Project

Double Intraperitoneal Artificial Pancreas

Goal: To do basic research and make a device for automatic delivery of insulin, i.e. an artificial pancreas for use in patients with diabetes. The system will measure glucose and deliver insulin in the peritoneal cavity.

Updates
0 new
2
Recommendations
0 new
0
Followers
0 new
15
Reads
0 new
278

Project log

Anders Lyngvi Fougner
added a research item
An open source simulation model of the mechanical properties of a fully functional insulin pump was made in Matlab Simscape. The model simulates realistic behavior of an insulin pump, parts of which are validated against real-world systems. Simulations include mechanical forces and internal pressures, and the following fluid dynamics. Failure modes, such as occlusions, can be simulated and the resulting simulations can give new insights on how these failures affect the pump and how to detect them.Clinical relevance- Realistic pump simulations can be used to analyze how pump failures affect the system and in turn how to most effectively detect them before posing a hazard to the user, increasing the safety and reliability of the system.
Anders Lyngvi Fougner
added a research item
Objective: The design of an Artificial Pancreas to regulate blood glucose levels requires reliable control methods. Model Predictive Control has emerged as a promising approach for glycemia control. However, model-based control methods require computationally simple and identifiable mathematical models that represent glucose dynamics accurately, which is challenging due to the complexity of glucose homeostasis. Methods: In this work, a simple model is deduced to estimate blood glucose concentration in subjects with Type 1 Diabetes Mellitus. Novel features in the model are power-law kinetics for intraperitoneal insulin absorption and a separate glucagon sensitivity state. Profile likelihood and a method based on singular value decomposition of the sensitivity matrix are carried out to assess parameter identifiability and guide a model reduction for improving the identification of parameters. Results: A reduced model with 10 parameters is obtained and calibrated, showing good fit to experimental data from pigs where insulin and glucagon boluses were delivered in the intraperitoneal cavity. Conclusion: A simple model with power-law kinetics can accurately represent glucose dynamics submitted to intraperitoneal insulin and glucagon injections. Importance: The parameters of the reduced model were not found to lack of local practical or structural identifiability.
Anders Lyngvi Fougner
added a research item
The intraperitoneal route of administration accounts for less than 1% of insulin treatment regimes in patients with diabetes mellitus type 1 (DM1). Despite being used for decades, a systematic review of various physiological effects of this route of insulin administration is lacking. Thus, the aim of this systematic review was to identify the physiological effects of continuous intraperitoneal insulin infusion (CIPII) compared to those of continuous subcutaneous insulin infusion (CSII) in patients with DM1. Four databases (EMBASE, PubMed, Scopus and CENTRAL) were searched beginning from the inception date of each database to 10 th of July 2020, using search terms related to intraperitoneal and subcutaneous insulin administration. Only studies comparing CIPII treatment (≥ 1 month) with CSII treatment were included. Primary outcomes were long-term glycaemic control (after ≥ 3 months of CIPII inferred from glycated haemoglobin (HbA1c) levels) and short-term (≥ 1 day for each intervention) measurements of insulin dynamics in the systematic circulation. Secondary outcomes included all reported parameters other than the primary outcomes. The search identified a total of 2242 records; 39 reports from 32 studies met the eligibility criteria. This meta-analysis focused on the most relevant clinical end points; the mean difference (MD) in HbA1c levels during CIPII was significantly lower than during CSII (MD = -6.7 mmol/mol, [95% CI: -10.3 –-3.1]; in percentage: MD = -0.61%, [95% CI: -0.94 –- 0.28], p = 0.0002), whereas fasting blood glucose levels were similar (MD = 0.20 mmol/L, [95% CI: -0.34–0.74], p = 0.47; in mg/dL: MD = 3.6 mg/dL, [95% CI: -6.1–13.3], p = 0.47). The frequencies of severe hypo- and hyper-glycaemia were reduced. The fasting insulin levels were significantly lower during CIPII than during CSII (MD = 16.70 pmol/L, [95% CI: -23.62 –-9.77], p < 0.0001). Compared to CSII treatment, CIPII treatment improved overall glucose control and reduced fasting insulin levels in patients with DM1.
Silje Fuglerud
added a research item
Near infrared spectroscopy (NIR) is a promising technique for continuous blood glucose monitoring for diabetic patients. Four interferents, at physiological concentrations, were introduced to study how the glucose predictions varied with a standard multivariate calibration model. Lactate and ethanol were found to interfere strongly with the glucose predictions unless theywere included in the calibration models. Lactate was mistaken for glucose and gave erroneously high glucose predictions, with a dose response of 0.46 mM/mM. The presence of ethanol resulted in too low glucose predictions, with a dose response of ‐0.43 mM/mM. Acetaminophen (APAP), a known interferent in the glucose monitoring devices used for diabetes management today, was not found to be an interferent in NIR spectroscopy, nor was caffeine. Thus, interferents that may appear in high concentrations, such as ethanol and lactate, must be included in the calibration or model building of future NIR‐based glucose measurement devices for diabetes monitoring. This article is protected by copyright. All rights reserved.
