Share this article on:

Comparative study of pressure- and volume-controlled ventilation on pulse pressure variation in a model of hypovolaemia in rabbits

Fonseca, E. B.*; Otsuki, D. A.*; Fantoni, D. T.; Bliacheriene, F.; Auler, J. O. C.

European Journal of Anaesthesiology (EJA): May 2008 - Volume 25 - Issue 5 - p 388–394
doi: 10.1017/S0265021508003554
Original Article

Background and objective Dynamic indices represented by systolic pressure variation and pulse pressure variation have been demonstrated to be more accurate than filling pressures in predicting fluid responsiveness. However, the literature is scarce concerning the impact of different ventilatory modes on these indices. We hypothesized that systolic pressure variation or pulse pressure variation could be affected differently by volume-controlled ventilation and pressure-controlled ventilation in an experimental model, during normovolaemia and hypovolaemia.

Method Thirty-two anaesthetized rabbits were randomly allocated into four groups according to ventilatory modality and volaemic status where G1-ConPCV was the pressure-controlled ventilation control group, G2-HemPCV was associated with haemorrhage, G3-ConVCV was the volume-controlled ventilation control group and G4-HemVCV was associated with haemorrhage. In the haemorrhage groups, blood was removed in two stages: 15% of the estimated blood volume withdrawal at M1, and, 30 min later, an additional 15% at M2. Data were submitted to analysis of variance for repeated measures; a value of P < 0.05 was considered to be statistically significant.

Results At M0 (baseline), no significant differences were observed among groups. At M1, dynamic parameters differed significantly among the control and hypovolaemic groups (P < 0.05) but not between ventilation modes. However, when 30% of the estimated blood volume was removed (M2), dynamic parameters became significantly higher in animals under volume-controlled ventilation when compared with those under pressure-controlled ventilation.

Conclusions Under normovolaemia and moderate haemorrhage, dynamic parameters were not influenced by either ventilatory modalities. However, in the second stage of haemorrhage (30%), animals in volume-controlled ventilation presented higher values of systolic pressure variation and pulse pressure variation when compared with those submitted to pressure-controlled ventilation.

* University of São Paulo, School of Medicine, Laboratory of Medical Investigation LIM/08-Anaesthesiology, São Paulo, Brazil

University of São Paulo, School of Veterinary Medicine, Department of Surgery, São Paulo, Brazil

University of São Paulo, School of Medicine, Hospital das Clínicas, Division of Anaesthesia, São Paulo, Brazil

University of São Paulo, School of Medicine, Hospital das Clínicas, Heart Institute, Anaesthesia Department and Surgical ICU, São Paulo, Brazil

Correspondence to: José Otávio Costa Auler Junior, Instituto do Coração (Incor) – Hospital das Clínicas da FMUSP, Av. Dr. Enéas de Carvalho Aguiar 44, 2° and. CEP 05403-000 São Paulo, Brazil. E-mail:; Tel/Fax: +55 11 3069 5232

Accepted for publication 1 January 2008

First published online 8 February 2008

Back to Top | Article Outline


The concept of fluid responsiveness by using dynamic parameters has become very popular over the last few years, most likely because it provides a pragmatic approach to fluid therapy under different clinical and experimental situations [1–3].

Dynamic parameters have been described and tested in previous studies, which have shown that systolic pressure variation (SPV) is well correlated with left ventricular preload during haemorrhage or intravascular volume loading. SPV was described in a more comprehensive manner by Perel and colleagues [4], through its components, Δup and Δdown, which are obtained through a systolic pressure value of reference after a short period of apnoea. The fundamentals of arterial pressure variation during a mechanical ventilatory cycle were extensively discussed by Michard [5]. In summary, three mechanisms are involved in this phenomenon: (a) right ventricular afterload increases during inspiration due to an increase in alveolar pressure; (b) left ventricular preload increases during inspiration by a balance in pleural pressure, favouring a blood squeezing out of pulmonary capillaries towards the left atrium; and (c) positive pleural pressure increases the systolic extracardiac pressure and decreases the systolic intracardiac pressure through a reduction in thoracic blood volume, causing a decrease in left ventricular afterload.

