2014 Diabetes Technology Meeting Abstracts

doi:10.1177/1932296815570777

. 2015 Mar 1;9(2):342–485. doi: 10.1177/1932296815570777

2014 Diabetes Technology Meeting Abstracts

PMCID: PMC4604580

Abdel-Tawab	Dose Accuracy of Reusable Insulin Pens in India: NovoPen 4, NovoPen 3, HumaPen…	A1
Ali	From Bench to Bedside: Conquering the Translational Gap by Humanizing Type 2…	A2
Angelides	Development and Use of a ¹³C6-Glucose Isotope-Dilution Liquid Chromatography…	A3
Bailey	Predictors of Nocturnal Hypoglycemia in Patients with Type 1 Diabetes: Effects of…	A4
Banerjee	Accurate Prediction and Avoidance of Hypoglycemia in an Artificial Pancreas Using…	A5
Barclay	A Synthetic Fusion Protein for Epigenetic Control of Pancreatic Cell Function	A6
Bashan	Service Evaluation of the Diabetes Insulin Guidance Service (DIGS)	A7
Basu	Acetaminophen Interference with Subcutaneous Glucose Sensing	A8
Ben Aharon	Acoustic Methods for Transmitting Blood Glucose Measurements	A9
Benesch	Improved Automated Glucose Clamps Using Algorithm Variations with the ClampArt	A10
Bitton	Effect of Local Warming on Regular Human Insulin Pharmacokinetics	A11
Block	Early Results Suggest Effectiveness of an Online Diabetes Prevention Program	A12
Bode	Comparison of Insulin Pump Therapy and Multiple Daily Injection Therapy in…	A13
Botti	Buccal Delivery of the Glucagon-Like Peptide Agonist Exenatide Using ArisCrown…	A14
Brahim	Designing a Decision Support System to Prevent Postexercise Hypoglycemia in Type…	A15
Brey	A Novel Imaging Techniques for Monitoring Stability of Encapsulated Islets	A16
Cahn	Correlation of Lens Autofluorescence and Signs of Diabetic Retinopathy in Wide…	A17
Cahn	Comparison of Distribution of Lens Autofluorescence in Patient Populations from…	A18
Cameron	Predicting Sleep and Meals for Improved Artificial Pancreas Performance	A19
Card	Influence of Glucose Set Point and Premeal Bolus on Proportional-Integral-Derivative…	A20
Clements	Leveraging Glooko for Real-Time Glucometer Data Sharing Between Patients with…	A21
Collier	The SMART Pilot Project: The Utility of a Tex-Messaging-Based Intervention for…	A22
Dagliati	Cardiovascular-Risk-Associated Qualitative Pathways in Type 2 Diabetes Patients	A23
Davis	Pilot Baseline Results from DeviceChoices.com: A Diabetes Device Patient Registry	A24
Dehais	A Smartphone-Based Carbohydrate Estimation System Using Computer Vision…	A25
Dickson	Gut Absorption Rates for Glucose in Term and Premature Neonatal Intensive Care…	A26
Dierick	Assessing the Relationship between Patient Compliance to Blood Glucose Monitoring…	A27
Driscoll	White Coat Adherence in Pediatric Insulin Pump Users	A28
Dutil	Adoption of Insulin Pumps and Continuous Glucose Monitors: Patient Perceptions of…	A29
Edoh	Antihyperglycemic Effects of γ-Hydroxy-δ-Valerolactone Isolated from Ethanolic…	A30
Edoh	Computerization of Patient Records at a Public Herbal Clinic in Ghana	A31
El Youssef	Inclusion of Glucagon Dynamics into a Model of Glucose Homeostasis	A32
Fabris	Are Risk Indices Derived from Continuous Glucose Monitoring Interchangeable with…	A33
Freckmann	Analytical and User Performance Evaluation of a Blood Glucose Monitoring System…	A34
Freckmann	Low Pain Perception across Two Generations of Infusion Se Systems for Continuous…	A35
Gainey	Glucosense: A Low-Cost Glucometer System for Resource-Poor Settings	A36
Gal	A Novel Noninvasive Glucose Monitor for Home Use: Assessing the Learning Curve…	A37
Gal	Enabling Frequent Blood Glucose Monitoring at Home Using a Truly Noninvasive…	A38
González-Pérez	MDiasite Smart Website for Diabetes Education	A39
Gu	Transcutaneous Glucose Sensor Performance in a Diabetic Animal Mode	A40
Haluzik	Continuous Exenatide Infusion Reduces Glycemic Variability and Improves Glucose…	A41
Henderson	Glycemic Variability and Glycemic Goals in a Pilot of Young Children with Type 1…	A42
Henderson	Use of the eGlycemic Management System by Glytec Provides Safe and Effective…	A43
Herrero	Bolus Calculator Autotuning	A44
Hinnen	Use of Diabetes Data Management Software Reports by Health Care Providers…	A45
Hinnen	Use of Diabetes Data Management Software Reports by Patients with Diabetes and…	A46
Hoi	Detection of Peripheral Arterial Disease in Diabetic Patients Using Diffuse Optical…	A47
Horowitz	The Diabetes Snapshot: Using the Electronic Medical Record to Improve Glycemic…	A48
Huyett	Application of Fuzzy Anti-Reset Windup for Proportional-Integral-Derivative Control…	A49
Jiang	Insulin Sensitivity Changes Assessed by Continuous Glucose Monitoring and Insulin…	A50
Joseph	Noninvasive Continuous Glucose Monitoring in Postsurgical Diabetic and Nondiabetic…	A51
Kaineder	A Novel Focal Drug Delivery System to the Hypothalamus to Assess the Neuronal…	A52
Kaplan	Design of the ENDO Trial: A Randomized, Sham-Controlled, Double-Blind Trial of…	A53
Kastellorizios	A Semi-Implantable Biosensor for Continuous Lactate Monitoring	A54
Kay	A U.K. Patient’s Perspective on Graphical Display of Glycemic Control Data	A55
Kidron	Dose-dependent Rise in Plasma Insulin Levels Following Preprandial Treatment of…	A56
Kircher	Improved Predication Accuracy for Dynamic Bayesian Network Glucoregulatory…	A57
Klonoff	Results from an Interview-Based Survey of Gla-300 SoloSTAR Compared with Three…	A58
Knab	Sensor Fusion and Zone Model Predictive Controller for Targeted Glucose Control in…	A59
Kordi Yoosefinejad	Effects of the Application of Whole-Body Vibration on the Balance of Type 2…	A60
La Belle	Multimarker Electrochemical Sensor for Artificial Pancreas	A61
Lande	Practical Lifelong A/V Access for a Forearm-Worn, Closed-Loop, Blood-Based…	A62
Larson	Artificial Pancreas Database	A63
Lee	A Novel Run-to-Run Algorithm for Enhanced Continuous Glucose Monitor Accuracy…	A64
Li	Microfabricated Microporous Membranes Prolong the Functional Lifetime of a Closed…	A65
Li	Photo-Patternable Hydrogels for Implantable Glucose Sensors	A66
Link	Measurement Accuracy of Two Blood Glucose Moniroting Systems with Built-In…	A67
Lipman	Technology-Enabled Type 1 Diabetes Education and Support System	A68
Lucisano	Patient Preference Assessment of a Long-Term Fully Implantable Continuous Glucose…	A69
Lv	Introducing an Exercise Model for the Artificial Pancreas	A70
Mader	Glycemic Control and Risk of Hypoglycemia in Hospitalized Patients with Type 2…	A71
Maria Isabel	Exploring Continuous Glucose Monitoring on the Frequency Domain to Identify Risk…	A72
Marini	A Multivariate Data-Driven Model to Investigate the Arising of Complications in Type…	A73
Marvin	GlucoCare-140: A Target, Not a Target Range, for Glycemic Control in the Intensive…	A74
Mauseth	Smart Low Glucose Suspend Capabilities of the Dose Safety Fuzzy Logic Dosing…	A75
Mehta	Telephone-Based Insulin Management Program	A76
Meo	Inhaled Insulin: Recovered from Deep Coma	A77
Montgomery	MyDiaText: Feasibility of a Text Messaging System for Youth with Type 1 Diabetes	A78
Mucci	Measurement of Glucose Concentrations in Saliva via Infrared Spectroscopy	A79
Muchmore	Continuous Glucose Monitoring Greatly Improves Sensitivity for Detection of…	A80
Neubauer	Health Care Professionals’ Adherence and Satisfaction Using the GlucoTab, a…	A81
Nnamoko	Evaluation of a Fuzzy Inference Model for Continuous Regimen Alterations in Type 2…	A82
Norrie	Reducing Risk with Continuous Glucose Monitoring in Intensive Care Unit Patients	A83
Osokina	Efficacy of Continuous Glucose Monitoring in Children with Type 1 Diabetes	A84
Pardo	Using Radar Plots to Evaluate Five Blood Glucose Monitoring Systems for Accuracy…	A85
Pardo	Evaluating the Accuracy and Precision of Six Blood Glucose Monitoring Systems…	A86
Patwardhan	CardioChek Plus: A Wireless Point-of-Care Analyzer Enabling Rapid, Simultaneous…	A87
Pflug	Comparative Accuracy of Insulin Dosing Based on Results from Two Clinical Trials…	A88
Pflug	Comparative Accuracy of Insulin Dosing Based on Results from Five Blood Glucose…	A89
Pfützner	Preliminary Results of a Crossover Study on Usage Time for Insulin Pump Infusion…	A90
Pfützner	Use of InsuPad in Daily Practice in Patients with Type 1 Diabetes: Post Hoc Analysis…	A91
Pfützner	Glycemic Variability Is Associated with Frequency of Blood Glucose Testing and…	A92
Pfützner	Long-Term Use of InsuPad Is Associated with High Treatment Adherence and Results…	A93
Prestrelski	Clinical Development of XeriSol Glucagon: A Stable, Nonaqueous Liquid Glucagon…	A94
Rankin	Illustration of Potential Cost-Saving Implication of Lower-Extremity Nerve…	A95
Rastogi	Improved Glycemic Management with an Implantable Continuous Monitoring System…	A96
Rebec	iSense Mastery Utility Demonstrated in 7-Day Clinical Study	A97
Resalat	Integrating Exercise Compensation into a Bihormonal Artificial Pancreas System	A98
Robbins	The I-KAN Study: Internet Initiation of Insulin for Type 2 Diabetes in Kansas	A99
Rodbard	Graphical Display of Quality of Glycemic Control: New Methods	A100
Rollins	A Continuous Glucose Estimation System on a Smart Phone	A101
Roy	Evaluating Safety and Efficacy of a Closed-Loop Glucose Control System with Biased…	A102
Ruiz	A Method for Safe Real-Time Retuning of General Artificial Pancreas Systems	A103
Sabrina	Effect of Diary Milk Powder as a Source of Vitamin D on Total Cholesterol Levels in…	A104
Sakai	Engineering of Fungi-Derived Flavin Adenine Dinucleotide Glucose Dehydrogenase…	A105
Scheiner	Remote Education and Management of Patients with Diabetes	A106
Schiavon	Modeling the Effect of Physical Activity on Postprandial Glucose Turnover in Healthy…	A107
Scholtes	A Novel Approach toward Noninvasive Glucose Monitoring	A108
Schunk	Simulating the Effects of Lifestyle choices on Blood Glucose Control: Implications of…	A109
Segagni	Temporal Association Rules for Stratification of Type 2 Diabetes	A110
Sekimoto	In Vitro/In Vivo Characterization of the Continuous Glucose Monitoring Sensor Using…	A111
Sieber	Evaluation of a Methodology for Estimating HbA1c Values by a New Glucose Meter	A112
Sierra	Hypoglycemia Reduction with Threshold Suspend and Its Economic Impact in Insulin…	A113
Signal	Intermittent Versus Continuous Glucose Monitoring to Detect Hypoglycemia in…	A114
Sink II	Clinical Experience with the V-Go Disposable Insulin Delivery Device	A115
Slage	The Use of Big Data Online Crowdsourcing to Improve Outcome and Reduce…	A116
Sølvik	Performance of Systems for Self-Monitoring of Blood Glucose Compared with the…	A117
Stahl	Individualized Rapid-Acting Insulin Action Estimation from Insulin Pump and…	A118
Standard	A Device to Support the Use of a Glucose Monitor by Individuals Who Are Blind or…	A119
Stern	Checking In: Feasibility of a Physician-Delivered Intervention to Improve Blood…	A120
Stewart	Stochastic Model Predictive Glycemic Control for the Intensive Care Unit	A121
Suminaka	Screening for Glucose Intolerance Using Area under the Curve after Glucose Loading	A122
Tauschmann	Factors Influencing Overnight Closed-Loop Performance during Free Living in…	A123
Tella	Antihyperglycemic Potential of Psidium guajava Leaf in Streptozotocin-Induced…	A124
Thomas	Accuracy and Precision of Three Common Glucose Meters in Intensive Care Unit	A125
Tomlinson	Glycemic and Nutrition Delivery: Performance of the Stochastic Targeted Protocol	A126
Turksoy	Detection of Continuous Glucose Monitor Faults for Patients with Type 1 Diabetes	A127
Unruh	Evaluation of Luminescent Copolymer Hydrogels as Fully Implantable Sensors	A128
v.Lilienfeld-Toal	Noninvasive Glucose Measurement at the Finger with mid-Infrared Spectroscopy and…	A129
Vehi	A Fuzzy Adaptive Strategy for Blood Glucose Controllers	A130
Virdi	Physician Acceptance of an Innovative Digital Feedback System Depends on Their…	A131
Viswanathan	Use of iPro2 in Real-Life Management of Type 2 Diabetes Patients in India	A132
Wang	Artificial Pancreas: A Novel Coping Strategy for Big Dietary Ingestion	A133
Wang	Assessing Diabetic Foot Ulcer Healing at Wound Clinics: Development of a Tracking…	A134
Wang	Benefit of an Implantable Continuous Glucose Monitoring System for Nocturnal…	A135
Yanping	Effect of Dumpling Cooking Method on Postprandial Glucose Level	A136
Zhang	Clinical Accuracy Evaluation of San MediTech’s Real-Time Continuous Glucose…	A137
Zhao	Online Hyper/Hypoglycemia Alert Based on Adaptive Kalman Filter for Type 1 Diabetes	A138
Zijlstra	A New Three-Step Clamp Method for the Evaluation of Blood Glucose Meters	A139
Zijlstra	Blood Glucose Meter Performance: A Comparison between Two Test Procedures	A140

