- Email: [email protected]
Predicting fluid responsiveness
Trends in Anaesthesia and Critical Care 2 (2012) 15e19
Contents lists available at SciVerse ScienceDirect
Trends in Anaesthesia and Critical Care journal homepage: www.elsevier.com/locate/tacc
REVIEW
Predicting fluid responsiveness Zubair U. Mohamed, Jost W. Mullenheim* Anaesthetics and Intensive Care Medicine, James Cook University Hospital, Marton Road, Middlesbrough, Cleveland TS4 3BW, UK
s u m m a r y Keywords: Cardiovascular monitoring Fluid therapy Fluid optimisation Fluid challenge
Fluid therapy is a key component of resuscitation of critically ill patients. However, inadvertent administration of intravenous fluids can have deleterious effects on the patient outcome. Thus, the ability to identify patients who would respond to fluid administration by increasing stroke volume and hence cardiac output is of vital importance. This article attempts to define ‘fluid challenge’ and ‘fluid responsiveness‘ and also looks at the advantages and limitations of currently used strategies. The recent increase of research interest in this field reflects the evidence that early fluid optimisation of critically ill patients improves outcome. This concept has subsequently been extended into the peri-operative setting. A brief summary of the latest research in these fields is given. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Hypotension is common in critically ill patients. The current internationally recommended first line therapy for hypotensive critically ill patients is a fluid challenge.1 Assessment of intravascular volume and need for fluid resuscitation in patients who present to accident and emergency with acute fluid loss or severe sepsis, using clinical signs like skin turgor, heart rate, blood pressure, urine output and chest examination is relatively straight forward. However, clinical assessment of intravascular volume depletion in an intensive care environment or in patients undergoing major surgery can often be inaccurate.2 Patients admitted to the intensive care unit are usually fluid resuscitated to a degree and hypotension in this setting can be due to various causes like cardiac failure, actual or relative hypovolemia, vasoplegia etc occurring either in isolation or in combination. While fluid administration is the treatment for hypovolemia, it may not be the appropriate treatment option for other causes of hypotension. Excessive fluid administration can cause iatrogenic volume overload leading to tissue oedema with consecutive tissue hypoxia, cardiac dysfunction, worsening gas exchange and haemodilution. There is evidence to suggest that critically ill patients who receive more fluids have worse outcomes.2,3 Only about 50% of haemodynamically unstable patients respond to a fluid challenge.2,4,5 Hence, fluid should be administered only to patients who are preload dependent, i.e., respond to a fluid
* Corresponding author. Tel.: þ44 1642854643. E-mail address: [email protected] (J.W. Mullenheim). 2210-8440/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tacc.2011.10.003
challenge by increasing stroke volume (SV) and hence cardiac output (CO). The challenge is to identify these patients, in other words, predict fluid responsiveness. 2. Physiologic basis CO is a function of heart rate (HR) and SV. SV, in turn, is dependent on preload, afterload and contractility. Arterial blood pressure is dependent on CO and systemic vascular resistance (SVR). According to FrankeStarling’s Law, ‘the energy of contraction is proportional to the initial length of the cardiac muscle fibre’.6 Ventricular preload is defined as the myocardial fibre length at the end of diastole. In practical terms, this relates to the ventricular end-diastolic volume. Hence we aim for a ventricular end-diastolic volume that will achieve the maximum contractility, thereby increasing SV, CO and thus oxygen delivery. These will be the patients on the steep part of the FrankeStarling curve who respond to volume expansion more than those in the plateau phase (Fig. 1). In physiological terms, predicting fluid responsiveness seeks to identify patients who are on the steep part of the FrankeStaling curve, who would increase their SV and hence CO in response to a fluid challenge. 3. Heartelung interaction in mechanically ventilated patients Mechanically ventilated patients without any spontaneous breathing activity who are in sinus rhythm present a unique opportunity to assess intravascular volume status. Intermittent positive pressure ventilation causes cyclical changes in SV of the
16
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
Table 1 Parameters involved in a fluid challenge. Choice of fluid Amount to be infused Duration of infusion Adequacy of challenge Target parameter Assessment time frame Positive response
2.
Fig. 1. FrankeStarling relationship. FrankeStarling curve showing the relation between ventricular preload and stroke volume. A given change in preload induces a greater change in stroke volume in A (Preload dependent), when compared to B (Preload independent). Figure adapted from the following article: Michard and Teboul. Using heartelung interactions to assess fluid responsiveness during mechanical ventilation. Critical care 2000; 4:282e289. Permission to reproduce granted under BioMed Central’s general terms.
ventricles. During inspiration, positive intrathoracic pressure impedes venous return to the right ventricle (RV), thereby decreasing RV preload. At the same time, the transpulmonary pressure is increased, leading to an increase in the RV afterload. This decrease in RV preload and increase in RV afterload, leads to a decrease in RV SV. This, in turn, leads to a decrease in the left ventricular (LV) preload after 2e3 heart beats (time taken for the blood to reach the LV, i.e., pulmonary blood transit time). This LV preload reduction causes a consecutive decrease in LV SV, which is at its minimum during the expiratory phase assuming a respiratory rate (RR) within physiological limits.7 The LV SV is maximum at the early inspiratory phase which is due to one or more of the following reasons: increased pulmonary venous return, increased LV compliance due to decreased RV dimensions, decreased LV afterload, and/ or external pressure on the RV.8 These cyclical changes are more marked if the ventricles are operating on the steep portion of the FrankeStarling curve and the magnitude of these changes is used as an indicator of biventricular preload dependence. 4. Defining ‘fluid challenge’ and ‘fluid responsiveness’ ‘Fluid responsiveness’ is defined as the ability of SV to increase in response to a fluid infusion or a “fluid challenge”.9 Weil and Henning introduced the concept of a “fluid challenge”.10 They described a method of administering a predetermined amount of fluid over a fixed amount of time and assessing the body’s response in terms of changes in central venous pressure (CVP), pulmonary artery diastolic pressure or pulmonary artery occlusion pressure (PAOP). Subsequent studies investigating fluid responsiveness used different measurements, fluids, parameters and indices. There is no current international consensus on what defines a ‘fluid challenge’ and many authors have recently attempted to standardise the different parameters involved in a fluid challenge.11,12 A fluid challenge involves the following parameters (Table 1): 1. Choice of fluid: Colloids tend to be more popular in Europe while crystalloids are preferred in North America. Since the aim of a fluid challenge is to administer the minimum amount of fluid necessary to identify fluid responsiveness while avoiding fluid overload at the same time, colloids may be a better choice.
3. 4.
5. 6.
Colloid (usually) 250 ml or 3 ml/kg 5e10 min Change in CVP of 2 cm H2O MAP, SV, CO Variable depending on CO monitor used Increment of 10e15%
Colloids have, however, been suggested to have a haemodynamic benefit that is unrelated to the FrankeStarling Law.13 Amount and duration of infusion: Surviving Sepsis campaign advises 500e1000 ml crystalloid or 300e500 ml colloid over 30 min for fluid resuscitation in septic patients.1 However, small fluid boluses of 250 ml or 3 ml/kg infused over 5e10 min may be more appropriate to assess fluid responsiveness12 as fluid administered redistributes rapidly into extravascular compartments in these critically ill patients particularly when crystalloids are used. Target parameters: Usually a predefined mean arterial pressure (MAP), SV or CO. Assessing adequacy of fluid challenge: It is necessary to ensure that the amount of fluid administered is of sufficient volume to increase the RV end-diastolic volume significantly so as to affect a change in SV. Otherwise there has not been a valid test of FrankeStarling’s Law and false negative tests may occur. Measurement of end-diastolic dimensions on echocardiogram or duration of aortic blood flow measured by oesophageal doppler can be used for this purpose.14 However, a simpler surrogate may be a change in CVP of at least 2 cm H2O.12,15 An increase in SV/CO of 10e15% has been suggested as a positive response.5,12 Response time: This is the time taken for the monitor to detect a change in the parameters measured. The short response time (s) of the newer flow based monitors facilitates rapid data output and real time calculation of SV and CO.
