INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hypoxic pulmonary vasoconstriction (HPV) was first described by von Euler and Liljestrand in 1946, and has since been extensively studied (1). It is characterized by vasoconstriction of pulmonary vessels of underventilated and thus hypoxic lung areas, and by diversion of blood flow away from these lung areas to better oxygenated alveoli. In this way, ventilation- perfusion (A/) mismatch is minimized and systemic hypoxemia is prevented. The vasoconstriction most likely takes place in the small pulmonary arteries (2). It begins within seconds of ventilation with hypoxic gas, reaches its maximum after minutes, and is sustained for hours. The final effector and the exact mechanism that mediates HPV, however, remain unknown.

HPV is altered in several disease states including sepsis and adult respiratory distress syndrome (3), and many different substances modulate HPV, among them inhalational anesthetics, nitrates, nitroprusside, calcium-channel blockers, metabolites of the cyclooxygenase pathway as well as cyclooxygenase inhibitors, and -blockers. Inhalation of nitric oxide (NO) decreases increased pulmonary vascular resistance (4, 5) and blunts HPV (6) in a dose-dependent manner. Nitrates and nitroprusside act as NO donors and thus reduce HPV. Inhibitors of NO synthase (NOS) augment HPV under normal conditions (10, 11). Endotoxin abolishes HPV (12).

During endotoxemia and sepsis, NO is produced in large amounts by the inducible form of NOS (iNOS). NO is a major mediator of hemodynamic changes associated with endotoxemia and sepsis. Hyperdynamic sepsis is characterized by generalized vasodilation with hypotension and increased cardiac output. Blockade of NOS reverses the hyperdynamic circulatory state (15).

Our goal for this study was to determine whether NO was in part responsible for the blunted pulmonary vasoconstrictive response to hypoxia during sepsis. We tested our hypothesis in an ovine model of chronic sepsis produced by continuous infusion of Pseudomonas aeruginosa. HPV was repeatedly induced by isolated unilateral hypoxia of the left lung, with the right lung receiving 100% oxygen to prevent systemic hypoxemia. HPV was assessed by change in left pulmonary blood flow during hypoxia. After sepsis had developed, we investigated the effect of a commonly used NOS inhibitor, N-monomethyl-L-arginine (L-NMMA), on HPV.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nineteen adult female sheep (weight: 26 to 42 kg; mean: 34 ± 1 kg) underwent general anesthesia with halothane and were chronically instrumented. Catheters were placed in the femoral artery and vein. A pulmonary artery thermodilution catheter (93A131 7F or 746H8F; Baxter-Edwards Critical Care, Irvine, CA) was introduced via the right jugular vein into the pulmonary artery. A left thoracotomy was then performed at the level of the 5th intercostal interspace, a left atrial catheter was inserted, and an ultrasonic transit-time flow probe (Transonic Systems, Ithaca, NY) was positioned around the left pulmonary artery. The animals were awakened after surgery and allowed 1 wk to recover. They were maintained in metabolic cages with free access to food and water.

The study was conducted according to the Guidelines of the American Physiological Society, and was approved by the Animal Care and Use Committee of The University of Texas Medical Branch. The Investigative Intensive Care Unit is an American Association of Laboratory Animal Care-approved facility.

On the day before the experiment began, the arterial line, the distal and proximal ports of the pulmonary artery catheter, and the left atrial catheter were connected to pressure transducers with continuous flushing devices (heparin-saline solution, 3 U · ml¹). The flow probe was connected to a flowmeter (Transonic Systems). An infusion of Ringer's lactate was started at 2 ml · kg¹ · h¹ via the femoral vein, and was continued throughout the experiment. The flow rate was adjusted to maintain left atrial pressure within 2 mm Hg of its baseline level.