Open in a new tab

Dose Accuracy of Reusable Insulin Pens in India: NovoPen 4, NovoPen 3, HumaPen Ergo, Glaritus Pen Royale, INSUPen, and AllStar

Mona Abdel-Tawab, PhD; Ingo Fritz, PhD; Volker Korger, PhD

Central Laboratory of German Pharmacists

Eschborn, Hessen, Germany

m.tawab@zentrallabor.com

Objective:

Currently several reusable insulin pens from different manufacturers are available in India and in some other emerging countries in Asia and Latin America. We assessed the dose accuracy of the insulin pens NovoPen 4 (NP4) and NovoPen 3 (NP3) for NovoMix, HumaPen Ergo (HPE) for Humulin, Glaritus Pen Royale (WR) for Glaritus, INSUPen (IP) for Basalog, and AllStar (AS) for Lantus.

Method:

In total, 15 pens of each type were used for the 4-fold, nonrandomized delivery of low (5 units), midpoint (30 units), and high (60 units) dosage levels, respectively, generating 180 values per pen type. Dose accuracy was evaluated gravimetrically as recommended by DIN EN ISO 11608-1:2000.

Result:

All tested insulin pens delivered doses within the International Organization for Standardization limits at all dosage levels. However, WR and IP revealed decreased dose-accuracy performance at all dosage levels, reflected in higher coefficients of variation (6.5 and 4.6%, respectively, at the 5-unit level, compared with 3.6% for AS, 3.5% for HPE/NP3, and 3.3% for NP4). The mechanical last-dose-stop feature of NP4, NP3, IP, and AS, preventing the dialed dose from exceeding the remaining content and providing greater patient safety, is not available in HPE and WR.

Conclusion:

While all pens delivered doses within ISO limits, slight dose-accuracy variations, as well as last-dose-stop features, confer notable differences between some pens.

Duality of Interest:

This study was supported by Sanofi.

From Bench to Bedside: Conquering the Translational Gap by Humanizing Type 2 Diabetes Research

Zeeshan Ali, PhD; P. Charukeshi Chandrasekera, PhD; Neal D. Barnard MD

Physicians Committee for Responsible Medicine Washington DC

zali@pcrm.org

Background:

Obesity and type 2 diabetes (T2D) have reached epidemic proportions worldwide, and thus considerable research effort has been dedicated to developing strategies to understand and treat these complex diseases. The two hallmark features of T2D, insulin resistance and pancreatic dysfunction, have been studied extensively using various animal models. Despite the wealth of knowledge hitherto acquired from animal models, many details of human disease pathogenesis remain unknown, and pharmacotherapeutic options for humans remain limited, with adverse effects associated with many widely used T2D drugs. From gene regulation to pancreatic cytoarchitecture to glucose sensing and transport to regulation of insulin secretion, emerging human data have raised concern regarding the immutable species differences and the subsequent limitations to translatability.

Method:

This study addresses the challenges and opportunities in T2D research and how it can be transformed with existing and novel future methodologies using human-based data acquisition.

Results:

We discuss how state-of-the-art in vitro, in vivo, and in silico technologies can be used to delineate disease mechanisms and drug responses. Insulin resistance and pancreatic dysfunction can be studied at many levels from gene transcription/regulation to functional proteomics to cellular signaling and cell-cell interactions to organ cultures and whole-body glucose regulation to human population–based studies.

Conclusion:

T2D research is in dire need of a paradigm shift, with increased emphasis on human-based data acquisition and decreased reliance on animal models. With continued support from scientists and funding agencies alike, this is the clear-cut path to address the challenges in the translational barrier and create opportunities for diabetes research and treatment.

Development and Use of a ¹³C6-Glucose Isotope-Dilution Liquid Chromatography–Mass Spectrometry Method to Determine System Accuracy and Precision of the In Touch Cellular Glucose Meter System

A Synthetic Fusion Protein for Epigenetic Control of Pancreatic Cell Function

David K. Barclay; Karmella A. Haynes, PhD

School of Biological and Health Systems Engineering

Arizona State University

Tempe, Arizona

david.barclay@asu.edu

Objective:

Type 1 diabetes is caused by the failure of beta cells to produce sufficient insulin in pancreatic tissue. Many recent treatment strategies have focused on replenishing functional beta cells. We focus on a strategy to manipulate the DNA-protein complex known as chromatin to transdifferentiate glucagon-producing alpha cells into insulin-producing beta cells. Chromatin modifications store a secondary level of information that determines the expression states of developmental genes that control cell fates.

Method:

Others have demonstrated that indirectly disrupting gene-silencing chromatin induces a beta-cell gene expression profile in alpha cells. In order to directly target chromatin-silenced genes for specific activation, we used a synthetic fusion protein called PcTF, which binds to the histone H3 lysine 27 trimethylation mark (H3K27me3). Previously, we have demonstrated that the PcTF protein reactivates epigenetically silenced target genes in human bone-derived cells. We aim to use PcTF to regulate genes in pancreatic cells in order to advance the treatment of diabetes.

Result:

We transfected mouse TC1 alpha cells with PcTF-expressing plasmid DNA and confirmed expression by quantitative reverse-transcription PCR (qRT-PCR). Thus far, we have observed upregulation of H3K27me3-associated genes MAFA and IAPP via qRT-PCR in PcTF-expressing cells, compared with nonexpressing cells. Our preliminary bioinformatics analysis has identified ~4,950 potential PcTF target genes in human alpha cells through the analysis of chromatin immunoprecipitation data.

Conclusion:

Taken together, our proof-of-concept results from mouse cells and bioinformatics analysis of human chromatin suggest the translational potential of the synthetic PcTF protein. PcTF could be an effective agent for the epigenetic conversion of human alpha cells to insulin-producing cells.

Service Evaluation of the Diabetes Insulin Guidance Service (DIGS)

Eran Bashan, PhD; Israel Hodish, MD, PhD

Hygieia Inc.

Ann Arbor, Michigan

bashan@hygieia.com

Objective:

Insulin therapy has been available for almost a century. However, its success rate is still disappointing where the majority of users sustain destructively elevated A1C. The key element essential for effective and safe insulin therapy, even for simple regimens like basal or premixed insulin, is frequent dosage titration to overcome constant variations in insulin requirements. In reality, dosage titration is done sporadically during clinic visits. Hygieia Inc. has developed a scalable solution to this problem. A diabetes-nurse service improves glycemic control without overburdening the health system. The service relies on a device called d-Nav, which provides patients with an insulin dose recommendation for each injection while using the device to monitor glucose. By analyzing stored glucose trends, in a similar approach that providers use during clinical encounters, d-Nav titrates insulin dosage without care providers’ constant supervision. A service evaluation was conducted at the Diabetes Consult Ward, Ulster Hospital, Dundonald, Belfast, Northern Ireland.

Method:

Patients aged ≥21 years with A1C ≥7.0% (53 mmol/mol) who were receiving insulin therapy for at least 1 year were recruited to the service and were followed for 1 year. A1C and information on hypoglycemia were collected during scheduled 3-monthly clinic visits.

Results:

A total of 126 patients were recruited, while 96 completed the evaluation. The mean (±SD) A1C for the 96 active users decreased from 9.2 ± 1.4% (77 ± 15 mmol/mol) at baseline to 7.8 ± 1.2% (62 ± 13 mmol/mol) at the 3–5-month clinic visit and to 7.5 ± 1.2% (58 ± 13 mmol/mol) at the 6–12-month clinic visit (P < 0.05). The frequency of minor hypoglycemia (glucose <3.7 mmol/L or <65 mg/dl) was 0.6 ± 0.2 events per week.

Conclusion:

The Diabetes Insulin Guidance Service (DIGS) was shown to be an effective, scalable solution for the management of insulin-treated patients without overburdening the health care system.

Acetaminophen Interference with Subcutaneous Glucose Sensing

Designing a Decision Support System to Prevent Postexercise Hypoglycemia in Type 1 Diabetes

Influence of Glucose Set Point and Premeal Bolus on Proportional-Integral-Derivative–Based Closed-Loop System Performance

Casey Card, BA; Jennifer L. Sherr, MD, PhD; Eda Cengiz, MD, MHS; Camille I. Michaud, MD; Neha Patel, DO; Miladys M. Palau-Collazo, MD; Lori Carria, BS; Eileen Tichy, MS; William V. Tamborlane, MD; Stuart A. Weinzimer, MD

Department of Pediatrics

Yale University

New Haven, Connecticut

c.card2@nuigalway.ie

Objective:

Proportional-integral-derivative (PID)–based algorithms are employed in the design of closed-loop (CL) systems, but the optimal glucose set point and need for manual premeal boluses (PMBs) have not been established. We analyzed glucose profiles from five CL protocols using a PID controller with insulin feedback to explore these questions.

Method:

Data for 41 subjects (23 female; age 19 ± 4 years; A1C 7.3 ± 0.7%) were pooled to assess the effect of glucose set point (100 mg/dL [G100] vs. 120 mg/dL [G120]) and use of PMB (PMB+ or PMB–) on CL system performance. Mean blood glucose (BG); BG levels within (70–180 mg/dL), below (<70 mg/dL), or above (>180 mg/dL) target range; prandial BG excursion and prandial BG area under the curve; and hypoglycemia (BG <60 mg/dL) and hyperglycemia (BG >250 mg/dL) were calculated for each condition.

