Laboratory evaluation of the Greiner G450 discrete analyser

A report of an evaluation of the Greiner G450 selective multichannel analyser is presented. Thirty different single-reagent tests can be carried out in one run at a rate between 200 and 400 tests/h. A double beam photometer allows kinetic and end-point measurements. A Radiometer FLM3 flame photometer has been included for the determination of electrolytes. The G450 analyser demonstrated excellent precision, linearity, accuracy and no carry-over. Results for 16 different analytes as determined with the Greiner G300, Astra 8 Beckman and Progress Kone correlated well with those obtained with the G450.


Introduction
The G450 analyser is a selective multichannel system designed for determining enzyme activities and concentrations of substrates and electrolytes. The performance of the analyser was evaluated for 16 variables and the results were compared with those obtained with other instruments in routine use in the authors' laboratory.

Instrumentation
The sample and the reagents are dispensed into process tubes which are advanced step-by-step in the incubator thermostatted at /30C and are submitted to an oscillating mixer. The sample volume can vary from 5 to 100 tl. Reagents (up to 30) are kept in a refrigerated compartment. Up to four different reagents can be used for each determination. The volume distributed can vary from 50 to 500 tl. The process tubes are reusable after washing but must be replaced after 10000 tests. The reaction mixture is transferred to the reading cell with a thermostatted needle. The G450 is equipped with a double-beam filter photometer (eight wavelengths from 334 to 650 nm).
The light of the mercury lamp is split into two beams and a thermostatted carroussel with eight quartz cuvettes passes through both light beams. A rotating filter wheel allows the analyser to monitor seven readings at s intervals for each beam and wavelength. With this procedure, kinetic measurements can be performed with polychromatic readings.
The G450 has been integrated with the authors' laboratory computer system-MicroMega 32 R Alcatel Thomson ].
The speed of the analyser varies from 200 to 400 tests/h, depending on whether a sample (or reagent) blank is used. A Radiometer photometer was included for the determination of the electrolytes. concentrations between 0 and 4 absorbance units (A). The coefficient of variation (CV) was less than 1.00% between 0 and 2"5 A and 1"00 to 5"13% for the upper absorbances.

Methods
The linearity of the photometer was tested for all eight wavelengths by Degiampetro et al. [2]: the absorbance varied linearly between 0 and 2 A with a CV of less than 0"4% for all wavelengths.

Accuracy of sample pipetting
This assay was performed by measuring the absorbance at 405 nm after sampling various volumes of PNP (0"1 mmol/1 in NaOH 10 mmol/1) :5 to 100 tl. The volume of diluant was also varied (distilled water: 50 to 200 1).
Each measurement was performed 10 times. The CVs were less than 1"6% and the percentage of the theoretical volume was between 94 and 102"5%.

Accuracy of reagent dispenser
This assay was performed by measuring the absorbance at 405 nm after dispensing various volumes of PNP (0"216 mmol/l in NaOH 10 mmol/1):50 to 500 1. Each measurement was performed five times. The CVs were less than 1% and the percentage of theoretical volume was between 97 and 99%.
Within-run and between-run accuracy Within-run accuracy was tested using three pools of serum (or plasma for glucose analysis) and urine from patients at the hospital: L (low), M (medium) and H (high). Thirty consecutive measurements were made for each pool and each parameter. Results (means, standard deviations [SD] and CVs) are given in tables 2(a) and 2(b). Between-run precision was tested using the above pools in 30 aliquots, for the determination of analyte concentrations, stored at -80C. For the determination of enzyme activities 20 aliquots, were used with the exception of ALT and LD for which a decrease was   [3]. For each parameter, a specimen with a low concentration (L 1) and one with a high concentration (H 1) were analysed 10 times consecutively. The H2, H3, L2, L3 sequence was then carried out five times. The means and SD were calculated for L1, L2, L3 and H1, H2, H3. There was no significant difference between H1 and H2, H2 and H3, L1 and L2, L2 and L3 (data not shown).
Transfer dispenser and cuvettes: this assay was performed according to the recommendations of the French Society of Clinical Chemistry [3]. One ALT test was inserted between every four LD measurements. This procedure was repeated 10 times. No interference was observed in the determination of LD.

Linearity of methods
For enzymatic activities, sera with abnormally high activities were diluted in 9 g/1 NaC1. For the other determinations, the analytes were dissolved and diluted in distilled water or 9 g/1 NaC1 (determination of T. Proteins). All measurements were carried out in duplicate. Results are given in table 4.

Discussion
Both within-run and between-run precision was good and in agreement with results obtained using other analysers [2 and 4]. Less good results were obtained in the within-run assays for ALT (CV for L pool: 10"05%) and CK (CV for L poo1:9"32%). With regard to the between-run assays we obtained the highest CV in 4 series:iron (CV for L pool: 5"35%), uric acid (CV for L poo1:8.46%), AST (CV for L pool: 7"30%) and ALT (CV for L pool 7"85%).
These relatively high CVs can be explained by the very low level of analytes present.
Being both a high-rate routine and stat analyser, the G450 is particularly suited to large public health laboratories.
The results obtained with the G450 correlated well with data from other routine instruments; the slope of the regression curve was between 0"89 and 1.11. Efficient cleaning of the cuvettes and sample and transfer dispensers minimized carry-over.
Daily start-up requires 20 min for cuvette calibration, reagent preparation and an additional 20 min for zero setting and daily quality control.
The 30 reagent dispensers need to be cleaned regularly-at intervals related to the stability of the reagents.