INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The cardiovascular changes, particularly modification of the venous system, that occur in sepsis syndrome are poorly understood. Experimental studies in canine models have indicated a biphasic hemodynamic response to acute endotoxemia, representing the combined effects of changes in cardiac performance, peripheral vasomotor tone, and circulating blood volume (1, 2). However, maintaining either blood volume or ventricular filling pressure constant after endotoxin infusion invariably results in a hyperdynamic, hypotensive circulatory state (3). Such a hemodynamic pattern suggests a disproportionate impairment in peripheral vasoregulation affecting the arterioles, veins, or combination of the two. Within this framework, peripheral vascular compliance was found to be unmodified in experimental septic shock, whereas unstressed vascular volume was suspected to be increased (2).

Because the early cardiovascular abnormalities in response to endotoxin may differ significantly from those that evolve over time, and because a time-dependent assessment of cardiovascular dysfunction in human sepsis is difficult to obtain under emergency conditions, little is known about the role of capacitance vessels in subjects with clinical sepsis syndrome. However, it is important to consider the status of capacitance vessels because changes in vascular compliance and distensibility affect cardiac filling pressure and may contribute to the control of cardiac performance and neurohumoral adjustments of extracellular fluid volume and its distribution (4, 5). In recent years, adequate methods have been developed in normal and hypertensive populations to obtain operational quantitative indices of vascular compliance (6, 7). These methods, which require the rapid infusion of iso-osmotic and iso-oncotic fluids, may be difficult to optimize in subjects admitted to intensive care units (ICU), for two reasons. Firstly, in order to obtain reliable data, the investigation must be performed in the shortest time possible and under conditions of stable systemic blood pressure (6, 7). Secondly, the specific role of mechanical ventilation on capacitance vessels must be assessed and distinguished from the disease responsible for the emergency (8).

The purpose of the present study was to determine the value of effective compliance and distensibility in mechanically ventilated patients with sepsis syndrome in comparison with control subjects and mechanically ventilated patients without sepsis syndrome.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

The study was performed in 60 subjects (43 men and 17 women), including 46 mechanically ventilated and 14 nonventilated patients. Mean age was 56 ± 17 yr (range: 20 to 88 yr). Mean weight and height were 70 ± 14 kg and 171 ± 8 cm, respectively. The patient population was divided into three groups as follows.

Group I was composed of 14 patients admitted to the ICU for drug overdose (n = 11), acute respiratory failure (n = 2), and postoperative period following vascular surgery (n = 1). None of these patients were treated with mechanical ventilation, and no patients had signs of sepsis syndrome during their ICU stay as listed below. This group was used to validate the investigative procedure designed to evaluate the role of capacitance vessels in subjects with sepsis syndrome.

Group II was composed of 21 patients requiring mechanical ventilation for drug overdose (n = 7), coma (n = 6), postoperative period following coronary artery bypass grafting (n = 6), and acute respiratory failure (n = 2). No patients were suspected of having sepsis syndrome as defined by the criteria listed below, and no patients had suspected heart failure during their ICU stay.

Group III was composed of 25 patients with sepsis syndrome, who were treated by mechanical ventilation. Sepsis syndrome consisted of sepsis with clinical evidence of infection, associated with evidence of organ dysfunction as defined by Bone (9). Sepsis was due to postoperative wound infection (n = 7), pneumonia (n = 7), peritonitis (n = 3), and other causes (n = 8).

Severity of illness was evaluated using the APACHE II score (10). Clinical characteristics of Groups II and III are listed in Table 1.

The study protocol was approved by the institution's clinical investigation committee, and informed consent was obtained from all subjects or relatives after a detailed description of the procedure.

Blood Volume Expansion

The study was performed between the second and tenth day of hospitalization, after resolution of the acute phase of severe hypotension, when present. A jugular or subclavian vein was catheterized with a 20G catheter in all subjects. The catheter was advanced into the right atrium. The correct position of the catheter was confirmed by typical pressure recordings and chest radiograph. Central venous pressure (CVP) was measured with a pressure transducer (PVB, Kirchsceon, Germany). Pressures were obtained after calibration, zeroing to atmospheric pressure and using the midchest level as reference. Transducers were connected to bedside amplifiers (HP M10469102B; Hewlett Packard, Evry, France). Room temperature was 23 to 24° C.