Result:

G100 subjects had a significantly lower average BG than G120 (132 ± 33 vs. 146 ± 33 mg/dL; P < 0.01). G100 subjects spent significantly more time in the target range (83.8 ± 9.8 vs. 74.9 ± 10.5%; P < 0.02) and less time above target range (13.0 ± 9.6 vs. 23.7 ± 11.4%; P < 0.01) than G120 subjects, with no significant increase in time below range (3.2 ± 3.5 vs. 1.4 ± 2.4%; P = 0.07) or hypoglycemia (1.1 vs. 0.3%; P = 0.11). Time in hyperglycemia was not different (1.7% G100 vs. 3.6% G120). PMB+ subjects had smaller prandial BG excursion

and BG area under the curve than PMB– subjects (94 vs. 105 mg/dL and 207.4 vs. 254.2 mg⋅h/dL, respectively), but these differences did not reach statistical significance.

Conclusion:

Using a sensor glucose set point of 100 mg/dL rather than 120 mg/dL in a PID-based CL system results in a lower average BG and improved time in target range without a significant increase in hypoglycemia. The use of a small PMB did not equivocally demonstrate benefit in mitigating prandial glycemic excursions, although further studies are needed to definitively answer this question.

Leveraging Glooko for Real-Time Glucometer Data Sharing Between Patients with Type 1 Diabetes and Their Providers: A “Smart” System

Mark A. Clements, MD, PhD, CPI, FAAP

Children’s Mercy Hospital

Kansas City, Missouri

maclements@cmh.edu

Objective:

We aim is to improve data sharing and communication between patients with diabetes and their health care providers.

Method:

We leverage the Glooko web service to stream glucometer data into the Cerner Electronic Health Record. Provider-defined alerts are created to notify the diabetes care team when a clinically actionable pattern emerges in the blood glucose data stream. Alerts based on blood glucose data trends are created via event processing engines. Such engines permit the identification of complex trends in “in-flight” data without creating transactional burden on medical record databases.

Result:

Multiple custom alerts are demonstrated, including alerts for postprandial hyperglycemia at breakfast, lunch, and dinner; decreases in the frequency of blood glucose monitoring; mild to moderate hypoglycemia occurring repeatedly within the same 3-h window; and increasing average daily glucose. The number of days over which the event of interest must occur to create an alert can be customized by provider or by patient. For instance, an alert can be generated if postbreakfast blood glucose is persistently 40 mg/dL greater than the preprandial level for 4 consecutive days. Alternately, the percentage change in a specific metric can be used to trigger an alert (i.e., a 20% increase in the average daily glucose observed over a 14-day period compared with the previous 14-day period).

Conclusion:

Glooko is an enabling technology that supports real-time sharing of glucometer data between patients and health care providers. Health care providers can provide “value added” to Glooko by delivering Glooko data into the electronic health record and leveraging event processing engines to alert the diabetes care team to clinically actionable changes in the quality of glycemic control.

The SMART Pilot Project: The Utility of a Text-Messaging-Based Intervention for Adolescents with Type 1 Diabetes

Suzanne Collier, BA; Alexa Stern, BA; Catherine Gillespie, PhD; Linda Herbert, PhD; Celia Henderson, RN, CDE; Randi Streisand, PhD

Children’s National Medical Center

Washington DC

scollier@cnmc.org

Objective:

Type 1 diabetes (T1D) is a lifelong condition that is particularly challenging during adolescence. Most adolescents use cell phones; thus mobile health interventions may be an effective method of T1D education and adherence promotion. The aims of this study are 1) to examine the utility of a text-message-based intervention program to promote self-care skills and 2) to determine which adolescents are most likely to benefit from the intervention.

Method:

Twenty-three participants (mean age = 15.13 years; range = 13–17 years) were enrolled in the SMART Project, a 6-week text-message program during which adolescents received blood glucose prompts and T1D education texts 2–3 times/day. Adolescents completed a demographic/medical history questionnaire, the Diabetes Behavior Rating Scale, the Self-Care Inventory, and the International Physical Activity Questionnaire at baseline and upon completion of the intervention. Glucometer data regarding the past 30 days were downloaded, and medical charts were reviewed for HbA1c from the clinic appointments immediately preceding and following the intervention.

Result:

Ninety-six percent of participants completed the follow-up questionnaires. Potentially due to a limited sample size, there were no statistically significant differences in any of the measures between the two time points. However, the intervention did appear to be useful among specific subgroups of adolescents, such as those who reported better self-care and physical activity levels at baseline, such that they had improved blood glucose and HbA1c levels after the intervention.

Conclusion:

A text -based intervention appears feasible for teens with T1D, yet more research is needed to understand for whom such interventions would be most beneficial. Mobile health interventions may be a useful method to provide health reminders and education to certain subgroups of adolescents with T1D.

Cardiovascular-Risk-Associated Qualitative Pathways in Type 2 Diabetes Patients

Arianna Dagliati, MS; Lucia Sacchi, PhD; Riccardo Bellazzi, PhD

The feasibility study in a laboratory environment using GoCARB and real meals was promising. Currently, the system is used by adult individuals with type 1 diabetes at the Bern University Hospital within the framework of a preclinical evaluation trial.

Gut Absorption Rates for Glucose in Term and Premature Neonatal Intensive Care Unit Infants Using Continuous Glucose Monitor Data

Inclusion of Glucagon Dynamics into a Model of Glucose Homeostasis

Joseph El Youssef, MD, MCR; Jessica Castle, MD; Peter Jacobs, PhD; Navid Resalat, MS; W. Kenneth Ward, MD

Oregon Health and Science University

Portland, Oregon

elyoussj@ohsu.edu

Objective:

Many currently used models of glucose dynamics were developed with insulin action in mind, and only recently has glucagon action been considered for inclusion in such models. We describe the updating of an accepted glucose homeostasis model with absorption and action components for glucagon for use in a bihormonal artificial pancreas system, with identification of parameters for the model.

Method:

Based on data obtained from a recent physiology study, we identify six individual models of glucagon action on glucose homeostasis for identification. Using Matlab and Simulink, these models were devised and are being prepared for testing.

Result:

High levels of plasma insulin reduce the effect of glucagon action to elevate blood glucose. However, whether this reduction can be overcome by a sufficiently high dose of glucagon could not be determined. Therefore, two models of glucagon action (immediate and delayed actions) were developed with a threshold insulin level above which glucagon effect is nullified, while two were built with the ability of glucagon to override insulin. The final two use a derivative effect of glucagon on endogenous production.

Conclusion:

Identification and validation of the most appropriate model of glucagon action allows us to update a previously validated model of glucose homeostasis for use in a bihormonal artificial pancreas controller.

Are Risk Indices Derived from Continuous Glucose Monitoring Interchangeable with Self-Monitoring of Blood Glucose–Based Indices?

Chiara Fabris, MSc; Stephen Patek, PhD; Marc Breton, PhD

Center for Diabetes Technology

University of Virginia

Charlottesville, Virginia

chiara.fabris@hotmail.it

Objective:

High blood glucose index (HBGI) and low blood glucose index (LBGI) are popular metrics used to quantify the risk of hyper - and hypoglycemia from sparse self-monitoring of blood glucose (SMBG) data and have been shown to be predictive of future glycemic events, e.g., LBGI of severe hypoglycemia. Continuous glucose monitoring (CGM) systems, however, allow almost continuous monitoring of blood glucose, providing a large amount of data. Using indices such as LBGI and HBGI, and their cutoff values, on CGM data stream has never been validated. The aim of this work is to model the relationship between HBGI/LBGI values obtained from SMBG and CGM, providing a correction to apply when these indices are computed from CGM time series.

Method:

Twenty-eight subjects with type 1 diabetes were monitored for up to 4 weeks with both SMBG and CGM systems. CGM time series were collected using the Dexcom G4 Platinum. HBGI and LBGI were calculated from SMBG and CGM, and a number of models were used to describe the relationship between them.

Result:

A linear model with unitary slope and no intercept is reliable for HBGI. Thus no correction is needed to compute this index from CGM time series. For LBGI, values obtained from CGM did not closely match SMBG-based values, with clear underestimation at low values; a nonlinear transform without saturation performed best to match CGM-based LBGI to SMBG-based LBGI.

Conclusion:

We propose alternate versions of the well-known HBGI and LBGI indices that are adapted to the characteristics of CGM time series. Using such transformations, we enable references to previous studies defining clinically relevant cutoffs for HBGI and LBGI.

Analytical and User Performance Evaluation of a Blood Glucose Monitoring System Following ISO 15197:2013

MDiasite Smart Website for Diabetes Education

Salvador González-Pérez, PhD; Javier Fernández-Cañete, PhD; Elena Alenina, PhD; Victor Grigorenko, PhD

Department of System Engineering and Automation

University of Malaga

Malaga, Spain

sgp@isa.uma.es

Objective:

MDiasite is an update of a web educational platform (Diasite) for course management system (CMS), a free, open-source software package designed using sound pedagogical principles to help diabetes educators and health care professionals in order to create effective online learning communities. It is an initiative that involves public and private partners to improve the treatment and outcomes for people with diabetes, to promote early diagnosis, and to prevent or delay the onset of diabetes.

Method:

This SW is based in the Moodle platform as open-source software under the GNU General Public License and has the following features: promotes a social constructionist pedagogy (e.g., collaboration, activities, critical reflection); suitable for 100% online classes as well as supplementing face-to-face learning; simple, lightweight, efficient, compatible, low-tech browser interface; and easy to install on almost any platform that supports PHP. Nowadays, mobile devices are very important for patients, so this update is related to use MDiasite on mobile devices: users can open MDiasite on their mobile web browsers and can download native applications for their mobile devices. MDiasite is mobile accessible through server extensions in IOS, Android, Windows Mobile, and Blackberry.

Result:

This platform can show all descriptions for every course on the server, including accessibility to guests. Contents can be edited using an embedded WYSIWYG HTML editor. Goals are to reduce administrator involvement to a minimum, while retaining high security. It supports a range of authentication mechanisms through plug-in authentication modules, allowing easy integration with existing systems. It offers standard e-mail methods where patients can create their own login accounts. For security, educators can add an “enrollment key” to their courses to keep out nonpatients. They can give out this key face-to-face or via personal e-mail. Educators can enroll and unenroll patients manually if desired even after a certain period of inactivity. Patients are encouraged to build an online profile. Personal data can be protected from display if required. Every user can specify their own time zone. Every user can choose the language used for the MDiasite interface. The modules include a chat module, choice module, forum module, quiz module, resource module, survey module, and workshop module.

Conclusion:

Use of this SW improves the range of diabetes educators, bringing content to mobile devices so that patients can take courses from anywhere.

Transcutaneous Glucose Sensor Performance in a Diabetic Animal Model

Bing Gu, MS; Sagar Vaddiraju, PhD; Fotios Papadimitrakopoulos, PhD; Diane J. Burgess, PhD

Department of Pharmaceutical Sciences

University of Connecticut

Storrs, Connecticut

bing.gu@uconn.edu

Objective:

The purpose of this work is to test amperometric glucose biosensor performance in a diabetic animal model.

Method:

Each streptozotocin-induced diabetic rat was transcutaneously implanted with one polyvinyl alcohol hydrogel/poly(lactic- co-glycolic acid) microsphere composite–coated amperometric glucose sensor. Six units/kg insulin was injected intravenously via the tail vein to induce hypoglycemia. Meanwhile, the blood glucose concentration was determined periodically. Two dextrose injections were then performed intraperitoneally under hypoglycemic conditions to test the sensor response. In order to test the effect of our composite coating (designed to prevent inflammation and fibrosis while still allowing glucose diffusion toward the sensor) on glucose diffusion, polyvinyl alcohol hydrogel/poly(lactic-co-glycolic acid) microsphere composite–coated microdialysis probes were also implanted into diabetic rats. Relative recovery of glucose from the microdialysis probe was determined at different days.

