Background: Recently, the GEM Premier blood gas analyser was upgraded to the GEM Premier 3000. In addition to pH, pCO2, pO2, Na+, K+, Ca2+, and hematocrit measurement, glucose and lactate can be measured on the GEM Premier 3000. In this prospective clinical study, the analytical performance of the GEM Premier 3000 was compared with the Ciba Corning 865 analyser for blood gas/electrolytes/metabolites, and for hematocrit with the Sysmex XE 2100 instrument.
Methods: During a 6-month period, 127 blood samples were analysed on both the GEM Premier 3000 analyser and our laboratory analysers (Ciba Corning 865/Sysmex 2100 instrument), and compared using the agreement analysis for quantitative data.
Results: With the exception of K+, the other parameters (pCO2, pO2, Na+, Ca2+, hematocrit, glucose, and lactate) can be described in terms of the mean and standard deviation of the differences. For K+ measurement, a clear linear trend (r=0.79, p<0.001) in the deviation of the GEM Premier 3000 from the Ciba Corning was noticed, ie, in the lower or upper K+ reference range, the GEM Premier 3000 measured systematically too low or too high, respectively. Furthermore, in comparison with the other parameters, a therapeutically unacceptable systematic difference (mean of difference: -2.2%, p=0.05) in hematocrit measurement on the GEM Premier 3000 was observed for hematocrit values below 30%. The variance of the readings for the GEM Premier 3000 measurements was at clinically acceptable levels.
Conclusion: The GEM Premier 3000 analyser seems to be suitable for point-of-care testing of electrolytes, metabolites, and blood gases during cardiopulmonary bypass. However, its downward bias in hematocrit values below 30% suggests that using the GEM Premier 3000 as a transfusion trigger leads to overtreatment with packed red cells.