Efficacy and Safety of Different Hybrid Closed Loop Systems for Automated Insulin Delivery in People With Type 1 Diabetes: A Systematic Review and Network Meta-Analysis.

AIMS: To compare the efficacy and safety of different hybrid closed loop (HCL) systems in people with diabetes through a network meta-analysis. METHODS: We searched MEDLINE, EMBASE, CENTRAL and PubMed for randomised clinical trials (RCTs) enrolling children, adolescents and/or adults with type 1 or type 2 diabetes, evaluating Minimed 670G, Minimed 780G, Control-IQ, CamAPS Fx, DBLG-1, DBLHU, and Omnipod 5 HCL systems against other types of insulin therapy, and reporting time in target range (TIR) as outcome. RESULTS: A total of 28 RCTs, all enrolling people with type 1 diabetes, were included. HCL systems significantly increased TIR compared with subcutaneous insulin therapy without continuous glucose monitoring (SIT). Minimed 780G achieved the highest TIR ahead of Control IQ (mean difference (MD) 5.1%, 95% confidence interval (95% CI) [0.68; 9.52], low certainty), Minimed 670G (MD 7.48%, 95% CI [4.27; 10.7], moderate certainty), CamAPS Fx (MD 8.94%, 95% CI [4.35; 13.54], low certainty), and DBLG1 (MD 10.69%, 95% CI [5.73; 15.65], low certainty). All HCL systems decreased time below target range, with DBLG1 (MD -3.69%, 95% CI [-5.2; -2.19], high certainty), Minimed 670G (MD -2.9%, 95% CI [-3.77; -2.04], moderate certainty) and Minimed 780G (MD -2.79%, 95% CI [-3.94; -1.64], high certainty) exhibiting the largest reductions compared to SIT. The risk of severe hypoglycaemia and diabetic ketoacidosis was similar to other types of insulin therapy. CONCLUSIONS: We show a hierarchy of efficacy among the different HCL systems in people with type 1 diabetes, thus providing support to clinical decision-making. TRIAL REGISTRATION: PROSPERO CRD42023453717.

Mealtime prediction using wearable insulin pump data to support diabetes management.

Many patients with diabetes struggle with post-meal high blood glucose due to missed or untimely meal-related insulin doses. To address this challenge, our research aims to: (1) study mealtime patterns in patients with type 1 diabetes using wearable insulin pump data, and (2) develop personalized models for predicting future mealtimes to support timely insulin dose administration. Using two independent datasets with over 45,000 meal logs from 82 patients with diabetes, we find that the majority of people ( ∼ 60%) have irregular and inconsistent mealtime patterns that change notably through the course of each day and across months in their own historical data. We also show the feasibility of predicting future mealtimes with personalized LSTM-based models that achieve an average F1 score of > 95% with less than 0.25 false positives per day. Our research lays the groundwork for developing a meal prediction system that can nudge patients with diabetes to administer bolus insulin doses before meal consumption to reduce the occurrence of post-meal high blood glucose.

Clinical evidence for high-risk CE-marked medical devices for glucose management: A systematic review and meta-analysis.

AIMS: To conduct a systematic review and meta-analysis, within the Coordinating Research and Evidence for Medical Devices (CORE-MD) project, evaluating CE-marked high-risk devices for glucose management. MATERIALS AND METHODS: We identified interventional and observational studies evaluating the efficacy and safety of eight automated insulin delivery (AID) systems, two implantable insulin pumps, and three implantable continuous glucose monitoring (CGM) devices. We meta-analysed randomized controlled trials (RCTs) comparing AID systems with other treatments. RESULTS: A total of 182 studies published between 2009 and 2024 were included, comprising 166 studies on AID systems, six on insulin pumps, and 10 on CGM devices; 26% reported industry funding; 18% were pre-market; 37% had a comparator group. Of the studies identified, 29% were RCTs, 24% were non-randomized trials, and 47% were observational studies. The median (interquartile range) sample size was 48 (28-102), age 34.8 (14-44.2) years, and study duration 17.5 (12-26) weeks. AID systems lowered glycated haemoglobin by 0.5 percentage points (absolute mean difference [MD] = -0.5; 21 RCTs; I2 = 86%) and increased time in target range for sensor glucose level by 13.4 percentage points (MD = 13.4; 14 RCTs; I2 = 90%). At least one safety outcome was assessed in 71% of studies. CONCLUSIONS: High-risk devices for glucose monitoring or insulin dosing, in particular AID systems, improve glucose control safely, but evidence on diabetes-related end-organ damage is lacking due to short study durations. Methodological heterogeneity highlights the need for developing standards for future pre- and post-market investigations of diabetes-specific high-risk medical devices.

