Associations of blood glucose dynamics with antihyperglycemic treatment and glycemic variability in type 1 and type 2 diabetes.

AIMS: The dynamical structure of glucose fluctuation has largely been disregarded in the contemporary management of diabetes. METHODS: In a retrospective study of patients with diabetes, we evaluated the relationship between glucose dynamics, antihyperglycemic therapy, glucose variability, and glucose exposure, while taking into account potential determinants of the complexity index. We used multiscale entropy (MSE) analysis of continuous glucose monitoring data from 131 subjects with type 1 (n = 18), type 2 diabetes (n = 102), and 11 nondiabetic control subjects. We compared the MSE complexity index derived from the glucose time series among the treatment groups, after adjusting for sex, age, diabetes duration, body mass index, and carbohydrate intake. RESULTS: In type 2 diabetic patients who were on a diet or insulin regimen with/without oral agents, the MSE index was significantly lower than in nondiabetic subjects but was lowest in the type 1 diabetes group (p < 0.001). The decline in the MSE complexity across the treatment groups correlated with increasing glucose variability and glucose exposure. Statistically, significant correlations existed between higher MSE complexity indices and better glycemic control. In multivariate regression analysis, the antidiabetic therapy was the most powerful predictor of the MSE (ß = -0.940 ± 0.242, R 2 = 0.306, p < 0.001), whereas the potential confounders failed to contribute. CONCLUSIONS: The loss of dynamical complexity in glucose homeostasis correlates more closely with therapy modalities and glucose variability than with clinical measures of glycemia. Thus, targeting the glucoregulatory system by adequate therapeutic interventions may protect against progressive worsening of diabetes control.

Model-based decision support in diabetes care.

The model-based Karlsburg Diabetes Management System (KADIS®) has been developed as a patient-focused decision-support tool to provide evidence-based advice for physicians in their daily efforts to optimize metabolic control in diabetes care of their patients on an individualized basis. For this purpose, KADIS® was established in terms of a personalized, interactive in silico simulation procedure, implemented into a problem-related diabetes health care network and evaluated under different conditions by conducting open-label mono- and polycentric trials, and a case-control study, and last but not least, by application in routine diabetes outpatient care. The trial outcomes clearly show that the recommendations provided to the physicians by KADIS® lead to significant improvement of metabolic control. This model-based decision-support system provides an excellent tool to effectively guide physicians in personalized decision-making to achieve optimal metabolic control for their patients.

Relationships between glucose variability and conventional measures of glycemic control in continuously monitored patients with type 2 diabetes.

Given the importance of glucose variability in the development of diabetic complications, the present study used continuous glucose monitoring (CGM) to determine various indices of glucose variability and to investigate their relationships with conventional measures of chronic sustained hyperglycemia. We examined 53 women and 61 men, aged 36-79 years afflicted with type 2 diabetes for 1-24 years. The following indices of glycemic variability were computed from CGM data sets: mean amplitude of glycemic excursions (MAGE), CGM glucose range, interquartile range (IQR), SD-score, and average daily risk range (ADRR). CGM measurements and self-monitored blood glucose (SMBG) records were used to calculate mean CGM sensor glucose and mean SMBG, respectively. In simple correlation analysis, the indices of glucose variability showed weak correlations with HbA1c: MAGE (r=0.27, p <0.01), CGM glucose range (r=0.21, p <0.05), IQR (r=0.31, p <0.01), SD-score (r=0.34, p<0.001), and ADRR (r=0.24, p<0.05). These indices were found to differ at identical HbA1c among several patients, as reflected by diurnal excursions of different frequency and magnitude. With the exception of ADRR, stronger correlations were found between mean SMBG and the other variability indices (r=0.51-0.63, p<0.01 for all). CGM provides various indices of glycemic variability not captured by conventional measures of glycemic control. Detection of the location and the magnitude of glucose fluctuations by CGM should aid in optimal treatment of glycemic disorders in type 2 diabetes.

Impact of cytokine- and FasL-induced apoptosis in the beta-cell line NIT-1.

