Reducing Hypoglycemia in the Real World: A Retrospective Analysis of Predictive Low-Glucose Suspend Technology in an Ambulatory Insulin-Dependent Cohort

Automating insulin delivery through pump and continuous glucose monitoring connectivity: Maximizing opportunities to improve outcomes

The development of automated insulin delivery (AID) systems, which connect continuous glucose monitoring (CGM) systems with algorithmic insulin delivery from an insulin pump (continuous subcutaneous insulin infusion, [CSII]), has led to improved glycaemia and quality of life benefits in those with insulin-treated diabetes. This review summarizes the benefits gained by the connectivity between insulin pumps and CGM devices. It details the technical requirements and advances that have enabled this, and highlights the clinical and user benefits of such systems. Clinical trials and real-world outcomes from the use of AID systems in people with type 1 diabetes (T1D) will be the focus of this article; outcomes in people with type 2 diabetes (T2D) and other diabetes subtypes will also be discussed. We also detail the limitations of current technological approaches for connectivity between insulin pumps and CGM devices. While recognizing the barriers, we discuss opportunities for the future.

Neonatal birth weight percentile following the use of sensor-augmented pump therapy in women with pre-gestational diabetes

AIMS

To assess the effect of using sensor-augmented pump therapy (SAP) during pregnancy on neonatal birth weight percentile and other neonatal and pregnancy outcomes.

METHODS

This retrospective cohort study included consecutive women with pregestational diabetes mellitus (PGDM) treated with an insulin pump and sensor that enabled the SAP feature during pregnancy. SAP use was defined as utilization of either low-glucose suspend (LGS) or predictive LGS technology. Utilization of SAP was according to physician discretion. Differences in neonatal birth weight percentile and in other neonatal and pregnancy outcomes were compared between those who did and not use SAP.

OUTCOMES

Of 142 women, 136 had type 1 diabetes, 5 type 2 diabetes and one diabetes due to pancreatectomy. 83 women used SAP and 59 did not. For the neonates of the mothers of the respective groups, the median birth weight percentiles were similar (79 and 80, pV = 0.96), as were the other neonatal outcomes assessed. The rate of cesarean section was higher in the SAP group. However, after adjusting for maternal age, BMI, and a history of severe hypoglycemic events before pregnancy, the relation between mode of delivery and the use of SAP was no longer statistically significant.

CONCLUSION

In women with PGDM treated with an insulin pump and sensor, SAP use during pregnancy was not associated with higher neonatal birth weight percentile or the occurrences of other adverse neonatal or pregnancy outcomes.

Sensor-Augmented Insulin Pump with Predictive Low-Glucose Suspend (PLGS): Determining Optimal Settings of Pump and Sensor in a Multicenter Cohort of Patients with Type 1 Diabetes

INTRODUCTION

The use of predictive low-glucose suspend (PLGS) sensor-augmented pumps has been shown to lead to a significant reduction in hypoglycemic episodes in patients with type 1 diabetes (T1D), but their effects on hyperglycemia exposure are heterogeneous. The aim of this study was to determine the settings of the Medtronic 640G system to obtain the optimal balance between occurrence of both hypoglycemia and hyperglycemia.

METHODS

The hypo- and hyperglycemia area under the curve (AUC), as well as system settings [hypoglycemic threshold, mean insulin total daily dose (TDD), mean basal insulin percentage, and mean daily duration of PLGS] were collected between 2 and 12 times during 1 year in patients from four university hospital centers. Univariate/multivariate analyses and receiver operating characteristics (ROC) curves were performed to determine factors associated with hyper- and hypoglycemia AUC.

RESULTS

A total of 864 observations were analyzed from 110 patients with T1D. Two preselected settings predictive of low hyperglycemia AUC were a basal insulin percentage < 52.0% [sensitivity (Se) = 0.66 and specificity (Sp) = 0.53] and a PLGS duration > 157.5 min/day (Se = 0.47 and Sp = 0.73). The preselected setting predictive of a low hypoglycemia AUC was a PLGS duration ≤ 174.4 min (Se = 0.83 and Sp = 0.51). Between-visit variation of PLGS and TDD was positively correlated (r = 0.61; p < 0.0001).

