Reliability and accuracy of continuous blood gas monitoring

The current method for blood gas measurement consists of intermittent blood sampling and analysis in a remote laboratory. The drawbacks of this method include the need lor heparinized syringes, wasted blood during sampling and testing, and wasted time due to sample transportation and result reporting. Continuous (in vivo) blood gas (CBG) measurement has the potential to eliminate all these drawbacks. However, its reliability and accuracy need clinical validation. We performed this prospective study to determine whether or not CBG monitoring was reliable and accurate. We hypothesized that the CBG technology was less reliable and less accurate than the laboratory blood gas (LBG) method. Between September and November of 1996, new admissions to the SICU requiring mechanical ventilation were studied. Upon placing an arterial catheter, a CBG sensor (Paratrend-7, Biomedical Sensors, Malvern, Pennsylvania and High Wycombe, England) was calibrated, inserted and connected to a Paratrend blood gas monitor. The patient's subsequent management was guided by the CBG data. Whenever the CBG values were used for therapy, they were also recorded on the data collection sheet. A simultaneous LBG value was also obtained and recorded. Additional data collection included the duration of CBG sensor use, the time saved by CBG monitoring, and problems with the arterial line, CBG sensor, and the laboratory. The results were analysed by descriptive statistics and Student t test. A P value of < 0.05 was considered to be significant.

Using 17 CBG sensors in 13 patients, 1478 h of CBG monitoring was carried out. Ninety-six pairs of CBG and LBG values were available for comparison. The mean pH, PaCO₂ (Torr) and PaO₂ (Torr) by CBG and LBG measurements were 7.40, 43.1 and 122.6 and 7.41, 41.4 and 120.5, respectively and the differences between them were significant. CBG sensors were placed in the femoral artery in eight patients and four of them required a second sensor. In five patients, CBG sensors were placed through the radial artery. Six of 12 femoral and four of five radial artery sensors had problems with either pH or PaO₂ measurements. The mean durations for the femoral and the radial arterial sensors were 100 + 85(SD) and 56 + 103(SD) h, respectively. CBG monitoring saved 35 min per therapy decision or 50.8 h during the study period.

Conclusion

Continuously measured blood gases are as accurate as the laboratory measured blood gases. They shorten therapeutic decision time. Their reliability can be improved by using the femoral artery for sensor placement.