Result:

The rats were induced to be diabetic with blood glucose concentrations of ~400 mg/dL. Amperometric sensor output closely followed the trend of the blood glucose concentration from hyperglycemic to hypoglycemic region. Under the hypoglycemic conditions, the glucose sensor was able to differentiate minor glucose concentration changes within a short time period. This indicates the amperometric glucose sensor has a fast response time and good sensitivity. The relative recovery of glucose from composite-coated microdialysis probes can be maintained at ~12% for up to 1 week.

Conclusions:

The amperometric glucose biosensor can perform at both hyper- and hypoglycemic conditions in diabetic animals. The sensor has a fast response time and good sensitivity. The composite coating was able to inhibit the foreign body reaction to transcutaneously implanted microdialysis probes over a 1-week test period.

Funding:

This study was supported by U.S. Army Medical Research (W81XWH0910711, W81XWH0710688), the National Institutes of Health (1R21HL09045801, R43EB011886, 9R01EB014586), and the National Science Foundation/Small Business Innovation Research (1046902, 1230148).

Continuous Exenatide Infusion Reduces Glycemic Variability and Improves Glucose Control in Cardiac Surgery Patients: The EXECUTIVE Trial

Martin Haluzik, MD, DSc; Michal Lipš, MD; Jana Drapalova, MSc; Milos Mraz, MD, PhD; Milos Dobias, MD; Petr Kopecky, MD; Jaroslav Lindner, MD, PhD; V. Burda, BSc; Daniel Novak, Ing, PhD; Michaela Diamant, MD, DSc; Stepan Svacina, MD, DSc

Third Department of Medicine

Department of Endocrinology and Metabolism

General University Hospital

Charles University in Prague

First School of Medicine

Prague, Czech Republic

mhalu@lf1.cuni.cz

Objective:

We performed a randomized control trial with the GLP-1 receptor agonist exenatide as add-on to standard perioperative insulin therapy in subjects undergoing elective cardiac surgery to intensify perioperative glucose control while minimizing the risk of hypoglycemia.

Method:

Insulin Sensitivity Changes Assessed by Continuous Glucose Monitoring and Insulin Pump System During Menstrual Cycles in Women with Type 1 Diabetes

Boyi Jiang, MA; Sue Brown, MD; Marc D. Breton, PhD

Center for Diabetes Technology

University of Virginia

Charlottesville, Virginia

bj6h@virginia.edu

Objective:

We quantify the insulin sensitivity (SI) during menstrual cycles in type 1 diabetic women using continuous glucose monitoring and pump data, targeting a description of pattern in SI changes.

Method:

We introduce a mathematical description of glucose homeostasis in log space based on the minimal model of glucose kinetics using feed-forward models for meal and insulin appearance. Using a Kalman filter, we are able to extract a physiologically sound estimate of SI. Twelve women with type 1 diabetes participated in a 3-month clinical trial during which they were asked to wear a Dexcom G4 sensor and track phases of their menstrual cycle (e.g., menstruation and ovulation). Self-monitoring of blood glucose, continuous glucose monitoring, and pump were recorded for all 3 months. For each subject, we computed the average nocturnal SI (1:00 A.M. to 6:00 A.M.) to minimize the impact of the model discrepancy on the Kalman filter performance around meals. Phases of the menstrual cycle were based on menstruation and ovulation for each cycle and confirmed a posteriori by blood draws. The SI in different cycle phases was then compared after being averaged across the three cycles.

Result:

The SI (e-4/min per mU/L) in the early follicular phase, the early-mid luteal phase, and the late luteal phase were 4.6 ± 0.5, 4.0 ± 0.9, and 4.0 ± 1. In the standard t test, early follicular phase versus early-mid luteal phase P = 0.05 and early follicular phase versus late luteal phase P = 0.03.

Conclusion:

The SI in the early-mid luteal phase and the late luteal phase deceases significantly compared with that in the early follicular phase in women with type 1 diabetes. The perimenstrual glycemic control is expected to be improved by applying the developed SI tracking technique.

Noninvasive Continuous Glucose Monitoring in Postsurgical Diabetic and Nondiabetic Intensive Care Unit Patients

Jeffrey I. Joseph, DO; Marc Torjman, PhD; John Reich, MD; Anthony Furnary, MD; Stanley A. Nasraway, MD; Mary McNamara-Cullinane, BA, RAC; Christine Olimpio, MS; David Olson, PhD; David Walton, MBA

Thomas Jefferson University Hospital

Philadelphia, Pennsylvania

jeffrey.joseph@jefferson.edu

Objective:

We assess the performance of a noninvasive, transdermal continuous glucose monitoring (CGM) system (Symphony CGM System, Echo Therapeutics, Philadelphia, PA) in the intensive care unit (ICU) setting.

Method:

Adult surgical patients with planned ICU admission of ≥24 hours at four medical centers were enrolled. A CGM sensor containing glucose oxidase was applied to a 6-mm diameter skin site that was prepared with controlled microabrasion using the Symphony SkinPrep device. Up to 30 blood samples from each patient were obtained from a radial artery catheter and measured using a YSI 2300 STAT Plus Glucose Analyzer as a reference. CGM was prospectively calibrated every 4 h. Safety was assessed by visual inspection of the site.

Result:

Thirty-two subjects completed the study. The study cohort was 19% female and 28% diabetes patients with a mean age of 65 ± 13 years. Overall mean absolute relative difference between CGM and reference blood glucose was 12.5% overall (12.2% for diabetic patients and 12.6% for nondiabetic patients). The post–warm-up, 8-h interval mean absolute relative differences were 13.2, 12.2, and 12.1%. Glucose values ranged from 49 to 324 mg/dL. No device/study-related adverse events were reported. Ninety-seven percent of subjects experienced no or mild irritation from the microabrasion skin preparation process, and 94% of subjects experienced no or mild irritation following the CGM sensor removal.

Conclusion:

Symphony CGM demonstrated clinically relevant accuracy consistently throughout the CGM session and excellent safety in diabetic and nondiabetic patients. Future studies are needed to determine whether Symphony CGM can be used to improve blood glucose control.

A Novel Focal Drug Delivery System to the Hypothalamus to Assess the Neuronal Role of GLP-1 Receptor in Energy Homeostasis

Katrin Kaineder, MSc; Thomas Birngruber, PhD; Thomas R. Pieber, MD

Center for Basic Medical Research

Department of Endocrinology

Graz, Styria, Austria

katrin.kaineder@medunigraz.at

Objective:

Obesity is a main risk factor in the development of type 2 diabetes. GLP- 1 analogs are currently used for the systemic treatment of type 2 diabetes, as they are known to reduce nutrient ingestion and body weight. We aim to investigate the pharmacological effects of systemic GLP-1 application versus GLP-1 application directly into the brain in terms of metabolic changes and GLP-1 receptor expression. Therefore, we developed a focal delivery system for drug administration directly into the brain, specifically into the hypothalamus of rats.

Method:

We assessed body weight changes in response to the implanted probe. The distribution pattern of the injected liquid was analyzed using dextran blue as a marker to verify precise application of high-molecular-weight substances into the hypothalamus. GLP-1 receptor distribution in the hypothalamus was examined via in situ hybridization and fluorescently labeled GLP-1 peptide. We assessed metabolic changes after systemic and intrahypothalamic GLP-1 administration by using the PhenoMaster cage system.

Result:

We found no changes in body weight in response to probe implantation and only minimal histological tissue reaction in the immediate vicinity of the probe. The results indicated GLP-1 receptor expression in the hypothalamus and also metabolic changes triggered by those. We examined difference between intrahypothalamic and intraperitoneal-administered GLP-1 in terms of energy homeostasis.

Conclusion:

We were able to successfully establish a focal delivery system in a preclinical pilot study. First PCR results are promising and will further be used to investigate precise signal transduction pathways triggered by intrahypothalamic GLP-1. Current experiments include the analysis of the signal transduction pattern with real-time PCR array analysis in situ after intrahypothalamic GLP-1 administration.

Design of the ENDO Trial: A Randomized, Sham-Controlled, Double -Blind Trial of the EndoBarrier for the Treatment of Type 2 Diabetes in Subjects with Obesity

Lee Kaplan, MD, PhD; John Buse, MD, PhD; Kenneth Malomo, BS; Keith Gersin, MD; Jennifer Cormier, BS; Steven Edmundowicz, MD; Louis Aronne, MD; Paul Visintainer, PhD; Lauren Baker, PhD; David Maggs, MD

Massachusetts General Hospital

Boston, Massachusetts

lmkaplan@partners.org

Objective:

Oramed Pharmaceuticals

Jerusalem, Israel

miriam@oramed.com

Objective:

We determine whether oral insulin (ORMD-0801) has a dose-dependent effect in type 1 diabetes mellitus (T1DM) patients.

Method:

ORMD-0801 was preprandially administered (t = −15 min) to 10 fasting T1DM patients, to final doses of 8, 16 (two 8-mg capsules), and 24 (one 8-mg and one 16-mg capsule) mg insulin, at three independent visits. Plasma glucose and insulin concentrations were monitored throughout the ensuing 5-h period. Pairwise comparisons of the responses to the various treatments were performed.

Result:

No adverse events were reported throughout the study period. Two patients were irresponsive to treatment and were excluded from all analyses. A dose-dependent response to ORMD-0801 was consistently observed, with significantly greater responses to the 24-mg dose. Mean insulin levels, Cmax, and area under the curve were, respectively, 47, 11.3, and 46 higher following 24-mg treatment when compared with the 8-mg dose (P ≤ 0.05) and, respectively, 35.9, 10.3, and 35.7% higher when compared with the 16-mg dose (P ≤ 0.01). In addition, the rate of decline in plasma insulin was slower following the 24-mg dose, with mean end-of-session insulin levels (1.6 ± 1.6 mU/mL) well above baseline values when compared with the other doses (−1.0 ± 1.1 and 0.7 ± 1.6 mU/mL, for 8 mg and 16 mg, respectively). Peak insulin levels were obtained within 105–120 min of dosing and coincided with peak glucose concentrations. In all three sessions, glucose concentrations gradually declined to baseline or near-baseline levels by the end of the monitoring period.

Conclusion:

The basic ORMD-0801 formulation provides for a dose-dependent bioavailability and bioefficacy of insulin preprandially delivered to T1DM patients.

Improved Prediction Accuracy for Dynamic Bayesian Network Glucoregulatory System Model

Robert Kircher, MS; Justin Lee, BS; Don Matheson, MS; Richard Mauseth, MD

Dose Safety Inc.

Redmond, Washington

bob@dosesafety.com

Objective:

We aim to improve the prediction accuracy of the statistical glucoregulatory system model. The prediction algorithm will be used within the hypoglycemia prevention module of the Dose Safety artificial pancreas controller.

Method:

We described the prediction algorithm and initial prediction accuracy results in a 2014 Advanced Technologies and Treatments for Diabetes poster. Changes to the algorithm included new blood glucose input signal-smoothing filter and revised feature vectors (FVs) for improved pattern matching of the glucose and insulin on board input signals and normalizing of the lambda and gamma model vectors to numerically bound the probability calculations. The blood glucose signal-smoothing filter uses a weighted average of the previous six glucose values to better reflect the trend of the glucose signal. The inherent delay caused by the filter is acceptable. The glucose and insulin on board 4-tuple (45 min) glucose FVs were revised to better conform to the rates and accelerations inherent in the input signals. The combined effects of the smoothing filter and the revised FVs produced improved FV matches within the perceptron-learning algorithm. The FV matching is reflected in the 1 × 15,000 real-valued lambda and gamma vectors representing the statistical model combining glucose and insulin dynamics. The improved FV matching caused the probability calculations to be unbounded, which were corrected by normalizing the values.

Result:

The revised algorithm produced mean prediction errors (standard deviation) for the 10 test subjects of 5.2% (4.8), 8.0% (6.9), and 10.7% (9.4) for predictions of 30, 45, and 60 min, respectively. This represents prediction accuracy improvements of 20.5, 22.4, and 21.5%, respectively, over the data previously presented.

Conclusion:

The prediction accuracy of the statistical glucoregulatory system model was substantially improved.

