The influence of positive end-expiratory pressure on stroke volume variation and central blood volume during open and closed chest conditions

Functional residual capacity, the volume of air in the lung at the end of expiration, decreases during induction of anesthesia and mechanical ventilation because of alveolar collapse and closure of airways [1,2]. The loss of respiratory muscle tone and gas resorption seem to be the underlying mechanisms for atelectasis [2,3]. Positive end-expiratory airway pressure (PEEP) prevents alveolar collapse thereby improving functional residual capacity. However, high inspiratory plateau pressures and positive end-expiratory pressures go along with decreased cardiac output (CO) in closed chest situations [4,5], which may outweigh the benefit of increased oxygenation. The optimum PEEP provides the maximum oxygen transport which is when cardiac output and arterial oxygen content are best balanced [6]. PEEP influences cardiac preload and afterload depending on the level of end-expiratory pressure applied [7,8]. Further, the effect of PEEP on cardiac performance depends on chest wall compliance [9]. Opening the thoracic cavity alters chest wall compliance extensively. It was hypothesized that during open chest conditions the adverse hemodynamic effects of intermittent positive pressure ventilation and PEEP are, therefore, less pronounced. Immediately after cardiac surgery, occasionally chest closure has to be postponed because of severe hemodynamic instability [10]. Further, PEEP is applied in this clinical scenario in order to optimize pulmonary function. For guiding fluid therapy in those patients, left ventricular stroke volume variation (SVV), which can be clinically assessed by pulse-contour analysis [11,12], may be a useful tool. It has been described a valuable parameter to assess fluid responsiveness in patients undergoing major surgery and in critically ill patients [13–15]. However, experimental data on right and left ventricular SVV during mechanical ventilation with PEEP in open chest conditions are lacking. Therefore, we investigated the effect of PEEP on right and left ventricular SVV during open chest conditions in comparison to the closed chest condition in a standardized animal model. We further assessed the effect of PEEP on parameters reflecting central blood volume, such as the right ventricular end-diastolic volume (RVEDV) and the global end-diastolic volume (GEDV). The experimental protocol was approved by the local Governmental Commission on the Care and Use of Animals. The animals received care in compliance with the ‘Guide for the Care and Use of Laboratory Animals’ (NIH publication No. 86–23, revised 1985). Fourteen piglets, weighing 25–40 kg, were studied. After premedication with intramuscular ketamine (500 mg) and azaperone (4 mg), a peripheral vein was cannulated. Following preoxygenation, anesthesia was induced with fentanyl (0.02 mg kg−1), propofol (1.5–2 mg kg−1), and atracurium besilate (2 mg kg−1). The animals were intubated and ventilated with 50% oxygen in air, a constant tidal volume of 12 ml kg−1 and an inspiration to expiration ratio of 1:2. Throughout the study, oxygenation was continuously monitored by pulse oxymetry. For maintenance of anesthesia, fentanyl (0.045 mg kg−1 h−1), midazolam (2.5 mg kg−1 h−1) and propofol (10 mg kg−1 h−1) were continuously administered. Bolus injection of atracurium besilate (1 mg kg−1) was repeated prior to surgery. After surgery, the continuous rate of midazolam and propofol was reduced by 30%. Hydration was maintained by continuous saline infusion at a rate of 10 ml kg−1 h−1. Arterial blood gases were regularly monitored throughout the study period. The animals were ventilated using a mechanical ventilator (Servo 900 D, Siemens, Elema, Sweden). The jugular veins, one common carotid artery and one femoral artery were exposed and cannulated with 8.5 F and 5 F introducer sheaths, respectively. In the right carotid artery, arterial pressure was measured using a micro-tip catheter (Millar Instruments, Houston, USA). An 8 Fr central venous catheter for drug administration was inserted in the right external jugular vein and a 7 Fr thermistor tipped pulmonary artery catheter (VoLEF, Catheter, PV2047, Pulsion Medical Systems, Munich, Germany) was introduced via the left external jugular vein. Finally, a 5 Fr thermistor-tipped catheter (PiCCO, PV 2015L20, Pulsion Medical Systems) for transcardiopulmonary thermodilution was placed in the femoral artery. Then, the heart was exposed in the pericardial cradle via a sternotomy and two ultrasonic flow probes (Medi-Stim AS, Grefsen, Norway), diameter 14–16 mm, were placed around the pulmonary artery and the ascending thoracic aorta, respectively. The pericardium was left open. The pleura, however, was not opened. Prior to the measurements, cardiac preload was optimized by fluid loading using 100 ml of a colloid solution (Voluven, Hydroxyethylstarch 6% (HES 130/0,4), Fresenius Kabi, Bad Homburg, Germany) until no further increase in cardiac output could be achieved by two subsequent fluid loading steps. First, hemodynamic and thermodilution measurements were performed with a PEEP of 15 cmH2O and no PEEP while the chest was open. Then, after closing the thorax airtight, the