BJA
Correspondence
Reply from the authors
A
B
Measured
Aortic pressure (mmHg)
p 100
Interpretation problem Ees(sb)
Before After
84% 75%
Before
After
0 Time
V0=0
V
Fig 1 (A) Aortic pressure waveforms before and after cardiopulmonary bypass of a patient undergoing coronary artery bypass graft surgery. End-systolic pressures, defined as dicrotic notch pressure, make up different percentages of systolic arterial pressure in these situations, leading to erronous estimates of (single-beat) Ees when assuming Pes ¼ 0.9 × systolic arterial pressure (B).
579
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on February 3, 2015
Editor—We thank Bertini and colleagues for their comments on our work1 and we are intrigued by their efforts to translate the concept of ventriculo-arterial coupling into their clinical practice. We agree that their suggested estimation of ‘arterial elastance’ (Ea) in addition to ventricular elastance (Ees) can further contribute to improved understanding of cardiac efficiency and work perioperatively. We can inform the authors that Ea can readily be calculated using our approach, combining stroke volume estimates by transoesophageal echocardiogram or Nexfin2 – 4 and end-systolic pressure (Pes) obtained from the non-invasive arterial pressure waveform. While we appreciate the non-invasive potential of currently available single-beat methods for Ees estimation,5 – 7 we would like to add a few comments on the physiological concepts that should be kept in mind using these methods. Importantly, approaches, as those used by Bertini and colleagues, have moved away from the concept of Ees as the slope of an incrementally determined end-systolic pressure –volume relation, and are empirical approximations which compromise several important physiological aspects of Ees.2 8 Single-beat methods either (i) assume a constant volume axis intercept V0 or (ii) impute a V0 value based on reference group statistics and multiple apparently easy-to-measure variables.5 6 In the first case (i), especially when V0 is considered 0, Ea/Ees simply degenerates to 1/EF21,2 which has two implications: first, consideration of Pes is unnecessary, arguing against the view and experience of many including Bertini and colleagues, and secondly, EF or in effect SV/Ved (stroke volume/end-diastolic volume) is exposed as a ventriculoarterial coupling parameter by itself, which does not come as a surprise when considering the well-known load-dependency of EF.2 In the second case (ii), the circulatory dynamics of the actual individual patient are equalled to peers, which may
lead to significant (i.e. on the order of 40%) under- or overestimations of Ees.6 8 The above illustrates why single-beat methods may fail to reflect changes in inotropic state when compared with multi-beat Ees estimation.2 8 If we do consider Pes as informative, then estimating endsystolic pressure as 0.9×systolic arterial pressure may be problematic in the perioperative period where the arterial pressure wave shape can change considerably because of haemodynamic variations. End-systolic pressure is best identified by the dicrotic notch, whose value is not a given fraction of the systolic maximum. This is illustrated in Figure 1 showing aortic pressure waveforms measured before and after cardiopulmonary bypass of a patient undergoing coronary artery bypass graft, illustrating markedly different end-systolic pressures at 84% and 75% of systolic arterial pressure, respectively, while the latter differs only slightly (1 mm Hg difference). Ignoring these pressure changes obviously results in erroneous conclusions (shown in Fig. 1B), especially when using single-beat approximations. While single-beat methods may be regarded as acceptable in relatively stable haemodynamic situations,6 8 we expect that perioperative fluctuations do influence the rather load-dependent input parameters of the empirical models they use and, therefore, may lead to less valid estimation of Ees. To conclude, we would like to thank Bertini and colleagues once again for their comments, contributing to the critical discussion on the clinical use of ventriculo-arterial coupling. Moreover, our colleagues should be applauded for their efforts to translate this important physiological concept in their clinical practice. However, the proposed single-beat methods may be useful to this end, but the assumptions on which these approximations are based are easily violated, especially in the perioperative period. Non-invasive pressure–volume
BJA
Correspondence
measurements based on an appropriate but safe loading intervention address the abovementioned problems that limit patient-specific haemodynamic monitoring and optimization, which is, in our opinion, an important advantage.
Declaration of interest None declared.
