Abstract: The prevalence of type 1 diabetes is escalating worldwide. Novel therapies and management strategies are needed to reduce associated morbidity. Aggressive blood glucose lowering using conventional insulin replacement regimens is limited by the risk of hypoglycemia. Even the most motivated patients may struggle to manage day-to-day variability in insulin requirements. The artificial pancreas or closed-loop insulin delivery may improve outcomes, building on recent technological progress and combining continuous glucose monitoring with insulin pump therapy. So far, closed-loop prototypes have been evaluated under controlled conditions suggesting improved glucose control and a reduced risk of hypoglycemia. Limitations include suboptimal accuracy and reliability of continuous glucose monitors and delays associated with subcutaneous insulin delivery. Outpatient evaluation is required as the next step, leading to deployment into clinical practice.
Type 1 diabetes, which accounts for approximately 5 to 10% of all diabetes cases, is characterized by autoimmune destruction of insulin-secreting beta cells in genetically predisposed individuals resulting in complete or near-complete loss of insulin production. Insulin, administered via multiple daily injections or continuous subcutaneous insulin infusion (CSII; insulin pump therapy), remains the only treatment option. Potentially curative therapies, including immunomodulatory agents and islet cell transplantation, have demonstrated some success, but neither approach has resulted in long term insulin independence (Ryan et al., 2005). Intensive insulin therapy aimed at achieving normal glucose levels has been shown to decrease the risk of microvascular (retinopathy, neuropathy, and nephropathy) and macrovascular (coronary and peripheral vessel disease) complications, but is associated with an increased risk of hypoglycemia (Nathan et al., 1993). This risk is even greater in those with reduced awareness of hypoglycemia, associated with recurrent episodes and diabetes-related autonomic dysfunction (Cryer, 2004). Glucose variability may increase diabetes-related morbidity (Weber and Schnell, 2009). Regular self-monitoring of blood glucose and the use of insulin analogues is an essential component of modern management approaches.
Closed-loop Insulin Delivery
The emergence of new technologies including smart insulin pumps and continuous glucose monitoring has contributed to improved diabetes care. Use of “sensor-augmented pumps” have been shown to improve glycemic control, compared with multiple daily injection therapy (Bergenstal et al., 2010). However, this still demands considerable patient participation including blood glucose testing, counting carbohydrates, and estimating insulin doses to be administered. Even the most diligent patients may struggle to achieve optimal glycemic control. Fine tuning of insulin regimens is feasible during the daytime whilst awake. The requirement for patients to do multiple testing and adjusting of insulin infusions may be minimized with use of an artificial pancreas or closed-loop insulin delivery, which combines glucose sensing via a minimally invasive continuous glucose monitor (CGM) and insulin pump delivery directed by a control algorithm, based on real-time glucose readings (Figure 1). The system imitates the feedback loop inherent to the pancreatic beta cells to control glucose levels (Figure 2).
The concept of closed-loop glucose control was first introduced in 1964 (Kadish, 1964), with follow-up pioneering work by Albisser and Pfeiffer (Albisser et al., 1974; Pfeiffer et al., 1974) resulting in the development of the first commercial artificial pancreas in 1977 (Clemens et al., 1977). The device known as Biostator (Miles Laboratories, Elkhart, IN, USA) combined minute-by-minute glucose monitoring from whole blood via a glucose oxidase sensor and intravenous infusions of insulin and dextrose. Drawbacks were related to simplified algorithms, risk of infection and venous thrombosis, wastage of venous blood, and the need for constant technical supervision to maintain patency of the sampling catheter. Nowadays, the focus has turned to other body access routes to bypass these limitations.
Body access routes
The Biostator established feasibility of closed-loop glucose control, but adopted the intravenous route for glucose sensing and insulin delivery limiting its use to research settings. The intraperitoneal delivery route has been tested in a system combining a sensor placed in the superior vena cava connected via a subcutaneous lead to an implanted intraperitoneal insulin pump (Renard et al., 2006). Feasibility of a closed-loop approach employing intraperitoneal insulin delivery with subcutaneous glucose sensing has been carried out (Renard et al., 2010). Intraperitoneal insulin delivery more closely mimics physiologic insulin delivery but requires surgery and is associated with risk of infection and catheter occlusion due to insulin aggregation. Subcutaneous glucose sensing and subcutaneous insulin delivery is the most viable approach given the widespread use of insulin pumps (Pickup and Keen, 2002) and commercial availability of subcutaneous glucose monitors (Klonoff, 2005). As this approach is used by most closed-loop systems currently under development, the review will focus on this particular configuration (Figure 3).
Continuous glucose monitor
Measurement of glucose concentration in the interstitial fluid is the most promising minimally invasive route for outpatient monitoring of glucose. The existing commercial sensors are inserted directly into the subcutaneous tissue and measure electric current generated by the oxidation of glucose via the enzyme glucose oxidase. Although less accurate than capillary glucose, CGM has several advantages over discrete finger stick measurements including continuous measurements of glucose in real time, prediction of future glycemic excursions and trends, and the ability to set alarms alerting patients to hypo- and hyperglycemia. Widespread use of CGM has been limited by its relatively high cost in countries without healthcare reimbursement schemes (Hermanides and DeVries, 2010), and device-related issues such as bulkiness and adhesive problems (Halford and Harris, 2010). Several studies have demonstrated a significant reduction in HbA1c with CGM use, especially in adults and those with worse glycemic control at baseline (Bergenstal et al., 2010; Hirsch et al., 2008; Pickup et al., 2011; Tamborlane et al., 2008). However, the clinical benefits seen are dependent on compliance with wearing the device which tends to decline over time, particularly in the younger age group (Chase et al., 2010; Kordonouri et al., 2010). Older age, increased frequency of blood glucose monitoring, and higher initial use of CGM are strong predictors of ongoing use of the device (Beck et al., 2009).
