• Steven Teerenstra, Jacoline Steinfelder-Visscher, Jacqueline Klein Gunnewiek, and Patrick W Weerwind.
• Department for Health Evidence, Section of Biostatistics (133), Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB, Nijmegen, The Netherlands, z824116@umcn.nl.
• J Clin Monit Comput. 2014 Apr 1;28(2):133-8.

AbstractMost point-of-care testing analyzers use the conductivity method to measure hematocrit (hct). During open-heart surgery, blood-conductivity is influenced by shifts in electrolyte and colloid concentrations caused by infusion media used, and this may lead to considerable bias in the hct measurement. We evaluated to what extent different analyzers correcting for 0, 1, 2, or 3 factors, respectively, compensated for this electrolyte/colloid interference: (1) the conductivity method with no correction (IRMA), (2) with a [Na(+)]-correction (GEM Premier 3000), (3) with a [Na(+)]/[K(+)]-correction (i-STAT), and (4) with a [Na(+)]/[K(+)]-correction in combination with an algorithm that estimates the protein dilution [i-STAT in cardiopulmonary bypass (CPB)-mode]. Bias in hct was measured during three consecutive stages of a CPB procedure: (I) before CPB, (II) start of CPB and (III) after cardioplegia. In order of high to low electrolyte/colloid interference: the analyzer with no correction, [Na(+)]-correction, [Na(+)/]/[K(+)]-correction, and [Na(+)/]/[K(+)]/estimated protein-correction showed a change of bias from stage I to stage III of -3.9 ± 0.5, -3.4 ± 0.4, -2.1 ± 0.5, -0.3 ± 0.5%. We conclude that correcting for more parameters (Na(+), K(+), estimated protein) gives less bias, but residual bias remains even after [Na(+)/]/[K(+)]/estimated protein-correction. This suggests that a satisfactory algorithm should also correct for other colloidal factors than protein.

16 keywords...

hide…