INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Previous investigations have shown that inhaled nitric oxide (NO) can alleviate the pulmonary hypertension and derangement in gas exchange subsequent to endotoxin lung injury (1, 2), but the influence on lung fluid filtration has not been specifically addressed. The purpose of the present experiments was to study the effect of inhaled NO on lung lymph flow (L) and protein concentration during the late phase of endotoxemia in awake sheep. In sheep, bolus injection of endotoxin is followed by a biphasic response: an early phase, lasting for approximately 1 h, of rapid increase in pulmonary vascular pressure and resistance, paralleled by an increase in protein-poor L. This response is probably more extreme than is seen in humans. During the late phase, lasting for four or more hours, the hemodynamic changes gradually return toward baseline while L increases further. Concomitantly, the lymph/plasma protein concentration ratio (L/P) returns to normal, or even higher than baseline, due to increasing lung vascular permeability (3, 4).

The mechanisms underlying the vascular abnormalities in endotoxemia are incompletely understood. There is substantial evidence that the endotoxin reaction involves formation of arachidonic acid metabolites and cytokines (5, 6). Furthermore, support is accumulating that endotoxin stimulates endothelial and inflammatory cell NO synthase, which generates NO production from L-arginine (7). Vasodilation secondary to NO generation, with or without impairment of cardiac output, is believed to be one of the main reasons for the systemic arterial hypotension after endotoxemia (8). The hypotension, but not the impaired cardiac output, is effectively blocked by the NO synthase inhibitor N^G-methyl-L-arginine (9). During endotoxemia, a possible pulmonary vasodilator effect of endogenous NO may be obscured by opposing vasoconstrictive forces, mainly due to thromboxane A₂ in the early phase and release of leukotrienes and the potent endothelium-derived vasoconstrictor protein, endothelin, during the late phase (5, 10). Attention has been paid recently to the potential of inhaled NO to counteract pulmonary hypertension and the derangement in gas exchange in acute respiratory distress in humans (11). Because NO is rapidly inactivated by hemoglobin, the effects of inhaled NO are limited to small vessels adjacent to ventilated areas (14, 15), giving rise to more favorable ventilation/perfusion ratios and improved gas exchange. In isolated rabbit lungs exposed to oxidant injury, it was postulated that inhaled NO may reverse the increase in microvascular permeability (16). Permeability is difficult to assess in situ without an accurate measure of microvascular pressure and assessment of the permeability surface product (17).

We hypothesized that NO might reduce lung vascular filtration primarily in small vessels by attenuating endotoxin- induced pulmonary hypertension, and secondarily by lowering microvascular permeability where most lung filtration occurs. Additionally, we expected that NO would enhance arterial oxygenation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Fourteen unanesthetized yearling sheep (25-30 kg) with lung lymph cannulas were studied. In addition, we investigated effects of inhaled NO on pulmonary microcirculation in six non-lymphing sheep. The procedures described in this article have been carried out in accordance with the specifications of the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and the protocol was approved by the Vanderbilt University Animal Use Committee.

Surgical Preparation

Sheep were anesthetized and catheters were inserted as previously described (3, 4). Through a right thoracotomy, the caudal mediastinal lymph node was isolated and the efferent duct was cannulated with a silastic catheter for lymph sampling. To prevent contamination with nonpulmonary lymph, the distal tail of the lymph node, along with all lymphatic tissue traversing the diaphragm, was interrupted with two ligatures via a second right-sided thoracotomy, just above the diaphragm. Via a left thoracotomy, silicone catheters were implanted directly into the pulmonary artery and the left atrium. The six non-lymphing sheep were prepared similarly with the exception of the right-sided thoracotomies.

After surgery, the animals were allowed 5-7 d of recovery before they underwent tracheotomy, performed under anesthesia through the second and fourth midline cartilages. A silicone catheter was inserted into the thoracic aorta via the right common carotid artery, and a number 7.5 Cordis introducer (Cordis, Miami, FL) was placed in the right external jugular vein for the later introduction of a Swan-Ganz pulmonary artery thermal dilution catheter (model 93A-131-7F; Edwards Laboratories, Anasco, PR) on the day prior to experimentation.

Animals were maintained in clean pens with free access to food and water. Sheep were injected with gentamicin 80 mg (Schering-Plough, NJ) and the vascular catheters were flushed with heparin (1,000 IU/ml) 3-4 ml daily. A recovery time of 2-3 d was allowed after the tracheotomy before entering the study protocol.

