39 patients) (age, mean ± DS = 65.9 years ± 12.5) affected by COVID-19 pneumonia and with DM (n = 31) or prediabetes (n = 11) and/or fasting glycemia > 125 mg% (n = 9).

Main outcomes were: median glucose (compared to goal) (MG) across 5 different daily time intervals (hours 03-08; 08-12; 12-18; 18-22; 22-03); detection of episodes of hyperglycemia (> 180 mg% for > 15 minutes) or hypoglycemia (< 70 mg% for 15 minutes) by routinely performed blood glucose monitoring (BGM; pre-breakfast, pre-lunch, pre-dinner and 2 h post-meals) and by is-CGM.

Results

MG were significantly > than goal and > than 20% and 40 mg% greater than the whole day median in: 3/51 (hours 03-08), 4/51 (hours 08-12), 18/51 (hours 12-18), 13/51 (hours 18-22) and 2/51 (hours 22-03) patients (p < 0.0001).

We detected significantly more episodes of hyperglycemia and hypoglycemia with CGM than BGM (respectively: 11.3 ± 8.5 vs 7.7 ± 6.2, p < 0.0001; 3.4 ± 5.4 vs 1.7 ± 3.2; p < 0.0001).

Conclusions

In COVID-19 inpatients with diabetes, prediabetes or fasting hyperglycemia is-CGM showed peculiar data about glycemic variability across the day and about hyper- and hypoglycemic episodes, indispensable to optimize the glycemic management.

Background and Aims

Digital diabetes self-management tools can support glycaemic control in people with type 1 diabetes (T1D). The Hedia Diabetes Assistant (HDA), an mHealth-based medical device, was designed to assist insulin dosing and daily decision making. Our aim was to investigate if T1D users of HDA can achieve clinically relevant improvements in glycaemic control after 12 weeks use.

Methods

Anonymized data from engaged users (≥10 logs/week for 12 weeks) were extracted from the HDA database. Outcomes were changes in eA1c (estimated HbA1c), and estimated time-in-range (eTIR, proportion of blood glucose level [BGL] measurements within 3.9-10.0mmol/l) after 12 weeks. We applied GLMM to data from all users, and to a subgroup of more poorly controlled users. Sensitivity analyses included t-tests, and GLMM leaving out weeks 0-1 from the analyses to mitigate for regression toward the mean.

Results

Results: The 234 engaged users had a mean age of 45.5 (SD 16.3) years and 116 (49.6%) were female. After 12 weeks, the more poorly controlled users (mean baseline BGL≥10mmol/l, eA1c≥7.9% in week 0; n=84) experienced statistically significant improvements in eA1c and eTIR. Improvements across all users (n=234) were not significant. Sensitivity analyses confirmed our results.

Conclusions

Statistically significant and clinically relevant improvements in glycaemic control were observed after 12 weeks of using the HDA in individuals with poorly controlled T1D. Findings support the relevance of mHealth in T1D, however, further studies including randomized controlled trials are needed to substantiate our observational findings.

Background and Aims

The objective of this study was to identify the baseline demographic and clinical characteristics associated with Diabetes-related distress in people living with Type 1 diabetes.

Methods

The study was performed using baseline data from the ABCD nationwide FreeStyle Libre audit. We collected diabetes-related distress scores at baseline with two items diabetes-related distress score (DDS) and follow up. An average item score of ≥3 (moderate distress) discriminated high from low-distressed subgroups. We used an unsupervised gradient boosting machine learning model (GBM) to identify the relative influence of baseline parameters on two components of DDS. The results of the GBM model were confirmed using linear regression analysis.