CONCLUSION

The most important Medtronic 640G setting was the mean daily PLGS duration, where a value between 157.5 and 174.4 min/day was associated with the best reduction in both hypo- and hyperglycemia AUC. In this study, we showed that PLGS duration could be indirectly modified through total daily insulin dose adaptation.

TRIAL REGISTRATION

This study is registered in clinicaltrials.gov (NCT03047486).

Artificial intelligence in critically ill diabetic patients: current status and future prospects

Recent years have witnessed increasing numbers of artificial intelligence (AI) based applications and devices being tested and approved for medical care. Diabetes is arguably the most common chronic disorder worldwide and AI is now being used for making an early diagnosis, to predict and diagnose early complications, increase adherence to therapy, and even motivate patients to manage diabetes and maintain glycemic control. However, these AI applications have largely been tested in non-critically ill patients and aid in managing chronic problems. Intensive care units (ICUs) have a dynamic environment generating huge data, which AI can extract and organize simultaneously, thus analysing many variables for diagnostic and/or therapeutic purposes in order to predict outcomes of interest. Even non-diabetic ICU patients are at risk of developing hypo or hyperglycemia, complicating their ICU course and affecting outcomes. In addition, to maintain glycemic control frequent blood sampling and insulin dose adjustments are required, increasing nursing workload and chances of error. AI has the potential to improve glycemic control while reducing the nursing workload and errors. Continuous glucose monitoring (CGM) devices, which are Food and Drug Administration (FDA) approved for use in non-critically ill patients, are now being recommended for use in specific ICU populations with increased accuracy. AI based devices including artificial pancreas and CGM regulated insulin infusion system have shown promise as comprehensive glycemic control solutions in critically ill patients. Even though many of these AI applications have shown potential, these devices need to be tested in larger number of ICU patients, have wider availability, show favorable cost-benefit ratio and be amenable for easy integration into the existing healthcare systems, before they become acceptable to ICU physicians for routine use.

Erste Anwendungserfahrung eines neuen, Glukosesensor-unterstützten Pumpensystems mit vorausschauender Insulin-Abschaltung zum Hypoglykämieschutz bei pädiatrischen Patienten in Deutschland

Einleitung Die prädiktive Insulinabschaltung ist als System zur Prävention von Hypoglykämien in Deutschland etabliert (Smartguard). Seit 2020 ist in Deutschland ein zweites System verfügbar (Basal-IQ). Unterschiede betreffen eine nicht veränderbare prädiktive Abschaltgrenze von 80 mg/dl (vs. 50–90 mg/dl), eine Abschaltzeit von minimal 5 Minuten (vs. 30 Minuten) sowie die Festlegung der Wiedereinschaltung des Insulin bei einem höheren Wert als zuvor (vs. einem Abstand von 20 mg/dl über der Abschaltgrenze und höherer Prädiktion). Die Systeme wurden in einer Altersgruppe, die besonders von Unterzuckerungen bedroht ist, verglichen.

Methodik Pädiatrische Patienten (Alter 6–13 Jahre), mit Pumpen- und Sensorerfahrung (kein AID) wurde die Erprobung von Basal-IQ für eine Dauer von 3 Monaten angeboten. Betrachtet wurden die CGM-Parameter Zeit unter Zielbereich (TBR < 70mg/dl), im Zielbereich (TIR 70–180 mg/dl), glykämische Variabilität (Varianzkoeffizient CV%) und HbA1c. Patienten-bezogene Outcomes (PROʼs) wurden mit dem Diabskids-Elternfragebogen und einem Gerätefragebogen erfasst.