Results from an Interview-Based Survey of Gla-300 SoloSTAR Compared with Three Commercialized Disposable Insulin Pens

David Klonoff, MD; Irina Nayberg, RN, CDE, CDTC; Frank Erbstein, PhD; Anna MG Cali, MD, MSc; Claire Brulle-Wohlhueter, MD; Thomas Haak, MD

Diabetes Research Institute

Mills-Peninsula Health Services

San Mateo, California

dklonoff@diabetestechnology.org

Objective:

We evaluate perceptions of people with diabetes (users) and health care providers (trainers) on usability of Gla-300 SoloSTAR disposable insulin pen, compared with Lantus SoloSTAR, FlexPen, and KwikPen.

Method:

Tempe, Arizona

jeffrey.labelle@asu.edu

Objective:

Current diabetes management is limited, measuring only one marker, glucose, which can be monitored rapidly and continuously. No similar technology exists for detection of glucagon, insulin, or a biomarker panel. Current artificial pancreases are based on measurements of a system output (glucose) rather than inputs (insulin, glucagon). Incorporating all biomarker panels as parameters in nonlinear models of closed-loop insulin delivery systems would improve accuracy and precision.

Method:

Electrochemical impedance spectroscopy (EIS) requires the covalent immobilization to attach a molecular recognition element to the sensor. Then the appropriate target can be introduced in a concentration gradient while EIS is run. Data are analyzed to determine responsivity and degree of fit at each frequency, and the optimal frequency is selected per marker.

Result:

A variety of purified markers pertinent to diabetes have been studied (replicated) and pertinent parameters taken, namely, dynamic range, optimal binding frequency, and lower limits of detection. The current biomarkers under study include 1,5-anhydroglucitol, glucose, insulin, glycated albumin, and HbA1c at their physiological levels.

Conclusion:

Currently, a biomarker panel has been measured via EIS in purified form. Expectations include deepening knowledge of the dynamic behavior and feedback control of insulin (and its analogs) for glucagon-glucose-insulin regulation models and improving point-of-care test strips for multibiomarker diabetes management.

Practical Lifelong A/V Access for a Forearm-Worn, Closed-Loop, Blood-Based Artificial Pancreas

Arnold J. Lande, MD; John Erickson, MS

Northport Navigable Waters Institution

Northport, Michigan

alande718@gmail.com

Objective:

We reference several clinically proven components of a blood-based artificial pancreas.

Method:

1) The bedside, blood-based (not subcutaneous) Biostator has since 1976 (Food and Drug Administration approved in 1983) demonstrated the ability to place a clamp on blood glucose level. 2) When a gold standard hemodialysis arteriovenous fistula (AVF) fails to mature or a matured AVF fails, emergency obstruction of flow-and-pressure-diverting accessory veins is routinely performed to route the highest available blood flow through a single outflow vein. This demonstrated, efficacious response to an emergency predicts that an identical elective in situ debranched vein fistula graft (VFG) will result in lifelong (with revisions) durable blood access. 3) Intermittent external compressions, over the VFG and between the in and outflow cannulae, make available the A/V pressure differential and thus eliminate the need for bulky, dangerous blood pumps and their necessary safety adjuncts. Turbulent flow-through at the arteriovenous anastomosis is interrupted frequently, reducing tissue reaction. There is no downstream anastomosis, so no downstream scarring and constriction. 4) Like the spider-web -appearing AVF, which is widely accepted by hemodialysis patients, technicians, nurses, and physicians, a more cosmetic single outflow VFG is even more likely to be tolerated. The blood-contacting disposable, with medication cartridges manufactured in, is easily replaced twice weekly by the patient at home to avoid biofilm, clotting, thrombosis, and the looming threat of drug-resistant septicemia.

Result:

Lifelong access to extracorporeal blood flow and A/V pressure differential is desirable for a forearm-wearable artificial pancreas, kidney, liver, congestive heart failure aquapheresis, and a variety of other phereses.

Conclusion:

Lifelong vascular access is already available in the form of several clinically proven components. Immense personal and public benefits stand to accrue from the integration and implementation of these components.

Artificial Pancreas Database

Gaven Larson, BS; Lauren Huyett, BS; Joon Bok Lee, BS; Eyal Dassau, PhD; Frank Doyle, PhD

Microfabricated microporous membranes act as a passive barrier to cell migration to the immunogenic surface of a chemically driven closed-loop insulin delivery device, thereby improving implant biocompatibility and prolonging implant lifetime.

Photo-Patternable Hydrogels for Implantable Glucose Sensors

Zhe Li, MS; Liangliang Qiang, PhD; Sagar Vaddiraju, PhD; Diane J. Burgess, PhD; Fotios Papadimitrakopoulos, PhD

University of Connecticut

Storrs, Connecticut

papadim@mail.ims.uconn.edu

Objectives:

The function and lifetime of implantable electrochemical-based continuous glucose monitoring (CGM) devices are intimately linked to the stability of the glucose oxidase (GOx) enzyme responsible for glucose detection. Enzyme denaturation and leaching, as well as biofouling-induced pore clogging, gradually decrease device performance. Polyethylene glycol (PEG)-ylated hydrogels offer an opportune venue to minimize biofouling while retaining their highly hydrated state to prevent enzyme denaturation. Here we report the use of a novel PEGylated hydrogel, which is capable of being photocrosslinked in place in its fully hydrated state. This hydrogel allows for CGM sensors with long-term stability and optimal performance.

Methods:

A PEGylated copolymer was synthesized via a free-radical polymerization using an azobisisobutyronitrile initiator. The polymer was mixed with GOx enzyme and photopatterned on the electrode (via ultraviolet exposure) followed by 24-h incubation to allow chemical coupling of GOx to the PEGylated hydrogel.

Results:

Glucose sensors using GOx-grafted hydrogels showed excellent stability during continuous testing for 30 days. Control experiments indicated no enzyme leaching over a test period of 30 days. The high molar ratio of PEG functionality affords good biofouling resistance, as investigated using BSA as model protein. The glucose sensors exhibited sensitivities as high as 320 nA/(mmol/L⁻¹)/mm⁻² with linearity beyond 25 mmol/L of glucose and limit of detection less than 1 µmol/L.

Conclusions:

Photopatternable PEGylated hydrogels were demonstrated as ideal matrices for covalent immobilization of GOx enzyme with high spatial tolerances, enabling the facile fabrication of CGM devices. Long-term in vivo studies are currently underway to determine the ultimate stability of these photopatternable CGM devices.

Measurement Accuracy of Two Blood Glucose Monitoring Systems with Built-In Bolus Calculators

Introducing an Exercise Model for the Artificial Pancreas

Dayu Lv, PhD; Marc D. Breton, PhD

Center for Diabetes Technology

University of Virginia

Charlottesville, Virginia

dayu@virginia.edu

Objective:

Exercise has been a great challenge to the tight control of glycemia because of its multiple and variable effects on glucose homeostasis. An accurate model of exercise would provide a useful tool to predict the impact of exercise and to improve artificial pancreas design. As similar efforts have been addressed on in silico populations, the aim here is to 1) establish exercise impacts on insulin-dependent/independent utilization and 2) validate the model.

Method:

A bout of mild exercise stimulates the translocation of GLUTs to muscle membranes, increasing both insulin-independent and insulin-dependent glucose uptake with different time and magnitude scales. These effects of exercise were incorporated into a model of glucose/insulin dynamics as actions at the sites of insulin-independent glucose utilization and insulin-dependent maximal glucose transport rate. Simulations were performed using the so modified University of Virginia/Padova type 1 diabetes mellitus (T1DM) simulator for 100 adult in silico subjects. The distribution of glycemic drop during exercise was then compared with glucose concentration in 32 data sets of frequent YSI measurements in T1DM during mild to moderate exercise.

Result:

Simulation results were obtained for 30 min of mild exercise for each subject. A metric was created that used the slope of the blood glucose (BG) values during the 1 h before exercise, the starting BG, the BG drop after exercise, and insulin on board at the start of exercise. We were able to demonstrate similar distributions of this metric between the simulations and the clinical data (two-sample Kolmogorov-Smirnov test = 0; P = 0.79831).

Conclusion:

A new model component of exercise has been developed for simulation of mild to moderate exercise and validated on clinical data. This could facilitate the design and analysis of decision support system for exercise in T1DM.

Glycemic Control and Risk of Hypoglycemia in Hospitalized Patients with Type 2 Diabetes Mellitus at Different Wards Using the GlucoTab System

Julia K. Mader, MD; Katharina M. Neubauer, MSc, BSc; Lukas Schaupp, PhD; Felix Aberer, MD; Klaus Donsa, MSc; Stephan Spat, MSc; Bernhard Höll, MSc; Peter Beck, PhD; Johannes Plank, MD; Thomas R. Piener, MD

Quick LLC

Farmington, Connecticut

dammd@comcast.net

Objective:

The objective is to develop a bloodless, noninvasive, accurate glucose monitoring system

Method:

The methodology of this glucose testing system uses state-of-the-art advanced capability infrared technology combined with a uniquely designed polymer saliva holding system. Infrared spectroscopy is used to directly measure the glucose concentrations in the saliva of patients with diabetes. A saliva collection device (Draw Wick) uses a microfluidic technique to obtain a saliva specimen filling the window of an infrared neutral strip. The Draw Wick is then inserted into the infrared analyzer, in a similar fashion as inserting strips into a blood glucose meter, to directly measure the glucose concentration in the sample. The result is available within seconds. Testing to date indicates diabetes patients must refrain from food and nonwater consumption prior to attaining a saliva sample with a Draw Wick.

Result:

Initial patient testing of saliva samples versus existing glucometer technology has achieved an R value of 0.94 when compared with the existing monitoring technology.

Conclusion:

Infrared spectroscopy combined with saliva sampling can be used as a bloodless, noninvasive means to determine glucose levels in diabetes patients. Further testing on larger populations of patients with diabetes is required to refine saliva analyzer calibration.

Continuous Glucose Monitoring Greatly Improves Sensitivity for Detection of Reduced Prandial Excursions in Response to Infusion-Site Pretreatment with Recombinant Human Hyaluronidase (rHuPH20)

Douglas B. Muchmore, MD; Simon R. Bruce, MD; Ya Huang, MS; Xionghua W. Wu, PhD; David Rodbard, MD

Halozyme Therapeutics Inc. San Diego, California

dmuchmore@halozyme.com

Objective:

We conducted a clinical trial comparing recombinant human hyaluronidase (rHuPH20) pretreatment of continuous subcutaneous insulin infusion (CSII) sites to usual CSII in subjects with type 1 diabetes. We sought to compare postprandial glucose (PPG) excursions obtained with the ultrafast insulin profile associated with rHuPH20 pretreatment (pre-Rx) to that of usual CSII with rapid-acting analog insulin alone. Here we compare the sensitivity of CGM profiles obtained in a 25% subset of subjects to a self-monitored blood glucose (SMBG) data set, including every SMBG value with >261,000 SMBG records from the 6-month study to assess improvement in PPG using rHuPH20.

Method:

Four hundred fifty-six subjects were randomized 3:1 to rHuPH20 pre-Rx at infusion set change or usual CSII; infusion set change was every 3 days. SMBG was mandated before meals and ~90 min after≥2 meals/day. SMBG values were uploaded to a central server with annotations to identify pre- versus postprandial values. Mean excursions were calculated for each subject. CGM data obtained from whatever system the subject was using prior to study were collected by direct uploads and merged with meal timing data obtained during 7 days prior to each of six visits.

Results:

CGM demonstrated a 20 mg/dL improvement in peak PPG occurring within 4 h after breakfast (P < 0.001), and mean postbreakfast excursion PPG0–4 h was reduced by 17 mg/dL with rHuPH20 pre-Rx (P < 0.001) as compared with usual CSII. In contrast, mean breakfast PPG excursion by SMBG showed minimal (<4 mg/dL) differences between treatments (P = 0.4).

Conclusion:

In type 1 diabetes, CGM was far more sensitive than SMBG in detecting reduced PPG excursions after pre-Rx of CSII infusion sites with rHuPH20 in the ambulatory care setting.

