Journal of Diabetes Science and Technology最新文献

Background: The growing adoption of diabetes devices has highlighted the need for integrated platforms to consolidate data from various vendors and device types, enhancing the patient experience and treatment. This shift could pave the way for a transition from conventional outpatient diabetes clinics to advanced home monitoring and virtual care methods. Overall, we wished to empower individuals with diabetes and healthcare providers to interpret and utilize information from diabetes devices more effectively.

Methods: Stenopool integrates most diabetes devices for glucose monitoring and insulin administration in our clinic. The platform was initially developed with inspiration from open-source software, and the current version is a unique digital platform for managing and analyzing diabetes device data. The development process, outcomes, and status are described.

Results: Since November 2021, Stenopool has been used in our outpatient clinic to integrate over 30 different diabetes devices from around 7000 individuals. Data are primarily uploaded via wired connections, but also using semi-automated and automated cloud-to-cloud data transfers. The platform offers a streamlined workflow for healthcare providers and displays data from various glucose meter, insulin pump, and continuous glucose monitor (CGM) vendors on a single screen in a manner that healthcare providers can modify. A data warehouse with data from Stenopool and electronical health records is nearing completion, preparing the development of tools for population health management, quality assessment, and risk stratification of patients.

Conclusion: Using Stenopool, we aimed to enhance diabetes device data management, facilitate the future for virtual patient care pathways, and improve outcomes. This article outlines the platform's development process and challenges.

Background: The study conceptualizes the application of particle filters for meal detection and shape estimation, aimed at assisting patients with type 1 diabetes (T1D) in meal announcements. It concentrates on leveraging continuous glucose monitoring (CGM) and insulin pump data to detect meals, which will eventually enhance prior research that relied on accelerometer data for event detection.

Method: The research employs glucose appearance rate (RA) curves derived from 21 triple-tracer studies. These curves are normalized and adjusted based on population-level meal size distributions obtained from the National Health and Nutrition Examination Survey (NHANES). Furthermore, glucose response curves resulting from insulin, as reported in existing literature focusing on fast-acting insulin analogs, are integrated into the analysis. This normalization of the population level probability distribution facilitates individualized scaling, taking into account insulin sensitivity and carbohydrate-to-insulin ratios.

Results: Preliminary findings from the Tidepool data set suggest that the algorithm is effective in meal detection.

Conclusions: This innovative approach holds the potential to help individuals with T1D better manage their blood glucose levels by providing information regarding glucose response to meals and estimated meal sizes for closed-loop control. Future research can focus on enhancing key components of the algorithm and incorporating additional data types to improve its performance further.

Background and aims: Glycemic control is crucial for people with type 2 diabetes. However, only about half achieve the advocated HbA1c target of ≤7%. Identifying those who will probably struggle to reach this target may be valuable as they require additional support. Thus, the aim of this study was to develop a model to predict people with type 2 diabetes not achieving HbA1c target after initiating fast-acting insulin.

Methods: Data from a randomized controlled trial (NCT01819129) of participants with type 2 diabetes initiating fast-acting insulin were used. Data included demographics, clinical laboratory values, self-monitored blood glucose (SMBG), health-related quality of life (SF-36), and body measurements. A logistic regression was developed to predict HbA1c target nonachievers. A potential of 196 features was input for a forward feature selection. To assess the performance, a 20-repeated stratified 5-fold cross-validation with area under the receiver operating characteristics curve (AUROC) was used.

Results: Out of the 467 included participants, 98 (21%) did not achieve HbA1c target of ≤7%. The forward selection identified 7 features: baseline HbA1c (%), mean postprandial SMBG at all meals 3 consecutive days before baseline (mmol/L), sex, no ketones in urine, baseline albumin (g/dL), baseline low-density lipoprotein cholesterol (mmol/L), and traces of protein in urine. The model had an AUROC of 0.745 [95% CI = 0.734, 0.756].

Conclusions: The model was able to predict those who did not achieve HbA1c target with promising performance, potentially enabling early identification of people with type 2 diabetes who require additional support to reach glycemic control.

Background: Despite remarkable progress in diabetes technology, most systems still require estimating meal carbohydrate (CHO) content for meal-time insulin delivery. Emerging smartphone applications may obviate this need, but performance data in relation to patient estimates remain scarce.

Objective: The objective is to assess the accuracy of two commercial CHO estimation applications, SNAQ and Calorie Mama, and compare their performance with the estimation accuracy of people with type 1 diabetes (T1D).

