INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sepsis is a common clinical syndrome in critically ill patients. Cardiovascular dysfunction resulting in septic shock develops in 40% of patients presenting with this syndrome (1), and circulatory insufficiency in sepsis (septic shock) is a grave prognostic sign (2). Furthermore, if cardiovascular collapse is not treated effectively, septic shock leads rapidly to ischemic organ dysfunction and death (2). Although loss of peripheral vasomotor tone represents an important aspect of the shock process, the role of impaired cardiac function in septic shock is poorly defined. Some animal models, such as the guinea pig and murine models, show immediate cardiac dysfunction in response to endotoxemia (3), whereas larger animal models, involving ovine and canine preparations, show an apparent lag time of more than 6 to 24 h between exposure to endotoxin and the subsequent development of cardiac dysfunction (4). Knowledge of the immediate effect of sepsis on cardiac function would be relevant in defining appropriate resuscitative therapies for septic shock.

In patients with septic shock of more than 24 h duration, decreased basal vasomotor tone, with hyporesponsiveness to vasopressor agents (5), as well as depression of both diastolic and systolic myocardial function, have been characterized (6). In sustained sepsis, the cardiovascular depression appears to be mediated initially through activation of proinflammatory cytokines, such as tumor necrosis factor (TNF)- (7), arachidonic acid metabolites (10), and excess production of nitric oxide (NO) (11, 12). Although NO may play a role in the myocardial depression of prolonged sepsis (13), its functions during the initial phases of septic shock are not known.

We previously documented that the cardiovascular determinants of the initial hemodynamic response to an acute bolus infusion of Escherichia coli endotoxin in the canine model are complex. If, after an initial 30-min transitory phase, if fluid resuscitation sustains cardiac output (CO), a hypotensive, hyperdynamic cardiovascular state develops that remains stable for at least 4 h. This state is characterized by a reduction in arterial outflow (critical closing) pressure, increased capillary permeability, lactic acidosis, and increased CO. However, using simple measures of pressure and flow, we were unable to identify either a change in systemic vascular conductance or left ventricular (LV) function during this cardiovascular state. In contrast, both canine and porcine models of sepsis studied more than 4 to 6 h after the induction of endotoxemia have shown decreases in peripheral vasomotor tone (change in the slope of the arterial pressure-flow relationship), vascular responsiveness, and LV contractility (5, 9).

Most prior studies of endotoxemic shock did not directly measure LV contractility (14, 15) because they did not assess ventricular pump function from preload- and afterload-inclusive measures of contractile function, such as the slope of the LV end-systolic pressure-volume relationship (ESPVR), known as end-systolic elastance (Ees) (16), and preload-recruitable stroke work (PRSW) (17). Thus, if LV diastolic compliance had increased, then depression of LV contractility could have occurred early in endotoxic shock, without being identified by studying the relation between LV filling pressure and stroke work, because similar LV filling pressures would have generated higher end-diastolic volumes. Moreover, reflex catacholamine release in response to the initial hypotensive episode may have sustained cardiac contractility despite a reduced basal contractile state. If this were the case, however, the subsequent increase in contractility in response to a fixed amount of dobutamine, a potent inotropic agent, would be blunted.

We therefore examined the effects of acute bolus-induced endotoxemia on LV function when changes in LV diastolic compliance, catecholamine stimulation, and NO synthesis were taken into account. To assess LV contractility we used the robust and highly sensitive measures of ventricular function described earlier: Ees, PRSW, and measures of diastolic relaxation. We reasoned that if increased catecholamine levels masked impaired LV contractility, there would be a reduced responsiveness of measures of cardiac contractility to a stepwise increase in exogenous catecholamine infusion. Similarly, if NO synthesis were impairing or supporting LV function, then its blockade by L-nitro arginine methyl ester (L-NMMA) or augmentation with L-arginine infusion would reciprocally alter LV function (18). Our data show that LV contractility in the dog was unaffected by bolus infusion of endotoxin over the initial 4-h interval despite systemic hypotension, lactic acidosis, and the expression of a hyperdynamic state. Thus, the initial cardiovascular depression seen in early canine endotoxemia primarily reflects a peripheral vascular response, with preserved cardiac contractility and contractile reserve.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surgical Preparation