Advantages of a resuscitation strategy involving fluid challenges include: 1. Testing preload reserve and quantification of the cardiovascular response to fluid administration.11,12 2. Prompt correction of fluid deficit. 3. Minimising the risk of fluid overload and its subsequent complications.11
4.1. Methods used to predict fluid responsiveness An ideal method to predict fluid responsiveness would be a cheap, direct, easy to perform, minimally invasive and continuous measurement with a high specificity and sensitivity. The fact that a multitude of methods is used to predict fluid responsiveness is a reflection of the lack of an ideal method. Currently used methods either use static or dynamic measurements. 5. Static measurements and its limitations 5.1. Central venous pressure and pulmonary artery occlusion pressure CVP is by far the most commonly used parameter to assess ‘filling’ in critically ill patients and guide fluid responsiveness.16 However, Marik et al. showed that there is no correlation
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
17
between CVP and blood volume and that CVP cannot be used to predict fluid responsiveness.4 The pooled correlation coefficient between baseline CVP and change in SV index/cardiac index after fluid administration was 0.18 (95% CI, 0.08e0.28). The pooled area under the receiver operating characteristics (ROC) curve was 0.56 (95% CI, 0.51e0.61). Thus CVP can accurately predict fluid responsiveness in only 56% of patients which is just slightly better than chance. The same limitation applies to PAOP.5,17e19 Similarly, changes in CVP or PAOP in response to a fluid challenge do not reflect changes in ventricular preload.17 These findings can be explained due to multiple reasons:
minimum values of the peak blood flow velocity over a respiratory cycle, Vpeakmax and Vpeakmin respectively, are used to calculate the respiratory changes in peak velocity (DVpeak). DVpeak is calculated as the difference between Vpeakmax and Vpeakmin divided by the mean of the two values and expressed as a percentage. Feissel et al., found that a threshold DVpeak of 12% had a 91% positive predictive value and 100% negative predictive value to differentiate between responders and non-responders in sedated ventilated patients.22
1. Patients with similar cardiac filling pressures may be on different parts of the FrankeStarling curve as regards to their cardiac function. Hence, those in the steep part of the curve may not demonstrate an increase in filling pressure to a fluid challenge while those on the flat part of the curve may do so.19 2. It is the transmural pressure and not the intracavitary pressure such as right atrial pressure (RAP) and PAOP that is related to end-diastolic volumes via the chamber compliance. Ventricular compliance is frequently altered in critically ill patients. 3. The ventricular diastolic compliance curves are non-linear.20 In patients with isolated RV dysfunction, a fluid challenge may increase the right heart filling pressure even with low LV preload. 4. RAP and PAOP have been shown to overestimate transmural pressures in patients with external or intrinsic positive end expiratory pressure (PEEP). 5. Filling pressures can paradoxically decline after fluid repletion as a result of decreased sympathetic stimulation.11
This non-invasive method using Doppler ultrasound measures the change in brachial artery peak velocity over a respiratory cycle. The difference in the maximum and minimum velocity over the respiratory cycle is divided by the mean of the two values and expressed as a percentage. A change of more than 10% over a respiratory cycle has been found to be 74% sensitive and 95% specific to predict fluid responsiveness in mechanically ventilated patients with acute circulatory failure.23
5.2. RV end-diastolic volume index (RVEDVI) and LV end-diastolic area (LVEDA) Studies looking at the relationship between end-diastolic volume index and volume responsiveness have found that patients with RVEDVI < 90 ml/m2 were more likely to respond to a fluid challenge and those with RVEDVI > 138 ml/m2 were less likely to respond to a volume expansion with those in between representing a group with uncertain predictability.21 The estimation of the LVEDA by echocardiography has also not been shown to predict fluid responsiveness accurately.5 This can be explained as follows: the preinfusion end-diastolic volume tells little about the diastolic chamber compliance. The rise in SV as a result of end-diastolic volume increase depends on ventricular function since a decrease in ventricular contractility decreases the slope of the relationship between end-diastolic volume and SV. A patient can be a non-responder to a fluid challenge because of high venous compliance, low ventricular compliance and/or ventricular dysfunction.5
6.2. Brachial artery peak velocity variation
6.3. Superior vena cava collapsibility index and inferior vena cava distensibility index The changes in RAP during positive pressure ventilation are reflected on to the vena cavae. The subsequent change in their diameter can be measured using echocardiography. In mechanically ventilated patients, the superior vena cava collapsibility index is calculated as the maximum diameter on expiration minus the minimum diameter on inspiration divided by the maximum diameter on expiration. Vieillard-Baron et al. have demonstrated that a threshold superior vena cava collapsibility index of 36% can reliably predict responders to fluid challenge with 90% sensitivity and 100% specificity in ventilated septic patients.24 The inferior vena cava distensibility index (dIVC) is calculated as follows: maximum diameter (Dmax) minus minimum diameter (Dmin) divided by Dmin. Barbier et al. found that a dIVC threshold of 18% can reliably predict a responder with 90% sensitivity and 90% specificity.25 Feissel et al. calculated the respiratory variation in the inferior vena cava diameter as (Dmax Dmin)/(Dmax þ Dmin)/2 expressed as a percentage. They found that a 12% threshold had a positive predictive value of 93% to predict fluid responsiveness and a negative predictive value of 92%, respectively.26 The limitations of all echocardiographic measurements include availability, skill to use the equipment, the practical aspects of getting independent measurements each time a fluid bolus is given and that echocardiography is not a monitoring tool. Intra-abdominal pressure and vasopressor agents can influence the dimensions of vena cavae and hence the data derived from these numbers. 6.4. Pulse pressure variation, stroke volume variation, systolic pressure variation and DDown
6. Dynamic measurements and its limitations The dynamic assessment of fluid responsiveness uses the principle of heartelung interaction in mechanically ventilated patients with no spontaneous breathing activity, which are in sinus rhythm, as mentioned above. It involves assessing the beat to beat variability of various haemodynamic parameters over a respiratory cycle. 6.1. Aortic blood flow velocity Aortic blood flow velocity can be measured via echocardiography or suprasternal/oesophageal doppler. The maximum and
Pulse pressure is the difference between systolic and diastolic arterial blood pressure (Fig. 2). Pulse pressure variation (PPV) is derived from two variables, namely maximum pulse pressure during mechanical inspiration (PPmax) and minimum pulse pressure at expiration (PPmin). PPV is calculated as (PPmax PPmin)/ (PPmax þ PPmin)/2, and is expressed as a percentage.5 The sensitivity and specificity of PPV to predict an increase in CO by 10e15% among patients admitted to intensive care was 89% (95% CI 82e94%) and 88% (95% CI 81e92%). The area under the ROC curve was 0.94 (95% CI 0.92e0.96).2,5 The average threshold value for PPV predicting fluid responsiveness including different groups of patients is 12.5 1.6%.2
18
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
6.6. Limitations of dynamic measurements Use of respirophasic variations in the above described haemodynamic parameters to predict fluid responsiveness is valid only when the following conditions are fulfilled.31,32
Fig. 2. Respiratory changes in airway pressure and arterial pressure/stroke volume in a mechanically ventilated patient. SPV e Systolic pressure variation, PPmax e Maximum pulse pressure, PPmin e Minimum pulse pressure, ΔUp and ΔDown e Increase and decrease in systolic pressure compared to baseline, SVmax e Maximum stroke volume, SVmin e Minimum stroke volume. Figure modified and adapted from the following article: Michard and Teboul. Using heartelung interactions to assess fluid responsiveness during mechanical ventilation. Critical care 2000; 4:282e289. Permission to reproduce granted under BioMed Central’s general terms.
