Using ventilation-induced aortic pressure and flow variation to diagnose preload responsiveness
The recent literature has documented that both arterial pulse pressure  and left ventricular stroke volume variations  induced by positive-pressure ventilation are sensitive and specific markers of preload responsiveness. The greater the degree of flow or pressure variation over the course of the respiratory cycle for a fixed tidal volume, the more likely the subject is to increase their cardiac output in response to a volume challenge and the greater that increase may be. Slama et al.  report in this issue that in a rabbit model of graded hemorrhage measures of aortic flow variation during positive-pressure ventilation cardiac output will increase in response to volume expansion. Furthermore, aortic flow variation was more sensitive than aortic pulse pressure variation in describing this effect further into the hemorrhagic state, presumably because arterial tone also increased. Since measures of descending aortic flow, using an esophageal Doppler probe, can be made easily and continuously without the need for complex echocardiographic techniques, this observation is clinically relevant. These concepts have been previously commented upon in this journal  and elsewhere ; however, the overarching principals of this clinical tool have not been previously described. In retrospect, the above observations may seem intuitively obvious based on simple principles of heart–lung interactions  and initial clinical observations . Still, there are several important caveats and limitations to this approach that need to be considered before the clinician proceeds to monitoring arterial pulse pressure or stroke volume variation during ventilation as a routine assessment of preload responsiveness.
Firstly, being preload responsive does not mean that the subject should be given a volume challenge. Normal subjects under general anesthesia without evidence of cardiovascular insufficiency are also preload responsive and do not need a volume challenge. The present author presumes that most of the people reading this article, if sedated and placed on mechanical ventilation, would also display aortic flow variation and be preload responsive; thus, the presence of positive-pressure-induced changes in aortic flow or arterial pulse pressure does not itself define therapy. Independent documentation of cardiovascular insufficiency needs to be sought before the clinician attempts fluid resuscitation based on these measures.
Secondly, these indices that quantify the variation in aortic flow, stroke volume, and arterial systolic and pulse pressures have routinely been demonstrated to outperform more traditional measures of left ventricular preload, such as pulmonary artery occlusion pressure, right atrial pressure, total thoracic blood volume, right ventricular end-diastolic volume, and left ventricular end-diastolic area [1, 2]. In fact, there appears to be little relation between these traditional measures of ventricular preload and preload-responsiveness. This finding also should hardly be surprising. Ventricular filling pressures poorly reflect ventricular volumes  and measures of absolute ventricular volumes do not define diastolic compliance . Patients with small left ventricles that are also stiff, as may occur with acute cor pulmonale, tamponade, left ventricular hypertrophy and myocardial fibrosis, will show poor volume responsiveness, whereas patients with large left ventricular volumes, as often occurs with congestive heart failure and afterload reduction, may be quite volume responsive; thus, the second concept is that preload does not equal preload responsiveness. The probable reason why traditional measures of preload have displayed such poor predictive value in defining preload responsiveness may be because they are indirect measures of the wrong parameter: left ventricular wall stress.
Thirdly, all the studies reported to date have used positive-pressure ventilation as a cyclic venous return forcing function. They all presume, though not all studies state this clearly, that for a subject to be preload responsive their intrathoracic cardiac performance must be globally responsive to dynamic changes in venous flow. Once again, this is intuitively obvious. If a patient has either severe right or left ventricular failure, then the small changes in venous pressure induced by the pressure forcing function of mechanical ventilation will not alter preload enough to cause changes in output; however, what is often forgotten is that for these aortic pressure and flow measures to detect any variations at all, the input forcing function must be great enough to alter the pressure gradients for right and left ventricular filling . Although controversy exists as to the extent to which positive-pressure ventilation selectively alters either pulmonary venous flow  or right ventricular filling , all agree that a cyclic change must occur. Accordingly, if the increase in lung volume with each tidal breath is either not great enough to induce changes in pulmonary venous flow , or if the positive-pressure breath is associated with spontaneous inspiratory efforts that minimize the changes in venous return , then the cyclic perturbations to cardiac filling may not be great enough to induce the cyclic variations in left ventricular filling needed to identify preload responsiveness. Furthermore, one can take a preload-unresponsive patient whose aortic flow variation is <10% and make their flow variation >20% by increasing tidal volume from 8 to 20 ml/kg [7, 8]; thus, the third concept is that the means by which cyclic changes in lung volume and intrathoracic pressure are induced and the consistency of their change over time is a central aspect of the ground-state requirements in the use of cyclic variations in aortic flow and arterial pressure to identify preload responsiveness.