measurements were repeated. Measurements were performed 5 min after changing the PEEP level and 15 min after sternal closure resulting in a total duration of the protocol of approximately 1 h. Using precalibrated ultrasonic flow probes (Medi-Stim AS, Oslo, Norway), we measured pulmonary artery and aortic blood flow over 60 s. The flow signals were registered with a volume flowmeter (CardioMed medical volume flowmeter, CM 1008, Medi-Stim AS, Oslo, Norway), which simultaneously transferred the signals to a personal computer. After amplification, the vascular pressure signals were digitalized and stored on the personal computer. Flow and pressure signals were recorded with a sample frequency of 250 Hz over a time period of 60 s. The stored signals were analyzed using a customized software (Flexpro Version 6.0.18, Weisang GmbH & Co. KG, Germany). The integrals of the pulmonary artery and aortic flow signal were automatically calculated and used for determination of stroke volume (SV). Right and left ventricular SVV and variation in pulmonary and aortic peak blood flow velocity (Δ peak velocity) were defined as variation in SV and peak blood flow velocity, respectively, determined by the flow signals, over one respiratory cycle. They were calculated as follows: SVV (%) = [(SVmax − SVmin)/((SVmax + SVmin)/2)] × 100 and variation in peak blood flow velocity (%) = [(peak flowmax − peak flowmin)/((peak flowmax + peak flowmin)/2)] × 100. The mean of the calculated SVV and Δ peak velocity values over 60 s was used for statistical analysis. As estimates of central blood volume we determined global end-diastolic volume and right ventricular end-diastolic volume. GEDV, the end-diastolic volume of the four heart chambers, was measured by transcardiopulmonary thermodilution (PiCCO system, Pulsion Medical Systems) [16]. RVEDV and right ventricular ejection fraction (RVEF) were measured by pulmonary artery thermodilution (VoLEF catheter, Pulsion Medical Systems). Both thermodilution measurements were performed concomitantly by three subsequent central venous injections of 10 ml cold saline solution (<8 °C), which were randomly administered throughout the respiratory cycle. A measurement was considered valid if the three single measurements differed less then 10% from the mean. If this difference was more than 10%, an additional measurement was performed. The mean of three valid measurements was used for the statistical analysis. Data were analyzed using SigmaStat for Windows 3.1 (Systat Software, Inc., Germany). Data were tested for normal distribution using the Kolmogorov–Smirnov test. Normally distributed data were analyzed with one-way repeated measures analysis of variance (ANOVA). Friedman's test was applied if the test for normal distribution failed. Post hoc testing was performed using the Tukey's test. Differences were considered significant for a P-value < 0.05. Data are reported as mean values ± standard deviation unless indicated otherwise. The application of PEEP significantly reduced mean arterial pressure (MAP) and cardiac output as illustrated in Table 1. Concomitantly, RVEDV, but not GEDV, was decreased during mechanical ventilation with PEEP in the open chest condition. The RVEF was not affected by applying PEEP while the chest was open (Table 1). Hemodynamic data and central blood volume during open and closed chest conditions SVV was calculated from an average of 24 ± 5 respiratory cycles. The results on left and right ventricular SV and SVV are illustrated in Figs. 1 and 22. Left and right ventricular stroke volume (SV) during open and closed chest conditions. (A) Left ventricular SV (LV SV) (14 animals); (B) right ventricular SV (RV SV) (12 animals). Thin lines: individual changes in SV during ventilation without and with positive end-expiratory pressure (PEEP) 15 cmH2O. Thick line, dots and error bars: mean value ± SEM. *P < 0.05, versus open chest, same PEEP level. †P < 0.05, versus no PEEP. Left and right ventricular stroke volume variation (SVV) during open and closed chest conditions. (A) Left ventricular SVV (LV SVV) (14 animals); (B) right ventricular SVV (RV SVV) (12 animals). Thin lines: individual changes in SVV during ventilation without PEEP (positive end-expiratory pressure) and with PEEP 15 cmH2O. Thick line, dots and error bars: mean value ± SEM. *P < 0.05, versus open chest, same PEEP level. †P < 0.05, versus no PEEP. During open chest conditions, the application of PEEP led to reduction of left and right ventricular stroke volumes (33 ± 8 vs 27 ± 8 and 35 ± 10 vs 28 ± 9, respectively, no PEEP vs PEEP) (Fig. 1). Left and right ventricular SVV were significantly higher during mechanical ventilation with PEEP when the chest was open: left ventricular SVV was 15 ± 5% and 8 ± 5% (with and without PEEP, respectively) (Fig. 2A), and right ventricular SVV was 17 ± 5% and 8 ± 3% (with and without PEEP, respectively) (Fig. 2B). The application of PEEP had no influence on the variation in peak blood flow velocity on the ascending aorta and pulmonary artery, respectively, during open chest conditions (Fig. 3). Variation in aortic and pulmonary artery peak blood flow velocity during open and closed chest conditions. Circle and solid line: variation of peak blood flow velocity on the ascending