1. Controlled mechanical ventilation with no spontaneous breathing and no active expiration 2. Tidal volume of 8 ml/kg 3. Sinus rhythm without frequent ventricular or supraventricular ectopics 4. Absence of cor pulmonale 5. HR/RR > 3.6 6. No change in autonomic nervous system activity (e.g. due to stimuli like pain, noise, anxiety) during measurements Thus, there are a significant number of intensive care patients who do not fulfil the above-mentioned criteria for using respirophasic variations in haemodynamic parameters to predict fluid responsiveness. 7. Passive leg raising
SV variation (SVV) is the difference between the maximum and minimum SV over the respiratory cycle. The sensitivity and specificity of SVV to predict an increase in CO by 10e15%, in a mixed post-surgical and intensive care population of patients was 82% (95% CI 75e98%) and 86% (95% CI 77e92%). The area under the ROC curve was 0.84 (95% CI 0.81e0.87).2 The average threshold value for SVV predicting fluid responsiveness investigated within different groups of patients is 11.6 1.9%.2 Systolic Pressure Variation (SPV) is the difference between maximal and minimal values of systolic blood pressure during one positive pressure mechanical breath. The correlation coefficient of SPV to predict fluid responsiveness in a mixed population of surgical and intensive care patients is 0.72 (95% CI 0.66e0.78) with an area under the ROC curve of 0.86 (95% CI 0.82e0.90).2 By using the systolic pressure at end expiration as baseline one can further divide SPV into two components: an increase (DUp) and a decrease (DDown) in systolic pressure compared to baseline. Tavernier et al. reported that, in septic patients, Ddown was a more predictive component of SPV, with a Ddown threshold of 5 mmHg having a positive predictive value and negative predictive value of 95% and 93%, respectively.27 The diagnostic accuracy of these arterial waveform derived variables is significantly better than those reported for CVP and LVEDA.2 Among the above-mentioned arterial waveform derived variables, the diagnostic accuracy of PPV is significantly better than SVV or SPV. This may be explained by the fact that SV calculation from the pulse waveform involves a number of assumptions and may not be as accurate when it comes to beat to beat analysis of small SV changes.2 6.5. Pulse oximeter plethysmograph A 9.5e15% respiratory variation in pulse oximeter plethysmography waveform amplitude (ΔPOP) has been shown to be a modest predictor of fluid responsiveness in mechanically ventilated patients, with a sensitivity of 81% and specificity of 78% and an area under ROC 0.88 (95% CI 0.80e0.96).28 Plethysmographic variability index (PVI) is an algorithm allowing for automated, non-invasive and continuous monitoring of ΔPOP, derived from the perfusion index. PVI has shown a good ability to predict fluid responsiveness both in the intraoperative and intensive care patients with circulatory failure.29,30
The currently available technique that is not affected by most of the above-mentioned limitations is passive leg raising (PLR).33,34 PLR induces an ‘autotransfusion’ of blood from the lower limbs and abdominal compartment into the central circulation. The shifted volume is higher if the patient is moved from a recumbent position into a supine position with the legs elevated. Assessment of the haemodynamic response induced by PLR requires a monitor which calculates SV and CO almost real time, i.e., every few seconds. A recent meta-analysis by Carvallo et al. showed that PLR induced changes in SV and CO is a good predictor of fluid responsiveness in critically ill patients, even in patients breathing spontaneously and patients with arrhythmias.9 A PLR induced increase in SV and/or CO was found to have a sensitivity and specificity of 89% (95% CI 84.1e93.4%) and 91% (95% CI 85.9e95.2%) to predict fluid responsiveness, respectively. The pooled area under the ROC was 0.95 (95% CI 0.92e0.97). This manoeuvre, however, cannot be performed in all critically ill patients, especially those with spine, pelvic or limb fractures. Elastic compression stockings and elevated intraabdominal pressure can influence the volume recruited by PLR.14 One specific advantage that PLR has over other techniques is that PLR is a ‘reversible self volume challenge’.35,36 8. Newer techniques Garcia et al. demonstrated that dynamic arterial elastance (Ea), as measured by the ratio of PPV to SVV, has a high sensitivity and specificity to predict fluid responsiveness in intensive care patients.37 Monnet et al. applied a 15-s long ventilatory pause at end expiration in mechanically ventilated intensive care patients. This period acted as a fluid challenge by preventing cyclical respiratory impediment to cardiac preload. They found that an increase in pulse pressure or cardiac index by 5% during this time predicted fluid responders with a sensitivity and specificity of around 90% and 100%, respectively.38 However, these measurements need further evaluation in varying clinical settings before they can be introduced to routine clinical practice. 9. Summary Even though static parameters are the most widely used, numerous studies have shown their inadequacy to predict fluid responsiveness. Dynamic parameters may be more useful in this
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
respect. Among the dynamic parameters, PPV seems to be the most accurate. However, most of the studies investigating the ability of dynamic parameters to predict fluid responsiveness were conducted in patients who are in sinus rhythm and either paralysed or deeply sedated receiving large tidal volumes without any spontaneous breathing activity. This does not reflect daily practice. PLR appears to solve this problem, though it has its own limitations. Further research needs to be done in this field. Multiplicity of measurements and lack of standardisation in the various parameters involved has made generalisation and translation to clinical practice difficult. 10. Outstanding questions Predicting fluid responsiveness is a fast developing field with an enormous amount of literature published in the last decade or two. The lack of a universally accepted definition for ‘fluid challenge’ and ‘fluid responsiveness’ make generalisability difficult. Nonuniformity of the measured target parameters and differences in the algorithms used by different CO monitors do not allow head-tohead comparison. A non-invasive, inexpensive, universally available, easy-to-use monitor providing continuous reliable data would be advantageous. Even more important is the fact that quoted cutoff values of dynamic measurements may not be universally applicable to a heterogeneous group of intensive care patients. Thus, future research may identify different threshold values for specific patient groups. Finally, it has yet to be proven that a treatment strategy using one of the above-mentioned haemodynamic parameters to predict fluid responsiveness for optimisation of intravascular volume will improve morbidity or mortality. It is likely that this approach will only improve outcome if performed early before the onset of organ failure. Conflict of interest statement Both authors have no conflict of interest to declare. References 1. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock. Crit Care Med 2008;36:296e327. 2. Marik PE, Cavallazzi R, Vasu T, Hirani A. Dynamic changes in arterial waveform derived variables and fluid responsiveness in mechanically ventilated patients: a systematic review of the literature. Crit Care Med 2009;37:2642e7. 3. Vincent JL, Sakr Y, Sprung CL, Ranieri VM, Reinhart K, Gerlach H, et al. Sepsis in European intensive care units: results of the SOAP study. Crit Care Med 2006;34:344e53. 4. Marik PE, Baram M, Vahid B. Does central venous pressure predict fluid responsiveness? A systematic review of the literature and the tale of seven mares. Chest 2008;134:172e8. 5. Michard F, Teboul JL. Predicting fluid responsiveness in ICU patients. A critical analysis of the evidence. Chest 2002;121:2000e8. 6. Ganong WF. Review of medical physiology. In: A Lange medical book. 22nd ed.. 2005. p. 572. 7. Michard F, Teboul JL. Using heart-lung interactions to assess fluid responsiveness during mechanical ventilation. Crit Care 2000;4:282e9. 8. Malhotra A, Eikermann M, Magder S. Is brachial artery peak velocity variation ready for prime time? Chest 2007;131:1279e81. 9. Cavallaro F, Sandroni C, Marano C, La Torre G, Mannocci A, De Waure C, et al. Diagnostic accuracy of passive leg raising for prediction of fluid responsiveness in adults: systematic review and meta-analysis of clinical studies. Intensive Care Med 2010;36:1475e83. 10. Weil MH, Henning RJ. New concepts in the diagnosis and fluid treatment of circulatory shock. Anesth Analg 1979;58:124e32. 11. Vincent JL, Weil MH. Fluid challenge revisited. Crit Care Med 2006;34:1333e7. 12. Cecconi M, Parsons AK, Rhodes A. What is a fluid challenge? Curr Opin Crit Care 2011;17:290e5.