Cytokine- and FasL-induced pathways contribute to beta-cell death in type 1 diabetes. It remains unclear, however, whether pro-apoptotic cyto-kines or FasL have more apoptotic impact. Cytokine- and FasL-induced apoptosis were simulated using IL-1beta/IFN-gamma, Super-FasLigand and the beta-cell line NIT-1. The role of caspases was addressed using the general caspase inhibitor ZVAD. Exposure to IL-1beta/IFN-gamma induced NIT-1 cell death. FasL augmented cytokine-induced cell death accompanied by increased caspase-3 activation, DNA fragmentation, and chromatin condensation. However, FasL mediated comparable effects on the mitochondrial transmembrane potential (Deltapsi (m)) and nitrite in cytokine- and untreated cells. The cytokine-induced sequence of apoptotic events was (1) Fas, nitrite, (2) Deltapsi (m), (3) DNA fragmentation, cell death, and (4) chromatin condensation. In the presence of FasL, cell death and chromatin condensation appeared earlier implicating a compression of the apoptotic time course. General caspase inhibition using ZVAD prevented cell death, Deltapsi (m), and DNA fragmentation; however, Fas expression and nitrite were increased. In conclusion, cytokines account for the major part of cell death induced by the simultaneously action of FasL + IL-1beta/IFN-gamma. Caspases are of central importance for beta-cell death.

Efficacy of the dipeptidyl peptidase IV inhibitor isoleucine thiazolidide (P32/98) in fatty Zucker rats with incipient and manifest impaired glucose tolerance.

AIM: Incretin enhancers are a new class of antidiabetic drugs with promising therapeutic potential for type 2 diabetes. Therapeutic intervention in prediabetes is an attractive strategy for preventing or delaying diabetes onset. The aim of the present study was to investigate the therapeutic effects of incretin enhancement on incipient impaired glucose tolerance (iIGT) and manifest IGT (mIGT) using the dipeptidyl peptidase IV (DPP-4) inhibitor P32/98- and fatty Zucker rat (ZR, fa/fa) as a model. METHODS: ZRs were classified into groups with iIGT and mIGT (n = 10 per group). P32/98 (21.61 mg/kg body weight) was administered orally to ZR with iIGT and mIGT once daily for 6 and 3 weeks respectively. Assessments included body weight, morning blood glucose and insulin, oral glucose tolerance test (oGTT; 2 g glucose/kg), plasma parameters and blood glucose day-night profile (DNP). In addition, glucose responsiveness of isolated islets and islet morphology were analysed. RESULTS: P32/98 decreased non-fasting morning blood glucose more effectively in ZR with iIGT than in ZR with mIGT. Compared with study entry, P32/98 improved DNP of blood glucose in ZR with mIGT and nearly normalized DNP in ZR with iIGT. An acute bolus of inhibitor reduced glucose load during oGTT in rats chronically treated with placebo or P32/98. In contrast to placebo-treated rats, rats receiving long-term treatment with P32/98 required less insulin during oGTT. This effect was larger in rats with iIGT vs. rats with mIGT. In isolated pancreatic islets of ZR with mIGT, treatment with P32/98 decreased pancreatic insulin content and increased glucose responsiveness, while the beta-cell volume density was unaffected. P32/98 significantly reduced triglycerides and non-esterified fatty acids. Intestinal growth was comparable between inhibitor- and placebo-treated fatty rats. CONCLUSIONS: Enhancement of incretin with the DPP-4 inhibitor P32/98 has therapeutic effects in hyperinsulinaemia, hyperglycaemia and IGT in ZR with iIGT and mIGT. Apparently, administration of P32/98 in ZR with iIGT results in more efficient beta-cell function, which is associated with less need for insulin to cope with deterioration of glucose tolerance. Importantly, P32/98 has a strong effect on dyslipidaemia in mIGT. P32/98 has no side effect on intestinal growth. Daily intake of P32/98 is a promising strategy for treatment of glucose intolerance and has the potential to prevent type 2 diabetes.