Anders Lyngvi Fougner
added 2 research items
Introduction The effect of intraperitoneal insulin infusion has limited evidence in the literature. Therefore, the aim of the study was to investigate the pharmacokinetics and pharmacodynamics of different intraperitoneal insulin boluses. There is a lack of studies comparing the insulin appearance in the systemic circulation after intraperitoneal compared with subcutaneous insulin delivery. Thus, we also aimed for a comparison with the subcutaneous route. Research design and methods Eight anesthetized, non-diabetic pigs were given three different intraperitoneal insulin boluses (2, 5 and 10 U). The order of boluses for the last six pigs was randomized. Endogenous insulin and glucagon release were suppressed by repeated somatostatin analog injections. The first pig was used to identify the infusion rate of glucose to maintain stable glucose values throughout the experiment. The estimated difference between insulin boluses was compared using two-way analysis of variance (GraphPad Prism V.8). In addition, a trial of three pigs which received subcutaneous insulin boluses was included for comparison with intraperitoneal insulin boluses. Results Decreased mean blood glucose levels were observed after 5 and 10 U intraperitoneal insulin boluses compared with the 2 U boluses. No changes in circulating insulin levels were observed after the 2 and 5 U intraperitoneal boluses, while increased circulating insulin levels were observed after the 10 U intraperitoneal boluses. Subcutaneously injected insulin resulted in higher values of circulating insulin compared with the corresponding intraperitoneal boluses. Conclusions Smaller intraperitoneal boluses of insulin have an effect on circulating glucose levels without increasing insulin levels in the systemic circulation. By increasing the insulin bolus, a major increase in circulating insulin was observed, with a minor additive effect on circulating glucose levels. This is compatible with a close to 100% first-pass effect in the liver after smaller intraperitoneal boluses. Subcutaneous insulin boluses markedly increased circulating insulin levels.
Objective: The design of an Artificial Pancreas to regulate blood glucose levels requires reliable control methods. Model Predictive Control has emerged as a promising approach for glycemia control. However, model-based control methods require computationally simple and identifiable mathematical models that represent glucose dynamics accurately, which is challenging due to the complexity of glucose homeostasis. Methods: In this work, a simple model is deduced to estimate blood glucose concentration in subjects with Type 1 Diabetes Mellitus. Novel features in the model are power-law kinetics for intraperitoneal insulin absorption and a separate glucagon sensitivity state. Profile likelihood and a method based on singular value decomposition of the sensitivity matrix are carried out to assess parameter identifiability and guide a model reduction for improving the identification of parameters. Results: A reduced model with 10 parameters is obtained and calibrated, showing good fit to experimental data from pigs where insulin and glucagon boluses were delivered in the intraperitoneal cavity. Conclusion: A simple model with power-law kinetics can accurately represent glucose dynamics submitted to intraperitoneal insulin and glucagon injections. Importance: The parameters of the reduced model were not found to lack of local practical or structural identifiability.
Anders Lyngvi Fougner
added 3 research items
Glucagon is a pancreatic hormone and increases the blood glucose levels. It may be incorporated in a dual hormone artificial pancreas, a device to automatically and continuously control blood glucose levels of individuals with diabetes. Artificial pancreas systems have been developed for use in the subcutaneous tissue; however, the systems are not fully automated due to slow dynamics. The intraperitoneal space is therefore investigated as an alternative location for an artificial pancreas. Glucose dynamics after subcutaneous and intraperitoneal glucagon delivery in ten anaesthetized pigs were investigated. The pigs received intraperitoneal boluses of 0.3 µg/kg and 0.6 µg/kg and a subcutaneous bolus of 0.6 µg/kg in randomized order. They also received an intraperitoneal bolus of 1 mg given at the end of the experiments to test the remaining capacity of rapid glucose release. Six pigs were included in the statistical analysis. The intraperitoneal glucagon bolus of 0.6 µg/kg gave a significantly higher glucose response from 14 to 30 min compared with the subcutaneous bolus. The results indicate that glucagon induces a larger glucose response after intraperitoneal delivery compared with subcutaneous delivery and is encouraging for the incorporation of glucagon in an intraperitoneal artificial pancreas.
