| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Departments of *Anaesthesiology and Intensive Care Medicine and Cardiothoracic and Vascular Surgery, University Hospital Schleswig-Holstein, Campus Kiel, Germany.
Address correspondence and reprints requests to Jochen Renner, MD, Department of Anaesthesiology and Intensive Care Medicine, University Hospital Schleswig-Holstein, Campus Kiel, Schwanenweg 21, D-24105 Kiel, Germany. Address e-mail to renner{at}anaesthesie.uni-kiel.de.
Abstract
BACKGROUND: Estimating volume status in infants and neonates is challenging. Global end-diastolic volume (GEDV) and dynamic variables of preload, such as pulse pressure variation (PPV), may be alternative variables for estimating cardiac preload and fluid responsiveness. Therefore, we designed the present study to evaluate whether GEDV and PPV are suitable variables of preload and fluid responsiveness during rapidly changing loading conditions in a pediatric animal model.
METHODS: Nineteen anesthetized and mechanically ventilated piglets (6.5 ± 0.8 kg) were studied during different loading conditions. Hemodynamic measurements, including central venous pressure, pulmonary capillary wedge pressure, PPV, GEDV, and cardiac output derived by transpulmonary thermodilution, cardiac output, and stroke volume index obtained by pulmonary artery thermodilution were performed at normovolemia, and after fluid administration, with 25 mL/kg of hydroxylethyl starch 6%.
RESULTS: There was a significant percentage change of GEDV after volume loading (25% ± 17%) that resulted in significant changes of all hemodynamic variables except of heart rate and systemic vascular resistance index. GEDV was the only preload variable that significantly correlated with volume-induced percentage change in stroke volume index (r = –0.61, P = 0.005). Area under the receiver operating characteristic curve was 0.8 for GEDV (P < 0.02) and 0.6 for PPV (P = ns).
CONCLUSIONS: In this pediatric animal model, GEDV derived from transpulmonary thermodilution was a reliable indicator of cardiac preload. Moreover, GEDV but not PPV, central venous pressure and pulmonary capillary wedge pressure accurately reflected fluid responsiveness.
Hypovolemia is a common cause of circulatory failure in the perioperative period in both infants and adults. Although a large proportion of critically ill children will respond to volume loading with a significant increase in cardiac output (CO), fluid administration may be harmful for some. Thus, for accurate variables of preload, assessment of potentially predictive factors of fluid responsiveness is important. Apart from clinical skills, mainly traditional variables, such central venous pressure (CVP) and mean arterial blood pressure (MAP), are used for guiding fluid therapy in infants and neonates (1). However, several investigations have shown that these variables are only of limited value for accurately reflecting preload (2,3).
Dynamic variables of preload, such as pulse pressure variation (PPV), stroke volume (SV) variation, and systolic pressure variation, have been shown to be good predictors of fluid responsiveness in adults (4–6). Currently, however, there are no data available for infants and neonates. The effect of different tidal volume (VT) ventilation on these dynamic variables is still under debate, and the use of these variables is limited to: 1) patients requiring mechanical ventilation without any spontaneous breathing efforts; 2) patients with sinus rhythm; and 3) patients without tachycardia. In this respect, global end-diastolic volume (GEDV) derived from transpulmonary thermodilution (TPTD) may be an alternative approach of monitoring cardiac preload and fluid responsiveness. TPTD enables bedside measurement of CO (TPTDCO), SV, and calculation of GEDV. The latter variable has been shown to be applicable in infants and neonates (7). GEDV has also been shown to be a reliable variable of cardiac preload in adults (8–10). The issue of fluid responsiveness, however, has not yet been addressed.
Therefore, the aim of the present study was to investigate whether GEDV is an indicator of cardiac preload and a variable of fluid responsiveness during acute changing loading conditions in comparison to PPV and static filling pressures, such as CVP and pulmonary capillary wedge pressure (PCWP), in a pediatric animal model.
METHODS
The project was approved by the local Animal Investigation Committee and the animals were managed in accordance with the American Physiologic Society and institutional guidelines. The study was performed according to the Utstein-style guidelines on healthy swine (German domestic pigs), ranging from 2- to 3-wk of age of either gender, weighing 6.5 ± 0.8 kg. All animals received humane care in compliance with the "Guide for the Care and Use of Laboratory Animals" published by the National Institute of Health (NIH Publication No. 88.23, revised 1996).
