Volume therapy is a key component in the treatment of critically ill and postoperative patients. However, to achieve an adequate volume status in hypovolaemic patients, most clinicians do not measure volumes but rely instead on surrogate variables such as cardiac filling pressures. Venous return, and hence the circulating volume, are deduced from central venous pressure (CVP) and pulmonary artery occlusion pressure (PAOP). Although its limited value for estimation of blood volume status is known , 'fluid challenge' guided by CVP or PAOP, as introduced by Weil and Shubin , is a widely used method to control volume therapy.
A frequently disregarded fact is that cardiac filling pressures reflect the appropriateness of the pump function of the heart rather than appropriateness of the intravascular volume . A low cardiac output indicates circulatory failure. The cause of that failure might be myocardial dysfunction as well as hypovolaemia. Myocardial dysfunction is characterized by elevated filling pressures, whereas hypovolaemia usually is associated with low filling pressures. On the other hand, 'normal' filling pressures do not exclude hypovolaemia if the functional status and the compliance of the myocardium (i.e. the slope and the position of the ventricular function curve) are unknown [4,5]. The intravascular volume has two major effects on cardiac output. First, it is in close relationship with intracardiac end-diastolic volumes and thus acts on cardiac output via the Frank-Starling relation between the preload and ventricular stroke work. Second, it determines venous return that is equal to cardiac output.
According to basic physiological principles, 'stressed vascular volume' is the main determinant of venous return to the heart, and thus of cardiac output . The stressed vascular volume represents that part of the total blood volume that stretches the vessel wall and creates the pressure that drives blood back to the heart. Venous return depends on vascular volume as well as on vascular compliance. Stressed vascular volume is regarded as the functional proportion of the intravascular volume that has a close relationship to venous return.
Direct measurement of total blood volume by different techniques [7-10] has not gained acceptance in clinical use because of complicated procedures or potential side-effects. A recently approved bedside-monitoring device for intravascular blood volumes is reliable and convenient for clinical practice . A fibreoptic thermistor catheter placed in the descending aorta, combined with central venous injections of ice-cold indocyanine green, allows for quantification of the total circulating blood volume (TBVcirc)  and the intrathoracic blood volume (ITBV)  by a double indicator dilution technique . This kind of monitoring enables physicians to think in terms of intravascular volumes instead of pressures in the management of volume therapy.
The aim of the present study was to describe the relationships between measured volumes (TBVcirc, ITBV) and cardiac-filling pressures (CVP, PAOP) by examining their ability to reflect preload dependence of cardiac output (i.e. to achieve the optimal position on the ventricular function curve). For that reason, we monitored the correction of a circulatory dysfunction (i.e. hypovolaemia) by volume replacement, indicated by an increase of cardiac index.
Written informed consent was given by each patient scheduled for cardiac surgery and the Institutional Ethics Committee approved the study protocol.
Haemodynamic measurements were performed with a standard thermodilution pulmonary artery catheter. CVP and PAOP were determined by pressure transducers for electronic measurement. Data for these pressures were recorded in supine patients at the end of expiration after calibrating the transducer to the mid-axillary line as the zero reference point. Cardiac output was measured by a conventional thermodilution technique, using the average of five consecutive 10 mL boluses with ambient temperature randomly distributed over the respiratory cycle. Cardiac outputs were measured independently by transpulmonary thermodilution and by a transpulmonary dye dilution technique and were not statistically different from thermodilutions obtained by the pulmonary artery catheter. The latter were used for calculations of correlations between SVI and other variables. Intravascular volumes were measured with a double indicator-dilution technique (COLD System®; Pulsion Corp., Munich, Germany). The method has been validated previously [11-13]. Of indocyanine green, 30 mg in 15 mL ice-cold normal saline was injected into the right atrium and the resulting indicator-dilution curves were registered with a fibreoptic catheter (4 FFT-Pulsiocath®; Pulsion Corp., Munich, Germany) placed in the abdominal aorta through the femoral artery. ITBV represents the volume of distribution between the injection and detection site  and is calculated from the mean transit time of the indocyanine green concentration-time course and the cardiac output. Total circulating blood volume (TBVcirc) was automatically calculated by the COLD System from the gently sloping section of the dye curve after a 3 min mixing-time of the indicator with blood [11,12].