Fast and accurate continuous glucose monitoring is needed in future systems for control of blood glucose levels in type 1 diabetes patients. Direct spectroscopic measurement of glucose in the peritoneal cavity is an attractive alternative to conventional electrochemical sensors placed subcutaneously. We demonstrate the feasibility of fast glucose measurements in peritoneal fluid using a fibre-coupled tuneable mid-infrared quantum cascade laser. Mid-infrared spectra (1200-925 cm-1) of peritoneal fluid samples from pigs with physiological glucose levels (32-426 mg/dL, or 1.8-23.7 mmol/L) were acquired with a tuneable quantum cascade laser employing both transmission and attenuated total reflection (ATR) spectroscopy. Using partial least-squares regression, glucose concentrations were predicted with mean absolute percentage errors (MAPEs) of 8.7% and 12.2% in the transmission and ATR configurations, respectively. These results show that highly accurate concentration predictions are possible with mid-infrared spectroscopy of peritoneal fluid, and represent a first step towards a miniaturised optical sensor for intraperitoneal continuous glucose monitoring.
Convolutional neural networks (CNNs) are widely used for image recognition and text analysis, and have been suggested for application on one-dimensional data as a way to reduce the need for pre-processing steps. Pre-processing is an integral part of multivariate analysis, but determination of the optimal pre-processing methods can be time-consuming due to the large number of available methods. In this work, the performance of a CNN was investigated for classification and regression analysis of spectral data. The CNN was compared with various other chemometric methods, including support vector machines (SVMs) for classification and partial least squares regression (PLSR) for regression analysis. The comparisons were made both on raw data, and on data that had gone through pre-processing and/or feature selection methods. The models were used on spectral data acquired with methods based on near-infrared, mid-infrared, and Raman spectroscopy. For the classification datasets the models were evaluated based on the percentage of correctly classified observations, while for regression analysis the models were assessed based on the coefficient of determination (R$^2$). Our results show that CNNs can outperform standard chemometric methods, especially for classification tasks where no pre-processing is used. However, both CNN and the standard chemometric methods see improved performance when proper pre-processing and feature selection methods are used. These results demonstrate some of the capabilities and limitations of CNNs used on one-dimensional data.
Anders Lyngvi Fougner
added 7 research items
Visible and near-infrared spectroscopy are widely used for sensing applications but suffer from poor signal-to-noise ratios for the detection of compounds with low concentrations. Enhancement by surface plasmon resonance is a popular technique that can be utilized to increase the signal of absorption spectroscopy due to the increased near-field created close to the plasmons. Despite interest in surface-enhanced infrared absorption spectroscopy (SEIRAS), the method is usually applied in lab setups rather than real-life sensing situations. This study aimed to achieve enhanced absorption from plasmons on a fiber-optic probe and thus move closer to applications of SEIRAS. A tapered coreless fiber coated with a 100 nm Au film supported signal enhancement at visible wavelengths. An increase in absorption was shown for two dyes spanning concentrations from 5 × 10−8 mol/L to 8 × 10−4 mol/L: Rhodamine 6G and Crystal Violet. In the presence of the Au film, the absorbance signal was 2–3 times higher than from an identically tapered uncoated fiber. The results confirm that the concept of SEIRAS can be implemented on an optical fiber probe, enabling enhanced signal detection in remote sensing applications.
Anders Lyngvi Fougner
added a research item
The development of rapid and accurate biomedical laser spectroscopy systems in the mid-infrared has been enabled by the commercial availability of external-cavity quantum cascade lasers (EC-QCLs). EC-QCLs are a preferable alternative to benchtop instruments such as Fourier transform infrared spectrometers for sensor development as they are small and have high spectral power density. They also allow for the investigation of multiple analytes due to their broad tuneability and through the use of multivariate analysis. This article presents an in vitro investigation with two fiber-coupled measurement setups based on attenuated total reflection spectroscopy and direct transmission spectroscopy for sensing. A pulsed EC-QCL (1200-900 cm −1) was used for measurements of glucose and albumin in aqueous solutions, with lactate and urea as interferents. This analyte composition was chosen as an example of a complex aqueous solution with relevance for biomedical sensors. Glucose concentrations were determined in both setup types with root-mean-square error of cross-validation (RMSECV) of less than 20 mg/dL using partial least-squares (PLS) regression. These results demonstrate accurate analyte measurements, and are promising for further development of fiber-coupled, miniaturised in vivo sensors based on mid-infrared spectroscopy.