Health Care Professionals’ Adherence and Satisfaction Using the GlucoTab, a Computerized Workflow and Decision Support System for In-Hospital Glucose Management

Katharina Maria Neubauer, BSc, MSc; Lukas Schaupp, PhD; Julia Katharina Mader, MD; Stephan Spat, MSc; Bernhard Höll, MSc; Felix Aberer, MD; Peter Beck, PhD; Thomas Augustin, PhD; Johannes Plank, MD; Thomas Rudolf Pieber, MD

Medical University of Graz

Graz, Austria

katharina.neubauer@medunigraz.at

Objective:

The development and evaluation of point-of-care computerized decision support systems are recommended to guide health care professionals (HCPs) in implementing evidence-based guidelines. The objective of this study was to investigate user adherence and satisfaction of the GlucoTab system.

Method:

The GlucoTab is a computerized workflow and decision support system for nurses and physicians with an integrated basal -bolus insulin algorithm for hospitalized noncritically ill patients with type 2 diabetes mellitus (T2DM). After testing the GlucoTab in 99 patients with T2DM at four different wards (cardiology, endocrinology, nephrology, plastic surgery), user adherence and satisfaction were assessed. Sixty-five HCPs (54 female, 85% nurses, 15% physicians, mean age 36 ± 11 years) participated in an anonymous questionnaire.

Result:

HCPs accepted 96.7 and 96.5% of the GlucoTab suggested basal and bolus insulin dosages. More basal insulin corrections by the users occur on treatment day 1. HCPs did not perform 4.2% of scheduled blood glucose measurements, 5.3% of scheduled bolus insulin injections, and 1.4% of basal insulin injections. Questionnaire data show that 53 HCPs (82%) believed that the GlucoTab is useful in daily routine; 54 HCPs (83%) stated that using the GlucoTab could prevent medication errors; 51 HCPs (78%) believed that they can handle the GlucoTab correctly; 49 HCPs (75%) answered that by using the GlucoTab, physicians have to be consulted less often; and 53 HCPs (82%) stated that by applying the GlucoTab, glucose control is more efficient.

Conclusion:

HCP adherence with the GlucoTab suggestions was very high, and the GlucoTab was used constantly in clinical practice. HCP satisfaction was high, which supports the implementation of the GlucoTab in other hospitals for evidence-based glucose management.

Funding:

The study was supported by the European Commission, Project REACTION (FP7 248590).

Evaluation of a Fuzzy Inference Model for Continuous Regimen Alterations in Type 2 Diabetes

Nonso Nnamoko, MPhil; Farath Arshad, PhD; David England, PhD; Jiten Vora, MD

School of Computing and Mathematical Science

Liverpool John Moores University

Liverpool, Merseyside, United Kingdom

n.a.nnamoko@2011.ljmu.ac.uk

Objective:

Conventional diabetes care delivery adopts a long-term feedback system, involving structured educational programs and care plan adjustments during routine checks (usually 3–6-month intervals). Here, goodness of fit evaluation was carried out on a model designed to provide short-term (daily) feedback with recommendations to compensate for any regimen disturbance.

Methods:

The prediction model is a multiple input–multiple output fuzzy inference system calibrated through expert knowledge and diabetes management algorithms; with seven inputs (age, sex, weight, height, average daily glucose, average heart rate, and calorie intake) and three output assessment predictions (glycated hemoglobin [A1C], exercise, and diet level). One day observation data sample was collected for 35 patients from Abbott Diabetes Care, which includes continuous blood glucose measurement, body weight, height, average daily heart rate, daily calorie intake, and most recent A1C result along with blind assessments of individual exercise and diet performance. The data sample consists of 13 male and 22 female patients aged between 18 and 65 years. Goodness of fit evaluation was conducted using mean square error, normalized mean square error (NMSE), and normalized root mean square error (NRMSE) between observed/advised and predicted data sets. A total of 35 × 3 output data pairs were compared.

Results:

Overall mean square error was 0.0899, showing good fit with little variation. NRMSE and NMSE evaluation allowed for individual results for each of the output assessment variables: A1C, exercise, and diet levels. NRMSE (0.7387, 0.7881, and 0.3716) and NMSE (0.9317, 0.9551, and 0.6051) results shows good fit for A1C and exercise level prediction but poor fit for diet level prediction.

Conclusions:

The model shows great potential in providing short-term interpretation of daily monitoring data, ultimately leading to continuous regimen alteration to improve diabetes management. As expected, prediction error was slightly high in some areas, but this could be due to a number of reasons, e.g., sample data veracity and assessment criteria used by Abbott. Testing in a real user environment is underway to generate more data sets that will guide future evaluations and relevant modifications to the model.

Reducing Risk with Continuous Glucose Monitoring in Intensive Care Unit Patients

John Norrie, MSc; Nicholas Barwell, PhD; Barry Crane

SumStats Ltd.

Edinburgh, Midlothian, United Kingdom

j.norrie@abdn.ac.uk

Objective:

University of Arizona

Tuscon, Arizona

timmrankin@gmail.com

Background:

The U.S. diabetic foot ulcer (DFU) incidence is 3 to 4% of 22.3 million diagnosed diabetes cases plus 6.3 million undiagnosed cases, 858,000 cases total. Risk of recurrence after healing is 30% annually. Lower-extremity multiple nerve decompression (ND) surgery reduces neuropathic DFU (nDFU) recurrence risk by >80%. Cost-effectiveness of hypothetical ND implementation to minimize nDFU recurrence is compared to the current annual $6.171 billion DFU expense.

Methods:

A literature review identified best estimates of annual incidence, recurrence risk, medical management expense, and noneconomic costs for DFU. Illustrative cost/benefit calculations were performed assuming widespread application of bilateral ND after wound healing to the nDFU problem, using Center for Medicare Services mean expense data of $1,143/case for unilateral lower-extremity ND. Calculations use conservative, evidence-based cost figures that are contemporary (2012) or adjusted for inflation.

Results:

Widespread adoption of ND after nDFU healing could reduce annual DFU occurrences at least 21% by the third program year, representing annual cost avoidance savings of $1.296 billion. This scenario proffers significant expense reduction and societal benefit and represents a minimum 1.9× return on the investment cost for surgical treatment. Further large cost savings would require reductions in initial DFU incidence, which ND may also achieve in advanced diabetic sensorimotor polyneuropathy.

Conclusions:

By minimizing the contribution of recurrences to yearly incidence rate of DFUs, ND has potential to significantly reduce the annual cost of treatment for nDFU in the U.S.

Improved Glycemic Management with an Implantable Continuous Glucose Monitoring System in a 90-Day Feasibility Study

Ravi Rastogi, PhD; Xiaolin Wang, PhD; Carrie Lorenz, PhD

Senseonics Inc.

Germantown, Maryland

ravi.rastogi@senseonics.com

Objective:

The Senseonics continuous glucose monitoring (CGM) system is composed of a fluorescence-based long-term implantable sensor and a smart wearable transmitter, which wirelessly powers the sensor and transmits data to a mobile medical application (MMA). The MMA displays the real-time glucose levels and rate of change and has features like logging calorie consumption and insulin dosage for improved glycemic management. For assessment of accuracy and clinical benefits of using a CGM system, a 90-day pilot study was performed.

Method:

A total of 30 sensors were implanted on the upper arms of 21 subjects with type 1 or type 2 diabetes. During the study, the CGM system measured glucose every 2 min. In-clinic sessions were performed every 14 days, where laboratory reference measurements were collected every 15 min over a duration of 8 h. Throughout the study, subjects followed the therapeutic regimen as prescribed by their doctor. Self-monitoring of blood glucose measurements were taken as per the protocol and entered in the MMA when requested for calibration (two per day). HbA1c levels of the patients were also monitored before and after the study.

Result:

The mean absolute relative difference was found to be 11.0%, and the mean absolute difference was 10.3 mg/dL. According to the Clarke error grid analysis, 99% of the data were in clinically accurate zone A (85.8%) and benign zone B (13.2%). We also found that the HbA1c levels of the patients on average decreased by 0.8% (SD = 0.7%) after the study.

Conclusion:

The results of this study showed that the Senseonics CGM system got a mean absolute relative difference of 11.0% and a reduction in HbA1c levels of 0.8%, showing that subjects were able to manage their therapeutic regimen better using CGM.

iSense Mastery Utility Demonstrated in 7-Day Clinical Study

Mihailo V. Rebec, PhD; Richard G. Sass, MBA

iSense CGM Inc.

Wilsonville, Oregon

rebec05@msn.com

Introduction:

iSense CGM is developing a continuous glucose monitoring (CGM) system that is called the Mastery CGM. The product development has advanced to the point that several clinical studies were completed. A 7-day clinical study was completed recently, and the results of that study will be presented in this poster.

Methods:

Sixteen subjects were enrolled in a clinical study. The CGM sensors were inserted by each subject on the right and left side of the abdomen. Each CGM system’s initial calibration was done 1 h postinsertion on day 1. Following calibration, Yellow Springs Instrument (YSI) venous plasma glucose assays were performed at 15-min intervals. Subjects returned home with instructions to calibrate their CGM system every 12 h and to keep a daily diary of living activities. Sensors were removed at the end of the seventh day. The accuracy of the sensors was determined with a comparison of paired senor glucose and YSI glucose values.

Results:

Fifteen subjects completed the 7-day study. A total of 22/34 sensors functioned for the entire 7 days. There was a total of 1,377 YSI reference measurements with valid CGM values associated with them. The overall (mean) absolute relative difference and (median) absolute relative difference for these was 18 and 15%, respectively. When optimal lag corrections were applied, the mean absolute relative difference was reduced to 14.8%.

Discussion:

The Mastery CGM has demonstrated performance that indicates its potential as an adjunct device for the treatment and control of diabetes.

Integrating Exercise Compensation into a Bihormonal Artificial Pancreas System

Engineering of Fungi-Derived Flavin Adenine Dinucleotide Glucose Dehydrogenase for Continuous Glucose Monitoring

Genki Sakai, BS; Kazushige Mori, MS; Katsuhiro Kojima, PhD; Stefano Ferri, PhD; Aimi Suzuki, BS; Seiya Tsujimura, PhD; Koji Sode, PhD

Department of Industrial Technology and Innovation

Graduate School of Engineering, Tokyo University of Agriculture and Technology

Koganei, Tokyo, Japan

50013648012@st.tuat.ac.jp

Objective:

Fungi-derived flavin adenine dinucleotide (FAD)-dependent glucose dehydrogenase (GDH) has received much attention for its glucose sensor application, focusing their narrow substrate specificity. GDH-FAD uses a variety of external electron acceptors, but not oxygen, making a sensor element for the blood glucose measurement that is not affected by the partial pressure of oxygen in the blood. However, stability of GDH-FAD is lower compared with glucose oxidase, which is the gold standard enzyme used in the sensor for the continuous glucose monitoring (CGM). This study aims to develop stable GDH-FAD suitable for CGM applications.

Method:

Based on the structural model of GDH-FAD, several amino acid substitutions were introduced. Thus, constructed engineered GDH-FADs were recombinantly produced using Escherichia coli as the host microorganisms, resulting in the nonglycosylated protein molecules, which is preferable for the combination with electrode-immobilized mediators. Stability of enzymes were evaluated based on their thermal stability as well as the stability of FAD binding.

Result:

The stability was greatly increased by introducing amino acid substitutions. The engineered GDH-FAD where one disulfide bond was introduced showed a 10-fold increase in the half-life of thermal inactivation compared with the wild-type enzyme. In addition, the holoenzyme stability, that is, the enzyme binding with its cofactor FAD, was also increased. Contrary to such drastic alteration of protein stability, the catalytic activity and the substrate specificity remained unchanged in the engineered GDH-FAD.

Conclusion:

The engineered GDH-FAD was constructed by amino acid substitutions without affecting its catalytic activity and substrate specificity. Thus, constructed molecules showed high thermal stability and holoenzyme stability, promising its application to the sensors for future CGM systems.