Hypovolaemia was studied in post-cardiac surgery patients. Fourteen patients (11 of whom were male) met the inclusion criteria: 11 of them underwent coronary artery bypass grafting and three had valve surgery. All patients had a pulmonary artery catheter in place and were anaesthetized and mechanically ventilated throughout the study period. The mean (range) age of the patients was 62.3 (32-79) yr, the mean height was 170.1 (156-180) cm, and the mean weight was 74.6 (55-95) kg. The mean body surface area was 1.86 (1.53-2.09) m2. Preoperative ejection fraction ranged from 0.49 to 0.62. Inclusion criteria were: hypovolaemia after cardiac surgery in adult patients during mechanical ventilation of the lungs. Exclusion criteria were: persisting active bleeding (>100 mL h−1 drained from the chest tubes), haematocrits < 30%, catecholamine therapy and extubation during the study period.
Patients were screened for hypovolaemia at admission to the intensive care unit. Hypovolaemia was by our definition a PAOP and CVP <10 mmHg, combined with a cardiac index (CI) <2.8 L min−1 m−2. These limits were supposed to be indicative for hypovolaemia during mechanical ventilation. Baseline measurements of CI, SVI, CVP, PAOP, ITBVI and TBVIcirc were performed in all hypovolaemic patients.
Patients who fulfilled the criteria of hypovolaemia received 1000 mL 6% hydroxyethyl starch (HES). The volume was infused over 30 min. Haemodynamic measurements were repeated 5 min after conclusion of infusion (volume loading phase). No further infusions were started during a steady-state phase of 60 min, during which all measurements were repeated.
The lung ventilators were set to intermittent positive pressure ventilation with tidal volumes of 8-10 mL kg−1 body weight, respiratory rates of 8-10 breaths min−1, and a positive end-expiratory pressure of 7 cmH2O. Blood gases were determined before study enrolment and ventilator settings remained unchanged in all patients throughout the measurement period without abnormal values of arterial blood gases.
Repeated measures ANOVA was used for comparisons of haemodynamic and volume parameters between different measurements. Post hoc comparisons were performed by the Scheffé test. P (5% level of significance) during volume loading indicated statistically significant differences of the means of the first measurements (hypovolaemia) compared with measurements after correction of hypovolaemia (end of infusion). During the steady-state phase, P indicated the significant differences between the beginning (end of infusion) and the end of the steady-state phase (1 h after the end of infusion).
Linear regression analyses of percental changes of the stroke volume index (SVI) and the variables CVP, PAOP, intrathoracic blood volume index (ITBVI) and total circulating blood volume index (TBVIcirc), respectively, were performed both in the volume loading and in the steady-state phase.
Volume loading phase
Cardiac performance improved significantly after administration of 1000 mL HES (Table 1) with a subsequent increase in CI, SVI and cardiac filling pressures (i.e. CVP and PAOP). Measured intravascular volumes (i.e. ITBVI and TBVIcirc) were also significantly augmented after volume administration.
The changes of filling pressures and measured volumes were compared with changes of corresponding stroke volumes (Fig. 1). No correlation was found between SVI and CVP (r2 = 0.06) and between SVI and PAOP (r2 = 0.03) as well. However, changes of SVI and ITBVI and changes of SVI and TBVIcirc correlated (r2 = 0.67 and 0.55, respectively).
CI, SVI, ITBVI and TBVIcirc remained nearly unchanged during the steady-state phase. In contrast, CVP and PAOP decreased significantly (Table 1).