The Animal Care and Use Committee of the University of Pittsburgh approved the study. After a 14-h fast during which they received only water, 14 male mongrel dogs (weight range: 18.2 to 21.6 kg) were anesthetized with pentobarbital sodium (30 mg/kg, intravenously). Each animal had its trachea intubated with a size 9-Fr cuffed endotracheal tube, and was ventilated (Servo 900B; Siemens Elema, Lund, Sweden) at a tidal volume of 12 ml/kg and a frequency sufficient to maintain a Pa_CO2 at 38 to 42 mm Hg. This was monitored continually by measurement of end-tidal CO₂ (ETCO₂) (Hewlett-Packard, Palo Alto, CA). Arterial blood gases were periodically sampled, and arterial pH was maintained at 7.37 to 7.42 by intravenous infusion of NaHCO₃ as necessary. A size 5-Fr catheter with multiple side holes was inserted into the right femoral artery for measurement of arterial pressure. A size 7-Fr polyvinyl chloride catheter was inserted into the intraabdominal inferior vena cava via the right femoral vein for continuous infusion of pentobarbital at 2 to 4 mg/kg/h and for venous blood sampling. A size 7-Fr balloon-tipped pulmonary artery thermodilution catheter, with a 15-cm proximal port (Model 93A-095; American Edwards, Irvine, CA), was advanced into the pulmonary artery through the right external jugular vein. Placement of the proximal port of the catheter in the right atrium was verified by waveform analysis of the pressure tracing, and placement of the catheter tip in a pulmonary artery was verified fluoroscopically. All vascular catheters were connected with MP-50 pressure transducers (Gould, Inc., Cleveland, OH) and the airway catheter from the Y-connector of the ventilator circuit to a 23-dB pressure transducer (Gould). A size 5-Fr high-fidelity pressure transducer (Micro-tip catheter, Model SPC-350; Millar, Houston, TX) was inserted via the left carotid artery into the left ventricle under fluoroscopic guidance to record LV pressure. To measure LV volume on a continuous basis, a size 6-Fr, 11-pole, multielectrode, dual-excitation conductance catheter (custom built by Webster Laboratories, Irvine, CA) was inserted under fluoroscopic guidance through the aortic valve and into the left ventricle via the right internal carotid artery. Positioning of the conductance catheter was verified through both two-plane fluoroscopy and repeated inspection of the regional sequential volume signals, as recommended by the manufacturer.

A midline laparotomy was performed. To minimize intravascular volume shifts during the study, splenectomy was done after maximal splenic contraction in response to 0.5 ml of topical epinephrine (1:10,000). The infrahepatic vena cava was isolated, and a hydraulic vascular occluder was placed around it. The splenic vein was cannulated with a size 5-Fr polyethylene catheter that was passed into the portal vein to the level of the porta hepatis. The portal vein was then isolated, and an ultrasonic flow probe (Transonic Systems, Ithaca, NY) was placed around it for the purpose of a parallel study. The abdomen was loosely closed with two layers of interrupted sutures. All animals were fluid-resuscitated with normal saline as necessary to maintain the right atrial pressure (Pra) at 2 to 5 mm Hg. After this, the animals were allowed to stabilize for at least 30 min. Stability was defined as a constant heart rate (HR), arterial pressure, and ETCO₂, and constant organ flow signals, for at least 30 min.