Methods: Carbohydrate estimates of 53 individuals with T1D (aged ≥16 years) were compared with those of SNAQ (food recognition + quantification) and Calorie Mama (food recognition + adjustable standard portion size). Twenty-six cooked meals were prepared at the hospital kitchen. Each participant estimated the CHO content of two meals in three different sizes without assistance. Participants then used SNAQ for CHO quantification in one meal and Calorie Mama for the other (all three sizes). Accuracy was the estimate's deviation from ground-truth CHO content (weight multiplied by nutritional facts from recipe database). Furthermore, the applications were rated using the Mars-G questionnaire.

Results: Participants' mean ± standard deviation (SD) absolute error was 21 ± 21.5 g (71 ± 72.7%). Calorie Mama had a mean absolute error of 24 ± 36.5 g (81.2 ± 123.4%). With a mean absolute error of 13.1 ± 11.3 g (44.3 ± 38.2%), SNAQ outperformed the estimation accuracy of patients and Calorie Mama (both P > .05). Error consistency (quantified by the within-participant SD) did not significantly differ between the methods.

Conclusions: SNAQ may provide effective CHO estimation support for people with T1D, particularly those with large or inconsistent CHO estimation errors. Its impact on glucose control remains to be evaluated.

Background: Self-monitoring of glucose is important to the successful management of diabetes; however, existing monitoring methods require a degree of invasive measurement which can be unpleasant for users. This study investigates the accuracy of a noninvasive glucose monitoring system that analyses spectral variations in microwave signals.

Methods: An open-label, pilot design study was conducted with four cohorts (N = 5/cohort). In each session, a dial-resonating sensor (DRS) attached to the wrist automatically collected data every 60 seconds, with a novel artificial intelligence (AI) model converting signal resonance output to a glucose prediction. Plasma glucose was measured in venous blood samples every 5 minutes for Cohorts 1 to 3 and every 10 minutes for Cohort 4. Accuracy was evaluated by calculating the mean absolute relative difference (MARD) between the DRS and plasma glucose values.

Results: Accurate plasma glucose predictions were obtained across all four cohorts using a random sampling procedure applied to the full four-cohort data set, with an average MARD of 10.3%. A statistical analysis demonstrates the quality of these predictions, with a surveillance error grid (SEG) plot indicating no data pairs falling into the high-risk zones.

Conclusions: These findings show that MARD values approaching accuracies comparable to current commercial alternatives can be obtained from a multiparticipant pilot study with the application of AI. Microwave biosensors and AI models show promise for improving the accuracy and convenience of glucose monitoring systems for people with diabetes.

Background: The glycemic characterization of congenital hyperinsulinism (HI), a rare disease causing severe hypoglycemia in childhood, is incomplete. Continuous glucose monitoring (CGM) offers deep glycemic phenotyping to understand disease burden and individualize patient care. Typically, CGM has been restricted to severe HI only, with performance being described in short-term, retrospective studies. We have described CGM-derived phenotyping in a prospective, unselected national cohort providing comprehensive baseline information for future therapeutic trials.

Methods: Glycemic frequency and trends, point accuracy, and patient experiences were drawn from a prospective, nationwide, observational study of unselected patients with persistent HI using the Dexcom G6 CGM device for 12 months as an additional monitoring tool alongside standard of care self- monitoring blood glucose (SMBG).

Findings: Among 45 patients with HI, mean age was six years and 53% carried a genetic diagnosis. Data confirmed higher risk of early morning (03:00-07:00 h) hypoglycemia throughout the study period and demonstrated no longitudinal reduction in hypoglycemia with CGM use. Device accuracy was suboptimal; 17 500 glucose levels paired with SMBG demonstrated mean absolute relative difference (MARD) 25% and hypoglycemia detection of 40%. Patient/parent dissatisfaction with CGM was high; 50% of patients discontinued use, citing inaccuracy and pain. However, qualitative feedback was also positive and families reported improved understanding of glycemic patterns to inform changes in behavior to reduce hypoglycemia.

Interpretation: This comprehensive study provides unbiased insights into glycemic frequency and long-term trends among patients with HI; such data are likely to influence and inform clinical priorities and future therapeutic trials.

Background: We investigated the risk of incident diabetic retinopathy (DR) among high glycator compared to low glycator patients based on the hemoglobin glycation index (HGI). Visit-to-visit variations in HGI also were assessed.