Funding
C. A. Boly* R. A. Bouwman K. D. Reesink Amsterdam, The Netherlands * E-mail: [email protected] 1 Boly CA, Reesink KD, van den Tol MP, et al. Minimally invasive intraoperative estimation of left-ventricular end-systolic elastance with phenylephrine as loading intervention. Br J Anaesth 2013; 111: 750–8 2 Westerhof N, Stergiopulos N, Noble MIM. Snapshots of Hemodynamics. New York, Dordrecht, Heidelberg, London: Springer, 2010 3 van Geldorp IE, Delhaas T, Hermans B, et al. Comparison of a noninvasive arterial pulse contour technique and echo Doppler aorta velocity–time integral on stroke volume changes in optimization of cardiac resynchronization therapy. Europace 2011; 13: 87 –95 4 Burkhoff D, Mirsky I, Suga H. Assessment of systolic and diastolic ventricular properties via pressure– volume analysis: a guide for clinical, translational, and basic researchers. Am J Physiol Heart Circ Physiol 2005; 289: H501– 12 5 Senzaki H, Chen C. Single-beat estimation of end-systolic pressure – volume relation in humans: a new method with the potential for noninvasive application. Circulation 1996; 94: 2497–506 6 Chen CH, Fetics B, Nevo E, et al. Noninvasive single-beat determination of left ventricular end-systolic elastance in humans. J Am Col Cardiol 2001; 38: 2028–34 7 Guarracino F, Cariello C, Danella A, et al. Effect of levosimendan on ventriculo-arterial coupling in patients with ischemic cardiomyopathy. Acta Anaesthesiol Scand 2007; 51: 1217–24 8 Kjørstad KE, Korvald C, Myrmel T. Pressure– volume-based singlebeat estimations cannot predict left ventricular contractility in vivo. Am J Physiol Heart Circ Physiol 2002; 282: H1739– 50
doi:10.1093/bja/aeu030
Predictability of the respiratory variation of stroke volume varies according to the definition of fluid responsiveness Editor—There has been abundant literature published on the ability of dynamic preload indices [such as the respiratory
580
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on February 3, 2015
R.A.B. is supported by a Dr E. Dekker research career grant of the Netherlands Heart foundation (grant 2008T003), a research career grant of the Netherlands Society of Anesthesiologists, and a ZonMW Clinical Fellow Stipendium (grant 40-0070397-305). K.D.R. is supported by a special purpose fund of the Imperial College Healthcare Charity (SPF 7037).
variation of stroke volume (DrespSV)] to predict fluid responsiveness, but the results have been inconsistent.1 2 One of the reasons could be the absence of a consensual definition of fluid responsiveness. Some authors have defined fluid responsiveness as an increase in SV, while others have used an increase in cardiac output (CO).1 2 In some cases, variations of CO may not be equal to variations of SV due to changes in heart rate (HR). We assumed that assessment of fluid responsiveness based on SV would provide more reliable information concerning the effect of fluid infusion than assessment of fluid responsiveness based on CO. After Institutional Review Board approval, we conducted a prospective, observational study, in which patients monitored by an oesophageal Doppler during general anaesthesia were included. Haemodynamic parameters [HR, mean arterial pressure (MAP)] and oesophageal Doppler indices (SV, CO, and DrespSV) were collected before and after volume expansion with 500 ml of crystalloid. Two definitions of responders were tested: .15% increase in SV and .15% increase in CO after volume expansion. Patients were classified into three groups: non-responders, responders according to SV and CO, and discordant group. An analysis of variance (ANOVA) analysis with the Bonferroni post hoc analysis was used. The Pearson rank method tested linear correlations. A receiver-operating characteristic curve was generated for DrespSV to predict either increase in CO or in SV. Of the 138 patients included in this study, 64 (47%) increased their SV and CO by more than 15%, and 15 (12%) increased only their SV by more than 15%. Patients in the discordant group had a significantly higher baseline HR, with a ,10% increase in CO (Table 1). Variation of CO with fluid expansion was correlated with variation of SV and HR (r¼0.78, P,0.001, r¼0.37, P,0.001, respectively). DrespSV presented a better ability to predict a .15% increase in SV than a .15% increase in CO [area under the curve of 0.89 (CI95% 0.82–0.95, P,0.0001) and 0.77 (CI95% 0.68–0.85, P,0.0001), respectively]. Using Oldham’s method, DrespSV and SV variations in response to fluid expansion were not correlated. Based on our results, a standardized fluid infusion gave different results according to the two definitions of fluid responsiveness. Some patients were not classified as CO responders because fluid infusion produced opposite effects on the two determinants of CO, as fluid expansion was sometimes associated with a significant decrease in HR resulting in a less marked increase in CO than the cut-off used to define fluid responsiveness. Baseline haemodynamic parameters and their variation in response to volume expansion differed between the responder and discordant groups (Table 1). Our observations may reflect two types of haemodynamic response to volume expansion depending on the baseline cardiovascular equilibrium. Assuming that all patients had no modification of drug dosage during the study period, baseline haemodynamic parameters of the discordant group may reflect adaptation of cardiovascular system, rather than in responders. Because CO is a physiologically controlled parameter dependent on several factors,3 CO variations in response to volume expansion must
Correspondence
Reply from the authors
A
B
Measured
Aortic pressure (mmHg)
p 100
Interpretation problem Ees(sb)
Before After
84% 75%
Before
After
0 Time
V0=0
V
Fig 1 (A) Aortic pressure waveforms before and after cardiopulmonary bypass of a patient undergoing coronary artery bypass graft surgery. End-systolic pressures, defined as dicrotic notch pressure, make up different percentages of systolic arterial pressure in these situations, leading to erronous estimates of (single-beat) Ees when assuming Pes ¼ 0.9 × systolic arterial pressure (B).
