Stable liquid glucagon formulations for rescue treatment and bi-hormonal closed-loop pancreas.

Small doses of glucagon given subcutaneously in the research setting by an automated system prevent most cases of hypoglycemia in persons with diabetes. However, glucagon is very unstable and cannot be kept in a portable pump. Glucagon rapidly forms amyloid fibrils, even within the first day after reconstitution. Aggregation eventually leads to insoluble gels, which occlude pump catheters. Fibrillation occurs rapidly at acid pH, but is absent or minimal at alkaline pH values of ~10. Glucagon also degrades over time; this problem is greater at alkaline pH. Several studies suggest that its primary degradative pathway is deamidation, which results in a conversion of asparagine to aspartic acid. A cell-based assay for glucagon bioactivity that assesses glucagon receptor (GluR) activation can screen promising glucagon formulations. However, mammalian hepatocytes are usually problematic as they can lose GluR expression during culture. Assays for cyclic AMP (cAMP) or its downstream effector, protein kinase A (PKA), in engineered cell systems, are more reliable and suitable for inexpensive, high-throughput assessment of bioactivity.

Measuring glucose at the site of insulin delivery with a redox-mediated sensor.

Automated insulin delivery systems for people with type 1 diabetes rely on an accurate subcutaneous glucose sensor and an infusion cannula that delivers insulin in response to measured glucose. Integrating the sensor with the infusion cannula would provide substantial benefit by reducing the number of devices inserted into subcutaneous tissue. We describe the sensor chemistry and a calibration algorithm to minimize impact of insulin delivery artifacts in a new glucose sensing cannula. Seven people with type 1 diabetes undergoing automated insulin delivery used two sensing cannulae whereby one delivered a rapidly-acting insulin analog and the other delivered a control phosphate buffered saline (PBS) solution with no insulin. While there was a small artifact in both conditions that increased for larger volumes, there was no difference between the artifacts in the sensing cannula delivering insulin compared with the sensing cannula delivering PBS as determined by integrating the area-under-the-curve of the sensor values following delivery of larger amounts of fluid (P = 0.7). The time for the sensor to recover from the artifact was found to be longer for larger fluid amounts compared with smaller fluid amounts (10.3 ± 8.5 min vs. 41.2 ± 78.3 s, P < 0.05). Using a smart-sampling Kalman filtering smoothing algorithm improved sensor accuracy. When using an all-point calibration on all sensors, the smart-sampling Kalman filter reduced the mean absolute relative difference from 10.9% to 9.5% and resulted in 96.7% of the data points falling within the A and B regions of the Clarke error grid. Despite a small artifact, which is likely due to dilution by fluid delivery, it is possible to continuously measure glucose in a cannula that simultaneously delivers insulin.

Elevation of transforming growth factor beta (TGFbeta) and its downstream mediators in subcutaneous foreign body capsule tissue.

Foreign body encapsulation represents a chronic fibrotic response and has been a major obstacle that reduces the useful life of implanted biomedical devices. The precise mechanism underlying such an encapsulation is still unknown. We hypothesized that, considering its central role in many other fibrotic conditions, transforming growth factor beta (TGFbeta) may play an important role during the formation of foreign body capsule (FBC). In the present study, we implanted mock sensors in rats subcutaneously and excised FBC samples at day 7, 21, and 48-55 postimplantation. The most abundant TGFbeta isoform in all tissues was TGFbeta1, which was expressed minimally in control tissue. The expression of both TGFbeta1 RNA and protein was significantly increased in FBC tissues at all time points, with the highest level in day 7 FBC. The number of cells stained for phosphorylated Smad2, an indication of activated TGFbeta signaling, paralleled the expression of TGFbeta. A similar dynamic change was also observed in the numbers of FBC myofibroblasts, which in response to TGFbeta, differentiate from quiescent fibroblasts and synthesize collagen. Type I collagen, the most prominent downstream target of TGFbeta in fibrosis, was found in abundance in the FBC, especially during the latter time periods. We suggest that TGFbeta plays an important role in the FBC formation. Inhibition of TGFbeta signaling could be a promising strategy in the prevention of FBC formation, thereby extending the useful life of subcutaneous implants.

