at SciVerse ScienceDirect

Trends in Anaesthesia and Critical Care 2 (2012) 115e122

Contents lists available

Trends in Anaesthesia and Critical Care

journal homepage: www.elsevier .com/locate/ tacc

REVIEW

To give or not to give fluid challenges!

Kanishka Indraratna*

Sri Jayewardenepura General Hospital, Sri Lanka

Keywords:Static parametersDynamic parametersFunctional haemodynamicsStroke volume variationTransoesophageal echocardiography

* Tel.:þ94 777578144; fax:þ94 112778213.E-mail address: kindraratna@yahoo.co.uk.

2210-8440/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.tacc.2012.02.008

s u m m a r y

Fluid challenges are used to improve cardiac output and oxygen delivery. This is done in the presence ofhypovolaemia. Hypovolaemia is generally diagnosed on static haemodynamic parameters, such asCentral Venous Pressure, Pulmonary Capillary Wedge Pressure etc. Only about half of patients admin-istered fluid in this manner, have benefitted. Recently interest has been directed at functional haemo-dynamic parameters. These are based on the relationship between mechanical ventilation and venousreturn due to the cyclical changes in intra thoracic pressure. These cause variation in the stroke volume,systolic blood pressure and pulse pressure during the respiratory cycle. There are factors which can affectthe accuracy and the interpretation of these parameters. Spontaneous respiration, the tidal volumes usedto ventilate, PEEP, lung and chest compliance, heart rhythm, right ventricular function are these factors.Passive leg raising can be used in the presence of these or in doubt.

This article attempts to set out, how to determine whether a fluid challenge will improve the cardiacoutput, and also to identify the problems in arriving at that decision.

� 2012 Elsevier Ltd. All rights reserved.

1. Haemodynamics and functional haemodynamics

Haemodynamics is the physiology concerned with the move-ment of blood and the forces and pressures associated with thatcirculation. Haemodynamics measures oxygen delivery to tissues.For this, static parameters such as Central Venous Pressure,Pulmonary Capillary Wedge Pressure, Systemic Vascular Resis-tance, Cardiac Output are measured. They do not tell us, however,whether increasing the volume status will improve the cardiacoutput and the oxygen delivery or compromise the chances ofsurvival. Only about 50% of patients responded to fluid adminis-tration based on these parameters.1 Functional Haemodynamicsattempts to address this dilemma, with the use of dynamicparameters.2 Therefore it is necessary to measure the haemody-namics to know the oxygen delivery and to monitor the functionalhaemodynamics to assess whether the oxygen delivery can beimproved.3

1.1. Cardiac output

DO2 ¼ CI(Hb*1.34*SaO2 þ .003PaO2)

All rights reserved.

DO2-Oxygen delivery to tissues, CI-Cardiac Index, Hb-Haemoglobin concentration, SaO2-Arterial oxygen saturation,PaO2-Arterial oxygen tension.

The oxygen delivery depends on the cardiac output, amount ofhaemoglobin, the saturation and the dissolved oxygen.

The cardiac output is affected by,

1. The pre load, this is the volume status of the patient.2. The myocardial contractility3. The after load or the resistance against which the left ventricle

pumps, which is measured by the systemic vascular resistance.

Of the above the Pre load or the fluid volume would be focusedupon, as attempting to improve the fluid status, and therebyincreasing the cardiac index and thus the oxygen delivery to tissuesis very often the first line of management when it is required toimprove the cardiac output.

The haemodynamic parameters available to measure the fluidstatus of a patient are,

Static parameters-A parameter measured under a single loadingcondition

� Pulmonary capillary wedge pressure� Central venous pressure� Right ventricular end diastolic volume� Left ventricular volume� Global end diastolic volume

Stroke

volume

Venous return

AB

1

2

CD

34

Fig. 1. Frank Starling curve.

K. Indraratna / Trends in Anaesthesia and Critical Care 2 (2012) 115e122116

However these are all measurements taken under single loadingconditions.

They, therefore do not indicate whether the patient is normovolaemic, hypovolaemic or over loaded. They are static parameters.These static indices of preload have a low predictive ability forhypovolaemia.2,4 It has been shown that the CVP does not accu-rately indicate the preload and also does not predict fluidresponsiveness.2,10,16

An inappropriate fluid administration, where the heart is unableto increase the cardiac output with a fluid bolus, can lead to oedemain both tissues and lungs causing further hypoxia. Therefore itbecomes important to identify those patients who will havea beneficial effect with a fluid administration.5 For this purposefunctional haemodynamics are required.

