Prandial insulin dosing using run-to-run control: application of clinical data and medical expertise to define a suitable performance metric

Bolus Insulin calculation without meal information. A reinforcement learning approach

In continuous subcutaneous insulin infusion and multiple daily injections, insulin boluses are usually calculated based on patient-specific parameters, such as carbohydrates-to-insulin ratio (CR), insulin sensitivity-based correction factor (CF), and the estimation of the carbohydrates (CHO) to be ingested. This study aimed to calculate insulin boluses without CR, CF, and CHO content, thereby eliminating the errors caused by misestimating CHO and alleviating the management burden on the patient. A Q-learning-based reinforcement learning algorithm (RL) was developed to optimise bolus insulin doses for in-silico type 1 diabetic patients. A realistic virtual cohort of 68 patients with type 1 diabetes that was previously developed by our research group, was considered for the in-silico trials. The results were compared to those of the standard bolus calculator (SBC) with and without CHO misestimation using open-loop basal insulin therapy. The percentage of the overall duration spent in the target range of 70-180 mg/dL was 73.4% and 72.37%, <70 mg/dL was 1.96 and 0.70%, and >180 mg/dL was 23.40 and 24.63%, respectively, for RL and SBC without CHO misestimation. The results revealed that RL outperformed SBC in the presence of CHO misestimation, and despite not knowing the CHO content of meals, the performance of RL was similar to that of SBC in perfect conditions. This algorithm can be incorporated into artificial pancreas and automatic insulin delivery systems in the future.

A Model-Based Insulin Dose Optimization Algorithm for People With Type 1 Diabetes on Multiple Daily Injections Therapy

OBJECTIVE

Multiple daily injections (MDI) therapy is the most common treatment for type 1 diabetes (T1D) including basal insulin doses to keep glucose levels constant during fasting conditions and bolus insulin doses with meals. Optimal insulin dosing is critical to achieving satisfactory glycemia but is challenging due to inter- and intra-individual variability. Here, we present a novel model-based iterative algorithm that optimizes insulin doses using previous-day glucose, insulin, and meal data.

METHODS

Our algorithm employs a maximum-a-posteriori method to estimate parameters of a model describing the effects of changes in basal-bolus insulin doses. Then, parameter estimates, their confidence intervals, and the goodness of fit, are combined to generate new recommendations. We assessed our algorithm in three ways. First, a clinical data set of 150 days (15 participants) were used to evaluate the proposed model and the estimation method. Second, 60-day simulations were performed to demonstrate the efficacy of the algorithm. Third, a sample 6-day clinical experiment is presented and discussed.

RESULTS

The model fitted the clinical data well with a root-mean-square-error of 1.75 mmol/L. Simulation results showed an improvement in the time in target (3.9-10 mmol/L) from 64% to 77% and a decrease in the time in hypoglycemia (< 3.9 mmol/L) from 8.1% to 3.8%. The clinical experiment demonstrated the feasibility of the algorithm.

CONCLUSION

Our algorithm has the potential to improve glycemic control in people with T1D using MDI.

SIGNIFICANCE

This work is a step forward towards a decision support system that improves their quality of life.

An Insulin Bolus Advisor for Type 1 Diabetes Using Deep Reinforcement Learning

(1) Background: People living with type 1 diabetes (T1D) require self-management to maintain blood glucose (BG) levels in a therapeutic range through the delivery of exogenous insulin. However, due to the various variability, uncertainty and complex glucose dynamics, optimizing the doses of insulin delivery to minimize the risk of hyperglycemia and hypoglycemia is still an open problem. (2) Methods: In this work, we propose a novel insulin bolus advisor which uses deep reinforcement learning (DRL) and continuous glucose monitoring to optimize insulin dosing at mealtime. In particular, an actor-critic model based on deep deterministic policy gradient is designed to compute mealtime insulin doses. The proposed system architecture uses a two-step learning framework, in which a population model is first obtained and then personalized by subject-specific data. Prioritized memory replay is adopted to accelerate the training process in clinical practice. To validate the algorithm, we employ a customized version of the FDA-accepted UVA/Padova T1D simulator to perform in silico trials on 10 adult subjects and 10 adolescent subjects. (3) Results: Compared to a standard bolus calculator as the baseline, the DRL insulin bolus advisor significantly improved the average percentage time in target range (70-180 mg/dL) from 74.1%±8.4% to 80.9%±6.9% (p<0.01) and 54.9%±12.4% to 61.6%±14.1% (p<0.01) in the the adult and adolescent cohorts, respectively, while reducing hypoglycemia. (4) Conclusions: The proposed algorithm has the potential to improve mealtime bolus insulin delivery in people with T1D and is a feasible candidate for future clinical validation.

A pilot non-inferiority randomized controlled trial to assess automatic adjustments of insulin doses in adolescents with type 1 diabetes on multiple daily injections therapy

BACKGROUND

Multiple daily injections (MDI) therapy for type 1 diabetes involves basal and bolus insulin doses. Non-optimal insulin doses contribute to the lack of satisfactory glycemic control. We aimed to evaluate the feasibility of an algorithm that optimizes daily basal and bolus doses using glucose monitoring systems for MDI therapy users.

METHODS

We performed a pilot, non-inferiority, randomized, parallel study at a diabetes camp comparing basal-bolus insulin dose adjustments made by camp physicians (PA) and a learning algorithm (LA), in children and adolescents on MDI therapy. Participants wore a glucose sensor and underwent 11 days of daily dose adjustments in either arm. Algorithm adjustments were reviewed and approved by a physician. The last 7 days were examined for outcomes.

RESULTS

Twenty-one youths (age 13.3 [SD, 3.7] years; 13 females; HbA1c 8.6% [SD, 1.8]) were randomized to either group (LA [n = 10] or PA [n = 11]). The algorithm made 293 adjustments with a 92% acceptance rate from the camp physicians. In the last 7 days, the time in target glucose (3.9-10 mmol/L) in LA (39.5%, SD, 20.7) was similar to PA (38.4%, SD, 15.6) (P = .89). The number of hypoglycemic events per day in LA (0.3, IQR, [0.1-0.6]) was similar to PA (0.2, IQR, [0.0-0.4]) (P = .42). There was no incidence of severe hypoglycemia nor ketoacidosis.

CONCLUSIONS

In this pilot study, glycemic outcomes in the LA group were similar to the PA group. This algorithm has the potential to facilitate MDI therapy, and longer and larger studies are warranted.

Preventing undesirable behavior of intelligent machines

Intelligent machines using machine learning algorithms are ubiquitous, ranging from simple data analysis and pattern recognition tools to complex systems that achieve superhuman performance on various tasks. Ensuring that they do not exhibit undesirable behavior-that they do not, for example, cause harm to humans-is therefore a pressing problem. We propose a general and flexible framework for designing machine learning algorithms. This framework simplifies the problem of specifying and regulating undesirable behavior. To show the viability of this framework, we used it to create machine learning algorithms that precluded the dangerous behavior caused by standard machine learning algorithms in our experiments. Our framework for designing machine learning algorithms simplifies the safe and responsible application of machine learning.