Passivating protein coatings for implantable glucose sensors: evaluation of protein retention.

The long-term function of implantable biosensors is limited by the foreign-body reaction (FBR). Since the acute phase of the FBR involves macrophage attachment mediated by adsorbed fibrinogen, preadsorption, and retention of other proteins might reduce the FBR. The retention of preadsorbed albumin, hemoglobin, von Willebrand's factor, and high-molecular-weight kininogen was therefore measured after exposure to plasma. The retention of preadsorbed proteins after incubation with monocyte cultures and implantation in rats was also measured. Fibrinogen adsorption from plasma to the preadsorbed surfaces was also measured. Hemoglobin adsorption was higher than that for other proteins, and it also had the greatest retention after exposure to blood plasma. When surfaces preadsorbed with hemoglobin were incubated with monocytes, more of the hemoglobin was displaced than that after incubation in plasma, while still more hemoglobin was displaced when the surfaces were implanted in vivo. Protein preadsorption on polystyrene greatly reduced fibrinogen adsorption. However, polyurethane surfaces used for glucose sensors had low fibrinogen adsorption compared with polystyrene, and this low level was not further reduced by preadsorption with other proteins. Preadsorbed proteins on polymers appear to be removed by passive exchange and/or displacement by plasma proteins and by proteases released by monocytes.

An Amperometric Glucose Sensor Integrated into an Insulin Delivery Cannula: In Vitro and In Vivo Evaluation.

BACKGROUND: Labeling prohibits delivery of insulin at the site of subcutaneous continuous glucose monitoring (CGM). Integration of the sensing and insulin delivery functions into a single device would likely increase the usage of CGM in persons with type 1 diabetes. METHODS: To understand the nature of such interference, we measured glucose at the site of bolus insulin delivery in swine using a flexible electrode strip that was laminated to the outer wall of an insulin delivery cannula. In terms of sensing design, we compared H2O2-measuring sensors biased at 600 mV with redox mediator-type sensors biased at 175 mV. RESULTS: In H2O2-measuring sensors, but not in sensors with redox-mediated chemistry, a spurious rise in current was seen after insulin lis-pro boluses. This prolonged artifact was accompanied by electrode poisoning. In redox-mediated sensors, the patterns of sensor signals acquired during delivery of saline and without any liquid delivery were similar to those acquired during insulin delivery. CONCLUSION: Considering in vitro and in vivo findings together, it became clear that the mechanism of interference is the oxidation, at high bias potentials, of phenolic preservatives present in insulin formulations. This effect can be avoided by the use of redox mediator chemistry using a low bias potential.

Prevention of hypoglycemia using risk assessment with a continuous glucose monitoring system.

Due to the lag between sugar intake and the beginning of recovery from hypoglycemia, it is necessary to intervene in an anticipatory way if one wants to prevent, not only detect, hypoglycemia. This article presents the principle of a hypoglycemia prevention system based on risk assessment. The risk situation can be defined as the moment when the system estimates that the glucose concentration is expected to reach a hypoglycemia threshold in less than a given time (e.g., 20 min). Since there are well-known discrepancies between blood and interstitial glucose concentrations, the aim of this experimental study performed in nondiabetic rats was first to validate this strategy, and second to determine whether it can work when the glucose concentration is estimated by a glucose sensor in subcutaneous tissue rather than in blood. We used a model of controlled decrease in blood glucose concentration. A glucose infusion, the profile of which mimicked the appearance of glucose from an intragastric load, was administered either when hypoglycemia was detected or on the basis of risk recognition. Despite the lag between the beginning of the load and that of the increase in blood glucose concentration, which was in all experiments 15-20 min, hypoglycemia was fully prevented without overshoot hyperglycemia in the groups of rats in which the glucose load was started when the hypoglycemia risk was detected, on the basis of either blood or interstitial glucose concentration. This was, of course, not the case when the same glucose load was infused at the detection of the hypoglycemia threshold.

