Multicentre evaluation of the Bayer DAX system.

The analytical performance of the DAX, a high-throughput random access analyser, was studied according to ECCLS guidelines (ECCLS Document Vol. 3, No. 2, Beuth Verlag, Berlin, 1986) in a multicentre evaluation involving four laboratories. The trial took about 4 months. Determinations of 12 analytes produced more than 60,000 data. The imprecision study on 3 control sera for all analytes gave a within-run CV (median of 4 laboratories) which never exceeded 3% and was below 2% in 94% of the results. The median between-day CV was less than 3% in 92% of the results, with a highest value of 5.0%. No significant drift was detected during the 5-hour work period. No relevant sample- and cuvette-related carry-over was found. The manufacturer's claims concerning linearity were fulfilled or exceeded. The recovery of the assigned values for the control sera (median of 4 laboratories) ranged from 94 to 106%. In the method comparison on patients' samples, deviations were statistically significant in some cases, due to differences either in the methods used or in the calibration of the systems used for comparison; the regression lines, as inspected visually, and the coefficients of correlation were, however, generally acceptable. Imprecision and inaccuracy were within the acceptability limits as recommended by the Société Française de Biologie Clinique (SFBC) (Biochim. Clin. 12 (1988) 284-327) and the Deutsche Gesellschaft für Klinische Chemie (DGKC) (Dt. Arztebl. 85 (11) (1988) A697-A712). The limits of acceptance, proposed more recently by Fraser et al. (Eur. J. Clin. Chem. Clin. Biochem. 30 (1992) 311-317), were met in thirty-three of thirty-six cases. The alpha-amylase assay was significantly affected by bilirubin and haemolysis; interferences for the remaining analytes were predictable and well-known from the literature. The rate of sample throughput was found to be in agreement with that claimed by the manufacturer. The software did not present problems and was readily accepted by the operators. The practicability of the instrument was rated very good. Since the DAX is the primary chemistry analyzer in all four participating laboratories, the present experiments were necessarily intermixed with a large routine workload. Therefore, the system performance was assessed under definitely "usual" conditions. Because of its high productivity and reliability, the DAX is highly suitable for routine use in medium to large sized hospitals.