The perioperative graft function of orthotopic liver transplantation (OLT) patients is particularly vulnerable , and its maintenance also depends on stable haemodynamics and the prevention of venous stasis. Extended monitoring during anaesthesia for liver transplantation usually includes a pulmonary artery catheter (PAC) for cardiac index (CIpa) determination and the monitoring of central venous pressure (CVP), pulmonary artery and pulmonary artery occlusion pressure (PAOP). These pressure-derived data, which often assist in determining initial resuscitative efforts, have little positive predictive value in guiding haemodynamic management to improve tissue perfusion . Inadequate cardiac filling leading to suboptimal tissue perfusion and multiple organ failure, or excessive filling resulting in pulmonary oedema and a worsening respiratory function may contribute to both perioperative morbidity and mortality rates in liver-transplanted patients. Because of the reluctance of using pulmonary artery catheterization with its inherent technical problems with the pressure data obtained, new preload variables have been investigated [3-5]. The transpulmonary indicator dilution technique directly measures the circulating blood volume. The PiCCO® system (Pulsion Medical System; Munich, Germany), which uses the single-indicator arterial thermodilution technique, can evaluate the cardiac index (CIart) to assess preload as the intrathoracic blood volume index (ITBVI) and the extravascular lung water index (EVLWI) [6-8].
The primary objective of the present study was to analyse two preload variables, PAOP and ITBVI, with respect to CIpa and stroke volume index (SVIpa) obtained from PAC, and the correlation between ITBVI and PAOP in patients undergoing liver transplantation. The secondary aim was to evaluate the relationship between the changes (Δ) of the preload variables (ITBVI, PAOP) and the changes (Δ) of CIpa and SVIpa, and between ΔITBVI and ΔPAOP. The reproducibility and precision of all CIart and CIpa measurements were also evaluated.
After approval of the study protocol by the Institutional Ethics Committee and having obtained patients' signed written informed consent, 60 OLT patients (46 male, 14 female) were consecutively enrolled into the study from June 1998 to December 2000. Patients with pre-existing pulmonary and/or cardiac diseases other than the common symptoms of end-stage liver dysfunction  and patients with fulminant hepatic failure, hepatopulmonary syndrome or pulmonary hypertension were excluded. These abnormalities were evaluated during preoperative clinical assessment of the liver transplant candidates.
Anaesthesia and mechanical ventilation during OLT
The same anaesthetic management was applied to all individuals to smooth out the influence of anaesthetic agents. Liver transplantation was performed with the piggyback technique. Intermittent positive-pressure ventilation of the lungs with positive end-expiratory pressure (PEEP, 5 cmH2O) was performed with a volumetric anaesthesia ventilator (CATO®, Dräger, Lübek, Germany). Tidal volume, respiratory rate and the mixture of O2 in air were adjusted to maintain normocapnia (PaCO2 = 4.7-5.3 kPa) and an arterial oxygen saturation >95%. End-tidal PCO2 and inspiratory and expiratory gas analysis were also monitored (CATO®). Body temperature was controlled to avoid hypothermia with a warming blanket (Gaymar Meditherm®; Orchard Park, NY, USA) and by warming all intravenous infusions (HOT LINE®; SIMS Medical System Inc, Rockland, MA, USA). In all patients, intravascular volume replacement was achieved with 0.9% saline (10 mL kg−1 h−1). Hydroxyethyl starch solution 6% (MW 200/0.5, HAES-steril®; Fresenius, Hamburg, Germany), 500 mL, was administered in all patients at the beginning of the procedure. Human albumin solution 20% was administered according to the degree of hypoalbuminaemia (2.0 g dL−1). Fresh frozen plasma (FFP) and platelet transfusions were given if the prothrombin time (judged as the international normalized ratio [INR]) > 1.5 and platelet count <5.104 mm3. Packed red cells (PRC) were transfused to maintain a haemoglobin concentration >8 g dL−1. Patients received continuous infusions of dopamine (2 μg kg−1 min−1) to maintain splanchnic perfusion and urine output, and tranexamic acid (7.5 mg kg−1 h−1, after a bolus of 15 mg kg−1) throughout the entire procedure. Furosemide (0.3-0.5 mg kg−1) or mannitol 18% (0.3 g kg−1) were used only when necessary to obtain a mean urine output > 1 mL kg−1 h−1. For guidance of volume resuscitation, ITBVI <800 mL m−2 was corrected by volume replacement; ITBVI > 1000 mL m−2 required the use of vasoactive drugs and/or diuretic therapy and EVLWI was limited to 7 mL kg−1.
