Journal of Diabetes Science and Technology最新文献

Introduction: Automated insulin delivery (AID) systems for type 1 diabetes (T1D) improve HbA1C, increase time-in-range, and reduce hypoglycemia. However, starting AID systems involves multiple steps, from decision to initiation. This study quantified time to AID initiation (TT-AID) and factors influencing the timeline.

Methods: This retrospective study included adults with T1D at an academic diabetes center in Baltimore, Maryland who were on multiple daily injections and initiated an AID system for the first time since diagnosis from May 2022 to March 2025. Demographics and dates of AID decision, AID selection visit (optional), prescription, training, and initiation were extracted from electronic medical records. Time to AID initiation was measured, with differences by insurance and AID selection visit assessed using Wilcoxon rank-sum and log-rank tests.

Results: Participants included 114 adults with T1D [median age 38.9 years, 57% male, 21% Black, 75% commercial insurance, median diabetes duration 10.2 years (IQR = 3.5, 18.1)]. The median TT-AID was 89.5 days (IQR = 49, 132). The longest delay was between decision and training [median: 82.5 days (IQR = 43, 122)]. Patients attending the optional AID selection visit had significantly longer TT-AID compared with those who did not [112 (IQR = 79, 144) vs 55 (IQR = 35, 98) days, P ≤ .0001]. Time to AID system initiation did not differ by AID type (P = .74). Patients with commercial insurance initiated AID systems sooner than those with public insurance, [86 days (IQR = 69, 98) vs 122 (IQR = 67, 195), P = .03] within 6 months of decision.

Conclusion: Adults took roughly 3 months to initiate AID, with longer delays among those with public insurance and those attending AID selection visits. Streamlining AID system initiation may reduce delays and optimize outcomes.

Background: Glycemic control (GC) is hard to implement safely in intensive care due to patient variability. GC has been wrongly blamed for increased hypoglycemic risk instead of protocol design, limiting its adoption. Stochastic TARgeted (STAR) is a model-based, patient-specific, risk-based GC framework modulating intravenous (IV) insulin and nutrition, accounting for both inter- and intra-patient variability. This study assesses STAR GC's ability to provide safe and effective control across a large cohort.

Methods: This study was performed in Christchurch Hospital Intensive Care Unit, New Zealand. Patients were treated with STAR GC between April 2019 and December 2024. The STAR GC episodes not complying with filtering criteria were excluded. Results are analyzed in terms of performance, safety, and workload.

Results: Of 1340 adult ICU patients totaling 1958 STAR GC episodes, 1085 patients and 1430 episodes (86 010 h of control) remained after filtering. In total, 71% of blood glucose (BG) measurements were in the target band for a median [interquartile range, IQR] BG of 124 [110-148] mg/dL. Only three (0.21%) severe hypoglycemia events (BG < 40 mg/dL) occurred, two unrelated to the control design. High median [IQR] nutrition delivery (89.0 [17.2-100.0]) %goal feed was achieved with median [IQR] insulin rate of 4.5 [2.0-6.0] U/h. Results were consistent per-patient and improved once in the target band.

Conclusions: STAR provides safe, effective control for all patients in this large cohort, with minimal hypoglycemia and high nutrition rates. The protocol adapts to patients' specific needs and tolerances, encouraging STAR's adoption in other ICUs. The quality of control also enables prospective assessment of the future of GC's impact on patient outcomes.

Background: This study aimed to investigate the decline over time in the proportion of total daily insulin delivered as boluses in newly diagnosed youth with type 1 diabetes using a hybrid closed-loop system.

Method: A secondary analysis was conducted using data from the CLOuD study, an open-label, multicenter, randomized, parallel hybrid closed-loop trial to investigate bolus patterns in youth with newly diagnosed type 1 diabetes.

Results: Over the 48-month trial period, the proportion of total daily insulin delivered as carbohydrate-related boluses decreased from 58% to 34%. There was a decreasing trend in the median (interquartile range) amount of carbohydrates entered per day from 236 (204, 253) g to 184 (127, 232) g, and the number of carbohydrate-related boluses per day from 5.5 (4.6, 6.5) to 3.7 (2.9, 5.2) over the 48 months. Mean ± SD daily carbohydrate-related bolus insulin increased from 15.1 ± 6.6 to 22.0 ± 9.0 units/d, and the amount of insulin delivered per 10 g of carbohydrate more than doubled from 0.6 (0.5, 0.8) units to 1.3 (0.9, 1.5) units. The postprandial change in glucose (measured as the difference between peak glucose 30 to 180 minutes post carbohydrate-related bolus and glucose on carbohydrate-related bolus delivery) changed from 49 (45, 54) to 59 (53, 66) mg/dL.

