Search

Heavy ion irradiation testing was performed on AlGaN/GaN High Electron Mobility Transistors (HEMTs), both under DC and RF operation. Single gate finger radiation test structures and multiple gate finger RF power cells were tested and failure mechanisms analyzed. Chip level peak voltage measurements on L-band power AlGaN/GaN HEMTs are compared to Single Event Burn-out voltage levels extracted from DC Single Event Effect (SEE) test results.

118 male and 68 female type I diabetics and 25 male and 23 female nondiabetic healthy controls were compared during submaximal bicycle ergometer tests using four work stages (duration: 6 min each) of 50 W, 75 W, 100 W and a submaximal stage producing a heart rate of 170 min-1. We found that male type I diabetics with and without retinopathy who had significantly higher than normal heart rates both at the start and during the ergometer test also had above normal ergometric blood pressures. Female type I diabetics, in contrast, did not differ significantly from the controls in terms of heart rate before and during the ergometer tests. The PWC170 of all male type I diabetics was, in contrast to that of the female patients, lower than that of the controls. In diabetics with diabetic nephropathy there was no correlation between exercised acceleration of the heart rate and the degree of nephropathy despite their higher heart rates at the beginning and during the ergometer test, but a correlation was found between the ergometrically increased blood pressure and the severity of the diabetic nephropathy. In male type I diabetics cardiocirculatory adaptation to muscular work was reduced, and this reduction became more marked as the degree of diabetic microangiopathy increased. Compared with the controls, the increase in systolic blood pressure evoked by activity on the ergometer in male and female type I diabetics was disproportionate to the actual load and correlated with the degree of diabetic microangiopathy.

Based upon the National Diabetes Registry the mortality rates were assessed annually between 1961 and 1987 in the total diabetic population of the GDR. The rise of diabetes prevalence from 724/10(5) up to 3988/10(5) during the 27-year observation period was associated with an increase of relative mortality rates from 466% to 600% in insulin-treated diabetics, from 352% up to 528% in non-insulin-treated diabetics. By calculation of standardized mortality ratios (SMR) it could be shown that excess mortality is dependent on age but not at all on sex. Insulin-treated diabetics exhibited their maximum SMR of 650% to 750% at ages 25 to 45 years, while in non-insulin-treated diabetics the maximum SMR amounted to 450% at ages 25 to 35 years. In contrast to trends of the total relative mortality rates that of the overall age structure adjusted SMR of diabetics was characterized by a declining tendency, which may be a reflection of the improvement in diabetes care in our country, and which underscores the dependence of mortal, ty rates on the methods used for evaluation.

Two types of insulin pens MADI and MD, were connected to subcutaneous catheters. These "catheter-pens" were used like hand-driven insulin pumps. Results after 1 year of treatment in 30 type 1 diabetics (HCP-negative; age at onset of diabetes 16.5 +/- 1.7 years; duration of diabetes 18.5 +/- 1.6 years, on multiple insulin injections before catheter-pen application): 1. better quality of life (reduction of frequency of needle pricks, more flexibility, inconspicuous application of insulin in public); 2. daily insulin--increased number of "injections" (4.2 +/- 0.1 vs 5.8 +/- 0.1, p less than 0.01), reduction of units per kg BW (0.70 +/- 0.02 vs 0.60 +/- 0.01, p less than 0.01), reduction of intermediate-acting insulin (14.1 +/- 1.3 vs 9.2 +/- 1.2 U/d, p less than 0.05); 3. no change of HbA1 (10.8 +/- 0.8 vs 10.2 +/- 0.2%, normal range 7.7 to 8.4%), mean blood glucose (MBG) in stress situation (8.4 +/- 0.4 vs 7.7 +/- 0.3 mmol/l), serum cholesterol and body weight, both within normal range; 4. improvement (p less than 0.05) of serum triglycerides, serum HDL-cholesterol, ratio of apolipoprotein A1/B; 5. rare skin reactions at the needle site. Conclusion. Catheter-pens offer a very convenient alternative for insulin administration in intensified conventional insulin treatment with multiple injections in type 1 diabetics.

