Arterial Pressure Variation during Positive-pressure Ventilation

(1)

Ventilation

A. Perel, S. Preisman, and H. Berkenstadt

Introduction

The hemodynamic status of critically ill patients may include the whole spectrum of volume states and myocardial performance. Accurate diagnosis of this cardiovascular status is therefore mandatory for the achievement of optimal preload conditions and optimal cardiac performance. However, the hemodynamic parameters that are most often used in clinical practice to assess preload, namely the central venous pressure (CVP) and the pulmonary artery occlusion pressure (PAOP), are far from perfect. A recent literature analysis has clearly demonstrated that the CVP and the PAOP are poor predictors of the response of the cardiac output to fluid loading and cannot differentiate between patients that respond to volume loading (responders) and patients that do not (non-responders) [1]. This inadequacy of filling pressures is due to the fact that, besides being determined by the end-diastolic volume of the heart chambers, they are also directly affected by ventricular compliance, which can be quite variable. Hence pressures are limited in their ability to reflect volumes. However, `volumes’ themselves are often me- diocre predictors of fluid-responsiveness, namely the degree by which the cardiac output responds to a volume load, since their relationship to the stroke volume depends on ventricular contractility, which can be extremely variable and cannot be directly measured in clinical practice. This is why even more accurate measures of preload, like the global end-diastolic volume or the left ventricular (LV) end- diastolic area (LVEDA), have a limited capability of predicting fluid responsiveness.

Traditionally fluid responsiveness is assessed by a time-consuming and invasive graded volume loading, which includes repetitive measurements of cardiac output and filling pressures and the construction of actual Frank-Starling LV function curves. However, although fluid loading is one of the most common therapeutic steps taken in the intensive care unit (ICU), it fails to increase the cardiac output in about 50% of the patients [2]. The resulting unnecessary fluid administration may be harmful especially in patients with respiratory, renal, and/or cardiac failure.

Overzealous fluid administration may indeed be an underestimated occult source of mortality in the ICU, since the excess fluid may increase interstitial edema in various organs, increase lung water content, postpone weaning, and increase the risk of sepsis. Part of this excess fluid administration may also stem from the fact that the end-point of fluid resuscitation is frequently unclear. Hence the impor-

(2)

tance of an accurate assessment of fluid responsiveness lies not only in the detection of latent hypovolemia or a meticulous ‘prophylactic optimization’, but also in the withholding of fluids when their administration may not be of benefit.

In mechanically ventilated patients, the hemodynamic effects of the increase in intrathoracic pressure offer dynamic information about fluid responsiveness. This direct clinical application of the physiological principles of heart-lung interaction during mechanical ventilation is gaining ever-growing interest and has been the topic of many reviews and editorials, most of which are quite recent [2–14]. In this chapter, we will describe the basic physiological principles of this monitoring approach, review the various parameters that have been developed and delineate the usefulness as well as the limitations of this technique.

The Hemodynamic Effects of the Mechanical Breath

The main hemodynamic effect of the increase in the intrathoracic pressure during positive-pressure ventilation is normally a decrease in right ventricular (RV) filling due to a decreased venous return. This decrease in RV filling has been shown to be about 20% in calves with artificial hearts [15], and to result in a decrease in RV stroke output of about 20% as well in patients ventilated with tidal volumes of 10–15 ml/kg [16]. The inspiratory decrease in RV outflow may be much more significant (about 70%) in the presence of hypovolemia, and is then associated with a high degree of inspiratory collapse of the superior vena cava [17]. The inspiratory decrease in RV outflow can be normalized (to about 25%) once the vena caval ‘zone 2’ conditions are corrected by volume expansion [17].

The reduction in venous return during the mechanical breath may be so significant that large tidal volumes were shown to decrease the LV end-diastolic volume (LVEDV) only slightly less than with the inferior vena cava occlusion maneuver [18].

The effect of the mechanical breath on venous return may be more complicated in the presence of hypervolemia or congestive heart failure. In these conditions the inspiratory diaphragmatic descent and the associated increase in abdominal pressure may cause squeezing of the abdominal venous compartment and the congested liver (which are under ‘zone 3’ conditions), causing an increase in the venous return during the mechanical breath [19]. This mechanism may be responsible in part for the lack of the decrease in cardiac output when positive end-expiratory pressure (PEEP) is applied to hypervolemic patients. In patients with ARDS, the inspiratory decrease in RV stroke output may occur also because of an increase in RV outflow impedance [20].

The transient inspiratory decrease in RV output leads to a decrease in the LV stroke output after a few beats. However, the first and immediate effect of the rise in intrathoracic pressure on the LV is normally an augmentation of the LV stroke volume [15, 16, 21–24]. This augmentation is due to the inspiratory squeezing of the pulmonary blood volume, an increase in pulmonary venous flow [23] (Fig. 1), significant increases in left atrial and LV dimensions, and increased LV stroke volume [16, 21–24]. This effect is more pronounced in the presence of congested (‘zone 3’) lungs, but when the lungs are in zone 2 conditions the opposite occurs, namely pulmonary venous flow decreases during inspiration [24].

(3)

Another suggested mechanism for the early inspiratory increase in LV stroke volume is a decrease in the transmural aortic pressure reflecting an effective decrease in LV afterload [22]. However, Vieillard-Baron et al. found recently that LV systolic wall stress, an index of LV afterload, significantly increased during tidal ventilation [23]. In addition, these authors found it difficult to imagine that the small increase in pleural pressures during lung inflation in their patients would have any measurable effect on LV ejection pressure. The conclusion therefore of this excellent study is that the inspiratory augmentation of LV stroke volume is due first and foremost due to increased LV preload and not a decrease in afterload [23].