The Use of a Smart Bolus Calculator Informed by Real-time Insulin Sensitivity Assessments Reduces Postprandial Hypoglycemia Following an Aerobic Exercise Session in Individuals With Type 1 Diabetes

OBJECTIVE

Insulin dosing in type 1 diabetes (T1D) is oftentimes complicated by fluctuating insulin requirements driven by metabolic and psychobehavioral factors impacting individuals' insulin sensitivity (IS). In this context, smart bolus calculators that automatically tailor prandial insulin dosing to the metabolic state of a person can improve glucose management in T1D.

RESEARCH DESIGN AND METHODS

Fifteen adults with T1D using continuous glucose monitors (CGMs) and insulin pumps completed two 24-h admissions in a hotel setting. During the admissions, participants engaged in an early afternoon 45-min aerobic exercise session, after which they received a standardized dinner meal. The dinner bolus was computed using a standard bolus calculator or smart bolus calculator informed by real-time IS estimates. Glucose control was assessed in the 4 h following dinner using CGMs and was compared between the two admissions.

RESULTS

The IS-informed bolus calculator allowed for a reduction in postprandial hypoglycemia as quantified by the low blood glucose index (2.02 vs. 3.31, P = 0.006) and percent time <70 mg/dL (8.48% vs. 15.18%, P = 0.049), without increasing hyperglycemia (high blood glucose index: 3.13 vs. 2.09, P = 0.075; percent time >180 mg/dL: 13.24% vs. 10.42%, P = 0.5; percent time >250 mg/dL: 2.08% vs. 1.19%, P = 0.317). In addition, the number of hypoglycemia rescue treatments was reduced from 12 to 7 with the use of the system.

CONCLUSIONS

The study shows that the proposed IS-informed bolus calculator is safe and feasible in adults with T1D, appropriately reducing postprandial hypoglycemia following an exercise-induced IS increase.

Continuous Glucose Monitors and Activity Trackers to Inform Insulin Dosing in Type 1 Diabetes: The University of Virginia Contribution

Objective: Suboptimal insulin dosing in type 1 diabetes (T1D) is frequently associated with time-varying insulin requirements driven by various psycho-behavioral and physiological factors influencing insulin sensitivity (IS). Among these, physical activity has been widely recognized as a trigger of altered IS both during and following the exercise effort, but limited indication is available for the management of structured and (even more) unstructured activity in T1D. In this work, we present two methods to inform insulin dosing with biosignals from wearable sensors to improve glycemic control in individuals with T1D. Research Design and Methods: Continuous glucose monitors (CGM) and activity trackers are leveraged by the methods. The first method uses CGM records to estimate IS in real time and adjust the insulin dose according to a person’s insulin needs; the second method uses step count data to inform the bolus calculation with the residual glucose-lowering effects of recently performed (structured or unstructured) physical activity. The methods were tested in silico within the University of Virginia/Padova T1D Simulator. A standard bolus calculator and the proposed “smart” systems were deployed in the control of one meal in presence of increased/decreased IS (Study 1) and following a 1-hour exercise bout (Study 2). Postprandial glycemic control was assessed in terms of time spent in different glycemic ranges and low/high blood glucose indices (LBGI/HBGI), and compared between the dosing strategies. Results: In Study 1, the CGM-informed system allowed to reduce exposure to hypoglycemia in presence of increased IS (percent time < 70 mg/dL: 6.1% versus 9.9%; LBGI: 1.9 versus 3.2) and exposure to hyperglycemia in presence of decreased IS (percent time > 180 mg/dL: 14.6% versus 18.3%; HBGI: 3.0 versus 3.9), tending toward optimal control. In Study 2, the step count-informed system allowed to reduce hypoglycemia (percent time < 70 mg/dL: 3.9% versus 13.4%; LBGI: 1.7 versus 3.2) at the cost of a minor increase in exposure to hyperglycemia (percent time > 180 mg/dL: 11.9% versus 7.5%; HBGI: 2.4 versus 1.5). Conclusions: We presented and validated in silico two methods for the smart dosing of prandial insulin in T1D. If seen within an ensemble, the two algorithms provide alternatives to individuals with T1D for improving insulin dosing accommodating a large variety of treatment options. Future work will be devoted to test the safety and efficacy of the methods in free-living conditions.

Model individualization for artificial pancreas

BACKGROUND AND OBJECTIVE

The inter-subject variability characterizing the patients affected by type 1 diabetes mellitus makes automatic blood glucose control very challenging. Different patients have different insulin responses, and a control law based on a non-individualized model could be ineffective. The definition of an individualized control law in the context of artificial pancreas is currently an open research topic. In this work we consider two novel identification approaches that can be used for individualizing linear glucose-insulin models to a specific patient.

METHODS

The first approach belongs to the class of black-box identification and is based on a novel kernel-based nonparametric approach, whereas the second is a gray-box identification technique which relies on a constrained optimization and requires to postulate a model structure as prior knowledge. The latter is derived from the linearization of the average nonlinear adult virtual patient of the UVA/Padova simulator. Model identification and validation are based on in silico data collected during simulations of clinical protocols designed to produce a sufficient signal excitation without compromising patient safety. The identified models are evaluated in terms of prediction performance by means of the coefficient of determination, fit, positive and negative max errors, and root mean square error.

RESULTS

Both identification approaches were used to identify a linear individualized glucose-insulin model for each adult virtual patient of the UVA/Padova simulator. The resulting model simulation performance is significantly improved with respect to the performance achieved by a linear average model.

CONCLUSIONS

The approaches proposed in this work have shown a good potential to identify glucose-insulin models for designing individualized control laws for artificial pancreas.

A Dual Mode Adaptive Basal-Bolus Advisor Based on Reinforcement Learning

Self-monitoring of blood glucose (SMBG) and continuous glucose monitoring (CGM) are commonly used by type 1 diabetes (T1D) patients to measure glucose concentrations. The proposed adaptive basal-bolus algorithm (ABBA) supports inputs from either SMBG or CGM devices to provide personalised suggestions for the daily basal rate and prandial insulin doses on the basis of the patients' glucose level on the previous day. The ABBA is based on reinforcement learning, a type of artificial intelligence, and was validated in silico with an FDA-accepted population of 100 adults under different realistic scenarios lasting three simulated months. The scenarios involve three main meals and one bedtime snack per day, along with different variabilities and uncertainties for insulin sensitivity, mealtime, carbohydrate amount, and glucose measurement time. The results indicate that the proposed approach achieves comparable performance with CGM or SMBG as input signals, without influencing the total daily insulin dose. The results are a promising indication that AI algorithmic approaches can provide personalised adaptive insulin optimization and achieve glucose control-independent of the type of glucose monitoring technology.