Calibration of the Conductance Catheter

The conductance catheter method for measuring ventricular volume has been described and validated previously (19, 20). Briefly, a 20-kHz, constant-amplitude current of 30 µA (root mean square) is passed between the electrodes of the distal and proximal extremes of the catheter in a dual-field format. This dual-field format reduces changes in parallel conduction artifact, and stabilizes the volume signal over time (21). An electrical field is generated within the ventricular cavity in which equipotential planes, at right angles to the long axis of the catheter, are defined between each of the intervening electrodes. The volume of blood measured between any two sensing electrodes is considered to be a disc with boundaries defined by the endothelial surfaces through to the electrodes. The change in conductance sensed during ventricular contraction in any one of these discs is caused by a change in resistance in the cross-sectional area of the disc. The sum of all such discs within a chamber such as the ventricle reflects total chamber volume. Parallel conductance artifact was determined with the saline dilution method (22), using 5 ml of 10 N NaCl injected into the right atrium. We and others (19, 21) have previously demonstrated in dogs that conductance catheter-derived measures of LV volumes and the parameters derived from them accurately follow both echocardiographic LV area signals and aortic flow-probe stroke-volume measures, and are unaltered by catecholamine infusions (23).

Data Processing

We recorded the following experimental variables: Pra, pulmonary artery pressure (Ppa), arterial blood pressure (Pa), LV pressure (Plv), LV volume, and the signal from lead II of the electrocardiogram (ECG). All data were recorded on an eight-channel physiologic recorder (Gould). Additionally, all data were digitized and recorded on magnetic disc for subsequent analysis as described in the following discussion. Signals were collected and processed with a conductance catheter data processor and signal conditioner (Sigma 5DF; Leycom, Leiden, The Netherlands). A four-electrode calibration chamber was used to determine blood conductivity ( value). Intermittent CO values were determined in triplicate, using a thermodilution technique with a 10-ml bolus of ice 0.9% saline (Model 9520 cardiac output computer; American Edwards) as a guide to cross-reference the conductance catheter-derived measurements of stroke volume. The signal acquisition and analysis system consisted of a Hewlett-Packard/Apollo DN4000 Unix workstation and a SignifiCAT RTS-132 A/D subsystem.

The LV pressure-volume loops were compared simultaneously on a large video screen to aid in defining steady-state conditioning. Post hoc analysis of myocardial function was done by replaying digitized data stored on magnetic hard disc during both apneic steady-state and rapid inferior vena cava (IVC) occlusion-induced decreases in LV end-diastolic volume. The LV ESPVR was calculated by the iterative technique, using data from IVC occlusion runs and the method of least squares as previously described (24, 25). The slope of the ESPVR curve was taken to reflect Ees, whereas the change in the slope of the ESPVR curve in response to pharmacologic intervention was taken to reflect immediate changes in contractility caused by exogenously induced changes in the cardiovascular state. We also calculated PRSW, as the slope of the LV stroke work-to-LV end-diastolic volume relationship during a rapid IVC occlusion maneuver, as previously described (17). Both the absolute PRSW value and its change in response to dobutamine infusion were calculated for all animals. Total myocardial oxygen demand was assumed to reflect the LV pressure-volume area (PVA), which encompasses the area inside the LV pressure-volume relationship reflected by the pressure-volume loop during steady-state apnea, plus the Ees-defined triangular area to its left, with volume on the x axis defined by the ESPVR, LV isometric relaxation, and an extension of the LV diastolic compliance relation toward the origin. LV ejection efficiency was calculated as the ratio of the LV stroke work to the LV PVA. LV diastolic function was also assessed, by measuring parameters of relaxation, including the maximal rate of LV pressure decay during isometric relaxation (the maximum negative change in left ventricular filling pressure versus time; LV dP/dt) and the maximal rate of LV filling (the change in left-ventricular volume versus time; LV dV/dt) as previously described (26). We also calculated arterial elastance (Ea), as an estimate of arterial tone, from the LV pressure-volume relationship, as the ratio of the stroke volume to the mean Pa over a 5-s apneic period, as previously described (27) and previously validated by us (28). We chose not to calculate vascular resistance, because of the inaccuracies to which its calculation may be subject and because we have previously shown it to be inaccurate in the model used in our study (15). Additionally, to assess ventriculoarterial coupling, we examined the Ea-to-Ees ratio. We assumed that ratios between 0.8 to 1.2 reflected normal ventriculoarterial coupling.