There was no correlation (Fig. 1) between changes in SVI and changes in CVP (r2 = 0.17) and between changes in SVI and changes in PAOP (r2 = 0.00). Changes in SVI correlated well with changes in measured intravascular volumes (r2 = 0.83 between SVI and ITBVI and r2 = 0.55 between SVI and TBVIcirc).
We investigated whether measured volumes or pressures would be more reliable to monitor the effects of 1000 mL HES in hypovolaemic patients after cardiac surgery. The infusion of a fixed amount of volume resembles the approach of 'volume challenge' . HES was chosen as the volume replacement fluid because of its known ability to guarantee a constant intravascular volume effect throughout the study . The intravascular volumes in the study correspond to the ranges found under similar conditions from other researchers. Before volume therapy, ITBV was 21.8 ± 7.0 mL kg−1 in our hypovolaemic patients. Nearly identical values (20.4 ± 4.0 mL kg−1) were determined in a previous study on hypovolaemic patients after cardiac surgery . A pronounced decrease of intrathoracic blood volume may also be the result of the vasodilatory effects of general anaesthesia or due to the increase of intrathoracic pressure by mechanical ventilation. Hedenstierna and colleagues  found a change of ITBV before (23.3 ± 1.8 mL kg−1) and after induction of anaesthesia (19.3 ± 1.6 mL kg−1). Similar to previous data , we found TBVcirc = 45.8 ± 13.6 mL kg−1 (1806 ± 502 mL m−2) in our hypovolaemic patients and 56.1 ± 10.4 mL kg−1 (2232 ± 396 mL m−2) after the steady-state phase. Hoeft and colleagues  found that the circulating blood volume was 43.00 ± 7.50 mL kg−1, which they measured 1 h after patients' admission to the intensive care unit. This is equivalent to the TBVcirc in the present study. Hoeft and colleagues  pointed out that the circulating blood volume, measured by a tracer distribution time of 3 min, represents a quickly perfused compartment of the vasculature and not the whole blood volume of the circulation.
The results of the present study demonstrate that the circulatory function is more closely related with measured intravascular volumes than with right/left-sided cardiac filling pressures, at least in mechanically ventilated patients after cardiac surgery. The infusion of 1000 mL HES increased filling pressures and intravascular volumes significantly. However, filling pressures did not correlate with the effects of fluid administration, which was an increase of stroke volume in hypovolaemic patients. This is not surprising, as the cardiac function curve, which is the relationship of cardiac output to the preload, varies from patient to patient and plateaus at an individual level . Depending on the position of the cardiac function curve, a small increase of cardiac output may correspond to large or small increases of filling pressures. According to the Frank-Starling concept, this holds also true for the relationship between myocardial contractile force and end-diastolic volume. A possible explanation for the observed discrepancy between changes of pressures and changes of volumes is that the individual myocardial compliance of the patients interferes with the correlation of intracardial volumes and pressures.
Transoesophageal echocardiography is thought to be one of the best means to determine left ventricular preload in clinical practice and could have been taken into consideration as a reference method. However, there is wide interindividual variability in the end-diastolic area index that jeopardizes the superiority of this parameter in comparison with PAOP .
During the steady-state period of 1 h, no intravascular fluids were added or lost. Means of either cardiac output or intravascular volumes did not change significantly during this period. The observed significant decreases of mean CVP and mean PAOP cannot be explained by alterations of myocardial compliance. It is assumed that in the absence of ischaemic events or administration of catecholamines, the compliance of the myocardium remained constant during that time. The decrease of the means of filling pressures represents a decrease of preload without an effect on cardiac output. On the other hand, haemodynamic changes under these conditions may be attributed to changes in vascular compliance. Changes of vascular compliance are common during rewarming and recovery from anaesthesia. The total blood volume comprises of an 'unstressed volume' that just allows the vessels to form their natural round shape and of the 'stressed volume' that stretches the vessel wall. By that concept, only the stressed vascular volume [3,6] determines venous return and, hence, cardiac output. Changes in vascular resistance or venous capacity affect cardiac output and cardiac filling pressures to various degrees, depending on the shape of the cardiac function curve. In contrast to filling pressures, relative changes of measured intravascular volumes ITBV and TBVcirc correlated well with relative changes of stroke volumes, both during volume loading and steady-state conditions.