Automated closed-loop control of diabetes: the artificial pancreas

The incidence of Diabetes Mellitus is on the rise worldwide, which exerts enormous health toll on the population and enormous pressure on the healthcare systems. Now, almost hundred years after the discovery of insulin in 1921, the optimization problem of diabetes is well formulated as maintenance of strict glycemic control without increasing the risk for hypoglycemia. External insulin administration is mandatory for people with type 1 diabetes; various medications, as well as basal and prandial insulin, are included in the daily treatment of type 2 diabetes. This review follows the development of the Diabetes Technology field which, since the 1970s, progressed remarkably through continuous subcutaneous insulin infusion (CSII), mathematical models and computer simulation of the human metabolic system, real-time continuous glucose monitoring (CGM), and control algorithms driving closed-loop control systems known as the "artificial pancreas" (AP). All of these developments included significant engineering advances and substantial bioelectronics progress in the sensing of blood glucose levels, insulin delivery, and control design. The key technologies that enabled contemporary AP systems are CSII and CGM, both of which became available and sufficiently portable in the beginning of this century. This powered the quest for wearable home-use AP, which is now under way with prototypes tested in outpatient studies during the past 6 years. Pivotal trials of new AP technologies are ongoing, and the first hybrid closed-loop system has been approved by the FDA for clinical use. Thus, the closed-loop AP is well on its way to become the digital-age treatment of diabetes.

Continuous Glucose Monitoring and Insulin Informed Advisory System with Automated Titration and Dosing of Insulin Reduces Glucose Variability in Type 1 Diabetes Mellitus

BACKGROUND

Glucose variability (GV) remains a key limiting factor in the success of diabetes management. While new technologies, for example, accurate continuous glucose monitoring (CGM) and connected insulin delivery devices, are now available, current treatment standards fail to leverage the wealth of information generated. Expert systems, from automated insulin delivery to advisory systems, are a key missing element to richer, more personalized, glucose management in diabetes.

METHODS

Twenty four subjects with type 1 diabetes mellitus (T1DM), 15 women, 37 ± 11 years of age, hemoglobin A1c 7.2% ± 1%, total daily insulin (TDI) 46.7 ± 22.3 U, using either an insulin pump or multiple daily injections with carbohydrate counting, completed two randomized crossover 48-h visits at the University of Virginia, wearing Dexcom G4 CGM, and using either usual care or the UVA decision support system (DSS). DSS consisted of a combination of automated insulin titration, bolus calculation, and CHO treatment advice. During each admission, participants were exposed to a variety of meal sizes and contents and two 45-min bouts of exercise. GV and glucose control were assessed using CGM.

RESULTS

The use of DSS significantly reduced GV (coefficient of variation: 0.36 ± 08. vs. 0.33 ± 0.06, P = 0.045) while maintaining glycemic control (average CGM: 155.2 ± 27.1 mg/dL vs. 155.2 ± 23.2 mg/dL), by reducing hypoglycemia exposure (%<70 mg/dL: 3.8% ± 4.6% vs. 1.8% ± 2%, P = 0.018), with nonsignificant trends toward reduction of significant hyperglycemia overnight (%>250 mg/dL: 5.3% ± 9.5% vs. 1.9% ± 4.6%) and at mealtime (11.3% ± 14.8% vs. 5.8% ± 9.1%).

CONCLUSIONS

A CGM/insulin informed advisory system proved to be safe and feasible in a cohort of 24 T1DM subjects. Use of the system may result in reduced GV and improved protection against hypoglycemia.

Personalized Adaptive CBR Bolus Recommender System for Type 1 Diabetes

Type 1 diabetes mellitus (T1DM) is a chronic disease. Those who have it must administer themselves with insulin to control their blood glucose level. It is difficult to estimate the correct insulin dosage due to the complex glucose metabolism, which can lead to less than optimal blood glucose levels. This paper presents PepperRec, a case-based reasoning (CBR) bolus insulin recommender system capable of dealing with an unrestricted number of situations in which T1DM persons can find themselves. PepperRec considers several factors that affect glucose metabolism, such as data about the physical activity of the user, and can also cope with missing values for these factors. Based on CBR methodology, PepperRec uses new methods to adapt past recommendations to the current state of the user, and retains updated historical patient information to deal with slow and gradual changes in the patient over time (concept drift). The proposed approach is tested using the UVA/PADOVA simulator with 33 virtual subjects and compared with other methods in the literature, and with the default insulin therapy of the simulator. The achieved results demonstrate that PepperRec increases the amount of time the users are in their target glycaemic range, reduces the time spent below it, while maintaining, or even reducing, the time spent above it.

Automatic Adaptation of Basal Insulin Using Sensor-Augmented Pump Therapy

BACKGROUND

People with insulin-dependent diabetes rely on an intensified insulin regimen. Despite several guidelines, they are usually impractical and fall short in achieving optimal glycemic outcomes. In this work, a novel technique for automatic adaptation of the basal insulin profile of people with diabetes on sensor-augmented pump therapy is presented.

METHODS

The presented technique is based on a run-to-run control law that overcomes some of the limitations of previously proposed methods. To prove its validity, an in silico validation was performed. Finally, the artificial intelligence technique of case-based reasoning is proposed as a potential solution to deal with variability in basal insulin requirements.

RESULTS

Over a period of 4 months, the proposed run-to-run control law successfully adapts the basal insulin profile of a virtual population (10 adults, 10 adolescents, and 10 children). In particular, average percentage time in target [70, 180] mg/dl was significantly improved over the evaluated period (first week versus last week): 70.9 ± 11.8 versus 91.1 ± 4.4 (adults), 46.5 ± 11.9 versus 80.1 ± 10.9 (adolescents), 49.4 ± 12.9 versus 73.7 ± 4.1 (children). Average percentage time in hypoglycemia (<70 mg/dl) was also significantly reduced: 9.7 ± 6.6 versus 0.9 ± 1.2 (adults), 10.5 ± 8.3 versus 0.83 ± 1.0 (adolescents), 10.9 ± 6.1 versus 3.2 ± 3.5 (children). When compared against an existing technique over the whole evaluated period, the presented approach achieved superior results on percentage of time in hypoglycemia: 3.9 ± 2.6 versus 2.6 ± 2.2 (adults), 2.9 ± 1.9 versus 2.0 ± 1.5 (adolescents), 4.6 ± 2.8 versus 3.5 ± 2.0 (children), without increasing the percentage time in hyperglycemia.

CONCLUSION

The present study shows the potential of a novel technique to effectively adjust the basal insulin profile of a type 1 diabetes population on sensor-augmented insulin pump therapy.

Bolus Advisors: Sources of Error, Targets for Improvement

Bolus advisors that are designed to improve the accuracy of individual bolus doses relative to a meal's carb content and the current glucose have not substantially changed since they were introduced 15 years ago despite an obvious need for enhancement and innovation. Although some glycemic benefits have been demonstrated, bolus advisors largely ignore the large amounts of clinical data they gather that could have a significant impact on glucose outcomes. Concerns have also been raised regarding the aggressive nature of largely unpublished or poorly explained bolus advisor algorithms. Hypoglycemia and hyperglycemia remain significant risks due to inaccurate bolus advisor settings and the absence of tracking or an inappropriate handling of bolus on board. This review covers common sources for bolus advisor error such as the selection of physiologically inappropriate bolus advisor settings, the use of short duration of insulin action times, poor algorithm logic that tends to cover all carb intake fully, and an excessive reliance on simplistic dosing algorithms. As well as discussing these areas, we provide 21 ways to improve current bolus calculators.