On the following day the animals were anesthetized with ketamine (5 mg · kg¹ intravenously) and a tracheostomy was performed. A modified left-sided double-lumen endotracheal tube (Broncho-Cath size 35 to 39 French [depending on the size of the animal]; Mallinckrodt, Inc., St. Louis, MO) was positioned through the stoma with its bronchial tip in the left main bronchus to accomplish separate ventilation of the right and left lung. In sheep, the right-upper-lobe bronchus (tracheal bronchus) emerges directly from the trachea. This special anatomic situation requires the tracheal balloon of the double-lumen tube to be moved 5 cm more proximally from its original position on the tube. To ensure correct positioning of the tube, its location was verified by bronchoscopy (Olympus URF Type P; Olympia Corporation, New York, NY). When this was done, the tube was held in place with a gauze dressing. The sheep were connected to a ventilator (Servo 900 C; Siemens-Elema, Solna, Sweden), through which they spontaneously breathed humidified air with an FI_O2 of 0.3 and a positive end-expiratory pressure (PEEP) of 5 cm H₂O. During the experimental period, the tubes were removed several times and cleaned to avoid obstruction by secretions, after which they were reinserted and their position corrected bronchoscopically. The two lumens of the endotracheal tube were frequently suctioned, and both lungs were manually inflated to mobilize secretions and prevent atelectasis.

Four hours after the sheep had fully regained consciousness and had adapted to the endotracheal tube, baseline hemodynamic data and blood samples were collected with the sheep standing and breathing humidified air with an FI_O2 of 0.3. The PEEP had been reduced to 0 cm H₂O at least 5 min before the measurements, which were made after all hemodynamic variables were stable. The measurements included heart rate (HR); mean arterial, mean pulmonary, left atrial, and central venous pressures; cardiac output by the thermodilution method; left pulmonary artery flow; and arterial and mixed venous (pulmonary arterial) blood-gas tensions.

Only the tracheal port of the double-lumen tube was connected to the ventilator, and the FI_O2 was increased to 1.0. The FI_O2 was repeatedly checked with an in-line oximeter (Miniox I oxygen analyzer; Catalyst Research Corporation, Owings Mills, MD) at the tracheal port of the double-lumen tube. The left bronchial port of the double-lumen endotracheal tube was connected to a one-way-valve T-piece that prevents rebreathing of CO₂, and was continuously flushed at high flow from a tank with 100% nitrogen. FI_O2 was checked in the same manner as described previously, in order to ensure an FI_O2 concentration of less than 0.1. After 15 min, hemodynamic measurements were made and blood samples were collected. Both tubes were then again connected to the ventilator and the FI_O2 was returned to 0.3.

Each hypoxic challenge was preceded by baseline hemodynamic measurements and arterial and mixed venous blood-gas determinations. Hypoxic challenges were performed at 0 h (baseline) and at 24 and 28 h, and then every 2 h until Hour 38, and again at Hour 48. Hemodynamic and blood-gas data without a hypoxic challenge were additionally collected at Hours 1 and 2, every 4 h, and at Hours 40 and 44.

Arterial and pulmonary arterial blood samples were drawn every 4 h for quantitative blood cultures to ensure stable bacterial blood levels.

The animals were divided into three groups: Group I (n = 5) served as a control group and received no bacteria or drugs but only the carrier solutions. Group II (sepsis group, n = 6) received a continuous intravenous bacterial infusion of 6 × 10⁶ cfu · kg¹ · h¹ P. aeruginosa. The infusion was started after the baseline hypoxic challenge. Group III (L-NMMA group, n = 8) also received the continuous bacterial infusion. After 24 h the animals received an 8 h continuous infusion of L-NMMA 6.6 mg · kg¹ · h¹.

At completion of the experiment, the animals were anesthetized with ketamine (10 mg · kg¹) and killed by intravenous injection of 50 ml of saturated KCl solution.

Stock cultures of P. aeruginosa were maintained in 0.9% saline at 70° C. Twenty-four hours before bacterial infusion, 1 ml of the stock solution was inoculated into 150 ml trypticase soy broth (Difco Laboratories, Detroit, MI). Bacteria were harvested by centrifugation and quantitated with a Pentroff-Hausser counting chamber (Hausser Scientific, Blue Bell, PA). The motility of the bacteria was observed under a microscope. The bacteria were washed twice during the preparation to minimize contamination by bacterial products, including endotoxin. Standard agar pour-plate technique was used to verify the bacterial concentration. After final centrifugation, the pellet was resuspended in 0.9% saline to yield a concentration of 6 × 10⁶ cfu · kg¹ · h¹ P. aeruginosa at an infusion rate of 2 ml · kg¹ · h¹.

L-NMMA was dissolved in normal saline and then infused at a dose of 6.6 mg · kg¹ · h¹; the infusion rate was 1 ml · kg¹ · h¹. This dose of L-NMMA has been shown in prior studies to be sufficient to return mean arterial blood pressure to presepsis values. We found no further increase in HPV with higher doses of L-NMMA. Lingnau and colleagues (15) demonstrated that plasma levels with this dose remain stable for 24 h.

