Aqueous and serum-based materials compared for use as simulated calibrators for three ionized calcium analyzers.

To determine if bias between different ionized calcium analyzers could be decreased, we analyzed 10 control fluids during a study in which ionized calcium was measured in more than 150 serum and whole-blood samples. After calibrating three ionized calcium analyzers (Radiometer ICA 1, Nova 8, and AVL 980) with the manufacturers' respective calibrators, we used the between-instrument differences of the control fluids to simulate recalibration of the analyzers during each analytical run. A filtered human serum pool containing ionized calcium at 1 mmol/L concentration, with CO2 removed and having no added buffer, was the only material that consistently decreased between-analyzer bias of both serum and whole blood. Another human serum pool containing about 1.3 mmol of ionized calcium and about 10 mmol of bicarbonate per liter was even better at minimizing analyzer biases for serum samples, but was not as effective for whole-blood samples. Some additives used to buffer pH apparently adversely affected both the accuracy and precision of some, but not other, calcium ion electrodes. We conclude that if a reference material is developed for calibration of ionized calcium analyzers, it should be tested on several analyzers for use with both serum and whole blood, and it should be at least as effective as a human serum material, such as that used here.