Extremum Seeking Control for Personalized Zone Adaptation in Model Predictive Control for Type 1 Diabetes

Zone model predictive control has proven to be an effective closed-loop method to regulate blood glucose for people with type 1 diabetes (T1D). In this paper, we present a universal model-free optimization scheme for adapting the zone for T1D patients individually. The adaptation is based on a clinical glycemic risk index named relative regularized glycemic penalty index (rrGPI), which is calculated from glucose measurements by a continuous glucose monitor. The scheme's objective is to minimize rrGPI by simultaneously modulating a controller's blood glucose target zone's upper bound and lower bound. The adaptation mechanism is based on extremum seeking control, in which the zone boundaries are driven by gradient estimation obtained by continuously sinusoidally modulating and demodulating the rrGPI readings. To improve the adaptation method's robustness against uncertainties, a decaying feedback gain and a vanishing dither signal are employed. in-silico trials suggested that the personalized optimized zone can be reached within a week of adaptation. Both for announced and unannounced meals, the proposed method outperforms the fixed zone [80, 140] mg/dL, which has been employed in the authors' clinical trials. It is also shown that the developed method has strong robustness against real-life uncertainties.

Twelve-Week 24/7 Ambulatory Artificial Pancreas With Weekly Adaptation of Insulin Delivery Settings: Effect on Hemoglobin A1c and Hypoglycemia

OBJECTIVE

Artificial pancreas (AP) systems are best positioned for optimal treatment of type 1 diabetes (T1D) and are currently being tested in outpatient clinical trials. Our consortium developed and tested a novel adaptive AP in an outpatient, single-arm, uncontrolled multicenter clinical trial lasting 12 weeks.

RESEARCH DESIGN AND METHODS

Thirty adults with T1D completed a continuous glucose monitor (CGM)-augmented 1-week sensor-augmented pump (SAP) period. After the AP was started, basal insulin delivery settings used by the AP for initialization were adapted weekly, and carbohydrate ratios were adapted every 4 weeks by an algorithm running on a cloud-based server, with automatic data upload from devices. Adaptations were reviewed by expert study clinicians and patients. The primary end point was change in hemoglobin A_1c (HbA_1c). Outcomes are reported adhering to consensus recommendations on reporting of AP trials.

RESULTS

Twenty-nine patients completed the trial. HbA_1c, 7.0 ± 0.8% at the start of AP use, improved to 6.7 ± 0.6% after 12 weeks (-0.3, 95% CI -0.5 to -0.2, P < 0.001). Compared with the SAP run-in, CGM time spent in the hypoglycemic range improved during the day from 5.0 to 1.9% (-3.1, 95% CI -4.1 to -2.1, P < 0.001) and overnight from 4.1 to 1.1% (-3.1, 95% CI -4.2 to -1.9, P < 0.001). Whereas carbohydrate ratios were adapted to a larger extent initially with minimal changes thereafter, basal insulin was adapted throughout. Approximately 10% of adaptation recommendations were manually overridden. There were no protocol-related serious adverse events.

CONCLUSIONS

Use of our novel adaptive AP yielded significant reductions in HbA_1c and hypoglycemia.

A Personalized Week-to-Week Updating Algorithm to Improve Continuous Glucose Monitoring Performance

BACKGROUND

Continuous glucose monitoring (CGM) systems are increasingly becoming essential components in type 1 diabetes mellitus (T1DM) management. Current CGM technology requires frequent calibration to ensure accurate sensor performance. The accuracy of these systems is of great importance since medical decisions are made based on monitored glucose values and trends.

METHODS

In this work, we introduce a calibration strategy that is augmented with a weekly updating feature. During the life cycle of the sensor, the calibration mechanism periodically estimates the parameters of a calibration model to fit self-monitoring blood glucose (SMBG) measurements. At the end of each week of use, an optimization problem that minimizes the sum of squared residuals between past reference and predicted blood glucose values is solved remotely to identify personalized calibration parameters. The newly identified parameters are used to initialize the calibration mechanism of the following week.

RESULTS

The proposed method was evaluated using two sets of clinical data both consisting of 6 weeks of Dexcom G4 Platinum CGM data on 10 adults with T1DM (over 10 000 hours of CGM use), with seven SMBG data points per day measured by each subject in an unsupervised outpatient setting. Updating the calibration parameters using the history of calibration data indicated a positive trend of improving CGM performance.

CONCLUSIONS

Although not statistically significant, the updating framework showed a relative improvement of CGM accuracy compared to the non-updating, static calibration method. The use of information collected for longer periods is expected to improve the performance of the sensor over time.

Enhancing automatic closed-loop glucose control in type 1 diabetes with an adaptive meal bolus calculator - in silico evaluation under intra-day variability

BACKGROUND AND OBJECTIVE

Current prototypes of closed-loop systems for glucose control in type 1 diabetes mellitus, also referred to as artificial pancreas systems, require a pre-meal insulin bolus to compensate for delays in subcutaneous insulin absorption in order to avoid initial post-prandial hyperglycemia. Computing such a meal bolus is a challenging task due to the high intra-subject variability of insulin requirements. Most closed-loop systems compute this pre-meal insulin dose by a standard bolus calculation, as is commonly found in insulin pumps. However, the performance of these calculators is limited due to a lack of adaptiveness in front of dynamic changes in insulin requirements. Despite some initial attempts to include adaptation within these calculators, challenges remain.

METHODS

In this paper we present a new technique to automatically adapt the meal-priming bolus within an artificial pancreas. The technique consists of using a novel adaptive bolus calculator based on Case-Based Reasoning and Run-To-Run control, within a closed-loop controller. Coordination between the adaptive bolus calculator and the controller was required to achieve the desired performance. For testing purposes, the clinically validated Imperial College Artificial Pancreas controller was employed. The proposed system was evaluated against itself but without bolus adaptation. The UVa-Padova T1DM v3.2 system was used to carry out a three-month in silico study on 11 adult and 11 adolescent virtual subjects taking into account inter-and intra-subject variability of insulin requirements and uncertainty on carbohydrate intake.

RESULTS

Overall, the closed-loop controller enhanced by an adaptive bolus calculator improves glycemic control when compared to its non-adaptive counterpart. In particular, the following statistically significant improvements were found (non-adaptive vs. adaptive). Adults: mean glucose 142.2 ± 9.4vs. 131.8 ± 4.2mg/dl; percentage time in target [70, 180]mg/dl, 82.0 ± 7.0vs. 89.5 ± 4.2; percentage time above target 17.7 ± 7.0vs. 10.2 ± 4.1. Adolescents: mean glucose 158.2 ± 21.4vs. 140.5 ± 13.0mg/dl; percentage time in target, 65.9 ± 12.9vs. 77.5 ± 12.2; percentage time above target, 31.7 ± 13.1vs. 19.8 ± 10.2. Note that no increase in percentage time in hypoglycemia was observed.