Accuracy of the commercially available CGM devices, as measured by the median relative absolute difference between sensor and reference glucose, ranges from 11 to 14% (Garg et al., 2009; Kamath et al., 2009). This is related to calibration errors, sensor artifacts, and a physiological time lag of transport of glucose between blood and the interstitial fluid as well as a device-dependent delay associated with sensor signal processing and filtering out measurement noise. Lag times of 6 minutes and 13 minutes have been reported for the DexCom (DexCom, San Diego, CA, USA) and FreeStyle Navigator (Abbott Diabetes Care, Alameda, CA, USA) CGM devices, respectively (Kamath et al., 2009; Kovatchev et al., 2009b).
Insulin pumps deliver insulin in a manner mimicking the pancreatic beta cell and have been available commercially for the past 30 years (Pickup and Keen, 2002). Current devices allow setting of multiple basal rates and incorporate sophisticated bolus calculators which advise on boluses to be administered for carbohydrates consumed (Bruttomesso et al., 2009). The major limitation of CSII is its inability to respond to glucose levels in real time, instead relying on manual adjustments of pre-set infusion rates and carbohydrate ratios. Delivery of the insulin bolus itself may be overlooked — a study in Swedish adolescents found that over a third had missed more than 15% of boluses, which was correlated with worse glycemic control (Olinder et al., 2009).
The recent emergence of “sensor-augmented pumps” (SAP), including the Paradigm Veo (Medtronic, Northridge, CA, USA) and the Animas Vibe (Johnson & Johnson, New Brunswick, NJ, USA), has provided patients with real time information on glucose values and trends based on CGM, as well as the ability to set alarms for impending hypo- or hyperglycemia. Manual input and adjustment of insulin basal rates and boluses are still required.
There are two main groups of control algorithm used in artificial pancreas prototypes — model predictive control (MPC) and proportional integral derivative (PID) (Bequette, 2005). Fuzzy logic control is another algorithmic approach, which approximates reasoning to replicate conventional insulin dosing instructions by diabetes practitioners.
Model predictive control
The MPC approach is at the forefront of current research as it accommodates delays associated with insulin absorption and can also easily account for meal intake and prandial insulin boluses instigated by the patient (Hovorka et al., 2004). The vital ingredient of MPC is a model that links insulin delivery and meal ingestion to glucose excursions. This may include a physiological model representing fundamental glucoregulatory processes or a “black-box” model which learns the insulin-glucose relationships using pattern recognition techniques. Information on glucose measurements may be used to update the model parameters such as insulin sensitivity. Prediction capabilities make MPC a suitable approach for use with longer time delays, such as with subcutaneous insulin delivery.
Proportional integral derivative algorithm
The PID algorithm, used widely in control systems in industrial settings, calculates insulin delivery based on three terms (Steil et al., 2004). The proportional (P) term adjusts insulin delivery in response to current glucose levels. The integral (I) component adjusts insulin delivery according to the area under the curve between measured and target glucose. The derivative (D) term delivers insulin in response to the rate of change of blood glucose over time. PID control is limited in its ability to anticipate future glucose excursions, and to overcome the inherent delays between insulin delivery and detection of its effect associated with the sc-sc approach. A variation on PID control which replaces the integral component with a “fading memory” where more recent glucose levels have a higher weighting has been investigated (Castle et al., 2010; El Youssef et al., 2009). Addition of insulin feedback has been proposed to improve the ability of PID to accommodate delayed insulin absorption (Steil et al., 2011).
Several common lifestyle factors pose distinct challenges to closed-loop insulin delivery.
Consumption of carbohydrates is associated with accelerated transient glycemic excursions, the magnitude of which is dependent on several factors including meal size and composition, as well as timing and amount of insulin administered. The delay in absorption of subcutaneously delivered insulin often results in postprandial hyperglycemia, which may inadvertently lead to administration of correction boluses and insulin stacking, thus increasing the risk of delayed hypoglycemia. This is also applicable to closed-loop insulin delivery. Administration of rapid acting insulin as a simple bolus 15 minutes prior to eating has been shown to result in lower peak glucose and greater time in target during conventional pump therapy (De Palma et al., 2011; Luijf et al., 2010).
Physical activity may have complex and unpredictable effects on glucose control, which together with the fear of hypoglycemia, often results in avoidance of exercise despite awareness of its positive health benefits (Brazeau et al., 2008). These effects may be immediate or delayed up to several hours after exercise is completed (McMahon et al., 2007). The presence of exogenously administered insulin, exercise-induced enhanced peripheral insulin sensitivity, and blunted counter-regulatory responses all contribute to increased occurrence of hypoglycemia (Cryer, 2009). Strenuous exercise, in contrast, may result in elevated glucose levels (Marliss and Vranic, 2002). Ideally, physical activity should be carefully planned with consumption of carbohydrates and/or reduction in insulin doses according to glucose testing performed prior to and following exercise. This is not always practical, particularly in the younger age group where physical activity is more spontaneous.
Acute illness in patients with diabetes is associated with elevated glucose levels, attributed to increased secretion of counter-regulatory hormones including catecholamines, cortisol, and glucagon which increase hepatic glucose production and peripheral insulin resistance (Sheetz and King, 2002). Simulation of this stress response by administration of synthetic corticosteroids in ten patients with type 1 diabetes resulted in a 69% higher daily insulin requirement (Bevier et al., 2008).