The present data confirm with the results of Lichtwarck-Aschoff and colleagues , who found a good association of changes of ITBV with those of cardiac output, but not with changes of cardiac filling pressures. In contrast to the present study, Lichtwarck-Aschoff and colleagues made their observations in critically ill patients under different modes of ventilation. In an investigation on 30 uncomplicated patients who underwent coronary artery bypass grafting, Gödje and colleagues  found the following results: a high correlation between changes of SVI and changes of ITBV and a similar high correlation between changes of CI and ITBV; the absence of a correlation between changes of CVP and ITBV and between changes of PAOP and ITBV. Hoeft and colleagues  also reported a high correlation between measured blood volume and cardiac function (r2 = 0.76 between circulating blood volume and SVI).
Presently, sufficient data about the correlation between measured volumes and cardiac output on the one hand and the missing correlation between cardiac filling pressures and cardiac output on the other exist, but physiological explanation is still lacking. Our proposed association of measured intravascular volumes and stressed vascular volume may augment the understanding of that issue. In this context, it is also apparent that measurement of intravascular volumes is superior to preload variables for the assessment of the volume status in hypovolaemic patients after cardiac surgery.
The first data of measured stressed vascular volume for human beings  (20.2 ± 1.0 mL kg−1) were obtained from five patients during hypothermic circulatory arrest. This measured volume represented a proportion of 30 ± 17% of the total predicted blood volume . This is similar to the range of ITBV in the present study. There is some evidence that this similarity might not be by chance. Stressed volume is defined as the volume that contributes to the mean systemic pressure [23,24]. It is a basic concept that volume, due to vascular wall stress of the venous system, is driven intrathoracically (i.e. increasing ITBV) and that volume influenced by arteriolar constriction is retained in the thorax (i.e. increasing afterload). Although ITBV and stressed vascular volume are not likely to represent an identical morphologic entity, both volumes have strong functional interrelations.
Stressed vascular volume determines cardiac output and ITBV correlates with cardiac output more closely than any other haemodynamic variable. Both do not correlate with right or left heart filling pressure because these variables reflect cardiac function and are linked to cardiac output by the individual cardiac function curve . Changes of stroke volume during volume loading as well as changes during steady-state correlate with ITBV as expected from stressed vascular volume.
Although ITBV resembles many features of stressed vascular volume, further clinical studies are necessary to confirm the proposed functional identity of the two entities. Nevertheless, the bedside assessment of intravascular volume status by fibreoptic dye measurements has proven to be useful and reliable in patients with critical preload requirements . The considerations derived from the data of this investigation may add another aspect to the ongoing debate on the use of right heart catheterization. A possible explanation for the observed higher mortality in critically ill patients with pulmonary artery catheters  is that the information provided by measurement of haemodynamic variables (CVP, PAOP, cardiac output) may lead to incorrect therapeutic decisions. Aggressive use of fluids and inotropic agents guided by pulmonary artery catheters are potentially harmful to critically ill patients . This must be viewed from the standpoint of the limited reliability of intravascular pressures to reflect intravascular volume status and thus, the actual requirements of intravenous fluids.
1. Shippy CR, Appel PL, Shoemaker WC. Reliability of clinical monitoring to assess blood volume
in critically ill patients. Crit Care Med
2. Weil MH, Shubin H. The VIP approach to the bedside management of shock. J Am Med Assoc
3. Guyton AC, Jones CE, Coleman TG. Circulatory Physiology: Cardiac Output and its Regulation.
Philadelphia, USA: WB Saunders, 1973.