19
13. Magder S, Lagonidis D. Effectiveness of albumin versus normal saline to test volume responsiveness in post-cardiac surgery patients. J Crit Care 1999;14:164e71. 14. Monnet X, Teboul JL. Passive leg raising. Intensive Care Med 2008;34:659e63. 15. Lakhal K, Ehrmann S, Runge I, Benzekri-Lefèvre D, Legras A, Dequin PF, et al. Central venous pressure measurements improve the accuracy of leg raisinginduced change in pulse pressure to predict fluid responsiveness. Intensive Care Med 2010;36:940e8. 16. Boldt J, Lenz M, Kumle B, Papsdorf M. Volume replacement strategies on intensive care units: results from a postal survey. Intensive Care Med 1998;24:147e51. 17. Kumar A, Anel R, Bunnell E, Habet K, Zanotti S, Marshall S, et al. Pulmonary artery occlusion pressure and central venous pressure fail to predict ventricular filling volume, cardiac performance, or the response to volume infusion in normal subjects. Crit Care Med 2004;32:691e9. 18. Marx G, Cope T, McCrossan L, Swaraj S, Cowan C, Mostafa SM, et al. Assessing fluid responsiveness by stroke volume variation in mechanically ventilated patients with severe sepsis. Eur J Anaesthesiol 2004;21:132e8. 19. Osman D, Ridel C, Ray P, Monnet X, Anguel N, Richard C, et al. Cardiac filling pressures are not appropriate to predict hemodynamic response to volume challenge. Crit Care Med 2007;35:64e8. 20. Pinsky MR. Assessment of indices of preload and volume responsiveness. Curr Opin Crit Care 2005;11:235e9. 21. Diebel LN, Wilson RF, Tagett MG, Kline RA. End-diastolic volume: a better indicator of preload in the critically ill. Arch Surg 1992;127:817e22. 22. Feissel M, Michard F, Mangin I, Ruyer O, Faller JP, Teboul JL. Respiratory changes in aortic blood velocity as an indicator of fluid responsiveness in ventilated patients with septic shock. Chest 2001;119:867e73. 23. García MIM, Cano AG, Monrové JCD. Brachial artery peak velocity variation to predict fluid responsiveness in mechanically ventilated patients. Crit Care 2009;13:R142. doi:10.1186/cc8027. 24. Vieillard-Baron A, Chergui K, Rabiller A, Peyrouset O, Page B, Beauchet A, et al. Superior vena caval collapsibility as a gauge of volume status in ventilated septic patients. Intensive Care Med 2004;30:1734e9. 25. Barbier C, Loubieres Y, Schmit C, Hayon J, Ricôme JL, Jardin F, et al. Respiratory changes in inferior vena cava diameter are helpful in predicting fluid responsiveness in ventilated septic patients. Intensive Care Med 2004;30:1740e6. 26. Feissel M, Michard F, Faller JP, Teboul JL. The respiratory variation in inferior vena cava diameter as a guide to fluid therapy. Intensive Care Med 2004;30:1834e7. 27. Tavernier B, Makhotine O, Lebuffe G, Dupont J, Scherpereel P. Systolic pressure variation as a guide to fluid therapy in patients with sepsis-induced hypotension. Anesthesiology 1998;89:1313e21. 28. Marano CM, Cavallaro FC, Sandroni CS, Dell’Anna AD, Antonelli MA. Plethysmographic indices as predictors of fluid responsiveness: a meta-analysis. Crit Care 2010;14(Suppl. 1):P122. doi:10.1186/cc8354. 29. Cannesson M, Desebbe O, Rosamel P, Delannoy B, Robin J, Bastien O, et al. Pleth variability index to monitor the respiratory variations in the pulse oximeter plethysmographic waveform amplitude and predict fluid responsiveness in the operating theatre. Br J Anaesth 2008;101(2):200e6. 30. Loupec T, Nanadoumgar H, Frasca D, Petitpas F, Laksiri L, Baudouin D, et al. Pleth variability index predicts fluid responsiveness in critically ill patients. Crit Care Med 2011;39:294e9. 31. Brennan JM, Blair JE, Hampole C, Goonewardena S, Vasaiwala S, Shah D, et al. Radial artery pulse pressure variation correlates with brachial artery peak velocity variation in ventilated subjects when measured by internal medicine residents using hand-carried ultrasound devices. Chest 2007;131: 1301e7. 32. Heenan S, DeBacker D, Vincent JL. How can the response to volume expansion in patients with spontaneous respiratory movements be predicted? Crit Care 2006;10:R102. doi:10.1186/cc4970. 33. Maizel J, Airapetian N, Lorne E, Tribouilloy C, Massy Z, Slama M. Diagnosis of central hypovolemia by using passive leg raising. Intensive Care Med 2007;33:1133e8. 34. Préau S, Saulnier F, Dewavrin F, Durocher A, Chagnon JL. Passive leg raising is predictive of fluid responsiveness in spontaneously breathing patients with severe sepsis or acute pancreatitis. Crit Care Med 2010;38:819e25. 35. Monnet X, Teboul JL. Volume responsiveness. Curr Opin Crit Care 2007;13:549e53. 36. Teboul JL, Monnet X. Prediction of volume responsiveness in critically ill patients with spontaneous breathing activity. Curr Opin Crit Care 2008;14:334e9. 37. Monge García MI, Gil Cano A, Gracia Romero M. Dynamic arterial elastance to predict arterial pressure response to volume loading in preload-dependent patients. Crit Care 2011;15:R15. 38. Monnet X, Osman D, Ridel C, Lamia B, Richard C, Teboul JL. Predicting volume responsiveness by using the end-expiratory occlusion in mechanically ventilated intensive care unit patients. Crit Care Med 2009;37:951e6.
