Less Nocturnal Hypoglycemia but Equivalent Time in Range Among Adults with Type 1 Diabetes Using Insulin Pumps Versus Multiple Daily Injections

Machine Learning and Deep Learning Models for Nocturnal High- and Low-Glucose Prediction in Adults with Type 1 Diabetes

Glucose management at night is a major challenge for people with type 1 diabetes (T1D), especially for those managed with multiple daily injections (MDIs). In this study, we developed machine learning (ML) and deep learning (DL) models to predict nocturnal glucose within the target range (3.9-10 mmol/L), above the target range, and below the target range in subjects with T1D managed with MDIs. The models were trained and tested on continuous glucose monitoring data obtained from 380 subjects with T1D. Two DL algorithms-multi-layer perceptron (MLP) and a convolutional neural network (CNN)-as well as two classic ML algorithms, random forest (RF) and gradient boosting trees (GBTs), were applied. The resulting models based on the DL and ML algorithms demonstrated high and similar accuracy in predicting target glucose (F1 metric: 96-98%) and above-target glucose (F1: 93-97%) within a 30 min prediction horizon. Model performance was poorer when predicting low glucose (F1: 80-86%). MLP provided the highest accuracy in low-glucose prediction. The results indicate that both DL (MLP, CNN) and ML (RF, GBTs) algorithms operating CGM data can be used for the simultaneous prediction of nocturnal glucose values within the target, above-target, and below-target ranges in people with T1D managed with MDIs.

Lifestyle in Emerging Adults with Type 1 Diabetes Mellitus: A Qualitative Systematic Review

Emerging adulthood is a transitional stage with significant lifestyle changes, making it especially challenging for those living with type 1 diabetes mellitus. This systematic review synthesizes qualitative research to explore how emerging adulthood (18-29 years) influences lifestyle behaviors in individuals with type 1 diabetes mellitus. CINAHL, Cochrane Library, Global Health, Nursing & Allied Health Premium, PsycINFO, PubMed, Scopus, and WOS were searched for original qualitative studies addressing the lifestyle of 18-31-year-olds with type 1 diabetes mellitus, published between January 2010 and March 2021 following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Thirty-five studies met the inclusion criteria and their findings were categorized into eight topics (emotions and feelings, nutrition, perceptions, risky behaviors, self-care, sleep, social relationships, and stigma) using meta-aggregation, as outlined in the Joanna Briggs Institute Manual for Evidence Synthesis. The spontaneity characteristic of emerging adulthood can undermine self-care. This is because new environments, schedules, and relationships encountered during this life stage often lead to the neglect of diabetes management, owing to the various social, academic, and occupational demands. This review highlights the necessity of creating health promotion strategies tailored to the unique lifestyle aspects of emerging adults with type 1 diabetes mellitus.

Combining uncertainty-aware predictive modeling and a bedtime Smart Snack intervention to prevent nocturnal hypoglycemia in people with type 1 diabetes on multiple daily injections

OBJECTIVE

Nocturnal hypoglycemia is a known challenge for people with type 1 diabetes, especially for physically active individuals or those on multiple daily injections. We developed an evidential neural network (ENN) to predict at bedtime the probability and timing of nocturnal hypoglycemia (0-4 vs 4-8 h after bedtime) based on several glucose metrics and physical activity patterns. We utilized these predictions in silico to prescribe bedtime carbohydrates with a Smart Snack intervention specific to the predicted minimum nocturnal glucose and timing of nocturnal hypoglycemia.

MATERIALS AND METHODS

We leveraged free-living datasets collected from 366 individuals from the T1DEXI Study and Glooko. Inputs to the ENN used to model nocturnal hypoglycemia were derived from demographic information, continuous glucose monitoring, and physical activity data. We assessed the accuracy of the ENN using area under the receiver operating curve, and the clinical impact of the Smart Snack intervention through simulations.

RESULTS

The ENN achieved an area under the receiver operating curve of 0.80 and 0.71 to predict nocturnal hypoglycemic events during 0-4 and 4-8 h after bedtime, respectively, outperforming all evaluated baseline methods. Use of the Smart Snack intervention reduced probability of nocturnal hypoglycemia from 23.9 ± 14.1% to 14.0 ± 13.3% and duration from 7.4 ± 7.0% to 2.4 ± 3.3% in silico.

DISCUSSION

Our findings indicate that the ENN-based Smart Snack intervention has the potential to significantly reduce the frequency and duration of nocturnal hypoglycemic events.

CONCLUSION

A decision support system that combines prediction of minimum nocturnal glucose and proactive recommendations for bedtime carbohydrate intake might effectively prevent nocturnal hypoglycemia and reduce the burden of glycemic self-management.

Effect of diabetes technologies on the fear of hypoglycaemia among people living with type 1 diabetes: a systematic review and meta-analysis

Background

Fear of hypoglycaemia (FOH) significantly disrupts the daily management of type 1 diabetes (T1D) and increases the risk of complications. Recent technological advances can improve glucose metrics and reduce hypoglycaemia frequency, yet their impact on FOH is unclear. This systematic review and meta-analysis (SRMA) aimed to synthesize the current literature to understand the impact of diabetes technologies on FOH in T1D.