CONCLUSION

Using an adaptive meal bolus calculator within a closed-loop control system has the potential to improve glycemic control in type 1 diabetes when compared to its non-adaptive counterpart.

Clinical Safety and Feasibility of the Advanced Bolus Calculator for Type 1 Diabetes Based on Case-Based Reasoning: A 6-Week Nonrandomized Single-Arm Pilot Study

BACKGROUND

The Advanced Bolus Calculator for Diabetes (ABC4D) is an insulin bolus dose decision support system based on case-based reasoning (CBR). The system is implemented in a smartphone application to provide personalized and adaptive insulin bolus advice for people with type 1 diabetes. We aimed to assess proof of concept, safety, and feasibility of ABC4D in a free-living environment over 6 weeks.

METHODS

Prospective nonrandomized single-arm pilot study. Participants used the ABC4D smartphone application for 6 weeks in their home environment, attending the clinical research facility weekly for data upload, revision, and adaptation of the CBR case base. The primary outcome was postprandial hypoglycemia.

RESULTS

Ten adults with type 1 diabetes, on multiple daily injections of insulin, mean (standard deviation) age 47 (17), diabetes duration 25 (16), and HbA1c 68 (16) mmol/mol (8.4 (1.5) %) participated. A total of 182 and 150 meals, in week 1 and week 6, respectively, were included in the analysis of postprandial outcomes. The median (interquartile range) number of postprandial hypoglycemia episodes within 6-h after the meal was 4.5 (2.0-8.2) in week 1 versus 2.0 (0.5-6.5) in week 6 (P = 0.1). No episodes of severe hypoglycemia occurred during the study.

CONCLUSION

The ABC4D is safe for use as a decision support tool for insulin bolus dosing in self-management of type 1 diabetes. A trend suggesting a reduction in postprandial hypoglycemia was observed in the final week compared with week 1.

Advanced Insulin Bolus Advisor Based on Run-To-Run Control and Case-Based Reasoning

This paper presents an advanced insulin bolus advisor for people with diabetes on multiple daily injections or insulin pump therapy. The proposed system, which runs on a smartphone, keeps the simplicity of a standard bolus calculator while enhancing its performance by providing more adaptability and flexibility. This is achieved by means of applying a retrospective optimization of the insulin bolus therapy using a novel combination of run-to-run (R2R) that uses intermittent continuous glucose monitoring data, and case-based reasoning (CBR). The validity of the proposed approach has been proven by in-silico studies using the FDA-accepted UVa-Padova type 1 diabetes simulator. Tests under more realistic in-silico scenarios are achieved by updating the simulator to emulate intrasubject insulin sensitivity variations and uncertainty in the capillarity measurements and carbohydrate intake. The CBR(R2R) algorithm performed well in simulations by significantly reducing the mean blood glucose, increasing the time in euglycemia and completely eliminating hypoglycaemia. Finally, compared to an R2R stand-alone version of the algorithm, the CBR(R2R) algorithm performed better in both adults and adolescent populations, proving the benefit of the utilization of CBR. In particular, the mean blood glucose improved from 166 ± 39 to 150 ± 16 in the adult populations (p = 0.03) and from 167 ± 25 to 162 ± 23 in the adolescent population (p = 0.06). In addition, CBR(R2R) was able to completely eliminate hypoglycaemia, while the R2R alone was not able to do it in the adolescent population.

An Advanced Bolus Calculator for Type 1 Diabetes: System Architecture and Usability Results

This paper presents the architecture and initial usability results of an advanced insulin bolus calculator for diabetes (ABC4D), which provides personalized insulin recommendations for people with diabetes by differentiating between various diabetes scenarios and automatically adjusting its parameters over time. The proposed platform comprises two main components: a smartphone-based patient platform allowing manual input of glucose and variables affecting blood glucose levels (e.g., meal carbohydrate content and exercise) and providing real-time insulin bolus recommendations; and a clinical revision platform to supervise the automatic adaptations of the bolus calculator parameters. The system implements a previously in silico validated bolus calculator algorithm based on case-based reasoning, which uses information from similar past events (i.e., cases) to suggest improved personalized insulin bolus recommendations and automatically learns from new events. Usability of ABC4D was assessed by analyzing the system usage at the end of a six-week pilot study (n = 10). Further feedback on the use of ABC4D has been obtained from each participant at the end of the study from a usability questionnaire. On average, each participant requested 115 ± 21 insulin recommendations, of which 103 ± 28 (90%) were accepted. The clinical revision software proposed a total of 754 case revisions, where 723 (96%) adaptations were approved by a clinical expert and updated in the patient platform.

Method for automatic adjustment of an insulin bolus calculator: in silico robustness evaluation under intra-day variability

BACKGROUND AND OBJECTIVE

Insulin bolus calculators are simple decision support software tools incorporated in most commercially available insulin pumps and some capillary blood glucose meters. Although their clinical benefit has been demonstrated, their utilisation has not been widespread and their performance remains suboptimal, mainly because of their lack of flexibility and adaptability. One of the difficulties that people with diabetes, clinicians and carers face when using bolus calculators is having to set parameters and adjust them on a regular basis according to changes in insulin requirements. In this work, we propose a novel method that aims to automatically adjust the parameters of a bolus calculator. Periodic usage of a continuous glucose monitoring device is required for this purpose.

METHODS

To test the proposed method, an in silico evaluation under real-life conditions was carried out using the FDA-accepted Type 1 diabetes mellitus (T1DM) UVa/Padova simulator. Since the T1DM simulator does not incorporate intra-subject variability and uncertainty, a set of modifications were introduced to emulate them. Ten adult and ten adolescent virtual subjects were assessed over a 3-month scenario with realistic meal variability. The glycaemic metrics: mean blood glucose; percentage time in target; percentage time in hypoglycaemia; risk index, low blood glucose index; and blood glucose standard deviation, were employed for evaluation purposes. A t-test statistical analysis was carried out to evaluate the benefit of the presented algorithm against a bolus calculator without automatic adjustment.

RESULTS

The proposed method statistically improved (p<0.05) all glycemic metrics evaluating hypoglycaemia on both virtual cohorts: percentage time in hypoglycaemia (i.e. BG<70 mg/dl) (adults: 2.7±4.0 vs. 0.4±0.7, p=0.03; adolescents: 7.1±7.4 vs. 1.3±2.4, p=0.02) and low blood glucose index (LBGI) (adults: 1.1±1.3 vs. 0.3±0.2, p=0.002; adolescents: 2.0±2.19 vs. 0.7±1.4, p=0.05). A statistically significant improvement was also observed on the blood glucose standard deviation (BG SD mg/dL) (adults: 33.5±13.7 vs. 29.2±8.3, p=0.01; adolescents: 63.7±22.7 vs. 44.9±23.9, p=0.01). Apart from a small increase in mean blood glucose on the adult cohort (129.9±11.9 vs. 133.9±11.6, p=0.03), the rest of the evaluated metrics, despite showing an improvement trend, did not experience a statistically significant change.