Development of a highly stable, nonaqueous glucagon formulation for delivery via infusion pump systems.

Despite a vigorous research effort, to date, the development of systems that achieve glucagon stability in aqueous formulations (without reconstitution) has failed to produce any clinical candidates. We have developed a novel, nonaqueous glucagon formulation based on a biocompatible pharmaceutical solvent, dimethyl sulfoxide, which demonstrates excellent physical and chemical stability at relatively high concentrations and at high temperatures. This article reports the development of a novel, biocompatible, nonaqueous native human glucagon formulation for potential use in subcutaneous infusion pump systems. Data are presented that demonstrate physical and chemical stability under presumed storage conditions (>2 years at room temperature) as well as "in use" stability and compatibility in an Insulet's OmniPod(®) infusion pump. Also presented are results of a skin irritation study in a rabbit model and pharmacokinetics/pharmacodynamics data following pump administration of glucagon in a diabetic swine model. This nonaqueous glucagon formulation is suitable for further clinical development in pump systems.

A novel, stable, aqueous glucagon formulation using ferulic acid as an excipient.

Commercial glucagon is unstable due to aggregation and degradation. In closed-loop studies, it must be reconstituted frequently. For use in a portable pump for 3 days, a more stable preparation is required. At alkaline pH, curcumin inhibited glucagon aggregation. However, curcumin is not sufficiently stable for long-term use. Here, we evaluated ferulic acid, a stable breakdown product of curcumin, for its ability to stabilize glucagon. Ferulic acid-formulated glucagon (FAFG), composed of ferulic acid, glucagon, L-methionine, polysorbate-80, and human serum albumin in glycine buffer at pH 9, was aged for 7 days at 37°C. Glucagon aggregation was assessed by transmission electron microscopy (TEM) and degradation by high-performance liquid chromatography (HPLC). A cell-based protein kinase A (PKA) assay was used to assess in vitro bioactivity. Pharmacodynamics (PD) of unaged FAFG, 7-day aged FAFG, and unaged synthetic glucagon was determined in octreotide-treated swine. No fibrils were observed in TEM images of fresh or aged FAFG. Aged FAFG was 94% intact based on HPLC analysis and there was no loss of bioactivity. In the PD swine analysis, the rise over baseline of glucose with unaged FAFG, aged FAFG, and synthetic native glucagon (unmodified human sequence) was similar. After 7 days of aging at 37°C, an alkaline ferulic acid formulation of glucagon exhibited significantly less aggregation and degradation than that seen with native glucagon and was bioactive in vitro and in vivo. Thus, this formulation may be stable for 3-7 days in a portable pump for bihormonal closed-loop treatment of T1D.

Effect of Repeated Glucagon Doses on Hepatic Glycogen in Type 1 Diabetes: Implications for a Bihormonal Closed-Loop System.