Conclusions: The decline in the proportion of total daily insulin delivered for as bolus is likely attributable to a combination of missed boluses and under-bolusing, while the closed-loop algorithm compensates for the missed or insufficient carbohydrate-related insulin delivery by increasing basal insulin delivery.

Aims: To evaluate the outcomes of prolonged religious Jewish fasting in individuals with type 1 diabetes using automated insulin delivery (AID) systems.

Methods: This cross-sectional, non-interventional study assessed the effects of a 25-hour Jewish fast in individuals using AID systems. Data was collected on the day of the fast, one week before, and one week after.

Results: The study included data from 109 fasting days involving 80 adolescents and young adults with type 1 diabetes. The mean age of participants was 17.4 ± 4.1 years; 47.5% were male, and the average duration of diabetes was 7.2 ± 4.3 years. A total of 67.6% of participants modified their AID system settings during the fasting period, with the most common modification being a change in the target glucose level. Overall, 71.5% completed the fast without complications. Fasts were primarily broken because of sensor-detected hypoglycemia. No cases of severe hypoglycemia or diabetic ketoacidosis were reported during or after the fasting period. During the fast, the mean blood glucose level was 135 ± 28.6 mg/dL, time in range (70-180 mg/dL) was 80.7%, and time spent in hypoglycemia (<70 mg/dL) was 2.6%.

Conclusions: Prolonged fasting appears to be safe for adolescents and young adults with type 1 diabetes using AID systems. However, individualized adjustments to system settings are often necessary to maintain glycemic stability during fasting. To our knowledge, this is the first report of the effects of using an AID system during Jewish religious fasting.

Background: To identify diurnal glycemic patterns in adults with type 2 diabetes (T2D) using continuous glucose monitoring (CGM)-based machine learning and examine their association with diabetes distress, a key psychosocial outcome.

Methods: In this observational study, 137 adults with T2D wore blinded CGM (FreeStyle Libre Pro), yielding 1657 days of data. Glycemic patterns were identified using unsupervised machine learning via Gaussian mixture modeling, validated with Bayesian information criterion and silhouette scores. Diabetes distress was assessed with the 17-item Diabetes Distress Scale and analyzed through analysis of covariance (ANCOVA), adjusting for age, sex, body mass index, diabetes duration, and glucose management indicator.

Results: Clustering identified four distinct glycemic profiles: Cluster 1 (suboptimal control, nocturnal hypoglycemia; 15.8%), Cluster 2 (suboptimal control, nocturnal hyperglycemia; 27.1%), Cluster 3 (poorly controlled, prolonged hyperglycemia; 21.1%), and Cluster 4 (well controlled; 36.1%). Diabetes distress scores varied significantly: participants in Cluster 3 reported the highest distress (mean = 2.37, 95% CI = 1.99-2.76), while Cluster 4 reported the lowest (mean = 1.67, 95% CI = 1.48-1.86; P = .03). Effect sizes indicated differences corresponded to clinically meaningful categories of "little or no distress" vs "moderate distress."

Conclusions: CGM-based machine learning identified physiologically distinct glycemic phenotypes that were also associated with psychosocial burden. This work demonstrates the added value of integrating CGM-derived profiles with patient-reported outcomes. These findings highlight the potential of CGM phenotyping to support precision diabetes care by enabling early identification of high-risk subgroups, guiding tailored behavioral and psychosocial interventions, and informing technology-enabled decision tools that connect physiological monitoring with emotional well-being in T2D management.

Background: Hypoglycemia is a critical challenge for insulin-dependent people with diabetes using multiple daily injections (MDI), who rely on reactive responses to continuous glucose monitoring (CGM) alerts. To meet the need for a proactive safety tool, we evaluated the performance of the Low Glucose Predict (LGP) feature in the Accu-Chek SmartGuide Predict App.

Methods: This retrospective analysis pooled data from three prospective trials, including 85 subjects over 2709 recording days. The LGP feature uses a XGBoost model to predict low glucose events up to 30 minutes in advance. Performance was assessed rigorously against both capillary blood glucose (BG) and CGM values, including an analysis with "close-call" predictions (+10 mg/dL above the threshold). Metrics included sensitivity, specificity, and ROC-AUC.

Results: Against the stringent capillary BG reference, LGP showed high performance: sensitivity of 87.13% and specificity of 97.43% (ROC-AUC 0.9787). Including close-call events improved sensitivity to 91.89% and specificity to 98.09%. Referenced against CGM, sensitivity was 94.40% and specificity was 98.25%. The system provided an actionable mean lead time of 14.71 ± 8.30 minutes (CGM reference), with a low average daily true notification rate of 1.31 (2.60 including close-calls).

