Search PubMed⌕ Search

Biomedical subjects

Z Trajanoski

Publications and source records attributed to Z Trajanoski.

24 records · Page 2Linked to original sources
PubMed1996

Fuzzy filter for state estimation of a glucoregulatory system.

A filter based on fuzzy logic for state estimation of a glucoregulatory system is presented. A published non-linear model for the dynamics of glucose and its hormonal control including a single glucose compartment, five insulin compartments and a glucagon compartment was used for simulation. The simulated data were corrupted by an additive white noise with zero mean and a coefficient of variation (CV) of between 2 and 20% and then submitted to the state estimation procedure using a fuzzy filter (FF). The performance of the FF was compared with an extended Kalman filter (EKF) for state estimation. Both the FF and the EKF were evaluated in the following cases: (a) five state variables are measurable; three plasma variables are measurable; only plasma glucose is measurable; (b) for different measurement noise levels (CV of 2-20%); and (c) a mismatch between the glucoregulatory system and the given mathematical model (uncertain or approximate model). In contrast to the FF, in the case of approximate model of the glucose system, the EKF failed to achieve useful state estimation. Moreover, the performance of the FF was independent of the noise level. In conclusion, the FF approach is a viable alternative for state estimation in a noisy environment and with an uncertain mathematical model of the glucoregulatory system.

Blood Glucose↗
PubMed1996

Portable device for continuous fractionated blood sampling and continuous ex vivo blood glucose monitoring.

The objective of the study was to develop and evaluate a portable device for continuous fractionated blood sampling and continuous ex vivo monitoring of blood glucose. The inner lumen of a double lumen catheter (18 gauge x 45 mm) was placed in a peripheral vein and perfused with heparin solution (1.4 U min-1). The outer lumen was used to collect heparinized blood into 48 vacuum tubes at programmable sample volumes and time intervals (0.2-2 ml in 2.5-30 min). A sensor flow chamber with an internal volume of 1 mm3 incorporating a miniaturized thin-film amperometric glucose sensor was placed in the sampling line for continuous ex vivo blood glucose monitoring. Blood glucose and plasma insulin were measured during a frequently sampled intravenous glucose tolerance test (250 mg kg-1) and a subsequent oral glucose tolerance test (150 g) over 6 h in eight healthy volunteers (BMI 24.5 +/- 3.2 kg m-2). Additionally, in four experiments blood glucose was measured on-line using the glucose sensors. The overall correlation coefficients for whole blood glucose and plasma insulin between the manually drawn samples and the vacuum tubes were 0.73 and 0.87, respectively (p < 0.001). The miniaturized glucose sensor exhibited a linear measuring range of 25 mmol-1 glucose concentration and 95% response times of less than 30 s. Sensor readings and laboratory analyser results for the blood glucose measurement correlated between 0.93 and 0.98 (p < 0.001). In summary, continuous fractionated blood sampling and ex vivo blood glucose monitoring in ambulatory subjects is possible with a portable device.

Adult↗
PubMed1996

Accuracy of home blood glucose meters during hypoglycemia.

OBJECTIVE: To evaluate the accuracy of home blood glucose meters during hypoglycemia. METHODS: Six blood glucose meters-One Touch II (LifeScan, Milpitas, CA), Companion II (Medisense, Cambridge, U.K.), Reflolux (Boehringer Mannheim, Mannheim, Germany), Accutrend (Boehringer Mannheim), Elite (Bayer, Munich, Germany), and HemoCue (HemoCue, Angelholm, Sweden)-were compared with a reference method (Beckman Glucose Analyzer 2). Glucose concentrations from arterialized venous blood samples were measured using all glucose meters (whole blood) and the reference method (plasma) during hypoglycemic-hyperinsulinemic clamps in 15 subjects. RESULTS: In total, 663 blood glucose monitor readings and 119 reference values ranging from 2.28 to 3.89 mmol/l were analyzed. The correlation coefficients and the percentage of measurements within 20% and outside 40% of the reference values for each glucose meter were as follows: One Touch II: 0.91, 99.2% and 0%; Companion II: 0.81, 88.2% and 2.5%; Reflolux: 0.78, 85.0% and 0.9%; Accutrend: 0.88, 46.0% and 6.6%; Elite: 0.78, 75.6% and 4.2%; and HemoCue: 0.93, 96.6% and 0% (P < 0.001). CONCLUSIONS: There were substantial differences between the blood glucose meters during hypoglycemia, and none of the devices met the latest criteria recommended by the American Diabetes Association.

Blood Glucose↗
PubMed1995

Numerical approximation of mathematical model for absorption of subcutaneously injected insulin.

A pharmacokinetic model is modified to enable quantitation of subcutaneous insulin absorption following insulin injections of soluble insulin and monomeric insulin analogues. The model for soluble insulin includes diffusion, equilibration between hexameric and dimeric insulin and absorption of dimeric insulin molecules. Numerical approximation is carried out by modelling the whole system as a capacitor-resistor network with lumped elements and discrete sources and sinks. By means of the analytical solution for monomeric-insulin absorption, it can be shown that the approximation scheme yields sufficiently accurate results. The modified model for soluble insulin demonstrates dose- and concentration-dependent insulin absorption within the range of therapeutic concentrations and volumes. Additionally, parameters are estimated from published glucose-clamp data. The results of the data fitting indicate that the model presented is adequate for pharmacological studies. The model is suitable for individual parameter estimation from the time course of plasma insulin or from the disappearance curves of radiolabelled injected insulin.

Humans↗
PubMed1993

Pharmacokinetic model for the absorption of subcutaneously injected soluble insulin and monomeric insulin analogues.

A subcutaneous insulin absorption model is presented for parameter estimation from the time course of plasma insulin. Modifications of a published model were made for the absorption of soluble insulin and monomeric insulin analogues in the range of therapeutic concentrations and volumes. The modified diffusion-dissociation model with distributed parameters was approximated by a multiple-compartment model. Subcutaneous absorption of soluble insulin and monomeric insulin analogues with various volumes, concentrations, and injection depths was simulated. The model for soluble insulin exhibits volume, concentration, and injection depth dependent absorption, as experimentally observed. It was found that binding of soluble insulin in the subcutaneous tissue is negligible for U-40 and U-100 strengths. The absorption of identical doses (10 U) of soluble U-40 insulin was markedly faster (T-50% = 159.4 min) than the absorption of U-100 (T-50% = 196.2 min). According to the simulation results, the absorption rate of monomeric analogues is not dependent on concentration. No significant chances of the absorption rate could also be observed by varying volume and injection depth of the monomeric analogues.

Computer Simulation↗