Ergebnisse Neun Teilnehmer (alle männlich, Mittelwerte: 9.7 Jahre, Diabetesdauer 6.1 Jahre, HbA1c 6.8%, Time in Range (TIR) 61.9%, Time below Range (TBR) 4.5%, mittlere Glukose (MW) 164 mg/dl, (CV) 40) wurden gefunden. Nach 3 Monaten konnten Verbesserungen der glykämischen Parameter beobachtet werden (HbA1c 6.5%, TIR 69.2%, TBR 2.8%, MW 159, CV 40; Kontrollen HbA1c 7.2%, TIR 64.9%, TBR 4.3%, MW 158, CV 39), die sich von einer zeitgleich mit Smartguard behandelten Kindern nicht unterschieden. Die Erfassung der PROʼs zeigte einen Rückgang der Diabetes- und Therapiebelastung, sowie eine Zufriedenheit mit dem System.

Diskussion Das neue System mit prädiktiver Abschaltung zeigte nach 3 Monaten eine Verbesserung der glykämischen Parameter und PROʼs. Ein statistischer Vergleich vorher/nachher ist aufgrund der geringen Patientenzahl nicht erfolgt, aber die Daten zeigen zumindest die Nichtunterlegenheit gegenüber dem Baseline-Zeitpunkt und den Daten, die aus einer Gruppe von Patienten mit kontinuierlicher Systemnutzung stammen. Somit stehen in Deutschland aktuell zwei verschiedene effiziente Systeme mit prädiktiver Insulinabschaltung für Kinder und Jugendliche mit Diabetes zur Verfügung, so dass diese nach fundierter Beratung auswählen können.

An In Silico Head-to-Head Comparison of the Do-It-Yourself Artificial Pancreas Loop and Bio-Inspired Artificial Pancreas Control Algorithms

BACKGROUND

User-developed automated insulin delivery systems, also referred to as do-it-yourself artificial pancreas systems (DIY APS), are in use by people living with type 1 diabetes. In this work, we evaluate, in silico, the DIY APS Loop control algorithm and compare it head-to-head with the bio-inspired artificial pancreas (BiAP) controller for which clinical data are available.

METHODS

The Python version of the Loop control algorithm called PyLoopKit was employed for evaluation purposes. A Python-MATLAB interface was created to integrate PyLoopKit with the UVa-Padova simulator. Two configurations of BiAP (non-adaptive and adaptive) were evaluated. In addition, the Tandem Basal-IQ predictive low-glucose suspend was used as a baseline algorithm. Two scenarios with different levels of variability were used to challenge the algorithms on the adult (n = 10) and adolescent (n = 10) virtual cohorts of the simulator.

RESULTS

Both BiAP and Loop improve, or maintain, glycemic control when compared with Basal-IQ. Under the scenario with lower variability, BiAP and Loop perform relatively similarly. However, BiAP, and in particular its adaptive configuration, outperformed Loop in the scenario with higher variability by increasing the percentage time in glucose target range 70-180 mg/dL (BiAP-Adaptive vs Loop vs Basal-IQ) (adults: 89.9% ± 3.2%* vs 79.5% ± 5.3%* vs 67.9% ± 8.3%; adolescents: 74.6 ± 9.5%* vs 53.0% ± 7.7% vs 55.4% ± 12.0%, where * indicates the significance of P < .05 calculated in sequential order) while maintaining the percentage time below range (adults: 0.89% ± 0.37% vs 1.72% ± 1.26% vs 3.41 ± 1.92%; adolescents: 2.87% ± 2.77% vs 4.90% ± 1.92% vs 4.17% ± 2.74%).

CONCLUSIONS

Both Loop and BiAP algorithms are safe and improve glycemic control when compared, in silico, with Basal-IQ. However, BiAP appears significantly more robust to real-world challenges by outperforming Loop and Basal-IQ in the more challenging scenario.

Real-world outcomes of two different sensor-augmented insulin pumps with predictive low glucose suspend function in type 1 diabetes patients

AIM

To analyse the real-life outcomes of two sensor-augmented pumps (SAP) with predictive low glucose suspend (PLGS) function, Medtronic Minimed 640G™ with SmartGuard (MM640G) and Tandem T Slim X2™ with Basal-IQ™ (TTSX2), in Type 1 Diabetes Mellitus (T1DM) patients.