Magder, on the contrary, has suggested even more recently that the inspiratory increase in LV stroke volume is not due to increase in LV preload but rather due to the aortic valve opening earlier and staying open longer during the inspiratory increase in pleural pressure [14]. Other hypothetical mechanisms that have been mentioned as being able to contribute to the early inspiratory increase in LV stroke volume include a higher external pressure exerted on the LV by the increased lung volume, better LV contractility due to the decreased size of the RV, and lung inflation-induced adrenergic discharge. The potential clinical implications of the inspiratory augmentation of the stroke volume have been repeatedly explored, since using the mechanical breath as a form of LV assist device is both sound physiologically and appealing clinically. However, to date there is no clinically accepted method that takes advantage of this phenomenon.

The second phase of the response of the LV to the mechanical breath is normally a decrease in LV stroke output, which is the result of the earlier decrease in RV stroke output. In summary, the mechanical breath induces cyclic changes in the output of the right and left ventricles. These normally include an early increase in LV stroke output and a simultaneous decrease in RV output. The expiratory phase is normally associated with an increased RV output and a decrease in LV output.

Fig. 1. A. Airway pressure. B. Pulmonary ve- nous flow. Doppler pulmonary venous flow velocity during the respiratory cycle. The transient inspiratory increase is associated with increased left atrial and LV dimensions, transient increase in LV stroke output and a prominent∆Up (see later). Adapted from [23]

with permission

(4)

Under normal conditions, the output of both ventricles evens out at end-expiration [16].

Basic Principles of Arterial Pressure Waveform Analysis (SPV, ∆Up, ∆Down)

The previously described respiratory fluctuations in the LV stroke output are reflected in the arterial pressure waveform. The early inspiratory augmentation of the LV stroke output is reflected as an increase in the systolic blood pressure termed delta up (dUp, ∆Up), while the later decrease in LV stroke output is reflected in a decrease in the systolic blood pressure termed delta down (dDown,

∆Down) [26] (Fig. 2). The ∆Up is measured as the difference between the maximal value of the systolic blood pressure and the systolic blood pressure during a long end-expiratory pause or a short (5 seconds) apnea, while the∆Down is measured as the difference between the reference end-expiratory systolic blood pressure and the minimal systolic blood pressure value. The sum of the∆Up and the ∆Down, which is the difference between the maximal and the minimal systolic blood pressure values during one mechanical breath, is termed the systolic pressure variation (SPV) [26].

It is important to note again that the∆Up and the ∆Down represent two different hemodynamic events. The∆Down is due to the decrease in venous return during the mechanical breath, and its magnitude reflects fluid responsiveness, namely, the degree by which LV stroke output decreases in response to a transiently decreased preload. The∆Up, on the other hand, reflects the early inspiratory augmentation of the LV stroke output, and has been originally described as ‘reversed pulsus paradoxus’ [21]. Theoretically the∆Up can be influenced by some partial transmission of the airway pressure to the LV and aorta during the mechanical breath and thus not be necessarily representative of augmented LV stroke volume [26–28].

The actual degree of this transmission seems however to be minimal [14, 23], which is not surprising in view of the close correlation of the SPV to the variations in the pulse pressure [27, 28] and in the stroke volume itself [29, 30]. However, in a study that was designed to test the hypothesis that changes in the systolic blood pressure

Fig. 2. The normal changes in the arterial pressure during a mechanical breath. The difference between the maximal and minimal values of the systolic BP is the systolic pressure variation (SPV).

The systolic BP value during a short apnea is used as the reference pressure to measure the∆Up (delta up) and the∆Down (delta down), which are the two components of the SPV.

(5)

are induced solely by in-phase changes in intrathoracic pressure, and which was done in patients with relatively small systolic blood pressure values in both closed and open chest conditions, Denault et al did not find a consistent relationship between the changes in the systolic blood pressure and the LVEDA as determined by an automated border detection software [26]. They therefore claimed that

“changes in systolic arterial pressure reflect changes in airway pressure better than they reflect concomitant changes in LV hemodynamics”, and that the∆Down is not related to a decrease in preload [26]. However, besides a long list of self-admit- ted study limitations, the recorded analog signals that accompany this publication are suggestive of a possible problem of synchronicity between the arterial pressure and the LV area recordings, since the recorded transient decrease in LV area seems to be associated with a simultaneous increase in arterial pulse pressure and hence with an increase in LV stroke volume (see figure 2 in ref [26]).

In contrast with the results of Denault et al. [26], many others have found, using a variety of techniques, that indeed the changes in the arterial pressure do corre- spond to real changes in the LV stroke volume. The SPV and∆Down were repeatedly found to either very significantly correlate with, or to behave in exactly the same way as, changes in stroke volume measured or estimated by aortic velocity- time integral [31–33], by Doppler echocardiography [16, 23], by the arterial pulse pressure [27, 28], and by the pulse contour method [29, 30]. A recent study using conductance volumetry has also shown that the arterial pressure and the LV volume change simultaneously during a mechanical breath [18]. For all practical purposes, therefore, the SPV,∆Down and ∆Up should be perceived as representing true changes in the LV stroke output during the mechanical breath.