Compared with spontaneous ventilation, mechanical ventilation alters the phase of the arterial pressure changes, because inspiration is associated with a fall in arterial pressure during spontaneous inspiration, and with a rise during mechanically applied breaths [6].

The magnitude of the arterial pressure changes has been suggested as a clinical way of assessing fluid responsiveness [5]. The effects of mechanical ventilation over arterial pressure are well known; blood pressure increases during the inspiratory phase and decreases during the expiratory one. This occurs by cyclic fluctuation over intrathoracic and transpulmonary pressures generating preload and afterload changes in both ventricles [7]. Such effects are still dependent on baseline cardiovascular state and are more prominent in hypovolaemia [4,8].

Many studies [5] have shown that arterial pulse pressure variation (PPV) induced by mechanical ventilation is much more accurate than cardiac filling pressures and volumetric markers of preload in predicting fluid responsiveness (i.e. the haemodynamic effects of volume loading). PPV is also more reliable than other well-known dynamic parameters such as systolic pressure variation [9,10] or pulse contour stroke volume variation [10]. However, the behaviour of dynamic parameters under different ventilatory modes has not been tested.

Volume-controlled ventilation (VCV) ensures the delivery of a defined tidal volume, producing a square waveform flow, but high airway pressures can be achieved in low pulmonary compliance states [11]. On the other hand, pressure-controlled ventilation (PCV) requires strict monitoring of the tidal volume because it is not an adjusted parameter as in VCV, but is dependent on the interaction between the patient's respiratory mechanics and applied peak inspiratory pressure (PIP). Gas flow is free in the inspiratory phase and decreases in a logarithmic manner when airway pressure achieves the adjusted value. The limitation in plateau pressure is a theoretical advantage of this mode of ventilation, which promotes fewer ventilation-induced lung injuries and a more homogeneous tidal volume distribution [11].

As mechanical ventilation is frequently used in clinical practice in patients who need to be tested for fluid responsiveness, we have two questions: first, are dynamic parameters herein represented by SPV and PPV affected by the most commonly utilized ventilatory modalities such as PCV and VCV? Second, if it occurs, under which conditions?

Back to Top | Article Outline


This study was approved by Ethical Committee in Animal Research of School of Medicine, University of São Paulo. All institutional protocols recommended to handle animals under experimentation were rigidly observed.

Back to Top | Article Outline

Anaesthesia and animal preparation

Thirty-two female New Zealand rabbits (body weight 2.71 ± 0.21 kg) were premedicated with acepromazine (0.3 mg kg−1) and meperidine (10.0 mg kg−1) intramuscularly before admission into the laboratory. Anaesthesia induction was done with propofol (10 mg kg−1 intravenous). In the same venous access (auricular vein), continuous maintenance fluid was administered (NaCl 0.9%) at a rate of 7 mL kg−1 h−1 (Anne® infusion pump; Abbott Laboratories, North Chicago, IL, USA), together with pancuronium (0.1 mg kg−1) boluses every 40 min to avoid inspiratory drive. A neuro-muscular activity monitor (TOF Watch-SX; Organon Teknika, Oss, The Netherlands) was used to ensure complete muscle paralysis and the absence of any inspiratory effort. Tracheotomy was performed and a 2.5 mm internal diameter tube was inserted in the trachea. Isoflurane (0.5 MAC) and oxygen (1.0 inspired fraction) were used for maintenance of anaesthesia. Mechanical ventilation and anaesthesia were obtained with Cicero EM (Dräger, Lübeck, Germany). Body temperature was maintained with a warmed blanket (TP-500 T/Pump; Gaymar Industries, Orchard, NY, USA). Great care was taken to avoid any leak of air around the intratracheal cannula, as well as to ensure its correct positioning and the correct ventilation of both lungs.

