Evaluation of a blood gas and chemistry monitor for use during surgery

An observational study was performed to evaluate a new blood gas and chemistry monitor (GEM-6 Diamond Sensor Systems, Ann Arbor, Michigan) in nine patients during cardiac surgery. Paired blood samples were analyzed by the instrument under test and by standard clinical laboratory instruments. The differences between the measurements of the new and the standard instruments are summarized as follows (mean +/- standard deviation, units of measure, number of samples): pH (-0.039 +/- 0.02, 154); PCO2 (2.63 +/- 1.8 mmHg, 154); venous PO2 (-2.0 +/- 3.0 mmHg, 72), hematocrit (4.7 +/- 2.7%, 98), potassium (0.18 +/- 0.13 mmol, 100), and ionized calcium (0.195 +/- 0.11 mmol, 100). Because the differences in arterial PO2 measurements were markedly heteroscedastic, a logarithmic transformation was employed, which upon retransformation gave the test instrument's 95% confidence limits as within 5.1% below to 46% above the nominal value on 82 samples. However, on the 14 samples having nominal values below 165 mmHg (the upper limit of the calibrated range of the GEM-6) the 95% confidence limits were from 5.4% below to 23.6% above the nominal reading. No failures of the test instrument occurred during the evaluation, and quality control standards run before, midway through, and again after sampling from each patient all gave readings within the manufacturer's tolerance. For all variables except hematocrit and ionized calcium, this instrument matches the values from the laboratory well enough over the clinically important range to supplant it for intraoperative monitoring purposes.