CONCLUSIONS

A novel method for automatically adjusting the parameters of a bolus calculator has the potential to improve glycemic control in T1DM diabetes management.

Preliminary evaluation of a new semi-closed-loop insulin therapy system over the prandial period in adult patients with type 1 diabetes: the WP6.0 Diabeloop study

There is room for improvement in the algorithms used in closed-loop insulin therapy during the prandial period. This pilot study evaluated the efficacy and safety of the Diabeloop algorithm (model predictive control type) during the postprandial period. This 2-center clinical trial compared interstitial glucose levels over two 5-hour periods (with/without the algorithm) following a calibrated lunch. On the control day, the amount of insulin delivered by the pump was determined according to the patient's usual parameters. On the test day, 50% or 75% of the theoretical bolus required was delivered, while the algorithm, informed of carbohydrate intake, proposed changes to insulin delivery every 15 minutes using modeling to forecast glucose levels. The primary endpoint was percentage of time spent at near normoglycemia (70-180 mg/dl). Twelve patients with type 1 diabetes (9 men, age 35.6 ± 12.7 years, HbA1c 7.3 ± 0.8%) were included. The percentage of time spent in the target range was 84.5 ± 20.8 (test day) versus 69.2 ± 33.9% (control day, P = .11). The percentage of time spent in hypoglycemia < 70 mg/dl was 0.2 ± 0.8 (test) versus 4.4 ± 8.2% (control, P = .18). Interstitial glucose at the end of the test (5 hours) was 127.5 ± 40.1 (test) versus 146 ± 53.5 mg/dl (control, P = .25). The insulin doses did not differ, and no differences were observed between the 50% and 75% boluses. In a semi-closed-loop configuration with manual priming boluses (25% or 50% reduction), the Diabeloop v1 algorithm was as successful as the manual method in determining the prandial bolus, without any exposure to excessive hypoglycemic risk.

The mathematician's control toolbox for management of type 1 diabetes

Blood glucose levels are controlled by well-known physiological feedback loops: high glucose levels promote insulin release from the pancreas, which in turn stimulates cellular glucose uptake. Low blood glucose levels promote pancreatic glucagon release, stimulating glycogen breakdown to glucose in the liver. In healthy people, this control system is remarkably good at maintaining blood glucose in a tight range despite many perturbations to the system imposed by diet and fasting, exercise, medications and other stressors. Type 1 diabetes mellitus (T1DM) results from loss of the insulin-producing cells of the pancreas, the beta cells. These cells serve as both sensor (of glucose levels) and actuator (insulin/glucagon release) in a control physiological feedback loop. Although the idea of rebuilding this feedback loop seems intuitively easy, considerable control mathematics involving multiple types of control schema were necessary to develop an artificial pancreas that still does not function as well as evolved control mechanisms. Here, we highlight some tools from control engineering used to mimic normal glucose control in an artificial pancreas, and the constraints, trade-offs and clinical consequences inherent in various types of control schemes. T1DM can be viewed as a loss of normal physiologic controls, as can many other disease states. For this reason, we introduce basic concepts of control engineering applicable to understanding pathophysiology of disease and development of physiologically based control strategies for treatment.

Diabetes technology: markers, monitoring, assessment, and control of blood glucose fluctuations in diabetes

People with diabetes face a life-long optimization problem: to maintain strict glycemic control without increasing their risk for hypoglycemia. Since the discovery of insulin in 1921, the external regulation of diabetes by engineering means has became a hallmark of this optimization. Diabetes technology has progressed remarkably over the past 50 years-a progress that includes the development of markers for diabetes control, sophisticated monitoring techniques, mathematical models, assessment procedures, and control algorithms. Continuous glucose monitoring (CGM) was introduced in 1999 and has evolved from means for retroactive review of blood glucose profiles to versatile reliable devices, which monitor the course of glucose fluctuations in real time and provide interactive feedback to the patient. Technology integrating CGM with insulin pumps is now available, opening the field for automated closed-loop control, known as the artificial pancreas. Following a number of in-clinic trials, the quest for a wearable ambulatory artificial pancreas is under way, with a first prototype tested in outpatient setting during the past year. This paper discusses key milestones of diabetes technology development, focusing on the progress in the past 10 years and on the artificial pancreas-still not a cure, but arguably the most promising treatment of diabetes to date.

A genetic algorithm tuned optimal controller for glucose regulation in type 1 diabetic subjects

An optimal state feedback controller is designed with the objective of minimizing the elevated glucose levels caused by meal intake in Type 1 diabetic subjects, by the minimal infusion of insulin. The states for the controller based on linear quadratic regulator theory are estimated from noisy data using Kalman filter. The controller designed for a physiological relevant mathematical model is coupled with another model for simulating meal dynamics, which converts meal intake into glucose appearance rate in the plasma. The tuning parameters (weighting matrices) of the controller and the design parameters (noise covariance matrices) of the Kalman filter are optimized using genetic algorithm. The controller based on the combined framework of evolutionary computing and state estimated linear quadratic regulator is found to maintain normoglycemia for meal intakes of varying carbohydrate content. The proposed approach addresses noisy output measurement, modeling error and delay in sensor measurement.

Anticipating the next meal using meal behavioral profiles: a hybrid model-based stochastic predictive control algorithm for T1DM

Automatic control of Type 1 Diabetes Mellitus (T1DM) with subcutaneous (SC) measurement of glucose concentration and subcutaneous (SC) insulin infusion is of great interest within the diabetes technology research community. The main challenge with the so-called "SC-SC" route to control is sensing and actuation delay, which tends to either destabilize the system or inhibit the aggressiveness of the controller in responding to meals and exercise. Model predictive control (MPC) is one strategy for mitigating delay, where optimal insulin infusions can be given in anticipation of future meal disturbances. Unfortunately, exact prior knowledge of meals can only be assured in a clinical environment and uncertainty about when and if meals will arrive could lead to catastrophic outcomes. As a follow-on to our recent paper in the IFAC symposium on Biological and Medical Systems (MCBMS 2009), we develop a control law that can anticipate meals given a probabilistic description of the patient's eating behavior in the form of a random meal (behavioral) profile. Preclinical in silico trials using the oral glucose meal model of Dalla Man et al. show that the control strategy provides a convenient means of accounting for uncertain prior knowledge of meals without compromising patient safety, even in the event that anticipated meals are skipped.

Smart telemedicine support for continuous glucose monitoring: the embryo of a future global agent for diabetes care

Although current systems for continuous glucose monitoring (CGM) are the result of progressive technological improvement, and although a beneficial effect on glucose control has been demonstrated, few patients are using them. Something similar has happened to telemedicine (TM); in spite of the long-term experience, which began in the early 1980s, no TM system has been widely adopted, and presential visits are still almost the only way diabetologists and patients communicate. The hypothesis developed in this article is that neither CGM nor TM will ever be routinely implemented separately, and their consideration as essential elements for standard diabetes care will one day come from their integration as parts of a telemedical monitoring platform. This platform, which should include artificial intelligence for giving decision support to patients and physicians, will represent the core of a more complex global agent for diabetes care, which will provide control algorithms and risk analysis among other essential functions.