Arterial, central venous, pulmonary arterial, and left atrial pressures were measured with transducers (Baxter-Edwards Critical Care) and recorded on a hemodynamic monitor (Model 78304; Hewlett Packard, Santa Clara, CA). A horizontal plane 12 cm above the sternum was taken as a zero reference point for vascular pressures. Cardiac output was determined in triplicate with the thermodilution technique on a cardiac output computer (Model 9530; Baxter-Edwards Critical Care). Ten milliliters of ice-cold 5% dextrose solution were used as the indicator.

Left pulmonary arterial blood flow was measured continuously with the transit-time ultrasonic flow probe connected to an ultrasonic blood flowmeter (Transonic Systems). Blood flow bypassing the right heart via the bronchial artery (anastomotic flow) could therefore not be detected. Cardiac index was calculated with the following formula for body-surface area of the sheep: body surface area (m²) = body weight (kg)^2/3 × 0.087.

HR was determined from the arterial tracing. Pulmonary and systemic vascular-resistance indices, stroke-volume index (SVI), and right- and left-ventricular stroke-work indices were calculated according to standard formulas. Left pulmonary vascular resistance was calculated as the difference between mean pulmonary arterial pressure (Ppa) and left atrial pressure multiplied by 80 and divided by left pulmonary blood flow. HPV was expressed as the percent decrease in left pulmonary blood flow during left lung hypoxia as compared with normoxia. Left pulmonary blood flows during hypoxia and normoxia were corrected for cardiac output at the time point of measurement.

Samples of arterial and mixed venous blood were analyzed for gas tension and pH (System 1302; Instrumentation Laboratory, Lexington, MA), and the data were corrected for core body temperature by the apparatus. Oxyhemoglobin saturation, carboxyhemoglobin, and methemoglobin concentrations were determined with an oximeter (CO-Oximeter 282; Instrumentation Laboratory). Shunt blood flow was estimated from a standard shunt equation.

Statistical analysis was performed with analysis of variance (ANOVA) for repeated measures to detect differences within a group, and with factorial ANOVA for repeated measures to detect differences between groups, with post hoc analysis done with Fisher's least-significant-difference procedure. Bonferroni's correction for multiple comparisons was used when appropriate. Statistical significance was set at p < 0.05. Data are presented as means and SEM.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The hemodynamic values found in the study are shown in Tables 1 and 2. The control group (Group I) showed stable values throughout the study period. After 24 h of bacterial infusion, the animals in the sepsis group (Group II) had developed a hyperdynamic circulatory state, with increased cardiac output, tachycardia (from 80 ± 4 beats · min¹ to 117 ± 9 beats · min¹ at 24 h, p < 0.05), systemic hypotension, and decreased systemic vascular resistance without further change throughout the remainder of the study. Although pulmonary artery pressures increased, pulmonary vascular resistance was not elevated. There was no difference in either Ppa or vascular resistance during normoxia and left lung hypoxia. Group III also developed a hyperdynamic circulatory state 24 h after the P. aeruginosa infusion began. Four hours after the start of the L-NMMA infusion, CI, mean arterial pressure, and systemic vascular resistance had reached values similar to their baseline levels prior to the bacterial infusion. HR, which had increased from 91 ± 4 beats · min¹ prior to the bacterial infusion to 120 ± 6 beats · min¹ at 24 h (p < 0.05), decreased to 108 ± 4 beats · min¹ after L-NMMA (p < 0.05 versus 24 h). Ppa and pulmonary vascular resistance were significantly increased over their 24-h values, without differences between normoxia and hypoxia. At 8 h after discontinuing L-NMMA, its effects had worn off. Left atrial pressures and central venous pressures showed comparable results in all three groups (data not shown).

Table 3 shows the oxygenation variables measured in the study. There were no changes in pH, Pa_O2, or Pa_CO2 at any time or between groups. These values did not change during left lung hypoxia, except that the Pa_O2 in Group II was significantly higher at 0 h than at the following time points. This was caused by high Pa_O2 values at 0 h in two animals with a profound decrease in left pulmonary blood flow during hypoxia. Pv_O2 remained stable throughout the study period, except that it was increased in Group III at 24 h as compared with its baseline value. There was, however, no difference to the other two groups at this time point.