Individual variability in insulin sensitivity
Insulin absorption and action varies within and between individuals, rendering individual tailoring of insulin regimens to achieve glycemic targets more challenging (Heinemann, 2002). Reasons include age, diurnal rhythms, physical activity and concurrent illness, as well as local factors such as insulin type and site of administration.
Clinical Studies with Closed-loop Systems
Low glucose suspend and pump shut-off
Suspension of insulin delivery during hypoglycemia is the most straightforward application of an early first generation closed-loop system. The Paradigm Veo (Medtronic, Northridge, CA, USA), approved for commercial use in Europe since June 2009 but awaiting FDA authorization, can be linked with the SOF-SENSOR or Enlite sensor (Medtronic, Northridge, CA, USA), suspending insulin delivery for up to two hours when glucose falls below a threshold. A user evaluation study in 31 adults showed that the low glucose suspend feature was activated 166 times (mean 1.9 times per patient per week), with 76% occurring in the daytime (Choudhary et al., 2011). Nocturnal hypoglycemia (≤2.2 mmol/l) was reduced from 46.2 to 1.8 minutes/day in those with highest baseline risk. Recent assessment of the low glucose suspend feature in children demonstrated reduction of time spent in hypoglycemia from 101 to 58 minutes/day (Danne et al., 2011). Almost 25% of pump suspension events lasted the maximum two hours, with the majority occurring overnight.
Pump shut-off during the daytime using two different hypoglycemia prediction algorithms was evaluated in 22 adult and young patients — at a threshold of 4.4 mmol/l, hypoglycemia was prevented on 60% of occasions using a statistical algorithm, and 75% using a linear algorithm (Buckingham et al., 2009). A combination of up to five hypoglycemia prediction algorithms was evaluated overnight, using a threshold of 4.4 mmol/l with a 35 min prediction horizon (Buckingham et al., 2010). Hypoglycemia (<3.3 mmol/l) was prevented on 60% of nights when three algorithms were used to predict pump suspension, and on 75% of nights when two algorithms were used.
The main safety concern with interruption of insulin delivery, which may occur under normal closed-loop control, is the risk of hyperglycemia and associated metabolic ketosis. There were seven episodes of mild ketosis in the two aforementioned studies in total, none of which were associated with symptoms or clinical sequelae (Buckingham et al., 2010; 2009). Insulin suspension for up to 240 minutes during closed-loop studies using MPC in children resulted in a peak glucose of 11.6 mmol/l with no metabolic derangement (Elleri et al., 2010b). Similarly, during closed-loop studies employing PID control in children, there were 18 occurrences of insulin suspension for at least 60 minutes with no occurrence of significant hyperglycemia or ketosis (Cengiz et al., 2009).
Improving glucose control whilst asleep, when fear of hypoglycemia is greatest and the counter regulatory response to hypoglycemia is blunted, may be of considerable benefit. The DCCT trial showed that 55% of severe hypoglycemia episodes occurred during sleep (Nathan et al., 1991). Overnight closed-loop using an MPC algorithm has been evaluated in hospital in randomized crossover studies in 19 children and 24 adults (Kumareswaran et al., 2011) (Figure 4). Compared with conventional pump therapy, closed-loop in children increased time in target glucose range (3.9-8.0 mmol/l) from 40% to 60%, and halved the time spent in hypoglycemia (<3.9 mmol/l) (Hovorka, et al., 2010). In adults, time in target improved from 50% to 76% during closed-loop, and hypoglycemia was reduced from 7% to 3% (Hovorka et al., 2011). Overnight closed-loop was also tested in a multicenter trial in 20 adults using an MPC controller developed in silico (Kovatchev et al., 2010). Time in target glucose range (3.9-7.8 mmol/l) increased from 64% to 78% and hypoglycemic episodes (<3.9 mmol/l) were reduced from 23 to 5, compared with conventional pump treatment. A prototype of fully automated closed-loop system was evaluated in 8 young children, with no differences in glycemic control with early (18:00h) versus late (21:00h) initiation of overnight closed-loop insulin delivery (Elleri et al., 2011b).
Closed-loop with meal announcement
Meal announcement may help overcome the delays associated with absorption of subcutaneously delivered insulin and appearance of meal-related glucose in the bloodstream. In this approach, information on size and timing of meals is provided to the algorithm and prandial insulin is administered manually, with fully closed-loop control at other times (Figure 5). Twelve adolescents were evaluated in a randomized crossover study using MPC-driven closed-loop over 36 hours, mimicking a typical day at school (Elleri et al., 2011a). Compared with conventional pump therapy, closed-loop improved overall time in target glucose (3.9-10.0 mmol/l) from 49% to 84%, with 100% time in target overnight. Although there was no difference in daytime frequency of hypoglycemia between interventions, closed-loop prevented nocturnal episodes. This approach was also evaluated in 12 women with type 1 diabetes during pregnancy using closed-loop over 24 hours in a randomized crossover design, demonstrating similar efficacy to conventional pump therapy (Figure 4) (Murphy et al., 2011b). Prior feasibility testing in this patient cohort demonstrated safety of overnight closed-loop in early and late gestation (Murphy et al., 2011a).
