INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hypotension is a common manifestation of sepsis. The low blood pressure is associated with tissue hypoperfusion and organ dysfunction, and the management of septic patients therefore usually requires infusion of a vasopressor agent. Norepinephrine (NE) is a commonly used vasopressor in the clinical setting. It is a potent vasoconstrictor with strong -adrenergic agonist activity, but also has important -adrenergic agonist activity (1, 2).

To be truly effective as a pressor, an agent must not only increase arterial pressure, but must also improve cardiac output (CO) so that overall tissue perfusion is improved. An increase in CO can occur from an increase in cardiac function as well as from an increase in circuit function (i.e., venous-return relationship) (3, 4). Circuit function depends upon stressed vascular volume, venous resistance, and venous capacitance. To properly use NE clinically, it is therefore important to know how NE affects circuit parameters under nonseptic and septic conditions, and whether these effects of NE contribute to an improvement in CO. Previous studies of the circuit effects of catecholamines have largely concentrated on epinephrine (5, 6), and there is less information about the circuit effects of NE (7), particularly in sepsis (1, 11, 12).

-Receptor agonists decrease venous capacitance, which will increase venous return (1, 13). These drugs are therefore of potential benefit in septic shock. However, -receptor agonists also increase the resistance to venous return, which decreases venous return and CO. On the other hand, -agonists decrease venous resistance (14), which benefits CO (1). Since NE has both - and -agonist activities, it could decrease vascular capacitance, which would benefit CO and decrease venous resistance, which would also tend to improve CO. Our first objective in this study was therefore to characterize the cardiac and circuit effects of NE at usual therapeutic concentrations in pigs, under baseline conditions and during endotoxemia. We hypothesized that NE would decrase vascular capacitance but not affect resistance to venous return. We also predicted that NE would increase cardiac function so that the net hemodynamic effect would be an improvement in CO.

The predominant hypothesis for the hypotension that occurs in sepsis is that there is an increase in nitric oxide (NO) production from induction of the inducible form of NO synthase (iNOS) (15). Increased NO production is believed to result in dilatation of vascular smooth muscle by activation of soluble gaunylyl cyclase and increased cyclic guanosine-3':5'-monophosphate (cGMP). Of importance is that the hypotension in sepsis is associated with a decrease in responsiveness to catecholamines in rats (19), and that inhibition of NOS in rats restores responsiveness to catecholamines (20).

We have been studying a porcine model of endotoxemia that has many of the features of human sepsis, including a decrease in systemic vascular resistance (21, 22). However, the pigs in these studies have only a small increase in NO production and no physiologically significant induction of iNOS (22). Our second objective in the present study was therefore to determine whether, in these animals with no significant iNOS induction, there is still an NO-dependent loss of vascular responsiveness to NE, as observed in rats. If so, this would argue for a role of constitutive NO in the loss of vascular responsiveness in sepsis.

A final objective in the present study was to determine regional differences in the response to NE during sepsis with and without NO synthase inhibition, including effects on systemic resistance, CO, heart rate (HR), resistance to venous return, mean circulatory filling pressure (MCFP), and pulmonary arterial pressure (Ppa). The purpose of this part of the study was to determine whether inhibition of NOS affects only the responsiveness of arteries or also affects responses in veins and the pulmonary vasculature (i.e., whether there are regional differences in the effects of NOS inhibition).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General Methods

All procedures were performed according to the guidelines of the animal care committee of McGill University. Domestic pigs (n = 8) weighing 29.9 ± 4 kg (range: 25 to 38 kg) were sedated with ketamine at 30 mg/kg, atropine at 1.0 mg/kg, and xylazine at 2 mg/kg. Twenty minutes later, the animals were anesthetized with sodium thiopental at 10 to 15 mg/kg; anesthesia was maintained with a continuous intravenous infusion of sodium thiopental at 5 mg/kg/h. The animals were placed in the supine position in a V-shaped support, intubated with a cuffed endotracheal tube, and ventilated with a volume respirator at a tidal volume of 12 ml/kg and a frequency of 12 to 15 breaths/min, at 5 cm H₂O positive end-expiratory pressure. Through a midline incision in the neck, the left common carotid artery was isolated and cannulated with a polyethylene catheter for pressure measurements. The right internal jugular vein was isolated and a pulmonary artery flotation catheter was passed into the pulmonary artery. A size 12-French balloon-tipped catheter with a 50-ml capacity (Prewitt aortic occlusion catheter No. 10; Pramel Inc., Lonueuil, PQ, Canada) was placed in the right atrium through the external jugular vein. When inflated to 40 ml, this balloon transiently obstructed the circulation and was used to stop venous return and measure MCFP. The right femoral vein was cannulated with a polyethylene catheter (Cole Palmer, Anjou, PQ, Canada) for the administration of drugs. Blood gases were monitored, and we tried to keep PCO₂ between 30 and 40 mm Hg by adjusting the ventilator and PO₂ above 90 mm Hg by giving supplemental oxygen. CO was measured by the thermodilution method (model 3300; Abbott, Inc., North Chicago, IL) by injecting 3 ml of 5% dextrose in water at room temperature into the right atrial port of the pulmonary catheter.

