An evaluation of the Eppendorf EPOS 5060 biochemistry autoanalyser

Up to 30 different methods may be memorized, with one or two reagents available for each. Reagent is containecl in a vessel with 80 ml maximum volume, whose cover is formed in such a way that it can take, at any one time, up to six standards of different concentrations, three control sera and a cleaner. These may also be set up on the sampler chain. The reagent 2 container is located above the rotor and is positioned automatically according to the preselected times. Its maximum volume is 20 ml.


Introduction
The EPOS 5060 is a selective batch analyser for routine and special work in clinical chemistry, with a theoretical throughput of 300 tests/h. The evaluation reported here was done according to the protocol of the Sociedad Espafiola de Quimica Clinica (SEQC) [1]. Different components of the analyser were evaluated for representative analytes: creatinine, total protein, glucose, AST and alkaline phosphatase; comparing results with those from the Ultrolab-Aurora analyser [2].

The instrument
The EPOS 5060 is manufactured by Eppendorf Gmbh (Hamburg, FR Germany) and is distributed in Spain by Merck-Igoda, S.A.
It consist of a single module (70 x 110 x 70 cm), and it weighs approximately 140 kg. All measuring procedures and calculations are completed in the one unit.
A data processor (6410) is available for rapid request input, work list preparation (also for additional analysers) and the preparation of patient reports and qualitycontrol data. The present evaluation did not include the data processor.
The reaction rotor consist of 20 quartz glass cuvettes with a 10 + 0"01 mm lightpath. Temperature is controlled through a Peltier system and is adjustable between 20 and 40 C (+ C). An appropriate washing solution is used for cleaning and drying by a filtered air system. Up to 30 different methods may be memorized, with one or two reagents available for each. Reagent is containecl in a vessel with 80 ml maximum volume, whose cover is formed in such a way that it can take, at any one time, up to six standards of different concentrations, three control sera and a cleaner. These may also be set up on the sampler chain. The reagent 2 container is located above the rotor and is positioned automatically according to the preselected times. Its maximum volume is 20 ml.
Since the instrument operates in batch mode, the reagent containers must be changed between different methods. At the beginning of a method the dispensing system operates always for filling and washing the tube system. Dispensing volumes are freely selectable within the preselected ranges and are set automatically. The dispensers use a Hamilton pipetting system, with plastic tubes for reagent and metallic tubes for reagent 2 and sampler.
The sample dispenser drifts down the sample tube lid before suction to avoid vacuum effects. Its range is between 2"5 and 50 1, adjustable in steps of 0"25 1. The reagent dispensing system has a range between 200 and 500 tl, in steps of 1.0 1 for reagent 1, and between 5 and 50 1, adjustable in steps of 0"25 1 for reagent 2.
Mixing takes place in the reaction cuvette by a compressed air system.
The last calibration is stored-so it is unnecessary to calibrate at each process batch.

Reagents
The following reagents were used:  For correlation with the Ultrolab-Aurora" Testomar creatinine combipack (Behring OVPC 10/11); Testomar total protein combipack (Behring OSOA 10/11); Cromatest glucose (Hexokinase) (Knickerbocker 250 B106); Testomar ASAT (GOT) IFCC (Behring OUCZ 10/11); Testomar alkaline phosphatase (DEA) ( From solutions prepared as previously, 30 successive measurements were obtained in the same cuvette, and from these were calculated the mean, standard deviation and coefficient of variation at both 340 and 405 nm [3]. Photometric linearity Using serial dilutions prepared as before, three successive determinations were made for each absorbance, always in the same cuvette [4]. Theoretical absorbances were calculated from the coefficient of molar absortivities of NADH and PNP. Photometric drift Photometric stability was studied over the first 30 min and at a free 12 h at 405 nm with PNP in NaOH solution of theoretical absorbance 1"00. Three successive determinations were made in the same cuvette. Sample pipette delivery system imprecision The reagent delivery system was set to dispense a constant amount of NaOH (20 mmol/l), and the sampler (sample pipette) to dispense volumes ranging from 2 to 50 zl of PNP solution. In each case, final absorbance was arranged to be around 0"5. Standard deviation and coefficient of variation were calculated from 30 determinations [5]. 78 Reagent pipettes delivery system imprecision The sampler pipette was blocked and NaOH (20 mmol/1) dispensed by the reagent pipette, and PNP solution by the reagent 2 pipette.
Five experiments were carried out with volumes between 100 and 300 1 for the reagent pipette, and to 60 1 for the reagent 2 pipette. Standard deviation and the coefficient of variation were calculated from 30 successive determinations [5].
Temperature control Warm-up time was studied making readings every 15 s, until three consecuti'e readings with a deviation of + C were obtained. 30 readings were then made at 20s intervals for 10min. The mean and variance were calculated.

Imprecision
Within the same run, 30 samples of control sera were tested at three levels, in order to study the within-run imprecision. To evaluate between-run imprecision, a further 30 samples were distributed in different runs.

Carry-over
Following a permutation order, according to the Comi-si6n de Instrumentaci6n de la SEQC [1], three control samples with different concentrations were distributed along the sample chain.
Relative inaccuracy Using 100 samples from patients, results from the EPOS 5060 were compared with those from the ULTROLAB-AURORA, applying regression analysis.

Photometric linearity
The obtained linearity is right for NADH at 340 nm and for PNP at 405 nm. The results are shown in figures and 2. There is some dispersion at low absorbances for both reagents.

Photometric drift
For the photometric stability test, PNP was read at 405nm for 30min, using a solution with a mean absorbance of 1"094. The coefficient of variation obtained was 0"24%. With the same solution read at 12 h, the mean absorbance was 1.039 and the coefficient of variation was 0"60%.

Pipette delivery system imprecision
For the reagent pipette, an imprecision of 0"78% was found for the largest volume for both the first and second reagents. For the smaller volumes, imprecision was 0"20.
Reagent volume therefore affects imprecision (see table  3). Sample pipette imprecision is also directly related to the volume dispensed (see table 3). The photometric imprecision can be considered to be negligible when compared with the delivery system.

Temperature control
Testing each 15 s, warm-up times were found to be as follows: to attain 25 C 240 s; to attain 30 C 150,s; to attain 37 C 330 s.
The variances found were smaller than those of the thermometer used.

Imprecision
Within-run and between-run imprecision is shown in

Conclusions
System practicability The Eppendorf EPOS 5060 is suitable for medium-size laboratories, or as a second instrument for large laboratories. It is especially useful for enzyme assays and is easy to maintain and operate with a short start-up time. The only external connection needed is a power supply. A waste container needs to be emptied periodically when indicated by an alarm. No special training is required for users. Any commercial reagents may be used and method change is not possible to introduce stat samples.
It is provided with an outline for the on-line linking to a central ordinator. Alarm flags appear with the results, giving information on linearity, and level of substrate for the central computer is provided.
The photometric linearity is acceptable, with some variation at low absorbances. The drift was negligible during the first 30 min, increasing slightly up to 12 h. The pipette delivery system has acceptable imprecision, except when operating with less than 2 1.

Working conditions system evaluation
There is no carry-over between samples. The imprecision for enzyme measurements is low and it is acceptable for the other analytes. Relative inaccuracy for total protein is not good, perhaps due to a calibration problem, since the reference instrument was calibrated with lyophilized serum, while an aqueous standard was used for the EPOS 5060. This aspect needs further study. The precision of the sample dilution system was found to be unsatisfactory but this problem has been resolved with new software. It will be the subject of a further evaluation.