Figure 1 shows HPV. The control group showed a reproducible response at all time points. At 48 h, HPV was slightly reduced. Both groups receiving P. aeruginosa demonstrated depressed but not completely abolished pulmonary vasoconstriction during left lung hypoxia. In Group III (L-NMMA), HPV was improved at 4 h after L-NMMA administration as compared with 24 h and with the sepsis group, but was again blunted at 48 h after L-NMMA had been stopped. Left pulmonary vascular resistance during normoxia (Figure 2) increased in Group III during sepsis, and increased further after L-NMMA. During hypoxia, left pulmonary vascular resistance was decreased at 24 h and increased after L-NMMA administration. Pulmonary shunt during the hypoxic challenges (Figure 3) increased during sepsis, but was significantly improved during L-NMMA administration and increased again after its discontinuation.

View larger version (30K): [in this window] [in a new window]   Figure 1.   Hypoxic pulmonary vasoconstriction. HPV during left lung hypoxia, expressed as percent decrease in left pulmonary blood flow compared with normoxic conditions. Control group (open bars), sepsis group (stippled bars), L-NMMA group (filled bars). * p < 0.05 versus 0 h;  p < 0.03 versus 24 h (sepsis);  p < 0.05 versus 28 h (L-NMMA); § p < 0.05 versus sepsis group. After 24 h of sepsis, HPV is significantly blunted in both groups receiving the bacterial infusion (sepsis and L-NMMA group). Four hours after the beginning of L-NMMA infusion, HPV had increased as compared with the 24-h value and the sepsis group at 28 h, indicating partial restoration. Sixteen hours after stopping the L-NMMA infusion, its effect had worn off.

View larger version (40K): [in this window] [in a new window]   Figure 2.   Left pulmonary vascular resistance index (LPVRI). LPVRI expressed as percent of baseline LPVRI at 0 h; normoxic conditions (A) and left lung hypoxia (B). Control group (open bars), sepsis group (stippled bars), L-NMMA group (filled bars). * p < 0.07 versus 0 h;  p < 0.05 versus 24 h (sepsis);  p < 0.05 versus 28 h (L-NMMA); § p < 0.05 versus sepsis group. During normoxia, LPVRI rose in the L-NMMA group during sepsis, and further after L-NMMA administration. After 24 h of sepsis, LPVRI rose during left lung hypoxia to a smaller extent than at 0 h, but was significantly increased after L-NMMA. At 48 h the value had fallen again to a level similar to that during sepsis.

View larger version (35K): [in this window] [in a new window]   Figure 3.   Intrapulmonary shunt. Shunt in percent of baseline. Control group (open bars), sepsis group (stippled bars), L-NMMA group (filled bars). *p < 0.05 versus 0 h; p < 0.05 versus 24 h (sepsis);  p < 0.05 versus 28 h (L-NMMA); § p < 0.05 versus sepsis group. After 24 h of sepsis, shunt was increased; after L-NMMA infusion, shunt decreased.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Infusion of P. aeruginosa in our ovine model of chronic sepsis induced a hyperdynamic circulatory state, with decreased systemic vascular resistance, hypotension, and increased cardiac output. HPV was significantly blunted but not completely abolished at 24 h after beginning the bacterial infusion, and pulmonary shunt was increased. L-NMMA reversed the hyperdynamic circulation to presepsis baseline levels. Pulmonary shunt decreased. HPV was eased but not fully reversed.

The sheep model of chronic sepsis was well suited for our experiments. Sheep have hemodynamic values similar to those of humans, and comparable collateral ventilation. Other animals, such as pigs, have almost no collateral ventilation and have a much more pronounced vasoconstrictive response to hypoxia (18). Our experiments were performed in awake animals without the use of anesthetics, since inhalational anesthetics are known to blunt HPV, and even pentobarbital, a commonly used drug in anesthetized animals, has been reported to attenuate HPV in isolated sheep lungs (19). The size of sheep allowed easy placement of our modified double-lumen endotracheal tube for separate lung ventilation. After sepsis had begun, meticulous bronchial toilette was necessary to remove secretions, minimize the work of breathing, and avoid atelectasis, which would alter hypoxic vasoconstriction. At autopsy, we found minimal atelectasis, equally distributed in both lungs. It is possible, however, that at the end of the experiments some mucus plugging occurred, since a small degree of atelectasis was seen.

