Carbon dioxide and oxygen partial pressure measurements in the cerebrospinal fluid in a conventional blood gas analyzer: analysis of bias and precision.

Cerebrospinal fluid (CSF) gas tension measurements have been used as a marker of cerebral oxygenation in animal models and in human studies. Discrepancies in the measurement of PCO2 and PO2 in non-blood solutions by standard blood gas analyzers have been described. The CSF is a physiological non-blood fluid with an ability to carry only dissolved O2 or CO2. The aim of this experiment was to determine the bias and precision of CSF pCO2 and pO2 measurements of a contemporary blood gas analyzer and a continuous gas sensor. CSF from human patients was tonometered to known PCO2 and PO2 using standard calibration gases in a bubble tonometer. Following equilibration, a continuous measurement of the partial pressure of gases in the CSF in the tonometer solution was made using a calibrated Paratrend 7 gas sensor (Biomedical Sensors, Bucks., UK) inserted into the equilibration chamber of the tonometer and continuous CSF pH, PCO2 and PO2 measurements recorded at 1-min intervals. After equilibration, a 3-ml CSF sample was aspirated anaerobically from the tonometer and analysed in duplicate using an ABL 620 (Radiometer, Copenhagen) blood gas analyser. The measured pCO2 and pO2 from the ABL 620 blood gas analyser and the Paratrend 7 sensor were paired with the expected values to calculate bias and precision. Small offsets in PCO2 and large offsets in PO2 measurement were seen with the ABL 620 blood gas analyser. This study brings into question clinical decisions based on CSF PO2 measurements. The implications for calibrating CSF sensors against CSF gas measurements are discussed.