4. Raper R, Sibbald WJ. Misled by the wedge. The Swan-Ganz catheter and left ventricular preload. Chest
5. Hansen RM, Viquerat CE, Matthay MA, et al.
Poor correlation between pulmonary arterial wedge pressure and left ventricular end-diastolic volume after coronary artery bypass graft surgery. Anesthesiology
6. Magder S. Shock physiology. In: Pinsky MR, Dhainaut JF (eds). Pathophysiologic Foundations of Critical Care.
Philadelphia, USA: Williams & Wilkins, 1993: 140-160.
7. Keith NM, Rowntree LG. A method for the determination of plasma and blood volume
. Arch Intern Med
8. Gibson JG, Evans WA. Clinical studies on the blood volume
. I. Clinical application of a method employing the azo dye 'Evans Blue' and the spectrophotometer. J Clin Invest
9. Fine J, Seligman AM. Traumatic shock IV. A study of the problem of the 'lost plasma' in hemorrhagic shock by the use of radioactive iodoplasma protein. J Clin Invest
10. Bradley EC, Barr JW. Determination of blood volume
using indocyanine green dye. Life Sci
11. Hoeft A, Schorn B, Weyland A, et al.
Bedside assessment of intravascular volume status in patients undergoing coronary bypass surgery. Anesthesiology
12. Kisch H, Leucht S, Lichtwarck-Aschoff M, Pfeiffer UJ. Accuracy and reproducibility of the measurement of actively circulating blood volume
with an integrated fiberoptic monitoring system. Crit Care Med
13. Lichtwarck-Aschoff M, Zeravik J, Pfeiffer UJ. Intrathoracic blood volume
accurately reflects circulatory volume status in critically ill patients with mechanical ventilation. Int Care Med
14. Lewis FR, Elings VB, Sturm JA. Bedside measurement of extravascular lung water. J Surg Res
15. 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 Critical Care Monitoring.
Berlin, Germany: Springer, 1990: 114-125.
16. Brock H, Rapf B, Necek S, et al.
Vergleichende untersuchungen zur postoperativen volumentherapie bei kardiochirurgischen patienten. Anaesthesist
17. Hedenstierna G, Strandberg A, Brismar B, et al.
Functional residual capacity, thoracoabdominal dimensions, and central blood volume
during general anesthesia with muscle paralysis and mechanical ventilation. Anesthesiology
18. Magder S. More respect for the CVP. Int Care Med
19. Tavernier B, Makhotine O, Lebuffe G, Dupont J, Scherpereel P. Systolic pressure variation as a guide to fluid therapy in patients with sepsis-induced hypotension. Anesthesiology
20. Gödje O, Peyerl M, Seebauer T, et al.
Central venous pressure, pulmonary capillary wedge pressure and intrathoracic blood volume
as preload indicators in cardiac surgery patients. Eur J Cardiothorac Surg
21. Magder S, De Varennes B. Clinical death and the measurement of stressed vascular volume. Crit Care Med
22. Isselbacher KJ, Braunwald E, Wilson JD, et al.
(eds). Harrison's Principles of Internal Medicine,
13th edn. New York, USA: McGraw-Hill, 1994: 2489-2496.
23. Guyton AC, Lindsey AW, Kaufmann BN. Effect of mean circulatory filling pressure and other peripheral circulatory factors on cardiac output. Am J Physiol
24. Rothe CF. Reflex control of veins and vascular capacitance. Physiol Rev
25. Connors AF Jr, Speroff T, Dawson NV, et al.
The effectiveness of right heart catheterization in the initial care of critically ill patients. SPORT investigators. J Am Med Assoc
26. Hayes MA, Timmins AC, Yau EH, et al.
Elevation of systemic oxygen delivery in the treatment of critically ill patients. N Engl J Med