In order to evaluate temporal trend variations of incidence and prevalence rates of Type 1 (insulin-treated) diabetes mellitus in the East-German population (19.6 million) 85,904 incidence cases, recorded by the Diabetes Registry between 1960 and 1989, were analyzed by calculating the regression coefficients for defined time intervals. The total incidence trend was characterized by a rise from 18/10(5) in 1960 to 29/10(5) in 1965, a decrease to 14/10(5) in 1975, and a slight increase to 15/10(5) in 1989. This trend was reflected only by patients with insulin treated diabetes mellitus agedor = 30 yrs. Patients of the age groups 0-9 yrs., 10-19 yrs., and 20-29 yrs. exhibited increasing rates of 0.21/10(5)/year (1960-75), 0.22/10(5)/year, and 0.08/10(5)/year (1960-89), respectively. The rise in the prevalence rate from 187/10(5) (1960) to 658/10(5) (1989) demonstrated temporal trend variations during the periods 1960-71 (+19.3/10(5)/year), 1972-77 (-0.8/10(5)/year), and 1978-89 (+25.6/10(5)/year). The age-specific increasing prevalence trend varied between +0.81/10(5)/year (0-9 yrs.) and +48.89/10(5)/year (60-69 yrs.). The differences observed between incidence and prevalence trends were mainly due to changes in relative mortality rates of patients with insulin treated diabetes and the percentage of secondarily insulinized Type 2 diabetic patients. In summary, the incidence of Type 1 diabetes mellitus in East-Germany increased over the past 3 decades of registration by about 94% in children and adolescents, by 38% in adults aged 20-29 yrs., but not in people agedor = 30 yrs.(ABSTRACT TRUNCATED AT 250 WORDS)

At present in Germany, data on the prevalence of diabetes can only be obtained by modelling health insurance data. The National Diabetes Register of the former (East) German Democratic Republic which, between 1960 and 1990, monitored approximately 98% of all diabetic subjects, provides a tool for evaluating epidemiological estimates from other data sources. Therefore, the following data bases were compared for the year 1988: (1) a 5% random-sample (n = 6478) of all subjects insured at a local statutory health insurance company in the city of Dormund; (2) related data from the population-based diabetes register of former East-Berlin and (3) of the former German Democratic Republic. All data were standardized by sex and age according to the 1988 population statistics of the Federal Republic of (West) Germany thus resulting in the apparent diabetes prevalence of the Western part of Germany at that time.total prevalence rates were (1) 4.8%, (2) 4.9%, and (3) 4.4% (p0.05). The percentages of insulin-treated patients were (1) 18%, (2) 19%, and (3) 16%, respectively. 54% (1), 37% (p0.05) (2), and 42% (p0.05) (3), of the patients received oral antidiabetic drugs. It is concluded that the three samples are comparable and that the diabetes prevalence rates as estimated from health insurance data and from the two population-based registers give corresponding conclusions. Sample-based health insurance data may provide a useful and reliable tool for epidemiological studies on diabetes mellitus.

The pancreaticoduodenal and portal venous blood flows were recorded electromagnetically in anaesthetized dogs. Blood glucose and IRI were measured in the arterial, portal, and peripheral venous as well as in the intestinal venous blood. By a mathematical model the actual net IRI output of the whole pancreas was estimated. Under basal conditions it is 10.2 +/- 2.4 mU/min (n = 30; 26 kg mean body wt.). After i.v. glucose injection, IRI output is rapidly enhanced. The biphasic nature of this reaction was unequivocally demonstrated by consideration of the ratio IRI output : blood glucose. Pancreaticoduodenal blood flow increases transiently in relation to the increased blood glucose concentration. The IRI secretion rate is well correlated with the blood glucose concentration and to the amounts of glucose or of blood reaching the whole pancreas. It is also correlated with the portal IRI Concentration. The overall peripheral venous or arterial IRI concentrations are correlated with the IRI secretion rate, but not in all individual experiments. The different phases of IRI output (basal rate, stimulated output 1-10 min and 10-60 min) show no influence on each other, nor are they correlated with the peripheral IRI concentration area. Basal IRI output is negatively correlated with the glucose assimilation constants. These constants or the peripheral BG areas, however, are independent of the stimulated IRI output rate. However, both the assimilation constants and the peripheral BG areas are related to the peripheral IRI concentration areas. Hepatic uptake of insulin and dynamics of pancreatic blood flow seem to contribute considerably to the estimated correlation pattern.

