Technical note: Evaluation of three autoanalysers for use in chemical pathology

Three autoanalysers, the EPOS 5060, ERIS 6170 and the Analyst, were evaluated for their adequacy for use in laboratories attached to Saudi ‘polyclinics’. All the analysers showed comparable within-batch imprecision. The Analyst was found to be the most useful because it was simple and practical and because its throughput time was faster than the other two analysers. The EPOS 5060 would be more suitable for screening large numbers of samples for single parameters; and the ERIS 6170 essentially suits normal routine chemical laboratory work.


Introduction
A number of 'polyclinics' with clinical laboratories were set up in many parts of the Kingdom of Saudi Arabia by the Government in 1984 (these 'polyclinics' are similar in concept to the UK's Health Centres). In th Riyadh health region alone, 80 of the 120 such governmentowned 'polyclinics' have clinical laboratories. There are also as many private 'polyclinics' with laboratories in the Riyadh region.
In these 'polyclinics', chemical pathology tests were to be performed with autoanalysers, the results of which were to be reliable and comparable to those produced by the larger hospital laboratories in major cities. An analyser that would suit these 'polyclinics' should be durable; show good, consistent imprecision and accuracy; provide a reasonably quick turn-round time; and should require a very short training period.
The Central Laboratory and Blood Bank in the Central Hospital in Riyadh is a large laboratory. Between 1984 and 1986, various biochemical autoanalysers were evaluated by the Biochemistry Department at the Central Laboratory and Blood Bank. Amongst these were the EPOS 5060, ERIS 6170 (both sold by Eppendorf Geratebau, FR Germany); and the Analyst (Du Pont, USA). Comprehensive studies of the EPOS 5060 and ERIS 6170 have already been published [1,2]. Only those features of interest to this study are discussed here. So far, no independent study of the Analyst has been published.
The purpose of this study was to determine the adequacy of these autoanalysers for use in 'polyclinics' in Saudi Arabia. GmbH), the autoanalyser in current use in the department for daily routine work. Samples are introduced for assay via a chain link, which is similar to the Epos. However, unlike the Epos, any combination of multiple choice of test assays on any one patient's sample can be analysed. It is also an open system, but during this study reagents from E. Merck (Darmstadt, FR Germany), which were supplied by the agent were used. A built-in computerized quality-control scheme facility is available, however this was not tested nor was it utilized rigorously during this study. The cost price of the instrument was about 600 000 SR and between 0"04-1"40 SR per test.

The Analyst
This is a bench-top, compact, sophisticated multichannel discrete profiler analyser. It measures about 216 x 635 x 330 mm and weighs about 19 kg. Reagents are supplied by Du Pont in prepacked disc rotors, which also contain the reaction cuvettes. At the time of this study, the following disc rotors were available.
(a) A 12  All chem rotors used in this study still had at least 9 months' shelf-life before their recommended expiry dates.
It is a completely closed system with respect to reagent utility. However, it requires only 90 btl of serum for the assay of all the parameters in the 12 chem rotor. The Analyst does not require daily calibration like most other multichannel discrete autoanalysers. It also has a built-in quality-control system, but this facility was neither tested nor used rigorously during this study. The Analyst costs about 40 000 SR and approximately 1-2"5 SR per test.

Quality control
Throughout this study two sets of quality-control (QC) assessment schemes were maintained. One set was physiological and pathological internal (QC) sera sup-   The effect of high concentrations of bilirubin and lipids were evaluated using highly icteric and turbid samples chosen randomly from patients' specimens. The throughput time of each analyser was studied by taking the average time to assay, in triplicate, 12 identical parameters.

Results
Table l(a) shows the variation in the mean accuracy values quoted for the quality control (Biotrol) sera for the parameters indicated between the three autoanalysers tested. On the basis of these results, accuracy can be classified as good. Imprecision was also, on average, satisfactory, see table 2. The dynamic throughput times are given in table 6. Although the times observed during this study are slightly longer than the manufacturers' claims, they are quite satisfactory.
Correlation scatter plots between the Hitachi 705 and ERIS 6170 and Hitachi 705 and Analyst are shown in figure 1; and those between the ERIS 6170 and Analyst in figure 2 for the parameters indicated. The corresponding calculated correlation coefficients are shown in tables 4 and 5. Good linear positive correlations were observed for the parameters indicated. The correlation observed for creatinine between the ERIS 6170 and the Analyst is interesting since the study was not carried out simultaneously on the same day, as were those between the Hitachi 705 and ERIS 6170, and Hitachi 705 and Analyst.
Imprecision, on the whole, was best in the Analyst, even with highly icteric and lipaemic samples. In general, imprecision was observed to be very consistent in all three analysers. The within-batch imprecision values, shown in table 2, are satisfactory for the three analysers.
The overall performance of the three analysers was quite satisfactory. Practically, the Analyst, a bench-top analyser would be preferred in Saudi's 'polyclinics' and in small paediatric laboratories. The Analyst is compact, portable and much easier to operate, than either of the other two analysers. The suitability of the Analyst for the UK's National Health Service Health Centres has also been discussed in Mires Magazine [5] (the Health Centres are similar to the 'polyclinics').
The EPOS 5060 would be suitable for hospital laboratories which serve daily disease-related clinics, for example diabetic clinics. The ERIS 6170, on the other hand, is a reliable analyser that could prove invaluable in any large routine clinical chemical laboratory.