Back to Top | Article Outline

Monitoring and data collection

Respiratory parameters were adjusted in the ventilator to keep a tidal volume of 10–12 mL kg−1 and normocapnia and inspiratory-to-expiratory ratio was set at 1 : 2. These parameters were monitored by a polygraph (Gould Inc. Instruments Division, Cleveland, OH, USA) integrated to a personal computer and a software (BioBench; National Instruments, Austin, TX, USA). Inspiratory pressure was obtained with an airway pressure transducer and flow was measured with a pneumotachograph (Hans Rudolph, KS, MO, USA) to ensure the same tidal volume during either ventilatory modalities. Dynamic compliance (mL cmH2O−1) was calculated as

Back to Top | Article Outline

Pleural pressure

In order to assess pleural pressure under different ventilatory modalities, a catheter with multiple holes in its extremity was inserted in the thoracic lateral region to gain pleural space. The catheter was filled with normal saline and connected to a transducer zeroed at atmospheric pressure and positioned at the heart level. The transducer was connected to a polygraph, which continuously registered the pleural pressure tracing. The skin around the catheter insertion was carefully sealed to avoid leaks, and a simple X-ray was taken to assure the absence of pneumothorax and to verify whether the catheter in the pleural cavity was properly located without any leaks.

Back to Top | Article Outline

Arterial pressure acquisition

Both auricular arteries had a catheter inserted to monitor arterial pressure and to collect/withdraw blood. The right one was connected to a pressure transducer and a polygraph and the left one to another pressure transducer and a multiparameter monitor (Viridia CMS; Hewlett Packard, Palo Alto, CA, USA).

SPV, Δup, Δdown and PPV were defined as recommended [4,12] during five consecutives breaths and through polygraph-obtained data.

Cardiac output was measured in triplicate using echoDopplercardiography (Image Point Ultrasound System, S5010; Agilent, Andover, MA, USA; 5 MHz transducer).

To measure central venous pressure, a catheter was inserted into the jugular vein and connected to a pressure transducer of a multiparameter monitor. The correct position was checked observing the classical tracings in the screen monitor as well as by means of X-ray, being assured that the tip filled with iodine contrast was inside the right atrium.

Back to Top | Article Outline

Experimental protocol

After the instrumentation of the animals, they were randomly allocated in one of the established groups: two control groups; one ventilated in PCV (G1-ConPCV) and the other in VCV (G3-Con VCV); and two submitted to graded haemorrhage, ventilated in PCV (G2-HemPCV) and VCV (G4-HemVCV). In the haemorrhage groups, the calculated blood volume was drawn within 10 min, through the left auricular artery, in steps of 15% of the estimated blood volume (65 mL kg−1) at M1 and another 15% at M2. Three moments were evaluated: M0, 30 min after anaesthesia stabilization; M1, 30 min after the end of M0; and M2, 30 min after the end of M1. In the haemorrhage groups, the measurements were coincident: M0, 30 min after anaesthesia; M1, after blood withdrawal (15%); and M2, after the second step of haemorrhage.

Back to Top | Article Outline

Statistical analysis

Data were submitted to analysis of variance for repeated measures (ANOVA) with Tukey's post hoc test; significance level was P < 0.05 and results were expressed as mean ± SD.

Back to Top | Article Outline


Baseline haemodynamic and respiratory parameters were similar among groups and did not change with time in control ones.

Haemodynamic changes are summarized in Table 1. Heart rate (HR) increased significantly (208 ± 38 to 260 ± 29 beats min−1) in G4-HemVCV after blood withdrawal, and this value was higher than any value documented in the control groups. Mean arterial pressure and cardiac output did not change at hypovolaemic states. A decrease in central venous pressure was observed in hypovolaemic animals: G2-HemPCV (3.6 ± 1.2 to 2.8 ± 1.0 and 1.6 ± 0.9 cmH2O); G4-HemVCV (3.2 ± 1.0 to 2.4 ± 1.3 and 1.8 ± 0.7 cmH2O), but no difference was observed between these groups.