Functional haemodynamics depend on the heartelunginteraction.

Dynamic parameters - Heartelung interaction on preload indices

Variations in� Stroke volume� Systolic pressure� Pulse pressure

With mechanical ventilation

With mechanical ventilation the pre load will change with eachcycle and therefore, there will be cyclical variations in strokevolume, systolic pressure and pulse pressure.18

With mechanical ventilation in a patient with normal leftventricular function, during inspiration, the intra thoracic pressurerises. The venous return is therefore decreased, and cardiac outputwill fall. In expiration, the intra thoracic pressure falls and thevenous return will increase, thereby increasing the cardiac output.6

Because of this heartelung interaction during mechanical ventila-tion, it is possible without administering extraneous fluid to obtaintwo points on the Frank-Starling curve. One, when the preload isdecreased and the other when it is increased. The changes in thecardiac output will manifest as variations in Stroke Volume, SystolicBlood Pressure, and Pulse Pressure with inspiration and expiration.If these variations are present it means that the patient is on thesteep part of the curve and if there is no variation on the plateausection. With a fluid challenge the cardiac output can increase onlyif the patient is on the steep part of the Frank-Starling curve.

As shown in the diagram (Fig. 1), in a person with normalcardiac function, who will be on the rising steep part of the curve,when there is a reduction in the venous return from A to B, thestroke volume will also fall from 1 to 2. This will manifest as a fallin stroke volume, systolic blood pressure or pulse pressure. Ina person with impaired left ventricular function, who will there-fore be on the plateau part of the Frank Starling curve, a corre-sponding change in venous return from C to D, will not showa reduction in stroke volume. There is hardly any differencebetween 3 and 4 in the figure. Therefore a patient with ventriculardysfunction, who will be on the plateau part of the curve will notshow any changes in stroke volume, systolic blood pressure orpulse pressure.

The CVPmeasurement on the other hand, on such a patient, willjust indicate the fluid volume of the patient, but not whether it isadequate or not. It will also not predict, as it is just one measure-ment under static conditions, how the patient will respond toa fluid challenge.10 Marik et al. in a literature survey in 2008,have said the CVP does not indicate the blood volume or the fluidresponsiveness.8 Use of CVP measurements to assess whether ornot a patient’s cardiac output will increase significantly inresponse to an infusion of intravenous fluid cannot therefore berecommended.9,10,19

Stroke volume variation can be easily observed with trans-oesophageal echocardiography.5,7,11

The picture shows the continuous wave, with trans-oesophageal echo, at the aortic valve. The variation is easilyobserved. The peak velocity at the aortic valve is an accurateindicator of predicting haemodynamic effects of volume expan-sion.17 This is because the stroke volume is the product of theaortic valve area and the velocity time integral at the aortic valve.The aortic valve area can be assumed to remain constant duringthe respiratory cycle.17 The peak velocity at the aortic valve iseasily observed with transoesophageal echocardiography asshown in Fig. 2. There is a variation in the peak velocity (strokevolume) in Fig. 2.

During mechanical ventilation, at inspiration, the venousreturn falls thereby making the stroke volume less, and duringexpiration the venous return increases thus making the strokevolume more. This patient is therefore in the steep part of theFrank-Starling curve and will be able to increase the cardiac outputfollowing a fluid challenge. These patients are termedresponders.2

This is a picture of the continuous wave transoesophageal waveform at the aortic valve of another patient (Fig. 3). As explainedbefore, with mechanical ventilation, there is cyclical change in thevenous return, however, there is no corresponding variation in thestroke volume in this patient. This means that this patient is on theflat part of the Frank Starling curve and is unable to increase thecardiac output in response to a fluid challenge. This is a non-responder.

The above two figures show the peak velocity, at the aorticvalve, with mechanical variation. As discussed before, this is anindication of the stroke volume. The arterial blood pressure, whenmonitored with an arterial cannula and wave form will also show,whether the systolic blood pressure varies with mechanical venti-lation. Responders and non-responders can be identified in thismanner too.

As explained above heart-lung interactions enable two points tobe determined on the Frank Starling curve, with the venous returnat two different levels without the administration of extraneousfluid.

However this same heart-lung interaction is subject to inter-ference by other factors. These will cause difficulties in interpre-tation and may have an effect on the accuracy of the estimationsand assessments.

Respiratory system issues

1. Tidal volume2. PEEP

Fig. 2. Continuous wave at the aortic valve with transoesophageal echocardiography-stroke volume variation. a, c eshow decreased velocity. beshows increased velocity.