OBJECTIVE: To evaluate subjects with type 1 diabetes for hepatic glycogen depletion after repeated doses of glucagon, simulating delivery in a bihormonal closed-loop system. RESEARCH DESIGN AND METHODS: Eleven adult subjects with type 1 diabetes participated. Subjects underwent estimation of hepatic glycogen using (13)C MRS. MRS was performed at the following four time points: fasting and after a meal at baseline, and fasting and after a meal after eight doses of subcutaneously administered glucagon at a dose of 2 µg/kg, for a total mean dose of 1,126 µg over 16 h. The primary and secondary end points were, respectively, estimated hepatic glycogen by MRS and incremental area under the glucose curve for a 90-min interval after glucagon administration. RESULTS: In the eight subjects with complete data sets, estimated glycogen stores were similar at baseline and after repeated glucagon doses. In the fasting state, glycogen averaged 21 ± 3 g/L before glucagon administration and 25 ± 4 g/L after glucagon administration (mean ± SEM) (P = NS). In the fed state, glycogen averaged 40 ± 2 g/L before glucagon administration and 34 ± 4 g/L after glucagon administration (P = NS). With the use of an insulin action model, the rise in glucose after the last dose of glucagon was comparable to the rise after the first dose, as measured by the 90-min incremental area under the glucose curve. CONCLUSIONS: In adult subjects with well-controlled type 1 diabetes (mean A1C 7.2%), glycogen stores and the hyperglycemic response to glucagon administration are maintained even after receiving multiple doses of glucagon. This finding supports the safety of repeated glucagon delivery in the setting of a bihormonal closed-loop system.

Fabrication of a Flexible Amperometric Glucose Sensor Using Additive Processes.

This study details the use of printing and other additive processes to fabricate a novel amperometric glucose sensor. The sensor was fabricated using a Au coated 12.7 µm thick polyimide substrate as a starting material, where micro-contact printing, electrochemical plating, chloridization, electrohydrodynamic jet (e-jet) printing, and spin coating were used to pattern, deposit, chloridize, print, and coat functional materials, respectively. We have found that e-jet printing was effective for the deposition and patterning of glucose oxidase inks with lateral feature sizes between ~5 to 1000 µm in width, and that the glucose oxidase was still active after printing. The thickness of the permselective layer was optimized to obtain a linear response for glucose concentrations up to 32 mM and no response to acetaminophen, a common interfering compound, was observed. The use of such thin polyimide substrates allow wrapping of the sensors around catheters with high radius of curvature ~250 µm, where additive and microfabrication methods may allow significant cost reductions.

The effect of microgeometry, implant thickness and polyurethane chemistry on the foreign body response to subcutaneous implants.

We addressed the effect of implant thickness, implant porosity, and polyurethane (PU) chemistry on angiogenesis and on the foreign body response in rats. The following materials were implanted subcutaneously for 7 weeks then excised for histologic analysis: a solid PU; a solid polyurethane with silicone and polyethylene oxide (PU-S-PEO); porous expanded polytetrafluoroethylene (ePTFE); and porous polyvinyl alcohol sponge (PVA). Two thicknesses of PU-S-PEO were compared: 300 microns (thin) and 2000 microns (thick). Foreign body capsule (FBC) thickness was much less in PU-S-PEO implants than in PU implants. In addition, FBC were thinner in thin implants than in thick implants. FBC was much more dense in solid implants than in porous implants. As compared with solid implants, porous implants (PVA and ePTFE) led to a marked increase in the number of microvessels that developed adjacent to the implant, as observed both with hematoxylin/eosin staining and with an immunohistochemical anti-endothelial stain. We conclude that the polyethylene oxide and silicone moieties in PU reduce the thickness of the subsequent FBC. In addition, thin implants lead to a thin FBC. Porous implants (PVA and ePTFE) cause more angiogenesis than solid implants. These results may have implications for the measurement of blood-derived analytes by biosensors.

A new amperometric glucose microsensor: in vitro and short-term in vivo evaluation.