Volume expansion was performed according to the procedure of Echt and associates (6), and modified by ourselves (7, 11). In order to minimize adverse effects due to capillary filtration and delayed compliance, the study was carried out in the shortest time possible. Briefly, 450 ml of iso-osmotic and iso-oncotic gelatin fluid (Plasmion; Roger Bellon Laboratory, Neuilly sur Seine, France) was infused over 6 min using a Jouvelet pump through the central venous catheter. Blood pressure was determined by arterial catheters in a radial or femoral artery or by a semi-automatic blood pressure device. Heart rate was recorded continuously by the electrocardiogram (HP M10469102B; Hewlett Packard). In all subjects, systemic blood pressure was verified to be stable for 24 h before and during the investigation. CVP was recorded at end-expiration immediately after infusion of 0, 75, 150, 250, 375, and 450 ml and plotted against volume changes (Figure 1). CVP measurements were accurate to the nearest 1 mm Hg. Based on correlation coefficients constantly  0.95, the pressure (y-axis)-volume (x-axis) relationship could be considered to be linear within the limits of the investigation. Elasticity coefficient (E)which characterizes the elasticity of the entire vascular systemwas calculated as the slope of the relationship, and the inverse of the slope (1/ E ± TBV/CVP) was calculated and standardized to body weight, with TBV representing the change in total blood volume (ml) and CVP the change in CVP (mm Hg) during volume expansion (Figure 1). The reproducibility of the method was tested in 24 patients. A first TBV/CVP ratio was determined after 150 ml of blood volume expansion. Twenty minutes later, a second TBV/CVP ratio was determined after 300 ml of blood volume expansion. The mean variation was 5.6%, similar to that observed in a previous study (7).

View larger version (8K): [in this window] [in a new window]   Figure 1.   A typical plot illustrating the time course of CVP changes following rapid volume infusion in a mechanically ventilated patient with sepsis syndrome. Total "effective" compliance is the inverse of the slope.

Echocardiography

In 25 patients, echocardiography (12, 13) was performed using a Hewlett-Packard Sonos 100 ultrasound device equipped with a 2.5 MHz probe. In order to exclude patients with cardiac abnormalities, M-Mode, two-dimensional, and Doppler echocardiography were performed before inclusion. Patients with left atrial dilation (> 4.0 cm), left ventricular dilation (left ventricular end-diastolic internal dimensions > 5.7 cm), decreased shortening fraction (< 31%) (14), regional wall motion abnormalities, valvular heart disease, cardiomyopathy, and pericardial disease were excluded from the study. Cardiac output (CO), using a pulsed Doppler two-dimensional method, was calculated by the formula: time-velocity integrals × cross-sectional area of the aortic annulus × heart rate. From the apical five-chamber view, the sample volume was placed in the middle of the left ventricular outflow immediately proximal to the leaflet of the aortic valve. The outflow velocity curves were digitized following the contour of the darkest portion of the curve. Measurements were taken from frozen images at the end of expiratory time and averaged five beats for each examination. From the long axis view, the cross-sectional area of the aortic annulus was calculated as 3.14r², where r represents half of the annular diameter during early systole (average of five cardiac cycles). Intra-observer variability was < 10%. Cardiac output was expressed either in ml/min or in ml × min¹ × m² (cardiac index: CI), and systemic vascular resistance was calculated according to the standard formula: SVR (mm Hg × L¹ × min¹) = (mean blood pressure CVP)/ CO. Because these measurements were performed just before and at the end of volume expansion, and because the cardiac output-blood volume relationship is known to be linear within such volume changes (7, 11), the CI/CVP ratio was used as an index of cardiac performance, where CI (ml × min¹ × m²) represented the change in CI, and CVP (mm Hg) the change in CVP. The SVI/CVP ratio was also calculated from the stroke volume index (SVI).