Fas ligand down-regulates cytokine-induced Fas receptor expression on insulinoma (NIT-1), but not islet cells, from autoimmune nonobese diabetic mice.

In the pathogenesis of autoimmune type 1 diabetes, the apoptosis receptor Fas appears de novo on the surface of insulin-producing beta-cells. Fas expression is thought to be induced by proinflammatory cytokines, such as IL-1beta, interferon-gamma (IFNgamma), and TNFalpha, released by islet-infiltrating mononuclear cells. To determine whether beta-cells can modulate their sensitivity to apoptosis at the level of Fas, we investigated the effect of Fas ligand (FasL) on surface expression of Fas in NIT-1 insulinoma cells from nonobese diabetic (NOD) mice prone to autoimmune diabetes and islet cells from NOD and nonautoimmune BALB/c mice. In NIT-1 insulinoma cells, Fas expression induced by the cytokine combination IL-1beta and IFNgamma was reduced in the presence of FasL, whereas in islet cells Fas expression was unaffected by FasL. The effect of FasL on NIT-1 cells was evident during and after the induction of Fas expression by IL-1beta and IFNgamma. Thus, FasL down-regulates cytokine-induced Fas expression in NOD mouse-derived NIT-1 cells, but not in NOD or BALB/c mouse islets. The ability of NIT-1 cells to down-regulate Fas receptor in response to ligation is similar to that of a variety of tumor cells, which may use this mechanism to escape destruction by cytotoxic T cells. Islets apparently cannot protect themselves against FasL-induced apoptosis by down-regulating the Fas receptor. Understanding how NIT-1 insulinoma cells down-regulate Fas receptor in response to ligation by FasL has therapeutic implications for protecting normal beta-cells in autoimmune type 1 diabetes.

Surface and intracellular Fas expression associated with cytokine-induced apoptosis in rodent islet and insulinoma cells.

During the process of insulitis in the pathogenesis of type I (insulin-dependent) diabetes mellitus, proinflammatory cytokines induce expression of the death receptor Fas on the surface of pancreatic beta-cells and thereby contribute to the enhanced susceptibility of beta-cells for apoptosis. The aim of this study was to compare cell-surface and intracellular Fas expression associated with cytokine-induced apoptosis in commonly used beta-cell models such as isolated islets and insulinoma lines derived from mouse and rat. The cell line NIT-1 responded to the interleukin (IL)-1beta+interferon (IFN)-gamma stimulus with translocation of Fas to the cell surface. Likewise, islet cells from non-obese diabetic (NOD) mice and BB/OK rats expressed increasing amounts of the Fas receptor on their surfaces after exposure to IL-1beta in combination with IFN-gamma and tumour necrosis factor-alpha. Moreover, islets obtained from BB/OK rats at an age near the onset of diabetes had an increased surface expression of Fas compared with young rats. In contrast, western blot analysis of cell lysates from cytokine-exposed islets and insulinoma cells revealed total Fas expression levels comparable to those of untreated controls. In conclusion, islets from BB/OK rats and NOD mice, in addition to NIT-1 insulinoma cells, responded to cytokine exposure with surface expression of the Fas receptor, whereas in cell lysates the levels of expression of Fas were found to be independent of cytokine exposure. Taken together, the findings indicate that cytokine-treated beta-cells might possess two pools of Fas protein, one of which is inducible by cytokines and accounts for surface Fas expression, whereas the other is constitutively expressed in cytoplasmic compartments. The underlying mechanisms, including possible interactions between these two sources of cellular Fas expression, need to be investigated in future studies.

Cell surface trafficking of Fas in NIT-1 cells and dissection of surface and total Fas expression.