Contents lists available at SciVerse ScienceDirect
Trends in Anaesthesia and Critical Care journal homepage: www.elsevier.com/locate/tacc
REVIEW
Predicting fluid responsiveness Zubair U. Mohamed, Jost W. Mullenheim* Anaesthetics and Intensive Care Medicine, James Cook University Hospital, Marton Road, Middlesbrough, Cleveland TS4 3BW, UK
s u m m a r y Keywords: Cardiovascular monitoring Fluid therapy Fluid optimisation Fluid challenge
Fluid therapy is a key component of resuscitation of critically ill patients. However, inadvertent administration of intravenous fluids can have deleterious effects on the patient outcome. Thus, the ability to identify patients who would respond to fluid administration by increasing stroke volume and hence cardiac output is of vital importance. This article attempts to define ‘fluid challenge’ and ‘fluid responsiveness‘ and also looks at the advantages and limitations of currently used strategies. The recent increase of research interest in this field reflects the evidence that early fluid optimisation of critically ill patients improves outcome. This concept has subsequently been extended into the peri-operative setting. A brief summary of the latest research in these fields is given. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Hypotension is common in critically ill patients. The current internationally recommended first line therapy for hypotensive critically ill patients is a fluid challenge.1 Assessment of intravascular volume and need for fluid resuscitation in patients who present to accident and emergency with acute fluid loss or severe sepsis, using clinical signs like skin turgor, heart rate, blood pressure, urine output and chest examination is relatively straight forward. However, clinical assessment of intravascular volume depletion in an intensive care environment or in patients undergoing major surgery can often be inaccurate.2 Patients admitted to the intensive care unit are usually fluid resuscitated to a degree and hypotension in this setting can be due to various causes like cardiac failure, actual or relative hypovolemia, vasoplegia etc occurring either in isolation or in combination. While fluid administration is the treatment for hypovolemia, it may not be the appropriate treatment option for other causes of hypotension. Excessive fluid administration can cause iatrogenic volume overload leading to tissue oedema with consecutive tissue hypoxia, cardiac dysfunction, worsening gas exchange and haemodilution. There is evidence to suggest that critically ill patients who receive more fluids have worse outcomes.2,3 Only about 50% of haemodynamically unstable patients respond to a fluid challenge.2,4,5 Hence, fluid should be administered only to patients who are preload dependent, i.e., respond to a fluid
* Corresponding author. Tel.: þ44 1642854643. E-mail address: [email protected] (J.W. Mullenheim). 2210-8440/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tacc.2011.10.003
challenge by increasing stroke volume (SV) and hence cardiac output (CO). The challenge is to identify these patients, in other words, predict fluid responsiveness. 2. Physiologic basis CO is a function of heart rate (HR) and SV. SV, in turn, is dependent on preload, afterload and contractility. Arterial blood pressure is dependent on CO and systemic vascular resistance (SVR). According to FrankeStarling’s Law, ‘the energy of contraction is proportional to the initial length of the cardiac muscle fibre’.6 Ventricular preload is defined as the myocardial fibre length at the end of diastole. In practical terms, this relates to the ventricular end-diastolic volume. Hence we aim for a ventricular end-diastolic volume that will achieve the maximum contractility, thereby increasing SV, CO and thus oxygen delivery. These will be the patients on the steep part of the FrankeStarling curve who respond to volume expansion more than those in the plateau phase (Fig. 1). In physiological terms, predicting fluid responsiveness seeks to identify patients who are on the steep part of the FrankeStaling curve, who would increase their SV and hence CO in response to a fluid challenge. 3. Heartelung interaction in mechanically ventilated patients Mechanically ventilated patients without any spontaneous breathing activity who are in sinus rhythm present a unique opportunity to assess intravascular volume status. Intermittent positive pressure ventilation causes cyclical changes in SV of the
16
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
Table 1 Parameters involved in a fluid challenge. Choice of fluid Amount to be infused Duration of infusion Adequacy of challenge Target parameter Assessment time frame Positive response
2.
Fig. 1. FrankeStarling relationship. FrankeStarling curve showing the relation between ventricular preload and stroke volume. A given change in preload induces a greater change in stroke volume in A (Preload dependent), when compared to B (Preload independent). Figure adapted from the following article: Michard and Teboul. Using heartelung interactions to assess fluid responsiveness during mechanical ventilation. Critical care 2000; 4:282e289. Permission to reproduce granted under BioMed Central’s general terms.
ventricles. During inspiration, positive intrathoracic pressure impedes venous return to the right ventricle (RV), thereby decreasing RV preload. At the same time, the transpulmonary pressure is increased, leading to an increase in the RV afterload. This decrease in RV preload and increase in RV afterload, leads to a decrease in RV SV. This, in turn, leads to a decrease in the left ventricular (LV) preload after 2e3 heart beats (time taken for the blood to reach the LV, i.e., pulmonary blood transit time). This LV preload reduction causes a consecutive decrease in LV SV, which is at its minimum during the expiratory phase assuming a respiratory rate (RR) within physiological limits.7 The LV SV is maximum at the early inspiratory phase which is due to one or more of the following reasons: increased pulmonary venous return, increased LV compliance due to decreased RV dimensions, decreased LV afterload, and/ or external pressure on the RV.8 These cyclical changes are more marked if the ventricles are operating on the steep portion of the FrankeStarling curve and the magnitude of these changes is used as an indicator of biventricular preload dependence. 4. Defining ‘fluid challenge’ and ‘fluid responsiveness’ ‘Fluid responsiveness’ is defined as the ability of SV to increase in response to a fluid infusion or a “fluid challenge”.9 Weil and Henning introduced the concept of a “fluid challenge”.10 They described a method of administering a predetermined amount of fluid over a fixed amount of time and assessing the body’s response in terms of changes in central venous pressure (CVP), pulmonary artery diastolic pressure or pulmonary artery occlusion pressure (PAOP). Subsequent studies investigating fluid responsiveness used different measurements, fluids, parameters and indices. There is no current international consensus on what defines a ‘fluid challenge’ and many authors have recently attempted to standardise the different parameters involved in a fluid challenge.11,12 A fluid challenge involves the following parameters (Table 1): 1. Choice of fluid: Colloids tend to be more popular in Europe while crystalloids are preferred in North America. Since the aim of a fluid challenge is to administer the minimum amount of fluid necessary to identify fluid responsiveness while avoiding fluid overload at the same time, colloids may be a better choice.
3. 4.
5. 6.
Colloid (usually) 250 ml or 3 ml/kg 5e10 min Change in CVP of 2 cm H2O MAP, SV, CO Variable depending on CO monitor used Increment of 10e15%
Colloids have, however, been suggested to have a haemodynamic benefit that is unrelated to the FrankeStarling Law.13 Amount and duration of infusion: Surviving Sepsis campaign advises 500e1000 ml crystalloid or 300e500 ml colloid over 30 min for fluid resuscitation in septic patients.1 However, small fluid boluses of 250 ml or 3 ml/kg infused over 5e10 min may be more appropriate to assess fluid responsiveness12 as fluid administered redistributes rapidly into extravascular compartments in these critically ill patients particularly when crystalloids are used. Target parameters: Usually a predefined mean arterial pressure (MAP), SV or CO. Assessing adequacy of fluid challenge: It is necessary to ensure that the amount of fluid administered is of sufficient volume to increase the RV end-diastolic volume significantly so as to affect a change in SV. Otherwise there has not been a valid test of FrankeStarling’s Law and false negative tests may occur. Measurement of end-diastolic dimensions on echocardiogram or duration of aortic blood flow measured by oesophageal doppler can be used for this purpose.14 However, a simpler surrogate may be a change in CVP of at least 2 cm H2O.12,15 An increase in SV/CO of 10e15% has been suggested as a positive response.5,12 Response time: This is the time taken for the monitor to detect a change in the parameters measured. The short response time (s) of the newer flow based monitors facilitates rapid data output and real time calculation of SV and CO.