Bei 73 insulinbedurftigen Diabetikern mit klinisch differentem Stoffwechselverhalten und 7 Gesunden wurden eine 30-minutige Nicotinsaureinfusion durchgefuhrt und die Blutparameter von Glucose, Pyruvat, Lactat, FFS, Glycerin, Katecholaminen und 11-OH-CS uber einen Zeitraum von 180 min analysiert. Auserdem wurde die Insulinsekretion in vivo bei einigen diabetischen und gesunden Probanden mittels der ILA, in vitro an isolierten Langerhans'schen Inseln aus Mausepankreata mittels dem IRI unter Nikotinsaureeinwirkung untersucht. Es konnte demonstriert werden, das Nikotinsaure in vitro zu einer significanten Steigerung der Insulin-ausschuttung fuhrt. In vivo ist eine Stimulation der ILA nur bei Gesunden und stoffwechselstabilen Diabetikern nachweisbar. Das Ausmas des sekundaren Nicotinsaure-induzierten FFS-Rebound steht in negativer Relation zur Stimulierbarkeit der ILA. Es gestattet unter Einbeziehung des Blutglucoseverhaltens und des exogenen, zur Stoffwechselkompensation erforderlichen taglichen Insulinbedarfs eine metabolische Differenzierung des stoffwechsellabilen und stoffwechselstabilen Diabetestyps.

Bei 1000 stationaren Patienten unseres Instituts wurden Einstellungsversuche auf D 860 vorgenommen. Wir untersuchten die Beeinflussung der Erfolgsquote durch folgende Umstande: Konstitutionstyp (vereinfacht erfast als Korpergewicht), Manifestationsalter und Dauer der Zuckerkrankheit, Dosis und Dauer der vorangehenden Insulintherapie, Lebensalter und Geschlecht der Patienten. Gleichzeitig wurde den Wechselwirkungen zwischen diesen Faktoren nachgegangen. Hierbei stellten sich als selbstandige Faktoren der Konstitutionstyp, die erforderliche Insulindosis, das Manifestationsalter des Diabetes und (sehr schwach) die Dauer der Insulinanwendung heraus. Wir deuten diese Gestzmasigkeiten als bedingt durch die Grose des Insulindefizits, einschlieslich der Proportion zwischen dessen absolutem und relativem Anteil bzw. durch das Ausmas der Zuckerhaushaltsstorung, einschlieslich der Proportion zwischen Minderverwertung und Uberproduktion von Dextrose.

Long-term hyperglycaemia in diabetics favours the development of diabetic microangiopathy. The clinical research, therefore, aims to find new opportunities to obtain normoglycaemia in insulin-dependent diabetics mainly by following three patts: -transplantation of pancreatic tissue or isolated islets of Langerhans - implantation of so-called hybrid pancreas and -construction of an artificial B-cell. This paper gives a review of the development, principle of function and clinical application of the glucose controlled insulin infusion system (closed loop system) and of the so-called open loop system without feedback to the glucose values. Results of our own investigations with GCIIS "Biostator", the development of an algorithm of our for the correlation between glucose and insulin behaviour and the first results with the insulin therapy with micropumps (open loop system) were demonstrated and discussed critically.

Within 3 days the 2-hour glucose infusion test (GIT) and the 50gm oral glucose tolerance test (oGTT) were performed in 113 normal weight and 33 obese persons suspected to protodiabetes and in 14 control subjects respectively. The results are compared with criteria from a group of healthy persons without any heredity of diabetes worked out in our laboratory. In about 66 per cent of the investigated subjects a concordance between both tests could observed in carbohydrate tolerance. Abnormal results were found more frequently after the oral glucose application. From these finding it was concluded a higher sensitivity of the oGTT. On the other hand followup studies of the disagreed diagnosis have shown that in 91 per cent the test results of the GIT were reproduced. The insulin secretion pattern agreed in 70 per cent between both tests. Whereas the insulin secretion pattern during the oGTT does not allowed to differ between the groups using the glucose infusion test we were able to observe a significant diminished hormone release in the initial as well as in the late phases, if the carbohydrate tolerance was pathologically. Summarizing the results we concluded that the GIT is characterized by a good reproducibility and a higher diagnostic importance than the 50 gm oGTT.

Twenty unselected unstable type I diabetic inpatients whose blood glucose control was insufficient employing three daily s.c. injections of regular insulin supplemented by intermediate acting insulin were subjected to a 48-hour treatment with the Biostator -GCIIS when both diet and muscular exercise were kept as close as possible to the conditions at home. The s.c. injection regimen was adjusted to the insulin dose pattern required by the artificial beta cell. There was significant metabolic improvement in 16 out of the 20 patients on discharge, in comparison to the pre-Biostator conditions. This improvement was still present when the patients were re-admitted after an average of seven months. It is concluded that in certain cases of unstable type I diabetes mellitus the metabolic re-arrangement based on intercalary days on an extracorporal artificial beta cell might be useful if the control constants are adapted to minimize the insulin requirement by the machine.

