artificial pancreas
As artificial pancreas or artificial beta cell a medical device is called, provides the patient with diabetes mellitus as a function of continuous measurements of blood glucose levels with insulin. So that it simulates the operation of occurring in the islets of Langerhans of the pancreas beta cells, the insulin production and release realize in the body and are destroyed in diabetics or limited in their function. The function of an artificial pancreas which equals the natural insulin release more than the conventional insulin therapy, or treatment by means of an insulin pump.
The essential components of an artificial pancreas are a continuously measuring glucose sensor for determination of blood glucose, a pump for the controlled release of insulin as well as a miniature computer that analyzes the sensor data from the sensor and controlled by an algorithm for the simulation of glucose-insulin-loop pump. The artificial pancreas is located since about 1970 in the development, from the late 1970s such systems were first tested in humans. The goal is the long-term use as an implant or as a portable device from the patient. At present, the artificial pancreas is still the subject of research and, apart from stationary devices for short-term use, not yet commercially available.
History
For the first time, a control loop was described in 1964 by Kadish. From 1974, several research groups tried to meet the goal of world-regulated insulin infusion in more detail, including Albisser (USA), Kreagen (Australia), Mirouze (France) and Shishiri (Japan). In Germany, by E. F. Pfeiffer († 23 January 1997) worked in Ulm on the development. In Karl castle in the GDR was researched by U.Fischer and colleagues simultaneously on the same system with a mathematical model and a glucose sensor. In Japan brought at the same time M. Shishiri a subcutaneously measuring device out. But both were not in the clinical development of miniaturized devices. Great devices with non-implanted glucose sensor and counter-infusion were at this time brought into commercial production (Biostator and Nikkiso STG-22 Blood Glucose Controller). From 1978, operated the first subcutaneous insulin pumps in USA by Pickup in England and Tamborlane. Only after reliable and implantable continuous glucose sensors could be produced in large numbers came from 2004, primarily in the US, back to numerous development priorities (Table). Due to the recognition of a simulator by the US Food and Drug Administration FDA to substitute animal testing 2008, experiments with the control loop could easily simulate ("in silico") what a new thrust in the development provoked. The full portability of control could be achieved from 2011 through transfer of the complete software on Android smartphones (DIAS system).
Subgroups types
There are different embodiments which differ in
-Alternatively the compartment for location of the sensor and the catheter insulin, subcutaneously, intravenously or intraperitoneally
-Degree of automation
-Designed as miniaturization, full-implanted, large external device
-Place of application: intensive care unit, inpatient outpatient,
-Target area or exact target value
-Insulin alone or with glucagon or bi-hormonal pramlintide
-Type of control algorithm, heart rate
The most important are discussed below in detail
Threshold interrupt system
The goal of a threshold-interrupt system for insulin is to reduce the severity or the reversal of a dangerous Blutgluckoseabfalls (hypoglycemia) by a temporary interruption of insulin delivery by means of pump when the glucose level reaches a threshold or this approach. A further illustrate systems which achieve the threshold predict (prediction, Eng. Predictive low-glucose suspend PLGS).
Range control systems
A range control system reduces the likelihood of hyper- or hypoglycaemic episodes, the fact that it changed the dose of insulin when blood glucose approaches a high or low threshold. People who choose this type of system, must continue to inject insulin themselves perform a blood glucose self-monitoring and adjust the insulin dose these values.
Setpoint control systems
Setpoint control systems try this nominal or target value to be reached at any time. The system is fully automatic and requires no calibration except involvement of the user.
Hybrid system with feedforward control
A hybrid system allows the patient an additional dose of insulin before meals to administer. This additional dose reduces the risk of hyperglycemia after a meal. The usual in control theory method of pilot control (engl. Foreward feed) for compensation of measurable disturbances is implemented partially here. Since the disorder can not be precisely measured, but only estimated by the user (or carbohydrate KE estimation) is usually a partial pre-control operated, in which, for example, 50% of the calculated dose of insulin to be injected as a bolus in advance.