All patients were monitored before induction of anaesthesia with a lead II/V5 electrocardiogram for the derivation of the heart rate, pulse oximetry (SpO2), and radial artery catheterization for measurement of invasive arterial pressure (PCM SpaceLabs, Inc, Redmond, WA, USA). After induction of anaesthesia, a 7.5 Fr PAC was placed in all patients via an introducer (8.5 Fr; all Intellicath®; Edwards Laboratories, Irvine, CA, USA) through the right internal jugular vein using the Seldinger technique. The catheter was connected to the Vigilance® system (Edwards Laboratories) for CIpa, CVP, PAOP, mean pulmonary arterial pressure (MPAP) and body temperature (°C) monitoring, and to obtain continuous mixed venous saturation (SvO2), continuous cardiac output evaluation and SVIpa. The zero measurement reference for the supine position was the midpoint of the axilla. Intermittent pulmonary artery thermodilution CIpa measurement was performed with the Vigilance® monitor by manual injection of iced normal saline. The average of three CIpa - measured within a 10% range - distributed randomly over the respiratory cycle was calculated. To avoid inter-operator variability, the same investigator always injected the bolus. Mechanical ventilation settings were kept the same during all haemodynamic determinations. Intracardiac and pulmonary artery pressures were monitored continuously to ensure the correct placement of the catheter.
A 4-Fr gauge thermistor-tipped catheter (Pulsiocath PV2014L®; Pulsion Medical System) was placed via a 5-Fr gauge introducer (Adam Spence Europe Ltd, Abbeytown, Boyle, Ireland) through the right femoral artery and connected to the PiCCO® System (Pulsion Medical System) to monitor the transpulmonary cardiac index (CIart) and ITBVI (normal = 800-1000 mL m−2) and EVLWI (normal = 4-7 mL kg−1). The CIart and volumetric variables were obtained from a transpulmonary indicator dilution technique; the mean of three subsequent CIart measurements was used. The measurements were performed by injection of 15 mL ice-cooled saline at 4-7°C via a central venous catheter with subsequent detection by the thermistor embedded into the wall of the arterial catheter. The Stewart-Hamilton algorithm calculated CIart from the thermodilution curve . ITBVI and EVLWI were calculated by the mean transit time approach (thermodilution technique) as described . The PiCCO® system revealed the volumes as: EQUATION 1
where GEDV is the global end-diastolic volume, COart is the arterial cardiac output, MTtTD is the mean transit time (s) of the cold indicator from the point of injection at the determination site and the down-slope time and DStTD is the constant of decay of the arterial thermodilution curves. Based on the linear relation found in previous experimental and clinical studies, the PiCCO® system could derive the intrathoracic blood volume (ITBV) with the single indicator: EQUATION 2
where a and b are two predefined coefficients (respectively 1.16 and 86 mL m−2).
After induction of anaesthesia and when stable haemodynamic conditions were complete, baseline values of haemodynamic data and intra- and extravascular thoracic volumes were determined. All volumetric and pressure-derived variables were indexed to body surface area to improve the interindividual comparison. The patients were studied during three different phases: after induction of anaesthesia (A), during the anhepatic phase (B) and at the end of surgery (C).
All haemodynamic and volumetric data obtained were analysed using SAS® (SAS Macro®; SAS Inc, Cary, NC, USA) software (for Windows PC, v.6.01). Means ± SD were calculated.