For biosensor fabrication, it is important to optimize materials and methods in order to create predictable function in vitro and in vivo. For this reason, we designed a new glucose sensor ('revised protocol') that utilized an outer permselective membrane made of amphiphobic polyurethane which allows glucose passage through hydrophilic segments. An inner polyethersulfone membrane, stabilized with a trimethoxysilane, provided specificity. Before application of the inner membrane, it was necessary to etch the platinum electrode with a radio frequency oxygen plasma. The revised protocol sensors (n=185) were compared with sensors fabricated with an earlier ('original') protocol (n=204) which used an outer polyurethane without hydrophilic segments and a complex inner membrane of cellulose acetate and Nafion. The function of revised protocol sensors was more predictable in vitro as evidenced by a much lower variation of glucose sensitivity than the original protocol sensors. Revised and original protocol sensors were nearly linear up to a glucose concentration of 20 mM. In vitro interference from 0.1 mM acetaminophen was minimal in both groups of sensors and would be expected to represent about 2% of the total sensor response at normal glucose levels for revised protocol sensors. Prolonged testing of the revised protocol sensors for 11 days during immersion in buffer revealed stable sensitivities (day 1: 6.12+/-1.34 nA/mM; day 3: 6.33+/-1.40; day 8: 7.13+/-1.39; and day 11: 7.56+/-1.47; sensitivity for day 1 vs. each other day: not significant) and no critical loss of glucose oxidase activity. The response of the revised protocol sensors (n=7) to intraperitoneal glucose was tested in rats approximately one day after subcutaneous implantation and the sensors tracked glucose closely with a slight lag of 3-6 min.

Vascularizing the tissue surrounding a model biosensor: how localized is the effect of a subcutaneous infusion of vascular endothelial growth factor (VEGF)?

Implantable continuous biosensors would improve disease management but long term function of such devices have been limited by a hypovascular foreign body capsule that inhibits influx of analytes. To assess whether capsule vascularity could be increased, we studied the histologic effects of a 28-day continuous infusion of vascular endothelial growth factor (VEGF) (0.45 microg/day) vs. saline from the surface of a model disk biosensor that was implanted subcutaneously in rats. At day 40, tissue was obtained at varying distances from the infusion port and capsular microvessels were counted using two histologic techniques. VEGF treatment led to a marked increase in capillary density. In tissue located 1 mm away from the infusion site, capillary density in VEGF-treated animals was 200-300% higher than in saline controls. Tissue located 13 mm away, but not 25 mm away, also demonstrated neovascularization. Serum obtained from a distant vein during the infusion did not show an elevated concentration of VEGF. These data demonstrate that a subcutaneous infusion of VEGF creates localized neovascularization of the foreign body capsule and suggest that systemic effects of VEGF are avoidable. Vascularization of a foreign body capsule surrounding a subcutaneous biosensor might well extend its useful life.

A wire-based dual-analyte sensor for glucose and lactate: in vitro and in vivo evaluation.

Continuous measurement of lactate is potentially useful for detecting physical exhaustion and for monitoring critical care conditions characterized by hypoperfusion, such as heart failure. In some conditions, it may be desirable to monitor more than one metabolic parameter concurrently. For this reason, we designed and fabricated twisted wire-based microelectrodes that can measure both lactate and glucose. These dual-analyte sensors were characterized in vitro by measuring their response to the analyte of interest and to assess whether they were susceptible to interference from the other analyte. When measured in stirred aqueous buffer, lactate sensors detected a very small amount of crosstalk from glucose in vitro, although this signal was less than 3% of the response to lactate. Glucose sensors did not detect crosstalk from lactate. Sensors were implanted subcutaneously in rats and tested during infusions of lactate and glucose. Each sensing electrode responded rapidly to changes in its analyte concentration, and there was no evidence of in vivo crosstalk. This study constitutes proof of the concept that oxidase-based, amperometric wire microsensors can detect changes in glucose and lactate during subcutaneous implantation in rats.

The effect of local subcutaneous delivery of vascular endothelial growth factor on the function of a chronically implanted amperometric glucose sensor.

