Objective: To evaluate the routine clinical performance of a new intra-arterial fiberoptic blood gas sensor that provides continuous PO2, PCO2, and pH monitoring.
Design: Criterion standard study under routine clinical conditions.
Setting: Intensive care unit (ICU) in a university hospital.
Patients: Twenty-two sensors were tested in 13 patients with acute respiratory failure, including two patients receiving veno-venous extracorporeal lung assist. Patient selection was based on the necessity of frequent blood gas monitoring.
Measurements: Sensor-deprived PO2, PCO2, and pH values were compared with values obtained using two different conventional laboratory blood gas analyzers located in the ICU. The median study period was 72 hrs per sensor (range 8 to 170 hrs). The quality of blood pressure readings with the sensor introduced through the arterial catheter was assessed by a grading system.
Results: Mean differences between sensor-derived values and the average values of the two conventional blood gas analyzers were as follows: PO2 -2.4 +/- 6.5 (SD) torr (-0.3 +/- 0.9 kPa), PCO2 -2.9 +/- 3.9 torr (-0.4 +/- 0.5 kPa), and pH -0.04 +/- 0.03. Correlation coefficients were 0.99 (PO2), 0.94 (PCO2), and 0.89 (pH), respectively. The agreement between the two methods for PO2 measurement was better for the clinically important range of values (PO2 < 150 torr [< 20 kPa]) than for all measured PO2 values (range 30 to 522 torr [4 to 69.6 kPa]). Blood withdrawal and pressure readings were not adversely affected by the sensor. No side effects due to the insertion of the sensor were observed.
Conclusions: The degree of agreement of intra-arterial blood gas sensor values with conventional blood gas analysis is within an acceptable range for routine clinical purposes. Acute changes in measured values are detected reliably. Continuous intra-arterial blood gas analysis can add substantially to the safety of patients with acute respiratory failure and can reduce blood sampling requirements for blood gas analysis.