Anders Lyngvi Fougner
added a research item
A typical artificial pancreas depends only on the continuous glucose monitoring (CGM) value for insulin dosing. However, both the insulin infusion and the glucose sensing are subject to time delays and slow dynamics. An automated and reliable meal onset information could enhance the control outcome of artificial pancreas by making it possible to infuse insulin earlier and thereby avoid large postprandial glucose excursions. In this study we employ abdominal sounds recorded in two healthy volunteers with a condenser microphone and propose an automated approach for meal onset detection from abdominal sounds. We use the Mel-frequency cepstral coefficients (MFCCs) and wavelet entropy extracted from the abdominal sounds as features. These features are fed to a simple feed forward neural network for discriminating meal from no-meal abdominal sounds. This approach detects meal onset with an average delay of 4.3 minutes in our limited number of subjects. More importantly, it provides lesser response delay than the state-of-the-art CGM based approach, which achieved a response delay ranging from 30-40 minutes. The preliminary results indicate that the proposed abdominal sound-based approach may provide early meal onset information. This can be exploited in an artificial pancreas through allowable earlier meal insulin boluses, resulting in improved glycemic control.
Anders Lyngvi Fougner
added an update
The APT group, Dept. of Engineering Cybernetics and Dept. of Electronic Systems, NTNU, currently have two vacant postdoc positions within sound processing. Applications are to be submitted through JobbNorge and the deadline is 3 November 2019:
These positions will develop sound processing methods related to health diagnostics and monitoring, with main focus on sounds from the digestive system and from the lungs. The primary target groups are patients with diabetes mellitus and patients with lung disease.
Main requirements:
  • Completion of a Norwegian doctoral degree (or an equivalent foreign doctoral degree) within physics, computer science, electrical or electronics engineering, control engineering, biomedical engineering, or similar fields.
  • Strong proficiency in oral and written English.
  • Strong skills in at least two of the following fields:
- Digital signal processing.
- Pattern recognition.
- Machine learning.
- Data driven (statistics based) modelling.
- Mechanistic (physics based) modelling.
For more information, please contact: anders.fougner@ntnu.no
 
Anders Lyngvi Fougner
added a research item
Background In classical approaches for an artificial pancreas, continuous glucose monitoring (CGM) is the only measured variable used for insulin dosing and additional control functions. The CGM values are subject to time delays and slow dynamics between blood and the sensing location. These time lags compromise the controller's performance in maintaining (near to) normal glucose levels. Meal information could enhance the control outcome. However, meal announcement by the user is not reliable, and it takes 30 min to 40 min from meal onset until a meal is detected by methods based on CGM. In this pilot study, the use of bowel sounds for meal detection was investigated. In particular, we focused on whether bowel sounds change qualitatively during or shortly after meal ingestion. Methods After fasting for at least 4 h, 11 healthy volunteers ingested a lunch meal at their usual time. Abdominal sound was recorded by a condenser microphone that was attached to the right upper quadrant of the abdomen by medical tape. Features that describe the power distribution over the frequency spectrum were extracted and used for classification by support vector machines. These classifiers were trained in a leave-one-out cross-validation scheme. Results Meals could be detected on average after less than 10 min with the best parameter choice. Conclusion This shows that abdominal sound monitoring could provide an early meal detection. Further studies should investigate this possibility on a larger population in more general settings.
Anders Lyngvi Fougner
added a research item
Glucose-insulin metabolism models are useful tools for research on diabetes, in development of diabetes-related medical devices like artificial pancreas systems, insulin pumps and continuous glucose monitors, and may also play a role in personalized decision support tools for people with diabetes. Such models are often highly nonlinear with many parameters that are person dependent. An example is the model used in the UVa/Padova T1DM simulator, which has a large number of states and parameters. It is desirable to be able to personalize such models through parameter identification based on limited glucose, meal and insulin data obtainable from free-living settings, as opposed to clinical research settings that have traditionally been required. In this paper we use the UVa-Padova T1DM simulator model in a case study to investigate observability of the model under different measurements, and the identifiability of its parameters as a function of the model's inputs and outputs. Structural identifiability is discussed and briefly investigated using the nonlinear Observability Rank Condition. Practical identifiability is discussed and investigated using sensitivity and Fisher information matrix analysis. We show how such analyses can be used to guide model reduction for improved identifiability, or to select the most proper subset of parameters to estimate.
Anders Lyngvi Fougner
added 2 research items
The artificial pancreas requires fast and reliable glucose measurements. The peritoneal space has shown promising results, and in one of our studies we detected glucose changes in the peritoneal space already at the same time as in the femoral artery. The peritoneal lining is highly vascularised, covered by a single layer of mesothelial cells and therefore easily accessible for proper sensor technology, e.g. optical technology. We hypothesize that the rapid intraperitoneal glucose dynamics observed in our study was possible because the sensors were located directly at the peritoneal lining, at the point where the glucose molecules entered the peritoneal space. Glucose travels slowly in fluids by diffusion, and a longer distance between the sensor and the peritoneal lining would consequently result in slower dynamics. We therefore propose to place the glucose sensor in an artificial pancreas as closely to the peritoneal lining as possible, or even utilize appropriate sensor technology to measure glucose in the peritoneal lining itself.