The SPV and ∆Down Reflect Volume Status and Predict Fluid Responsiveness

The respiratory changes in the systolic blood pressure were first quantified by Coyle et al. in 1983 [34]. Since then many experimental [25, 29, 35–41] and clinical [28, 30, 31, 42–47] reports have repeatedly shown that the SPV and the∆Down are sensitive indicators of changes in blood volume. The first experimental report was done in an animal model of graded hemorrhage and retransfusion [25]. The rate of the hemorrhage in this model was relatively slow so that arterial blood pressure and heart rate remained practically unchanged even when the animals were exan- guinated by 30% of their estimated blood volume and when the cardiac output was significantly reduced [25]. In addition, a vest was inflated around the dogs’ chest so as to bring their normally elevated chest-wall compliance to more human-like values. This work was the first to clearly demonstrate the ability of the SPV and the

∆Down to detect occult hypovolemia, as well as their very significant correlation with the degree of hemorrhage and with the changes in cardiac output. Other animal studies of graded hemorrhage have produced the same results, namely that the ∆Down increases gradually with each step in the hemorrhage, that during hypovolemia it is the main component of the SPV, and that it decreases back to normal values following restitution of intravascular volume. In one study, where pigs were exsanguinated to a mean arterial pressure (MAP) of 30 mmHg, the

(6)

∆Down was not considered to reflect the degree of hemorrhage better than the MAP and other hemodynamic parameters [40]. However, when the blood pressure is changing rapidly or when it is very low, it is highly recommended that the SPV and the∆Down are expressed as percentages of the systolic pressure value during end-expiration, namely as %SPV and %∆Down [3,6]. If in the above mentioned study [40] the %∆Down had been used rather than the ∆Down in absolute mmHg, then it would have changed much more significantly during the hypotensive period and would have become more significant.

In another experimental study, Pizov et al found that the application of PEEP in normovolemic dogs caused a significant reduction in cardiac output that was associated with significant increases in the SPV and∆Down [38]. The same level of PEEP, however, did not affect cardiac output in hypervolemic dogs with induced myocardial depression, nor did it change the SPV and the ∆Down. Hence the presence of a significant∆Down should prevent the augmentation of PEEP without prior fluid loading or without the application of more advanced hemodynamic monitoring, while the absence of the∆Down means that the expected hemodynamic effects of PEEP will most probably be negligible. In critically ill patients the pulse pressure variation (PPV) values prior to PEEP application were shown to significantly correlate with the PEEP-induced changes in cardiac output, which also correlated with the changes in PPV following PEEP [27].

The sensitivity of the SPV and the∆Down to changes in intravascular volume can also be seen from their response to passive leg raising. In patients with acute circulatory failure the changes in the respiratory-induced PPV during passive leg raising were significantly correlated with changes in stroke volume during passive leg raising and following rapid fluid expansion [47]. Our own data on the effects of passive leg raising on the SPV and ∆Down in patients following induction of anesthesia for cardiac surgery are shown in Table 1. It is important to note that the SPV and∆Down change much more significantly compared to other parameters denoting their surprising sensitivity to changes in effective blood volume.

The SPV and the∆Down have also been found to correlate with other parameters that reflect the volume status, like the intrathoracic blood volume (ITBV) [39], the echocardiographic LVEDA [41, 42], and even the PAOP [48]. However, the main Table 1. Immediate effects of passive leg raising (PLR) in18 patients following induction of anesthesia for cardiac surgery.

Baseline PLR* Change (%)

Cardiac output (l/min) 4.5 ± 1.1^# 5.7 ± 1.1 23

PAOP (mmHg) 12.9 ± 4.5 14.1 ± 4.8 10

SPV (mmHg) 11.3 ± 5.1 5.9 ± 2.4 48

∆Down (mmHg) 7.5 ± 3.7 3.3 ± 2.0 56

*All PLR values significantly different than baseline.^#Mean ± SD

(7)

value of the SPV and∆Down lies in their accuracy as predictors of fluid responsiveness. A number of clinical studies have shown that these parameters have much better correlation to the change in the cardiac output following volume loading, than the CVP and the PAOP [28,42,45,50]. Moreover, an excellent study done by Tavernier et al. in a group of septic patients, found that the∆Down was more sensitive and more specific than both filling pressures and the LVEDA [44]. A

∆Down component of more than 5 mmHg indicated that the stroke volume index would increase in response to a subsequent fluid challenge with positive and negative predictive values of 95% and 93% respectively [44]. One study only, done recently in patients undergoing cardiac surgery, found that the PAOP predicted the response to a fluid bolus better than the SPV and∆Down, while echocardiographic- derived values had no predictive value at all [50]. However most of the volume- loading steps were done in the presence of very low SPV (3–5 mmHg) and∆Down (1–2.6 mmHg) values, denoting reduced baseline fluid responsiveness. Baseline values of SPV and∆Down were however significantly higher in those patients in whom the cardiac output increased following fluid expansion [50].

In humans the reported values for the SPV vary between 7–16 mmHg and between 2–11 mmHg for the∆Down [23, 28, 30, 31, 42–46, 49–50]. This large spectrum of ‘normal’ values is due to a variety of filling conditions and the tidal volumes employed. However, human studies that examined the effects of hemorrhage on the SPV and∆Down found that a decrease of 500 ml (or 10%) in the blood volume resulted in an increase of about 5 mmHg in the SPV and∆Down [43, 45, 46]. Varying degrees of fluid expansion in humans have always shown the SPV to decrease significantly by anywhere from 2.5–10 mmHg [28, 30, 31, 42, 44], while experimental data have repeatedly shown that hypervolemia and/or congestive heart failure were associated with a relatively small SPV value and a practically non-existent∆Down segment [25,36–38]. One has to note that in cases of severe right ventricular failure the lack of∆Down may be due to the RV, and not the LV, being non-responsive to volume loading. Rarely, severe RV failure may be associated with a∆Down, though with no fluid responsiveness. This may be due to further loading of the RV during inspiration, as well as a possible leftward septal shift.