K. Indraratna / Trends in Anaesthesia and Critical Care 2 (2012) 115e122 117

3. Lung compliance4. Chest compliance

How the respiratory system behaves will have an impact onthe stroke volume variation and other functional haemodynamicparameters. These are tidal volumes, the amount of PEEP and thelung and chest compliance.

Tidal volumes - The larger the tidal volumes deliveredby mechanical ventilation, the greater the rise in inspiratory

Fig. 3. Transoesophageal echocardiogr

pressure and therefore the greater the fall in venous return.However small tidal volumes may not have an influence on theintra thoracic pressure, and therefore will not have an impact onvenous return.12 A patient who did not appear to be a responderwith small tidal volumes can appear to be a responder because ofthe use of large tidal volumes. In ARDS, where low tidal volumes areused for ventilation, it may still have an impact because the poorcompliance of the lung will cause the generation of large pressureseven with small tidal volumes.14

aphy-No stroke volume variation.

Fig. 4. Stroke volume with transoesophageal echocardiography-tidal volume 6 ml/kg.

K. Indraratna / Trends in Anaesthesia and Critical Care 2 (2012) 115e122118

Also hyperinflation can by constricting the alveolar vessels canincrease the right ventricular afterload and reduce its output,13

therefore affecting left sided filling and output. Therefore strokevolume can vary.

It has been found that stroke volume variation is best assessed attidal volumes of 8 ml/kg14,15.

Fig. 4 shows the peak velocity at the aortic valve when thepatient was ventilated with a tidal volume of 6 ml/kg. On this, thereis no stroke volume variation, and the patient appears as a non-responder to fluid challenges.

In the same patient, the tidal volume was increased to 8 ml/kg(Fig. 5). The stroke volume variation can be seen clearly now in thefigure. The patient now appears as a responder. It is therefore

Fig. 5. Stroke volume with transoesophageal

important to take into consideration, the tidal volumewhich is beingused to ventilate the patient, before interpretation and conclusions.

PEEP - Positive End Expiratory Pressure will increase the intra-thoracic pressure and thereby reduce venous return, causinga decrease in both right ventricular and left ventricular strokevolumes. This effect will however be present both in inspirationand expiration. This will therefore give an exaggerated effect on thestroke volume variation. A benefit of this effect is that, if the strokevolume variation increases with PEEP, it would predict that thecardiac index would be reduced by PEEP20e22 and is fluidresponsive23,24.

Fig. 6 shows the peak velocity at the aortic valve with trans-oesophageal echocardiography on a patient on mechanical

echocardiography-tidal volume 8 ml/kg.

Fig. 6. Stroke volume variation with transoesophageal echocardiography without PEEP.

K. Indraratna / Trends in Anaesthesia and Critical Care 2 (2012) 115e122 119

ventilation without PEEP. There is no variation in the peak velocity(Stroke volume as discussed before).

In the same patient a PEEP of 5 was applied (Fig. 7). Nowa variation in the peak velocity is seen. The patient now appears asfluid responsive.

A patient who drops his blood pressure or cardiac output, withthe application of PEEP, can be given fluid, if his stroke volumeshows variation, as he is responsive.23,24 If the patient is inventricular failure, PEEP will not induce variation in strokevolume.22

Lung compliance - In situations such as ARDS the lung compli-ance is very low, that the generated pressure is not transmitted tothe intra thoracic pressure. In addition small tidal volumes are usedto ventilate patients with ARDS. Therefore an actual fluid respon-siveness may be masked. However high intra alveolar pressures are

Fig. 7. Stroke volume variation with transoes

generated in ARDS even with small tidal volumes, because of thehigh PEEP used and the low lung compliance. If this pressure istransmitted to the intra thoracic pressure the venous return isreduced during inspiration and increased during expiration.25

Generally the cyclical changes in intra thoracic pressure are suffi-ciently high to predict fluid responsiveness in ARDS.26 Therefore ifstroke volume variation is present in a patient with ARDS, while onmechanical ventilation, it possibly is an accurate predictor of fluidresponsiveness.27

Chest compliance - The chest compliance will contribute to theamount of intra thoracic pressure generated during mechanicalventilation. If the compliance is very high, the pressure generatedwill be low, and therefore may not have an impact on venousreturn. Stroke volume variation will therefore decrease with thechest open and increase with closure of chest.28 Therefore if stroke

ophageal echocardiography with PEEP 5.