Anders Lyngvi Fougner
added a research item
Accurate continuous glucose monitoring (CGM) is essential for fully automated glucose control in diabetes mellitus type 1. State-of-the-art glucose control systems automatically regulate the basal insulin infusion. Users still need to manually announce meals to dose the prandial insulin boluses. An automated meal detection could release the user and improve the glucose regulation. In this study, patterns in the postprandial CGM data are exploited for meal detection. Binary classifiers are trained to recognize the postprandial pattern in horizons of the estimated glucose rate of appearance and in CGM data. The appearance rate is determined by moving horizon estimation (MHE) based on a simple model. Linear discriminant analysis (LDA) is used for classification. The proposed method is compared to methods that detect meals when thresholds are violated. Diabetes care data from twelve free-living pediatric patients was downloaded during regular screening. Experts identified meals and their start by retrospective evaluation. The classification was tested by cross-validation. Compared to the threshold-based methods, LDA showed higher sensitivity to meals with a low rate of false alarms. Classifying horizons outperformed the other methods also with respect to time of detection. The onset of meals can be detected by pattern recognition based on estimated model states and consecutive CGM measurements. No individual tuning is necessary. This makes the method easily adopted in the clinical practice.
Anders Lyngvi Fougner
added a research item
The bulkiness of common transmission spectroscopy probes prevents applicability at remote locations such as within the body. We present the fabrication and characterization of lensed fibers for transmission spectroscopy in the near-infrared. Eigenmode simulations and measurements of the coupling efficiency are presented and applied to design the setup corresponding to the sample absorption. Sensing capabilities are demonstrated on aqueous glucose samples ranged 80 to 500 mM, obtaining a mean absolute percentage error of calibration of 4.3%. With increased flexibility, transmission spectroscopic sensors at remote locations may be achievable, for example, applied to in vivo continuous glucose monitoring.
Anders Lyngvi Fougner
added an update
The APT research group (Artificial Pancreas Trondheim) and Department of Engineering Cybernetics (ITK) currently have a vacant PhD position within Meal Detection by Analysis of Bowel Sounds. Applications are to be submitted through JobbNorge and the deadline is 1 May 2019:
This PhD project aims to analyse recorded sounds from the upper body (mostly targeted at the intestines/bowel) in order to detect meals in patients with diabetes. The recorded sound will also be combined with continuous glucose sensor data and motion data in order to achieve larger robustness to e.g. ambient noise. The equipment is being built by SINTEF, but the PhD candidate will be involved in the design. Methods may include sensor fusion, calculation of time/frequency domain features, pattern recognition and machine learning. The APT group has access to facilities, staff and other necessary resources for acquisition of data from healthy people and diabetes patients, and the candidate will be involved in the planning and conduction of such experiments as part of a larger experienced team.
For more information, please contact: anders.fougner@ntnu.no
 
Anders Lyngvi Fougner
added a research item
Background The analysis of abdominal sounds can help to diagnose gastro-intestinal diseases. Sounds originating from the stomach and the intestine, the so-called bowel sounds, occur in various forms. They are described as loose successions or clusters of rather sudden bursts. Realistic recordings of abdominal sounds are contaminated with noise and artifacts from which the bowel sounds must be differentiated. Methods The proposed intrinsic mode function-fractal dimension (IMF-FD) filtering utilizes the property of the multivariate empirical mode decomposition (MEMD) to behave as a series of band pass filters. The MEMD decomposes the abdominal signal into its different frequency components. The resulting intrinsic mode functions (IMFs) are modulated in amplitude and frequency where transient sonic events occur. Based on the complexity of the IMFs, measured by their fractal dimension (FD) in sliding windows, the information-carrying IMFs are selected. The filtered signal is formed as the superposition of all selected IMFs. The IMF-FD filter not only enhances the non-linear components of the original signal but also segments them from the rest. Another important aspect of this work is that typical artifacts that occur in the same frequency range as bowel sounds can be subsequently eliminated by heuristic rules. Conclusions The method is tested on a realistic, contaminated data set with promising performance: close to 100% of the manually labeled bowel sounds are identified.
Anders Lyngvi Fougner
added 5 research items
People with diabetes mellitus type 1 could benefit from fully automated systems for glucose control. However, faults in any component of the system can severely compromise the safety of the user. An increasing degree of automation also increases the risk that faults remain undiscovered for longer periods - unless automated routines for fault detection are implemented at the same time. The aim of this article is to give a categorized overview of methods for fault detection in glucose control systems. This overview targets at disclosing hidden potentials for improvement and unresolved issues. Methods for fault detection in glucose control systems have been reviewed and classified with respect to categories such as the type of method and the exploited data basis. Both journal and conference papers were taken into account. Compared to the number of studies on glucose control algorithms, only a few articles have been published on fault detection. Surprisingly few of them consider system information beyond the standard diabetes care data.