Advantages of a resuscitation strategy involving fluid challenges include: 1. Testing preload reserve and quantification of the cardiovascular response to fluid administration.11,12 2. Prompt correction of fluid deficit. 3. Minimising the risk of fluid overload and its subsequent complications.11
4.1. Methods used to predict fluid responsiveness An ideal method to predict fluid responsiveness would be a cheap, direct, easy to perform, minimally invasive and continuous measurement with a high specificity and sensitivity. The fact that a multitude of methods is used to predict fluid responsiveness is a reflection of the lack of an ideal method. Currently used methods either use static or dynamic measurements. 5. Static measurements and its limitations 5.1. Central venous pressure and pulmonary artery occlusion pressure CVP is by far the most commonly used parameter to assess ‘filling’ in critically ill patients and guide fluid responsiveness.16 However, Marik et al. showed that there is no correlation
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
17
between CVP and blood volume and that CVP cannot be used to predict fluid responsiveness.4 The pooled correlation coefficient between baseline CVP and change in SV index/cardiac index after fluid administration was 0.18 (95% CI, 0.08e0.28). The pooled area under the receiver operating characteristics (ROC) curve was 0.56 (95% CI, 0.51e0.61). Thus CVP can accurately predict fluid responsiveness in only 56% of patients which is just slightly better than chance. The same limitation applies to PAOP.5,17e19 Similarly, changes in CVP or PAOP in response to a fluid challenge do not reflect changes in ventricular preload.17 These findings can be explained due to multiple reasons:
minimum values of the peak blood flow velocity over a respiratory cycle, Vpeakmax and Vpeakmin respectively, are used to calculate the respiratory changes in peak velocity (DVpeak). DVpeak is calculated as the difference between Vpeakmax and Vpeakmin divided by the mean of the two values and expressed as a percentage. Feissel et al., found that a threshold DVpeak of 12% had a 91% positive predictive value and 100% negative predictive value to differentiate between responders and non-responders in sedated ventilated patients.22
1. Patients with similar cardiac filling pressures may be on different parts of the FrankeStarling curve as regards to their cardiac function. Hence, those in the steep part of the curve may not demonstrate an increase in filling pressure to a fluid challenge while those on the flat part of the curve may do so.19 2. It is the transmural pressure and not the intracavitary pressure such as right atrial pressure (RAP) and PAOP that is related to end-diastolic volumes via the chamber compliance. Ventricular compliance is frequently altered in critically ill patients. 3. The ventricular diastolic compliance curves are non-linear.20 In patients with isolated RV dysfunction, a fluid challenge may increase the right heart filling pressure even with low LV preload. 4. RAP and PAOP have been shown to overestimate transmural pressures in patients with external or intrinsic positive end expiratory pressure (PEEP). 5. Filling pressures can paradoxically decline after fluid repletion as a result of decreased sympathetic stimulation.11
This non-invasive method using Doppler ultrasound measures the change in brachial artery peak velocity over a respiratory cycle. The difference in the maximum and minimum velocity over the respiratory cycle is divided by the mean of the two values and expressed as a percentage. A change of more than 10% over a respiratory cycle has been found to be 74% sensitive and 95% specific to predict fluid responsiveness in mechanically ventilated patients with acute circulatory failure.23
5.2. RV end-diastolic volume index (RVEDVI) and LV end-diastolic area (LVEDA) Studies looking at the relationship between end-diastolic volume index and volume responsiveness have found that patients with RVEDVI < 90 ml/m2 were more likely to respond to a fluid challenge and those with RVEDVI > 138 ml/m2 were less likely to respond to a volume expansion with those in between representing a group with uncertain predictability.21 The estimation of the LVEDA by echocardiography has also not been shown to predict fluid responsiveness accurately.5 This can be explained as follows: the preinfusion end-diastolic volume tells little about the diastolic chamber compliance. The rise in SV as a result of end-diastolic volume increase depends on ventricular function since a decrease in ventricular contractility decreases the slope of the relationship between end-diastolic volume and SV. A patient can be a non-responder to a fluid challenge because of high venous compliance, low ventricular compliance and/or ventricular dysfunction.5
6.2. Brachial artery peak velocity variation
6.3. Superior vena cava collapsibility index and inferior vena cava distensibility index The changes in RAP during positive pressure ventilation are reflected on to the vena cavae. The subsequent change in their diameter can be measured using echocardiography. In mechanically ventilated patients, the superior vena cava collapsibility index is calculated as the maximum diameter on expiration minus the minimum diameter on inspiration divided by the maximum diameter on expiration. Vieillard-Baron et al. have demonstrated that a threshold superior vena cava collapsibility index of 36% can reliably predict responders to fluid challenge with 90% sensitivity and 100% specificity in ventilated septic patients.24 The inferior vena cava distensibility index (dIVC) is calculated as follows: maximum diameter (Dmax) minus minimum diameter (Dmin) divided by Dmin. Barbier et al. found that a dIVC threshold of 18% can reliably predict a responder with 90% sensitivity and 90% specificity.25 Feissel et al. calculated the respiratory variation in the inferior vena cava diameter as (Dmax Dmin)/(Dmax þ Dmin)/2 expressed as a percentage. They found that a 12% threshold had a positive predictive value of 93% to predict fluid responsiveness and a negative predictive value of 92%, respectively.26 The limitations of all echocardiographic measurements include availability, skill to use the equipment, the practical aspects of getting independent measurements each time a fluid bolus is given and that echocardiography is not a monitoring tool. Intra-abdominal pressure and vasopressor agents can influence the dimensions of vena cavae and hence the data derived from these numbers. 6.4. Pulse pressure variation, stroke volume variation, systolic pressure variation and DDown
6. Dynamic measurements and its limitations The dynamic assessment of fluid responsiveness uses the principle of heartelung interaction in mechanically ventilated patients with no spontaneous breathing activity, which are in sinus rhythm, as mentioned above. It involves assessing the beat to beat variability of various haemodynamic parameters over a respiratory cycle. 6.1. Aortic blood flow velocity Aortic blood flow velocity can be measured via echocardiography or suprasternal/oesophageal doppler. The maximum and
Pulse pressure is the difference between systolic and diastolic arterial blood pressure (Fig. 2). Pulse pressure variation (PPV) is derived from two variables, namely maximum pulse pressure during mechanical inspiration (PPmax) and minimum pulse pressure at expiration (PPmin). PPV is calculated as (PPmax PPmin)/ (PPmax þ PPmin)/2, and is expressed as a percentage.5 The sensitivity and specificity of PPV to predict an increase in CO by 10e15% among patients admitted to intensive care was 89% (95% CI 82e94%) and 88% (95% CI 81e92%). The area under the ROC curve was 0.94 (95% CI 0.92e0.96).2,5 The average threshold value for PPV predicting fluid responsiveness including different groups of patients is 12.5 1.6%.2
18
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
6.6. Limitations of dynamic measurements Use of respirophasic variations in the above described haemodynamic parameters to predict fluid responsiveness is valid only when the following conditions are fulfilled.31,32
Fig. 2. Respiratory changes in airway pressure and arterial pressure/stroke volume in a mechanically ventilated patient. SPV e Systolic pressure variation, PPmax e Maximum pulse pressure, PPmin e Minimum pulse pressure, ΔUp and ΔDown e Increase and decrease in systolic pressure compared to baseline, SVmax e Maximum stroke volume, SVmin e Minimum stroke volume. Figure modified and adapted from the following article: Michard and Teboul. Using heartelung interactions to assess fluid responsiveness during mechanical ventilation. Critical care 2000; 4:282e289. Permission to reproduce granted under BioMed Central’s general terms.
