POCUS: Feissel


Respiratory Changes in Aortic Blood Velocity as an Indicator of Fluid Responsiveness in Ventilated Patients With Septic Shock

Feissel. Chest 2001; 119:867-873

Clinical Question

  • In patients with septic shock who are mechanically ventilated does indexed LV end-diastolic area or respiratory changes in aortic blood velocity predict the haemodynamic effects of volume expansion?


  • Prospective observational diagnostic study
  • Presumed convenience sampling (not stated)
  • Blinding not stated
  • Statistics
    • No sample size calculation performed
    • Nonparametric Wilcoxon rank sum test used to asses the effects of volume expansion on haemodynamic parameters
    • Nonparametric Mann-Whitney test used to compare haemodynamic parameters prior to volume expansion in responder and non-responder patients
    • Linear correlations tested using Spearman rank method


  • 2 intensive care units in France (1 university and 1 non-university)
  • Dates of data collection not stated


  • Inclusion criteria:
  • Exclusion criteria:
    • Arrhythmias, PaO2/FiO2 ratio <100mmHg, contraindication to transoesophageal echocardiography (TOE), aortic valvulopathy, LV systolic dysfunction (fractional area of contraction <30%), doppler LV output recordings using transgastric view could not be obtained
  • n=19
    • All received inotropic and vasopressor drugs
    • TOE performed between 12-72 hours after diagnosis of septic shock
    • All had already received colloids (3±2 L) and/or crystalloids (1±1 L)

Tests of Interest

  • LV end-diastolic area index (EDAI)
    • TOE used to obtain a transgastric, short-axis, cross-sectional view of the LV at the mid-papillary muscle level
    • End-diastole defined as the frame corresponding to the largest LV cross-sectional area immediately after the R-wave peak on the ECG
    • The anterolateral + posteromedial papillary muscles were included within the ventricular area
    • LV areas divided by body surface area (BSA) to obtain EDAI
    • Used mean of 5 measurements performed at the end of expiration
  • Respiratory changes in aortic blood velocity
    • Aortic blood flow measured over single respiratory cycle to allow measurement of peak velocity (Vpeak) and determination of maximum Vpeak values (Vpeakmax) and minimum Vpeak values (Vpeakmin)
    • Change in peak velocity (ΔVpeak %) = [100 * [Vpeakmax – Vpeakmin]] / [[Vpeakmax + Vpeakmin]/2]
    • Mean value of 5 determination used for statistical analysis

Gold standard investigation

  • Cardiac output measurement
    • Using TOE, LV outflow tract and ascending aorta imaged when parallel to the ultrasound beam
    • Aortic blood flow measured by pulse wave-doppler at level of aortic valve
    • Diameter of aortic orifice measured at the insertion of the aortic cusp
    • Aortic valve area = π * (aortic diameter/2)
    • Stroke volume = aortic valve area * the velocity time integral of aortic blood flow
    • Cardiac output = stroke volume * heart rate
    • Stroke volume and Cardiac output were divided by the BSA to obtain the stroke volume index and cardiac index
    • If cardiac index increased by ≥15% with fluid challenge classified as responder. If <15% classified as non-responder.
    • Mean of 5 measurements performed at end of expiration used for statistical analysis
All patients
  • Mechanically ventilated in volume-controlled mode with tidal volumes of 8-10ml/kg
  • Paralysed if significant inspiratory effort detected on visual inspection of airway pressure curve
  • Measurements performed prior to volume expansion, then repeated immediately post volume expansion (8ml/kg of 6% hydroxyethylstarch over 30 minutes)
  • Ventilatory settings and dosages of inotropic and vasopressor drugs held constant
Intra-observer and inter-observer variability
  • Determined in 10 randomised patients
  • Mean % error for measurements of EDAI and cardiac output were 6±3% and 8±4% by the same observer and 7±4% and 9±5% between 2 different observers


  • Primary outcome: Not defined
  • EDAI
    • Before volume expansion
      • No significant difference between responders vs. non-responders
        • 9.7±3.7 vs. 9.7±2.4 cm2/m2
    • Volume expansion led to a significant increase in EDAI
      • 9.7±3.1 cm2/m2 to 11.8±3.9 cm2/m2
      • The increase in EDAI significantly correlated with the % increase in Cardiac index (r2 = 0.49, P<0.01) such that the greater the increase in EDAI, the higher the increase in cardiac index in response to volume expansion
  • ΔVpeak
    • Before volume expansion
      • Significantly higher in responder vs. non-responder
        • 20±6% vs. 10±3%, P <0.001
      • Using threshold for ΔVpeak of 12%, gave a sensitivity of 100% (95% C.I. 69-100) and specificity of 89% (95% C.I. 52-100) for discriminating between responder and non-responder
        • >12% in all 10 out of 10 responders
        • ≤ 12% in 8 out of 9 non-responders
    • Volume expansion led to a significant decrease in ΔVpeak
      • 15±7% to 8±4%, P<0.001
    • A positive linear correlation found between ΔVpeak before volume expansion and volume expansion-induced changes in cardiac index (r2 = 0.83, P <0.001) i.e. the higher the ΔVpeak before volume expansion, the greater the increase in cardiac index in response to fluid challenge

aortic blood velocity

Authors’ Conclusions

  • In patients with septic shock analysis of respiratory changes in aortic blood velocity is an accurate method for assessing fluid responsiveness


  • Clearly defined test of interest and gold standard investigations
  • Tested for inter and intra-observer variability
  • Controlled known confounding variables


  • No sample size calculation
  • Presumed convenience sample
  • Did not state who performed TOE and what training they had received
  • Aortic blood velocity used in test of interest and for gold standard investigation
  • Small sample size led to wide confidence intervals for sensitivity and specificity

The Bottom Line

  • In this small study, patients with septic shock who were mechanically ventilated with tidal volumes of 8-10ml/kg, the respiratory changes in peak aortic blood velocity predicted which patients would have a significant increase in cardiac index with a fluid challenge. The increase in LV end-diastolic area index with a fluid challenge significantly correlated with the increase in cardiac index; where as the baseline LV end-diastolic area index did not predict if a patient would be fluid responsive.

External Links


Summary author: David Slessor
Summary date: 7th December 2015
Peer-review editor: Duncan Chambler

Leave a Reply