Fig. 8. Transoesophageal echocardiography of constrictive pericarditis-right ventricle constricted.

K. Indraratna / Trends in Anaesthesia and Critical Care 2 (2012) 115e122120

volume variation is not seen, when the chest compliance is veryhigh i e-open chest during cardiac surgery, it does not necessarilymean that the patient is a non-responder. Reuter DA et al. foundthat Stroke volume variation can be accurately used during openchest conditions.29

1.2. Cardio vascular issues

1. Heart rhythm2. Right ventricular dysfunction3. Afterload4. Arterial compliance

Heart rate and rhythm - The diastolic times will differ from beatto beat when arrhythmias are present. This will affect the effect ofmechanical insufflations on stroke volume. Sinus rhythm should bepresent during assessment.30

Right ventricular dysfunction - When there is right ventriculardysfunction due to myocardial damage or severe pulmonaryhypertension the right ventricular output will be decreased. Thiswill therefore affect left ventricular filling and thus its strokevolume.31 During mechanical ventilation this will manifest asa stroke volume variation. However giving a fluid challenge in thissituation can be fraught with danger as the right ventricle may notbe able to handle any more volume and because of interventriculardependence it may affect and compromise left ventricular functionas well.

The two transoesphageal echocardiography pictures shows(Figs. 8 and 9), on top a patient with severe constrictive pericarditis,with compromise of the right ventricular cavity. The second pictureshows the stroke volume of the same patient. The stroke volumevariation is obvious.However givingfluid to this patient is extremelydangerous because of the small right ventricular cavity size.

In the presence of right ventricular dysfunction, stroke volumevariation may be seen, because left sided filling is affected. Thisdoes not mean however, that the patient will respond to a fluidchallenge.

Afterload - An increased afterload can reduce the left ventricularstroke volume. However it has been shown that variations in theafterload does not affect the accuracy of stroke volume variation.32

Arterial compliance - Variations in systolic pressure and pulsepressure may be subject to changes and inaccuracies because offluctuations in arterial compliance. These are difficult to predict andmay vary even in the same patient.33

1.3. Afterload reserve

In addition to improving cardiac output, assessment of how thevaso motor tone changes in response to changes in volume andpressure is also important to obtain the best results. This can bedone by assessing the ratio of pulse pressure variation to strokevolume variation.33 If the stroke volume variation is greater thanpulse pressure variation, there is vasodilatation and either fluid ora vaso constrictor can be used.30

1.4. Limitations

Stroke volume variation is easy to assess with the availabilityof transoesophageal echocardiography and oesophageal doppler.Systolic pressure variation, Pulse pressure variation and Strokevolume variation by pulse contour analysis are affected by arterialcompliance. Transoesophageal echocardiography has the addedadvantage of being able to identify the causative problemby being able to look at both ventricles. However even strokevolume variation assessed by echocardiography is subject tocertain limitations, which have to be considered duringinterpretation.

Stroke volume variation has to be measured only in mechan-ically ventilated patients. The patient should be in the continuousmandatory ventilation mode without any spontaneous breathingduring the time of measurement. The tidal volume should be atleast 8 ml/kg. In the presence of arrhythmias, the arrhythmia canbe the cause of the filling problems on the left ventricle. Rightheart dysfunction also causes filling deficiencies on the left side of

Fig. 9. The stroke volume variation in constrictive pericarditis.

K. Indraratna / Trends in Anaesthesia and Critical Care 2 (2012) 115e122 121

the heart making these patients appear as responders when theyare not. It is important to assess right heart function beforeinterpretation.

2. Limitations of SVV

� Mechanical ventilationSVV has to be assessed on patients who are mechanicallyventilated on CMV mode or without any spontaneousbreathing activity at the time of assessment, with tidalvolumes of more than 8 ml/kg.

� Spontaneous ventilationSVV cannot be used on patients who are breathingspontaneously

� ArrhythmiasSVV is not accurate in the presence of arrhythmia

� Right heart failureCan give misleading and inaccurate information.

� PEEP, chest and lung compliance should be considered beforeinterpretation.

3. Passive leg raising

� Elevation of legs to 45�

� 150 ml of blood shifted to intra thoracic compartment� A short response time of 30 s to 1 min� Therefore responsiveness needs to be measured immediately� Echo/doppler pulse contour analysis can be used� Reversible if patient is not responsive

When the limiting factors, for the functional haemodynamicindices are present, “Passive leg raising” can be done. This will autotransfuse about 150 ml of blood into the thoracic cavity. The changein stroke volume can be measured by transoesophageal echo orDoppler at the aorta. The response is rapid.34 If a change is observedthe patient can be deemed to be fluid responsive.35,3 On putting the

legs down the fluid returns back as this is a transient and temporaryfluid challenge.

