Assessment of stroke volume variation

Introduction

The specific interactions of the lungs and the cardiovascular system under mechanical ventilation cause cyclic variations of left ventricular stroke volume (SVV). Real-time measurement of SVV using arterial pulse contour analysis is useful to predict volume responsiveness and to monitor volume therapy in mechanically ventilated patients. In this study, SVV was evaluated using simultaneous recordings with the LiDCO plus system (LiDCO Ltd. Cambridge, UK) and the Vigileo (Edwards Lifesciences, Irvine, CA, USA) cardiac output monitor.

Methods

At baseline the LiDCO pulse contour cardiac output was calibrated using the bolus thermodilution technique. The setup and zeroing procedure of the Vigileo cardiac output monitor was carried out in compliance with product specifications. Nine postoperative cardiac surgical patients were included. All measurements were carried out during standard clinical care. In stable clinical conditions, changes in SVV were forced by changes in tidal volume (Vt), the level of PEEP and leg-raising procedures. In all patients a total of 134 data pairs were evaluated using linear regression and Bland-Altman statistics.

Results

The mean SVV measured with LiDCO was 10.1 ± 5.3% (SD), and that measured with the Vigileo was 11.7 ± 5.9%. The correlation coefficient regarding SVV measured with the two different devices was R² = 0.678 (slope 0.944, SE 0.057) (Fig. 1). The computed bias was significantly different from 0 (-1.70 ± 3.3%, P < 0.001) (95% CI -2.28 to -1.13). The upper and lower limits of agreement were 5.01 and -8.42% (Fig. 2). The calculated agreement between negative and positive changes in SVV using the LiDCO plus system and the Vigileo cardiac output monitor was correct in 80.6%.

Figure 1

Figure 2

Conclusion

The agreement in SVV measured with the LiDCO and Vigileo cardiac output computer is acceptable. We found a slight, but significant, difference in SVV between the LiDCO and Vigileo cardiac output monitoring systems. For adequate interpretation of differences of SVV, more disclosure of the used mathematical models is essential.