Measurement of MCFP

To measure MCFP, we rapidly inflated the balloon in the right atrium with 40 ml of air for 15 to 20 s (21). This transiently arrests venous return, and the venous pressure measured in the central vein is equal to the pressure upstream in the compliant region of the venous system. This procedure could be repeated frequently without an effect on the hemodynamic parameters or general condition of the animal. It was reproducible with < 0.5 mm Hg difference in MCFP on repeated measurements under the same conditions. The MCFP measurement was obtained after 15 s even though the arterial plateau pressure (APP) remains above the venous plateau pressure (VPP) at this point, because doing so avoids reflex changes. The APP remains high because volume continues to drain through the high arterial resistance created by the inflated balloon and also because there is an arterial critical closing pressure, which traps volume in the arterial vessels (24). This volume is given by the formula MCFP = VPP + (APP  VPP) (arterial compliance/venous compliance), where the ratio of arterial to venous compliance is assumed to be 1:30.

Hemodynamic Calculations

Systemic venous resistance (SVR) was calculated as SVR = (  Pra)/CO, where is mean arterial blood pressure, Pra is right atrial pressure, and CO is measured in L/min. Pulmonary vascular resistance (PVR) was calculated as PVR = (  Pcwp)/CO, where is mean pulmonary artery pressure and Pcwp is the pulmonary capillary wedge pressure. Resistance to venous return (RVR) was calculated as RVR = (MCFP  Pra)/CO. SVR, PVR, and RVR are in units of mm Hg · L¹ min.

Protocol

The animals were stabilized for 30 min and baseline hemodynamic values were obtained. We then infused NE at 3 µg/min, 9 µg/min, and 27 µg/min for 10 min each, in consecutive order. These doses correspond to commonly used clinical doses of NE. The doses were set for a 30-kg pig and adjusted according to size of the animal. When the third dose of NE was finished (i.e., after 30 min) and the animal had again stabilized, we infused 10 µg/kg/h of Escherichia coli endotoxin until the end of the experiment. After 105 min, measurements of the responses to the three doses of NE were repeated in the same way as under the control condition. Following the NE infusions (a total of 30 min), the animal's hemodynamics were again allowed to stabilize. We then gave 12.5 mg/kg N^G-nitro-L-arginine methyl esther (L-NAME) over a period of 10 min and repeated the infusions of the three doses of NE. This dose of L-NAME is similar to that used by others (21, 22). As in our previous studies, we infused normal saline or dextran to maintain the Pra at 3 to 4 mm Hg (21, 22).

Statistics

All values represent mean ± SD. A two-way analysis of variance (ANOVA) for repeated measures was used to evaluate significant differences among conditions, and where significance was found, post hoc analysis was performed with the Student-Newman-Keuls test. The values over time were evaluated with one-way ANOVA. Statistical significance was considered to exist at p < 0.05. Analyses were done with Sigma stat software (Jandel Scientific).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Initial blood gas measurements showed at PO₂ of 229 ± 151 mm Hg, PCO₂ of 33 ± 4 mm Hg, pH of 7.53, bicarbonate of 27 ± 3 mEq/L, and hemoglobin of 10.8 ± 1.8 g/dl. After 105 min of endotoxemia, the PO₂ was 214 ± 100 mm Hg, the PCO₂ was 37 ± 2 mm Hg, the pH was 7.42, the bicarbonate was 21 ± 2 mEq/L, and the hemoglobin was 9.7 ± 1.0 g/dl. There was a deterioration in the arterial blood gas tensions following infusion of L-NAME. PO₂ declined to 83 ± 46 (p < 0.05) mm Hg, PCO₂ remained at 37 ± 2 mm Hg, pH decreased to 7.27 and bicarbonate also decreased, to 18 ± 1.2 mEq/L (p < 0.05), and hemoglobin rose to 12.0 ± 1.9 g/dl (p = NS).