The appearance of Fas receptor at the surface of pancreatic beta-cells affected by progressive insulitis strongly suggests that Fas-mediated beta-cell apoptosis plays an important role in the pathogenesis of type 1 diabetes. In support of this concept, the present study has shown that islet cells from NOD mice and the beta-cell line NIT-1 respond to the proinflammatory cytokines IL-1beta and IFN-gamma with Fas surface expression in a dose- and time-dependent manner. Moreover, the prevention of cytokine-induced surface Fas expression by actinomycin D, cycloheximide, and brefeldin A demonstrated that trafficking of Fas to the beta-cell surface requires RNA and protein synthesis and, in addition is critically dependent on intracellular protein transport. Compared with total cellular Fas protein, the amount of Fas at the cell surface was relatively small and indicated that Fas is preferentially expressed in cytoplasmic compartments of NIT-1 cells. It is concluded that inflammatory insults specifically induce translocation of Fas to the beta-cell surface and that interference with cell surface Fas expression is a new strategy to improve beta-cell survival in inflamed islets.

Immunoadsorption of immunoglobulins alters intracytoplasmic type 1 and type 2 T cell cytokine production in patients with refractory autoimmune diseases.

Intracellular cytokine staining and flow cytometry were used to investigate whether immunoadsorption (IA) of immunoglobulins alters intracytoplasmic cytokine production in CD4+ and CD8+ T cells from the blood of patients with refractory rheumatoid arthritis (n = 7), membrane proliferative glomerulonephritis (n = 1), and Goodpasture's syndrome (n = 1). Four patients (Group 1) showed severely depressed production of TNF-alpha, IL-2, IFN-gamma, and IL-4 by CD4+ and CD8+ T cells and responded to 3 IA sessions with significant increases in CD4+TNF-alpha+, CD4+IL-2+, and CD8+IL-2+ T cells. Also, a tendency toward increased percentage levels of CD4+ T cells producing IFN-gamma or IL-4 and of CD8+ T cells producing either TNF-alpha or IFN-gamma was seen, but due to the small number of patients investigated, these differences did not attain statistic significance. Group 2 (n = 5) showed unimpaired intracellular cytokine levels and responded to IA with a heterogeneous pattern of changes in TNF-alpha, IL-2, IFN-gamma, and IL-4 production, but these alterations were smaller than those in Group 1. The present findings indicate that the extracorporeal removal of immunoglobulins by anti-IgG or protein A adsorber columns has an impact on T cell immunity and suggest that modulating effects on cellular immune system function are involved in the mode of action of IA.

Mixed graphical models for simultaneous model identification and control applied to the glucose-insulin metabolism.

In this paper a method for model identification of biological systems described by stochastic linear differential equations using a new computational technique for statistical Bayesian inference, namely mixed graphical models in the sense of Lauritzen and Wermuth, is presented. The model is identified in terms of biological model parameters and noise parameters. This non-linear estimation problem is solved by means of an exact inference algorithm. The parameter estimates are given as a-posteriori distributions which can be interpreted as fuzzy possibility distributions. For model-based simulations of the underlying biological system the model parameters are represented as uncertain parameters with the distributions obtained from the estimation procedure. We apply the presented methods to a model for the glucose-insulin metabolism: the Karlsburg model for type I diabetes.

KADIS: model-aided education in type I diabetes. Karlsburg Diabetes Management System.

Education and training in self-management of blood glucose control has become a permanent task for all people involved in the care of diabetic patients. Since this may be facilitated by applying state-of-the-art information technology, we have developed the decision support system KADIS (Karlsburg Diabetes Management System). It comprises computer-aided tools for (1) the evaluation (selection, aggregation, storage, statistics, graphics) of therapeutic data, e.g. from patients' logbooks, and (2) the simulation of daily profiles of glycaemia and insulinaemia on the basis of a mathematical model of the glucose-insulin regulatory system, parameters of which can be adapted to the characteristics of individual patients. The latter tool allows the patient to predict his response to any modification in the therapeutic regime and to learn how variations in timing, formulation and doses of insulin, in carbohydrate equivalents and absorption characteristics of meals, and in exercise may influence the daily pattern in glycaemia. This procedure has been well accepted as an educational tool by those patients who were 'self-managing' their metabolic control.

