Stroke Volume Variation as an Indicator of Fluid Responsiveness

To the Editor: We read with interest the recent article by Wiesenack et al. (1) suggesting, in contrast to other recent studies (2–4), that stroke volume variation (SVV) derived from pulse contour analysis could not serve as an indicator of fluid responsiveness in cardiac surgical patients. This conclusion was based on the lack of correlation between SVV at baseline and the percentage changes in stroke volume index (ΔSVI) after volume loading. The validity of this conclusion relies thus on the assumption that changes in preload were the only determinants of ΔSVI in all patients. However, achieving fluid challenge and measurements in an otherwise hemodynamic steady-state in every patient may be uncertain after induction of anesthesia and introduction of isoflurane. The large changes in SVI and in systemic vascular resistance observed following fluid loading were relatively unexpected in patients scheduled for elective surgery and may suggest that such steady-state was not obtained in all patients. In addition, it was not clearly stated in the article how SVI was measured. This is important because the validity of the pulse contour technique for the accurate quantification of changes in SVI in individual patients remains poorly established, especially when arterial compliance may have changed. In fact, we believe that much data in this study actually suggested that SVV is useful to the assessment of preload, and probably, preload responsiveness. The high correlation between SVV at baseline and its changes after volume replacement (ΔSVV) is apparently consistent with this hypothesis but was not reported in other studies on fluid responsiveness. We thus retrospectively calculated these correlations from a study where the ‘delta down‘ component of the systolic pressure variation was shown to be a reliable indicator of fluid responsiveness in patients with septic shock (5). Interestingly, the correlation was highly significant for ‘delta down‘ (r = 0.85; P < 0.001), but not for either PCWP (r = 0.01; P = 0.98) or the left ventricular end-diastolic area (r = 0.13; P = 0.66). It is also notable in the Wiesenack et al. (1) study that the change in mean SVV associated with volume replacement was much larger than that observed with either CVP or PCWP. Thus, at the ‘group level,‘ results suggest a close relationship between the increase in intravascular volume (due to volume loading), the large ΔSVV, and the large ΔSVI. Moreover, because profound functional hypovolemia would be very unlikely at this stage of surgery, it can be assumed that a second fluid challenge would have been associated with small, if any, ΔSVI. Thus, as compared with baseline measurements, the smaller values of SVV observed at the end of the study would have been associated with smaller ΔSVI, reinforcing the hypothesis of a preload responsiveness assessment by SVV. Nicolas Bouteau, MD Benoît Tavernier, MD, PhD