Diabetic Ketoacidosis Among Patients Treated With Continuous Subcutaneous Insulin Infusion

Supervised and Unsupervised Approaches for the Real-Time Detection of Undesired Insulin Suspension Caused by Malfunctions

BACKGROUND

Automated insulin delivery (AID) systems, permit improved treatment of type 1 diabetes (T1D). Unfortunately, malfunctioning in the insulin pump or in the infusion set can prevent insulin from being administered, reducing the AID efficacy and posing the patient at risk. Different data-driven methods available in the literature can be used to deal with the problem of automatically detecting complete insulin suspension in real-time. This article investigates both supervised and unsupervised strategies and proposes a fair comparison under either population or personalized settings.

METHODS

Several algorithms are compared using data generated through the UVA/Padova T1D simulator, a computer simulator widely used to test control strategies in silico and accepted by the Food and Drugs Administration (FDA) as a substitute to animal pre-clinical trials. Two synthetic data sets, each consisting of 100 virtual subjects monitored for 1 month, were generated. Occasional faults of the insulin pump are simulated as complete occlusions by suspending the therapy administration. Personalized algorithms are investigated with unsupervised approaches only, since personalized labels are hardly available.

RESULTS

In the population scenario, the supervised approach outperforms the unsupervised strategy. In particular, logistic regression and random forest achieves a recall of 72% and 82%, with 0.12 and 0.21 false positives (FP) per day, respectively. In the personalized setting scenario, the unsupervised algorithms are tailored on each patient and outperform the population ones, in particular isolation forest achieves a recall 80% and 0.06 FPs per day.

CONCLUSIONS

This article suggests that unsupervised personalized approach, by addressing the large variability in glucose response among individuals with T1D, is superior to other one-fits-all approaches in detecting insulin suspensions caused by malfunctioning. Population methodologies can be effectively used while waiting to collect sufficient patient data, when the system is installed on a new patient.

Frequency and Detection of Insulin Infusion Site Failure in the Type 1 Diabetes Exchange Online Community

Insulin infusion site (IIS) failures are a weakness in insulin pump therapy. We examined experience with IIS failures among U.S. individuals with diabetes on insulin pump through survey distributed to the T1D Exchange Online Community. Demographic factors, IIS characteristics, and diabetes-related perceptions were assessed by logistic regression to determine odds of higher (≥1 per month) or lower (<1 per month) reported IIS failure frequency. IIS failures were common; 41.4% reported ≥1 per month. IIS failure is usually detected through development of hyperglycemia rather than pump alarm. No assessed demographic factor or IIS characteristic was predictive; however, higher odds of ≥1 failure per month were associated with feelings of burnout (odds ratios [OR] 1.489 [1.024, 2.165]) and considering pump discontinuation (OR 2.233 [1.455, 3.427]). IIS failures are frequent and unpredictable, typically require hyperglycemia for detection, and are associated with negative perceptions. More should be done toward preventing IIS failures and/or detecting them sooner.

Insulin pump therapy is associated with higher rates of mild diabetic ketoacidosis compared to injection therapy: A 2-year Swedish national survey of children and adolescents with type 1 diabetes

OBJECTIVES

Diabetic ketoacidosis (DKA) in type 1 diabetes (T1D) can occur during both insulin pump therapy (continuous subcutaneous insulin infusion, CSII) and insulin injection therapy (multiple daily injections, MDI). The primary aim of this study was to compare CSII and MDI regarding DKA frequency. A secondary aim was to compare metabolic derangement between CSII and MDI at hospital admission for DKA. RESEARCH DESIGN AND METHODS: Children 0-17.99 years with established T1D admitted for DKA in Sweden from February 1, 2015 to January 31, 2017 were invited to participate. Data regarding demographics, laboratory data, CSII or MDI, and access to ketone meters and CGM were provided through questionnaires and medical records. The Swedish National Diabetes Registry (SWEDIABKIDS) was used to compare the distribution of CSII and MDI in the national population with the population admitted for DKA, using the chi-square goodness-of-fit test. Distribution of CSII and MDI was then categorized in clinical severity grades for mild (pH 7.20-7.29), moderate (pH 7.10-7.29) and severe DKA (pH <7.10).

RESULTS

The distribution of CSII at DKA admission was significantly larger than in the national pediatric population with T1D (74.7% vs. 59.7%, p = 0.002). CSII was overrepresented in mild DKA (85.2% vs. with CSII, p < 0.001), but not in moderate/severe DKA (57.9% with CSII, p = 0.82). Mean HbA1c at hospital admission was 73.9 mmol/mol with CSII and 102.7 mmol/mol with MDI.

CONCLUSIONS

CSII was associated with higher risk of mild DKA than MDI. MDI was associated with markedly higher HbA1c levels than CSII at hospital admission for DKA.