Mid-infrared spectroscopy using multivariate analysis for quantification has high potential in biosensing. This case study on aqueous glucose solutions yields improved prediction errors by optimising preprocessing and wavelength selection procedures.
We demonstrate a multimode optical fiber sensor for spectroscopic Raman measurements of glucose concentration for the application in intraperitoneal glucose detection in diabetic patients. A regression model with a RMSEC of 2.2 mM was obtained.
Anders Lyngvi Fougner
added 2 research items
Objective Hypoglycemia is a frequent and potentially dangerous event among patients with diabetes mellitus type 1. Subcutaneous glucagon is an emergency treatment to counteract severe hypoglycemia. The effect of intraperitoneal glucagon delivery is sparsely studied. We performed a direct comparison of the blood glucose response following intraperitoneally, subcutaneously and intravenously administered glucagon. Research design and methods This is a prospective, randomized, controlled, open-label, crossover trial in 20 octreotide-treated rats. Three interventions, 1 week apart, in a randomized order, were done in each rat. All 20 rats were given intraperitoneal and subcutaneous glucagon injections, from which 5 rats were given intravenous glucagon injections and 15 rats received placebo (intraperitoneal isotonic saline) injection. The dose of glucagon was 5 µg/kg body weight for all routes of administration. Blood glucose levels were measured before and until 60 min after the glucagon/placebo injections. Results Compared with placebo-treated rats, a significant increase in blood glucose was observed 4 min after intraperitoneal glucagon administration (p=0.009), whereas after subcutaneous and intravenous glucagon administration significant increases were seen after 8 min (p=0.002 and p<0.001, respectively). In intraperitoneally treated compared with subcutaneously treated rats, the increase in blood glucose was higher after 4 min (p=0.019) and lower after 40 min (p=0.005) and 50 min (p=0.011). The maximum glucose response occurred earlier after intraperitoneal compared with subcutaneous glucagon injection (25 min vs 35 min; p=0.003). Conclusions Glucagon administered intraperitoneally gives a faster glucose response compared with subcutaneously administered glucagon in rats. If repeatable in humans, the more rapid glucose response may be of importance in a dual-hormone artificial pancreas using the intraperitoneal route for administration of insulin and glucagon.
Frequent glucose monitoring is a fundamental part of diabetes management, and good glucose control is important for long-term health outcomes. New types of electrochemical sensors that allow for continuous glucose monitoring (CGM) have become an important tool for diabetes management, although they still have drawbacks such as short lifetime and a need for frequent calibration. Other technologies are still being researched for CGM, in an attempt to replace the electrochemical sensors. Optical methods have several advantages for CGM, including potentially long sensor lifetimes and short measurement times, and many developments have been made over the last decades. This paper will review optical measurement methods for CGM, their challenges, and the current research status. The different methods will be compared, and the future prospects for optical methods will be discussed.
Anders Lyngvi Fougner
added a research item
Closed-loop glucose control has the potential to improve the glycemic control in patients with diabetes mellitus type 1. Such an artificial pancreas (AP) should keep the user safe despite all disturbances and faults. The objective of this paper is to analyze those perturbations according to their effects on the glycemic status, and thereby supporting an informed design process of the control system. As suggested by the international standard ISO 14971 for risk management of medical devices, the well proven failure modes and effects analysis (FMEA) was chosen as instrument. An FMEA scheme was modified for this purpose and applied to a single-hormone system with subcutaneous and intraperitoneal routes for glucose sensing and insulin administration. Faults that imply urgent danger and thus require fast detection and diagnosis were identified and distinguished from disturbances that can be sufficiently addressed by basic control functions, e.g. by adaptive control algorithms. Requirements and testing criteria for basic control functions as well as fault detection and diagnosis functions can be derived from the provided overview.
Anders Lyngvi Fougner
added a research item
Freestyle Libre (FL) is a factory calibrated Flash Glucose Monitor (FGM). We investigated Mean Absolute Relative Difference (MARD) between Self Monitoring of Blood Glucose (SMBG) and FL measurements in the first day of sensor wear in 39 subjects with Type 1 diabetes. The overall MARD was 12.3%, while the individual MARDs ranged from 4% to 25%. Five participants had a MARD ≥ 20%. We estimated bias and lag between the FL and SMBG measurements. The estimated biases range from −1.8 mmol / L to 1.4 mmol / L , and lags range from 2 min to 24 min . Bias is identified as a main cause of poor individual MARDs. The biases seem to persist in days 2–7 of sensor usage. All cases of MARD ≥ 20% in the first day are eliminated by bias correction, and overall MARD is reduced from 12.3% to 9.2%, indicating that adding support for voluntary user-supplied bias correction in the FL could improve its performance.