Statistical Analysis

The results are expressed as mean ± 1 SD. Comparisons between the three groups were performed by analysis of variance (ANOVA) followed by a Scheffe F-test for two-by-two comparisons of quantitative variables. Quantitative variables were compared using the Student's t test or Mann-Whitney U test when appropriate. The chi-square test or Fisher exact test were used to compare qualitative variables. The effect of volume expansion was studied using two-way ANOVA. Linear regression analysis and analysis of covariance (ANCOVA) (15) were performed to adjust the TBV/CVP ratio to age and baseline CVP. A p value  0.05 was considered to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effective compliance in control nonventilated patients (Group I) was similar to values previously reported by us and others: 2.03 ± 0.21 ml × mm Hg¹ × kg¹ (6, 11, 16, 17) (Table 2; Figure 2, upper panel). Values of cardiac output were 4.4 ± 0.42 and 5.6 ± 0.62 L/min before and after volume expansion, respectively (p < 0.001). The CI/CVP ratio was 0.20 ± 0.006 L × min¹ × m² × mm Hg¹.

View larger version (10K): [in this window] [in a new window]   Figure 2.   (Upper panel ) Value of TBV/CVP ratio in the three groups. (Lower panel ) Values of CI/CVP ratio in the three groups.

The clinical characteristics of patients with (Group III) and without (Group II) sepsis syndrome under mechanical ventilation are shown in Table 1. Compared with patients of Group II, patients of Group III were significantly older (p < 0.04) and had a lower Pa_O2/FI_O2 ratio (p < 0.001). Subjects of Group III also received more antibiotics (p  0.0001), vasoactive agents (p < 0.01), and fluid loading (p < 0.0007) prior to blood volume expansion.

The changes in heart rate, mean blood pressure, and CVP after volume expansion are shown in Table 2. Whereas heart rate did not change, blood pressure increased slightly in the three groups, as previously reported (7). However, despite a higher fluid loading in Group III, blood pressure never returned to the values observed in Groups I or II before or after volume expansion. Baseline CVP did not differ between Group II and III (4.19 ± 3.74 versus 6.28 ± 3.9 mm Hg, respectively). The TBV/CVP ratio was significantly lower in Groups II and III than in Group I (p < 0.0001); the ratio was also significantly lower (p < 0.0001) in Group III than in Group II (Figure 2, upper panel).

Patients of Group III were divided into two groups according to their initial CVP: 13 patients with CVP within the normal range ( 5 mm Hg) (Group IIIa) and 12 patients with elevated CVP (> 5 mm Hg) (Group IIIb) (Table 3). Fluid loading during the previous 24 h represented 615 ± 711 ml for Group IIIa and 958 ± 1,096 ml for Group IIIb (p = NS). The TBV/CVP ratio did not differ significantly between Group IIIa and IIIb (0.94 ± 0.23 versus 0.93 ± 0.26 ml × mm Hg¹ × kg¹). The same finding was observed when subjects of Group III were classified according to the presence or absence of vasoactive agents (0.90 ± 0.24 and 0.98 ± 0.23 ml × mm Hg¹ × kg¹, respectively).

Regression analyses performed on the entire population (n = 60) demonstrated that the TBV/CVP ratio was negatively correlated with age (r² = 0.23, y = 17.06x + 78.92, p = 0.0001) and baseline CVP (r² = 0.27, y = 4.12x + 10.16, p = 0.0001). Comparison of mean TBV/CVP in the three different groups may therefore be affected by differences in their mean baseline CVP and/or mean age. A comparison is therefore needed that corrects for discrepancies between mean age and mean baseline CVP of the various groups. The three groups were then compared by ANCOVA in order to adjust the value of the TBV/CVP ratio to age and baseline CVP as identified by the results of regression analyses. By comparison with Group I, the TBV/CVP ratio remained significantly lower in Groups II and III and was lower in Group III than in Group II (1.0 ± 0.05 versus 1.41 ± 0.05 ml × mm Hg¹ × kg¹, p = 0.0001), even after adjustment for age and baseline CVP (Figure 3) by ANCOVA analysis (sum of squares = 5.2, variance = 57.9, p < 0.0001).