Machine Learning-Based Anomaly Detection Algorithms to Alert Patients Using Sensor Augmented Pump of Infusion Site Failures

BACKGROUND

Personal insulin pumps have shown to be effective in improving the quality of therapy for people with type 1 diabetes (T1D). However, the safety of this technology is limited by the possible infusion site failures, which are linked with hyperglycemia and ketoacidosis. Thanks to the large availability of collected data provided by modern therapeutic technologies, machine learning algorithms have the potential to provide new way to identify failures early and avert adverse events.

METHODS

A clinical dataset (N = 20) is used to evaluate a novel method for detecting real-time infusion site failures using unsupervised anomaly detection algorithms, previously proposed and developed on in-silico data. An adapted feature engineering procedure is introduced to make the method able to operate in the absence of a closed-loop (CL) system and meal announcements.

RESULTS

In the optimal configuration, we obtained a performance of 0.75 Sensitivity (15 out of 20 total failures detected) and 0.08 FP/day, outperforming previously proposed literature algorithms. The algorithm was able to anticipate the replacement of the malfunctioning infusion sets by ~2 h on average.

CONCLUSIONS

On the considered dataset, the proposed algorithm showed the potential to improve the safety of patients treated with sensor-augmented pump systems.

Rising Rates and Widening Socioeconomic Disparities in Diabetic Ketoacidosis in Type 1 Diabetes in Scotland: A Nationwide Retrospective Cohort Observational Study

OBJECTIVE

Whether advances in the management of type 1 diabetes are reducing rates of diabetic ketoacidosis (DKA) is unclear. We investigated time trends in DKA rates in a national cohort of individuals with type 1 diabetes monitored for 14 years, overall and by socioeconomic characteristics.

RESEARCH DESIGN AND METHODS

All individuals in Scotland with type 1 diabetes who were alive and at least 1 year old between 1 January 2004 and 31 December 2018 were identified using the national register (N = 37,939). DKA deaths and hospital admissions were obtained through linkage to Scottish national death and morbidity records. Bayesian regression was used to test for DKA time trends and association with risk markers, including socioeconomic deprivation.

RESULTS

There were 30,427 DKA admissions and 472 DKA deaths observed over 393,223 person-years at risk. DKA event rates increased over the study period (incidence rate ratio [IRR] per year 1.058 [95% credibility interval 1.054-1.061]). Males had lower rates than females (IRR male-to-female 0.814 [0.776-0.855]). DKA incidence rose in all age-groups other than 10- to 19-year-olds, in whom rates were the highest, but fell over the study. There was a large socioeconomic differential (IRR least-to-most deprived quintile 0.446 [0.406-0.490]), which increased during follow-up. Insulin pump use or completion of structured education were associated with lower DKA rates, and antidepressant and methadone prescription were associated with higher DKA rates.

CONCLUSIONS

DKA incidence has risen since 2004, except in 10- to 19-year-olds. Of particular concern are the strong and widening socioeconomic disparities in DKA outcomes. Efforts to prevent DKA, especially in vulnerable groups, require strengthening.

Causes of diabetic ketoacidosis among adults with type 1 diabetes mellitus: insulin pump users and non-users

INTRODUCTION

Insulin pumps are increasingly being used as a method of insulin delivery in patients with type 1 diabetes mellitus (T1DM). Diabetic ketoacidosis (DKA) is a serious complication of T1DM. This study aims to identify the causes of DKA in patients with T1DM on continuous subcutaneous insulin infusion (CSII) and to compare these with patients with T1DM on multiple daily insulin injections (MDIIs).

RESEARCH DESIGN AND METHODS

This is a prospective observational study between January and June 2019 at the Cleveland Clinic Fairview Hospital. Demographic, clinical, and biochemical data were obtained from chart review. A questionnaire to explore additional clinical data relating to DKA was administered, with additional items for patients on the insulin pump.

RESULTS

Seventy-four patients were admitted with a diagnosis of DKA between the period of January and June 2019. Of these, 45 met the inclusion criteria and 43 consented. These were divided into two groups: group 1 included patients on MDII and group 2 included CSII. Overall, the most common precipitating factor for developing DKA was insulin non-adherence, seen in 51.2% of the cases. The most common cause of DKA in group 2 was pump/tubing related to 55% of the cases.

CONCLUSION

Despite non-adherence being common in both CSII and MDII, a combination of social factors, education and insulin pump malfunction, such as pump/tubing problems, might be playing a pivotal role in DKA etiology in young adults with T1DM, especially in CSII users. Continued education on pump use may reduce the rate of DKA in pump users.