Protocol

Hemodynamic data were collected during the baseline preendotoxemic state and then every 30 min thereafter for 4 h or more. These data included all measured vascular and airway pressures and HR. The data were collected during 15-s periods of apneic steady-state conditions. First, the data were recorded over the initial apneic steady-state interval. The LV pressure-volume data were recorded during a brief interval of rapid IVC occlusion (5 to 7 s). From these data, we calculated LV ESPVR, PRSW, and Ea. To assess the initial humoral sympathetic state prior to endotoxin infusion, we gave a bolus infusion of esmolol (5 mg intravenously) and observed the hemodynamic response 1 min later. This dose of esmolol was chosen because it reproducibly impaired LV contractile function without inducing profound changes in arterial tone. Fifteen minutes after the esmolol bolus, and once the cardiovascular status had returned to baseline values, we collected an additional baseline set of data. Then, to ascertain the degree of catecholamine responsiveness of the heart, a subsequent set of data were collected during an infusion of dobutamine (4 µg/kg/min; Eli Lilly, Inc., Indianapolis, IN) with a Harvard infusion pump (Harvard Instruments, Cambridge MA). This dose of dobutamine was chosen, because it reproducibly induced increases in measures of LV contraction in both normal animals and animals with stunned myocardium after brief intervals of ischemia. These additional data were collected after 10 min of dobutamine infusion, once a hemodynamic steady state occurred. In practice, this steady state developed within 2 to 3 min. The dobutamine infusion was then stopped and the animal was allowed to recover. Identical paired data sets were collected again, before and during dobutamine infusion, at 2 h and 4 h after allocation of animals to endotoxin or control groups. Once these final data were collected, we infused L-NMMA (0.15 mg/kg/h), a competitive inhibitor of NO synthase (NOS), and collected another set of data after 15 min, in order to ascertain the degree to which NO metabolism was affecting cardiac contractility. Additionally, to ascertain whether augmented NO synthesis was possible and could alter contractility, we infused a bolus of L-arginine (150 mg) and collected a final data set 5 min later.

Animals were assigned to endotoxin or control groups through a lottery conducted after the initial data were collected before endotoxin administration. Eight animals received endotoxin and are referred to as the endotoxin group, and seven animals did not receive endotoxin and are referred to as the control group. Acute endotoxemia was induced by the infusion of E. coli endotoxin (L-2880 lipopolysaccharide; Sigma, St. Louis, MO) in a dose of 1 mg/kg over a 5-min period via the right atrial port of the thermodilution catheter. Fluid resuscitation with isoconductive fluid (a mixture of 0.9% NaCl and dextran blended to a similar resistivity as the dog's blood) was used to maintain LV end-diastolic volume at preendotoxin levels during the initial 30 min after endotoxin infusion. Bolus infusions of the isoconductive solution were given as needed to maintain an LV end-diastolic pressure of > 10 mm Hg.

Several procedures were completed before the IVC occlusion maneuvers. Mixed venous and arterial blood were collected to measure O₂ saturation, O₂ content, and hemoglobin concentration with a Co-oximeter calibrated for dog blood (Model 282; Instrumentation Laboratory, Lexington, MA). Arterial blood gases were analyzed with a blood gas analyzer (ABL-30; Radiometer, Copenhagen, Denmark). CO was then measured through the thermodilution technique, by averaging measurements from five bolus injections of 5 ml of 4° C saline given at random intervals in the respiratory cycle, using a CO computer (model 9520 A; American Edwards). Hematocrit was measured in duplicate, and whole-blood lactate was measured by the enzymatic method (2300 Stat Plus; Yellow Springs Instrument Co., Yellow Springs, OH).