Results

The study consisted of 9124 people with Type 1 diabetes, with a mean age of 45.1(±15.3) years, 50.3% female, mean BMI of 26.5(±6.2) kg/m² and mean baseline HbA1c 70.2(±18.3) mmol/mol. High diabetes-related distress was prevalent in 5476(60%) of people living with T1D at baseline. The two components of the DDS were significantly correlated (r2=0.73 P<0.0001). In the GBM model, baseline HbA1c (RI=51.4), GOLD score (RI=23.3), gender (RI=7.05) and fear of hypoglycaemia as an indication for starting on the FSL (RI=4.9) were associated with diabetes-related distress. The linear regression model confirmed that higher baseline HbA1c, higher GOLD score, female gender and fear of hypoglycaemia were significantly associated with DDS.

Conclusions

In this large UK cohort of people living with Type 1 diabetes, diabetes distress was prevalent and associated with high HbA1c, impaired awareness of hypoglycaemia and female gender.

Background and Aims

Aims

To explore the impact of the coronavirus disease lockdown on diabetes patients living in Jeddah, Saudi Arabia, in terms of their compliance with medication intake and lifestyle habits, and quality of life.

Methods

Methods

In this cross-sectional, qualitative prospective study, a questionnaire was administered over the telephone to diabetes patients who had attended National Guard primary care centers in Jeddah, Saudi Arabia. The survey included questions on demographic data, type of diabetes, medications used, comorbidities, medication compliance, and daily habits before and after the lockdown, and those assessing patients’ psychological parameters during the past month by using the Kessler Psychological Distress Scale (K10). Data analysis was performed using SPSS program version 26.

Results

Results

Totally, 394 patients participated. All of them had type 2 diabetes, and 37.6% had only one comorbidity. Antidiabetic monotherapy was used in 76.4% of the patients, while combination therapy was used in 23.6%. The compliance score before the lockdown was significantly higher (18.49 ± 3.05) than that after it (17.40 ± 3.25) (p-value <0.001). The average psychological assessment score was 9.78 ± 4.14 (range 8–35). Male participants and smokers had a significantly better psychological status than female participants (p-value = 0.002) and non-smokers (p value < 0.001), respectively.

Conclusions

Conclusions

The patients’ levels of compliance with medications and healthy lifestyle habits were significantly reduced after the lockdown. These findings highlight the need for healthcare professionals to encourage diabetes patients to adhere to healthy lifestyle habits and use telemedicine during lockdowns to ensure optimal blood glucose control and reduce the incidence of complications.

Background and Aims

The optimal method of monitoring glycemia in pregnant women with type 1 diabetes remains controversial. This study aimed to assess if continuous glucose monitoring (CGM) metrics and alternative biochemical markers of glycemia could improve ability to identify pregnancies at risk of suboptimal obstetric and neonatal outcomes compared to biochemical markers including HbA1c.

Methods

157 women from the CGM in pregnant women with type 1 diabetes trial (CONCEPTT) were included in this pre-specified secondary analysis. HbA1c, CGM data, and alternative biochemical markers (glycated CD59, 1,5 anhydroglucitol, fructosamine and glycated albumin) were compared at approximately 12, 24 and 34 weeks gestation using logistic regression and ROC curves to predict pregnancy complications (pre-eclampsia, preterm delivery, large-for-gestational-age, neonatal hypoglycemia, admission to neonatal intensive care unit).

Results

HbA1c, CGM metrics, and alternative laboratory markers were all significantly associated with obstetric and neonatal outcomes at 24 weeks gestation. HbA1c, Time-in-range (TIR; 63-140 mg/dl; 3.5-7.8 mmol/l) and time-above-range (TAR; >140 mg/dl; >7.8 mmol/l) were the most consistently predictive CGM metrics and showed good predictive ability for many outcomes. Some alternative laboratory markers showed promise, but overall, they had lower predictive ability than HbA1c.

Conclusions

CGM metrics TIR and TAR performed well and had comparable predictive ability to HbA1c for many outcomes. Alternative biochemical markers of glycemia and other CGM metrics did not substantially improve the prediction of pregnancy outcomes. CGM metrics TIR and TAR are able to predict suboptimal pregnancy outcomes in women with type 1 diabetes.