Accuracy assessment for three fiberoptic pulmonary artery catheters for$S\bar vO_2$ monitoring

Abstract

Objective

To compare values of$S\bar vO_2$ obtained by reflectance spectrophotometry continuous monitoring with those obtained from blood samples measurements by transmission spectrophotometry (Co-Oximetry).

Design

Values of$S\bar vO_2$ recorded by three pulmonary artery catheters for continuous monitoring, SAT 1, SAT 2 and Oximetrix 3 (OX 3), were compared in a prospective manner to those measured on blood samples by a Co-Oximeter, using the statistical analyis of Bland and Altman.

Setting

Adult intensive care unit in an University Hospital.

Patients

37 patients admitted for acute respiratory failure and/or shock who required hemodynamic monitoring.

Main results

The bias (average under- or overestimation) was small for all comparative measurements: +1.3, −0.2 and +1.0 sat% for SAT 1, SAT 2 and OX 3, respectively. However, limits of agreement were only acceptable for SAT 2 (−8.3 to +7.9 sat%) and OX 3 (−6.7 to +8.6 sat%), but not for SAT 1 (−23.3 to +25.9 sat%). No significant drift during 24 h was found with the three catheters. However, in vitro calibration was only found acceptable for SAT 2 and OX 3. The results were not influenced by the numbers of wavelengths of the device (2 for SAT 1 and SAT 2, and 3 for OX 3) nor did they correlate with any of the hemodynamic and biochemical variables tested.

Conclusion

For usual monitoring in the ICU, SAT 2 and OX 3, gave$S\bar vO_2$ values which are in acceptable agreement with$S\bar vO_2$ measured on blood samples by Co-Oximetry.