Although pulmonary artery pressure rose during sepsis in the sheep, pulmonary vascular resistance was not increased, which is in contrast to the finding that many septic patients with respiratory failure show increased pulmonary vascular resistance (20). However, it has been shown that pulmonary vascular resistance is inversely related to cardiac output (20), and since all of the sheep in our study were hyperdynamic, with an increased cardiac output, this may explain our findings. Other authors studying HPV in other species also found no increase in pulmonary vascular resistance after administration of endotoxin or tumor necrosis factor (TNF), or bacterial infection (13, 21). In contrast, McCormack and Paterson reported an increase in pulmonary vascular resistance in a rat model of chronic pneumonia, but cardiac output in their control and infected rats did not differ (22).

Many studies of HPV have used global hypoxia as a stimulus, and have measured the change in Ppa and resistance. However, homogeneous global hypoxia is much less often encountered clinically than is inhomogeneous A/ mismatch. Investigating HPV by making only parts of the lung hypoxic better mimics A/ mismatch. Redistribution of blood flow within the lung during our hypoxic challenges occurred without an increase in Ppa or resistance, indicating that although vasoconstriction occurred in hypoxic areas, vasodilation occurred in the well-oxygenated parts of the lung. The degree of flow diversion is inversely proportional to the size of the hypoxic segment, ranging from 75% with a small segment to zero when the whole lung is hypoxic (23). The degree of HPV is also dependent on the P O₂ of the blood in the bronchial artery. A low Pa_O2 in the bronchial arterial blood causes an increase in pulmonary vascular resistance, whereas at very low bronchial Pa_O2 values, pulmonary vascular resistance decreases (24). Since the bronchial artery supplies the vasa vasorum to all but the smallest pulmonary vessels, global hypoxia interferes with HPV. By ventilating the right lung with 100% oxygen, we were able to avoid global hypoxia in our sheep.

We demonstrated that HPV can be induced repeatedly and to the same extent in healthy animals over a 48-h period. After sepsis induction, HPV was blunted and showed no adaptation over time. We chose repeated exposures rather than a continuous challenge, in order to avoid oxygen toxicity of the right lung. Continuous ventilation with 100% oxygen has been shown to induce respiratory failure in sheep, and by itself inhibits HPV (25). Other variables that could influence HPV, such as acid-base status, alveolar PO₂, or central venous PO₂ were rather constant in our experiments. Increased cardiac output could interfere with HPV, but inhibition of HPV in endotoxemia has been documented with an increased and, as compared with the value before endotoxemia, decreased cardiac output (12). However, we cannot exclude the possibility that the decrease in cardiac output after L-NMMA administration contributed to the improvement in HPV.

Both Gibson and associates (26) and Zelenkov and coworkers (27) reported enhancement of HPV by endotoxin. The different response to endotoxin could be explained by the different study design. Gibson and associates administered a low-dose, gram-positive bacterial infusion to neonatal piglets, which did not have the detrimental effect on oxygenation reported with high-dose infusions. Data with which to assess the systemic circulation, other than cardiac output, were not provided. Zelenkov and coworkers studied isolated rat aortic and pulmonary rings, with and without exposure to endotoxin and under normoxic and hypoxic conditions. They reported enhanced vasoconstriction after administration of endotoxin. Their explanation for this finding was that hypoxia deprives NOS of oxygen, therefore inhibiting its action. Thus, hypoxia acts as an NOS inhibitor. After pretreatment with L-nitroarginine methyl ester (L-NAME), which induced a contraction of the vessels, hypoxia caused only a small additional contraction. In the intact animal without global hypoxemia, in contrast to isolated vessels, NOS in the pulmonary vasculature may be sufficiently supplied with oxygen via the bronchial circulation to maintain its full activity. These differences in study design may account for the different findings.

The NOS inhibitor L-NMMA reversed sepsis-induced hypotension in our study by increasing systemic vascular resistance. These findings are supported by prior work done in our laboratory (15, 28), as well as by other investigators (16, 17, 29, 30). Although L-NMMA has a greater affinity for iNOS than L-NAME, it remains a nonspecific NOS inhibitor. The use of a specific iNOS inhibitor might have elucidated the role of iNOS in relation to the constitutive NOS (cNOS) in sepsis. On the other hand, there is evidence for inhibition of NOS by NO itself, in the form of a negative feedback mechanism (31, 32). Whereas cNOS is very sensitive to NO, iNOS requires much higher concentrations of NO to be inhibited. In sepsis, in which iNOS is induced and produces large enough amounts of NO to cause systemic hemodynamic effects, it is likely that cNOS activity is suppressed, and it is unlikely that we may have picked up subtle hemodynamic differences with our experimental design.

