1995excursion in the abundance of that isotope. The natural abundance of 34S is on the order of 4.0% and that of I5N is about 0.37%. It will, therefore, be possible to follow the dilution of 16N down to approximately a factor of 10 lower abundance than will be possible with Calculations based on these estimates suggest that 1 kg of %SO4 would label about 1 mile3 of air at the detection limit given a background level of about 8 pg/m3 sulfate. Sir:In recent years, the fabrication and study of various miniaturized versions of ion-selective electrodes (ISEs) have been vigorously pursued (1). Once fully developed, it is hoped that such electrochemical devices could be routinely used for continuous in vivo monitoring of blood electrolytes (e.g., Kt, Nat, C1-, etc.) during surgical procedures or at the bedside of patients in critical care units. Band and co-workers have reported considerable success in preliminary animal studies with simple scaled down versions of conventional ion-exchanger or neutral carrier based polymer membrane type ISEs (with internal electrolyte solutions) (2,3). Others have described the fabrication and performance of ion-selective field effect transistors (ISFETs) which are, in effect, solid-state devices (no internal reference electrolyte) (4-8). The ISFET has attracted considerable attention because it is envisioned that a single miniaturized solid-state chip could contain multiple gates and be used to sense several ions simultane-ISFETs which are not covered with polymeric ion-selective membranes can be used directly as pH sensors. The insulating gate materials are typically silicon nitride and silicon oxide which, by their own intrinsic surface properties, develop phase boundary potentials proportional to the logarithm of hydrogen ion activity of any solution they are in contact with. However, potassium ISFETs, for example, are usually prepared by coating the gate region of these pH devices with the same polymeric ion-selective membranes used in conventional potassium ISEs. Thus, once the gate region is covered with the polymeric membrane material, two potentials are generated: a Donnan potential at the interface of the membrane and the sample, dependent on the analyte ion activity in the sample, and an unknown potential at the interface of the membrane and the gate which is dependent on the activity of hydrogen ions present in this region. It is our belief that, since the hydrogen ion activity is not fixed at this membrane-gate interface, species which can enter this region by passing through the membrane can alter the pH and therefore interfere with the measurement of the analyte ions.In this correspondence, we provide preliminary data which strongly supports this view. Indeed, it is clearly shown that ously, ISFETs based on polymer ion-selective membranes are subject to positive interference by carbon dioxide and organic acids in the sample solution. As a model, we describe experimental results obtained with potassium ISFETs and nonselective ISFETs coated with polymer membranes containing...

The Glucose-Controlled Insulin Infusion System (Biostator) is a modular, computerized, feedback control system for dynamic control of blood glucose concentrations in diabetics. This on-line glucose analyzer for use with whole blood utilizes a novel enzyme (glucose oxidase)-membrane configuration and an electrochemical cell to measure the H202 generated. The analyzer exhibits both short- and long-range stability, and instrument response and analyte concentration are linearly related over the full range of clinical interest. The response is fast, accurate, and precise, and permits determination of blood glucose within 2 min from the moment the blood leaves the patient. Correlation studies were completed to show the agreement between the Biostator Glucose Analyzer and the FDA's recommended hexokinase/glucose-6-phosphate dehydrogenase procedure on whole blood (e.g., average per cent recovered for 11 concentrations between 250 and 900 mg/liter was: hexokinase, 95.6%, Biostator Analyzer, 95.9%; bias and SDd, respectively, at low, normal, and high glucose values were: 12 and 41 mg/liter at the 500 mg/liter level; 4 and 52 mg/liter at the 1000 mg/liter level, and 4 and 128 mg/liter at the 4000 mg/liter level). No appreciable interference is observed with above-normal concentrations of bilirubin, uric acid, creatinine, sodium salicylate, or dextran. Platelet adhesion, which tends to decrease the useful life of the membrane, has been significantly decreased.

Medical devices, intended for blood contacting applications, undergo extensive in vitro testing followed by animal and clinical feasibility studies. Besides the use of materials known to be intrinsically blood-compatible, the surface of such devices is often modified with a coating in order to improve the performance characteristics during blood exposure. In vitro evaluation of blood-device interactions accompanies the product development cycle from the early design phase using basic material geometries until final finished-product testing. Specific test strategies can vary significantly depending on the end application, the particular study objectives and variables of interest, and cost. To examine the degree to which findings derived from two different in vitro approaches complement one another, this report contrasts findings from a simple multipass loop model with findings from a simulated cardiopulmonary bypass (CPB) model. The loop model consists of tubular test materials, with and without surface modification, formed into valved Chandler loops. The CPB model has an oxygenator with and without surface modification connected to a reservoir and a blood pump. The surface modifications studied in this report are the Carmeda BioActive Surface and Duraflo II heparin coatings. Common blood parameters in the categories of coagulation, platelets, hematology, and immunology were monitored in each model. Ideal models employ the optimal level of complexity to study the design variables of interest and to meet practical cost considerations. In the case of medical device design studies, such models should also be predictive of performance. In the more complex and realistic simulated CPB model, experimental design and cost factors prevented easy/optimum manipulation of critical variables such as blood donor (use of paired samples) and heparin level. Testing in the simpler loop model, on the other hand, readily offered manipulation of these variables, and produced findings which overlapped with observations from the more complex CPB model. Thus, the models described here complimented one another. Moreover, conclusions from consistent findings, such as favorable responses associated with the heparin coatings, between the two models were considered to be more robust.

No abstract

No abstract