The influence of the moment in pregnancy of insulin dependent diabetic women, at which normoglycemia by insulin therapy could be reached, on cord blood insulin concentration and neonatal morbidity was investigated. There is a positive correlation between the moment of normoglycemia (HbA1-values less than or equal to 8.5%) and the insulin concentration and furthermore the incidence of respiratory distress syndrome and macrosomia in the newborn. It is concluded that a tight metabolic control in insulin dependent diabetic women already prior to or early in pregnancy (at least until the 16th week) will be able to decrease the incidence of morbidity in infants of diabetic mothers.

The clinical importance of different constants used in the algorithm of a glucose-controlled insulin infusion system (Biostator) with respect to insulin requirement was studied in 9 insulin-dependent juvenile-onset type diabetics. After normalizing the fasting blood glucose level (around 4.5 mmol/l), 3 consecutive glucose infusions (12 mg . kg-1 . min-1) followed by a 1 h post-infusion period were carried out in all individuals. The blood glucose was controlled by the Biostator using three sets of constants: A (KR = 160, KF = 50, QI = 30, RI = 12), B (KR = 70, KF = 50, QI = 40, RI = 12) and C (KR = 20, KF = 50, QI = 50, RI = 20). The algorithm constants were changed in an arbitrary order before the start of each glucose infusion in the same patient. Mean blood glucose profiles were comparable whereas the mean total insulin dose infused by the Biostator amounted to 13 +/- 1.3 U/h, 7.8 +/- 1.2 U/h and 7.0 +/- 1.0 U/h when sets A, B and C were applied (A versus B and C p less than 0.01). The higher insulin amount during control of set A was mainly due to a high insulin infusion during the first 20 min. In summary, the results emphasize the clinical importance of appropriate algorithm constants of a glucose-controlled insulin infusion system for evaluation of the insulin requirement in insulin-dependent diabetics.

Simultaneous monitoring of pancreatic and hepatic blood flows and serial analyses of blood glucose and immunoreactive insulin activity in arterial, portal, intestinal and hepatic venous blood were performed in anaesthetized laparotomized dogs. Based on these measurements, the rates of pancreatic insulin secretion, of hepatic insulin uptake and the hepatic glucose balance were computed. Under basal conditions pancreatic output was about 40% of the total hepatic insulin input. 30% of hepatic insulin inflow were actually taken up in the liver. During a primed i.v. glucose infusion (steady blood glucose concentration about 330 mg/100 ml) pancreatic and hepatic blood flows increase rapidly and the insulin secretion exhibits a biphasic pattern. The relative part of hepatic insulin inflow taken up in the liver remained unchanged. A non-primed i.v. glucose infusion provoked a slow increase of blood flow and a rapid but monophasic increase of insulin release. Under these conditions the relative hepatic insulin uptake was transiently increased. During both loads hepatic glucose balance switched from release to uptake. However the non-primed infusion was more effective. When pancreatic and hepatic blood flows were increased without any blood glucose alteration by an infusion of mannitol, insulin secretion did not change; hepatic uptake of insulin decreased, but hepatic glucose output was clearly reduced. A correlation between the rates of insulin secretion, of hepatic insulin inflow and of insulin uptake was observed under all conditions. However, the hepatic glucose balance was correlated to the inflow of insulin or glucose and to the hepatic insulin uptake under certain conditions in individual experiments only.

Trends of prevalence and incidence rates of non-insulin-dependent diabetes mellitus (non-insulin-treated diabetes mellitus) were assessed in the population of the GDR based upon the National Diabetes Register and the Official Statistical Year Book as sources for the calculations. Within the 25-year follow-up period 1960-1984 the prevalence rose from 4.39%; to 31.95%; the incidence rate from 1.04%; to 3.57%. Age-dependence of the specific rates is characterized by their continuous rise above the age of 30 years reaching the peak prevalence of 146.6%; in 75- to 80-year-olds, that of 14.1%, for the incidence in people aged 70 to 75 years. A significant male preponderance was confirmed between the ages of 30 and 50 years, a significant overwhelming of female NIDDM in the age groups 60 to 90 years. Based on demonstrated correlations between the changes of living standard parameters and the epidemiological trend of NIDDM the conclusion is drawn that overnutrition and reduced muscular activity mainly account for the rise of diabetes morbidity in the population of the GDR.