The Inspiratory Increase in Arterial Pressure – ∆Up (Delta Up, DUp)

We have earlier described the∆Up as representing an inspiratory augmentation of the LV stroke volume. In experimental studies the∆Up has been repeatedly shown to increase during hypervolemia and/or congestive heart failure [25, 36, 37, 41]. In addition we have observed that the retransfusion stage that follows significant blood loss was associated with deterioration in LV function, as well as elevated

∆Up [41]. We have previously shown that experimental pharmacological induction of myocardial depression together with volume loading caused the SPV to decrease from 9 to 3 mmHg and the∆Up to increase from 0.6 to 2.7 mmHg, with the∆Up becoming responsible for practically all of the SPV [36]. In critically ill patients the∆Up has been shown to be a frequent, and at times the main, component of the SPV [23, 44]. A recent study has found that among the 31 septic patients studied, 23 had a∆Up (3.8 ± 1.8 mmHg), isolated in 7 cases, and associ-

(8)

ated with∆Down in 16 cases. Twenty-four patients had a ∆Down (8.6± 7.2 mm Hg), isolated in eight cases and associated with a∆Up in 16 cases [23]. According to this study the presence of the∆Up was associated with increased inspiratory pulmonary venous flow, and its absence associated with no change in pulmonary venous flow during inspiration [23].

The fact that fluid administration causes the∆Up to increase may be explained by more lung regions being converted from zone 2 to zone 3 conditions, causing a higher inspiratory increase in pulmonary venous flow [23, 24]. This is nicely demonstrated by the following case (Fig. 3), where fluid expansion had eliminated the∆Down and caused the ∆Up to significantly increase [44]. The fact that all these changes occurred without any concomitant change in the airway pressure, and the observation that the∆Up is associated with a significant increase in pulmonary venous flow during lung inflation, does not support the argument that the∆Up is mainly due to a transmitted airway pressure [26–28]. In addition to the lungs containing more blood, external pressure on the heart by the inflating lungs and a possible contribution of the decrease in afterload should also be considered as contributing to the∆Up in these conditions. It is of interest to note that in complete open-chest conditions there is normally a small∆Up that is most probably due to squeezing of pulmonary blood volume as well. It may well be that under these conditions the∆Up is indicative of the LV being fluid-responsive, but this has to be studied further.

The fact that the∆Up does not normally reflect fluid responsiveness and yet can be so significant has some very important implications. The first is that simple eyeballing of the arterial pressure fluctuations during mechanical ventilation without relating them to some reference pressure may be misleading. Second, a patient presenting with a prominent∆Up should be considered as being either hypervolemic or as having compromised LV function. The mechanical breath serves as a repetitive ‘assist device’ to the LV in such conditions. Weaning the patient from ventilatory support at this time without improving their cardiovascular function (e.g., diuretics, inotropes or afterload reduction) is probably not advisable. The last Fig. 3. A. Left panel depicts SPV of 15 mmHg in a critically ill patient, most of it being due to a

∆Down. B. Right panel depicts same patient after the administration of 1000 ml of plasma expander. The ∆Up increased to 6 mmHg comprising nearly all the SPV. From [46] with permission

(9)

and not least implication of the presence of a prominent∆Up is that the interpretation of the SPV, stroke volume variation (SVV), and PPV as parameters of fluid responsiveness should be done with caution, as these parameters include a component (∆Up) that is not directly related to fluid responsiveness.

The Pulse Pressure Variation and the Stroke Volume Variation

The PPV is the difference between the maximal and minimal pulse pressures during the mechanical breath cycle divided by the mean of these two values [27, 28]. The rationale of using the PPV rather than the SPV as a parameter of fluid responsiveness is that the pulse pressure is directly related to LV stroke volume and that it is not influenced by any transmission of pleural pressure since such transmission would affect the systolic and diastolic pressure to the same degree.

The PPV has indeed been shown to be an excellent predictor of fluid responsiveness during the application of PEEP [27] and in septic patients [28]. The application of PEEP increased PPV from 9±7% to 16±13% [27] while fluid loading in septic patients caused it to decrease from 14±10% to 7±5% [28]. A PPV value of 13% allowed discrimination between responders (increase in cardiac index [CI]=15%) and non-responders with a sensitivity of 94% and a specificity of 96%, which was somewhat better than that of the SPV, but much better than that of the PAOP and the CVP [28]. It is interesting to note that in another study, a similar threshold value of 12% in the respiratory variability of the aortic flow allowed discrimination between responders and non-responders with a sensitivity of 100% and a specificity of 89%, while the LVEDA index was not significantly different between the two groups [32]. Another recent study, examining the respiratory variations in the preejection period, has shown the PPV to correlate better with the change of stroke volume following fluid load than the SPV and even the

∆Down [49]. Theoretically, the performance of the PPV may be improved even further, especially in the presence of a significant isolated∆Down, if the difference between the maximal and minimal pulse pressure values would be related not to the mean of these two values, but rather to the pulse pressure during end- expira- tion or apnea. In the presence of a large∆Up, the PPV, like the SPV and SVV, will be less effective in predicting fluid responsiveness. In rare cases of severe RV failure, a considerable pulse pressure may be associated with lack of fluid responsiveness, as noted earlier.

Measuring the respiratory variation of the stroke volume itself rather than that of surrogate parameters has become possible with the renewed introduction of pulse contour analysis in the PiCCO monitor (Pulsion Medical Systems, Munich, Germany). The SVV is the difference between the maximal and minimal stroke volume during one mechanical breath, divided by the mean stroke volume value.