aorta (14 animals); Diamond and long dash: variation of peak blood flow velocity on the pulmonary artery (12 animals). Mean values ± SEM are presented. *P < 0.05, versus open chest, same PEEP level. †P < 0.05, versus no PEEP. Global cardiac function (MAP and CO) was significantly decreased during closed chest conditions when compared to the open chest condition. Concomitantly, GEDV and RVEDV were significantly reduced after chest closure. The effect of chest closure was comparable between the two PEEP levels investigated. The RVEF was significantly lower during closed chest conditions when compared to open chest conditions (Table 1). The closure of the chest reduced left and right ventricular SV (from 33 ± 8 ml to 26 ± 9 ml and from 35 ± 10 ml to 28 ± 11 ml, respectively) (P < 0.05, no PEEP applied) (Fig. 1). Concomitantly, right ventricular SVV increased (from 8 ± 3% to 33 ± 10%) (P < 0.05), whereas left ventricular SVV did not change significantly (from 8 ± 5% to 9 ± 3%) (Fig. 2). The variation in peak blood flow velocity on the ascending aorta and the pulmonary artery was significantly higher after closing the chest when compared to the open chest condition except for the aortic flow variation without PEEP. Pulmonary artery flow variation increased from 14 ± 7% to 31 ± 11% with PEEP applied and from 6 ± 3% to 18 ± 10% without PEEP (P < 0.05). Aortic flow variation increased from 5 ± 4% to 15 ± 5% with PEEP applied (P < 0.05) and from 3 ± 2% to 4 ± 3% (P > 0.05) without PEEP (Fig. 3). The application of PEEP led to a significant reduction of MAP, CO, and RVEDV during closed chest conditions too. GEDV, however, did not differ between ventilation with PEEP and without PEEP (Table 1). During closed chest conditions, left and right ventricular SV were significantly higher during ventilation without PEEP than with PEEP (13 ± 5 vs 26 ± 9 and 14 ± 5 vs 28 ± 11, respectively) (P < 0.05) (Fig. 1). Again, both left and right ventricular SVV were significantly higher if PEEP was applied. The left ventricular SVV was 19 ± 6% and 9 ± 3% (with and without PEEP, respectively) (all P-values < 0.05), right ventricular SVV was 54 ± 11% and 33 ± 10% (with PEEP and without PEEP, respectively) (Fig. 2). The variation in peak blood flow velocity on the ascending aorta and pulmonary artery were significantly increased during mechanical ventilation with PEEP: 15 ± 5% versus 4 ± 3% (PEEP vs no PEEP) for aortic blood flow velocity, and 31 ± 11% versus 18 ± 10% (PEEP vs no PEEP) for pulmonary artery blood flow velocity (P > 0.05) (Fig. 3). In this study, we investigated the effect of PEEP on biventricular stroke volume variation and central blood volume during open and closed chest conditions: left and right ventricular SVV were significantly increased during mechanical ventilation with PEEP in both conditions. Sternal closure following cardiac surgery often adversely affects cardiac performance and is, therefore, sometimes postponed in hemodynamically instable patients immediately after surgery [10,17]. In the present study, cardiac output and stroke volume decreased significantly when the chest was closed. In a previous study in pigs, pericardial closure has been attributed with more adverse effects on cardiac performance than closing the chest [18]. However, closing the pericardium and the chest resulted in a significant reduction in cardiac output and stroke volume comparable to our findings. The decrease of left ventricular end-diastolic dimensions following suture of the pericardium and the chest described by Angelini and colleagues [18] parallel in parts the presented results on global end-diastolic volume and stroke volume variation both reflecting a decrease in cardiac preload. Conversely, opening the chest and the pericardium has been shown to increase right ventricular stroke volume tremendously [19]. Therefore, in accordance with the cited literature, our results support delaying sternal closure in hemodynamically instable patients, especially if the patients are in respiratory failure requiring recruiting maneuvers and mechanical ventilation with PEEP. In the present study, the application of positive end-expiratory pressure led to a reduction in stroke volume and cardiac output both during open and closed chest conditions. The hemodynamic effects of PEEP are known to be dependent on respiratory system compliance [9]. PEEP augments the mechanical compressive force which the expansion of the lungs exerts on the heart during positive pressure ventilation. Further, PEEP reduces venous return by direct compression of the vena cava inferior [20]. Consequently, with high PEEP right ventricular end-diastolic volume and cardiac preload in general decrease [4,7,21]. The observed decrease in right ventricular end-diastolic volume and the increase in biventricular stroke volume variation in the present study demonstrate that this effect is also present during open chest conditions. Interestingly, a reduction in the load-dependent right ventricular ejection fraction following application of PEEP, as it is known from animal and clinical studies [4,7], was absent in the open chest situation in spite of a decrease in cardiac preload, which itself discloses right ventricular afterloading. Whether the unchanged RVEF, therefore, reflects