Closed-loop without meal announcement
Feasibility of fully closed-loop was first evaluated using a PID algorithm over 29 hours in 10 adults showing 75% time in target (3.9-10.0 mmol/l), compared with 63% during open loop control at home (Steil et al., 2006). A “hybrid” closed-loop system was tested in 17 children using a PID algorithm with manual administration of a priming (25-50% of the total) bolus 10-15 minutes before each meal, resulting in lower mean and postprandial peak glucose levels, compared with fully automated closed-loop (Weinzimer et al., 2008). This semi closed-loop approach was evaluated in 8 adults using a PID algorithm with insulin feedback over 30 hours, with a manual bolus of 2 U delivered at the start of each meal (Steil et al., 2011). Although satisfactory mean postprandial glucose levels were achieved, hypoglycemia was not eliminated and supplemental carbohydrates were administered on 8 occasions. A feasibility study employing fuzzy logic control in a fully closed-loop system (MD-Logic Artificial Pancreas system) over 8 (12 occasions) or 24 hours (two occasions), demonstrated 73% of sensor values between 3.9-10.0 mmol/l with no symptomatic hypoglycemia (Atlas et al., 2010).
Co-administration of glucagon, a hormone which counters the effect of insulin, as part of a closed-loop system was evaluated in 11 adults during 24 hours of fully closed-loop control using an adaptive MPC algorithm (El-Khatib et al., 2010). Application of a suitable model of insulin absorption prevented occurrence of hypoglycemia. The use of combined glucagon and insulin was evaluated in a semi-closed-loop system under fading memory proportional derivative control with manual delivery of 50-75% of the bolus pre-meals, demonstrating 63% lower time spent in hypoglycemia compared with insulin alone (Castle et al., 2010). Compared with low gain, glucagon delivered using high gain parameters (pulses over 5-10 minutes followed by a 50-minute off period) reduced frequency of hypoglycemia even further. There was only one report of nausea and vomiting associated with glucagon delivery during a low gain study. Limitations of glucagon use include its instability and tendency to form amyloid fibrils in solution (Pedersen, 2010), as well as potential depletion of liver glycogen stores with repeated administration.
Compared with clinical studies, which may be time consuming and costly to perform, computer-based simulated studies enable rapid systems evaluation without requirement for ethical or regulatory approvals. Simulators provide invaluable information on the effect of various parameters or scenarios that may be encountered in real life such as sensor errors, pump occlusion, hypoglycemic episodes, exercise, and unannounced meals (Wilinska et al., 2010). In January 2008, the FDA approved an in silico simulation environment as an alternative to animal trials for pre-clinical closed-loop studies (Kovatchev et al., 2009a).
Further research is required to address the current hurdles faced in improving system performance, including refinement of control algorithms to cope with variability in insulin requirements and development of more accurate glucose sensors. The immediate benefits of short-term (up to 48 hours) application of closed-loop on glycemic control have been demonstrated in studies carried out in the inpatient setting. Transition to outpatient studies will be required to assess efficacy of using the system over an extended period under free living conditions.
Combining insulin delivery and continuous glucose sensing into a single device is likely to increase device acceptability and hence compliance with wear. Stability of glucose levels measured at the site of insulin delivery was demonstrated in a study in humans (Lindpointner et al., 2010). However, glycemic variation was seen between sensors placed 0.5 cm and 3 cm from the site of insulin injection in healthy minipigs (Rodriguez et al., 2011).
One of the major limitations of subcutaneous insulin delivery is the delay in time to peak blood glucose lowering effect which may be 90-120 minutes with the currently available rapid acting insulin analogues (Hovorka, 2006). Potential solutions may include improving the molecule itself or its delivery. There are at least two new insulin formulations under development. VIAject (Biodel, Danbury, CT, USA) is an ultra rapid acting insulin analogue which retains insulin in its monomeric form allowing more rapid onset of action (Steiner et al., 2008). Co-administration of insulin with synthetic human hyaluronidase (Halozyme Therapeutics, San Diego, CA, USA) facilitates diffusion and thus faster insulin action via cleavage of hyaluronic acid, a space-filling substance in body tissues (Vaughn et al., 2009).
Insulin delivered via the intraperitoneal route achieves maximal action within 15 minutes with resulting reduced rate of severe hypoglycemia (Liebl et al., 2009). However, long term use is limited by catheter-related complications. Inhaled insulin (Technosphere; MannKind Corp., Valencia, CA, USA) has a more rapid onset and shorter duration of action than subcutaneous insulin (Rave et al., 2009). Intradermal insulin injection using 1.5 mm steel microneedles is another potential route being explored (Pettis et al., 2011). Application of a local heating device at the site of insulin injection has been shown to accelerate absorption of prandial insulin boluses (Raz et al., 2009). Another approach to overcome the delay in insulin action is to reduce the rate of appearance of meal-related glucose. Pramlintide, the synthetic analogue of the hormone amylin, acts to slow gastric emptying and secretion of digestive enzymes in addition to inhibiting release of glucagon, insulin’s counter-regulatory hormone. Co-administration of pramlintide with insulin at mealtimes may reduce postprandial hyperglycemia (Weyer et al., 2003).
Prior to employment of closed-loop systems in clinical practice, strict safety checks by regulatory bodies are essential. An audit of U.S. FDA medical device recalls found that the majority were approved via less rigorous processes, recommending stricter practices to ensure device safety (Zuckerman et al., 2011). Ongoing safety monitoring in addition to robust infrastructure to manage technical issues will be necessary. Telemedicine may play an increasing role in closed-loop systems, enabling remote monitoring and logging of data. A prototype CGM device equipped with global positioning system technology has been proposed as a way of alerting family and medical personnel of the location of patients in the event of severe or impending hypoglycemia (Dassau et al., 2009). Patient acceptance of future use of a closed-loop system is likely to be positive based on qualitative evaluation in adult patients (van Bon et al., 2010), and caregivers of children with type 1 diabetes (Elleri et al., 2010a).