The control constants for glucose-dependent insulin dosage in diabetic dogs were determined from test results in the opened system on the basis of a global blood glucose plasma-insulin control model. The controlled plant of the model consisted of the glucose and insulin subsystems; the entire insulin providing process was considered to be the controlling element, and the glucose-dependent insulin dose estimation, the controller. The constants obtained were employed in the extracorporeal artificial beta cell. The structure of the model and the numerical values of its state variables were verified by the prediction of blood glucose responses to intravenous glucose loads and by the correspondence between glucose balances and insulin doses as calculates and those as observed in diabetic animals. The application of the optimum control constant estimates in feedback-controlled insulin infusions provides improved blood glucose patterns but unchanged needs for insulin in comparison to the application of standard control parameters.

The appearance rate of insulin (calculated insulin secretion rate) in the circulating blood after subcutaneous injection was estimated in diabetic dogs from serial measurements of immunoreactive insulin concentrations using a simple mathematical model based on the insulin half-life and the distribution space. In the case of highly purified monocomponent porcine insulin, maximum concentrations occurred after 30-60 min. The duration of insulin appearance was dose-dependent and the rate of appearance could be described by a bi-exponential function. It was linearly dose-dependent but the effect on glycaemia showed saturation kinetics. The action of the injected dose on the fasting glycaemia diminished when the appearance rate became less than 0.3 mU X kg-1 X min-1. Fractional dose recovery was between 70% and 90% and was not different between depot and regular insulin. Appearance kinetics were not significantly affected by the initial glycaemia. The model presented provides a means for quantitative characterization of different insulin preparations.

For optimum long-term glycemic regulation using a miniaturized artificial beta cell it is indispensible to estimate control parameters suited to the individual requirements of each diabetic patient. To solve this problem, a strategy has been developed which is based on engineering optimum-control theory with a model involving glucose and insulin interactions. The model considers physiologically relevant unit processes like endogenous glucose production, insulin-independent glucose uptake from its apparent distribution space, insulin-dependent glucose utilization, glucose-dependent insulin supply, and insulin catabolism. The assumed model structure is validated by results obtained in experimentally diabetic dogs using partition analysis. The individual parameter values of the model are obtained by a digital computer procedure based on a simple test which involves a bolus injection of glucose + insulin when a constant basal insulin dose is being administered in the diabetic in whom normoglycemia was re-established before the test. The method presented is recommended for future use in all cases where an optimized insulin regimen is to be worked out.

The influence of fetal hyperinsulinemia on the development of diabetic fetopathy was investigated. Human fetal pancreatic slices of fetus between the 12th and 26th weeks of pregnancy from non diabetic (n = 32) and diabetic (n = 18) women were incubated in vitro for one hour. The insulin secretion results clearly demonstrate an age dependent increase in both groups investigated, but the increase is much more pronounced in fetuses from diabetic women. The differences between the groups are present already during the 12th week of pregnancy. The results support the concept to normalize the metabolic control in pregnancy as early as possible in order to prevent the diabetic fetopathy.

Insulin secretion rates after glucose loading were calculated from peripheral venous IRI concentrations considering half life and distribution space of exogenous insulin in normal men and dogs. The coefficients of multiple linear regression analysis between insulin secretion rates and plasma glucose (level and order and rate of change) were used as algorithm parameters in glucose-controlled insulin infusions. These were carried out in each dog based on individual estimations before the induction of diabetes but in the diabetic patients based on values derived from a group of normal subjects. Using this formula, nearly normal patterns of glucose and of insulin were observed in diabetic men and dogs under basal conditions and after IV glucose loading but not after meals. This algorithm enables selection of the parameters prospectively. The effect of a parameter combination depends on insulin sensitivity and it should be appropriately adapted. In the diabetic patients there was no predictable influence of the brittle or stable characteristics of the disease nor of insulin antibodies on the glucose curves obtained with glucose controlled insulin infusions.