Effectiveness of a hybrid closed-loop system for children and adolescents with type 1 diabetes during physical exercise: A cross-sectional study in real life.

OBJECTIVE: The aim of the study was to describe how physical exercise affects metabolic control, insulin requirements and carbohydrate intake in children who use hybrid closed-loop systems. METHODS: Cross-sectional study design. The sample included 21 children and adolescents diagnosed with type 1 diabetes. During the study, participants were monitored for a period of 7 days to gather comprehensive data on these factors. RESULTS: Nine participants (42.9%) had switched to exercise mode to raise the target glucose temporarily to 150 mg/dL. The HbA1c values ranged from 5.5% to 7.9% (median, 6.5%; IQR, 0.75). The percentage of time within the target range of 70-180 mg/dL was similar; however, there was an increased duration of hyperglycaemia and more autocorrections on exercise days. The time spent in severe hyperglycaemia (>250 mg/dL) increased by 2.7% in exercise compared to non-exercise days (P = .02). It is worth noting that hypoglycaemic episodes did not increase during the exercise days compared with non-exercise days. CONCLUSION: The hybrid closed-loop system was effective and safe in children and adolescents with type 1 diabetes during the performance of competitive sports in real life.

Use of a decision support tool and quick start onboarding tool in individuals with type 1 diabetes using advanced automated insulin delivery: a single-arm multi-phase intervention study.

BACKGROUND: Multiple clinician adjustable parameters impact upon glycemia in people with type 1 diabetes (T1D) using Medtronic Mini Med 780G (MM780G) AHCL. These include glucose targets, carbohydrate ratios (CR), and active insulin time (AIT). Algorithm-based decision support advising upon potential settings adjustments may enhance clinical decision-making. METHODS: Single-arm, two-phase exploratory study developing decision support to commence and sustain AHCL. Participants commenced investigational MM780G, then 8 weeks Phase 1-initial optimization tool evaluation, involving algorithm-based decision support with weekly AIT and CR recommendations. Clinicians approved or rejected CR and AIT recommendations based on perceived safety per protocol. Co-design resulted in a refined algorithm evaluated in a further identically configured Phase 2. Phase 2 participants also transitioned to commercial MM780G following "Quick Start" (algorithm-derived tool determining initial AHCL settings using daily insulin dose and weight). We assessed efficacy, safety, and acceptability of decision support using glycemic metrics, and the proportion of accepted CR and AIT settings per phase. RESULTS: Fifty three participants commenced Phase 1 (mean age 24.4; Hba1c 61.5mmol/7.7%). The proportion of CR and AIT accepted by clinicians increased between Phases 1 and 2 respectively: CR 89.2% vs. 98.6%, p < 0.01; AIT 95.2% vs. 99.3%, p < 0.01. Between Phases, mean glucose percentage time < 3.9mmol (< 70mg/dl) reduced (2.1% vs. 1.4%, p = 0.04); change in mean TIR 3.9-10mmol/L (70-180mg/dl) was not statistically significant: 72.9% ± 7.8 and 73.5% ± 8.6. Quick start resulted in stable TIR, and glycemic metrics compared to international guidelines. CONCLUSION: The co-designed decision support tools were able to deliver safe and effective therapy. They can potentially reduce the burden of diabetes management related decision making for both health care practitioners and patients. TRIAL REGISTRATION: Prospectively registered with Australia/New Zealand Clinical Trials Registry(ANZCTR) on 30th March 2021 as study ACTRN12621000360819.

[Position paper: Open-source technology in the treatment of people living with diabetes mellitus-an Austrian perspective : Technology Committee of the Austrian Diabetes Association]. / Positionspapier: Open-source-Technologie in der Behandlung von Menschen mit Diabetes mellitus ­ eine österreichische Perspektive : Technologieausschuss der Österreichischen Diabetes Gesellschaft.