CO, , and over the course of the experiment are shown in Figure 1. There was little change in baseline CO, , and between the measurement before and the measurement after the first set of NE infusions (30 to 0 min). After the infusion of endotoxin, there was an immediate increase in , no significant change in , and a small decrease in CO. began to fall after approximately 60 min, and CO returned to the baseline level. also fell by 60 min, but remained above the baseline for the duration of the experiment. CO, , and before and after the dose-response assessment of NE with endotoxin were also unchanged.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Time course of CO (A), Part (B), Ppa (C ) in response to endotoxin or control infusion. After the baseline measurements (30 min), we gave the first set of NE infusions. The first arrow indicates the start of endotoxin infusion. The second arrow indicates the beginning of the second set of NE dose-response measurements (see text for details). L-NAME was infused after 140 min (data not shown in figure). Stars represent significant difference from baseline (p < 0.05), except for CO, for which the difference is from the lowest CO.

Examples of the change in Part and Ppa under baseline conditions, after endotoxin, and after endotoxin and L-NAME, are shown in Figure 2 for the 9- and 27-µg/min infusions of NE. In the baseline condition, NE produced a marked increase in Part that lasted for several minutes. The response to NE at 27 µg/min was greater than the response to 9 µg/min. There was little change in Ppa. After the endotoxin infusion, the response of Part to NE was markedly blunted and there was a small increase in Ppa. When L-NAME was infused after endotoxin, the peak increase in Part was similar to that seen under the control condition, but it was of very short duration. The actual pressure pattern was biphasic and very different from the control response. There was an initial, very short increase followed by a decrease in pressure, followed by a second rise and then a decline. The pattern seen in Figure 2 was similar in all animals.

View larger version (21K): [in this window] [in a new window]   Figure 2.   Examples of changes in Part and Ppa with infusion of 9 µg/min (left) and 27 µg/min (right) of NE at baseline, after endotoxin, and after endotoxin and L-NAME. The beginnings of infusions are marked by arrows. The top set of tracings is the baseline condition, the middle set is after 105 min of endotoxemia, and the bottom set represents L-NAME with endotoxemia. The change in Part was greater with the 27-µg/min than with the 9-µg/min dose of NE. In fact, this tracing is saturated. There was no change in Ppa. With endotoxemia, NE had little effect on Part, although there was a small response in Ppa. Following L-NAME there was a very transient but large increase in Part (note the spike on the tracings) and a small increase in Ppa. There was then a marked decrease in pressure followed by another increase, with a smaller pulse pressure and a shorter duration than under the control condition before endotoxin. Note that the baseline pressure is higher with L-NAME than with endotoxin alone. The paper speed was either 1 mm/s or 100 mm/s.

Dose-response curves based on the maximum change in Part in response to NE are shown in Figure 3 for the 3-, 9-, and 27-µg/min doses of NE. A dose-response relationship was clearly evident under control conditions, but the response during endotoxemia was markedly reduced (p < 0.05). The infusion of L-NAME restored the response to almost the control level.

View larger version (11K): [in this window] [in a new window]   Figure 3.   Dose-response relationship to NE under baseline (control: closed circles) following 105 min of endotoxin (closed squares), and with endotoxin and L-NAME (open triangles) for the 3-, 9-, and 27-µg/min doses of NE. There was a dose-response effect for the change in Part with NE as determined by ANOVA (p < 0.05). The endotoxin group effect was markedly reduced as compared with the control condition (*p < 0.05 for endotoxin compared with control), but the response was restored with L-NAME, so that there was no difference between the endotoxin + L-NAME condition and the control condition (mean ± SD).

Table 1 shows the change in systolic arterial pressure and after 10 min of each infusion of NE. The dose-response relationship after 10 min was less clear than the maximum response shown in Figure 3. After 10 min of infusion of NE at 27 µg/kg, there was a wide variation in Part, and the after 10 min of 27 µg/min NE was not different from the baseline in the preendotoxin condition. in the endotoxin and L-NAME conditions were significantly different from that in the control condition. There was a significant increase in systolic Part at 9 µg/min and 27 µg/min NE in the endotoxin group, and at all levels of NE in the L-NAME group. , however, was only significantly increased at 3 µg/min and 9 µg/min NE in the control group.