METHODS

Observational cross-sectional study using data obtained from computerized clinical records. All T1DM patients on TTSX2 therapy were compared (1:1) with MM640G treated patients selected through stratified sampling. Primary efficacy outcome was to describe time in rage (TIR, 70-180 mg/dL, 3.9-10 mmol/L) interstitial glucose differences according to a non-inferiority hypothesis with TTSX2 compared to MM640G.

RESULTS

Forty-four patients were analyzed (female 66%). Mean age was 38.9 yrs. (range 23-59 yrs.) and mean diabetes duration was 23.4 ± 9.2 yrs. Patients treated with TTSX2 showed a numerically slightly lower, but non-statistically significantly different, TIR from the MM640G pump group (64.9 ± 16.4% vs. 72.4 ± 17.0%, P = 0.108). Similarly, we did no find differences in HbA1c between T1D patients treated with TTSX2 and MM640G (6.8 ± 1.0% vs. 7.0 ± 0.9%, 51 ± 11 mmol/mol vs. 53 ± 10 mmol/mol, P = 0.312). Moreover, rest of evaluated glycemic outcomes were similar between both treatment groups.

CONCLUSIONS

Patients using two different SAP with PLGS automatic function showed similar glycaemic control in a real-world scenario. NCT04741685.

Sustainable Use of a Real-Time Continuous Glucose Monitoring System from 2018 to 2020

We aimed to describe patterns of continuous glucose monitoring (CGM) system use and glycemic outcomes from 2018 to 2020 in a large real-world cohort by analyzing anonymized data from US-based CGM users who transitioned from the G5 to the G6 System (Dexcom) in 2018. The main end points were persistent use, within-day and between-day utilization, hypoglycemia, time in range (TIR, 70-180 mg/dL [3.9-10 mmol/L]), and use of the optional calibration feature in 2019 and 2020. In a cohort of 31,034 individuals, rates of persistent use were high, with 27,932 (90.0%) and 26,861 (86.6%) continuing to upload data in 2019 and 2020, respectively. Compared with G5 use, G6 use was associated with higher device utilization, less hypoglycemia, higher TIR (in 2020), and >80% fewer calibrations in both 2019 and 2020 (P's < 0.001). High persistence and utilization of the G6 system may contribute to sustainable glycemic outcomes and decreased user burden.

Real-World Patient-Reported Outcomes and Glycemic Results with Initiation of Control-IQ Technology

Background: The t:slim X2™ insulin pump with Control-IQ™ technology, an advanced hybrid closed-loop system, became available in the United States in early 2020. Real-world outcomes with use of this system have not yet been comprehensively reported. Methods: Individuals with type 1 diabetes (T1D) (≥14 years of age) who had ≥21 days of pump usage data were invited via email to participate. Participants completed psychosocial questionnaires (Technology Acceptance Scale [TAS], well-being index [WHO-5], and Diabetes Impact and Devices Satisfaction [DIDS] scale) at timepoint 1 (T1) (at least 3 weeks after starting Control-IQ technology) and the DIDS and WHO-5 at timepoint 2 (T2) (4 weeks from T1). Patient-reported outcomes (PROs) and glycemic outcomes were reviewed at each timepoint. Results: Overall, 9,085 potentially eligible individuals received the study invite. Of these, 3,116 consented and subsequently 1,435 participants completed questionnaires at both T1 and T2 and had corresponding glycemic data available on the t:connect® web application. Time in range was 78.2% (70.2%-85.1%) at T1 and 79.2% (70.3%-86.2%) at T2. PROs reflected high device-related satisfaction and reduced diabetes impact at T2. Factors contributing to high trust in the system included sensor accuracy, improved diabetes control, reduction in extreme blood glucose levels, and improved sleep quality. In addition, participants reported improved quality of life, ease of use, and efficient connectivity to the continuous glucose monitoring system as being valuable features of the system. Conclusions: Continued real-world use of the t:slim X2 pump with Control-IQ technology showed improvements in psychosocial outcomes and persistent achievement of recommended TIR glycemic outcomes in people with T1D.

Weight Management in Youth with Type 1 Diabetes and Obesity: Challenges and Possible Solutions

PURPOSE OF REVIEW

This review highlights challenges associated with weight management in children and adolescents with type 1 diabetes (T1D). Our purpose is to propose potential solutions to improve weight outcomes in youth with T1D.