The SVV has been shown to be a sensitive indicator of fluid responsiveness in anesthetized patients and to correlate well with the changes in cardiac output following volume loading [30, 51–54]. In healthy patients undergoing neurosur- gery, a SVV value of 9.5% was found to predict a positive (=5%) increase in cardiac output in response to only 100 ml of plasma expander with 79% sensitivity and 93%

specificity [51]. In patients with normal and impaired cardiac function undergoing

(10)

cardiac surgery, the same SVV value of 9.5% was found to predict a positive (=5%) increase in cardiac output in response to a much larger fluid load with a sensitivity of 71–79% and a specificity of 80–85% [53]. It is important to note that in this study the SVV was a better predictor of fluid responsiveness in patients with normal cardiac function than in patients with a low preoperative ejection fraction and higher intraoperative LV end-diastolic dimensions [53]. The most probable expla- nation for this difference is that the patients with impaired cardiac function may have had a characteristically higher∆Up, and that the prominence of the ∆Up had resulted in a lesser predictive ability of the SVV, as mentioned earlier, since the

∆Up is not directly related to fluid responsiveness.

Another study in cardiac patients has found no significant relationship between baseline SVV values and the percentage of change in the cardiac output following volume loading [55]. These results are indeed surprising in view of the fact that there was a very significant correlation between baseline SVV values and the change in SVV after volume loading [55]. In an accompanying editorial [11], and in another [9] that accompanied the studies by Reuter et al [52], Pinsky has repeatedly claimed that the use of the SVV for clinical decision-making cannot be recommended. As correctly pointed out by Pinsky, the SVV measured by the pulse contour method has not been validated on a beat-by-beat basis against a ‘gold-standard’ measurement of stroke volume. However, the SVV has been shown in the meantime to correlate extremely well with the SPV [29,30], attesting to the robustness of the algorithm used in the PiCCO monitor. The clinical results that have been published Fig. 4. Fluid loading (arrow) causing immediate increase in the continuous CO (PCCO, upper panel) and a simultaneous decrease in SVV (lower panel). Trended PiCCO parameters recorded on the MVICU patient data management system (iMDsoft, Israel). Courtesy of Dr Eran Segal, Sheba Medical Center, Tel Aviv, Israel.

(11)

so far, and our own clinical experience (Fig. 4), justify the clinical use of this parameter within its limitations.

The respiratory variations of other pulsatile parameters have been suggested to reflect fluid responsiveness in the same manner as the SPV, PPV, and SVV. These include the ventilation-induced changes in the pulse oximeter’s plethysmographic waveform [46], in the collapsibility of the superior vena cava [17], in the aortic blood flow velocity [31, 32] and the aortic velocity-time integral [33], and in the preejection period [49]. All these parameters were shown to reflect the status of fluid responsiveness better than existing ‘preload’ parameters.

Systolic Pressure Variations

during Programmed Positive-Pressure Ventilation

The change in arterial pressure following a mechanical breath is dependent not only on the status of fluid responsiveness but also on the magnitude of the tidal volume itself [37, 54], since larger increases in intrathoracic pressure will reduce venous return to a greater extent. In view of this fact and of the limitations of current functional hemodynamic monitoring (see below), we have developed the Respiratory Systolic Variation Test (RSVT) which is a ventilatory maneuver that is composed of three consecutive incremental pressure-controlled (10, 20, and 30 cmH2O) breaths (Fig. 5) [56–58]. This maneuver would normally produce respec-

Fig. 5. Upper panel shows the slope (line of best fit) connecting the three respective lowest systolic pressures following each of the three consecutive incremental breaths (10, 20, 30 cmH2O) of the RSVT maneuver (lower panel). Upper panel – arterial pressure Lower panel – airway pressure

(12)

tive incremental decreases in the venous return and hence in the LV stroke volume and in the systolic arterial pressure. Plotting the respective three lowest systolic pressure values (one after each breath respectively) against their respective airway pressure produces a slope that reflects fluid responsiveness and that is termed the RSVT slope.

In dogs subjected to graded hemorrhage, the RSVT slope increased significantly from 0.12±0.09 to 0.79±0.62 mmHg/cmH2O after 30% removal of blood volume, and showed an overall significant correlation (r = -.81) to the stroke volume index [56]. In 14 patients who underwent vascular surgery, a baseline RSVT slope of 0.30

± 0.18 decreased to 0.10 ± 0.09 mmHg/cmH2O (p<.001) after volume loading, and correlated significantly with the LVEDA (r = -0.817) and with the changes in stroke volume index following volume loading (r = 0.8492) [57]. An RSVT value of≥0.24 mmHg/cmH2O predicted a change of 15% in CI with a sensitivity of 87.5% and a specificity of 83%, while a value≥0.34 mmHg/cmH2O predicted a change of CI of 20% with a sensitivity of 83% and a specificity of 75% (Perel A et al, unpublished data). In another study in dogs with induced LV failure, the RSVT slopes correlated significantly (r = 0.7154) with the slopes of actual LV function curves [58]. In patients undergoing cardiac surgery, the RSVT slope compared favorably with other functional hemodynamic parameters, which as a group performed better than ‘preload’ indices (Fig. 6, Preisman et al, unpublished data).

Fig. 6. Receiver-operating characteristics curves comparing various functional hemodynamic and preload parameters in patients undergoing cardiac surgery. The respiratory systolic variation test (RSVT) slope is in the left upper corner. (Preisman et al., unpublished data) PPV: pulse pressure variation; SVV: stroke volume variation; CVP: central venous pressure; ITBV: intrathoracic blood volume; BP: blood pressure; EDAI: end-diastolic area index

(13)

The RSVT has the potential to become the preferred parameter of fluid responsiveness since it is highly standardized, it is not influenced by the∆Up and hence reflects fluid responsiveness only, it is potentially less influenced by changing compliance, and it produces a slope that correlates with the slope of the LV function curve, which is often mentioned but only rarely measured. Further studies are needed to establish the effectiveness of this parameter that is produced by a true automated linkage of the ventilator and the monitor.