4. Conclusion

It is important to remember that fluid responsiveness does notmean that the patient requires fluid. Responsiveness means that,given a fluid challenge the cardiac output will increase. The hae-modynamic indices such as cardiac index, blood pressure, CVP,SVCO2% and other clinical signs as blood pressure, heart rate,capillary refill, urine output need to be measured and assessedfirst.2 If these indicate that the patients cardiac output is notsufficient, functional haemodynamics should be used to assess fluidresponsiveness, provided that the patient is mechanically venti-lated with a tidal volume of at least 8 ml/kg, there is no rightventricular dysfunction and the patient has a normal cardiacrhythm. In addition other considerations, such as whether the chestis open, how much PEEP is on and the lung compliance, have to betaken into account during interpretation. Transoesophageal echo-cardiography can be easily used to assess functional haemody-namics. Those patients determined as responders can be given fluidchallenges to improve the cardiac output, but those identified asnon-responders should not be given fluid challenges.

Conflict of interest statement

The author declared that there is no conflict of interest.

References

1. Marik PE, Cavallazzi R, Vasu T, Hirani A. Dynamic changes in arterial waveformderived variables and fluid responsiveness in mechanically ventilated patients:a systematic review of the literature. Critical Care Medicine 2009;37:2642e7.

2. Michard F, Teboul JL. Predicting fluid responsiveness in ICU patients: a criticalanalysis of the evidence. Chest 2002;121:2000e8.

3. Pinsky MR, Payen D. Functional haemodynamic monitoring, foundations andfuture. In: Pinsky MR, Payen D, editors. Functional haemodynamic monitoring.Springer; 2006. p. 3. Update in intensive care medicine, Series ed., Vincent J.L.

K. Indraratna / Trends in Anaesthesia and Critical Care 2 (2012) 115e122122

4. Preisman S, Kogen S, Berkenstadt H, Perel A. Predicting fluid responsiveness inpatients undergoing cardiac surgery: functional haemodynamic parametersincluding respiratory systolic variation test and static preload indicators. BritishJournal of Anaesthesia 2005;95:746e55.

5. Eyre L, Breen A. Optimal volaemic status and predicting fluid responsiveness.continuing education in anaesthesia. Critical Care and Pain 2010;10(2):59e62.

6. Magder S. Clinical usefulness of respiratory variations in arterial pressure.American Journal of Respiratory Critical Care Medicine 2004;169:151e5.

7. Feinberg MS, Hopkins WE, Davila-Roman VG, Barziala B. Multiplane trans-oesophageal echocardiographic doppler imaging accurately determines cardiacoutput measurements in critically ill. Chest 1995;107:769e83.

8. Marik PE, Baram M, Vahid B. Does central venous pressure predict fluidresponsiveness. Chest 2008;134:172e8. July 2008.

9. Smith T, Grounds RM, Rhodes A. Central venous pressure. Uses and limitations.In: Pinsky MR, Payen D, editors. Functional haemodynamic monitoring. Springer;2005. p. 99. Update in intensive care medicine, Series ed. Vincent J.L.

10. Kumar A, Anel R, Bunnel E. Pulmonary artery occlusion pressure and centralvenous pressure fail to predict ventricular filling volume, cardiac performanceor the response to volume infusion in normal subjects. Critcal Care Medicine2004;32:691e9.

11. Darmon PL, Hillel Z, Mogtader A, Mindich B, Thys D. Cardiac output by trans-oesophageal using continuous wave doppler across aortic valve. Anaesthesiol-ogy 1994;80:796e805.

12. Reuter DA, Baverlein J, Goepfert MS, Weiss FC. Influence of tidal volume on leftventricular stroke volume variation measured by pulse contour analysis inmechanically ventilated patients. Intensive Care Medicine 2003;29(3):476e80.

13. Teboul JL, Monnet X, Richard C. Arterial pulse variation during positive pres-sure ventilation and passive leg raising. In: Pinsky MR, Payen D, editors.Functional haemodynamic monitoring. Springer; 2005.

14. De Backer D, Heenan S, Piagnerelli M, Koch M, Vincent JL. Pulse pressurevariation to predict fluid responsiveness;influence of tidal volume. IntensiveCare Medicine 2005;31:517e23.