View larger version (11K): [in this window] [in a new window]   Figure 3.   Graph showing baseline CVP and total "effective" compliance values in the three patient groups.

Analyses of the subgroup of seven Group II patients and 10 Group III patients with central hemodynamic measurements indicated that, under baseline conditions, patients of Group III had significantly lower values of mean blood pressure and systemic vascular resistance and a higher value of cardiac output. Cardiac performance, as evaluated by the CI/CVP ratio and SVI/CVP ratio, was not significantly different between the two groups (Figure 2, lower panel; Table 4). There was a significant increase in cardiac output in both groups after volume expansion. In these two subgroups, the TBV/ CVP ratio was 1.42 ± 0.16 ml × mm Hg¹ × kg¹ in the seven Group II patients and 1.0 ± 0.28 ml × mm Hg¹ × kg¹ in the 10 Group III patients (p = 0.003).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present investigation, the TBV/CVP ratio during rapid volume expansion was studied in normal subjects (Group I), in mechanically ventilated patients (Group II), and in mechanically ventilated patients with sepsis syndrome (Group III). Independently of age and baseline CVP, the TBV/CVP ratio was significantly lower in subjects receiving mechanical ventilation alone and was independently lower in mechanically ventilated subjects with sepsis syndrome.

Studies in laboratory animals and in humans have shown that blood volume changes and pressure values, measured in different segments of the venous circulation, are closely correlated under conditions of free flow (17, 18). The slope of the correlation between CVP and blood volume (TBV/CVP) under these circumstances has the units of compliance (6). However, cardiovascular physiology studies have extensively shown that the measurement of mean circulatory filling pressure under transient cardiac arrest is required (17, 18), in order to accurately determine the compliance of the total vascular system.

In the beating heart, CVP is not exclusively dependent on blood volume and the elastic properties of the vascular bed but also on the pumping capacity of the heart itself (17, 18). However, although changes in cardiac output induced by blood volume expansion exert only minimal effects on right atrial pressure and compliance measurements, reflex compensation with resulting neurohumoral effects on venous tone may affect evaluation of the TBV/CVP ratio (17). Temporary reflex blockade might be induced in order to determine whether reflex activity is important during blood volume change (20). In our study, heart rate did not change during volume expansion. To avoid any time-dependent variations in compliance and to minimize extravasation of the marker used to evaluate blood volume variations from the vascular volume to the interstitial space (6, 7, 11), all samples and parameters measured during TBV/CVP determinations must be performed in the shortest time possible. Under these conditions, the TBV/CVP ratio in humans behaves like a genuine biologic constant, ranging between 2.1 and 2.7 ml¹ × mm Hg¹ × kg¹ (16), as observed in our control patients. Such a measurement involves, in series, the various portions of the cardiovascular system including the veins and the right cavities, arterial system and left ventricle in diastole. However, the two latter territories have a very low compliance, 0.03 and 0.08 ml × mm Hg¹ × kg¹, respectively, in normal subjects (16). Finally, the TBV/CVP ratio must only be considered to be an "effective" index of total vascular compliance and in any case should be considered to be a marker of the intrinsic properties of the venous wall.

In the present study, mechanically ventilated patients (Group II) exhibited a substantial decrease in the TBV/ CVP ratio. Mechanical ventilation itself might be directly or indirectly responsible for a decrease in the "effective" compliance of the total venous bed. There are several arguments in favor of this interpretation. Firstly, intrathoracic pressure is increased in mechanically ventilated subjects (8), thereby contributing to an alteration of transmural pressure and CVP (see Table 2). Secondly, changes in CVP have been shown to vary substantially with ventilation (8). Finally, mechanical ventilation is known to be responsible per se for a slight decrease in venous return and cardiac output (8). Because reduced venous capacitance is the major factor contributing physiologically to maintenance of the filling pressure of the normal heart (18), it is suggested that the observed reduction in the TBV/CVP ratio in mechanically ventilated subjects (Group II) contributes significantly to maintenance of normal cardiac performance.