RECENT FINDINGS

A common barrier to weight management in T1D is reluctance to engage in exercise for fear of hypoglycemia. Healthcare practitioners generally provide limited guidance for insulin dosing and carbohydrate modifications to maintain stable glycemia during exercise. Adherence to dietary guidelines is associated with improved glycemia; however, youth struggle to meet recommendations. When psychosocial factors are addressed in combination with glucose trends, this often leads to successful T1D management. Newer medications also hold promise to potentially aid in glycemia and weight management, but further research is necessary. Properly addressing physical activity, nutrition, pharmacotherapy, and psychosocial factors while emphasizing weight management may reduce the likelihood of obesity development and its perpetuation in this population.

New and Emerging Technologies in Type 1 Diabetes

There has been a rapid advancement in the pace of development of new diabetes technologies and therapies for the management of type 1 diabetes over the past decade. The Diabetes Control and Complications Trial conclusively established that tight glycemic control with intensive insulin therapy decreases the rates of diabetes complications in proportion to glycemic control, and diabetes technologies have accordingly been developed to help patients reach these goals. In this review, the authors discuss new diabetes therapeutics and technologies, including new insulin analogues, insulin pumps, continuous glucose monitoring systems, and automated insulin delivery systems."

Closed-loop insulin delivery systems in children and adolescents with type 1 diabetes

Introduction: Optimal glycemic control remains challenging in children and adolescents with type 1 diabetes due to highly variable day-to-day and night-to-night insulin requirements. This hurdle could be addressed by glucose-responsive insulin delivery based on real-time continuous glucose measurements.Areas covered: This review summaries recent advances of closed-loop systems in children and adolescents with type 1 diabetes, using both single- and dual-hormone closed-loop systems. The main outcomes, proportions of time spent in target range 70-180 mg/dl, and time spent in hypoglycemia below 70 mg/dl, are assessed particularly during unsupervised free-living randomized controlled trials.Expert opinion: Noteworthy and clinically meaningful translation of experimental investigations from controlled in-hospital settings to unrestricted home studies have been achieved over the past years, resulting in the regulatory approval of the first hybrid closed-loop system also in the pediatric population and with several other advanced devices in the pipeline. Large multinational and pivotal clinical trials including broad age populations are underway to facilitate the use of closed-loop systems in routine clinical practice.

Real-World Improvements in Hypoglycemia in an Insulin-Dependent Cohort With Diabetes Mellitus Pre/Post Tandem Basal-Iq Technology Remote Software Update

OBJECTIVE

Software updatable insulin pumps, such as the t:slim X2 pump from Tandem Diabetes Care, enable access to new technology as soon as it is commercialized. The remote software update process allows for minimal interruption in therapy compared to purchasing a new pump; however, little quantitative data exist on the software update process or on pre/post therapeutic outcomes. We examined real-world usage and impact of a remote software updatable predictive low-glucose suspend (PLGS) technology designed to reduce hypoglycemic events in people with insulin-dependent diabetes.

METHODS

Approximately 15,000 U.S. Tandem pump users remotely updated their t:slim X2 software to Basal-IQ PLGS technology since its commercial release. We performed a retrospective analysis of users who uploaded at least 21 days of pre/post PLGS update usage data to the Tandem t:connect web application between August 28, 2018, and October 21, 2019 (N = 6,170). Insulin delivery and sensor-glucose values were analyzed per recent international consensus and American Diabetes Association guidelines. Software update performance was also assessed.

RESULTS

Median software update time was 5.36 minutes. Overall glycemic outcomes for pre and post software update showed a decrease in sensor time <70 mg/dL from 2.14 to 1.18% (-1.01; 95% confidence interval [CI], -0.97, -1.05; P<.001), with overall sensor time 70 to 180 mg/dL increasing from 57.8 to 58.5% (0.64; 95% CI, 0.04, 1.24; P<.001). These improvements were sustained at 3, 6, and 9 months after the update.

CONCLUSION

Introduction of a software updatable PLGS algorithm for the Tandem t:slim X2 insulin pump resulted in sustained reductions of hypoglycemia.