Limitations of Functional Hemodynamic Parameters

The main limitation of functional hemodynamic parameters is that their use is limited to patients who are on fully controlled mechanical ventilation. In patients who are breathing spontaneously or on partial ventilatory support, quantification of the respiratory changes in pulsatile parameters may be inaccurate and difficult to interpret. Other potential inaccuracies may be due to the lack of stand- ardization of the magnitude of the tidal volume employed, an exaggerated variation in the presence of large tidal volumes and too little variation when low tidal volumes are being used [59]. Exaggerated variations can also be seen in the presence of air-trapping or reduced chest wall compliance [6, 14]. Decreased lung compliance by itself should not affect the usefulness of the SPV and its derivatives if the tidal volume is unchanged, since the effects of increased airway pressure and its reduced transmission may cancel each other out. In fact some of the major clinical studies on functional hemodynamic parameters have been done in patients who were in respiratory failure [23, 27, 28, 44].

Since functional hemodynamic parameters rely on individually measured beats, any arrhythmias may cause significant inaccuracies. Nodal rhythm, however, may increase the SPV by effectively decreasing preload due to the loss of the ‘atrial kick’.

As mentioned before, the SPV, SVV, and PPV include the∆Up, a component that is unrelated to fluid-responsiveness and that may reduce their ability to accurately reflect fluid responsiveness. This may occur especially when they are in mid-range, since an SPV of 5–6 mmHg may be composed of only a∆Up, only a ∆Down or a combination of both. The∆Down has therefore a theoretical advantage in that it is a parameter of fluid responsiveness only. The measurement of the∆Down (and the

∆Up) necessitates however a long end-expiratory pause or the introduction of a short apnea followed by a careful analysis of the effects of the succeeding (or preceding) mechanical breath on the arterial pressure waveform. The apnea should preferably be achieved without disconnecting the patient from the ventilator, so as not to lose the prevailing PEEP or auto-PEEP. Some attempts have been made to automate the measurement of the SPV and even define automatically the reference systolic blood pressure measured at end-expiration for the automatic measurement of ∆Up and ∆Down [60–62], but the automated measurement of these parameters is still unavailable in commercial monitors. The SVV and PPV are however measured automatically in the PiCCO and the LidCo monitors.

(14)

Conclusion

We are often confronted with a variety of static parameters that do not provide a conclusive picture. Challenging the system with a standardized stimulus may provide new insights about the function of the whole system. The normal effects of this stimulus have to be well known, so that interpretation of the response to this stimulus is clear and preferably immediate. Confounding factors may decrease the usability of this approach.

These are general guidelines for the use of any diagnostic or therapeutic functional test, and directly apply to the use of the increase in intrathoracic pressure as a repetitive challenge of the circulation. In order to be used and interpreted correctly one must have a basic knowledge of the normal physiology of heart-lung interaction during mechanical ventilation. The resulting functional hemodynamics parameters can be of great value in the monitoring of ventilated patients, in which hemodynamic uncertainty and potential instability are often present.

Since the main hemodynamic effect of the mechanical breath is normally a decrease in venous return, the hemodynamic response to the transient decrease in preload (the∆Down) reflects the degree of fluid-responsiveness. For the SPV, PPV, and SVV parameters, values above 10–13% indicate, with very high sensitivity and specificity, that fluid loading will cause an increase in cardiac output. Since all these parameters are affected by the early inspiratory augmentation of LV stroke volume, their performance can be further standardized and improved by automated respiratory maneuvers in which fluid responsiveness alone is being analyzed, like the RSVT. Besides supplying an immediate estimation of fluid responsiveness, these parameters are extremely sensitive to changes in preload, and therefore are useful in following the response to fluid loading.

Since the normal healthy heart is fluid responsive, the presence of fluid responsiveness is not an indication by itself to administer fluids. In addition functional hemodynamic parameters do not offer an answer to the dilemma of cardiovascular

‘optimization’. However, by being able to detect occult hypovolemia, identify the presence of fluid responsiveness or its absence in low-flow states, and reflect the response to changes in effective blood volume, these parameters offer immediate, dynamic, and essential information about the cardiovascular function.

References

1. Michard F, Teboul JL (2002) Predicting fluid responsiveness in ICU patients: a critical analysis of the evidence. Chest 121:2000–2008

2. Michard F, Teboul JL (2000) Using heart-lung interactions to assess fluid responsiveness during mechanical ventilation. Crit Care 4:282–289

3. Perel A (1991) Cardiovascular assessment by pressure waveform analysis. In Annual Re- fresher Course Lectures, American Society of Anesthesiologists

4. Rooke GA (1995) Systolic pressure variation as an indicator of hypovolemia. Curr Opin Anesthesiol 8:511–515

5. Stoneham MD (1999) Less is more…using systolic pressure variation to assess hypovolemia.

Br J Anaesth 83:550–551

(15)

6. Perel A (1998) Assessing fluid responsiveness by the systolic pressure variation in mechanically ventilated patients. Anesthesiology 89:1309–1310

7. Gunn SR, Pinsky MR (2001) Implications of arterial pressure variation in patients in the intensive care unit. Curr Opin Crit Care 7:212–2178.Pinsky MR (2001) Functional hemodynamic monitoring: applied physiology at the bedside. In: Vincent JL (ed) Yearbook of Intensive Care and Emergency Medicine, Springer Verlag, pp 537–552

