Prediction of fluid responsiveness using respiratory variations in left ventricular stroke area by transoesophageal echocardiographic automated border detection in mechanically ventilated patients

Abstract

Background: Left ventricular stroke area by transoesophageal echocardiographic automated border detection has been shown to be strongly correlated to left ventricular stroke volume. Respiratory variations in left ventricular stroke volume or its surrogates are good predictors of fluid responsiveness in mechanically ventilated patients. We hypothesised that respiratory variations in left ventricular stroke area (DeltaSA) can predict fluid responsiveness.

Methods: Eighteen mechanically ventilated patients undergoing coronary artery bypass grafting were studied immediately after induction of anaesthesia. Stroke area was measured on a beat-to-beat basis using transoesophageal echocardiographic automated border detection. Haemodynamic and echocardiographic data were measured at baseline and after volume expansion induced by a passive leg raising manoeuvre. Responders to passive leg raising manoeuvre were defined as patients presenting a more than 15% increase in cardiac output.

Results: Cardiac output increased significantly in response to volume expansion induced by passive leg raising (from 2.16 +/- 0.79 litres per minute to 2.78 +/- 1.08 litres per minute; p < 0.01). DeltaSA decreased significantly in response to volume expansion (from 17% +/- 7% to 8% +/- 6%; p < 0.01). DeltaSA was higher in responders than in non-responders (20% +/- 5% versus 10% +/- 5%; p < 0.01). A cutoff DeltaSA value of 16% allowed fluid responsiveness prediction with a sensitivity of 92% and a specificity of 83%. DeltaSA at baseline was related to the percentage increase in cardiac output in response to volume expansion (r = 0.53, p < 0.01).

Conclusion: DeltaSA by transoesophageal echocardiographic automated border detection is sensitive to changes in preload, can predict fluid responsiveness, and can quantify the effects of volume expansion on cardiac output. It has potential clinical applications.

Figure 1

Transoesophageal echocardiographic transgastric, cross-sectional view of the left ventricle at midpapillary muscle level with automated border detection (ABD). Endocardial border of the left ventricle, including the papillary muscles, was circumscribed manually to define the region of interest (blue line). ABD quantifies the cardiac chamber areas instantaneously by detecting the blood-tissue interface (red line), which results in a continuous, beat-to-beat left ventricular area curve (green line). Left ventricular end-diastolic area (LVEDA) was defined as peak of the left ventricular area during diastole. Left ventricular end-systolic area (LVESA) was defined as minimum left ventricular area during systole. Stroke area (SA) was defined as LVEDA – LVESA over the same cardiac cycle.

Figure 2

Transoesophageal echocardiographic transgastric, cross-sectional views of the left ventricle at midpapillary muscle level with automated border detection at baseline (top panel) and after volume expansion induced by passive leg raising manoeuvre (bottom panel). Left ventricular area curve was displayed with electrocardiogram and respiratory curve. Stroke area (SA) was defined as the difference between the end-diastolic area (LVEDA) and the end-systolic area. Maximal (SA_max) and minimal (SA_min) values of pulse pressure were determined over the same respiratory cycle. Respiratory variations in left ventricular SA (ΔSA) were then calculated using the following formula: ΔSA = [(SA_max - SA_min)/([SA_max + SA_min]/2)] × 100%. Passive leg raising manoeuvre induced a decrease in ΔSA and an increase in LVEDA. Gain was held constant throughout the protocol.

Figure 3

Receiver operating characteristic curves comparing the ability of respiratory variations in left ventricular stroke area (ΔSA), respiratory variations in pulse pressure (ΔPP), left ventricular end-diastolic area index (LVEDAI), and central venous pressure (CVP) at baseline to predict response to volume expansion induced by passive leg raising manoeuvre.

Figure 4

Respiratory variations in stroke area (ΔSA) values at baseline in responders and non-responders to volume expansion induced by passive leg raising manoeuvre. A ΔSA threshold value of 16% allowed discrimination between responders and non-responders with a 93% sensitivity and an 82% specificity.

Figure 5

Relationship between respiratory variations in left ventricular stroke area (ΔSA) (top left panel), respiratory variations in pulse pressure (ΔPP) (top right panel), left ventricular end-diastolic area index (LVEDAI) (bottom left panel), and central venous pressure (CVP) (bottom right panel) at baseline and percentage increase in cardiac output (CO) after volume expansion (VE) induced by passive leg raising manoeuvre.