In the clinical laboratories, the same test could be analyzed by more than one assay
system or analyzer due to various requirements. This may cause a problem in patient monitoring. In our laboratory, stat TSH testing has been recently implemented in the stat biochemistry laboratory for
meet emergency request meanwhile routin analyses have been carried out in the endocrine
laboratory. This study was aimed to investigate the concordance between test results generated by
different asssay systems or analyzers, and to define whether the differences are in the acceptable
limits or not.
Materials and Methods: Fifthy three sera remaining from the routine analyses were used for the
method comparison study; of which TSH concentrations ranged from 0.029 µIU/ml to 165.0 µIU/ml.
After analyzed by Siemens Immulite 2000 XPi (instrument 1) in the endocrine laboratory, the samples
were analyzed by the other Siemens Immulite 2000 XPi (instrument 2) in the endocrine laboratory
and the Abbott Architect i2000 at the stat biochemistry laboratory on the same day. The concordance
between the results were evaluated by a regression analysis and Bland-Altman graphics by means of
MedCalc statistics software.
Results: Strong correlations between the TSH results obtained from the three analyzers were
obtained; the systematic difference between Siemens Immulite 2000 XPi analyzers was 4.7%, and the
systematic difference between Siemens Immulite 2000 XPi (instrument 1)-Abbott Architect i2000 was
22.8%, and finally it was 23.1% between Siemens Immulite 2000 XPi (instrument 2)-Abbott Architect
i2000. The TSH results generated by Abbott Architect i2000 were lower then those generated by
Siemens Immulite 2000 XPi.
Conclusion: Due to the high systematic diferences, higher than total allowable error, obtained
between Abbott Arhitect and Siemens Imulite systems, we assume the use of two diferrent systems
at the same time for the same test may cause problems in patientsÕ monitoring. If the use of two
different system at the same time is obligatory, clinicians should be informed about this situation.
The name of the analyzer or the system which generates the results should be involved on the
laboratory reports. In addition, our findings demonstrated that in case of a new method to be used in
the laboratory, there is a need for a comparison to the already existing method by means of patient
based samples and also for a standardization between methods.