Analysis of this situation is more complex in subjects of Group III, i.e., mechanically ventilated with sepsis syndrome, as their baseline hemodynamic pattern differed from that of subjects of Groups I and II, with a probably different level of sympathetic tone. The TBV/CVP ratio may be difficult to interpret in these subjects, and several questions should be successively addressed. Firstly, there was no alteration in cardiac performance in subjects of Group III, since cardiac output was high and the CI/CVP ratio was preserved (Table 4). Secondly, because subjects with sepsis syndrome are known to have altered capillary filtration (21), this mechanism could have interfered with determination of the TBV/CVP ratio. However, in such a case, increased capillary filtration during the 6 min of blood volume expansion might have led to underestimation of the amount of fluid infused and therefore underestimation (and not overestimation) of the decrease in the TBV/CVP ratio. Thirdly, although changes in vascular compliance might reasonably be considered to be synonymous with changes in venous smooth muscle activity (19), the shift to a different curve of the TBV/CVP ratio could be due to a change in either unstressed volume or compliance or both (17, 19, 20, 24). TBV/CVP could yield information not only about vascular compliance but also about a much broader system including local and central vascular regulation. Finally, our data suggest changes in the elastic characteristics of the circulation (unstressed volume and/or compliance are decreased) as previously suggested in experimental models (25).

Because the venous pressure-volume relationship is known to be curvilinear and because CVP measured under baseline conditions was significantly higher in subjects of Group III than in subjects of Group I, the reduced "effective" venous compliance observed in Group III could simply be a consequence of the curvilinearity of the normal (Group I) or ventilatory-induced (Group II) pressure-volume curve. However, the present study provides several strong arguments against this interpretation. Firstly, when subjects of Group III were classified according to the presence of normal (Group IIIa) or elevated (Group IIIb) baseline CVP, the two subgroups had exactly the same TBV/CVP ratio (Table 4). Secondly, the ratio was not influenced after adjustment for any difference in mean age and mean baseline CVP by ANCOVA analysis. Such findings imply that subjects with sepsis syndrome (Group III) had a reduced "effective" compliance of the overall venous bed and that their own pressure-volume curve was significantly different from that of subjects of Groups I and II. This assumption agrees with the recent finding that forearm venous tone is significantly increased in subjects with sepsis syndrome (26).

In sepsis syndrome, alterations of the veins may be due to passive or active mechanisms (4, 18, 19). Because arteriolar vasodilatation is present and even enhanced in the presence of rapid volume expansion, a passive response of the veins cannot be excluded (18, 19). On the other hand, although the drugs used in the clinical management of sepsis syndrome did not appear to significantly influence the value of the TBV/ CVP ratio (see RESULTS), numerous neurohumoral factors involved in the disease might have modified the active properties of the veins (18, 19, 27). This process may occur either as a primary defect or as an adaptive mechanism in patients with a severe underlying medical condition and physiologic disturbances, as suggested by the following two observations. First, few changes in vascular compliance have been previously observed during the early phase of experimental septic shock (2). Second, we observed that the TBV/CVP ratio was more markedly reduced in the presence of more severe disease, as assessed from the severity scoring system, and the disturbances of gas exchanges (Table 1). It is difficult to distinguish the direct effects of sepsis on venous compliance vessels from the reflex effects of hypotension, on the basis of these findings (28). The fall in blood pressure would be expected to have elicited increases in sympathetic nervous system tone, which in turn could decrease venous compliance even in the absence of sepsis.

Finally, a compensatory process appears to increase the effective circulating blood volume and venous return in sepsis syndrome and septic shock. Whatever the mechanisms involved, strong interactions appear to exist between reduced "effective" venous vascular compliance, degree of systemic hypotension, and severity of disease. For therapeutic purposes, knowledge of the total "effective" vascular compliance could be useful to guide fluid infusion and evaluate cardiac function. Moreover, regardless of the degree of fluid loading in septic patients, mean blood pressure appears to remain constantly lower than in control or mechanically ventilated nonseptic patients. Correction of arteriolar vasodilation and abnormalities of venous return curves would be a logical approach to the treatment of septic shock (29).