Even with the currently available sophisticated insulin pumps and continuous glucose monitoring devices, the majority of patients with diabetes struggle to achieve optimal glycemic control. The artificial pancreas may offer a more convenient and superior mode of insulin delivery. Introduction of closed-loop insulin delivery into clinical practice is likely to be gradual, commencing with simpler approaches such as pump suspension during hypoglycemia and application overnight, eventually progressing to more complex situations including control during meals and exercise.
Supported by Juvenile Diabetes Research Foundation (#22-2006-1113, #22-2007-1801, #22-2009-801), Diabetes UK (BDA07/0003549, BDA07/0003551), European Commission Framework Program 7 (247138), NIDDK (DK085621), and NIHR Cambridge Biomedical Research Centre.
K.K. has no competing interests. M.L.E. reports having received speaker honoraria from Eli Lilly and Takeda and serving on advisory panels for Medtronic, Roche, Sanofi-Aventis, and Cellnovo. R.H. reports having received speaker honoraria from Minimed Medtronic, Lifescan, and Novo Nordisk, serving on advisory panel for Animas and Minimed Medtronic, receiving license fees from BBraun and Becton Dickinson, and having served as a consultant to Becton Dickinson, BBraun, and Profil.
Roman Hovorka, Ph.D., University of Cambridge Metabolic Research Laboratories and NIHR Cambridge Biomedical Research Centre, Level 4, Institute of Metabolic Science, Box 289, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 0QQ, United Kingdom.
Albisser AM, Leibel BS, Ewart TG, Davidova Z, Botz CK, Zingg W. Artificial endocrine pancreas. Diabetes 23(5):389-396, 1974.
Atlas E, Nimri R, Miller S, Grunberg EA, Phillip M. MD-logic artificial pancreas system: a pilot study in adults with type 1 diabetes. Diabetes Care 33(5):1072-1076, 2010.
Beck RW, Buckingham B, Miller K, Wolpert H, Xing D, Block JM, Chase HP, Hirsch I, Kollman C, Laffel L, Lawrence JM, Milaszewski K, Ruedy KJ, Tamborlane WV. Factors predictive of use and of benefit from continuous glucose monitoring in type 1 diabetes. Diabetes Care 32(11):1947-1953, 2009.
Bequette BW. A critical assessment of algorithms and challenges in the development of a closed-loop artificial pancreas. Diabetes Technol Ther 7(1):28-47, 2005.
Bergenstal RM, Tamborlane WV, Ahmann A, Buse JB, Dailey G, Davis SN, Joyce C, Peoples T, Perkins BA, Welsh JB, Willi SM, Wood MA, Grp SS. Effectiveness of sensor-augmented insulin-pump therapy in type 1 diabetes. New Engl J Med 363(4):311-320, 2010.
Bevier WC, Zisser HC, Jovanovic L, Finan DA, Palerm CC, Seborg DE, Doyle FJ, 3rd. Use of continuous glucose monitoring to estimate insulin requirements in patients with type 1 diabetes mellitus during a short course of prednisone. J Diabetes Sci Technol 2(4):578-583, 2008.
Brazeau A, Rabasa-Lhoret R, Strychar I, Mircescu H. Barriers to physical activity among patients with type 1 diabetes. Diabetes care 31(11):2108, 2008.
Bruttomesso D, Costa S, Baritussio A. Continuous subcutaneous insulin infusion (CSII) 30 years later: still the best option for insulin therapy. Diabetes Metab Res Rev 25(2):99-111, 2009.
Buckingham B, Chase HP, Dassau E, Cobry E, Clinton P, Gage V, Caswell K, Wilkinson J, Cameron F, Lee H, Bequette BW, Doyle FJ. Prevention of nocturnal hypoglycemia using predictive alarm algorithms and insulin pump suspension. Diabetes Care 33(5):1013-1017, 2010.
Buckingham B, Cobry E, Clinton P, Gage V, Caswell K, Kunselman E, Cameron F, Chase HP. Preventing hypoglycemia using predictive alarm algorithms and insulin pump suspension. Diabetes Technol Ther 11(2):93-97, 2009.
Castle JR, Engle JM, Youssef JEL, Massoud RG, Yuen KCJ, Kagan R, Ward WK. Novel use of glucagon in a closed-loop system for prevention of hypoglycemia in type 1 diabetes. Diabetes Care 33(6):1282-1287, 2010.
Cengiz E, Swan KL, Tamborlane WV, Steil GM, Steffen AT, Weinzimer SA. Is an automatic pump suspension feature safe for children with type 1 diabetes? An exploratory analysis with a closed-loop system. Diabetes Technol Ther 11(4):207-210, 2009.
Chase HP, Beck RW, Xing DY, Tamborlane WV, Coffey J, Fox LA, Ives B, Keady J, Kollman C, Laffel L, Ruedy KJ. Continuous glucose monitoring in youth with type 1 diabetes: 12-month follow-up of the Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Randomized Trial. Diabetes Technol Ther 12(7):507-515, 2010.
Choudhary P, Shin J, Wang Y, Evans ML, Hammond PJ, Kerr D, Shaw JA, Pickup JC, Amiel SA. Insulin pump therapy with automated insulin suspension in response to hypoglycemia: reduction in nocturnal hypoglycemia in those at greatest risk. Diabetes Care 34(9):2023-2025, 2011.