Statistics

All parameters were examined graphically, using standardized normal probability plots to ensure that the data were normally distributed, with homoscedastic variance. All results are summarized as mean ± SD. Directly observed measures and derived hemodynamic variables were compared over time through repeated measures analysis of variance (ANOVA), and differences between control and endotoxic group were determined with between-group ANOVA, using a computer-based statistical package (Statistica V 6.0; Statsoft Inc., Tulsa, OK). Differences in hemodynamic variables in the 240-min data sets and the subsequent two data sets, with L-NMMA and L-arginine infusions, were also examined with a repeated measures ANOVA, to assess the role of NO metabolism over that time period. Significance is given in terms of an F value corresponding to a value of p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of Endotoxemia on Basic Hemodynamic Values

All animals completed the protocol without severe hemodynamic deterioration or the development of persistent ventricular arrhythmias. Seven animals received no endotoxin and served as a control group, and eight animals received endotoxin and served as the endotoxin group. The directly measured hemodynamic variables for these two groups over the observation period are summarized in Table 1. Except for an increase in mean arterial pressure (MAP) at 210 min, no changes were seen in the observed hemodynamic variables over the 240-min observation interval in the control group. The animals that received endotoxin developed a significant hypotensive hyperdynamic state associated with increased arterial lactate levels. MAP decreased from a baseline value of 142 ± 38 to 92 ± 37 mm Hg at 60 min, and remained below control values until the 240-min data collection point, when it was not significantly different from control values. CO reciprocally increased, from 4.21 ± 1.06 to 5.49 to 1.01 L/min at 60 min, and remained elevated until and including the 240-min data collection point. Arterial lactate levels at baseline were similar in the control and endotoxin groups (1.31 ± 0.12 and 1.32 ± 0.44 mmol/L, respectively). However, lactate levels increased only in the endotoxin group, becoming significant at 120 min and remaining elevated for the remainder of the observation period.

Effect of Endotoxemia on Derived Hemodynamic Values

Although the endotoxin group developed a systemic hypotensive, hyperdynamic, hyperlacticemic state, measures of LV systolic and diastolic function did not demonstrate evidence of deterioration. HR in this group also increased for the initial 150 min after endotoxin infusion, but returned to preinfusion levels by 180 min and remained constant for the remainder of the study interval. Neither Ees (Figure 1) nor PRSW (Figure 2) decreased over time in the endotoxin group, either as compared with the control group values or as compared with baseline Ees and PRSW values before endotoxin infusion. On the contrary, both Ees and PRSW tended to increase during the initial 60 min after induction of endotoxemia, although these changes were not significant. Because the variability in Ees data among animals was great over the course of the protocol, individual animal data for Ees are shown in Figure 3. Furthermore, at 120 and 240 min after the induction of acute endotoxemia, dobutamine infusion gave results no different between groups or from the data before endotoxin infusion for the endotoxin group when we defined cardiac reserve as the change in either Ees (Figure 4) or PRSW (Figure 5) from before-dobutamine infusion to values after dobutamine infusion. The surgical preparation also did not appear to alter baseline cardiac function, because neither bolus infusion of esmolol nor dobutamine infusion induced significant changes in either Ees or PRSW.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Mean LV Ees over time and after 240 min of bolus infusions of L-NMMA and L-arginine for controls (solid diamonds) (n = 7) and endotoxic dogs (solid squares) (n = 8).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Mean LV PRSW over time and after 240 min of bolus infusions of L-NMMA and L-arginine for controls (solid diamonds) (n = 7) and endotoxic dogs (solid squares) (n = 8).

View larger version (18K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   Figure 3.   Individual LV Ees values for control animals (A) (n = 7) and endotoxic animals (B) (n = 8) described in Figure 1.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Change in LV Ees in response to pharmacologic challenge. The change is given as mean ± SD. Dobut = dobutamine. Although all directional changes in Ees were significant (p < 0.05) for each group with each challenge, there were no significant differences in the change in Ees between control and endotoxin groups.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Change in LV PRSW in response to pharmacologic challenge. The change is given a mean ± SD. Dobut = dobutamine. Although all directional changes in PRSW were significant (p < 0.05) for each group with each challenge, there were no significant differences in the change in PRSW between control and endotoxin groups.