1. Controlled mechanical ventilation with no spontaneous breathing and no active expiration 2. Tidal volume of 8 ml/kg 3. Sinus rhythm without frequent ventricular or supraventricular ectopics 4. Absence of cor pulmonale 5. HR/RR > 3.6 6. No change in autonomic nervous system activity (e.g. due to stimuli like pain, noise, anxiety) during measurements Thus, there are a significant number of intensive care patients who do not fulfil the above-mentioned criteria for using respirophasic variations in haemodynamic parameters to predict fluid responsiveness. 7. Passive leg raising
SV variation (SVV) is the difference between the maximum and minimum SV over the respiratory cycle. The sensitivity and specificity of SVV to predict an increase in CO by 10e15%, in a mixed post-surgical and intensive care population of patients was 82% (95% CI 75e98%) and 86% (95% CI 77e92%). The area under the ROC curve was 0.84 (95% CI 0.81e0.87).2 The average threshold value for SVV predicting fluid responsiveness investigated within different groups of patients is 11.6 1.9%.2 Systolic Pressure Variation (SPV) is the difference between maximal and minimal values of systolic blood pressure during one positive pressure mechanical breath. The correlation coefficient of SPV to predict fluid responsiveness in a mixed population of surgical and intensive care patients is 0.72 (95% CI 0.66e0.78) with an area under the ROC curve of 0.86 (95% CI 0.82e0.90).2 By using the systolic pressure at end expiration as baseline one can further divide SPV into two components: an increase (DUp) and a decrease (DDown) in systolic pressure compared to baseline. Tavernier et al. reported that, in septic patients, Ddown was a more predictive component of SPV, with a Ddown threshold of 5 mmHg having a positive predictive value and negative predictive value of 95% and 93%, respectively.27 The diagnostic accuracy of these arterial waveform derived variables is significantly better than those reported for CVP and LVEDA.2 Among the above-mentioned arterial waveform derived variables, the diagnostic accuracy of PPV is significantly better than SVV or SPV. This may be explained by the fact that SV calculation from the pulse waveform involves a number of assumptions and may not be as accurate when it comes to beat to beat analysis of small SV changes.2 6.5. Pulse oximeter plethysmograph A 9.5e15% respiratory variation in pulse oximeter plethysmography waveform amplitude (ΔPOP) has been shown to be a modest predictor of fluid responsiveness in mechanically ventilated patients, with a sensitivity of 81% and specificity of 78% and an area under ROC 0.88 (95% CI 0.80e0.96).28 Plethysmographic variability index (PVI) is an algorithm allowing for automated, non-invasive and continuous monitoring of ΔPOP, derived from the perfusion index. PVI has shown a good ability to predict fluid responsiveness both in the intraoperative and intensive care patients with circulatory failure.29,30
The currently available technique that is not affected by most of the above-mentioned limitations is passive leg raising (PLR).33,34 PLR induces an ‘autotransfusion’ of blood from the lower limbs and abdominal compartment into the central circulation. The shifted volume is higher if the patient is moved from a recumbent position into a supine position with the legs elevated. Assessment of the haemodynamic response induced by PLR requires a monitor which calculates SV and CO almost real time, i.e., every few seconds. A recent meta-analysis by Carvallo et al. showed that PLR induced changes in SV and CO is a good predictor of fluid responsiveness in critically ill patients, even in patients breathing spontaneously and patients with arrhythmias.9 A PLR induced increase in SV and/or CO was found to have a sensitivity and specificity of 89% (95% CI 84.1e93.4%) and 91% (95% CI 85.9e95.2%) to predict fluid responsiveness, respectively. The pooled area under the ROC was 0.95 (95% CI 0.92e0.97). This manoeuvre, however, cannot be performed in all critically ill patients, especially those with spine, pelvic or limb fractures. Elastic compression stockings and elevated intraabdominal pressure can influence the volume recruited by PLR.14 One specific advantage that PLR has over other techniques is that PLR is a ‘reversible self volume challenge’.35,36 8. Newer techniques Garcia et al. demonstrated that dynamic arterial elastance (Ea), as measured by the ratio of PPV to SVV, has a high sensitivity and specificity to predict fluid responsiveness in intensive care patients.37 Monnet et al. applied a 15-s long ventilatory pause at end expiration in mechanically ventilated intensive care patients. This period acted as a fluid challenge by preventing cyclical respiratory impediment to cardiac preload. They found that an increase in pulse pressure or cardiac index by 5% during this time predicted fluid responders with a sensitivity and specificity of around 90% and 100%, respectively.38 However, these measurements need further evaluation in varying clinical settings before they can be introduced to routine clinical practice. 9. Summary Even though static parameters are the most widely used, numerous studies have shown their inadequacy to predict fluid responsiveness. Dynamic parameters may be more useful in this
Z.U. Mohamed, J.W. Mullenheim / Trends in Anaesthesia and Critical Care 2 (2012) 15e19
respect. Among the dynamic parameters, PPV seems to be the most accurate. However, most of the studies investigating the ability of dynamic parameters to predict fluid responsiveness were conducted in patients who are in sinus rhythm and either paralysed or deeply sedated receiving large tidal volumes without any spontaneous breathing activity. This does not reflect daily practice. PLR appears to solve this problem, though it has its own limitations. Further research needs to be done in this field. Multiplicity of measurements and lack of standardisation in the various parameters involved has made generalisation and translation to clinical practice difficult. 10. Outstanding questions Predicting fluid responsiveness is a fast developing field with an enormous amount of literature published in the last decade or two. The lack of a universally accepted definition for ‘fluid challenge’ and ‘fluid responsiveness’ make generalisability difficult. Nonuniformity of the measured target parameters and differences in the algorithms used by different CO monitors do not allow head-tohead comparison. A non-invasive, inexpensive, universally available, easy-to-use monitor providing continuous reliable data would be advantageous. Even more important is the fact that quoted cutoff values of dynamic measurements may not be universally applicable to a heterogeneous group of intensive care patients. Thus, future research may identify different threshold values for specific patient groups. Finally, it has yet to be proven that a treatment strategy using one of the above-mentioned haemodynamic parameters to predict fluid responsiveness for optimisation of intravascular volume will improve morbidity or mortality. It is likely that this approach will only improve outcome if performed early before the onset of organ failure. Conflict of interest statement Both authors have no conflict of interest to declare. References 1. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock. Crit Care Med 2008;36:296e327. 2. Marik PE, Cavallazzi R, Vasu T, Hirani A. Dynamic changes in arterial waveform derived variables and fluid responsiveness in mechanically ventilated patients: a systematic review of the literature. Crit Care Med 2009;37:2642e7. 3. Vincent JL, Sakr Y, Sprung CL, Ranieri VM, Reinhart K, Gerlach H, et al. Sepsis in European intensive care units: results of the SOAP study. Crit Care Med 2006;34:344e53. 4. Marik PE, Baram M, Vahid B. Does central venous pressure predict fluid responsiveness? A systematic review of the literature and the tale of seven mares. Chest 2008;134:172e8. 5. Michard F, Teboul JL. Predicting fluid responsiveness in ICU patients. A critical analysis of the evidence. Chest 2002;121:2000e8. 6. Ganong WF. Review of medical physiology. In: A Lange medical book. 22nd ed.. 2005. p. 572. 7. Michard F, Teboul JL. Using heart-lung interactions to assess fluid responsiveness during mechanical ventilation. Crit Care 2000;4:282e9. 8. Malhotra A, Eikermann M, Magder S. Is brachial artery peak velocity variation ready for prime time? Chest 2007;131:1279e81. 9. Cavallaro F, Sandroni C, Marano C, La Torre G, Mannocci A, De Waure C, et al. Diagnostic accuracy of passive leg raising for prediction of fluid responsiveness in adults: systematic review and meta-analysis of clinical studies. Intensive Care Med 2010;36:1475e83. 10. Weil MH, Henning RJ. New concepts in the diagnosis and fluid treatment of circulatory shock. Anesth Analg 1979;58:124e32. 11. Vincent JL, Weil MH. Fluid challenge revisited. Crit Care Med 2006;34:1333e7. 12. Cecconi M, Parsons AK, Rhodes A. What is a fluid challenge? Curr Opin Crit Care 2011;17:290e5.