We are thus on the principle of fully automatic control for better compensation of meals and calls for a the help of the user.
Controller types
Model predictive controller
Model predictive controllers are used in industry in refineries, waste incineration, etc., when integrated controller (PID) does not have the necessary quality and sufficient time to optimize at each sampling control. You can calculate a prediction of the control parameters based after each scan.
Proportional-integral-differential controller (PID)
This type of controller consists of three components: (a) the proportional share, in the manipulated variable (infusion rate) is proportional to the error variable (glucose value - glucose setpoint). (b) An integrator element, which ensures the steady-state accuracy, but slows the response. (c) a differentiating circuit which is responsive to changes, the controller makes fast, but increasingly fast disorders and can make the controller unstable. As of 2010, have such control - get a feedback from the (predicted) levels of insulin to prevent hypoglycemia by a Überinsulinierung - based on the physiological regulation of a healthy pancreas.
Fuzzy controller
Fuzzy control consists of a controller, which a number of discrete input values (eg glucose three areas:. High, normal, low) by fuzzy rules ('wenn'-' then '- rules) an output value (insulin infusion rate ) result. The fuzzy control is inspired by the common practice of bolus calculation in everyday life: The patient corrections of 140 to 170 mg / dL with an insulin unit (IU) and between 170 and 200 mg / dl with 2 iE etc.
Risk reduction and security architecture
An incorrect measurement by the sensor and / or an incorrect control by the control algorithm can in principle lead to a life-threatening hypoglycemia. A modular structure in which security modules can issue a warning or shutdown insulin independent of the control algorithm is required. Thus, for example, technical limitations provided by a separate hypoglycemia prediction algorithm (engl. low glucose dectection module) or by a and insulin-on-board calculation of the rate of infusion.
Peculiarities and difficulties
A special feature of the feedback are long delays in the process: a physiological and thus uncontrollable delay action is the action of insulin in the liver of about 100 min and in peripheral tissues (muscle) of about 20 min. In addition, in the present application's preferred SC-SC in addition at both ends of potentially variable delays: the grounded in the glucose sensor and the lag time of about 5-15 Gewebsdiffusion min, and the resulting delay in the absorption of insulin from 30-90 min (even at application faster insulin analogues). Classic controller types (PID) can compensate for such delays by increasing the D-link part, this is to the detriment of the amplification of the sensor noise and other disturbances. Although modern types of controller (MPC) can detect such delays better, but here encounter the same limitations. Solutions are currently being intensively searched for other insulin-application forms (intraperitoneally, by inhalation).
Goals and metrics for the performance
-complete or substantial transfer of all activities of the diabetes management of patients on the unit (discharge)
-Stabilization and reduction of mean glycemia with long-term reduction of consequential damages
-Avoidance or reduction of acute complications such as hypoglycemia and ketoacidotic comas,
It is proposed to detect the quality and variability by means of time in the target area, which (CGM) is through the continuous glucose measurement is possible. Hybrid systems (see above) currently achieve that in about 70% of the measurement time, the glucose levels are 70-180 mg / dl in the area. Further, the measurement of quality of life will play an increasingly important role. There are validated instruments to measure this example Timesheets for Diabetes Quality of Life or Fear of hypoglycemia. Ultimately, a compromise struck between the degree of automation and control performance; just as a compromise between achieving euglycemia and the number of undesirable hypoglycemia. The partial transition to the user involves risks that are due to unpredictable human behavior and give rise to safety concerns. ("Non-inferiority" by best current comparator therapy) are compared metrics through national registries or quality initiatives made. In Germany was demonstrated for example by a project initiated in the hospital combined therapy and training program with the methods of quality management, people with diabetes type 1 by intensive insulin therapy (ICT) can achieve an HbA1c of 7.3% on average and an HbA1c-independent number of severe hypoglycemia of 0.14 / Pat. / year.