15. Suehiro K, Okutani R. Influence of tidal volume for stroke volume variation topredict fluid responsiveness in patients undergoing one lung ventilation July 92011.

16. Godje O, Peyerl M, Seebauer T, Lamm P, Mair H, Reichart B. Central venouspressure, pulmonary capillary wedge pressure and intra thoracic bloodvolumes as preload indicators in cardiac surgery patients. European Journal ofCardiothoracic Surgery 1998;13(5):533e9.

17. Feissel M, Michard F, Mangin I, Ruyer O, Faller JP, Teboul JL. Respiratorychanges in aortic blood velocity as an indicator of fluid responsiveness inventilated patients with septic shock. Chest 2001;119:867e73.

18. Michard F, Teboul JL. Using heart lung interactions to assess fluid responsive-ness during mechanical ventilation. Critical Care 2000;4:282e9.

19. Marick PE, Monnet X, Teboul JL. Haemodynamic parameters to guide fluidtherapy. Annals of Intensive Care 2011;1:1.

20. Michard F, Chemla D, Richard C. Clinical use of respiratory changes in arterialpressure to monitor haemodynamic effects of PEEP. American Journal ofRespiratory Critical Care Medicine 1999;159:935e9.

21. Kublitz JC, Annecke T, Kemming GI, Fork S, Kronas N, Goetz AE, et al. Theinfluence of positive end expiratory pressure on stroke volume variation andcentral blood volume during open and closed chest conditions. EuropeanJournal of Cardiothoracic Surgery 2006;30(1):90e5.

22. Pizov R, Cohen M, Weiss Y, Segal E, Cotev S, Perel A. Positive end expiratorypressure induced haemodynamic changes are reflected in the arterial pressurewaveform. Critical Care Medicine 1996;24(8):1381e7.

23. Yogo H, Isona S, Ikeda A, Kasuya Y, Nishiro T. Positive end expiratorypressure induced pulse pressure changes predict stroke volume variation inanaesthetized patients. European Journal of Anaesthesiology 2011;28:58.

24. Edward Life Sciences. Stroke volume variation. can we use fluid to improvehaemodynamics? 2007.

25. Lamia B, Chemla D, Richard C, Teboul JL. Clinical review-Interpretation ofarterial pressure wave in shock states. Critical Care 2005;9:601e6.

26. Teboul JL, Monnet X. Meaning of pulse pressure variation during ARDS. In:Vincent JL, editor. Annual update in intensive care and emergency medicine 2011.p. 322e31.

27. Huang CC, Jui-Ying F, Han C. Prediction of fluid responsiveness in acuterespiratory distress syndrome patients ventilated with low tidal volume andhigh positive end expiratory pressure. Critical Care Medicine 2008;36:2810e6.

28. Rex S, Schalte G, Schroth S, De Waal EE, Metzelder S, Overbeck Y, et al. Limi-tations of arterial pulse pressure variation and left ventricular stroke volumevariation in estimating cardiac preload during open heart surgery. ActaAnaesthesiology Scandal 2007;51(9):1258e67.

29. Reuter DA, Goepfert MS, Goersch T, Schmoeckel M, Kilger E, Goetz AE.Assessing fluid responsiveness during open chest conditions. British Journal ofAnaesthesia 2005;94(3):318e23.

30. Sorbara C, Romangoli S, Rossi A, Romano SM. Circulatory failure;bedside func-tional haemodynamic monitoring. In: Atlee JL, Gullo A, Singara G, Vincent JL,editors. Peri operative critical care cardiology. 2nd ed. 2007. p. 89e108.

31. Viellard-Baron A. Pulse pressure variation in managing fluid requirements.Beware the pitfalls. In: Vincent JL, editor. Year book of intensive care andemergency medicine 2006. p. 185e91.

32. Kubitz JC, Annecke T, Fork S, Kemming GI, Kronas N, Goetz AE, et al. Validationof pulse contour derived stroke volume variation during modification ofcardiac afterload. British Journal of Anaesthesia 2007;98(5):591e7.

33. Pinsky MR. Functional haemodynamic monitoring, applied physiology at thebedside. In: Vincent JL, editor. Intensive care medicine annual update. Springer;2002. p. 537e52.

34. De Backer D. Can passive leg raising be used to guide fluid administration?Critical Care 2006;10(6):170.

35. Monnet X, Rienzo M, Osmann D, Angel N, Richard C. Passive leg raising predictsfluid responsiveness in the critically ill. Critical Care Medicine 2006;34:1402e7.