In healthy animals, NOS inhibitors augment HPV (11, 33). Very little is known, however, about the effect of NOS inhibitors on HPV in sepsis. McCormack and Paterson (22) induced chronic pneumonia in rats by injecting a P. aeruginosa slurry into a distal bronchus. Seven to 10 days later, HPV was assessed during global hypoxia as the change in Ppa and vascular resistance as compared with these values in uninfected control animals. As compared with the healthy controls, the infected animals showed a significantly smaller increase in Ppa but not in pulmonary vascular resistance. After a bolus of 50 mg · kg¹ L-NMMA, Ppa and vascular resistance increased in both groups of rats, and there was no significant difference in the increase between healthy and infected animals. McCormack and Paterson concluded that NO does not account for the attenuated hypoxic pressure response in pneumonia. Blood flow to the infected part of the lung constituted only 15% of total lung blood flow before and after L-NMMA. The degree of infection was assessed by bacterial colony counts; the degree of hypoxia of the infected lung area was not determined. Subtle changes would have been difficult to detect in McCormack and Paterson's study, especially when taking into account our finding that L-NMMA only partly restored HPV.

One explanation for why HPV was only partly restored is that other vasodilatory mediators than NO contribute to the reduction of HPV. For example, levels of atrial natriuretic peptide (ANP), which causes cyclic guanosine monophosphate (cGMP)-mediated vasodilation independent of NO, are increased in sepsis (34). Administration of an ANP receptor antagonist reversed ANP-mediated vasodilation in healthy and septic animals (35, 36). In addition, levels of ANP were increased in animals subjected to acute global hypoxia (37). Infusion of exogenous ANP reduced the hypoxia-induced increase in Ppa and vascular resistance (37). Whether ANP plays a role in loss of HPV in sepsis is unknown.

Endotoxin also causes the release of mediators such as TNF, platelet-activating factor (PAF), interleukin-1 (IL-1), or -interferon (IFN-), which in turn stimulate the iNOS and eventually lead to increased NO secretion (38). Administration of TNF- to rats (21) was associated with loss of HPV, which was restored neither by nitro-L-arginine nor by L-NMMA, suggesting that TNF- diminished HPV by a mechanism independent of NO. Endotoxin also stimulates the secretion of PAF. Administration of PAF blunted the increase in Ppa during global hypoxia in healthy rats (14). Its antagonist, CV3988, partly prevented the decrease in the pulmonary pressure response to hypoxia after endotoxin administration (14).

In addition, prostaglandins play a role in modulating HPV. Enhancement of the HPV response by prostaglandins has been reported in healthy animals (11, 39) and in endotoxemia (40). Endotoxin administration to perfused, isolated dog lungs abolished the hypoxic pressure response, which was, however, preserved after addition of indomethacin to the perfusing blood (41). In contrast, meclofenamate failed to restore HPV in endotoxemic sheep (42).

The role in sepsis of another candidate mediator of HPV, endothelin-1, is not clear. Endothelin-1 gene expression and secretion in cultured cells are stimulated by hypoxia and potentiated by the addition of an NOS inhibitor (43). In healthy rats, pulmonary vasoconstriction during acute global hypoxia, and pulmonary vascular and right-heart remodeling during chronic hypoxic exposure, were inhibited by pretreatment with a nonselective endothelin-1 antagonist (44). Contrastingly, pulmonary vasodilation and inhibition of HPV have also been reported after endothelin-1 injections (45). In vitro, lungs from endotoxemic rats showed an enhancement of the HPV response after administration of a nonspecific endothelin-1 antagonist (46). The effects of endothelin-1 in vivo during endotoxemia are yet to be elucidated.

In summary, continuous infusion of live P. aeruginosa to wakeful sheep induces a hyperdynamic circulatory state. HPV is blunted. Inhibition of NOS with L-NMMA reverses the circulatory state and partially restores HPV. An explanation for the incomplete response could be the complex interaction of multiple vasodilatatory and vasoconstrictive mediators of HPV that are also activated during sepsis. Although restoration of HPV and reduction of pulmonary shunt fraction are very desirable in critically ill patients, especially in those with concomitant respiratory failure, the benefits of NOS inhibitors for this purpose should be carefully weighed against their undesirable effects, such as pulmonary hypertension.