9. Pinsky MR (2002) Functional hemodynamic monitoring. Intensive Care Med 28:386–8 10. Michard F, Teboul JL (2002) Detection of fluid responsiveness. In: Vincent JL (ed) Yearbook

of Intensive Care and Emergency Medicine, Springer Verlag, pp 553–563

11. Pinsky MR (2003) Probing the limits of arterial pulse contour analysis to predict preload responsiveness. Anesth Analg 96:1245–1247

12. Bendjelid K, Romand JA (2003) Fluid responsiveness in mechanically ventilated patients: a review of indices used in intensive care. Intensive Care Med 29:352–360

13. Perel A (2003). The value of functional hemodynamic parameters in hemodynamic monitoring of ventilated patients (Editorial). Anaesthetist 52:1003–4

14. Magder S (2004) Clinical usefulness of respiratory variations in arterial pressure. Am J Respir Crit Care Med 169: 151–155

15. Fukamachi K, Irie H, Massiello A, et al (1992). Effects of mechanical ventilation and sponta- neous respiration on hemodynamics in calves with total artificial hearts. ASAIO J 38:M493–496

16. Jardin F, Farcot JC, Gueret P, Prost JF, Ozier Y, Bourdarias JP (1983) Cyclic changes in arterial pulse during respiratory support. Circulation 68:266–274

17. Vieillard-Baron A, Augarde R, Prin S, Page B, Beauchet A, Jardin F (2001) Influence of superior vena caval zone conditions on cyclic changes in right ventricular outflow during respiratory support. Anesthesiology 95:1083–1088

18. Sellgren J, Soderstrom S, Johansson G, Biber B, Haggmark S, Ponten J (2003) Preload changes by positive pressure ventilation can be used for assessment of left ventricular systolic function. Acta Anaesthesiol Scand 47:541–548

19. Van den Berg PCM, Jansen JRC, Pinsky MR (2002) Effect of positive pressure on venous return in volume loaded cardiac surgical patients. J Appl Physiol 92: 1223–1231

20. Vieillard-Baron A, Loubieres Y, Schmitt JM, Page B, Dubourg O, Jardin F (1999) Cyclic changes in right ventricular output impedance during mechanical ventilation. J Appl Physiol 87:1644–1650

21. Massumi RA, Mason DT, Zakauddin V, Zelis R, Otero J, Amsterdam EA (1973). Reversed pulsus paradoxus. N Eng’ J Med 289:1272–1275

22. Robotham JL, Cherry D, Mitzner W, Rabson JL, Lixfeld W, Bromberger-Barnea B (1983) A re-evaluation of the hemodynamic consequences on intermittent positive pressure ventilation. Crit Care Med 11:783–793

23. Vieillard-Baron A, Chergui K, Augarde R, et al (2003) Cyclic changes in arterial pulse during respiratory support revisited by Doppler echocardiography. Am J Respir Crit Care Med 168:671–676

24. Brower R, Wise RA, Hassapoyannes C, Bromberger-Barnea B, Pemutt S (1985) Effect of lung inflation on lung blood volume and pulmonary venous flow. J Appl Physiol 58:954–963 25. Perel A, Pizov R, Cotev C (1987) Systolic blood pressure variation is a sensitive indicator of

hypovolemia in ventilated dogs subjected to graded hemorrhage. Anesthesiology 67:498–502 26. Denault AY, Gasior TA, Gorscan J, Mandarino WA, Deneault LG, Pinsky MR (2000) Deter- minants of aortic pressure variation during positive pressure ventilation in man. Chest 116:176–186

27. Michard F, Chemla D, Richard C, et al (1999) Clinical use of respiratory changes in arterial pulse pressure to monitor the hemodynamic effects of PEEP. Am J Respir Crit Care Med 159:935–939

(16)

28. Michard F, Boussat S, Chemla D, et al (2000) Relation between respiratory changes in arterial pulse pressure and fluid responsiveness in septic patients with acute circulatory failure. Am J Respir Crit Care Med 62:134–138

29. Berkenstadt H, Friedman Z, Preisman S, Perel A (2001) The stroke volume variation (SVV) correlates with the stroke volume and the systolic pressure variation (SPV) during hemorrhage and retransfusion in dogs. Intensive Care Med 27:S258 (abst)

30. Reuter D, Felbinger TW, Kilger F, Schmidt C, Lamm P, Goetz AE (2002) Optimizing fluid therapy in mechanically ventilated patients after cardiac surgery by on-line monitoring of left ventricular stroke volume variations: a comparison to aortic systolic pressure variations.

Br J Anaesth 88:124–126

31. Beaussier M, Coriat P, Perel A, et al (1995) Determinants of systolic pressure variation after vascular surgery. J Cardiothoracic Vasc Anesth 9:547–551

32. Feissel M, Michard F, Mangin I, Ruyer O, Faller JP, Teboul JL (2001) Respiratory changes in aortic blood velocity as an indicator of fluid responsiveness in ventilated patients with septic shock. Chest 119:867–873

33. Slama M, Masson H, Teboul JL, et al (2002) Respiratory variations of aortic VTI: a new index of hypovolemia and fluid responsiveness. Am J Physiol Heart Circ Physiol 283:H1729–1733 34. Coyle JP, Teplick RS, Long MC (1983) Respiratory variations in systemic arterial pressure as

an indicator of volume status. Anesthesiology 59:A53 (abst)

35. Pizov R, Yaari Y, Perel A (1988) The systolic pressure variation is greater during hemorrhage than during sodium- nitroprusside induced hypotension in ventilated dogs. Anesth Analg 67:170–174

36. Pizov R, Ya’ari Y, Perel A (1989) The arterial pressure waveform during acute ventricular failure and synchronized external chest compression. Anesth Analg 68:150–156

37. Szold A, Pizov R, Segal E, Perel A (1989) The effect of tidal volume and intravasculer volume state on systolic pressure variation in ventilated dogs. Intensive Care Med 15:368–371 38. Pizov R, Cohen M, Weiss Y, Segal E, Cotev S, Perel A (1996) Positive end-expiratory pressure-

induced hemodynamic changes are reflected in the arterial pressure waveform. Crit Care Med 24:1381–1387

39. Preisman S, Pfeiffer U, Lieberman N, Perel A (1997) New monitors of intravascular volume:

a comparison of arterial pressure waveform analysis and the intrathoracic blood volume.