Table 1

Table 1

SPV, Δdown (Table 2) and PPV (###Fig. 1) increased with haemorrhage. This variation was more evident and statistically significant in G4-HemVCV.

Table 2

Table 2

Figure 1

Figure 1

Respiratory parameters such as tidal volume, respiratory rate, pleural pressure and dynamic compliance were not different among groups or evaluated moments (Table 3).

Table 3

Table 3

Back to Top | Article Outline


The main findings of this study were that in the first phase of haemorrhage (15%), a significant variation on arterial pressure dependent on mechanical ventilation was observed, represented by SPV, Δdown and PPV, when compared with the control groups. However, no difference was observed between modes of ventilation. After the second phase of haemorrhage, when it was possible to characterize severe hypovolaemia, animals under VCV exhibited higher SPV, Δdown and PPV, when compared with those under PCV. We could speculate that the pattern of inspiratory flow, typical of PCV that theoretically causes a smaller peak in PIP and a consequently lower impact in intrathoracic pressure, would be responsible for keeping PPV relatively unaltered even in hypovolaemic states. The strict control of pleural pressure during the study, which did not present any difference among groups, as well as the same tidal volume for both ventilatory modes rigorously assured by the pneumotachograph, associated with lack of change in respiratory compliance, may support this hypothesis. In fact, these results herein represented by high levels of PPV in the VCV mode that was particularly enhanced during haemorrhage allowed us to infer that, in shock states, the typical inspiratory flow generated during PCV did not move PPV from the flat to the steeper position in the Starling curve. Could PCV be a better choice as ventilatory mode in hypovolaemic states, or could this mode of ventilation keep the tool of fluid responsiveness undisclosed? The steadiness of PPV suggests that haemodynamics under PCV were relatively preserved, and necessity for fluid loading during the haemorrhage phase was not revealed. In contrast, measurements made under VCV could uncover the actual hypovolaemic state, where the real haemodynamic state was portrayed by high values of PPV. According to the literature [5], the elevation of PPV characterizes a state of positive fluid responsiveness to be corrected with fluids if necessary. The question that arises from such a result is whether the employment of PCV is better in hypovolaemic states because it provokes less haemodynamic interference or is VCV more appropriate because the need for fluid therapy could be uncovered. By stirring up cyclic modifications in intrathoracic and transpulmonary pressures, mechanical ventilation provokes recurring variations in the preload and the afterload of both ventricles. The resulting changes in stroke volume are reproduced by pulse pressure variation over the respiratory cycle, which is influenced, as shown by the literature, by the value of tidal volume [13–15], PEEP levels [12] and probable pattern in inspiratory flow according to the ventilatory modality.

In our study, the comparison of different indices of cardiac preload demonstrated that SPV, Δdown and PPV were able to detect hypovolaemia as shown in other studies [4,8] and to distinguish modes of ventilation after severe hypovolaemia. On the other hand, central venous pressure detected hypovolaemia but was not able to distinguish modes of ventilation in haemorrhage groups. Cardiac output measured by Doppler echocardiography did not decrease after blood withdrawal probably because compensatory mechanisms were activated, especially the increase in HR.

An unexpected result of our study was the absence of significant changes in pleural pressure between the modes of ventilation. We initially hypothesized that the increase in PPV under VCV could be determined by higher levels of PIP and, consequently, intrathoracic pressure. However, the values of pleural pressure measured under two different modalities were similar. A reasonable explanation could be related to the place where the pleural catheter was located. We know, from physiology, that pleural pressure varies along the thorax, being higher around the juxtacardiac pleural space [16]. It is possible that, in the position where the catheter was located, laterally in the thorax with the transducer zeroed to the same level of the heart, the real impact of the ventilatory mode upon pleural pressure could not be detected. Many factors contribute to stroke volume and arterial pressure variations during the respiratory cycle; some authors consider pleural pressure to be of paramount importance. Respiration affects the circulation by changing pleural pressure (Ppl) and thus the relationship of intrathoracic structures with extrathoracic structures [17], leading to a change in right and left ventricular loadings, which could be relevant in determining of SPV [18–20].