Clemens AH, Chang PH, Myers RW. The development of Biostator, a glucose controlled insulin infusion system (GCIIS). Horm Metab Res 7(Suppl):23-33, 1977.
Cryer PE. Current concepts: diverse causes of hypoglycemia-associated autonomic failure in diabetes. New Engl J Med 350(22):2272-2279, 2004.
Cryer PE. Exercise-related hypoglycemia-associated autonomic failure in diabetes. Diabetes 58(9):1951-1952, 2009.
Danne T, Kordonouri O, Holder M, Haberland H, Golembowski S, Remus K, Blasig S, Wadien T, Zierow S, Hartmann R, Thomas A. Prevention of hypoglycemia by using low glucose suspend function in sensor-augmented pump therapy. Diabetes Technol Ther 13(11):1129-1134, 2011.
Dassau E, Jovanovic L, Doyle FJ, 3rd, Zisser HC. Enhanced 911/global position system wizard: a telemedicine application for the prevention of severe hypoglycemia — monitor, alert, and locate. J Diabetes Sci Technol 3(6):1501-1506, 2009.
De Palma A, Giani E, Iafusco D, Bosetti A, Macedoni M, Gazzarri A, Spiri D, Scaramuzza AE, Zuccotti GV. Lowering postprandial glycemia in children with type 1 diabetes after Italian Pizza “Margherita” (TyBoDi2 Study). Diabetes Technol Ther 13(4):483-487, 2011.
El-Khatib FH, Russell SJ, Nathan DM, Sutherlin RG, Damiano ER. A bihormonal closed-loop artificial pancreas for type 1 diabetes. Sci Transl Med 2(27), 2010.
El Youssef J, Castle J, Ward W. A review of closed-loop algorithms for glycemic control in the treatment of type 1 diabetes. Algorithms 2(1):518-532, 2009.
Elleri D, Acerini Cl, Allen JM, Hayes J, Pesterfield C, Wilinska ME, Dunger DB, Hovorka R. Parental attitudes towards overnight closed-loop glucose control in children with type 1 diabetes. Diabetes Technol Ther 12(1):35-39, 2010a.
Elleri D, Allen J, Kumareswaran K, Leelarathna L, Nodale M, Caldwell K, Murphy H, Wilinska M, Acerini C, Dunger D, Hovorka R. Day-and-night closed-loop (CL) glucose control in adolescents with type 1 diabetes (T1D). Diabetes 60:A41, 2011a.
Elleri D, Allen JM, Nodale M, Wilinska ME, Acerini CL, Dunger DB, Hovorka R. Suspended insulin infusion during overnight closed-loop glucose control in children and adolescents with type 1 diabetes. Diabet Med 27(4):480-484, 2010b.
Elleri D, Allen JM, Nodale M, Wilinska ME, Mangat JS, Larsen AMF, Acerini CL, Dunger DB, Hovorka R. Automated overnight closed-loop glucose control in young children with type 1 diabetes. Diabetes Technol Ther 13(4):419-424, 2011b.
Elleri D, Dunger DB, Hovorka R. Closed-loop insulin delivery for treatment of type 1 diabetes. BMC Med 9(120):1-9, 2011c.
Garg SK, Smith J, Beatson C, Lopez-Baca B, Voelmle M, Gottlieb PA. Comparison of accuracy and safety of the SEVEN and the Navigator continuous glucose monitoring systems. Diabetes Technol Ther 11(2):65-72, 2009.
Halford J, Harris C. Determining clinical and psychological benefits and barriers with continuous glucose monitoring therapy. Diabetes Technol Ther 12(3):201-205, 2010.
Heinemann L. Variability of insulin absorption and insulin action. Diabetes Technol Ther 4(5):673-682, 2002.
Hermanides J, DeVries JH. Sense and nonsense in sensors. Diabetologia 53(4):593-596, 2010.
Hirsch IB, Abelseth J, Bode BW, Fischer JS, Kaufman FR, Mastrototaro J, Parkin CG, Wolpert HA, Buckingham BA. Sensor-augmented insulin pump therapy: results of the first randomized treat-to-target study. Diabetes Technol Ther 10(5):377-383, 2008.
Hovorka R. Continuous glucose monitoring and closed-loop systems. Diabet Med 23(1):1-12, 2006.
Hovorka R, Allen JM, Elleri D, Chassin LJ, Harris J, Xing DY, Kollman C, Hovorka T, Larsen AMF, Nodale M, De Palma A, Wilinska ME, Acerini CL, Dunger DB. Manual closed-loop insulin delivery in children and adolescents with type 1 diabetes: a phase 2 randomised crossover trial. Lancet 375(9716):743-751, 2010.
Hovorka R, Canonico V, Chassin LJ, Haueter U, Massi-Benedetti M, Orsini Federici M, Pieber TR, Schaller HC, Schaupp L, Vering T, Wilinska ME. Nonlinear model predictive control of glucose concentration in subjects with type 1 diabetes. Physiol Meas 25(4):905-920, 2004.
Hovorka R, Kumareswaran K, Harris J, Allen JM, Elleri D, Xing D, Kollman C, Nodale M, Murphy HR, Dunger DB, Amiel SA, Heller SR, Wilinska ME, Evans ML. Overnight closed loop insulin delivery (artificial pancreas) in adults with type 1 diabetes: crossover randomised controlled studies. BMJ 342:d1855, 2011.
Kadish AH. Automation control of blood sugar. I. Servomechanism for glucose monitoring and control. Am J Med Electron 3:82-86, 1964.