Animals were fasted overnight but had free access to water. Pigs were premedicated with the neuroleptic azaperone (8 mg/kg IM) and atropine (0.05 mg/kg IM) 1 h before surgery, and anesthesia was induced with a bolus dose of ketamine (2 mg/kg IM), propofol (2–4 mg/kg IV), and sufentanil (0.3 µg/kg IV) given via an ear vein. After intubation with a cuffed endotracheal tube (ID 4.5 mm), their lungs were ventilated in a volume-controlled mode (Draeger, EV-A, Lübeck, Germany) with a positive end-expiratory pressure of 5 cm H2O, a fixed Vt of Vt 10 mL/kg, an I:E ratio of 1:1, and a fraction of inspired oxygen (Fio2) of 0.35. Respiratory rate (18–22 breaths/min) was adjusted to maintain normocapnia (Pco2, 35–45 mm Hg). Anesthesia was maintained with a continuous infusion of propofol (6–8 mg · kg–1 · h–1), sufentanil (0.3 µg · kg–1 · h–1), and muscle relaxation was provided by a continuous infusion of pancuronium (0.2 mg · kg–1 · h–1) to ensure that the respiratory changes in PPV reflected only the effects of positive pressure ventilation. Ringer's solution (10 mL · kg–1 · h–1) was administered during instrumentation. A standard lead II electrocardiogram was used to monitor cardiac rhythm. Depth of anesthesia was judged according to MAP and heart rate. If cardiovascular variables indicated a reduced depth of anesthesia, additional propofol and sufentanil were given. In our experience, the animals do not respond to painful or auditory stimuli under this anesthetic regimen when the paralyzing drug is withheld.
Hemodynamic Monitoring
A 5F pulmonary artery catheter (PAC) (Baxter Healthcare Corporation, Irvine, CA) was inserted percutaneously in the right internal jugular vein for measurement of CVP, PCWP, and pulmonary artery thermodilution CO (COPAC). The catheter was advanced under continuous pressure recording into wedge position and then connected to a CO computer system (Vigilance Monitor, Baxter Edwards Critical Care, Irvine, CA). A 7-cm 3F thermistor-tipped catheter for arterial thermodilution and pulse contour (PC) analysis (Pulsion Medical Systems AG, Munich, Germany) was inserted percutaneously into the femoral artery. The arterial catheter allows discontinuous measurement of TPTDCO, SV, and GEDV as described previously (11). Additionally, MAP, PC CO (PCCO), SV, SV index (SVI), and PPV were monitored continuously based on a modified algorithm originally described by Wesseling et al. (12) (PiCCO Plus®, Version 6.0, Pulsion Medical Systems, Munich, Germany). This algorithm enables continuous calculation of SV by measuring the systolic portion of the aortic pressure wave form and dividing the area under the curve by the aortic impedance. Initially, the specific aortic impedance is determined by TPTD (11). Five milliliters ice-cold saline randomly assigned to the respiratory cycle was injected three times in the proximal port of the PAC to simultaneously assess TPTDCO, calibrate PC derived CO and obtain COPAC. All thermodilution curves were analyzed and accepted or, if necessary, rejected and calibration repeated.
The PiCCO monitor also calculates the mean transit time (mtt) and the down-slope time (dst) of the aortic thermodilution curve which enables GEDV calculation (13).
GEDV is calculated according to:
|
|
Additionally, using the PiCCO system, PPV can be determined during a continuously sliding time window of 30 s. This time window is further divided into four 7.5 s periods. Within each period, the largest and the smallest value of pulse pressure (i.e., the difference between systolic and diastolic arterial pressure) were determined and the average of the four 7.5 s intervals were used to calculate PPV:
|
|
The intravascular catheters were attached to pressure transducers (model 1290A, Hewlett Packard, Böblingen, Germany) that were aligned at the level of the right atrium; all pressure tracings were recorded with a data acquisition system (Dewetron port 2000, Graz, Austria). Body temperature was maintained between 38.0°C and 39.0°C with a heating blanket. End-tidal carbon dioxide was measured with an infrared absorption analyzer (Sirecust 960, Siemens, Erlangen, Germany).
Experimental Protocol
Pigs were studied at different experimental stages: normovolemia after induction of anesthesia and after performing a fluid loading with 25 mL/kg of hydroxylethyl starch 6% lasting approximately 10 min. Hemodynamic measurements at each experimental stage were obtained after a short period of stabilization and after recalibration of the PiCCO system and after assessment of GEDV and COPAC and SVIPAC. Hemodynamic data were recorded during a period of 3 min and mean values were calculated. Correlation between preload variables and volume-induced percentage change in SVI are based on SVIPAC. After finishing the experimental protocol, the animals were killed with an overdose of propofol, sufentanil, and potassium chloride. All piglets were then subjected to necropsy to check for correct positioning of the catheters, and damage to internal organs.