Additional derived LV hemodynamic data, describing arterial tone and LV diastolic function, are shown in Table 2. Except for an increase in Ea at 210 and 240 min, no changes were seen in the derived hemodynamic variables over the 240-min observation period in the control group. However, for the endotoxin group, measures of LV diastolic function showed small but significant changes. The peak LV filling rate increased from 163 ± 35 ml/s at baseline to 243 ± 105 ml/s at 60 min, and remained elevated until 210 min. This increased filling rate occurred despite a small and nonsignificant decease in LV end-diastolic pressure. However, the peak value of LV dP/dt, another marker of LV performance, decreased by 30 min (2,127 ± 429 to 1,732 ± 390 mm Hg/s) and remained depressed relative to baseline for the remainder of the observation period. In keeping with the observed decease in Pa and increase in CO, Ea also decreased by 60 min and remained depressed for the entire 240-min observation interval.

Effect of Modulation of NO Metabolism

When L-NMMA, a competitive inhibitor of NOS, or L-arginine were infused into the animals after the 240-min observation period, many hemodynamic changes occurred. First, in the control group, CO decreased during L-NMMA infusion, returning to preinfusion values after L-arginine infusion. Although MAP did not change during L-NMMA infusion, calculated Ea increased from 14.7 ± 7.8 to 18.6 ± 7.8 mm Hg/ml, whereas mechanical efficiency reciprocally decreased, from 32 ± 9% to 22 ± 7%, and LV stroke work decreased from 221 ± 41 to 160 ± 46 mm Hg/ml. Similar changes were seen in the endotoxin group, except that they occurred from different 240-min starting values. CO decreased during L-NMMA infusion, returning to values similar to those with L-arginine infusion (4.37 ± 1.6 to 3.46 ± 1.16 to 4.37 ± 1.79 L/min, respectively). Similarly, Ea increased during L-NMMA infusion whereas mechanical efficiency and stroke work declined, all returning to their pre-L-NMMA values during L-arginine infusion (Table 2).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our data show that in the canine model of acute endotoxemia, LV contractile function is preserved for the first 4 h despite the induction of a hyperdynamic, hypotensive, lactic acidotic state. Furthermore, the observed consistent increase over baseline in contractility in response to dobutamine after induction of endotoxemia did not appear to be due to a reflexively increased endogenous sympathetic tone associated with endotoxic shock, as has been suggested by others (29). This lack of cardiac dysfunction contrasted with the peripheral vascular response, as manifested by changes in Ea and CO. Indeed, although the changes in Ees and PRSW were not significantly different in the endotoxic and control groups, the increases in both measures of contractility in response to dobutamine were slightly greater in the endotoxic group than in the control group, suggesting that this observation did not reflect a  error. Our endotoxic dogs also displayed a complex pattern of diastolic dysfunction, characterized by an increased dynamic diastolic compliance (increase in maximal filling rate despite no change or a slight decrease in filling pressure) but a reduced isometric relaxation phase (maximal LV dP/dt). Although some of the reduction in the maximum LV dP/dt could reflect lower preload through the Frank-Starling mechanism, we cannot rule out the possibility that some degree of diastolic dysfunction occurs early in canine endotoxemia. However, such changes were minor, and did not affect either steady-state hemodynamics or the response of the left ventricle to rapid changes in loading conditions (IVC occlusion maneuver). Thus, the initial hemodynamic state of acute endotoxemia in the dogs is not associated with a measurable impairment of cardiac pump function. These data suggest that if the presenting expression of sepsis in a previously healthy individual is acute hemodynamic instability, then cardiac contractile function may well be preserved.