Experimental Protocol

A tracheal cannula was inserted, the cuff was inflated, and the sheep was placed in an experimental pen. Variables were measured for 2-3 h to establish a baseline. Escherichia coli endotoxin (Sigma Chemical, St. Louis, MO) 0.75 µg/kg dissolved in 30 ml of saline was infused for 20 min from time 0. Except for L, which was determined every 15 min, all other measurements were carried out at 30-min intervals. Non-lymphing sheep were exposed to endotoxin 1 µg/kg. In the latter, hemodynamic data were determined first at 30 min and then at 1-h intervals.

In lymphing sheep, 2 h after endotoxin, a gas mixture containing 25% oxygen in nitrogen (fraction of inspired oxygen [FI_O2] 0.25) was administered via a delivery system consisting of a 50-psi gas blender (Puritan-Bennett, Carlsbad, CA) in series with a rotameter. A 3-L anesthetic bag served as a reservoir and was attached to the tracheotomy tube via a corrugated ventilator tubing and a one-way valve to avoid rebreathing. Thirty minutes later, in the NO group, the nitrogen (N₂) of the inspired gas was replaced for 1 h with a mixture containing NO 50 ppm in N₂, while the inspiratory gas of the control group remained unchanged throughout the experiment. The non-lymphing sheep breathed FI_O2 0.25 throughout the experiment interrupted for 3-5 min intervals hourly for inhalation of the NO gas mixture.

Mean aortic (Psa), pulmonary arterial (Ppa), and left atrial (Pla) pressures were recorded continuously with the zero reference at a point on the shoulder corresponding to the level of the left atrium. Pressures were preamplified using a Validyne model ML1-10-871 (Validyne, Northridge, CA) and recorded on an Astro-Med MT 95000 recorder (Astro-Med, W. Warwick, RI). Cardiac output (CO) was determined every 30 min as the mean of six injections of 5 ml each iced saline using a cardiac output computer (Model 9520; Edwards, Santa Ana, CA). Pulmonary vascular resistance (PVR) was calculated as (Ppa  Pla)/CO and systemic vascular resistance (SVR) as Psa/CO. Pulmonary microvascular pressure (Pmv) was estimated as Pmv = Pla + 0.4(Ppa  Pla) (18).

In the six non-lymphing sheep, pulmonary microwedge pressure (Pmw) was determined by advancing the Swan-Ganz catheter into the wedge position in a distal pulmonary artery with the balloon deflated. The criteria for attainment of the microwedge position included: (1) easy retrograde aspiration of blood from the catheter; (2) pH, PO₂, and PCO₂ of aspirated blood (Pwedge) consistent with wedged position, i.e., PwedgeO₂ > Pa_O2 and PwedgeCO₂ < Pa_CO2; (3) microwedge pressure greater than proximal wedge pressure; and (4) microwedge pressure greater than Pla, with true zero confirmed by connecting the Pla catheter and microwedge catheter sequentially to the same fixed transducer. PVR in upstream (PVR upstream) and downstream (PVR downstream) vessels were calculated as follows: (Ppa  Pmw)/CO and (Pmw  Pla)/CO, respectively (19).

Lung lymph was collected in heparinized tubes and L was determined every 15 min. Every 30 min, lymph and blood samples were obtained for determination of lymph and plasma protein concentrations, respectively, using the modified biuret method with an automated analyzer (Autoanalyzer II; Technicon Instruments, Tarrytown, NY). L/P was determined, and lung lymph protein clearance (CL) was calculated as L × L/P (ml/15 min).

The concentration ratio of interstitial fluid to plasma protein has been shown to fall hyperbolically with increasing net fluid filtration and lymph flow, according to the equation L/P = PS/(PS + L), where PS is the permeability surface product for protein. A prerequisite for this assumption is that transcapillary protein transport is purely dissipative (diffusion or vesicle transport proportional to the transcapillary protein concentration difference) (17, 20). After reorganizing the formula, we calculated mean PS at baseline in each of the groups (values as of  0.5 and 0 h were presented together), at 2.5 h, and from 3.0 through 4.5 h (values as of 3.0 through 4.5 h were presented together). Iso-PS lines were drawn by keeping the calculated PS at the selected time intervals constant while varying L between 0 and 10 ml/15 min. The derived L/P values were plotted against L.