Statistical Analysis
Data are represented as mean ± sd. Statistical comparisons were performed using commercially available statistics software (GraphPad Prism 4, GraphPad Software Inc., San Diego, CA). A Student's t-test was used for comparison of data before and after volume challenge. Linear correlation was tested using Pearson product moment correlation. CO derived by different methods was compared with the method described by Bland and Altman (14), and the mean difference (bias) and two standard deviations of bias (limits of agreement) were calculated. Prediction of fluid responsiveness for preload variables was tested by calculating the area under the receiver operating characteristics (ROC) curve for a SVI increase of 
20% (area under the curve [AUC] = 0.5: not better than chance, no prediction possible; AUC = 1.0 = best possible prediction). P < 0.05 was considered significant.
RESULTS
Hemodynamic variables during the experiment are presented in Table 1. Comparing preload variables throughout the experiment, there was a close correlation between CVP and PCWP (r = 0.80, P < 0.0001), and a weak correlation between GEDV and PPV (r = 0.47, P < 0.01), whereas there was no correlation between GEDV and CVP/PCWP, respectively. There was a significant percentage change of GEDV after volume loading (25% ± 17%) that resulted in significant changes of all hemodynamic variables except of heart rate and systemic vascular resistance index (SVRI). The mean increase of SVI induced by volume loading was 24% ± 20%. TPTDCO significantly correlated with PACCO (r = 0.79, P < 0.0001). The mean difference between TPTDCO and PACCO was 0.18 ± 0.32 L/min, and limits of agreement were –0.45 to 0.82 L/min (Fig. 1). Ten piglets had an increase of SVI >20% (responder), and nine piglets showed an increase of SVI <20% (nonresponder).
|
|
GEDV was the only preload variable that showed a good correlation with SVI throughout the entire experiment. In contrast to PPV, GEDV showed less interindividual variance in response to volume loading (Fig. 2). Moreover, only GEDV significantly correlated with volume-induced percentage change in SVI (
SVI), whereas PPV, CVP, and PCWP did not correlate with SVI (Fig. 3). Results of ROC analysis and linear regression analysis are presented in Table 2. The area under the receiver operating characteristic curve (AUC), showing the ability of different hemodynamic variables to discriminate between responder and nonresponder, was greatest for GEDV (0.80), followed by PPV (0.60) (Fig. 4). ROC analysis yielded the highest sensitivity and specificity with respect to prediction of fluid responsiveness for GEDV values 
169 mL. In subjects with GEDV 169 mL at baseline an increase in SVI >20% in response to volume loading could be expected with a sensitivity of 71% and a specificity of 77%.
|
|
|
|
DISCUSSION
The main findings of our prospective experimental study are as follows: 1) in this pediatric animal model, GEDV was a reliable indicator of cardiac preload; 2) GEDV compared with PPV showed less interindividual variance in response to volume loading; 3) GEDV, but not PPV and static filling pressures, accurately reflected fluid responsiveness.
TPTD is a reliable method of calculating CO in both adults and children compared with other techniques, such as the well-established Fick method (15–17). Besides monitoring CO and SVI, TPTD allows estimation of volumetric variables of preload such as GEDV (9,18,19). In addition to volumetric and static variables of preload, the recently introduced dynamic variables, such as PPV, SV variation, and systolic pressure variation, have been shown to be good predictors of fluid responsiveness (6,20). However, only few studies are available in infants and neonates addressing the impact of volumetric and dynamic variables of preload on preload assessment and prediction of fluid responsiveness.
In the present investigation, GEDV proved to be a suitable variable of preload and showed less interindividual variability compared with PPV and static filling pressures, which is in agreement with recent investigations in adults (7,10,21). Currently, normal values for intracardiac and intrathoracic blood volumes in infants and neonates have not yet been defined. In a study from Egan et al. (22) in infants after cardiac surgery, values of indexed GEDV (427 ± 38 mL/m2) were comparable to values obtained in our study. In contrast, in a study on 10 critically ill infants with heart failure, GEDV values were different compared with our results (GEDV: 405 ± 129 mL) (10). The presence of congestion due to heart failure in most of the infants may explain this difference. To our knowledge, only few data are available addressing the ability of volumetric variables to reflect preload in infants and neonates. Moreover, no studies in this patient population are available investigating and comparing the ability of volumetric and dynamic variables of preload to predict fluid responsiveness.