a positive effect of PEEP on myocardial systolic function [22] requires further investigation. The decrease in RVEF during the closed chest situation is attributable both to the preload-reductive effects of PEEP [7] and to a decrease in coronary perfusion pressure reducing right ventricular contractility. This study is the first to present experimental data on left and right ventricular stroke volume variation and variation in peak blood flow velocity during open and closed chest conditions. Recently, our group showed that pulse-contour derived left ventricular SVV enables to predict fluid responsiveness during open chest conditions in cardiac surgical patients [23]. In the present study, a significant decrease in both left and right ventricular SVV was accompanied by a significant increase in RVEDV and CO both during open and closed chest conditions. This further underlines the usefulness of SVV for preload monitoring during open chest conditions. In this context, we further measured aortic peak blood flow variation, a parameter that can be clinically assessed by transesophageal echocardiography. This parameter has been described to accurately reflect cardiac preload in adult and pediatric patients [24,25]. In the present study, aortic peak blood flow variation increased significantly following application of PEEP during closed chest conditions but not when the chest was open. This indicates that in the open chest condition the variation in peak blood flow velocity increases less than the variation of the integral calculated from this flow signal, when cardiac preload decreases. Therefore, the variation in peak blood flow velocity would be less sensitive to changes in cardiac preload than the left ventricular SVV in this situation. This discrepancy between peak blood flow velocity and left ventricular SVV needs further investigation to confirm our results. Both closing the chest and applying PEEP significantly reduced RVEDV. Concomitantly, right ventricular SVV increased further indicating a significant reduction in right ventricular filling caused by those maneuvers. However, so far, no comparable data exist on right ventricular SVV and its relation to cardiac preload and chest wall compliance. Following chest closure, we observed an increase in right ventricular SVV without concomitant changes in left ventricular SVV. This suggests that closing the chest affects the left and the right ventricles in different ways: the cyclic variation in venous return and ventricular filling induced by the mechanical breath seems to be more pronounced for the right than for the left ventricle. An increased intrathoracic pressure due to hyperinflation or PEEP may compress the thoracic vena cava [20,26]. The open pericardium further interacts with ventricular coupling in disfavor of the less musculous right ventricle. These assumptions are further supported by a higher variation in peak blood flow velocity on the pulmonary artery than on the ascending aorta. During closed chest conditions, the results on the variation in peak blood flow velocity parallel the results on SVV, again with an increased variation for the right when compared to the left ventricle. Left and right ventricular stroke volumes, however, have to be almost equal. In this study, we investigated the influence of PEEP on biventricular SVV in an experimental model during open and closed chest conditions. Calculation of SVV from a flow signal is the most precise method of SVV determination, which, however, requires opening the thoracic cavity and the pericardium for placing the flow probes. The fact that we did not readapt the pericardium may have partly contributed to the differences in right and left ventricular SVV too. The pleura was left untouched and no pleural pressures were recorded in the animal model studied. As the pleural pressure is one determinant of right ventricular filling [19], it would have provided some additional and valuable information. We did not perform this study in a crossover design, so there might be a question of intravascular volume depletion during the protocol. As the protocol lasted no longer than 1 h and previous volume loading was performed with a colloid solution, a relevant loss of intravascular volume is not probable. Using this study design, we considered this investigation to be closer to the clinical situation of sternal closure after cardiac surgery. Within the limitations of the model, we consider our experimental data to provide information on functional preload assessment using the left ventricular SVV or the peak blood flow velocity transferable to some clinical situations in thoracic and, especially, cardiac surgery. In conclusion, we found that mechanical ventilation with PEEP is associated with increased right and left ventricular stroke volume variation, both during open and closed chest conditions. The reduction of RVEDV further indicates that PEEP has a preload reductive effect during open chest conditions too. Changes in stroke volume variation and in the variation in peak blood flow velocity behave similar during closed chest conditions, but may differ during open chest conditions. This discrepancy needs further investigation. This study was supported by a research grant from the University of Munich.