Kamath A, Mahalingam A, Brauker J. Analysis of time lags and other sources of error of the DexCom SEVEN continuous glucose monitor. Diabetes Technol Ther 11(11):689-695, 2009.
Klonoff DC. Continuous glucose monitoring: roadmap for 21st century diabetes therapy. Diabetes Care 28(5):1231-1239, 2005.
Kordonouri O, Pankowska E, Rami B, Kapellen T, Coutant R, Hartmann R, Lange K, Knip M, Danne T. Sensor-augmented pump therapy from the diagnosis of childhood type 1 diabetes: results of the Paediatric Onset Study (ONSET) after 12 months of treatment. Diabetologia 53(12):2487-2495, 2010.
Kovatchev B, Breton M, Dalla Man C, Cobelli C. In silico preclinical trials: a proof of concept in closed-loop control of type 1 diabetes. J Diabetes Sci Technol 3(1):43-55, 2009a.
Kovatchev B, Cobelli C, Renard E, Anderson S, Breton M, Patek S, Clarke W, Bruttomesso D, Maran A, Costa S, Avogaro A, Dalla Man C, Facchinetti A, Magni L, De Nicolao G, Place J, Farret A. Multinational study of subcutaneous model-predictive closed-loop control in type 1 diabetes mellitus: summary of the results. J Diabetes Sci Technol 4(6):1374-1381, 2010.
Kovatchev BP, Shields D, Breton M. Graphical and numerical evaluation of continuous glucose sensing time lag. Diabetes Technol Ther 11(3):139-143, 2009b.
Kumareswaran K, Elleri D, Allen JM, Harris J, Xing D, Kollman C, Nodale M, Murphy HR, Amiel SA, Heller SR, Wilinska ME, Acerini CL, Evans ML, Dunger DB, Hovorka R. Meta-analysis of overnight closed-loop randomized studies in children and adults with type 1 diabetes: the Cambridge cohort. J Diabetes Sci Technol 5(6):1352-1362, 2011.
Liebl A, Hoogma R, Renard E, Geelhoed-Duijvestijn PH, Klein E, Diglas J, Kessler L, Melki V, Diem P, Brun JM, Schaepelynck-Belicar P, Frei T. A reduction in severe hypoglycaemia in type 1 diabetes in a randomized crossover study of continuous intraperitoneal compared with subcutaneous insulin infusion. Diabetes Obes Metab 11(11):1001-1008, 2009.
Lindpointner S, Korsatko S, Kohler G, Kohler H, Schaller R, Schaupp L, Ellmerer M, Pieber TR, Regittnig W. Glucose levels at the site of subcutaneous insulin administration and their relationship to plasma levels. Diabetes Care 33(4):833-838, 2010.
Luijf YM, Van Bon AC, Hoekstra JB, DeVries JH. Premeal injection of rapid-acting insulin reduces postprandial glycemic excursions in type 1 diabetes. Diabetes Care 33(10):2152-2155, 2010.
Marliss EB, Vranic M. Intense exercise has unique effects on both insulin release and its roles in glucoregulation: implications for diabetes. Diabetes 51(Suppl 1):S271-S283, 2002.
McMahon SK, Ferreira LD, Ratnam N, Davey RJ, Youngs LM, Davis EA, Fournier PA, Jones TW. Glucose requirements to maintain euglycemia after moderate-intensity afternoon exercise in adolescents with type 1 diabetes are increased in a biphasic manner. J Clin Endocrinol Metab 92(3):963-968, 2007.
Murphy HR, Elleri D, Allen JM, Harris J, Simmons D, Rayman G, Temple R, Dunger DB, Haidar A, Nodale M, Wilinska Me, Hovorka R. Closed-loop insulin delivery during pregnancy complicated by type 1 diabetes. Diabetes Care 34(2):406-411, 2011a.
Murphy HR, Kumareswaran K, Elleri D, Allen JM, Caldwell K, Biagioni M, Simmons D, Dunger DB, Nodale M, Wilinska ME, Amiel SA, Hovorka R. Safety and efficacy of 24-H closed-loop insulin delivery in well-controlled pregnant women with type 1 diabetes: A randomized crossover case series. Diabetes Care 34(12):2527-2529, 2011b.
Nathan D, Genuth S, Lachin J, Cleary P, Crofford O, Davis M, Rand L, Siebert C. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. The Diabetes Control and Complications Trial Research Group. N Engl J Med 329(14):977-986, 1993.
Nathan DM, Cleary PA, Backlund JY, Genuth SM, Lachin JM, Orchard TJ, Raskin P, Zinman B. Epidemiology of severe hypoglycemia in the diabetes control and complications trial. The DCCT Research Group. Am J Med 90(4):450-459, 1991.
Olinder AJ, Kernell A, Smide B. Missed bolus doses: devastating for metabolic control in CSII-treated adolescents with type 1 diabetes. Pediatr Diabetes 10(2):142-148, 2009.
Pedersen JS. The nature of amyloid-like glucagon fibrils. J Diabetes Sci Technol 4(6):1357-1367, 2010.
Pettis RJ, Hirsch L, Kapitza C, Nosek L, Hovelmann U, Kurth HJ, Sutter DE, Harvey NG, Heinemann L. Microneedle-based intradermal versus subcutaneous administration of regular human insulin or insulin lispro: pharmacokinetics and postprandial glycemic excursions in patients with type 1 diabetes. Diabetes Technol Ther 13(4):443-450, 2011.
Pfeiffer EF, Thum C, Clemens AH. Artificial beta cell — continuous control of blood-sugar by external regulation of insulin infusion. Horm Metab Res 6(5):339-342, 1974.