Quest for the artificial pancreas: combining technology with treatment

The various components of the artificial pancreas puzzle are being put into place. Features such as communication, control, modeling, and learning are being realized presently. Steps have been set in motion to carry the conceptual design through simulation to clinical implementation. The challenging pieces still to be addressed include stress and exercise; as integral parts of the ultimate goal, effort has begun to shift toward overcoming the remaining hurdles to the full artificial pancreas. The artificial pancreas is close to becoming a reality, driven by technology, and the expectation that lives will be improved.

Multinational study of subcutaneous model-predictive closed-loop control in type 1 diabetes mellitus: summary of the results

BACKGROUND

In 2008-2009, the first multinational study was completed comparing closed-loop control (artificial pancreas) to state-of-the-art open-loop therapy in adults with type 1 diabetes mellitus (T1DM).

METHODS

The design of the control algorithm was done entirely in silico, i.e., using computer simulation experiments with N=300 synthetic "subjects" with T1DM instead of traditional animal trials. The clinical experiments recruited 20 adults with T1DM at the Universities of Virginia (11); Padova, Italy (6); and Montpellier, France (3). Open-loop and closed-loop admission was scheduled 3-4 weeks apart, continued for 22 h (14.5 h of which were in closed loop), and used a continuous glucose monitor and an insulin pump. The only difference between the two sessions was that insulin dosing was performed by the patient under a physician's supervision during open loop, whereas insulin dosing was performed by a control algorithm during closed loop.

RESULTS

In silico design resulted in rapid (less than 6 months compared to years of animal trials) and cost-effective system development, testing, and regulatory approvals in the United States, Italy, and France. In the clinic, compared to open-loop, closed-loop control reduced nocturnal hypoglycemia (blood glucose below 3.9 mmol/liter) from 23 to 5 episodes (p<.01) and increased the amount of time spent overnight within the target range (3.9 to 7.8 mmol/liter) from 64% to 78% (p=.03).

CONCLUSIONS

In silico experiments can be used as viable alternatives to animal trials for the preclinical testing of insulin treatment strategies. Compared to open-loop treatment under identical conditions, closed-loop control improves the overnight regulation of diabetes.

Automatic bolus and adaptive basal algorithm for the artificial pancreatic β-cell

BACKGROUND

The current basal and bolus insulin pump therapy is dependent on user intervention; because of its open-loop nature, the therapy does not accommodate insulin variability and unmeasured meal disturbances. To conquer these challenges, an automatic bolus and adaptive basal (ABAB) therapy is proposed to regulate glucose levels for people with type 1 diabetes mellitus.

METHODS

The basal insulin profile is adjusted by the proposed algorithm every 30  min based on interstitial glucose level and its rate of change. An automated bolus is suggested by the system if a meal is detected or a hyperglycemia event occurs. A conservative insulin bolus is administered, the size of which is determined based on glucose prediction and the subject-specific correction factor. One hour later, the algorithm checks whether another bolus is needed. To prevent overdelivery, insulin-on-board is used as a safety constraint.

RESULTS

The ABAB therapy was compared with the optimal open-loop therapy and missed-bolus scenario on 100 adult subjects from the Food and Drug Administration-accepted University of Virginia/Padova Metabolic Simulator. The ABAB therapy presented superior performance according to the control-variability grid analysis. In addition, the ABAB therapy shows excellent robustness to insulin sensitivity rise: the hypoglycemia percentage was only 3.3% even when insulin sensitivity was increased by 20%. Independent of user intervention, the ABAB therapy is a good candidate for the first generation of an artificial pancreas. The proposed therapy shows excellent robustness to insulin dosing mismatches.

An advisory protocol for rapid- and slow-acting insulin therapy based on a run-to-run methodology

BACKGROUND

Emerging technology, such as an artificial pancreatic beta-cell, is not likely to be affordable to people who live in developing nations in the next 20-30 years. However, multiple-daily injection (MDI) therapy can be improved using similar advanced control algorithms designed for continuous glucose monitoring and continuous insulin infusion pumps.

METHODS

A simulation study of run-to-run control was developed for MDI therapy. Rapid- and slow-acting insulins were used in the protocol, which uses pre- and postprandial glucose measurements. The key information for the synthesis of the control algorithm is the subject insulin sensitivity that is calculated for two cases: (a) when the subject's glycemia and insulin dosing information is known (sensitivity response) and (b) when there is no previous information about the subject's response to the insulin protocol. In the latter case, this information needs to be estimated recursively using online data. After the sensitivity is recalculated, the run-to-run correction scheme is updated, obtaining an adaptive MDI therapy. The robustness of the advisory algorithm was evaluated by constant random parameter variations and superimposing sinusoidal oscillations on glucose-insulin model parameters to simulate intra-individual variability of the glucoregulatory system.

RESULTS

Optimal glycemic control has been achieved for both cases (a and b) despite variable meals (15% variation in carbohydrate content and 15-min variation in timing) and parametric variations in the glucose-insulin model. In Case (b), no profound hypoglycemic (<60 mg/dL) or hyperglycemic (>180 mg/dL) events were observed on average during all evaluations.

CONCLUSIONS

This work shows that the run-to-run framework for insulin updating can be successfully extended to an adaptive MDI protocol. These results motivate the practical implementation of this scheme in portable units such as personal digital assistants or smartphones.

Automatic data processing to achieve a safe telemedical artificial pancreas

BACKGROUND

The use of telemedicine for diabetes care has evolved over time, proving that it contributes to patient self-monitoring, improves glycemic control, and provides analysis tools for decision support. The timely development of a safe and robust ambulatory artificial pancreas should rely on a telemedicine architecture complemented with automatic data analysis tools able to manage all the possible high-risk situations and to guarantee the patient's safety.

METHODS

The Intelligent Control Assistant system (INCA) telemedical artificial pancreas architecture is based on a mobile personal assistant integrated into a telemedicine system. The INCA supports four control strategies and implements an automatic data processing system for risk management (ADP-RM) providing short-term and medium-term risk analyses. The system validation comprises data from 10 type 1 pump-treated diabetic patients who participated in two randomized crossover studies, and it also includes in silico simulation and retrospective data analysis.

RESULTS

The ADP-RM short-term risk analysis prevents hypoglycemic events by interrupting insulin infusion. The pump interruption has been implemented in silico and tested for a closed-loop simulation over 30 hours. For medium-term risk management, analysis of capillary blood glucose notified the physician with a total of 62 alarms during a clinical experiment (56% for hyperglycemic events). The ADP-RM system is able to filter anomalous continuous glucose records and to detect abnormal administration of insulin doses with the pump.

CONCLUSIONS

Automatic data analysis procedures have been tested as an essential tool to achieve a safe ambulatory telemedical artificial pancreas, showing their ability to manage short-term and medium-term risk situations.