Methods

In this SRMA, we searched PubMed, Medline, Scopus, and Web of Science from inception up to May 21st, 2023 for studies assessing the effect of using real-time or intermittently scanned continuous glucose monitors (rtCGM or isCGM); insulin pumps (CSII); and their combinations on FOH as the primary outcome, measured using the Hypoglycaemia Fear Survey (HFS; including total, worries [HFS-W], and behaviours [HFS-B] scores), in non-pregnant adults with T1D. Data was extracted by the first and second authors. Results were pooled using a random-effects model based on study design (RCT and non-RCT), with subgroup analysis based on the type of technology, reported change in hypoglycaemia frequency, and duration of use. Risk of bias was evaluated with Cochrane and Joanna Briggs Institute tools. This study is registered with PROSPERO, CRD42021253618.

Findings

A total of 51 studies (n = 8966) were included, 22 of which were RCTs. Studies on rtCGM and CSII reported lower FOH levels with ≥8 weeks of use. Studies on CSII and rtCGM combinations reported lower FOH levels after ≥13 weeks of automated insulin delivery (AID) use or 26 weeks of sensor-augmented pump (SAP) use. The meta-analysis showed an overall lower FOH with technologies, specifically for the HFS-W subscale. The RCT meta-analysis showed lower HFS-W scores with rtCGM use (standard mean difference [95%CI]: -0.14 [-0.23, -0.05], I2 = 0%) and AID (-0.17 [-0.33, -0.01], I2 = 0%). Results from non-RCT studies show that SAP users (-0.33 [-0.38, -0.27], I2 = 0%) and rtCGM users (-0.38 [-0.61, -0.14], I2 = 0%) had lower HFS-W.

Interpretation

We found consistent, yet small to moderate, effects supporting that diabetes technologies (specifically rtCGM, SAP, and AID) may reduce hypoglycaemia-related worries in adults with T1D. Current literature, however, has limitations including discrepancies in baseline characteristics and limited, mainly descriptive, statistical analysis. Thus, future studies should assess FOH as a primary outcome, use validated surveys, and appropriate statistical analysis to evaluate the clinical impacts of technology use beyond just glucose metrics.

Funding

Canadian Institutes of Health Research, Juvenile Diabetes Research Foundation Ltd.

Are nocturnal hypoglycemia prevention strategies influenced by diabetes technology usage? A BETTER registry analysis

AIM

To assess the association of nocturnal hypoglycemia prevention strategies (NH-PS) and diabetes technology usage (insulin pump and/or continuous glucose monitors [CGM]) in people with type 1 diabetes (PWT1D).

METHODS

Logistic regression models were used to describe associations between self-reported NH-PS and diabetes technology (pump with intermittently-scanned or real-time CGM (isCGM or rtCGM), or automated insulin delivery (AID)), hypoglycemia history, and fear of hypoglycemia (FOH).

RESULTS

Among 831 adults (65 % female, aged 44 ± 15 years, T1D duration 26 ± 15 years), 32 % reported HbA1c ≤ 7.0 %, 88 % used ≥ 1 diabetes technology, 66 % reported ≥ 1 symptomatic NH in the past month, and 64 % used ≥ 2 NH-PS. Compared to multiple daily injections (MDI) + capillary blood glucose (CBG), bedtime snack consumption was less likely among pump + isCGM (OR [95 %CI]: 0.55 [0.31, 0.98]), pump + rtCGM (0.40 [0.20, 0.81]), and AID (0.34 [0.17, 0.66]) users, while evening insulin basal reduction was associated with CSII + CBG (3.15 [1.25, 7.99]), pump + isCGM 4.00 [1.99, 8.01]), and pump + rtCGM 2.89 [1.28, 6.50] use. Elevated FOH was associated with snack consumption (1.37 [1.00, 1.89]), evening bolus insulin avoidance (1.77 [1.11, 2.83]), limiting exercise (2.50 [1.30, 4.82]), and limiting alcohol consumption (2.33 [1.15, 4.70]) as NH-PS.

CONCLUSION

Technology use and elevated FOH might influence PWT1D' choice of NH-PS.

Cell delivery systems: Toward the next generation of cell therapies for type 1 diabetes

Immunoprotection and oxygen supply are vital in implementing a cell therapy for type 1 diabetes (T1D). Without these features, the transplanted islet cell clusters will be rejected by the host immune system, and necrosis will occur due to hypoxia. The use of anti-rejection drugs can help protect the transplanted cells from the immune system; yet, they also may have severe side effects. Cell delivery systems (CDS) have been developed for islet transplantation to avoid using immunosuppressants. CDS provide physical barriers to reduce the immune response and chemical coatings to reduce host fibrotic reaction. In some CDS, there is architecture to support vascularization, which enhances oxygen exchange. In this review, we discuss the current clinical and preclinical studies using CDS without immunosuppression as a cell therapy for T1D. We find that though CDS have been demonstrated for their ability to support immunoisolation of the grafted cells, their functionality has not been fully optimized. Current advanced methods in clinical trials demonstrate the systems are partly functional, physically complicated to implement or inefficient. However, modifications are being made to overcome these issues.