According to Magder [16], the change in Ppl may be the foremost determinant of SPV and, consequently, of PPV. This assertion proposed by Magder is based on studies by Scharf and colleagues [21] and the presentation by Denault and colleagues [22] that opening the chest significantly reduces SPV. Should this hypothesis be accepted, our data suggest that VCV seems to interfere in pleural pressure dynamics by means of its higher PIP and should be, therefore, responsible for the changes in PPV. However, this expected influence in pleural pressure during VCV was not detected by our methodological approach.

The importance of Ppl is controversial. According to Michard, PPV is not affected by variations in pleural pressure. This can be explained because pleural pressure (i.e. in aortic extramural pressure) affects both systolic and diastolic pressures; therefore, PPV depends mainly on changes in aortic transmural pressure [23]. The arterial pulse pressure is directly proportional to stroke volume [24] and cyclic changes in stroke volume during a mechanical breath are reflected by proportional changes in pulse pressure [25]. Because it reflects the magnitude of changes in left ventricular stroke volume induced by mechanical inspiration, PPV is a marker of the position on the Frank–Starling curve [5]. Indeed, when the heart is operated on the flat portion of the Frank–Starling curve, cyclic changes in preload induced by mechanical inspiration do not induce any significant changes in stroke volume and pulse pressure. In contrast, when the heart is operating on the steep portion of the curve, cyclic changes in preload are turned into significant variations in stroke volume and pulse pressure. Michard and colleagues [9,12] tried to make SPV more specific by examining respiratory changes in pulse pressure. They calculated the difference between maximum and minimum pulse pressures during the ventilatory cycle and normalized the difference to the average of the maximum and minimum pulse pressure (PPV). Their rationale is based on the fact that pulse pressure is more related to stroke volume than to systolic pressure and should not be affected by the direct transmission of Ppl pressure to the aorta. Based on these fundaments, the explanation to our findings pertaining PPV would be centred not on the variation of pleural pressure but on the displacement of the subject in the Frank–Starling curve to a steeper portion by the VCV, because of its characteristic waveform determined by a higher PIP.

In conclusion, rabbits with normovolaemia and moderate haemorrhage showed similar behaviour of the studied parameters in either modes of ventilation. However, when rabbits were submitted to severe haemorrhage, PPV varied significantly less during PCV when compared with VCV. The explanation of the exact mechanism of these findings is beyond the scope of this study, which was to compare the effects of the two means of ventilation. Nonetheless, it is reasonable to infer that the different inspiratory flow patterns inherent to each employed ventilatory mode may have some contribution. Considering the limitations to extrapolate the results of an experimental study to the clinical scenario, according to the results herein presented, our opinion is that the influence of PCV or VCV on pulse pressure variation should also be investigated in daily medical practice.

Back to Top | Article Outline


The authors thank Mr Gilberto de Mello Nascimento (laboratory technician from LIM-08) and Dr Marcos N. Samano for their help in experimental procedures. This study was supported by FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) number 03/12967-, LIM/08-Anaesthesiology (Laboratory of Medical Investigation) and grants from CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior).