Intensive Care Med 23:651–657

40. Dalibon N, Schlumberger S, Saada M, Fischler M, Riou B (1999) Hemodynamic assessment of hypovolemia under general anesthesia in pigs submitted to graded hemorrhage and retransfusion. Br J Anaesth 82:97–103

41. Preisman S, DiSegni E, Vered Z, Perel A (2002) Left ventricular preload and function during graded hemorrhage and retranfusion in pigs: analysis of arterial pressure waveform and correlation with echocardiography. Br J Anaesth 88:716–718

42. Coriat P, Vrillon M, Perel A, et al (1994) A comparison of systolic blood pressure variations and echocardiographic estimates of end-diastolic left ventricular size in patients after aortic surgery. Anesth Analg 78:46–53

43. Rooke GA, Schwid HA, Shapira Y (1995) The effect of graded hemorrhage and intravascular volume replacement on systolic pressure variation in humans during mechanical and spon- taneous ventilation. Anesth Analg 80:925–932

44. Tavernier B, Makhotine O, Lebuffe G, Dupont J, Scherpereel P (1998) Systolic pressure variation as a guide to fluid therapy in patients with sepsis-induced hypotension. Anesthesi- ology 89:1313–1321

45. Ornstein E, Eidelman LA, Drenger B, Elami A, Pizov R (1998) Systolic pressure variation predicts the response to acute blood loss. J Clin Anesth 10:137–140

46. Shamir M, Eidelman LA, Floman Y, Kaplan L, Pizov R (1999) Pulse oximetry plethysmographic waveform during changes in blood volume. Br J Anaesth 82:178–181

(17)

47. Boulain T, Achard JM, Teboul JL, Richard C, Perrotin D, Ginies G (2002) Changes in blood pressure induced by passive leg raising predict response to fluid loading in critically ill patients. Chest 121:1245–1252

48. Marik PE (1993) The systolic blood pressure variation as an indicator of pulmonary capillary wedge pressure in ventilated patients. Anaesth Intensive Care 21:405–408

49. Benjelid K, Suter PM, Romand JA (2004) The respiratory change in preejection period: a new method to predict fluid responsiveness. J Appl Physiol 96:337–342

50. Bennett-Guerrero E, Kahn RA, Moskowitz DM, Falcucci O, Bodian CA (2002) Comparison of arterial systolic pressure variation with other clinical parameters to predict the response to fluid challenges during cardiac surgery. Mt Sinai J Med 69:96–100

51. Berkenstadt H, Margalit N, Hadani M, et al (2001) Stroke volume variation as a predictor of fluid responsiveness in patients undergoing brain surgery. Anesth Analg 92:984–989 52. Reuter DA, Felbinger TW, Schmidt C, et al (2002) Stroke volume variations for assessment of

cardiac responsiveness to volume loading in mechanically ventilated patients after cardiac surgery. Intensive Care Med 28:392–398

53. Reuter DA, Kirchner A, Felbinger TW, et al (2003) Usefulness of left ventricular stroke volume variation to assess fluid responsiveness in patients with reduced cardiac function. Crit Care Med 31:1399–1404

54. Reuter DA, Bayerlein J, Goepfert MS, et al (2003) Influence of tidal volume on left ventricular stroke volume variation measured by pulse contour analysis in mechanically ventilated patients. Intensive Care Med 29:476–480

55. Wiesenack C, Prasser C, Rodig G, Keyl C (2003) Stroke volume variation as an indicator of fluid responsiveness using pulse contour analysis in mechanically ventilated patients. Anesth Analg 96:1254–1257

56. Perel A, Preisman S, Baer R, Shneider A (1995) Respiratory Systolic Variation Test reflects preload during graded hemorrhage in ventilated dogs. Br J Anaesth 74 (Suppl 1):A134 (abst) 57. Perel A, Minkovich L, Abiad M, Coriat P, Viars P (1995) Respiratory systolic variation test -

A new method for assessing preload. Br J Anaesth 74 (Suppl 1):A137 (abst)

58. Perel A, Baer R, Minkovich L (1996) Respiratory systolic variation test reflects the slope of the LV function curve during LV failure in ventilated dogs. Br J Anaesth 76 (Suppl 2):39 (abst) 59. De Backer D (2003) Stroke volume variations. Minerva Anestesiol 69:285–288

60. Schwid HA, Rooke GA (2000) Systolic blood pressure at end-expiration measured by the automated systolic pressure variation monitor is equivalent to systolic blood pressure during apnea. J Clin Monit Comput 16:115–120

61. Soncini M, Manfred G, Redaelli A, et al (2002) A computerized method to measure systolic pressure variation (SPV) in mechanically ventilated patients. J Clin Monit Comput 17:141–146

62. Morelot-Panzini C, Lefort Y, Derenne JP, Similowski T (2003) Simplified method to measure respiratory-related changes in arterial pulse pressure in patients receiving mechanical ventilation. Chest 124:665–670