Barometric pressure was noted daily. FI_O2 was monitored continuously using a Puritan-Bennett 7820 oxygen monitor (Puritan-Bennett, Carlsbad, CA). In sheep with a lung lymph catheter, samples for blood pH and gas tensions were obtained every 30 min from the systemic artery (a) and pulmonary artery (v) lines, respectively, and analyzed for pH, PO₂ (mm Hg), PCO₂ (mm Hg), and oxygen saturation (SO₂,%) using a Ciba-Corning 238 pH/blood gas analyzer (Corning Medical Co., Medfield, MA). Hemoglobin samples were taken simultaneously from arterial blood. Alveolar oxygen tension (PA_O2) was calculated according to the alveolar gas equation: PA_O2 = (PB  PH₂O) · FI_O2  Pa_CO2/RQ, assuming a respiratory quotient (RQ) of 0.8. Alveolar-arterial oxygen tension difference AaPO₂ was calculated as well. The pulmonary end-capillary oxygen saturation (Sc'O₂) was calculated using a computer version of the Kelman and Nunn equation for hemoglobin oxygen saturation (21). Capillary (Cc'O₂), arterial (CaO₂), and venous blood oxygen contents (CvO₂) were calculated, assuming a hemoglobin oxygen binding capacity of 1.39 ml/g. By means of these calculations, we estimated oxygen delivery as CaO₂ × CO (ml/min), oxygen consumption (O₂) as (CaO₂  CvO₂) × CO (ml/min), and venous admixture (S/T) employing the shunt equation S/T = (Cc'O₂  CaO₂)/(Cc'O₂  CvO₂) (22). Arterial blood leukocyte counts were made every 30 min using a Coulter counter (Coulter Electronics, Inc., Hialeah, FL).

Statistics

Data are presented as the mean ± the standard error of the mean (SEM). Differences within and between groups (relative to intervention and time) were identified by one- and two-way analysis of variance. Wilcoxon rank sum test was used for evaluation of differences between iso-PS lines and L/P-L relationships. The null hypothesis was rejected for values of p < 0.05. Data sampling and statistics were computed using Excel 5.0 (Microsoft Corporation, Redmond, WA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hemodynamics

Pulmonary hemodynamic changes following bolus injection of endotoxin in lymphing sheep, with and without the 1-h period of inhaled NO, are displayed in Figure 1. In both groups, an approximately threefold rise in Ppa was paralleled by a nearly fivefold increase in PVR. The intragroup changes in Ppa and CO were maximal 30-45 min after endotoxin. After endotoxin, Ppa was significantly higher than at baseline for the first 2.5 h. With the onset of NO, Ppa and PVR dropped abruptly in comparison to the preceding intragroup values and to the control group. After withdrawal of NO, Ppa and PVR returned to pretreatment levels. In both groups, Pla decreased in parallel, reached nadir 1 h after endotoxin, and returned gradually toward baseline, displaying no intergroup differences (Table 1). Systemic vascular resistance rose transiently 0.5 h after endotoxin because of the reduction in CO, while Psa remained unchanged. Inhaled NO, interposed between 2.5 and 3.5 h, reduced calculated Pmv from 9.9 ± 1.1 cm H₂O to 3.9 ± 1.1 cm H₂O at 3.0 h and 4.9 ± 1.1 cm H₂O at 3.5 h, corresponding to 60% and 50%, respectively. Tables 2 and 3 show hemodynamic changes during periods of intermittent NO inhalation. In the late phase, 2-4 h after endotoxin, inhaled NO reduced Pmw 33-37%. Simultaneously, PVR upstream declined by 29- 37% and PVR downstream by 22-29%. CO and systemic hemodynamic variables did not vary from those noted in lymphing sheep.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Hemodynamic reactions to endotoxin (ENDO = 0.75 ug/kg) in awake sheep with and without inhalation of nitric oxide (NO = 37.6 ppm). In both the control (open squares) and the NO group (filled circles), each consisting of seven animals, ENDO was injected over 20 min at time 0 h. Sheep breathed FIO2 0.25 for 2 h (arrows). CO = cardiac output; Ppa = mean pulmonary artery pressure; PVR = pulmonary vascular resistance. Data are presented as means ± SEM. @ denotes p < 0.05 from intragroup baseline in both groups; * denotes p < 0.05 between groups; + denotes p < 0.05 from control intragroup baseline.