Conclusion: The LGP feature is an accurate, highly sensitive, and specific tool for timely, proactive low glucose prediction, validated against both capillary BG and CGM. This predictive intelligence is a crucial mechanism for people with diabetes to safely mitigate hypoglycemia risk, addressing a significant clinical gap and potentially reducing fear of hypoglycemia and diabetes distress.

Background: The automated assessment and prediction of diabetic foot ulcer (DFU) severity depends heavily on precise segmentation of the ulcer region. This approach avoided reliance on built-in segmentation tools, which often lacked the accuracy needed to delineate wound boundaries effectively. The objective of this study was to develop and evaluate an artificial intelligence (AI)-driven method for ulcer segmentation and severity classification of DFU using Wagner's grading system.

Methods: A novel method was introduced for segmenting the boundaries of DFUs, paired with a lightweight classification model for predicting ulcer severity as per Wagner's grade. This method was developed using a retrospective cohort of patients in India. A total of 1339 ulcer images were collected from 510 patients and augmented to 6579 images for AI-model generalizability. It incorporated an enhanced active contour model, combined with Sobel edge detection, to achieve precise delineation of ulcer edges. An AI-powered mobile application was developed to facilitate the real-time and remote assessment of the severity of DFUs.

Results: The proposed segmentation approach successfully delineated ulcer regions, achieving a Dice similarity coefficient of 0.99. The classification model attained an accuracy of 95.58%, with a sensitivity of 95.58%, a specificity of 99.16%, and an F1 score of 95.53%. The method also recorded a false-positive rate of 0.84% and a false negative rate of 4.83%, reflecting improved classification performance compared to existing methods.

Conclusions: The comparative analysis demonstrated that the proposed method significantly improved both segmentation and classification of DFUs, thereby supporting enhanced clinical management of the condition.

Sharing research code in an open access version-controlled repository offers significant benefits for both science as a whole and for individual researchers. In this article, we focus on this practice, which is fully aligned with the NIH's Gold Standard Science (GSS) program as well as FAIR (findable, accessible, interoperable, reusable) and TRUST (transparency, responsibility, user focus, sustainability, technology) principles. Gold Standard Science supports open science by emphasizing transparency, reproducibility, and the use of best practices that enable others to verify and extend research. Pairing a research article's cited data snapshot with a versioned, environment-specific code release, deposited in a companion code repository, ensures that, upon submission to a medical journal, readers and reviewers can directly verify results. An executable and updatable companion code repository complements, rather than replaces, established research data repositories. When code underlying medical research results is made openly available, then other scientists can inspect, run, and validate analyses. These activities enhance reproducibility, which is a core aim of GSS. Shared code also facilitates collaborative innovation by allowing researchers to extend the utility of the code to new datasets and applications. For researchers, code sharing can increase visibility, credibility, and citation impact. Demonstrating transparency through shared executable and updatable code builds trust with journal readers, peer reviewers, funders, and peers. Shared code in an open access repository signals adherence to high standards of scientific integrity and attracts opportunities for collaboration. A researcher who shares code receives recognition as a leader in reproducible, trustworthy research consistent with NIH's GSS principles.

Background: Despite established guidelines and increasing national hospital quality metrics, achieving consistent inpatient glycemic control remains challenging. A system-wide glucose data monitoring dashboard can help consolidate and visualize key metrics to support quality improvement (QI) and standardize care.

Methods: A web-based diabetes dashboard was implemented across 7 hospitals within a large health care system to monitor monthly data from the electronic health record. Metrics included patient-days with hypoglycemia (<70 mg/dL), hyperglycemia (mean >180 mg/dL), in-hospital mortality, hospital length of stay (LOS), and 30-day readmissions to the emergency department (ED) or inpatient/observation (IP/OBS). A total of 455 303 admissions were analyzed between January 2018 and March 2025, comparing pre-implementation (2018-2022) to post-implementation (2023-2025). Statistical analyses included t tests or Wilcoxon rank-sum tests. Given differences between the large academic site and 6 community hospitals, a difference-in-differences analysis was performed to evaluate impact by hospital type.

Results: After implementation of the dashboard, patient-days with hypoglycemia decreased from 4.81% to 4.15%, hyperglycemia from 25.30% to 23.46%, mortality from 2.69% to 2.13%, and LOS from 7.56 to 7.29 days (all P < .01). Emergency department and IP/OBS readmissions increased slightly (P < .01 and P = .01, respectively). Comparing the community hospitals to the academic, statistically significant reductions were observed in hypoglycemia, hyperglycemia, and mortality but with increased ED readmissions. There were no differences in LOS or IP/OBS readmission.

Conclusions: Implementation of a system-wide electronic dashboard was associated with improved glycemic control, mortality, and LOS. Dashboards can effectively support multidisciplinary collaboration and QI in diverse hospital settings.