579
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on February 3, 2015
Editor—We thank Bertini and colleagues for their comments on our work1 and we are intrigued by their efforts to translate the concept of ventriculo-arterial coupling into their clinical practice. We agree that their suggested estimation of ‘arterial elastance’ (Ea) in addition to ventricular elastance (Ees) can further contribute to improved understanding of cardiac efficiency and work perioperatively. We can inform the authors that Ea can readily be calculated using our approach, combining stroke volume estimates by transoesophageal echocardiogram or Nexfin2 – 4 and end-systolic pressure (Pes) obtained from the non-invasive arterial pressure waveform. While we appreciate the non-invasive potential of currently available single-beat methods for Ees estimation,5 – 7 we would like to add a few comments on the physiological concepts that should be kept in mind using these methods. Importantly, approaches, as those used by Bertini and colleagues, have moved away from the concept of Ees as the slope of an incrementally determined end-systolic pressure –volume relation, and are empirical approximations which compromise several important physiological aspects of Ees.2 8 Single-beat methods either (i) assume a constant volume axis intercept V0 or (ii) impute a V0 value based on reference group statistics and multiple apparently easy-to-measure variables.5 6 In the first case (i), especially when V0 is considered 0, Ea/Ees simply degenerates to 1/EF21,2 which has two implications: first, consideration of Pes is unnecessary, arguing against the view and experience of many including Bertini and colleagues, and secondly, EF or in effect SV/Ved (stroke volume/end-diastolic volume) is exposed as a ventriculoarterial coupling parameter by itself, which does not come as a surprise when considering the well-known load-dependency of EF.2 In the second case (ii), the circulatory dynamics of the actual individual patient are equalled to peers, which may
lead to significant (i.e. on the order of 40%) under- or overestimations of Ees.6 8 The above illustrates why single-beat methods may fail to reflect changes in inotropic state when compared with multi-beat Ees estimation.2 8 If we do consider Pes as informative, then estimating endsystolic pressure as 0.9×systolic arterial pressure may be problematic in the perioperative period where the arterial pressure wave shape can change considerably because of haemodynamic variations. End-systolic pressure is best identified by the dicrotic notch, whose value is not a given fraction of the systolic maximum. This is illustrated in Figure 1 showing aortic pressure waveforms measured before and after cardiopulmonary bypass of a patient undergoing coronary artery bypass graft, illustrating markedly different end-systolic pressures at 84% and 75% of systolic arterial pressure, respectively, while the latter differs only slightly (1 mm Hg difference). Ignoring these pressure changes obviously results in erroneous conclusions (shown in Fig. 1B), especially when using single-beat approximations. While single-beat methods may be regarded as acceptable in relatively stable haemodynamic situations,6 8 we expect that perioperative fluctuations do influence the rather load-dependent input parameters of the empirical models they use and, therefore, may lead to less valid estimation of Ees. To conclude, we would like to thank Bertini and colleagues once again for their comments, contributing to the critical discussion on the clinical use of ventriculo-arterial coupling. Moreover, our colleagues should be applauded for their efforts to translate this important physiological concept in their clinical practice. However, the proposed single-beat methods may be useful to this end, but the assumptions on which these approximations are based are easily violated, especially in the perioperative period. Non-invasive pressure–volume
BJA
Correspondence
measurements based on an appropriate but safe loading intervention address the abovementioned problems that limit patient-specific haemodynamic monitoring and optimization, which is, in our opinion, an important advantage.
Declaration of interest None declared.