Remote Education and Management of Patients with Diabetes

Gary Scheiner, MS, CDE

Integrated Diabetes Services LLC

Wynnewood, Pennsylvania

gary@integrateddiabetes.com

Objective:

The aim is to demonstrate the cost savings, practicality, and effectiveness of 1) managing patients’ glucose levels and 2) teaching self-management skills via remote means.

Method:

Our practice consists of a multidisciplinary team of Certified Diabetes Educators with personal and professional expertise in intensive insulin therapy for children and adults with type 1 and type 2 diabetes. Working in collaboration with patients’ primary care providers and endocrinologists, our team consults patients (and their care providers) on strategies for optimizing glucose control by applying self-management skills. Services are provided through a variety of remote means, including telephonic consultation; video chat consultation; live chat (text only) consultation; e-mail correspondence; text messaging; website-based record keeping; electronic transmission of educational materials; sharing of downloaded data from glucose meters, insulin pumps, and continuous glucose monitors; and shared documentation via smartphone applications. All services are assessment-based, individualized, and delivered in accordance with American Diabetes Association Standards of Care and the AADE-7 Self-Care Behaviors initiative. In addition to traditional fee-for-service consultations, our practice offers ongoing retainer-type services and specialized programs for those who are newly diagnosed, use insulin pumps, are contemplating pregnancy, have weight loss needs, or are training for athletic events. Those engaged in a retainer/program service are offered monthly consultations consisting of pertinent self-management education and a detailed review of current glucose management strategies. Between appointments, patients are encouraged to download/submit data, convey questions, and obtain assistance whenever troubleshooting is required. In some instances, hands-on training/assessment is traditionally preferred. For example, when training on a self-management apparatus, checking for lipohypertrophy, or performing a foot examination. In most cases, the lack of face-to-face contact can be overcome by using video chat or training the patient’s partner to perform simple procedures. When more detailed visualization or palpation is necessary, the patient is referred to their local physician.

Result:

Patients report a very high level of satisfaction with remote care for their diabetes. They experience cost and time savings and are able to receive care in a more comfortable/relaxed setting than the traditional health care environment. No-show rates are near zero for scheduled remote appointments. Data collected over several years indicate a significant improvement in glycemic control based on reduced A1C levels and incidents of severe hypoglycemia. No significant complaints or negative events have been reported as a result of the remote form of care.

Conclusion:

Virtually every aspect of diabetes self-management education and glucose control can be achieved remotely. This approach to care is safe, practical, cost-effective, and results-oriented for all parties. It behooves health care providers to offer remote services to their patients and third-party payors to recognize remote care as a reimbursable service.

Modeling the Effect of Physical Activity on Postprandial Glucose Turnover in Healthy Subjects

Duality of Interest:

This study was supported by AgaMatrix.

Hypoglycemia Reduction with Threshold Suspend and Its Economic Impact in Insulin-Treated Patients with Diabetes in a U.S. Health Plan

Joseph A. Sierra, BA; Matthew Sussman, MA; Richard M. Bergenstal, MD; Satish K. Garg, MD; Bruce W. Bode, MD; Molly Frean, BA; Mark Friedman, MD; Max S. Gill, MS, MBA; Francine R. Kaufman, MD; Joseph Menzin, PhD

Medtronic Diabetes

Northridge, California

joseph.a.sierra@medtronic.com

Objective:

Arlington, Virginia

mslage@healthengage.com

Objective:

Because of the global growth of online health tools and communities, research technologies and methods must be developed to effectively access the users and data being generated by these sites.

Method:

We examine the assessment, development, and implementation processes associated with using an online social community’s existing aggregate user and data assets for research and data applications.

Result:

Addressed will be the many technical, research, and business issues at play and what the implications are for the industry overall. Case studies will be discussed during the presentation to describe what works and what does not when using this type of data for research purposes.

Conclusion:

Given the instant nature of the web, a whole new category of products and services for research can be tapped. Researchers can access and export real-time and retrospective aggregate data sets in seconds or view a real-time dashboard to be able to see developing trends. Using these tools can improve outcomes for individual patients as well as generating global aggregate data that can be used to help all diabetes patients.

Performance of Systems for Self-Monitoring of Blood Glucose Compared with the Analytical Quality Specifications in ISO 15197:2013

Una Ørvim Sølvik, PhD; Marianne Risa, BLS; Camilla Eide Jacobsen, MS; Grete Monsen, BLS; Sverre Sandberg, MD, PhD

Department of Global Health and Primary Care

Faculty of Medicine and Dentistry

University of Bergen

Bergen, Norway

una.solvik@igs.uib.no

Objective:

The aim of this study was to assess the accuracy of 10 different self-monitoring of blood glucose (SMBG) systems against the quality goal specified in ISO 15197:2013 under optimal conditions performed by biomedical laboratory scientists (BLSs) in a hospital environment and by persons with diabetes. In addition, the SMBG systems were evaluated against the American Diabetes Association quality goal for imprecision (CV ≤5%).

Method:

Data from 10 SMBG systems evaluations performed by the Scandinavian evaluation of laboratory equipment for primary health care (www.skup.nu) were used. Approximately 90 persons with diabetes participated in each evaluation. All measurements on each SMBG system were compared with the glucose hexokinase method. Standard reference material from the National Institute of Standards and Technology was used to show traceability, and controls with target values determined at a reference laboratory were used to verify trueness of the comparison method. The imprecision of the SMBG systems was calculated using duplicate results.

Result:

Out of the 10 SMBG systems, 9 and 6 fulfilled the requirements for accuracy when a BLS and a person with diabetes performed the measurements, respectively. Out of the 10 SMBG systems, 10 and 8 were below the recommended limit for repeatability when a BLS and a person with diabetes performed the measurements, respectively.

Conclusion:

Nearly all the SMBG systems fulfill the analytical quality specification in ISO 15197:2013 and from American Diabetes Association, especially when the BLS performs the measurements. All SMBG systems fulfilled the requirement in ISO 15197:2003. It appears that the quality of SMBG systems has improved in past years. This contributes to more accurate and reliable results, which is essential for therapeutic decisions.

Individualized Rapid-Acting Insulin Action Estimation from Insulin Pump and Continuous Glucose Measurement Data

Stochastic Model Predictive Glycemic Control for the Intensive Care Unit

Kent W. Stewart; Christopher G. Pretty, PhD; Hamish Tomlinson, BE (Hons); Liam Fisk, BE (Hons); Geoffery M. Shaw, MbChB, FJFICM; J. Geoffrey Chase, PhD

Department of Mechanical Engineering

Centre for Bio-Engineering

University of Canterbury

Christchurch, Canterbury, New Zealand

kws21@uclive.ac.nz

Objective:

Critically ill patients often experience stress-induced hyperglycemia, which has been shown to result in increased morbidity and mortality. Stochastic model predictive (STOMP) control is a new model-based glycemic control protocol that implements the probabilistic, stochastic forecasting methods of its predecessor, stochastic targeted (STAR). STOMP formalizes the control methodology using model predictive control theory to improve the ability to tune the dynamic response of the controller.

Methods:

The STOMP controller uses a model of the glucose-insulin system combined with a model of insulin sensitivity variability to predict the response to a proposed insulin/nutrition intervention. An optimal intervention is selected through minimization of the weighted sum of 6-h forecasted performance metrics derived from penalty functions, enabling safe 4-hourly measurement intervals. STOMP was evaluated using the validated virtual trial method with 149 virtual patient profiles, comprising 17,610 h of clinical data collected from the Christchurch Hospital intensive care unit.

Results:

The glycemic performance of STOMP was very similar to STAR. Both protocols maintain 85% of time within the 4.4–8.0 mmol/L glycemic target band and have very low incidence of hypoglycemia (0.06% time <2.2 mmol/L). STOMP simultaneously maintained nutrition targets, with a median patient nutrition rate of 90% of the desired goal of 25 kcal/kg/day. Thus STOMP maintained equivalent glycemic and nutritional performance to STAR, with the advantage of 35% fewer measurements.

Conclusions:

STOMP formalizes the heuristic algorithm of STAR, allowing customization to patient-specific requirements and variation in clinical practices. STOMP’s robust long-term forecasting prioritizes glycemic stability, allowing fewer interventions, which reduces clinical burden. STOMP’s strong glycemic and nutrition delivery performance in this virtual trial demonstrates its potential as a valuable enhancement to the STAR glycemic control framework.

Screening for Glucose Intolerance Using Area under the Curve after Glucose Loading

Chihiro Suminaka, MSc; Toshihiro Watanabe, MSc; Aya Morimoto, MSc; Samiko Hosoya, BSc; Toshiyuki Sato, MSc

Central Research Laboratories

Sysmex Corporation

Kobe, Japan

suminaka.chihiro@sysmex.co.jp

Objective:

We developed a glucose area under the curve (AUC) measurement system using minimally invasive interstitial fluid extraction technology (MIET) that allows convenient monitoring of postprandial glucose levels without blood sampling. Although the glucose AUC is believed to reflect whole information about glycemic excursion after glucose loading, there are few reports regarding its application to screening. Here we evaluated the performance of screening for glucose intolerance using AUC.

Method:

The subjects were 570 patients who underwent oral glucose tolerance tests (75 g). In 50 patients, we performed simultaneous measurement of interstitial fluid glucose AUC (IG-AUC) using MIET. Plasma glucose (PG) levels were measured at predefined points for 2 h after glucose loading to calculate PG-AUC. MIET included a pretreatment step using a plastic microneedle array to enable painless transdermal interstitial fluid glucose extraction and an accumulation step with a hydrogel patch placed on the pretreated area for a predefined period.

Result:

We defined the cutoff level for PG-AUC during oral glucose tolerance tests as 290 mg·h/dL according to the receiver-operating-characteristic curve (sensitivity, 0.90; specificity, 0.93). The receiver-operating-characteristic curve showed that PG-AUC could identify glucose intolerance better than HbA1c, fasting PG, or 2-h PG levels. Moreover, IG-AUC measured by MIET could identify diabetes using the same cutoff value. Both PG-AUC and IG-AUC accurately reflected peak PG levels and could identify patients with high peak PG ≥180levels ( mg/dL).

Conclusion:

PG-AUC is a reliable index of glucose intolerance, and IG-AUC measured by MIET performed as well as PG-AUC without the need for blood sampling. We are conducting a feasibility study on the simultaneous measurement of insulin AUC using MIET and plan to report the results at the annual meeting.

Factors Influencing Overnight Closed-Loop Performance during Free Living in Children and Adults with Type 1 Diabetes

Martin Tauschmann, MD; Hood Thabit, MD; Daniela Elleri, PhD; Alexandra Lubina-Solomon, PhD; Marietta Stadler; PhD; Simon R. Heller, FRCP; Stephanie A. Amiel, FRCP; Mark L. Evans, FRCP; David B. Dunger, MD; Roman Hovorka, PhD

Wellcome Trust-MRC Institute of Metabolic Science

University of Cambridge

Cambridge, United Kingdom

mt614@medschl.cam.ac.uk

Objectives:

Overnight closed-loop (OCL) insulin delivery is feasible, safe, and effective in the home setting, but outcomes vary between individuals. Understanding factors influencing glucose control may help target those likely to benefit most and propose appropriate use characteristics.

Methods:

We combined data collected during two randomized at home studies in 16 adolescents and 24 adults with type 1 diabetes on insulin pump therapy. Participants underwent, in random order, two periods of sensor-augmented insulin pump therapy or sensor-augmented insulin pump therapy combined with OCL insulin delivery, each lasting 3 (adolescents) or 4 (adults) weeks. Associations between baseline characteristics (age, HbA 1c, duration of diabetes), utility (time of OCL start, duration of OCL application per night), and performance of OCL (time in target [TT], improvement in TT) were examined using Spearman correlation.