19
13. Magder S, Lagonidis D. Effectiveness of albumin versus normal saline to test volume responsiveness in post-cardiac surgery patients. J Crit Care 1999;14:164e71. 14. Monnet X, Teboul JL. Passive leg raising. Intensive Care Med 2008;34:659e63. 15. Lakhal K, Ehrmann S, Runge I, Benzekri-Lefèvre D, Legras A, Dequin PF, et al. Central venous pressure measurements improve the accuracy of leg raisinginduced change in pulse pressure to predict fluid responsiveness. Intensive Care Med 2010;36:940e8. 16. Boldt J, Lenz M, Kumle B, Papsdorf M. Volume replacement strategies on intensive care units: results from a postal survey. Intensive Care Med 1998;24:147e51. 17. Kumar A, Anel R, Bunnell E, Habet K, Zanotti S, Marshall S, et al. Pulmonary artery occlusion pressure and central venous pressure fail to predict ventricular filling volume, cardiac performance, or the response to volume infusion in normal subjects. Crit Care Med 2004;32:691e9. 18. Marx G, Cope T, McCrossan L, Swaraj S, Cowan C, Mostafa SM, et al. Assessing fluid responsiveness by stroke volume variation in mechanically ventilated patients with severe sepsis. Eur J Anaesthesiol 2004;21:132e8. 19. Osman D, Ridel C, Ray P, Monnet X, Anguel N, Richard C, et al. Cardiac filling pressures are not appropriate to predict hemodynamic response to volume challenge. Crit Care Med 2007;35:64e8. 20. Pinsky MR. Assessment of indices of preload and volume responsiveness. Curr Opin Crit Care 2005;11:235e9. 21. Diebel LN, Wilson RF, Tagett MG, Kline RA. End-diastolic volume: a better indicator of preload in the critically ill. Arch Surg 1992;127:817e22. 22. Feissel M, Michard F, Mangin I, Ruyer O, Faller JP, Teboul JL. Respiratory changes in aortic blood velocity as an indicator of fluid responsiveness in ventilated patients with septic shock. Chest 2001;119:867e73. 23. García MIM, Cano AG, Monrové JCD. Brachial artery peak velocity variation to predict fluid responsiveness in mechanically ventilated patients. Crit Care 2009;13:R142. doi:10.1186/cc8027. 24. Vieillard-Baron A, Chergui K, Rabiller A, Peyrouset O, Page B, Beauchet A, et al. Superior vena caval collapsibility as a gauge of volume status in ventilated septic patients. Intensive Care Med 2004;30:1734e9. 25. Barbier C, Loubieres Y, Schmit C, Hayon J, Ricôme JL, Jardin F, et al. Respiratory changes in inferior vena cava diameter are helpful in predicting fluid responsiveness in ventilated septic patients. Intensive Care Med 2004;30:1740e6. 26. Feissel M, Michard F, Faller JP, Teboul JL. The respiratory variation in inferior vena cava diameter as a guide to fluid therapy. Intensive Care Med 2004;30:1834e7. 27. Tavernier B, Makhotine O, Lebuffe G, Dupont J, Scherpereel P. Systolic pressure variation as a guide to fluid therapy in patients with sepsis-induced hypotension. Anesthesiology 1998;89:1313e21. 28. Marano CM, Cavallaro FC, Sandroni CS, Dell’Anna AD, Antonelli MA. Plethysmographic indices as predictors of fluid responsiveness: a meta-analysis. Crit Care 2010;14(Suppl. 1):P122. doi:10.1186/cc8354. 29. Cannesson M, Desebbe O, Rosamel P, Delannoy B, Robin J, Bastien O, et al. Pleth variability index to monitor the respiratory variations in the pulse oximeter plethysmographic waveform amplitude and predict fluid responsiveness in the operating theatre. Br J Anaesth 2008;101(2):200e6. 30. Loupec T, Nanadoumgar H, Frasca D, Petitpas F, Laksiri L, Baudouin D, et al. Pleth variability index predicts fluid responsiveness in critically ill patients. Crit Care Med 2011;39:294e9. 31. Brennan JM, Blair JE, Hampole C, Goonewardena S, Vasaiwala S, Shah D, et al. Radial artery pulse pressure variation correlates with brachial artery peak velocity variation in ventilated subjects when measured by internal medicine residents using hand-carried ultrasound devices. Chest 2007;131: 1301e7. 32. Heenan S, DeBacker D, Vincent JL. How can the response to volume expansion in patients with spontaneous respiratory movements be predicted? Crit Care 2006;10:R102. doi:10.1186/cc4970. 33. Maizel J, Airapetian N, Lorne E, Tribouilloy C, Massy Z, Slama M. Diagnosis of central hypovolemia by using passive leg raising. Intensive Care Med 2007;33:1133e8. 34. Préau S, Saulnier F, Dewavrin F, Durocher A, Chagnon JL. Passive leg raising is predictive of fluid responsiveness in spontaneously breathing patients with severe sepsis or acute pancreatitis. Crit Care Med 2010;38:819e25. 35. Monnet X, Teboul JL. Volume responsiveness. Curr Opin Crit Care 2007;13:549e53. 36. Teboul JL, Monnet X. Prediction of volume responsiveness in critically ill patients with spontaneous breathing activity. Curr Opin Crit Care 2008;14:334e9. 37. Monge García MI, Gil Cano A, Gracia Romero M. Dynamic arterial elastance to predict arterial pressure response to volume loading in preload-dependent patients. Crit Care 2011;15:R15. 38. Monnet X, Osman D, Ridel C, Lamia B, Richard C, Teboul JL. Predicting volume responsiveness by using the end-expiratory occlusion in mechanically ventilated intensive care unit patients. Crit Care Med 2009;37:951e6.