All haemodynamic and volumetric data collected were analysed using ANOVA for repeated measurements and paired Student's t-test with Bonferroni correction. The correlations between the variables as well as correlations between the changes (Δ) in these variables were studied using linear regression analysis. Changes in the variables were calculated by subtracting the first from the second measurement (Δ1 = B − A) and the second from the third (Δ2 = C − B). To reduce the influence of changes in contractility and afterload, each set of measurements was performed in a steady-state period, i.e. at least 15 min after the change in dosage of catecholamine or sedatives, infusion rate or ventilator settings. This avoided changes of the SVIpa by a factor unrelated to volume status. The SVIpa used in all analyses was based on pulmonary arterial thermodilution. The relationship between the two different preload variables (PAOP and ITBVI), and the CIpa and SVIpa, and between ITBVI and PAOP were also analysed at each predefined step by linear regression.
Agreement between cardiac index measurements obtained by the PAC and PiCCO® systems was analysed using the method of Bland and Altman . Bias between the methods was calculated as the mean difference between CIart and CIpa. The upper and lower limits of agreement were calculated as bias ± 2 SD, and defined the range in which 95% of the differences between the methods was expected to lie. The precision of the bias analysis and limits of agreement were assessed using 95% confidence intervals. Additionally, correlation analysis was performed between methods. P < 0.05 was considered significant.
Patients' characteristics and the underlying diseases necessitating liver transplantation are presented in Table 1. The mean (range) anaesthesia time was 595 ± 95 (340-795) min. Mean (range) FFP, PRC and human albumin administration were respectively 19.4 ± 9.8 (3-56) U, 5.1 ± 3 (1-11) U and 300 ± 150 (0-450) mL. Four patients received platelets (9 ± 1.4 U). The intraoperative volume of blood salvaged by the cell saver was 906 ± 875 (100-4230) mL. Mean urine output was 5.1 ± 2.3 mL kg−1 h−1. Haemodynamic and volumetric data are given as mean ± SD in Table 2.
Linear regression analysis revealed a significant correlation between ITBVI and CIpa (r2 = 0.47), and between ITBVI and SVIpa (r2 = 0.55) while PAOP poor correlated to CIpa (r2 = 0.02) and SVIpa (r2 = 0.015) (Fig. 1 and Table 3). In addition, correlation between ITBVI and PAOP was poor (r2 = 0.002) (Fig. 2 and Table 3). Only changes in ITBVI showed significant correlations to changes in CIpa (Δ1, r2 = 0.37, P < 0.0001; Δ2, r2 = 0.32, P < 0.0001) and in SVIpa (Δ1, r2 = 0.60, P < 0.0001; Δ2, r2 = 0.47, P < 0.0001), whereas PAOP failed (CIpa Δ1, r2 = 0.0009; Δ2, r2 = 0.03; SVIpa Δ1, r2 = 0.007; Δ2, r2 = 0.0001) as given in Table 4 and in Figures 3 and 4. Linear regression between analysed data for the three phases are given in Table 5.
The determination of CIart is required for the calculation of ITBVI. To verify the reliability and reproducibility of CIart, all simultaneous measurements of CIpa and CIart were compared. A close agreement among the two different techniques was observed at all measurements. Mean bias between CIart and CIpa was 0.13 L min−1 m−2 (2 SD of differences between methods = 1.04 L min−1 m−2). The correlation analysis showed a correlation coefficient (r2) of 0.86 (P < 0.0001) (Fig. 5).
Adequate volume management owing to rapidly changing volume conditions and the variety of haemodynamic disturbances during liver transplantation, in particular, without venovenous bypass during the anhepatic period and after graft reperfusion is vital. Tissue hypoperfusion during surgery has been shown to be a portent of poor outcome . Whatever the cause of this problem, either poor cardiovascular performance or reduced intravascular volume, the link between alterations in microvascular flow and the onset of multiple organ dysfunction is strong.
The application of the piggyback technique without temporary portocaval shunt is characterized by portal clamping without any surgical shunt with venous circulation that allows blood outflow from the mesenteric region only by pre-existing portocaval shunts. Therefore, splanchnic organs may become congested and therefore hypoxic. Under such conditions, cardiac output is reduced by at least 40-50%, and the increase in systemic vascular resistance resists a reduction in arterial pressure . Furthermore, by the analysis of classical cardiovascular variables, a better organ perfusion and oxygenation with the piggyback technique (with temporary portocaval shunt) than with venovenous bypass has been suggested . Therefore, fluid management in these patients is often a balancing act between the optimization of preload and the avoidance of pulmonary oedema, particularly during the postanhepatic phase when a postperfusion syndrome can occur .