The relationship between a volume challenge and the percentage change in SVI, recently introduced as "fluid responsiveness," displays the ability of the heart to improve SV in response to fluid administration. With respect to fluid responsiveness, GEDV, and not PPV, CVP or PCWP, significantly correlated with SVI. Further, our results suggest a threshold value of GEDV <169 mL to predict an increase in SVI >20% in response to a fluid bolus of 25 mL/kg of hydroxylethyl starch 6%. In pigs with a GEDV <169 mL at baseline, an increase in SVI >20% in response to volume loading could be expected with a sensitivity of 71% and a specificity of 77%. Consequently, the lower the GEDV, the more likely the animals increased SV in response to volume loading. In contrast, PPV, as well as CVP and PCWP, showed no predictive value in terms of fluid responsiveness. With regard to static filling pressures, this observation is in good agreement with previous investigations (3,6,21).
For PPV, threshold values of 13% and 13.5% reportedly predict an increase in cardiac index >15% or an increase in SVI >25%, respectively, after volume loading in adults (23,24). In this pediatric animal model, however, PPV was neither an accurate parameter of preload nor a sensitive and specific marker of fluid responsiveness. This seems counterintuitive at first, given the number of studies supporting PPV for guiding fluid administration in adults. However, there are important differences between adult and pediatric cardiovascular physiology. In the latter, heart rate, chest wall compliance, MAP, arterial vasomotor tone, and aortic compliance, all may influence pulse pressure in a differing manner than in adults. The interaction of these variables may explain, at least in part, why PPV does not work in infants and neonates for prediction of fluid responsiveness. Although PPV in our experimental setting decreased significantly after volume challenge, we could not demonstrate that the greater the PPV the more likely the subjects were to increase their SV in response to volume loading. Interestingly, there was a great interindividual variance in PPV values that may have prevented a more reliable prediction of fluid responsiveness. This may have been due to the used monitoring system we used that calculates and displays PPV over a period of several seconds, albeit without any synchronization to the ventilator or any kind of respiratory signal (25).
Although the ability of PPV to predict fluid responsiveness has been tested in a number of studies, it still remains only a surrogate of changes in left ventricular SV during positive-pressure ventilation. Further, dynamic variables are of limited value in patients with spontaneous breathing (26). Therefore, many patients in the intensive care unit cannot be monitored by dynamic variables of preload; consequently, variables without the aforementioned limitations, such as GEDV, may be particularly useful under these circumstances. From all variables studied, only GEDV was able to reflect both actual preload and fluid responsiveness, a fact that puts GEDV in a unique position for guidance of fluid administration. Further studies in infants undergoing surgical treatment of congenital heart defects and those undergoing major abdominal surgery such as liver transplantation are warranted to further define whether volumetric and dynamic variables of preload are as suitable as they have been shown to be in adults, especially under rapid and pronounced alterations of preload.
Some limitations of our study should be noted. One is that a real "gold standard," such as the direct Fick method or an ultrasonic flow probe was not used on the aorta to measure hemodynamic variables. Instead we used an accepted clinical gold standard, the PAC. We cannot completely exclude the possibility that the anesthetic regimen may have influenced the cardiovascular response in an unforeseen manner. Many hemodynamic studies on pigs, however, have used total IV anesthetics comparable to our study.
To our knowledge, GEDV has not been compared with standard methods of measuring heart volumes, such as magnetic resonance imaging. However, in a study in pigs during hemorrhage, Nirmalan et al. (27) were able to demonstrate that the algorithm used in our study for estimation of GEDV is robust even during extreme loading conditions. Another important issue relates to the possibility of mathematical coupling between GEDV and CO, since both variables are derived from the same thermodilution curve (28). However, McLuckie and Bihari (29) have shown that the mean transit time of an indicator may change independently of changes in CO, a finding supported by a study performed in patients during minimally invasive coronary artery bypass grafting (30).
In conclusion, in this pediatric animal model, GEDV derived from TPTD was a reliable indicator of cardiac preload and showed less interindividual variance in response to volume loading compared to PPV. Moreover, GEDV, but not PPV and static filling pressures, accurately reflected fluid responsiveness.
ACKNOWLEDGMENTS
The authors are indebted to Gunnar Kuschel, MS, for excellent technical assistance and logistic support, and to Juergen Hedderich, PhD, for statistical advice.
Footnotes
Accepted for publication July 5, 2007.
No author has a conflict of interest with regard to any device employed in this study.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