The foreign body capsule that forms around implanted devices such as glucose sensors is hypovascular and has limited permeability to glucose. Such a capsule may function better if well vascularized. We hypothesized that capsular vascularization achieved by local release of vascular endothelial growth factor (VEGF) would lead to enhanced function. Amperometric glucose sensor array disks, each with four indicating electrodes, were implanted into rats. Animals received local subcutaneous infusions of VEGF(165) via osmotic pumps at a location on the sensor face 2 mm from one of the electrodes ("near units"). "Intermediate" electrode units were 15 mm, and "distant" units were 22 mm, from the VEGF source. Every 2 weeks, a glucose infusion was given to assess sensor function by telemetry. Near units demonstrated a lower lag duration (delay after blood glucose) than intermediate and distant units. The mean absolute relative difference for near units was less than for distant units. The percentage of data pairs in the A region of the Clarke error grid of the near sensing units was greater than that of the distant units. Values for the functional measures for saline controls fell between near and distant VEGF values. Glucose sensor function was found to be more favorable in units immediately adjacent to the VEGF infusion port. The most likely cause for this finding is increased neovessel growth in the surrounding foreign body capsule. Slow release of angiogenic growth factors may be a potential method for chronically enhancing the function of a subcutaneously implanted biosensor.

A fully implantable subcutaneous glucose sensor array: enhanced accuracy from multiple sensing units and a median-based algorithm.

Although continuous electrochemical glucose monitoring holds promise in the management of diabetes, its utility is limited in part because of error of unclear origin. The use of redundant glucose sensors in an array might reduce such error. We hypothesized that in a subcutaneously implanted array, a median-based continuous computation that excludes outlying data would lead to more accurate glucose measurement than averaging of all signals. Each rat was implanted with an array of four sensing units, and each unit transmitted data independently to an external monitoring device. Animals underwent perturbation of glucose by insulin infusions in diabetic animals and glucose infusions in nondiabetic animals, and in both, capillary glucose monitoring was performed frequently. Repeat glucose perturbation studies were performed every 1-2 weeks. We observed that a median-based technique, the Z-score with Median Absolute Deviation (ZMAD), consistently led to greater sensing accuracy as compared with signal averaging. The ZMAD technique yielded a correlation coefficient of 0.93, and 96% of values fell in the A and B regions of the Clarke error grid, demonstrating a high degree of accuracy of the unified signal. When tested in an implanted array of glucose sensors, a median-based technique (ZMAD) yields an accurate unified signal, and its accuracy is superior to signal averaging.

Induction of type-1 diabetes mellitus in laboratory rats by use of alloxan: route of administration, pitfalls, and insulin treatment.

Uncertainties have existed regarding the systematic induction and management of drug-induced diabetes mellitus (DM). Issues have included the optimal route of administration of the drug, methods of reducing drug toxicosis and mortality, how to induce type-1 versus type-2 DM, and how to manage labile DM in rats. In attempting to induce type-1 DM in Sprague-Dawley rats, we classified hyperglycemic animals as having type-1 DM only if their post-treatment blood ketone concentration was high. We found that multiple doses of alloxan led to significantly higher mortality than did a single dose. A single high dose (200 mg/kg of body weight given intraperitoneally) was the best treatment and led to 70% incidence of type-1 DM and only 10% mortality. In contrast, intravenous administration of similar doses was toxic. Assiduous management of alloxan-induced DM is crucial to avoid severe hypoglycemia from massive insulin release and to avoid diabetic ketoacidosis. Frequent glucose monitoring and appropriate administration of carbohydrate and fluids is necessary during this stage. For long-term management, daily administration of long-acting insulin (glargine) appears to be safe and effective. Rapid-acting insulins reduce glucose concentration rapidly, and must be used with caution. If specific precautions are observed, intraperitoneal administration of high-dose alloxan to laboratory rats leads to a condition that closely resembles human type-1 DM.

Can glucose be monitored accurately at the site of subcutaneous insulin delivery?