Transoesophageal echocardiography, a semi-invasive and safe technique, has been helpful by providing additional information about preload and contractility and can help guide fluid replacement and catecholamine therapy during liver transplantation [16-18]. Measurement of left ventricular end-diastolic area by transoesophageal echocardiography, although considered to be the clinical 'gold standard' for the estimation of preload, is not routinely used in most operating rooms because of its high cost and dependency on skilled operators.
The PAC is still the most commonly used haemodynamic tool, but filling pressures can only serve as indirect indicators of filling volumes. Recently, owing to its relatively high degree of invasiveness, its use has been criticized [19-21]. As an alternative, a new monitoring system based on the transpulmonary thermal-dye dilution technique using cold indocyanine green dye (ICG) has been developed (COLD® system; Pulsion Medical System) . The method uses two indicators (i.e. ice-cold glucose and ICG). The cold indicator is distributed throughout both the intra- and extravascular volumes, whereas ICG remains in the intravascular volume. The aortic cardiac index is determined from the thermodilution curve. The calculation of the intrathoracic volumes, ITBVI and EVLWI, is performed by an analysis of the mean transit time of the indicators derived from the dilution curves that are recorded in the descending aorta as described by Pfeiffer and colleagues . Another parameter that may be measured by applying the double-indicator thermodilution technique is the indocyanine green elimination as an index of excretory liver function (PDRICG where PDR = plasma disappearance rate).
ITBVI, when compared with CVP and PAOP, directly reflects cardiac preload in many experimental and clinical studies and is considered to be an indicator of cardiac preload which is less influenced by changes in intrathoracic pressure and myocardial compliance [23-26]. Gödje and colleagues compared the conventional variables CVP and PAOP with ITBV in 30 patients after coronary artery bypass grafting, and in 40 patients during the postoperative period after heart transplantation [23,25]. After coronary artery bypass grafting, linear regression analysis was computed between changes of preload dependent left ventricular SVI and the corresponding, presumably preload indicating variables CVP, PAOP and ITBV. A good correlation was only obtained between ITBVI and SVI. Sakka and colleagues  confirmed that in comparison with cardiac-filling pressure, ITBVI seemed the best cardiac preload indicator in patients with sepsis or septic shock. Recently, Krenn and colleagues  monitored patients undergoing liver transplantation with the COLD® system and observed that an increase of ITBVI influenced pulmonary function after reperfusion; this was demonstrated by the increase in QS/QT without EVLWI and oxygenation impairment.
The data reported above [23-25,27] were obtained with the double-indicator technique. Although effective at the bedside, it is relatively time-consuming owing to the preparation of the indocyanine green solution, and is cumbersome and expensive. We used the single-indicator thermodilution technique, performed with the PiCCO® system, to measure ITBVI. To our knowledge, this is the first study that compares the value of each preload variable with the single-indicator thermodilution technique during anaesthesia for liver transplantation. In terms of preload variables, the main findings here showed a good correlation between ITBVI and SVIpa and CIpa, while no consistent correlation could be established between PAOP and SVIpa or CIpa (Table 3). Changes in ITBVI showed better correlations with changes in SVIpa or CIpa than changes in PAOP, which is also in agreement with earlier findings obtained with the double-indicator technique (Table 4)[18,23-26]. Statistically significant correlations were obtained by analysing the predefined steps: after induction of anaesthesia, during the anhepatic phase and at the end of surgery. This confirmed the validity of ITBVI as a preload index also during phases characterized by major haemodynamic changes: clamping of the inferior vena cava, unclamping of the anastomoses, and the graft reperfusion, bleeding and surgical manipulations (Table 5). The underlying measurement for the cardiac index by femoral arterial thermodilution (CIart) has been validated by a direct comparison with standard pulmonary artery thermodilution (CIpa). The results showed a close agreement between CIart and CIpa confirming previous data reported in a different clinical setting [28,29].