Primary health care of diabetic patients in a specialized outpatient setting: a DIABCARE-based analysis.

To assure health care quality requires a tool for establishing the feedback between parameters of patient management and related standards. To assess the current situation and to evaluate the DIABCARE system as a potential monitoring instrument, a retrospective study was performed in 85 randomly assigned insulin-treated patients (72% type 2) who were regularly attending the diabetes outpatient unit (total of 3,595 patients) in a township of approximately 90,000 inhabitants. 1,195 records of sequential medical visits during the years 1987 and 1990 were analyzed. -Selected results (averages in 1987 vs. 1990): (1) Visits per year 6.9 vs 7.1; (2) intensified insulin treatment in 14 vs. 27% of all patients, they were on 4.0 vs. 4.5 injections per day applying doses of 0.8 vs. 0.6 IU kg-1 d-1; (3) glycaemic control: random blood glucose 6.0 vs. 5.8 mmol/l on conventional and 6.4 vs. 5.7 mmol/l on intensified regimes, HbA1 regular measurements in 2 vs. 21% of the patients; (4) body mass index 26.4 vs. 26.6 (conventional) and 24.7 vs. 25.9 (intensified) kg/(m)2; (5) retinopathy prevalence 30 vs. 29%, in 4 vs. 29% of the patients no information; (6) nephropathy prevalence 7 vs. 11%, in 75 vs. 68% of the patients no information; (7) foot complications prevalence 6 vs. 9%, in 91 vs. 84% of the patients no pathological findings. -The DIABCARE monitor proved appropriate but too laborious. The general level of care showed a tendency towards improvement between the two investigated periods but dit not yet meet the standards which must be attained to attain the St. Vincent Declaration.

Diabetes prevalence from health insurance data: evaluation of estimates by comparison with a population-based diabetes register.

UNLABELLED: 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. RESULTS: total prevalence rates were (1) 4.8%, (2) 4.9%, and (3) 4.4% (p < 0.05). The percentages of insulin-treated patients were (1) 18%, (2) 19%, and (3) 16%, respectively. 54% (1), 37% (p < 0.05) (2), and 42% (p < 0.05) (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.

Artificial connection between glucose sensing and insulin delivery: implications of peritoneal administration.

The replacement of insulinogenic function in insulin-dependent diabetes has to restore the feedback between intracorporal glucose and insulin. This has been accomplished by the following approaches: (a) the so-called open-loop insulin treatment by means of injections or pumps, employing laboratory or other extracorporal analytical devices and closing the feedback at large intervals only; (b) transplantation of insulin producing tissue and the bioartificial pancreas, employing the natural beta-cell both for glucose sensing and insulin delivery; (c) implanted artificial drug delivery systems providing chemical feedback between intracorporal glucose and insulin release from a nonrefillable reservoir of limited capacity; (d) the intracorporal or paracorporal artificial beta-cell comprising a glucose sensor (electrochemical or other type) that permanently delivers the signal to the computer-controlled insulin pump. This artificial device works on the basis of an algorithm of glucose-dependent insulin provision, compensating for the lack of other regulators, for the site of insulin administration, which is usually posthepatic, and for the kinetic properties of sensing system, e.g., a subcutaneous inserted amperometric electrode. Present experimental studies show that the pharmacodynamics of peritoneally applied insulin may be implemented into a mathematical model of the overall glucose-insulin system. They include absorption nearly as fast as after intravenous application, predominant portal inflow and approximately 30% hepatic removal. Feedback-controlled peritoneal insulin administration by means of an artificial beta-cell working on peripheral-venous blood glucose monitoring results in normal glycemic profiles under basal conditions and during oral glucose loads, if the pharmacodynamic properties of the peritoneal route are implemented into the insulin dosage algorithm.

Computer-aided systems in the management of type I diabetes: the application of a model-based strategy.