Funding
C. A. Boly* R. A. Bouwman K. D. Reesink Amsterdam, The Netherlands * E-mail: [email protected] 1 Boly CA, Reesink KD, van den Tol MP, et al. Minimally invasive intraoperative estimation of left-ventricular end-systolic elastance with phenylephrine as loading intervention. Br J Anaesth 2013; 111: 750–8 2 Westerhof N, Stergiopulos N, Noble MIM. Snapshots of Hemodynamics. New York, Dordrecht, Heidelberg, London: Springer, 2010 3 van Geldorp IE, Delhaas T, Hermans B, et al. Comparison of a noninvasive arterial pulse contour technique and echo Doppler aorta velocity–time integral on stroke volume changes in optimization of cardiac resynchronization therapy. Europace 2011; 13: 87 –95 4 Burkhoff D, Mirsky I, Suga H. Assessment of systolic and diastolic ventricular properties via pressure– volume analysis: a guide for clinical, translational, and basic researchers. Am J Physiol Heart Circ Physiol 2005; 289: H501– 12 5 Senzaki H, Chen C. Single-beat estimation of end-systolic pressure – volume relation in humans: a new method with the potential for noninvasive application. Circulation 1996; 94: 2497–506 6 Chen CH, Fetics B, Nevo E, et al. Noninvasive single-beat determination of left ventricular end-systolic elastance in humans. J Am Col Cardiol 2001; 38: 2028–34 7 Guarracino F, Cariello C, Danella A, et al. Effect of levosimendan on ventriculo-arterial coupling in patients with ischemic cardiomyopathy. Acta Anaesthesiol Scand 2007; 51: 1217–24 8 Kjørstad KE, Korvald C, Myrmel T. Pressure– volume-based singlebeat estimations cannot predict left ventricular contractility in vivo. Am J Physiol Heart Circ Physiol 2002; 282: H1739– 50
doi:10.1093/bja/aeu030
Predictability of the respiratory variation of stroke volume varies according to the definition of fluid responsiveness Editor—There has been abundant literature published on the ability of dynamic preload indices [such as the respiratory
580
Downloaded from http://bja.oxfordjournals.org/ at Georgetown University on February 3, 2015
R.A.B. is supported by a Dr E. Dekker research career grant of the Netherlands Heart foundation (grant 2008T003), a research career grant of the Netherlands Society of Anesthesiologists, and a ZonMW Clinical Fellow Stipendium (grant 40-0070397-305). K.D.R. is supported by a special purpose fund of the Imperial College Healthcare Charity (SPF 7037).
variation of stroke volume (DrespSV)] to predict fluid responsiveness, but the results have been inconsistent.1 2 One of the reasons could be the absence of a consensual definition of fluid responsiveness. Some authors have defined fluid responsiveness as an increase in SV, while others have used an increase in cardiac output (CO).1 2 In some cases, variations of CO may not be equal to variations of SV due to changes in heart rate (HR). We assumed that assessment of fluid responsiveness based on SV would provide more reliable information concerning the effect of fluid infusion than assessment of fluid responsiveness based on CO. After Institutional Review Board approval, we conducted a prospective, observational study, in which patients monitored by an oesophageal Doppler during general anaesthesia were included. Haemodynamic parameters [HR, mean arterial pressure (MAP)] and oesophageal Doppler indices (SV, CO, and DrespSV) were collected before and after volume expansion with 500 ml of crystalloid. Two definitions of responders were tested: .15% increase in SV and .15% increase in CO after volume expansion. Patients were classified into three groups: non-responders, responders according to SV and CO, and discordant group. An analysis of variance (ANOVA) analysis with the Bonferroni post hoc analysis was used. The Pearson rank method tested linear correlations. A receiver-operating characteristic curve was generated for DrespSV to predict either increase in CO or in SV. Of the 138 patients included in this study, 64 (47%) increased their SV and CO by more than 15%, and 15 (12%) increased only their SV by more than 15%. Patients in the discordant group had a significantly higher baseline HR, with a ,10% increase in CO (Table 1). Variation of CO with fluid expansion was correlated with variation of SV and HR (r¼0.78, P,0.001, r¼0.37, P,0.001, respectively). DrespSV presented a better ability to predict a .15% increase in SV than a .15% increase in CO [area under the curve of 0.89 (CI95% 0.82–0.95, P,0.0001) and 0.77 (CI95% 0.68–0.85, P,0.0001), respectively]. Using Oldham’s method, DrespSV and SV variations in response to fluid expansion were not correlated. Based on our results, a standardized fluid infusion gave different results according to the two definitions of fluid responsiveness. Some patients were not classified as CO responders because fluid infusion produced opposite effects on the two determinants of CO, as fluid expansion was sometimes associated with a significant decrease in HR resulting in a less marked increase in CO than the cut-off used to define fluid responsiveness. Baseline haemodynamic parameters and their variation in response to volume expansion differed between the responder and discordant groups (Table 1). Our observations may reflect two types of haemodynamic response to volume expansion depending on the baseline cardiovascular equilibrium. Assuming that all patients had no modification of drug dosage during the study period, baseline haemodynamic parameters of the discordant group may reflect adaptation of cardiovascular system, rather than in responders. Because CO is a physiologically controlled parameter dependent on several factors,3 CO variations in response to volume expansion must