Twenty-one patients in severe diabetic coma were treated with small doses of insulin at a rate of 4.1 units per hour (total dose about 100 units per 24 hours). Using single doses of 4 to 10 units by the intravenous or intramuscular routes the fall of blood glucose was steady in all cases. In the treatment of diabetic coma this regimen of insulin administration has proved simple, safe and effective since 1946. Main dangers during recompensation of diabetic coma are: hypovolaemia with oliguria -- anuria, dysequilibrium syndrome with cerebral edema and hypokalaemia. Therefore early intensive and adequate intravenous fluid and electrolyte replacement is the most important part of treatment. Most of the cases in this study were undiagnosed diabetics (14) and elderly patients (9). Three patients older than 65 years and a 56-year old diabetic died. In this context the most important aspects of treatment to avoid death are: prevention of diabetic coma and adequate fluid and electrolyte replacement especially in geriatric patients.

Normal dogs were submitted to oral glucose loads or to intravenous glucose infusions. Insulin secretion rates (CISR) were calculated considering the resulting peripheral venous concentration differences in short intervals and the experimentally determined half life and apparent distribution space of exogenous insulin. Multiple regression analysis was done between CISR and both the level and the rate of change of plasma glucose. The regression coefficients were used as algorithm parameters for continuous plasma glucose dependent intravenous insulin administration in the same animals after induction of an insulin-dependent diabetes. Normal glycemic regulation over the day could be resotred by this sytem. The insulin responsiveness, however, varies from day to day; tusing this insulin dosage pattern we observed nearly normal plasma glucose curves and slightly elevated insulin reactions after glucose loading. This kind of algorithm could also be used in diabetic humans.

The blood glucose responses of diabetic dogs on the ABC are nearly identical if different algorithms, but equivalent sets of parameters are used to control the glucose-dependent insulin infusion. There are however some differences in the insulin doses applied, which obviously result from different response kinetics of the particular formula to the minor BG oscillations that always occur. In the practical application of closed-loop systems the control constants should be individually estimated on the basis of certain controlled plant characteristics. For use in patients the checking and modeling systems still have to be simplified. The limits of the study, i.e. the fact that normoglycemia was restored for a short time only using the IV instead of the physiological portal route, remain the subject of further investigation.

The thickness of the epithelial and endothelial basal laminae was determined in placentas from mothers without known metabolic disease (control) and mothers with diabetes mellitus. The measurements were carried out on electron micrographs. Frequency distributions and statistical parameters were calculated; in 20 cases, 16,331 values were measured. In all cases, the epithelial and endothelial basal laminae were thickest in edematous and fibrotic villi. They were thinner in placentas from diabetic mothers than in controls. Particularly thin basal laminae were found in regions with syncytiovascular membrances. The thinner basal laminae found in the placentas of diabetic mothers are considered to be the result of compensatory growth processes with formation of ontogenetically younger villous parts. The thicker basal laminae in fibrotic and edematous villi may result from a swelling of the basal laminae or possibly from an increased synthesis or a decreased decomposition of the basal laminae. The finding of very thin basal laminae in the region of the so-called syncytiovascular membranes may be of importance for the placental exchange function.

In 122 diabetic pregnancies the placental blood flow has been estimated determining the half life of the activity inflow (2 MBq 113 m In labelled transferrin) into the placental bed. We used a highly sensitive detector (modified pinhole collimator) and a computer supported evaluation, free from subjective influences. 259 flow measurements were compared to the risk of complication in the course of a diabetic pregnancy. - The half life values in the diabetic group, calculated by a gamma camera computer system by means of an iterative regression analysis, were significantly different compared to a control group (12 pregnancies without risk.) - Severe diabetic angiopathic complications (White classes D, F, and R) are accompanied by higher half life values (placental blood flow reductions) and perinatal complications. - Even in pregnant women with gestational diabetes or disturbances of the carbohydrate metabolism a disturbed placental hemodynamic is to be found.

On the basis of the estimation of the plasma glucose and the insulin concentrations in 10-minute intervals during 24 hours by the mathematical analysis (multiple regression) over relations between the behaviour of plasma glucose and insulin concentration insulin secretion rates independent of plasma glucose can be calculated. In non-diabetics a circadian rhythm of the disposal of insulin - independent of food - is established. Thus the secretion rate increases in the early morning hours still before before the first meal and is maintained in different size up to nearly 18 o'clock. Then it established itself to the level of the nocturnal basal rate. Adequate changes of the insulin need during night are proved in diabetics without endogenic residual insulin secretion by means of the artificial B-cell Biostator. The circadian rhythm of the disposal of insulin take place parallel with adequate changes of other biological parameters and is of importance for the practice of the conventional insulin therapy as well as in the future use of non-measuring value-regulated insulin infusions (open-loop system).