Results:

We analyzed data on 866 closed-loop nights. Longer OCL application was associated with greater improvement in TT between 3.9 to 8.0 mmol/L (r = 0.39; P = 0.013). Baseline HbA1c was not associated with TT, improvements in TT, or duration of OCL application. Being adolescent as opposed to being adult as well as shorter duration of diabetes was associated with higher TT (age group r = 0.47, P < 0.01; duration r = −0.46, P < 0.01). Higher age and longer diabetes duration was associated with later OCL start (age r = 0.58, P < 0.01; duration r = 0.47, P < 0.01) and shorter OCL application (age r = −0.58, P < 0.01; duration r = −0.44, P < 0.01).

Conclusions:

In adolescents and adults with type 1 diabetes, the extent of glucose improvements during OCL is seen irrespective of baseline HbA1c. Improved outcomes may be observed if OCL is started earlier and applied for longer, particularly in adults.

Antihyperglycemic Potential of Psidium guajava Leaf in Streptozotocin-Induced Diabetic Rats

Toluwani Tella, BSc, MSc; Samson Mukaratirwa, BVSc, MVSc, PhD; Bubuya Masola, BSc, MSc, PhD

Discipline of Biochemistry

School of Life Sciences

University of KwaZulu-Natal (Westville Campus)

KwaZulu-Natal, South Africa

212561170@stu.ukzn.ac.za

Objectives:

Psidium guajava (PG) leaf is known to have a blood-glucose-lowering effect in diabetic rats. The objectives of the present study were to carry out a phytochemical study of PG leaf extract and investigate its protective effect on the pancreas and its effect on muscle glycogen synthase and phosphorylase activities in streptozotocin-induced diabetic male Sprague-Dawley rats.

Methods:

Diabetes was induced in male Sprague-Dawley rats with a single dose of 40 mg/kg body weight streptozotocin. The aqueous extract of PG leaves was used to treat both normal and diabetic animals (400 mg/kg body weight) for 2 weeks while control animals were treated with the vehicle.

Results:

After 2 weeks of treatment, PG lowered blood glucose and protected pancreatic tissue from diabetic damage. The treatment restored glycogen synthase activity depressed by diabetes and decreased glycogen phosphorylase activity in skeletal muscle. Gas chromatography–mass spectrometry analysis of the aqueous extract of PG indicated the presence of phenolic compounds and triterpenoids.

Conclusions:

We conclude that PG has a significant antihyperglycemic effect and that this effect may be associated with the presence of phenolic compounds and triterpenoids. PG also protects the pancreas against diabetic damage and modulates the activity of enzymes in the insulin-signaling pathway.

Accuracy and Precision of Three Common Glucose Meters in the Intensive Care Unit

Felicity Thomas, BE (Hons); Matthew Signal, PhD; Chris Pretty, PhD; Geoffrey M. Shaw, MbChB FJFICM; J. Geoffrey Chase, PhD

Department of Mechanical Engineering

University of Canterbury

Christchurch, Canterbury, New Zealand

felicity.thomas@pg.cantebury.ac.nz

Objective:

Handheld glucometers have become standard in most intensive care units (ICUs) for monitoring patients with stress-induced hyperglycemia. Inaccuracies in these devices can lead to reduced glycemic control performance. This research quantifies and compares the performance of three glucometers (Abbott Optium Xceed, Roche Accu-Chek Inform II, and Nova StatStrip) in the ICU setting.

Methods:

Blood samples from 13 critically ill patients were analyzed for blood glucose (BG) concentration using a blood-gas analyzer (BGA; Radiometer ABL90 Flex). Aliquots from each sample were also distributed across up to five glucometers of each model, resulting in 724 Abbott BGA, 432 Nova BGA, and 481 Roche BGA paired measurements. Bias and precision were used to quantify performance. Bias was defined as the median of the discrepancies between glucometer and BGA for a given blood sample. Precision was characterized by the difference between the maximum and minimum glucometer values for a given blood sample.

Results:

The mean (bias) results for the glucometers were Abbott 0.4 (0.6) mmol/L, Nova −0.1 (0.5) mmol/L, and Roche −0.2 (0.3) mmol/L.

Conclusion:

The Nova and the Roche glucometers adjust for hematocrit when calculating plasma glucose concentration. The lack of patient-specific hematocrit adjustment may account for the increased bias and reduced precision of the Abbott meter. Both the Nova and Roche glucometers are designed for point-of-care testing in the hospital environment. Their negative bias means they will underestimate BG level, which is safe. In contrast, the Abbott, an inexpensive device designed for diabetes management, can overestimate BG. This overestimation may not be significant for many cases of diabetes management. However, in an ICU setting, with very variable patients, underestimation is preferred, as it would result in lower insulin doses and increased safety.

Glycemic and Nutrition Delivery: Performance of the Stochastic Targeted Protocol

Noninvasive Glucose Measurement at the Finger With Mid-Infrared Spectroscopy and Photoacoustic Detection Is Not Influenced by Body Water Content

Hermann v.Lilienfeld-Toal, MD; Alexander Bauer, BA; Otto Hertzberg, BA; Arne Küderle; Miguel Pleitez, PhD; Werner Mäntele, PhD

Elté Sensoric GmbH

Gelnhausen, Hesse, Germany

huovonlilienfeldtoal@t-online.de

Objective:

The noninvasive glucose measurement by means of mid-infrared (IR) spectroscopy combined with photoacoustic (PA) detection is based on the quantification of glucose within the epidermis. The composition of the different parts of the epidermis (such as stratum corneum and the stratum spinosum) is probably of critical importance for the result with this approach. Therefore we systematically analyzed the influence of different easily accessible locations with different thickness of stratum corneum for measurement to find an optimal position.

Method:

During oral glucose tolerance tests (OGTTs) in five volunteers we compared blood glucose with noninvasively predicted values based on a mid-IR PA system designed to evaluate the glucose concentration within the epidermis. A test was performed in four different locations on the hand (index finger, thumb, hypothenar, and arm inside). In order to study body water loss, two OGTTs, before and after physical exercise (jogging for 45 min), were performed. The root mean square error of cross validation was calculated as an indicator of the quality of prediction.

Result:

The results of these experiments show the lowest error in the OGTTs performed was at the index finger (16 mg/dL), and the physical exercise did not influence the quality of prediction (18 mg/dL before vs. 16 mg/dL after).

Conclusion:

Our data suggest that regardless of differences in stratum corneum composition, the mid-IR PA gives best results in the optimal perfused parts of the hand. Changes in body water content as it is produced by physical activity do not influence the quality of the results.

A Fuzzy Adaptive Strategy for Blood Glucose Controllers

Josep Vehi, PhD; Aleix Beneyto, BEng

University of Girona

Girona, Girona, Spain

josep.vehi@udg.edu

Objective:

A major challenge of any blood glucose control is to tune the controller and to keep it tuned. A strategy to adjust the parameters of the controller to an specific patient is presented. Main parameters of amount of bolus, limits of insulin on board, proportional gain, and basal profile are tuned using the information given by the controller itself and also the postprandial performance.

Method:

A new three-step clamp method with frequent capillary sampling was introduced, which provides valuable data for simultaneous investigations of BG meter accuracy, precision, and bias. In this trial, the random error of the tested BG meters is comparable, but a lower systematic error for Accu-Chek, BGStar, iBGStar, and MyStar Extra gives these meters a highly accurate performance at low, normal, and high BG levels.

Duality of Interest:

This investigator-initiated study was supported by Sanofi.

Blood Glucose Meter Performance: A Comparison between Two Test Procedures

Eric Zijlstra, PhD; Annette Baumstark, PhD; Cornelia Haug, MD; Lutz Heinemann, PhD; Guido Freckmann, MD; Christoph Kapitza, MD

Profil

Neuss, Germany

eric.zijlstra@profil.com

Objective:

We compare the measurement accuracy and bias of the BGStar blood glucose (BG) meter evaluated by a reduced scale standard test protocol and a new three-step clamp method.

Method:

Three BGStar systems were evaluated in two clinical trials. The first used a reduced scale protocol of the ISO 15197:2013 accuracy evaluation, including 35 subjects with type 1 or type 2 diabetes, two measurement series per subject, and no defined distribution of BG concentrations (instead of 100 subjects with one measurement series per subject and with a defined BG distribution). BG concentrations ranged from 56 to 397 mg/dL. The second trial, conducted with the same test materials at a different test center, evaluated meter accuracy and bias using a new three-step (60–100–200 mg/dL) clamp method with frequent capillary sampling (10 samples per clamp level). Nineteen subjects with type 1 diabetes participated in this trial. Reference samples were analyzed using the YSI2300 STAT Plus glucose analyzer at both sites.

Result:

Results are presented for the three BGStar systems compared between the two procedures, i.e., reduced scale standard test protocol versus clamp method. The mean absolute relative deviation was 4.3, 4.6, 5.3 vs. 4.6, 5.4, 4.8%. Results within accuracy limits of ISO 15197:2013 were 99.3, 98.6, 94.3 vs. 100, 99.1, 98.2%. The Clarke error grid zone A results were 100, 100, 98.6 vs. 99.4, 100, 100%. Bias was −2.4, −1.9, 3.7 vs. −0.9, −2.6, 2.9%.

Conclusion:

Similar performance results for the BGStar systems were obtained with the three-step clamp method and with a reduced scale standard test protocol.

Duality of Interest:

This investigator-initiated study was supported by Sanofi.

PERMALINK

2014 Diabetes Technology Meeting Abstracts

Dose Accuracy of Reusable Insulin Pens in India: NovoPen 4, NovoPen 3, HumaPen Ergo, Glaritus Pen Royale, INSUPen, and AllStar

Objective:

Method:

Result:

Conclusion:

Duality of Interest:

From Bench to Bedside: Conquering the Translational Gap by Humanizing Type 2 Diabetes Research

Background:

Method:

Results:

Conclusion:

Development and Use of a 13C6-Glucose Isotope-Dilution Liquid Chromatography–Mass Spectrometry Method to Determine System Accuracy and Precision of the In Touch Cellular Glucose Meter System

Introduction:

Methods:

Results:

Conclusion:

Predictors of Nocturnal Hypoglycemia in Patients with Type 1 Diabetes: Effects of Automatic-Threshold-Suspend Feature Activation

Objective:

Method:

Result:

Conclusion:

Accurate Prediction and Avoidance of Hypoglycemia in an Artificial Pancreas Using Spatiotemporal Models

Objective:

Method:

Result:

Conclusion:

A Synthetic Fusion Protein for Epigenetic Control of Pancreatic Cell Function

Objective:

Method:

Result:

Conclusion:

Service Evaluation of the Diabetes Insulin Guidance Service (DIGS)

Objective:

Method:

Results:

Conclusion:

Acetaminophen Interference with Subcutaneous Glucose Sensing

Objective:

Method:

Result:

Conclusion:

Acoustic Methods for Transmitting Blood Glucose Measurements

Objective:

Method:

Result:

Conclusion:

Improved Automated Glucose Clamps Using Algorithm Variations with the ClampArt Device

Objectives:

Methods:

Results:

Conclusions:

Effect of Local Warming on Regular Human Insulin Pharmacokinetics

Background:

Method:

Result:

Conclusion:

Early Results Suggest Effectiveness of an Online Diabetes Prevention Program

Objective:

Method:

Result:

Conclusion:

Comparison of Insulin Pump Therapy and Multiple Daily Injection Therapy in Suboptimally Controlled Type 2 Diabetes: The OpT2mise Trial

Objective:

Method:

Result:

Conclusion:

Buccal Delivery of the Glucagon-Like Peptide Agonist Exenatide Using ArisCrown Technology

Objective:

Method:

Result:

Conclusion:

Designing a Decision Support System to Prevent Postexercise Hypoglycemia in Type 1 Diabetes

Objective:

Method:

Results:

Conclusion:

A Novel Imaging Technique for Monitoring Stability of Encapsulated Islets

Objective:

Development and Use of a ¹³C6-Glucose Isotope-Dilution Liquid Chromatography–Mass Spectrometry Method to Determine System Accuracy and Precision of the In Touch Cellular Glucose Meter System