With minor quantitative changes, we believe our findings are relevant to clinical and research blood analyses. Because aqueous and FCE proficiency testing materials differ from fresh human blood in viscosity, oxygen capacity and content, oxygen halfsaturation pressure of hemoglobin values, and temperature dependence, they can be expected to differ from blood, but FCE has shown much less variability than aqueous materials in measuring Po2. Two studies using fresh tonometered blood at several Po2 intensities found minimal model differences between a few instruments. In contrast, four studies using fresh human blood demonstrated major model differences in measuring Po2 at several intensities. Our comparison of tonometered blood data from the four instruments used by Scuderi et al with FCE data from approximately 150 instruments of the same models discloses that blood deviations averaged 81% of the FCE deviations we detected in the same models at 11 Po2 levels. If we combine data from two of our own studies, using multiple models and seven levels of Po2 intensity, blood deviations averaged 77% of FCE deviations between Po2 of 42 and 92 mm Hg and 100% for Po2 values of 145 mm Hg. Thus, one can estimate that fresh human blood Po2 deviations below the hyperoxic range are likely to be roughly four-fifths of those for FCE.
For Pco2 and pH comparisons, the differences between proficiency testing materials and fresh human blood have been studied less but are probably smaller, because both aqueous and FCE materials are stable and have similar buffering capacities. Blood is not a primary standard nor a good quality control material for pH, due to its inconvenience, infectivity, and variations in buffering capacity. One study using only two models of analyzers over a wide range of Pco2 showed blood deviations that were about 60% of those of ampules of several quality control materials. In contrast, another study using several analyzer models showed that blood deviations slightly exceeded FCE deviations for four Pco2 intensities from 22 to 65 mm Hg. Because Pco2 electrodes are modified pH electrodes, one would suspect that the pH deviations for blood are reasonably similar to those of either FCE or buffered aqueous materials. Because proficiency testing ampules can be manufactured in huge quantities and are stable over long periods, either aqueous or FCE materials seem preferable to blood and acceptable for comparing Pco2 and pH values between instruments and models. In the absence of other data, it seems reasonable to conclude that for Pco2 and pH measurement, FCE instrument and model differences are likely to be between IV3 and % of those for blood.