Because insulin promotes glucose uptake into adipocytes, it has been assumed that during measurement of glucose at the site of insulin delivery, the local glucose level would be much lower than systemic glucose. However, recent investigations challenge this notion. What explanations could account for a reduced local effect of insulin in the subcutaneous space? One explanation is that, in humans, the effect of insulin on adipocytes appears to be small. Another is that insulin monomers and dimers (from hexamer disassociation) might be absorbed into the circulation before they can increase glucose uptake locally. In addition, negative cooperativity of insulin action (a lower than expected effect of very high insulin concentrations)may play a contributing role. Other factors to be considered include dilution of interstitial fluid by the insulin vehicle and the possibility that some of the local decline in glucose might be due to the systemic effect of insulin. With regard to future research, redundant sensing units might be able to quantify the effects of proximity, leading to a compensatory algorithm. In summary, when measured at the site of insulin delivery, the decline in subcutaneous glucose level appears to be minimal, though the literature base is not large. Findings thus far support (1) the development of integrated devices that monitor glucose and deliver insulin and (2) the use of such devices to investigate the relationship between subcutaneous delivery of insulin and its local effects on glucose. A reduction in the number of percutaneous devices needed to manage diabetes would be welcome.

Modeling the glucose sensor error.

Continuous glucose monitoring (CGM) sensors are portable devices, employed in the treatment of diabetes, able to measure glucose concentration in the interstitium almost continuously for several days. However, CGM sensors are not as accurate as standard blood glucose (BG) meters. Studies comparing CGM versus BG demonstrated that CGM is affected by distortion due to diffusion processes and by time-varying systematic under/overestimations due to calibrations and sensor drifts. In addition, measurement noise is also present in CGM data. A reliable model of the different components of CGM inaccuracy with respect to BG (briefly, "sensor error") is important in several applications, e.g., design of optimal digital filters for denoising of CGM data, real-time glucose prediction, insulin dosing, and artificial pancreas control algorithms. The aim of this paper is to propose an approach to describe CGM sensor error by exploiting n multiple simultaneous CGM recordings. The model of sensor error description includes a model of blood-to-interstitial glucose diffusion process, a linear time-varying model to account for calibration and sensor drift-in-time, and an autoregressive model to describe the additive measurement noise. Model orders and parameters are identified from the n simultaneous CGM sensor recordings and BG references. While the model is applicable to any CGM sensor, here, it is used on a database of 36 datasets of type 1 diabetic adults in which n = 4 Dexcom SEVEN Plus CGM time series and frequent BG references were available simultaneously. Results demonstrates that multiple simultaneous sensor data and proper modeling allow dissecting the sensor error into its different components, distinguishing those related to physiology from those related to technology.

Automated control of an adaptive bihormonal, dual-sensor artificial pancreas and evaluation during inpatient studies.

Automated control of blood glucose in patients with type-1 diabetes has not yet been fully implemented. The aim of this study was to design and clinically evaluate a system that integrates a control algorithm with off-the-shelf subcutaneous sensors and pumps to automate the delivery of the hormones glucagon and insulin in response to continuous glucose sensor measurements. The automated component of the system runs an adaptive proportional derivative control algorithm which determines hormone delivery rates based on the sensed glucose measurements and the meal announcements by the patient. We provide details about the system design and the control algorithm, which incorporates both a fading memory proportional derivative controller (FMPD) and an adaptive system for estimating changing sensitivity to insulin based on a glucoregulatory model of insulin action. For an inpatient study carried out in eight subjects using Dexcom SEVEN PLUS sensors, prestudy HbA1c averaged 7.6, which translates to an estimated average glucose of 171 mg/dL. In contrast, during use of the automated system, after initial stabilization, glucose averaged 145 mg/dL and subjects were kept within the euglycemic range (between 70 and 180 mg/dL) for 73.1% of the time, indicating improved glycemic control. A further study on five additional subjects in which we used a newer and more reliable glucose sensor (Dexcom G4 PLATINUM) and made improvements to the insulin and glucagon pump communication system resulted in elimination of hypoglycemic events. For this G4 study, the system was able to maintain subjects' glucose levels within the near-euglycemic range for 71.6% of the study duration and the mean venous glucose level was 151 mg/dL.