In the control condition, SVR fell significantly with the 27 µg/min infusion of NE (Figure 4). There was no significant change in SVR in response to NE with endotoxin. Following L-NAME, SVR increased to a similar extent with each dose of NE, and all these values were significantly different from those seen both under control conditions and with endotoxin. SVR in the endotoxin and endotoxin + L-NAME conditions was significantly different from that in the control condition.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Bar graph of the change in SVR in the baseline condition (control), with endotoxin, and with endotoxin + L-NAME for the three doses of NE (open boxes = 3 µg/min, hatched boxes = 9 µg/ min, filled boxes = 27 µg/min). There was a significant decrease in SVR in the control condition with 27 µg/min of NE (p < 0.05). There was no change in SVR with NE during endotoxin infusion. NE produced a rise in SVR with L-NAME + endotoxin. The endotoxin and endotoxin + L-NAME conditions were significantly different (p < 0.01) from the control condition by ANOVA (indicated by the plus sign [+]). *Significantly different from baseline for the condition (i.e., 0 µg NE) (mean ± SD).

The changes in CO with NE are shown in Figure 5. There was a dose effect under the baseline condition, and this was markedly flattened with endotoxin, but the effect with endotoxin was not significantly different from the control condition by ANOVA. The addition of L-NAME to endotoxin did not restore the response of CO to NE, and there was a tendency toward a decrease in CO with NE at 27 µg/kg. The L-NAME condition was significantly different from the control condition. Similarly, endotoxin markedly reduced the response of HR to NE (Figure 6), and the HR response was not altered by the addition of L-NAME.

View larger version (10K): [in this window] [in a new window]   Figure 5.   Dose-response curves of change in CO at baseline (control: closed circles; endotoxin: closed squares; and endotoxin + L-NAME: open triangles) with 3, 9, and 27 µg/min of NE. The plus sign indicates a significant increase (p < 0.05) in CO with 27 µg/min NE in the control condition. There was no change in CO with endotoxin or endotoxin + L-NAME at any dose. *The L-NAME condition was significantly different from control (p < 0.05) at baseline, whereas the endotoxin condition was not. Values are mean ± SD.

View larger version (11K): [in this window] [in a new window]   Figure 6.   Dose-response relationship of change in HR with 3, 9, and 27 µg/min of NE. The dose effect was significant by ANOVA (p < 0.05). There was a marked flattening of the dose-response relationship with endotoxemia, and this was not altered with endotoxin + L-NAME. *The latter two conditions were significantly different from the control condition (p > 0.05) (control: closed circles; endotoxin: closed squares; and endotoxin + L-NAME: open triangles). Values are mean ± SD.

Changes in Ppa with NE are shown in Figure 7. There was a small increase in Ppa with NE at 3 µg/min in all three conditions. There were no further significant changes with higher doses of NE in any of the three conditions, except at the 27-µg/min dose during endotoxemia.

View larger version (10K): [in this window] [in a new window]   Figure 7.   Dose-response relationship of change in Ppa with 3, 9, and 27 µg/min NE. There was a significant increase in Ppa from baseline with endotoxin at 27 µg/min NE (asterisk), but not with the other two conditions. The responses in the three conditions were not significantly different by ANOVA (control: closed circles; endotoxin: closed squares; and endotoxin + L-NAME: open triangles). Values are mean ± SD.

Figure 8 shows venous return curves for the baseline condition, after endotoxin, and after endotoxin + L-NAME. NE produced a shift to the right of the venous return curve. This was due to a significant increase in MCFP without a change in the slope of the relationship (RVR). With endotoxin, the change in MCFP was smaller but still significant (p < 0.05), and there was again a shift to the right of the venous return curve. NE had no effect on the venous return curve when L-NAME was added to the endotoxin condition.

View larger version (15K): [in this window] [in a new window]   Figure 8.   Venous return cures under control, endotoxin, and endotoxin (endo) + L-NAME conditions. The y axis represents flow and the x axis right atrial pressure. The x intercept represents MCFP. Conditions at baseline and with a 27-µg/min dose of NE are shown. Under the control condition, NE produced a shift to the right of the venous return curve, an increase in CO, and no change in slope of the curve. A similar response was seen with endotoxemia, but of a smaller magnitude. There was no change in the position of the curve with endotoxin + L-NAME. *p < 0.05 compared with baseline for the condition (mean ± SD).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We found that in anesthetized pigs, NE increased CO by increasing MCFP and shifting the venous return curve to the right without changing RVR under both control and endotoxic conditions. However, the effect of NE in endotoxemia was less than that seen under control conditions. The peak pressor response to NE also was markedly reduced during endotoxemia. Infusion of the NOS inhibitor L-NAME restored the initial arterial pressure response to NE, but did not restore the increase in MCFP, HR, CO, or Ppa. Furthermore, the peak increase in arterial pressure with NE after both infusion of L-NAME and of endotoxin was of a shorter duration and had a different pattern than what was seen under the control condition.