People living with diabetes mellitus can be supported in the daily management by diabetes technology with automated insulin delivery (AID) systems to reduce the risk of hypoglycemia and improve glycemic control as well as the quality of life. Due to barriers in the availability of AID-systems, the use and development of open-source AID-systems have internationally increased. This technology provides a necessary alternative to commercially available products, especially when approved systems are inaccessible or insufficiently adapted to the specific needs of the users. Open-source technology is characterized by worldwide free availability of codes on the internet, is not officially approved and therefore the use is on the individual's own responsibility. In the clinical practice a lack of expertise with open-source AID technology and concerns about legal consequences, lead to conflict situations for health-care professionals (HCP), sometimes resulting in the refusal of care of people living with diabetes mellitus. This position paper provides an overview of the available evidence and practical guidance for HCP to minimize uncertainties and barriers. People living with diabetes mellitus must continue to be supported in education and diabetes management, independent of the chosen diabetes technology including open-source technology. Check-ups of the metabolic control, acute and chronic complications and screening for diabetes-related diseases are necessary and should be regularly carried out, regardless of the chosen AID-system and by a multidisciplinary team with appropriate expertise.

Excellence in the management of Advanced Hybrid Closed-Loop Systems: Lessons from the Polish cohort.

BACKGROUND: The aim of the study was to analyze the real-world performance of MiniMed 780G (MM780G) Advanced Hybrid Closed Loop (AHCL) system users from Poland (PL) and compare it to the European region excluding Poland (EU-PL) in order to identify factors contributing to potential differences. The former achieved some of the best Time in Range (TIR) results globally using this technology. METHODS: CareLink Personal data uploaded by MM780G system users from August 2020 to December 2022 were analyzed. RESULTS: The Polish users (N=1304) on average reached to TIR of 79.1 ± 8.7 % (vs 73.0 ± 10.0 % for EU-PL, N=55659), a TBR<54 mg/dL of 0.6 ± 0.7 % (vs 0.4 ± 0.6 %) and a TBR<70 mg/dL of 2.9 ± 2.1 % (vs 2.1 ± 1.8 %). The adoption rate of optimal settings (i.e, GT=100 mg/dL, AIT=2hr) in PL was high (19.7 % vs 6.3 %), and filtering on optimal setting users led to less pronounced differences in glycemic control between PL and EU-PL. A univariable analysis with post-AHCL TIR showed that geography itself (PL vs EU-PL) is not a significant contributor to a high post-AHCL TIR (p = 0.15), and that much of the Polish post-AHCL TIR can be explained by the high pre-AHCL TIR. CONCLUSION: The Polish MM780G users achieved better glycemic control than the general European population (excluding Poland). This is largely attributable to the adoption of optimal settings in Poland and the already high glycemic outcomes at system start. As these characteristics can be implemented elsewhere, we believe this outstanding result can be obtained in other countries as well.

Unmet needs in the treatment of type 1 diabetes: why is it so difficult to achieve an improvement in metabolic control?

The development of advanced diabetes technology has permitted persons with type 1 diabetes mellitus to improve metabolic control significantly, particularly with the development of advanced hybrid closed-loop systems which have improved the quality of life by reducing hypoglycemia, decreasing macroangiopathy and microangiopathy-related complications, ameliorating HbA1c and improving glycemic variability. Despite the progression made over the past few decades, there is still significant margin for improvement to be made in terms of attaining appropriate metabolic control. Various factors are responsible for poor glycemic control including inappropriate carbohydrate counting, repeated bouts of hypoglycemia, hypoglycemia unawareness, cutaneous manifestations due to localized insulin use and prolonged use of diabetes technology, psychosocial comorbidities such as eating disorders or 'diabulimia', the coexistence of insulin resistance among people with type 1 diabetes and the inability to mirror physiological endogenous pancreatic insulin secretion appropriately. Hence, the aim of this review is to highlight and overcome the barriers in attaining appropriate metabolic control among people with type 1 diabetes by driving research into adjunctive treatment for coexistent insulin resistance and developing new advanced diabetic technologies to preserve ß cell function and mirror as much as possible endogenous pancreatic functions.

Patients with type 2 diabetes who achieve reduced postprandial glucose levels during insulin intensive therapy may have a better recovery of ß-cell function.

OBJECTIVES: To explore parameters that may determine the improvement in C-peptide levels in patients with type 2 diabetes (T2D) receiving continuous subcutaneous insulin infusion (CSII) therapy. METHODS: The trial included a lead-in period for collecting baseline parameters and correcting hyperglycemia, a 4-day CGM period, and a 2-3 weeks treatment period. After screening, patients were hospitalized and randomized to the metformin add-on NovoRapid group or the Prandilin group. Once the glycemic target was reached, all patients underwent a 4-day CGM, with treatments maintained for 2-3 weeks. OGTTs were performed at baseline and endpoint. The primary endpoint was identifying factors contributing to better ß-cell function recovery after CSII therapy. RESULTS: A total of 99 recruited patients were admitted as inpatients and achieved glycemic control within 3.8 ± 1.1 days. Of these, 83 (84 %) patients showed improvement in C-peptide levels, while 16 (16 %) did not show any change in C-peptide levels at the endpoint. Pearson analysis showed a negative correlation between the incremental AUC of glucose concentration (from 0700 to 1000) and the increase in incremental AUC of C-peptide levels (r = -0.199, P < 0.05). CONCLUSIONS: Drug-naïve T2D patients with lower postprandial glucose concentration during CSII therapy exhibit better ß-cell function recovery.