One approach of improving metabolic control in type I diabetic patients is the application of computer-aided procedures aimed at supporting the decision on optimal therapeutic regimens. To accomplish this, a complex strategy was developed which in an individual patient permits (1) the evaluation of metabolic data by means of statistical and graphical methods, and (2) the prediction of the outcome in feedback and in non-feedback-controlled insulin therapy. The latter is realized by means of simulation, employing a structured model of the glucose-insulin control system where the model parameters can either be identified individually or be taken at random. The practical applicability was validated in C-peptide-negative type I diabetic patients who were on intensified insulin injection therapy. The comparison between theoretical predictions and daily glycaemic profiles measured by the patients under ambulatory conditions showed close correspondence which justifies the application of this method as a clinical decision support.

Experimental validation of a glucose-insulin control model to stimulate patterns in glucose turnover.

To verify a structured model of the glucose-insulin system, metabolic measurements were compared with model-based simulations in insulin-dependent diabetic dogs which had been previously identified in terms of model parameters. Glycaemia, glucosuria, plasma insulin, and the rates of appearance Ra and disappearance Rd of glucose (kinetics of double-labelled glucose, evaluated according to Steele's equation in its non-steady-state version) were observed under the following conditions, starting from normoglycaemia during glucose-controlled insulin infusion (GCII): (I) insulin withdrawal, (II) insulin withdrawal and glucose infusion, (III) constant i.v. infusion of glucose and insulin, (IV) glucose infusion during GCII. After fitting the patterns of glycaemia, simulations of the other state variables were accomplished, employing the individual model parameters, the preset experimental inputs, and the GCII control constants (test IV only). Under nearly all conditions, correspondence was excellent between measured and simulated data. There were, however, the following exceptions: incomplete representation by the model of kinetics in glucose utilisation after interruption of insulin supply, overestimation of glucosuria by the model in the presence of insulin. It is concluded that the model provides a reasonable representation of metabolic processes which are of importance in the treatment of insulin-dependent diabetes mellitus and that it might thus appropriately simulate the outcome of metabolic regimens.

KADIS--a computer-aided decision support system for improving the management of type-I diabetes.

Despite the introduction of new therapeutic aids such as insulin pumps and injectors, blood glucose test tapes, particular insulin formulations, and the physiological basis-bolus principle of insulin dosage regimes, the metabolic care of most insulin-dependent patients is still insufficient. One potential tool of further improving the results in diabetes treatment consists in the application of computer-aided procedures to estimate individually optimal regimes. Employing a validated mathematical model of the glucose/insulin metabolic control system and individual sets of data from patients' self-monitoring, a software package was developed on a micro-computer which allowed both the retrospective analysis of data resulting from the therapeutic process, and the prospective simulation of the outcome of alterations in the regime in terms of glycaemia and insulinaemia. The two parts of the programme provide either for the patient or for the physician an interactive mode of working with the computer. The system is now being validated by means of a long-term follow-up study in type-I diabetic patients. It may be used mainly in diabetic outpatient centers and as a tool of educating, training, and motivating patients.

A model-based system for the individual prediction of metabolic responses to improve therapy in type I diabetes.

Despite recent achievements such as home glucose monitoring and intensified injection regimens or insulin pumps, the metabolic care in diabetic patients is still mostly insufficient. One approach of further improving the management is the application of computer-aided procedures to estimate individually the optimal regimes. To accomplish this, a model-based strategy was developed which permits the prospective assessment of the metabolic outcome. This strategy comprises the following components: (1) a validated model of the physiological glucose-insulin regulatory system; (2) a procedure for identifying the metabolic situation of a given patient in terms of the model parameters; (3) methods of estimating the pharmacokinetics of insulin and its effect on glycemia, the absorption profiles of ingested glucose equivalents, and the effect of exercise as expressed in equivalents of insulin action; (4) computer procedures of prospective simulation of glycemic profiles around the day under the influence of selected or proposed therapeutic regimes. The entire method has been validated in C-peptide negative type I diabetic patients by comparing experimental results with theoretical predictions from model-based simulations over up to one year. This model-based simulation may be applied by ambulatory patients together with their physicians as a decision-support system in selecting appropriate individually suited regimes.