Before these results are discussed a number of technical factors need to be considered. NE has mixed - and -agonist activity, which makes more difficult the physiologic interpretation of the responses to it, since responses could be due to either its -receptor activity, -receptor activity, or both. This was evident under the baseline condition, in which the highest dose of NE produced a smaller change in Part than did the other doses. However, NE is the most commonly used catecholamine clinically, and a better characterization of the response to NE therefore has important practical implications. The doses we used are commonly used clinically, although higher doses are sometimes used.

The experiments in our study were performed by necessity in a cumulative manner, which could have affected the response to NE given after L-NAME. Furthermore, L-NAME produced an immediate and obvious hemodynamic deterioration, including a decrease in CO and pH in all animals. Despite this deterioration in overall status, loss of the arterial pressure response to NE that was seen with endotoxin was reversed with L-NAME, which indicates that the response with L-NAME was a valid one and not just due to changes with time. CO and arterial pressure were also similar before and after the infusion of NE, both under the baseline condition and during endotoxemia.

A very important variable in these studies was the volume needed for resuscitation, for in a previous study we found that animals that were not volume resuscitated had a marked decrease in cardiac function (21). We were therefore careful to try to keep the central venous pressure constant by giving fluid as necessary.

A potential complicating variable in analyzing a vasopressor response is the baseline arterial pressure. If the starting pressure is too high, further vasoconstriction may not be expected. Fortuitously, L-NAME restored the blood pressure to the same baseline level as before endotoxin, and changes in arterial pressure therefore, began from the same baseline level.

Our first objective was to analyze the effects of NE on cardiac and circuit interactions. As hypothesized, NE increased MCFP, which implies a decrease in vascular capacitance. This most likely occurs through a change in the splanchnic venous system (25). NE also increased CO at a given right atrial pressure. Since the arterial pressure also increased, this increased output with a given preload and increased afterload implies an increase in cardiac function. Under the baseline condition, NE did not decrease the resistance to venous return, as we had predicted, although there was a tendency for resistance to venous return to decrease after infusion of endotoxin. This latter observation might reflect a shift to more -receptors than -receptors in endotoxemia (28). Imai and colleagues (1) compared the effect on venous return of isoproterenol (a pure -agonist), NE, and methoxamine (a pure -agonist) in dogs. They found that NE increased venous return and proposed that it decreased the resistance to venous return. However, they did not measure MCFP and thus could not separate capacitance from resistance. As in our study, NE had a minimal effect on the pulmonary vasculature. Imai and colleagues also found that methoxamine decreased venous return, and suggested that this was because of an increase in venous resistance.

The response to NE in our study was similar before and after endotoxemia, although the response was smaller after endotoxemia. The rise in CO and lack of effect on venous resistance with NE makes it a very favorable agent for treating septic patients. This is supported by the work of Breslow and colleagues (29), who compared NE, phenylephrine, and dopamine in porcine endotoxemia and found that only NE restored CO. Two clinical studies have also shown an advantage of NE over dopamine (7, 30).

The rise in MCFP that comes from a decrease in vascular capacitance with NE is equivalent to the physiologic effect of giving fluids (3, 31). The advantage, however, of producing an increase in MCFP with NE rather than by giving fluids is that fluids do not have to be administered. Thus, when the patient recovers, excess fluid does not have to be removed. Furthermore, although it might be argued that arterial pressure and CO could be raised by giving more fluid, this approach increases pressures throughout the vascular tree, including the capillaries. This will then contribute to the leak that is observed in septic animals and patients. However, it must also be appreciated when treating patients that a decrease in vascular capacitance with NE depends upon the presence of an adequate volume in the vasculature. When total vascular volume is low, and reflex mechanisms have already decreased vascular capacitance, NE will not further decrease vascular capacitance. This is the case in cardiogenic shock, especially in patients who have been previously diuresed.

