Background: Systemic arterial respiratory variation has been shown to be a reliable predictor of changes in cardiac output after fluid administration. Arterial respiratory variation is often estimated from visual examination of the arterial waveform tracing. Our goal in this study was to assess the ability of anesthesia providers to visually estimate systolic pressure variation (SPV) as a percentage of systolic blood pressure (SPV).
Methods: Fifty anesthesia providers were recruited and asked to visually examine 10 recorded arterial waveform tracings (played in real time), to estimate SPV, and to state whether or not a fluid bolus was indicated. After completion of the examination, the participants were shown the original tracings, the true value for SPV, and their estimate. The percentage of incorrect physician decisions to administer or not administer additional fluid was analyzed using a binomial proportion confidence interval. Clinical utility was also assessed using clinical significance analysis. Limits of agreement were analyzed using the nonparametric approach recommended by Bland and Altman.
Results: The mean bias was +1.2%. The nonparametric limits of agreement were -5.1 and 7.5%, and contained 82% of values. Actual physician decisions were incorrect 4.4% of the time (95% confidence interval [CI] 2.8% to 6.6%). On the basis of the clinical significance analysis, only 1% of treatments based on the visual estimation would have been incorrect.
Conclusion: Visual estimates of respiratory variation are within clinically reasonable limits 82% of the time and lead to erroneous management decisions in 4.4% of measurements.