A simple flow-injection method for the determination of blood glucose using a Technicon immobilized enzyme coil

The applicability of a single-channel flow-injection system with immobilized enzyme coil (Technicon) and UV detection to the determination of glucose is described. The method was used for a pure glucose solution and for serum. The detection limit was 0.10 mM, the rate of determination was 20-40 per hour and the precision was satisfactory. The system is very simple and practical when many analysis are to be determined periodically.

. Manifoldfor the flow injection analysis of glucose in biologicalJluids. The manifold consists of a peristaltic pump (P), injector (I), enzyme coil (C) and UV detector (D).

Introduction
The determination of glucose in serum, plasma and other biological samples is frequently required in clinical chemistry. In recent years various methods for the determination of glucose have been based on flow-injection analysis (FIA) with enzyme columns that are not commercially available ]. This paper describes a simple determination of glucose by FIA in combination with a Technicon immobilized enzyme coil for the Auto-Analyzer II system [2,3].

Apparatus
The manifold was build as shown in figure 1. An SHS 200 peristaltic pump (FIAtron Systems, 510 S. Worthington Street, Oconomowog WI 53066, USA), an actuator provided by Bitbk (Box 124, Malmv/igen 28, S-19122, Sollentuna, Sweden) and an Ultrospec 4050 UV spectrophotometer (Pharmacia LKB Biotechnology, Herredsvejen 2, DK-3400 Hillerod, Denmark), fixed at a wavelength of 340 nm and coupled with an RE 511 Kompensationschreiber (Bruwn Boviri, Norm Elec. Horsvinget 7, DK-2630, Taastrup, Denmark) recorder were used. Polyethylene tubing of 0"50 mm i.d. was used throughout the system. experiments were performed at room temperature. After injection, the mixture was led through the enzyme coil cell where the enzymatic reactions took place (figure 2), through a flow cell and from there to waste. The retention time was about 3 rain.
The sample concentration was calculated on the basis of the peak height, by reference to a calibration graph obtained by linear regression.

Results and discussion
Calibration data for glucose standards were measured at concentrations 0"00, 0"56, 1.11, 1"67, 2"78 and 3"33 mM with six determinations at each level. The correlation coefficient was 0"994 and the mean coefficient ofvariation (CV) was 1"5%, ranging from 0"0 to 3"6%. Carry-over was tested as described by Andersen and Hannibal [3] and was found to be 2%. The detection limit was found to be 0.10 mM, which covers the normal range of plasma glucose (4"2-6"7 m) and hypo and hyperglycaemic levels resulting from metabolic disorders. The mean glucose concentration from a serum standard was found to be 6"3 mM (CV 2-75%, n 6), compared with the declared concentration of 6"03-6"49 mM. When the glucose level was determined in serum or plasma, no interference of proteins was observed. The rate ofdeterminations was 20 per hour because of the high viscosity, but dilution (3 5) with 70% ethanol increased the capacity to 40 determinations per hour. When full blood samples are analysed, addition of heparin (15 bd for a ca 200-bd sample) is needed, followed by centrifugation.
In conclusion, the system is simple and practical when many analyses are to be performed periodically.