Fourthly, although the primary determinant of arterial pulse pressure variation over a single breath is left ventricular stroke volume variation because changes in aortic impedance and arterial tone cannot change that rapidly (12 of 15), over time this limitation no longer applies. As arterial tone decreases, for example, then for the same aortic flow and stroke volume both mean arterial pressure and pulse pressure will be less. Similarly, if arterial tone were to increase, as often occurs in response to hemorrhage, then for the same aortic flow arterial pressure will be greater. This is the probable reason why Slama et al.  observed that flow variation became more sensitive than pulse pressure variation as hemorrhage progressed. Finally, since aortic cross-sectional diameter is not fixed but varies during left ventricular ejection and with changes in aortic input impedance, aortic peak velocity changes do not equate with left ventricular stroke volume or descending aortic flow changes. Although most studies have shown that flow velocity variation thresholds of >15% predict preload responsiveness, and that this threshold is very close to the 13% threshold described for pulse pressure variation, these two measures have never been compared with each other in the same study. Accordingly, the limits of flow or pressure variation needed to identify preload responsiveness will be different if one uses peak aortic flow velocity, mean aortic flow, left ventricular stroke volume, or arterial pulse pressure measures to define preload responsiveness, especially if arterial tone is varying.
Fifthly, like many clinical indices of cardiovascular state, the ventilation-induced changes in left ventricular output and pulse pressure have been analyzed in retrospective studies. No study yet published has prospectively predicted which patients would be preload responsive based on these measures and then shown that their predictions were correct. This procedural limitation should be resolved shortly, as prospective clinical trials of these indices to predict preload responsiveness have recently been completed.
Finally, no monitoring tool, no matter how accurate, by itself has improved patient outcome. It is only with the thoughtful and titrated use of therapies known to improve outcome, but requiring information derived from such monitoring tools, will improved patient outcome realized by the use of these tools. We know that occult tissue hypoperfusion exists in critically ill patients and that appropriate and timely resuscitation improves outcome in all critically ill patients, whereas inappropriate and/or delayed resuscitation increases mortality. Accordingly, this new tool needs to be used in concert with proven therapeutic approaches that utilize its diagnostic power correctly. Thus, we are left with a new and robust, powerful, and clinically relevant diagnostic tool whose thoughtful application in the clinic is gaining acceptance and whose influence on future clinical trials using resuscitation therapies will need to be considered. Clinicians are urged to consider the above physiological limitations when applying these deceivingly simple measures at the bedside to define resuscitation decisions.
- 1.Michard F, Boussat S, Chemla D, Anguel N, Mercat A, Lecarpentier Y, Richard C, Pinsky MR, Teboul JL (2000) Relation between respiratory changes in arterial pulse pressure and fluid responsiveness in septic patients with acute circulatory failure. Am J Respir Crit Care Med 162:134–138PubMedGoogle Scholar
- 3.Slama M, Masson H, Teboul JL, Arnould ML, Nait-Kaoudjt R, Colas B, Peltier M, Tribouilloy C, Susic D, Frohlich E, Andrejak M (2004) Monitoring of respiratory variations of aortic flow velocity using oesophageal Doppler. Intensive Care Med (http://dx.doi.org/10.1007/s00134-004-2190-z)
- 5.Pinsky MR (2003) Probing the limits of arterial pulse contour analysis to predict preload-responsiveness. Anes Analg 96:1245–1247Google Scholar
- 8.Hansen RM, Viquerat CE, Matthay MA, Wiener-Kronish JP, DeMarco T, Bahtia S, Marks JD, Botvinick EH, Chatterjee K (1986) Poor correlation between pulmonary arterial wedge pressure and left ventricular end-diastolic volume after coronary artery bypass graft surgery. Anesthesiology 64:764–770PubMedGoogle Scholar
- 11.Vieillard-Baron A, Loubieres Y, Schmitt JM, Page B, Dubourg O, Jardin F (1999) Cyclic changes in right ventricular output impedance during mechanical ventilation. J Appl Physiol 87:1644–1650Google Scholar