Pickup J, Keen H. Continuous subcutaneous insulin infusion at 25 years — evidence base for the expanding use of insulin pump therapy in type 1 diabetes. Diabetes Care 25(3):593-598, 2002.
Pickup JC, Freeman SC, Sutton AJ. Glycaemic control in type 1 diabetes during real time continuous glucose monitoring compared with self monitoring of blood glucose: meta-analysis of randomised controlled trials using individual patient data. BMJ 343:d3805, 2011.
Rave K, Potocka E, Heinemann L, Heise T, Boss AH, Marino M, Costello D, Chen R. Pharmacokinetics and linear exposure of AFRESA compared with the subcutaneous injection of regular human insulin. Diabetes Obes Metab 11(7):715-720, 2009.
Raz I, Weiss R, Yegorchikov Y, Bitton G, Nagar R, Pesach B. Effect of a local heating device on insulin and glucose pharmacokinetic profiles in an open-label, randomized, two-period, one-way crossover study in patients with type 1 diabetes using continuous subcutaneous insulin infusion. Clin Ther 31(5):980-987, 2009.
Renard E, Costalat G, Chevassus H, Bringer J. Closed loop insulin delivery using implanted insulin pumps and sensors in type 1 diabetic patients. Diabetes Res Clin Pract 74:S173-S177, 2006.
Renard E, Place J, Cantwell M, Chevassus H, Palerm CC. Closed-loop insulin delivery using a subcutaneous glucose sensor and intraperitoneal insulin delivery feasibility study testing a new model for the artificial pancreas. Diabetes Care 33(1):121-127, 2010.
Rodriguez LT, Friedman KA, Coffman SS, Heller A. Effect of the sensor site-insulin injection site distance on the dynamics of local glycemia in the minipig model. Diabetes Technol Ther 13(4):489-493, 2011.
Ryan EA, Paty BW, Senior PA, Bigam D, Alfadhli E, Kneteman NM, Lakey JRT, Shapir AMJ. Five-year follow-up after clinical islet transplantation. Diabetes 54(7):2060-2069, 2005.
Sheetz MJ, King GL. Molecular understanding of hyperglycemia’s adverse effects for diabetic complications. JAMA 288(20):2579-2588, 2002.
Steil GM, Palerm CC, Kurtz N, Voskanyan G, Roy A, Paz S, Kandeel FR. The effect of insulin feedback on closed loop glucose control. J Clin Endocrinol Metab 96(5):1402-1408, 2011.
Steil GM, Panteleon AE, Rebrin K. Closed-loop insulin delivery-the path to physiological glucose control. Adv Drug Deliv Rev 56(2):125-144, 2004.
Steil GM, Rebrin K, Darwin C, Hariri F, Saad MF. Feasibility of automating insulin delivery for the treatment of type 1 diabetes. Diabetes 55(12):3344-3350, 2006.
Steiner S, Hompesch M, Pohl R, Simms P, Flacke F, Mohr T, Pfutzner A, Heinemann L. A novel insulin formulation with a more rapid onset of action. Diabetologia 51(9):1602-1606, 2008.
Tamborlane WV, Beck RW, Bode BW, Buckingham B, Chase HP, Clemons R, Fiallo-Scharer R, Fox LA, Gilliam LK, Hirsch IB, Huang ES, Kollman C, Kowalski AJ, Laffel L, Lawrence JM, Lee J, Mauras N, O’grady M, Ruedy KJ, Tansey M, et al. Continuous glucose monitoring and intensive treatment of type 1 diabetes. New Engl J Med 359(14):1464-U1465, 2008.
Van Bon AC, Kohinor MJ, Hoekstra JB, Von Basum G, DeVries JH. Patients’ perception and future acceptance of an artificial pancreas. J Diabetes Sci Technol 4(3):596-602, 2010.
Vaughn DE, Yocum RC, Muchmore DB, Sugarman BJ, Vick AM, Bilinsky IP, Frost GI. Accelerated pharmacokinetics and glucodynamics of prandial insulins injected with recombinant human hyaluronidase. Diabetes Technol Ther 11(6):345-352, 2009.
Weber C, Schnell O. The assessment of glycemic variability and its impact on diabetes-related complications: an overview. Diabetes Technol Ther 11(10):623-633, 2009.
Weinzimer SA, Steil GM, Swan KL, Dziura J, Kurtz N, Tamborlane WV. Fully automated closed-loop insulin delivery versus semiautomated hybrid control in pediatric patients with type 1 diabetes using an artificial pancreas. Diabetes Care 31(5):934-939, 2008.
Weyer C, Gottlieb A, Kim DD, Lutz K, Schwartz S, Gutierrez M, Wang Y, Ruggles JA, Kolterman OG, Maggs DG. Pramlintide reduces postprandial glucose excursions when added to regular insulin or insulin lispro in subjects with type 1 diabetes: a dose-timing study. Diabetes Care 26(11):3074-3079, 2003.
Wilinska ME, Chassin LJ, Acerini CL, Allen JM, Dunger DB, Hovorka R. Simulation environment to evaluate closed-loop insulin delivery systems in type 1 diabetes. J Diabetes Sci Technol 4(1):132-144, 2010.
Zuckerman DM, Brown P, Nissen SE. Medical device recalls and the FDA approval process. Arch Intern Med 171(11):1006-1011, 2011.
[Discovery Medicine; ISSN: 1539-6509; Discov Med 13(69):159-170, February 2012. Copyright © Discovery Medicine. All rights reserved.]