Anders Lyngvi Fougner
added a research item
Background In diabetes research, the development of the artificial pancreas has been a major topic since continuous glucose monitoring became available in the early 2000’s. A prerequisite for an artificial pancreas is fast and reliable glucose sensing. However, subcutaneous continuous glucose monitoring carries the disadvantage of slow dynamics. As an alternative, we explored continuous glucose sensing in the peritoneal space, and investigated potential spatial differences in glucose dynamics within the peritoneal cavity. As a secondary outcome, we compared the glucose dynamics in the peritoneal space to the subcutaneous tissue. Material and methods Eight-hour experiments were conducted on 12 anesthetised non-diabetic pigs. Four commercially available amperometric glucose sensors (FreeStyle Libre, Abbott Diabetes Care Ltd., Witney, UK) were inserted in four different locations of the peritoneal cavity and two sensors were inserted in the subcutaneous tissue. Meals were simulated by intravenous infusions of glucose, and frequent arterial blood and intraperitoneal fluid samples were collected for glucose reference. Results No significant differences were discovered in glucose dynamics between the four quadrants of the peritoneal cavity. The intraperitoneal sensors responded faster to the glucose excursions than the subcutaneous sensors, and the time delay was significantly smaller for the intraperitoneal sensors, but we did not find significant results when comparing the other dynamic parameters.
Anders Lyngvi Fougner
added 2 research items
Comparison of Flash Glucose Monitor (FGM) measurements to frequent Self Monitor Blood Glucose measurements (SMBG) in the first day of use. Investigation of bias and lag.
A method for preprocessing a time series of glucose measurements based on Kalman smoothing is presented. Given a glucose data time series that may be irregularly sampled, the method outputs an interpolated time series of glucose estimates with mean and variance. The method can provide homogenization of glucose data collected from different devices by using separate measurement noise parameters for differing glucose measurement equipment. We establish a link between the ISO 15197 standard and the measurement noise variance used by the Kalman smoother for Self Monitoring of Blood Glucose (SMBG) measurements. The method provides phaseless smoothing, and it can automatically correct errors in the original datasets like small fallouts and erroneous readings when surrounding data allows. The estimated variance can be used for deciding at which times the data are trustworthy. The method can be used as a preprocessing step in many kinds of glucose data processing and analysis tasks, such as computing the Mean Absolute Relative Deviation (MARD) between measurement systems, or estimating the plasma-to-interstital fluid glucose dynamics of continuous glucose monitor (CGM) or Flash Glucose Monitor (FGM) signals. The method is demonstrated on SMBG and FGM glucose data from a clinical study.
Anders Lyngvi Fougner
added 3 research items
Meals are most challenging in the regulation of blood glucose levels (BGL) in diabetes mellitus type 1, whether it is automated, semi-automated or manually controlled. The common subcutaneous (SC) route for glucose sensing and insulin administration suffers from large latencies. This paper investigates the impact of glucose sensing and insulin absorption dynamics on the achievable glucose regulation when insulin boluses are triggered by a meal detection system. In silico patients from the academic version of the UVa/Padova simulator are studied. The sub-models of glucose sensing and insulin absorption are adjusted to allow simulations with different time delays and time constants. Meals are detected with published methods based on threshold-checking of continuous glucose monitoring data. Slow glucose sensing dynamics delay the meal detection. Delayed meal detection can be compensated to some extent by exact knowledge about the insulin absorption. The combination of slow glucose sensing and slow insulin administration reduces the effect of insulin boluses on the postprandial BGL. The classical SC approach is, therefore, at high risk of large BGL excursions despite meal detection.
Anders Lyngvi Fougner
added a research item
Introduction Patients with diabetes type 1 (DM1) struggle daily to achieve good glucose control. The last decade has seen a rush of research groups working towards an artificial pancreas (AP) through the application of a double subcutaneous approach, i.e., subcutaneous (SC) continuous glucose monitoring (CGM) and continuous subcutaneous insulin infusion. Few have focused on the fundamental limitations of this approach, especially regarding outcome measures beyond time in range. Methods Based on insulin physiology, the limitations of CGM, SC insulin absorption, meal challenge, and physical activity in DM1 patients, we discuss the limitations of the double SC approach. Finally, we discuss safety measures and the achievements reported in some recent AP studies that have utilized the double SC approach. Results Most studies show that a double SC AP increases the time in range compared to a sensor-augmented insulin pump and shortens the time in hypoglycemia. Despite these achievements, the proportion of time spent in hyperglycemia is still roughly 20–40%, and hypoglycemia is still present 1–4% of the time. The main factors limiting further progress are the latency of SC CGM (at least 5–10 min) and the slow pharmacokinetics of SC-delivered fast-acting insulin. The maximum blood insulin level is reached after 45 min and the maximum glucose-lowering effect is observed after 1.5–2 h, while the glucose-lowering effect lasts for at least 5 h. Conclusions Although using a double SC AP leads to significant improvements in glucose control, the SC approach has severe limitations that hamper further progress towards a robust AP.