Our second objective was to determine whether responsiveness to NE is decreased in porcine endotoxemia, and whether an NOS inhibitor restores the response to NE. As in previous studies (29), we found that the pressor response to NE was markedly reduced after 90 min of endotoxemia. Furthermore, as in previous rat studies, an NOS inhibitor restored responsiveness to NE. However, the pattern of the pressor response to NE after NOS inhibition was very different from what was observed before infusion of endotoxin. There was a rapid initial transient increase in blood pressure, followed by a decrease and then a further short rise in blood pressure. The peak response is thus somewhat misleading. L-NAME also had no beneficial effect on the cardiac response. If anything, CO tended to worsen with NE and L-NAME. There was also no change in MCFP with NE.

The predominant systemic arterial effect of NE after L-NAME, with little effect on the venous circulation, heart, and pulmonary circuit, suggests that the benefit of L-NAME might be limited to areas with a greater -receptor density, since it has been shown that the density of ₁-adrenergic receptors is greater in arteries than in veins (2). It is possible that this more predominant effect on arteries could have been due simply to a greater signal-to-noise ratio in the arterial response than in other vascular beds, where changes were proportionally smaller. This, however, seems unlikely, because CO actually worsened and showed no tendency for improvement with NE after L-NAME and after endotoxin. The response of HR to NE after L-NAME was no different from the response during endotoxin infusion alone.

The responses to L-NAME were somewhat complicated by the profound effect of L-NAME on the vasculature. There was a marked decrease in CO, and acidosis worsened. The decrease in CO with L-NAME was similar to what we previously observed with the same model (21), and even occurs in nonendotoxic pigs (32). As in our previous studies, L-NAME produced a marked flattening of the cardiac function curve (data not shown) and increased the resistance to venous return. Together, these effects greatly limit CO. A decrease in CO upon inhibition of NOS has been a common observation in previous animal and human studies. The major benefit of NOS inhibition has been a reduction in the amount of vasopressor needed to maintain arterial pressure (33). This, however, will be of limited benefit if inhibition of NOS decreases CO and reduces total tissue perfusion.

The transient rise in vascular responsiveness to NE after L-NAME in septic pigs indicates that adrenergic receptors, second messengers, and contractile machinery are all intact in these animals, since they can respond. The question then arises of why the response is not sustained. Suba and associates (34) studied -receptors in aortic rings of septic rats. They found a rightward shift of the dose-response curve for NE in aortic rings ex vivo after 18 h of endotoxemia, and a decrease in phosphoinositide hydrolysis. However, the receptors themselves appeared to be functional. It is possible that alterations in membrane function from endotoxic injury affect ion channels and calcium handling. This could explain why the increase in blood pressure with NE following L-NAME is not maintained as long in sepsis as without sepsis. An effect on calcium handling is supported by the work of Bigaud and coworkers (19), who found that increased extracellular calcium restored contractile responses in aortic rings from endotoxic rats. Of interest was that Suba and associates (34) found no decrease in the contractile response of aortic rings at 2 h after NE followed by endotoxin in rats, although there is clear evidence from other studies and our study of decreased contractile responsiveness to vasopressors when they are followed by endotoxin in vivo (29). Unfortunately, Suba and associates did not determine whether the contractile response was depressed in vivo as well. However, Paya and coworkers did find that ex vivo, aortic rings of endotoxic rats responded normally to NE, although in vivo responses were abnormal. These data raise the possibility that circulating factors are important for maintaining the decreased responsiveness in the early phase of endotoxemia, as suggested by others.

In rat studies, the decreased responsiveness to NE occurs after only 60 min of endotoxemia, which is well before the induction of iNOS. We also found no evidence of iNOS induction in pigs through Western blot analysis and NOS activity assay for up to 4 h of endotoxemia (22, 35). The partial restoration by L-NAME of the response to NE during this early phase of endotoxemia does therefore not occur through the inhibition of iNOS. However, L-NAME could be acting through its effect on endothelial NOS (ecNOS). Szabo and colleagues found an increase in total NOS activity in rat aorta after 60 min of endotoxin infusion, and we also found an increase in NOS activity in the aorta of septic pigs after 2 h of endotoxemia (35).

In conclusion, NE is an effective agent for supporting cardiac function in sepsis. It can restore arterial pressure, decrease vascular capacitance, and shift the venous return curve to the right, thereby improving CO. NE has little effect on Ppa, which should make it safe even in cases with increased Ppa. Endotoxin decreased the vascular responsiveness to NE, but the arterial pressor response could be restored with L-NAME. However, the pressor response to NE during endotoxemia with infusion of L-NAME was biphasic and more transient. L-NAME also caused a marked deterioration in cardiac function and an overall worsening of the animals to which it was given, which limits its therapeutic usefulness.