Closed-loop artificial pancreas using subcutaneous glucose sensing and insulin delivery and a model predictive control algorithm: preliminary studies in Padova and Montpellier

New effort has been made to develop closed-loop glucose control, using subcutaneous (SC) glucose sensing and continuous subcutaneous insulin infusion (CSII) from a pump, and a control algorithm. An approach based on a model predictive control (MPC) algorithm has been utilized during closed-loop control in type 1 diabetes patients. Here we describe the preliminary clinical experience with this approach. Six type 1 diabetes patients (three in each of two clinical investigation centers in Padova and Montpellier), using CSII, aged 36 +/- 8 and 48 +/- 6 years, duration of diabetes 12 +/- 8 and 29 +/- 4 years, hemoglobin A1c 7.4% +/- 0.1% and 7.3% +/- 0.3%, body mass index 23.2 +/- 0.3 and 28.4 +/- 2.2 kg/m(2), respectively, were studied on two occasions during 22 h overnight hospital admissions 2-4 weeks apart. A Freestyle Navigator(R) continuous glucose monitor and an OmniPod insulin pump were applied in each trial. Admission 1 used open-loop control, while admission 2 employed closed-loop control using our MPC algorithm. In Padova, two out of three subjects showed better performance with the closed-loop system compared to open loop. Altogether, mean overnight plasma glucose (PG) levels were 134 versus 111 mg/dl during open loop versus closed loop, respectively. The percentage of time spent at PG > 140 mg/dl was 45% versus 12%, while postbreakfast mean PG was 165 versus 156 mg/dl during open loop versus closed loop, respectively. Also, in Montpellier, two patients out of three showed a better glucose control during closed-loop trials. Avoidance of nocturnal hypoglycemic excursions was a clear benefit during algorithm-guided insulin delivery in all cases. This preliminary set of studies demonstrates that closed-loop control based entirely on SC glucose sensing and insulin delivery is feasible and can be applied to improve glucose control in patients with type 1 diabetes, although the algorithm needs to be further improved to achieve better glycemic control.

Clinical update on optimal prandial insulin dosing using a refined run-to-run control algorithm

BACKGROUND

This article provides a clinical update using a novel run-to-run algorithm to optimize prandial insulin dosing based on sparse glucose measurements from the previous day's meals. The objective was to use a refined run-to-run algorithm to calculate prandial insulin-to-carbohydrate ratios (I:CHO) for meals of variable carbohydrate content in subjects with type 1 diabetes (T1DM).

METHOD

The open-labeled, nonrandomized study took place over a 6-week period in a nonprofit research center. Nine subjects with T1DM using continuous subcutaneous insulin infusion participated. Basal insulin rates were optimized using continuous glucose monitoring, with a target fasting blood glucose of 90 mg/dl. Subjects monitored blood glucose concentration at the beginning of the meal and at 60 and 120 minutes after the start of the meal. They were instructed to start meals with blood glucose levels between 70 and 130 mg/dl. Subjects were contacted daily to collect data for the previous 24-hour period and to give them the physician-approved, algorithm-derived I:CHO ratios for the next 24 hours. Subjects calculated the amount of the insulin bolus for each meal based on the corresponding I:CHO and their estimate of the meal's carbohydrate content. One- and 2-hour postprandial glucose concentrations served as the main outcome measures.

RESULTS

The mean 1-hour postprandial blood glucose level was 104 +/- 19 mg/dl. The 2-hour postprandial levels (96.5 +/- 18 mg/dl) approached the preprandial levels (90.1 +/- 13 mg/dl).

CONCLUSIONS

Run-to-run algorithms are able to improve postprandial blood glucose levels in subjects with T1DM.

Benefits of a bolus calculator in pre- and postprandial glycaemic control and meal flexibility of paediatric patients using continuous subcutaneous insulin infusion (CSII)

AIMS

To assess the efficacy in pre- and postprandial glycaemic control and the impact on treatment satisfaction of a bolus calculator (Bolus Wizard) incorporated into the insulin pump in Type 1 diabetic (T1D) paediatric patients using continuous subcutaneous insulin infusion (CSII) treatment at various stages of pubertal development.

METHODS

Thirty-six T1D patients on CSII treatment (19 males; mean age 13.9 +/- 3.5 years; range 4.9-17.8 years), were prospectively enrolled into this two-period crossover study. Eighteen patients were randomized to begin phase A using the Bolus Wizard, followed by phase B, using their current conventional insulin dosing method to determine pre-meal boluses. The remaining subjects were randomized to begin with phase B followed by phase A. Each study period lasted 2 weeks. A questionnaire assessing treatment satisfaction with the Bolus Wizard was compiled.

RESULTS

There was a significant reduction in blood glucose levels before and 2 h after meals and in the number of correction boluses during phase A with respect to phase B of the study. The lower frequency of hypoglycaemic events during phase A did not reach statistical significance. There were no differences between the two phases: insulin requirement, daily bolus rate (%) and meal bolus quantity.

CONCLUSIONS

When paediatric patients used CSII, the bolus insulin dose calculated using the Bolus Wizard was more effective in improving pre- and postprandial glycaemic control with fewer correction boluses, without differences in the prandial insulin requirements and without restriction in the carbohydrate content of meals. The use of the Bolus Wizard was easy and was associated with a high level of satisfaction in these patients.

Detection of a meal using continuous glucose monitoring: implications for an artificial beta-cell

OBJECTIVE

The purpose of this study was to introduce a novel meal detection algorithm (MDA) to be used as part of an artificial beta-cell that uses a continuous glucose monitor (CGM).

RESEARCH DESIGN AND METHODS

We developed our MDA on a dataset of 26 meal events using records from 19 children aged 1-6 years who used the MiniMed CGMS Gold. We then applied this algorithm to CGM records from a DirecNet pilot study of the FreeStyle Navigator continuous glucose sensor. During a research center admission, breakfast insulin was withheld for 1 h, and discrete glucose levels were obtained every 10 min after the meal.

RESULTS

Based on the Navigator readings, the MDA detected a meal at a mean time of 30 min from the onset of eating, at which time the mean serum glucose was 21 mg/dl above baseline (range 2-36 mg/dl), and >90% of meals were detected before the glucose had risen 40 mg/dl from baseline.

CONCLUSIONS

The MDA will enable automated insulin dosing in response to meals, facilitating the development of an artificial pancreas.

A Run-to-Run Control Strategy to Adjust Basal Insulin Infusion Rates in Type 1 Diabetes

Maintaining good glycemic control is a daily challenge for people with type 1 diabetes. Insulin requirements are changing constantly due to many factors, such as levels of stress and physical activity. The basal insulin requirement also has a circadian rhythm, adding another level of complexity. Automating the adjustment of insulin dosing would result in improved glycemic control, as well as an improved quality of life by significantly reducing the burden on the patient. Building on our previous success of using run-to-run control for prandial insulin dosing (a strategy adapted from the chemical process industry), we show how this same framework can be used to adjust basal infusion profiles. We present a mathematical model of insulin-glucose dynamics which we augment in order to capture the circadian variation in insulin requirements. Using this model, we show that the run-to-run framework can also be successfully applied to adjust basal insulin dosing.