Nils K Skjaervold
added a research item
Rapid, accurate and robust glucose measurements are needed to make a safe artificial pancreas for the treatment of diabetes mellitus type 1 and 2. The present gold standard of continuous glucose sensing, subcutaneous (SC) glucose sensing, has been claimed to have slow response and poor robustness towards local tissue changes such as mechanical pressure, temperature changes, etc. The present study aimed at quantifying glucose dynamics from central circulation to intraperitoneal (IP) sensor sites, as an alternative to the SC location. Intraarterial (IA) and IP sensors were tested in three anaesthetized non-diabetic pigs during experiments with intravenous infusion of glucose boluses, enforcing rapid glucose level excursions in the range 70--360 mg/dL (approximately 3.8--20 mmol/L). Optical interferometric sensors were used for IA and IP measurements. A first-order dynamic model with time delay was fitted to the data after compensating for sensor dynamics. Additionally, off-the-shelf Medtronic Enlite sensors were used for illustration of SC glucose sensing. The time delay in glucose excursions from central circulation (IA) to IP sensor location was found to be in the range 0--26 s (median: 8.5 s, mean: 9.7 s, SD 9.5 s), and the time constant was found to be 0.5--10.2 min (median: 4.8 min, mean: 4.7 min, SD 2.9 min). IP glucose sensing sites have a substantially faster and more distinctive response than SC sites when sensor dynamics is ignored, and the peritoneal fluid reacts even faster to changes in intravascular glucose levels than reported in previous animal studies. This study may provide a benchmark for future, rapid IP glucose sensors.
Anders Lyngvi Fougner
added 9 research items
Objective: To assess whether 4 week’s use of a continuous glucose monitoring (CGM) system improves glucose control, treatment satisfaction or health status, as compared to intensified conventional finger-prick measurements (ICFM) in patients with type 1 diabetes mellitus (DM1). Method: Thirty patients suffering from DM1 for more than three years and treated with either insulin pumps or multiple daily insulin injections, were included in a randomised controlled cross-over trial. They were Caucasians of both genders, between 18 and 50 years, and had moderately well controlled diabetes. The participants performed either ICFM or CGM for 4 weeks, followed by an 8 week’s observation period. Thereafter they were crossed over to the opposite intervention. HbA1c, hypoglycaemic episodes, treatment satisfaction and health status were assessed at all meetings, although HbA1c was the primary endpoint. Results: At inclusion mean HbA1c was 7.8 ± 0.9 %. The mean change in HbA1c was −0.2 ± 0.1% and −0.2 ± 0.1% for the CGM and the ICFM periods, accordingly (p = 0.91). The mean changes in HbA1c during the combined treatment and observation periods were −0.1 ± 0.1% and −0.2 ± 0.1% for the CGM and the ICFM period, accordingly (p = 0.86). The frequency of severe hypoglycaemic episodes, treatment satisfaction and health status was also equal between the two interventions. No adverse events were observed.
The aim of this study was to construct a glucose regulatory algorithm by employing the natural pulsatile pattern of insulin secretion and the oscillatory pattern of resting blood glucose levels and further to regulate the blood glucose level in diabetic pigs by this method. We developed a control algorithm based on repetitive intravenous bolus injections of insulin and combined this with an intravascular blood glucose monitor. Four anesthetized pigs were used in the study. The animals developed a mildly diabetic state from streptozotocin pretreatment. They were steadily brought within the blood glucose target range of 4.5-6.0 mmol/L in 21 to 121 min and kept within that range for 128 to 238 min (hypoglycemic values varied from 2.9 to 51.1 min). The study confirmed our hypotheses regarding the feasibility of this new principle for blood glucose control, and the algorithm was constantly improved during the study to produce the best results in the last animals. The main obstacles were the drift of the IvS-1 sensor and problems with the calibration procedure, which calls for an improvement in the sensor stability before this method can be applied fully in new studies in animals and humans.
Sven Magnus Carlsen
added a project goal
To do basic research and make a device for automatic delivery of insulin, i.e. an artificial pancreas for use in patients with diabetes. The system will measure glucose and deliver insulin in the peritoneal cavity.