Diabetic ketoacidosis

Diabetic ketoacidosis (DKA) is the most common acute hyperglycaemic emergency in people with diabetes mellitus. A diagnosis of DKA is confirmed when all of the three criteria are present - 'D', either elevated blood glucose levels or a family history of diabetes mellitus; 'K', the presence of high urinary or blood ketoacids; and 'A', a high anion gap metabolic acidosis. Early diagnosis and management are paramount to improve patient outcomes. The mainstays of treatment include restoration of circulating volume, insulin therapy, electrolyte replacement and treatment of any underlying precipitating event. Without optimal treatment, DKA remains a condition with appreciable, although largely preventable, morbidity and mortality. In this Primer, we discuss the epidemiology, pathogenesis, risk factors and diagnosis of DKA and provide practical recommendations for the management of DKA in adults and children.

Detection of Insulin Pump Malfunctioning to Improve Safety in Artificial Pancreas Using Unsupervised Algorithms

BACKGROUND

Recent development of automated closed-loop (CL) insulin delivery systems, the so-called artificial pancreas (AP), improved the quality of type 1 diabetes (T1D) therapy. As new technologies emerge, patients put increasing trust in their therapeutic devices; therefore, it becomes increasingly important to detect malfunctioning affecting such devices. In this work, we explore a new paradigm to detect insulin pump faults (IPFs) that use unsupervised anomaly detection.

METHODS

We generated CL data corrupted with IPFs using the latest version of the T1D Padova/UVA simulator. From the data, we extracted several features capable to describe the patient dynamics and making more apparent suspicious data portions. Then, a feature selection is performed to determine the optimal feature set. Finally, the performance of several popular unsupervised anomaly detection algorithms is analyzed and compared on the identified optimal feature set.

RESULTS

Using the identified optimal configuration, the best performance is obtained by the Histogram-Based Outlier Score (HBOS) algorithm, which detected 87% of the IPF with only 0.08 false positives per day on average. Isolation forest is the best algorithm that offers more conservative performances, detection of 85% of the faults but only 0.06 false positives per day on average.

CONCLUSION

Unsupervised anomaly detection algorithms can be used effectively to detect IPFs and improve the safety of the AP. Future studies will be dedicated to test the presented method inside dedicated clinical trials.

Current Diabetes Technology: Striving for the Artificial Pancreas

Diabetes technology has continually evolved over the years to improve quality of life and ease of care for affected patients. Frequent blood glucose (BG) checks and multiple daily insulin injections have become standard of care in Type 1 diabetes (T1DM) management. Continuous glucose monitors (CGM) allow patients to observe and discern trends in their glycemic control. These devices improve quality of life for parents and caregivers with preset alerts for hypoglycemia. Insulin pumps have continued to improve and innovate since their emergence into the market. Hybrid closed-loop systems have harnessed the data gathered with CGM use to aid in basal insulin dosing and hypoglycemia prevention. As technology continues to progress, patients will likely have to enter less and less information into their pump system manually. In the future, we will likely see a system that requires no manual patient input and allows users to eat throughout the day without counting carbohydrates or entering in any blood sugars. As technology continues to advance, endocrinologists and diabetes providers need to stay current to better guide their patients in optimal use of emerging management tools.

Real-Time Detection of Infusion Site Failures in a Closed-Loop Artificial Pancreas

BACKGROUND

As evidence emerges that artificial pancreas systems improve clinical outcomes for patients with type 1 diabetes, the burden of this disease will hopefully begin to be alleviated for many patients and caregivers. However, reliance on automated insulin delivery potentially means patients will be slower to act when devices stop functioning appropriately. One such scenario involves an insulin infusion site failure, where the insulin that is recorded as delivered fails to affect the patient's glucose as expected. Alerting patients to these events in real time would potentially reduce hyperglycemia and ketosis associated with infusion site failures.

METHODS

An infusion site failure detection algorithm was deployed in a randomized crossover study with artificial pancreas and sensor-augmented pump arms in an outpatient setting. Each arm lasted two weeks. Nineteen participants wore infusion sets for up to 7 days. Clinicians contacted patients to confirm infusion site failures detected by the algorithm and instructed on set replacement if failure was confirmed.

RESULTS

In real time and under zone model predictive control, the infusion site failure detection algorithm achieved a sensitivity of 88.0% (n = 25) while issuing only 0.22 false positives per day, compared with a sensitivity of 73.3% (n = 15) and 0.27 false positives per day in the SAP arm (as indicated by retrospective analysis). No association between intervention strategy and duration of infusion sets was observed ( P = .58).

CONCLUSIONS

As patient burden is reduced by each generation of advanced diabetes technology, fault detection algorithms will help ensure that patients are alerted when they need to manually intervene. Clinical Trial Identifier: www.clinicaltrials.gov,NCT02773875.