These experimental and clinical data demonstrate that a single arterial thermodilution-derived ITBVTD correlates well with the values measured by the double-indicator technique (ITBVDD) [5,8]. Neumann , in an experimental model of lung injury performed in 13 mechanically ventilated pigs, evaluated the accuracy of single-thermodilution ITBVTD and extravascular lung water (EVLWTD) versus these data obtained with the double-indicator technique. The single-thermodilution estimation of ITBVTD (r2 = 0.87) and EVLWTD (r2 = 0.98) is reasonably accurate and therefore provides useful information about the cardiac preload and severity of lung injury. Buhre and colleagues  compared ITBVDD versus the derived single-indicator ITBVTD in 10 neurosurgical patients and concluded that changes in ITBVTD were similar to those assessed by double-indicator dilution; the repeatability of both methods was comparable. Sakka and colleagues  confirmed the accuracy of the single-indicator ITBVITD and EVLWITD in septic patients.
ITBVI is uninfluenced by mechanical ventilation of the lungs because the recording of transpulmonary thermodilution is within the aorta and not within the pulmonary artery - thus avoiding interference due to chest wall compliance . The limitations of the single-thermodilution technique are similar to those of the double-indicator technique. Volumes will be overestimated in the presence of large aortic aneurysms or catheters placed too far peripherally, i.e. in the radial artery, owing to a prolonged mean transit time. Furthermore, intracardiac shunts may limit the use of this technique. Additionally the accuracy of EVLWITD may be reduced in the presence of extreme pulmonary perfusion abnormality .
Based on data from the literature, the assessment of changes in ITBVI by the single-thermodilution method is effective for clinical purposes and is relatively less expensive when compared with the double-indicator technique. It can be used in the current clinical setting to monitor the cardiac index (intermittent and continuous values), preload data (intrathoracic blood volume) and lung oedema as extravascular lung water. The PiCCO® system essentially needs only a central venous approach and is less invasive than a PAC. The necessary haemodynamic and volumetric monitoring can be performed with good reproducibility independent of operator ability during anaesthesia in critically ill patients or during weaning from ventilatory support.
In conclusion, the ITBVI is a more reliable indicator of preload than the PAOP method during liver transplantation. Optimization of the fluid balance and vasoactive drugs administration based on volumetric monitoring makes the PiCCO® an effective monitoring system, particularly during anaesthesia for liver transplantation where intravascular volume management is a primary objective.
1. Seltzer H, Hiesmajr M, Tuchy G, Zimpfer M. Perioperative liver graft function: the role of oxygen transport and utilization. Anesth Analg
2. Shippy CR, Apple PL, Shoemaker WC. Reliability of clinical monitoring to assess blood volume in critically ill patients. Crit Care Med
3. Cusack RJ, Rhodes A. Pulmonary artery catheters - to use or not to use; that is the question? Clin Intens Care
4. Hedenstierna G. What value does the recording of intrathoracic blood volume have in clinical practice? Intensive Care Med
5. Gödje O, Peyerl M, Seebauer T, Dewald O, Reichart B. Reproducibility of double indicator dilution measurements of intrathoracic blood volume compartments, extravascular lung water, and liver function. Chest
6. Buhre W, Bendyk K, Weyland A, et al.
Assessment of intrathoracic blood volume. Thermo-dye dilution technique vs single thermodilution technique. Anaesthesist
7. Neumann P. Extravascular lung water and intrathoracic blood volume: double versus single indicator dilution technique. Intensive Care Med
8. Sakka SG, Rühl CC, Pfeiffer UJ, et al.
Assessment of cardiac preload and extravascular lung water by single transpulmonary thermodilution. Intensive Care Med
9. Hourani JM, Bellamy PE, Tashkin DP, et al.
Pulmonary dysfunction in advanced liver disease: frequent occurrence of an abnormal diffusing capacity. Am J Med
10. Weyland A, Buhre W, Hoeft A, et al.
Application of transpulmonary double indicator dilution method for post-operative assessment of cardiac index, pulmonary vascular resistance index, and extra vascular lung water in children undergoing total cavo-pulmonary anastomosis: preliminary results in six patients. J Cardiothorac Vasc Anesth
11. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurements. Lancet
12. Nasraway SA, Klein RD, Spanier TB, Rohrer RJ, et al.
Haemodynamic correlates of outcome in patients undergoing orthotopic liver transplantation
. Evidence for early post-operative myocardial depression. Chest
13. Carmichael FJ, Lindop MJ, Farman JA. Anaesthesia for hepatic transplantation: cardiovascular and metabolic alterations and their management. Anesth Analg
14. Steib A, Saada A, Clever B, et al.
Orthotopic liver transplantation
with preservation of portocaval flow compared with veno venous by-pass. Liver Transpl Surg
15. Aggarwal S, Kang Y, Freeman JA, Fortunato F, Pinsky MR. Postreperfusion syndrome: cardiovascular collapse following hepatic reperfusion during liver transplantation
. Transpl Proc
(4 Suppl 3): 54-55.
16. De Wolf A. Transesophageal echocardiography and orthotopic liver transplantation
: general concepts. Liver Transpl Surg
17. Steltzer H, Blazek G, Gabriel A, et al.
Two-dimensional transesophageal echocardiography in early diagnosis and treatment of hemodynamic disturbances during liver transplantation
. Transplant Proc
18. Hinder F, Poelaert JI, Schmidt C, et al.
Assessment of cardiovascular volume status by transoesophageal echocardiography and dye dilution during cardiac surgery. Eur J Anaesthesiol
19. Dalen JE. The pulmonary artery catheter. Friend, foe, or accomplice? JAMA
20. Swann DG. The utility of pulmonary artery catheterization. Br J Anaesth
21. Murdoch SD, Cohen AT, Bellamy MC. Pulmonary artery catheterization and mortality in critically ill patients. Br J Anaesth
22. Pfeiffer UJ, Backus G, Blümel G, et al.
A fiberoptics-based system for integrated monitoring of cardiac output
, intrathoracic blood volume, extravascular lung water. O2
saturation, and a-v differences. In: Lewis FR, Pfeiffer UJ, eds. Practical Applications of Fiberoptics in Clinical Care Monitoring.
Berlin, Germany: Springer, 1990: 114-125.
23. Gödje O, Peyerl M, Seebauer T, Lamm P, Mair H, Reichart B. Central venous pressure, pulmonary capillary wedge pressure and intrathoracic blood volumes as preload indicators in cardiac surgery patients. Eur J Cardiothorac Surg
24. Sakka SG, Bredle DL, Reinhart K, Meier-Hellmann A. Comparison between intrathoracic blood volume and cardiac filling pressures in the early phase of hemodynamic instability of patients with sepsis or septic shock. J Crit Care
25. Gödje O, Seebauer T, Peyerl M, Pfeiffer UJ, Reichart B. Hemodynamic monitoring by double indicator dilution technique in patients after orthotopic heart transplantation. Chest
26. Buhre W, Buhre K, Kazmaier S, Sonntag H, Weyland A. Assessment of cardiac preload by indicator dilution and transoesophageal echocardiography. Eur J Anaesthesiol
27. Krenn CG, Plochl W, Nikolic A, et al.
Intrathoracic fluid volumes and pulmonary function during orthotopic liver transplantation
28. Buhre W, Weyland A, Kazmaier S, et al.
Comparison of cardiac output
assessed by pulse contour analysis and thermodilution in patients undergoing minimally invasive direct coronary artery bypass grafting. J Cardiothorac Vasc Anesth
29. Gödje O, Hoeke K, Lichtwarck-Aschoff M, Faltchauser A, Lamm P, Reichart B. Continuous cardiac output
by femoral arterial thermodilution calibrated pulse contour analysis: comparison with pulmonary arterial thermodilution. Crit Care Med
30. Lichtwarck-Aschoff M, Zeravick J, Pfeiffer UJ. Intrathoracic blood volume accurately reflects circulatory volume status in critically ill patients with mechanical ventilation. Intensive Care Med
31. Oppenheimer L, Elings VB, Lewis FR. Thermal-dye lung water measurements: effects of edema and embolization. J Surg Res