Methods: Glycated hemoglobin (HbA1c) and continuous glucose monitoring data were collected up to 7 years prior to the date of eye examination defining incident DR or no retinopathy (control). Hemoglobin glycation index was calculated as difference in measured HbA1c and an estimated A1c from sensor glucose (eA1c) to define high (HbA1c - eA1c >0%) or low (HbA1c - eA1c <0%) glycator. Stable glycators were defined as ≥75% of visits with same HGI category. Logistic regression was used to assess the association between glycation category and incident DR.

Results: Of 119 adults with type 1 diabetes (T1D), 49 (41%) were stable low glycator (HbA1c - eA1c <0%), 36 (30%) were stable high glycator (HbA1c - eA1c >0%), and 34 (29%) were unstable glycator. Using alternate criteria to define high vs low glycator (consistent difference in HbA1c - eA1c of > 0.4% or <0.4%, respectively), 53% of the adults were characterized as unstable glycator. Compared to low glycators, high glycators did not have a significantly higher risk for incident DR over time when adjusted for age, T1D duration and continuous glucose monitoring (CGM) sensor type (odds ratio [OR] = 1.31, 95% confidence interval [CI] = 0.48-3.62, P = .15).

Conclusions: The risk of diabetic retinopathy was not found to differ significantly comparing high glycators to low glycators in adults with T1D. Moreover, HbA1c - eA1c relationship was not stable in nearly 30% to 50% adults with T1D, suggesting that discordance in HbA1c and eA1c are mostly related either HbA1c measurements or estimation of A1c from sensor glucose rather than physiological reasons.

Background: Sodium glucose cotransporter 2 inhibitors (SGLT2i) prevent heart failure (HF) in patients with type 2 diabetes mellitus (T2DM) but prescription rates are low. The effect of an electronic health record (EHR) alert notifying providers of patients' estimated risk of developing HF on SGTL2i prescriptions is unknown.

Methods: This was a pragmatic, randomized clinical trial that compared an EHR alert and usual care among patients with T2DM and no history of HF or SGLT2i use at a single center. The EHR alert notified providers of their patient's HF risk and recommended HF prevention strategies. Randomization was performed at the provider level across general and subspecialty internal medicine as well as family medicine outpatient clinics. The primary outcome was proportion of SGLT2i prescriptions within 30 days. Proportion of natriuretic peptide (NP) tests within 90 days was also assessed.

Results: A total of 1524 patients (median age 75 years, 45% women, 23% Black) were enrolled between September 28, 2021, and April 29, 2022 from 189 outpatient clinics. SGLT2i were prescribed to 1.2% (9/780) of patients in the EHR alert group and 0% (0/744) of those in the usual care group (P value = 0.009). Natriuretic peptide testing was performed within 90 days among 10.8% (84/780) of patients in the EHR alert group and 7.3% (54/744) of patients in the usual care group (P value = 0.02).

Conclusions: In a single-center trial with low overall SGLT2i use, an EHR alert incorporating HF risk information significantly increased SGLT2i prescriptions and NP testing although the absolute rates were low.

Background: Automated insulin delivery (AID) has revolutionized glucose management. Next-generation AID systems focus on reducing user input, particularly for mealtime dosing, aiming for fully closed loop (FCL) control. Our goal was to assess the safety and feasibility of the next iteration of FCL control, using a miniature neural network to enable implementation within existing hardware capabilities.

Methods: In a randomized crossover trial, six adults with type 1 diabetes completed seven days of usual care and seven days using AIDANET in free-living conditions. AIDANET is designed to enable FCL control, but carbohydrate counting and a novel easy-bolus strategy were enabled for one day each to test the system in hybrid closed loop modalities.

Results: The mean glucose during usual care was 168 ± 24.3 mg/dL, compared to 161.3 ± 16.7 mg/dL using the AIDANET system. Time-in-range (TIR) 70 to 180 mg/dL was 63.3% ± 14.9% in usual care compared to 66.4% ± 8.3% using AIDANET, while time-below-range (TBR) 70 mg/dL remained within acceptable margins (0.9 ± 1 vs 1.6 ± 1.8). There were no serious adverse events during the study. The hybrid bolusing options provided safe glycemic control, with carbohydrate counting achieving 57.1% TIR with 0.6% TBR, and Easy Bolus achieving 70.5% TIR with 1.5% TBR.

Conclusion: This pilot-feasibility study demonstrates that the AIDANET system provides safe glycemic control. The small sample size (n = 6) limits overall generalizability, and further larger, statistically powered trials to validate these results are warranted.