Back to Top | Article Outline


1. Michard F, Lopes MR, Auler JO. Pulse pressure variation: beyond the fluid management of patients with shock. Crit Care 2007; 11: 131.
2. Lopes MR, Auler JOC, Michard F. Volume management in critically ill patients: new insights. Clinics 2006; 61: 345–350.
3. Bliacheriene F, Machado SB, Fonseca EB, Otsuki D, Auler Jr JOC, Michard F. Pulse pressure variation as a tool to detect hypovolaemia during pneumoperitoneum. Acta Anaesthesiol Scand 2007; 51: 1268–1272.
4. Perel A, Pizov R, Cotev S. Systolic blood pressure variation is a sensitive indicator of hypovolemia in ventilated dogs subjected to graded hemorrhage. Anesthesiology 1987; 67: 498–502.
5. Michard F. Changes in arterial pressure during mechanical ventilation. Anesthesiology 2005; 103: 419–428.
6. Pinsky MR. Assessment of indices of preload and volume responsiveness. Curr Opin Crit Care 2005; 11: 235–239.
7. Pinsky MR. The hemodynamic consequences of mechanical ventilation: an evolving story. Intensive Care Med 1997; 23: 493–503.
8. Perel A. The value of delta down during haemorrhage. Br J Anaesth 1999; 83: 967–968.
9. Michard F, Boussat S, Chemla D et al. Relation between respiratory changes in arterial pulse pressure and fluid responsiveness in septic patients with acute circulatory failure. Am J Respir Crit Care Med 2000; 162: 134–138.
10. Preisman S, Kogan S, Berkenstadt H, Perel A. Predicting fluid responsiveness in patients undergoing cardiac surgery: functional haemodynamic parameters including the Respiratory Systolic Variation Test and static preload indicators. Br J Anaesth 2005; 95: 746–755.
11. Davis K, Branson RD, Campbell RS, Porembka DT. Comparison of volume control and pressure control ventilation: Is flow waveform the difference? J Trauma 1996; 41: 808–814.
12. Michard F, Chemla D, Richard C et al. Clinical use of respiratory changes in arterial pulse pressure to monitor the hemodynamic effects of PEEP. Am J Respir Crit Care Med 1999; 159: 935–939.
13. Perel A, Minkovich L, Preisman S, Abiad M, Segal E, Coriat P. Assessing fluid-responsiveness by a standardized ventilatory maneuver: the respiratory systolic variation test. Anesth Analg 2005; 100: 942–945.
14. De Baker D, Heenen S, Piagnerelli M, Koch M, Vincent JL. Pulse pressure variations to predict fluid responsiveness: influence of tidal volume. Intensive Care Med 2005; 31: 517–523.
15. Charron C, Fessenmeyer C, Cosson C et al. The influence of tidal volume on the dynamic variables of fluid responsiveness in critically ill patients. Anesth Analg 2006; 102: 1511–1517.
16. Magder S. Clinical usefulness of respiratory variations in arterial pressure. Am J Respir Crit Care Med 2004; 169: 151–155.
17. Cournand A, Motley HL, Werko Jr L, Richards DW. Physiological studies of the effects of intermittent positive pressure breathing on cardiac output in man. Am J Physiol 1948; 152: 162–174.
18. Robotham JL, Rabson J, Permutt S, Bromberger-Barnea B. Left ventricular hemodynamics during respiration. J Appl Physiol 1979; 47: 1295–1303.
19. Robotham JL, Lixfeld W, Holland L et al. The effects of positive end-expiratory pressure on right and left ventricular performance. Am Rev Respir Dis 1980; 121: 677–683.
20. Bromberger-Barnea B. Mechanical effects of inspiration on heart functions. Fed Proc 1981; 40: 2172–2177.
21. Scharf SM, Caldini P, Ingram RH. Cardiovascular effects of increasing airway pressure in the dog. Am J Physiol 1977; 232: H35–H43.
22. Denault AY, Gastor TA, Goresan JI, Mandarino WA, Deneault LG, Pinsky MR. Determinants of aortic pressure variation during positive pressure ventilation in man. Chest 1999; 116: 176–186.
23. Michard F, Boussat S, Chemla D et al. Relation between respiratory changes in arterial pulse pressure and fluid responsiveness in septic patients with acute circulatory failure. Am J Respir Crit Care Med 2000; 162: 134–138.
24. Chemla D, Hebert JL, Coirault C et al. Total arterial compliance estimated by stroke volume-to-aortic pulse pressure ratio in humans. Am J Physiol 1998; 274: H500–H505.
25. Jardin F, Farcot JC, Gueret P, Prost JF, Ozier Y, Bourdarias JP. Cyclic changes in arterial pulse during respiratory support. Circulation 1983; 68: 266–274.


© 2008 European Society of Anaesthesiology