Association of continuous subcutaneous insulin therapy and diabetic retinopathy in type 1 diabetes: A national cohort study.

AIM: This study aimed to investigate the short-and long-term effect on diabetic retinopathy (DR) in individuals with type 1 diabetes treated with continuous subcutaneous insulin injections (CSII) compared to those using multiple daily injections (MDI). METHODS: We conducted a register-based matched cohort study utilizing data from the Danish Registry of Diabetic Retinopathy as well as several other national Danish health registers. Our cohort consisted of all individuals with type 1 diabetes who attended the Danish screening program for DR from 2013 to 2022. We included individuals registered with CSII treatment, and compared them to individuals using MDI, matched by age, sex, and DR level. Cox regression analysis was performed to evaluate the outcomes. RESULTS: The study included 674 individuals treated with CSII and 2006 matched MDI users. In our cohort 53.4 % were female and median age was 36 (IQR 27-47). Average follow-up risk-time was 4.8 years. There was no difference in the risk of DR worsening between the CSII group and MDI group (HR 1.05 [95%CI 0.91; 1.22], p = 0.49). However, an increased risk of focal photocoagulation was observed in the CSII group (HR 2.40 [95%CI 1.11; 5.19], p = 0.03). CONCLUSIONS: Our findings indicate that CSII treatment does not confer a significant difference in the overall short- and long-term risk of DR worsening or ocular intervention compared to MDI treatment. These results provide insights into the DR outcomes of CSII treatment in individuals with type 1 diabetes.

Single-arm, first-in-human feasibility study results for an ultra-low-cost insulin pump.

BACKGROUND: Use of Continuous Subcutaneous Insulin Infusion (CSII) has been shown to improve glycemic outcomes in Type 1 Diabetes (T1D), but high costs limit accessibility. To address this issue, an inter-operable, open-source Ultra-Low-Cost Insulin Pump (ULCIP) was developed and previously shown to demonstrate comparable delivery accuracy to commercial models in standardised laboratory tests. This study aims to evaluate the updated ULCIP in-vivo, assessing its viability as an affordable alternative for those who cannot afford commercially available devices. METHODS: This first-in-human feasibility study recruited six participants with T1D. During a nine-hour inpatient stay, participants used the ULCIP under clinical supervision. Venous glucose, insulin, and ß-Hydroxybutyrate were monitored to assess device performance. RESULTS: Participants displayed expected blood glucose and blood insulin levels in response to programmed basal and bolus insulin dosing. One participant developed mild ketosis, which was treated and did not recur when a new pump reservoir was placed. All other participants maintained ß-Hydroxybutyrate < 0.6 mmol/L throughout. CONCLUSION: The ULCIP safely delivered insulin therapy to users in a supervised inpatient environment. Future work should focus on correcting a pump hardware issue identified in this trial and extending device capabilities for use in closed loop control. Longer-term outpatient studies are warranted. TRIAL REGISTRATION: The trial was prospectively registered with the Australian New Zealand Clinical Trials Registry (ACTRN12623001288617) on the 11 December 2023.

A systematic review and meta-analysis comparing outcomes between using subcutaneous insulin and continuous insulin infusion in managing adult patients with diabetic ketoacidosis.