ABBREVIATIONS

ADA = American Diabetes Association; CGM = continuous glucose monitoring; CI = confidence interval; PLGS = predictive low-glucose suspend; SG = sensor glucose; T1D = type 1 diabetes; T2D = type 2 diabetes; TIR = time-in-range.

Artificial Intelligence and Machine Learning Applied at the Point of Care

Introduction

The increasing availability of healthcare data and rapid development of big data analytic methods has opened new avenues for use of Artificial Intelligence (AI)- and Machine Learning (ML)-based technology in medical practice. However, applications at the point of care are still scarce.

Objective

Review and discuss case studies to understand current capabilities for applying AI/ML in the healthcare setting, and regulatory requirements in the US, Europe and China.

Methods

A targeted narrative literature review of AI/ML based digital tools was performed. Scientific publications (identified in PubMed) and grey literature (identified on the websites of regulatory agencies) were reviewed and analyzed.

Results

From the regulatory perspective, AI/ML-based solutions can be considered medical devices (i.e., Software as Medical Device, SaMD). A case series of SaMD is presented. First, tools for monitoring and remote management of chronic diseases are presented. Second, imaging applications for diagnostic support are discussed. Finally, clinical decision support tools to facilitate the choice of treatment and precision dosing are reviewed. While tested and validated algorithms for precision dosing exist, their implementation at the point of care is limited, and their regulatory and commercialization pathway is not clear. Regulatory requirements depend on the level of risk associated with the use of the device in medical practice, and can be classified into administrative (manufacturing and quality control), software-related (design, specification, hazard analysis, architecture, traceability, software risk analysis, cybersecurity, etc.), clinical evidence (including patient perspectives in some cases), non-clinical evidence (dosing validation and biocompatibility/toxicology) and other, such as e.g. benefit-to-risk determination, risk assessment and mitigation. There generally is an alignment between the US and Europe. China additionally requires that the clinical evidence is applicable to the Chinese population and recommends that a third-party central laboratory evaluates the clinical trial results.

Conclusions

The number of promising AI/ML-based technologies is increasing, but few have been implemented widely at the point of care. The need for external validation, implementation logistics, and data exchange and privacy remain the main obstacles.

Effect of real-life insulin pump with predictive low-glucose management use for 3 months: Analysis of the patients treated in a Japanese center

Aims/Introduction

In Japan, an insulin pump with predictive low‐glucose management (PLGM) was launched in 2018. It automatically suspends insulin delivery when the sensor detects or predicts low glucose values. The aim of this study was to analyze the safety and efficacy of PLGM in patients treated in a Japanese center.

Materials and Methods

We carried out a retrospective observational analysis of 16 patients with type 1 diabetes mellitus and one patient after pancreatectomy. They switched from the MiniMed 620G device to the 640G device with PLGM. The primary outcome was the change in the percentage of time in hypoglycemia. The secondary outcome was the change in HbA1c (%) over a period of 3 months. We also explored the presence of “post‐suspend hyperglycemia” with the 640G device.

Results

After changing to the 640G device, the percentage of time in hypoglycemia (glucose <50 mg/dL) significantly decreased from 0.39% (0–1.51%) to 0% (0–0.44%; P = 0.0407). The percentage of time in hyperglycemia (glucose >180 mg/dL) significantly increased from 25.53% (15.78–44.14%) to 32.9% (24.71–45.49%; P = 0.0373). HbA1c significantly increased from 7.6 ± 1.0% to 7.8 ± 1.1% (P = 0.0161). From 1.5 to 4.5 h after the resumption of insulin delivery, the percentage of time in hyperglycemia was 32.23% (24.2–53.75%), but it was significantly lower, 2.78% (0–21.6%), when patients manually restarted the pump within 30 min compared with automatic resumption 31.2% (20–61.66%; P = 0.0063).

Conclusions

Predictive low‐glucose management is an effective tool for reducing hypoglycemia, but possibly elicits “post‐suspend hyperglycemia.” This information is useful for achieving better blood glucose control in the patients treated with PLGM.