Acute endotoxemia produces an immediate hypotensive, hyperdynamic response in the fluid-resuscitated mammal, characterized by progressive metabolic acidosis and activation via numerous inflammatory networks of circling and fixed tissue immune effector cells. The role of impaired cardiac function in this response has been the subject of continuing investigation and controversy. Although subjects with prolonged sepsis have impaired LV contractile function, their CO is usually increased and their LV filling pressures are often low. Presumably, the low outflow resistance in prolonged sepsis makes the associated impaired LV contractility less apparent, because treatment with NO inhibitors not only increases LV afterload but also markedly reduces CO in septic dogs with systemic hypotension (30). Acute endotoxemia also induces the sequential production and release of TNF- into the circulation (7), in addition to upregulating NO synthesis (31). Though the systemic release of TNF-, a potent proinflammatory cytokine, can alter LV contractile function (6), the mechanism by which this occurs is not well described. Both loss of autoregulation of coronary blood flow (9) and nonspecific vasodilation caused by widespread NO release from vascular smooth-muscle cells (32) have been described. Our data suggest that NO metabolism does not measurably influence cardiac contractile function during the immediate hyperdynamic, hypotensive, hyperlactecemic interval following endotoxin infusion. Furthermore, although other mediators, such as prostaglandin (PG) F₂, PGE₁, leukotriene (LT) C₄, LTD₄ and platelet activating factor (PAF) (33, 34) are rapidly released during endotoxemia, and are the primary substances initially altering peripheral vasomotor tone (35), they do not influence cardiac contractility. Similarly, upregulation of inducible NOS (iNOS) is not an explanation for the acute hemodynamic effects seen in the early phases of septic shock or endotoxic shock in the dog, because inducible isoforms of NOS are not expressed in tissues exposed to proinflammatory stimuli for up to 4 to 6 h (36), and blockade of PAF-induced activation in endotoxin-treated rats inhibits the acute-phase response but does not inhibit upregulation of NOS (37). Although endotoxin infusion stimulates iNOS activity in the heart that decreases LV systolic function (38), the expression of this iNOS activity occurs over a period of 6 to 8 h. Still, some degree of NO release may occur early in sepsis, via constitutively expressed NOS, and may induce combined peripheral vascular and LV dysfunction (32). Granton and coworkers (39) showed that the formed cellular elements of the blood binding in the microcirculation were the primary factors inducing the observed myocardial dysfunction in acute endotoxemia. The time course of this effect appears to be variable.

Study Limitations

We found an initial significant increase in HR in the endotoxemic group. Increases in HR in the range observed have been shown in dogs to increase contractility by a chronotropic mechanism (40). This potentially confounding effect could have masked an initial myocardial depression. However, if this were the case, then once HR returned to preendotoxin levels, by 180 min, any chronotropic effects would also have been lost, revealing underlying impaired myocardial contractility. We saw no differences in either basal Ees or basal PRSW in the endotoxemic group over time, either when HR was increased at an early point or when it returned to baseline values at 180 min.

Some of the apparent discrepancy in the cardiovascular effects of acute endotoxemia reported in the literature and highlighted in our data may reflect species differences in response, as well as the model of sepsis used or model-specific limitations. Endotoxin infusion in guinea pig models of septic shock reduces LV contractility by as early as 2 h after endotoxin exposure (26). Similarly, infusion of TNF- in an ovine model of septic shock induced myocardial depression from 3 to 4 h after infusion of TNF- (41), whereas in murine models of endotoxic shock, constitutive NOS activity appears capable of inducing systemic hypotension immediately after endotoxin challenge (42). Moreover, since septic shock induces a profound sympathetic response associated with increased circulating plasma levels of norepinephrine, some degree of myocardial dysfunction could be masked by concurrent increases in systemic catecholamine levels. However, if anything, cardiovascular responsiveness to catecholamines appears to be blunted during endotoxic shock (29), making this possibility unlikely. In the artificial environment of minidose endotoxin challenge in humans, Munchie and associates (7) demonstrated that the sequential release into the bloodstream of TNF- and interleukin-6 followed the pattern and time course seen in both canine and porcine endotoxic shock. Thus, differences in species in the effects of endotoxic shock may to a greater extent reflect sensitivity differences among species or the role of other processes, such as microcirculatory derangements, differential interactions of formed cellular elements with the activated vascular endothelium, and normal physiologic reserves within organ systems.