BACKGROUND: The purpose of this systematic review and meta-analysis was to synthesize the current literature to determine the safety and efficacy of using subcutaneous insulin compared to an intravenous (IV) insulin infusion in managing diabetic ketoacidosis (DKA). METHODS: We searched Ovid-Medline, EMBASE, SCOPUS, BIOSIS and CENTRAL from inception to April 26, 2024. Randomized controlled trials (RCTs) and observational studies that assessed the use of subcutaneous compared to intravenous insulin for the treatment of mild to moderate DKA were included. Data extraction and quality assessment were performed by two independent reviewers and disagreements were resolved through further discussion or by a third reviewer. The Cochrane Risk of Bias tool version 2.0 was used to evaluate the RCTs and the Risk of Bias in Non-randomized Studies of Interventions (ROBINS)-I tool was used to evaluate the observational studies. The quality of evidence was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) criteria. Meta-analyses were conducted using random-effects models. We followed the PRISMA guidelines for reporting our findings. RESULTS: Six RCTs (245 participants) and four observational studies (8444 patients) met our inclusion criteria. Some studies showed a decreased length of stay (Mean Difference [MD] in days: -0.39; 95% CI: -2.83 to 2.08; I2: 0%) among individuals treated with subcutaneous insulin compared to intravenous insulin. There was no difference in the risk of all-cause mortality, time to resolution of DKA (MD in hours: 0.17; 95% confidence interval [CI]: -3.45 to 3.79; I2: 0%) and hypoglycemia (Risk Ratio [RR]: 1.02; 95% CI: 0.88 to 1.19; I2: 0%) between the two groups. CONCLUSION: Treatment of DKA with subcutaneous insulin may be a safe and effective alternative to IV insulin in selected patients. The limited available evidence underscores the need for further studies to explore optimal dosing, patient selection criteria and long-term outcomes.

Three months use of Hybrid Closed Loop Systems improves glycated hemoglobin levels in adolescents and children with type 1 diabetes: A meta-analysis.

BACKGROUND: Longer outpatient studies have demonstrated that hybrid closed loop (HCL) use has led to a concomitant reduction in glycated hemoglobin(HbA1c) by 0.3%-0.7%. However, reports have also indicated that HbA1c levels are not declined in the long-term use of HCL. Therefore, we wonder that 3 months use of HCL could improve glycated hemoglobin levels in adolescents and children with T1D. METHODS: Relevant studies were searched electronically in the Cochrane Library, PubMed, and Embase utilizing the key words "Pediatrics or Child or Adolescent", "Insulin Infusion Systems" and "Diabetes Mellitus" from inception to 17th March 2024 to evaluate the performance of HCL on HbA1c in adolescents, and children with T1D. RESULTS: Nine studies involving 927 patients were identified. Three months use of HCL show a beneficial effect on HbA1c management (p <0.001) as compared to standard of care in adolescents and children with T1D, without evidence of heterogeneity between articles (I2 = 40%, p = 0.10). HCL did significantly increase the overall average percentage of hypoglycemic time between 70 and 180 mg/dL (TIR) (p <0.001; I2 = 51%). HCL did not show a beneficial effect on hypoglycemic time <70 mg/dL and <54 mg/dL (p >0.05). The overall percentage of hyperglycemic time was significantly decreased in HCL group compared to the control group when it was defined as >180 mg/dL (p <0.001; I2 = 83%), >250 mg/dL (p = 0.007, I2 = 86%) and >300 mg/dL (p = 0.005; I2 = 76%). The mean glucose level was significantly decreased by HCL (p <0.001; I2 = 58%), however, no significant difference was found in coefficient of variation of sensor glucose (p = 0.82; I2 = 71%) and daily insulin dose (p = 0.94; I2 <0.001) between the HCL group and the control group. CONCLUSIONS: HCL had a beneficial effect on HbA1c management and TIR without increased hypoglycemic time as compared to standard of care in adolescents and children with T1D when therapy duration of HCL was not less than three months. TRIAL NUMBER AND REGISTRY URL: CRD42022367493; https://www.crd.york.ac.uk/PROSPERO, Principal investigator: Zhen-feng Zhou, Date of registration: October 30, 2022.

Type 1 Diabetes Mellitus in the First Years of Life - Onset, Initial Treatment, and Early Disease Course.

OBJECTIVE: This study investigated the onset and the choice of treatment in children with very early onset of type 1 diabetes mellitus (T1D). METHODS: The study included 5,763 patients from the German Diabetes Patient Follow-up registry with onset of T1D in the first 4 years of life from January 2010 - June 2022. The analysis included diabetes-specific parameters, anthropometric data, and mode of treatment at onset, within the first and second year of T1D. Three groups were compared according to age at onset (G1: 223 patients 6-<12 months, G2: 1519 patients 12-<24 months, G3: 4001 patients 24-48 months). RESULTS: In 12.3% of all cases in childhood and adolescence, the incidence of diabetes in the first 4 years of life was rare. At the onset, clinical status was worse and diabetic ketoacidosis (DKA) rates were higher in G1 and G2 (52.3% and 46.5%, respectively) compared to G3 (27.3% (p<0.001)). G1 and G2 were significantly more likely to be treated with insulin pump therapy (CSII) 2 years after onset (98.1% and 94.1%, respectively)) compared to G3 (85.8%, p<0.001). Median HbA1c after 2 years did not differ between groups (G1: 7.27% (56.0 mmol/mol), G2: 7.34% (56.7 mmol/mol) and G3: 7.27% (56.0 mmol/mol)) or when comparing CSII vs MDI. The rate of severe hypoglycemia (SH) and DKA during the first 2 years of treatment did not differ among the three groups, ranging from 1.83-2.63/100 patient-years (PY) for DKA and 9.37-24.2/100 PY for SH. Children with T1D under 4 years of age are more likely to be diagnosed with celiac disease but less likely to have thyroiditis than older children with T1DM. CONCLUSIONS: Young children with T1D had high rates of DKA at onset and were predominantly treated with insulin pump therapy during the first 2 years. The median HbA1c for all three groups was<7.5% (58 mmol/mol) without increased risk of SH or DKA. The use of continuous glucose monitoring (CGM) was not associated with lower HbA1c in children under 48 months.

Identifying Patterns in Long Term CGM and Insulin Data - An Explorative Study.

The increased utilization of continuous glucose monitors (CGM) and smart insulin pens (SIP) among people with type 2 diabetes generates significant health data. This study explored possible patterns in long term CGM and SIP data.

Nocturnal Hypoglycemia in the Era of Continuous Glucose Monitoring.

Nocturnal hypoglycemia is a common acute complication of people with diabetes on insulin therapy. In particular, the inability to control glucose levels during sleep, the impact of external factors such as exercise, or alcohol and the influence of hormones are the main causes. Nocturnal hypoglycemia has several negative somatic, psychological, and social effects for people with diabetes, which are summarized in this article. With the advent of continuous glucose monitoring (CGM), it has been shown that the number of nocturnal hypoglycemic events was significantly underestimated when traditional blood glucose monitoring was used. The CGM can reduce the number of nocturnal hypoglycemia episodes with the help of alarms, trend arrows, and evaluation routines. In combination with CGM with an insulin pump and an algorithm, automatic glucose adjustment (AID) systems have their particular strength in nocturnal glucose regulation and the prevention of nocturnal hypoglycemia. Nevertheless, the problem of nocturnal hypoglycemia has not yet been solved completely with the technologies currently available. The CGM systems that use predictive models to warn of hypoglycemia, improved AID systems that recognize hypoglycemia patterns even better, and the increasing integration of artificial intelligence methods are promising approaches in the future to significantly minimize the risk of a side effect of insulin therapy that is burdensome for people with diabetes.

Enhancing the Capabilities of Continuous Glucose Monitoring With a Predictive App.

BACKGROUND: Despite abundant evidence demonstrating the benefits of continuous glucose monitoring (CGM) in diabetes management, a significant proportion of people using this technology still struggle to achieve glycemic targets. To address this challenge, we propose the Accu-Chek® SmartGuide Predict app, an innovative CGM digital companion that incorporates a suite of advanced glucose predictive functionalities aiming to inform users earlier about acute glycemic situations. METHODS: The app's functionalities, powered by three machine learning models, include a two-hour glucose forecast, a 30-minute low glucose detection, and a nighttime low glucose prediction for bedtime interventions. Evaluation of the models' performance included three data sets, comprising subjects with T1D on MDI (n = 21), subjects with type 2 diabetes (T2D) on MDI (n = 59), and subjects with T1D on insulin pump therapy (n = 226). RESULTS: On an aggregated data set, the two-hour glucose prediction model, at a forecasting horizon of 30, 45, 60, and 120 minutes, achieved a percentage of data points in zones A and B of Consensus Error Grid of: 99.8%, 99.3%, 98.7%, and 96.3%, respectively. The 30-minute low glucose prediction model achieved an accuracy, sensitivity, specificity, mean lead time, and area under the receiver operating characteristic curve (ROC AUC) of: 98.9%, 95.2%, 98.9%, 16.2 minutes, and 0.958, respectively. The nighttime low glucose prediction model achieved an accuracy, sensitivity, specificity, and ROC AUC of: 86.5%, 55.3%, 91.6%, and 0.859, respectively. CONCLUSIONS: The consistency of the performance of the three predictive models when evaluated on different cohorts of subjects with T1D and T2D on different insulin therapies, including real-world data, offers reassurance for real-world efficacy.