Lung Lymph

Lung lymph data are displayed in Figures 2 and 3. In both groups, L and CL increased significantly above baseline within 30 min after endotoxin. During inhalation of NO, L and CL fell significantly, and L remained below that of the control group even after withdrawal of NO, albeit a slight increase occurred at 4.5 h (NS). After endotoxin, L/P decreased gradually but not significantly, and then increased to a level near baseline 2 h later. Inhaled NO transiently increased L/P to a level significantly higher than at baseline, but despite reduction in Pmw by NO (Table 3), both L and CL remained significantly higher than at baseline. Figure 3 depicts the mean L/P-L relationship in the NO group (closed circles) and the control group (open squares), at baseline (left), at 2.5 h (start of arrows), and from 3.0-4.5 h (arrowheads). Iso-PS lines, dashed in the NO group and continuous in the control group, were calculated at the same time points. Inhaled NO slightly but significantly reduced the displacement of the L/P-L relationship, as compared with the control group, which was shifted beyond the most rightward iso-PS line.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Effects of 1 h of inhaled NO 37.6 ppm on lung lymph balance in awake endotoxemic sheep. Time from baseline in hours; L = lung lymph flow; L/P = lymph-to-plasma protein concentration ratio; CL = lymph protein clearance. CL =  L × L/P. Data are presented as means ± SEM. @ denotes p < 0.05 from intragroup baseline in both groups; * denotes p < 0.05 between groups; + denotes p < 0.05 from NO intragroup baseline. For further information, see Figure 1.

View larger version (14K): [in this window] [in a new window]   Figure 3.   Relationship between mean L/P and mean L at baseline (data at 0.5 h and 0 h are presented together), at 2.5 h, and from 3-4.5 h (presented together) in two groups of awake endotoxemic sheep. The corresponding mean permeability surface product (PS) iso-lines for the control group (solid line) and the NO group (dashed line) are superimposed. Iso-PS lines at baseline (left) were significantly different from those at and after 2.5 h. Both the control group (open squares) and the NO group (closed circles) received endotoxin from time zero; the latter additionally received inhaled NO, 37.6 ppm, for 1 h from time 2.5 h. Inhalation of NO shifted the L/P- L relationship significantly to the left (arrow to the left) compared with the time corresponding control value (arrow to the right). Data are presented as means ± SEM. For further information, see Figure 1.

Leukocytes

Leukocyte counts (Table 4) fell significantly after endotoxin in both groups, at one-half hour in the control group and 1 h in the NO group. A striking intergroup difference with higher counts was found after 1 h of NO inhalation, which persisted 30 min after withdrawal.

Gas Exchange and Metabolism

Figure 4 shows changes in gas exchange subsequent to endotoxin with and without inhalation of NO. Initially, Pa_O2 declined and AaPO₂ and S/T increased. At 2 h, when both groups were exposed to FI_O2 0.25, Pa_O2 was still lower than at baseline and there was no intergroup difference. In sheep exposed to NO, Pa_O2 rose significantly compared to baseline and to control subjects. After FI_O2 was changed to 0.25, AaPO₂ initially increased and then decreased during NO inhalation. Thirty minutes after withdrawal of NO, AaPO₂ returned to significantly above baseline values. S/T, increased significantly after endotoxin, but without intergroup differences for the first 2 h. At 2.5 h S/T was significantly higher in the animals subsequently given NO. NO caused an immediate drop compared to the preceding intragroup value as well as to the control group. Additional gas exchange and metabolic variables are depicted in Table 4 and Table 5, respectively. A fall in Sa_O2 was observed from 0.5 h through 2 h without differences between the groups. Likewise, venous oxygen saturation (Sv_O2 fell from 0.5 through 1 h. Although both Sa_O2 and Sv_O2 increased during NO inhalation, the changes did not reach statistical significance. The decrease in oxygenation induced a nadir in Pa_CO2 at 2 h, but pH remained unchanged (Table 5). Hemoglobin increased subsequent to endotoxin in both groups, but the increase was significant only at 1 and 1.5 h in the NO group. Oxygen delivery decreased at 0.5 h in both groups, but no significant intergroup differences were found. Oxygen consumption did not change significantly, neither within nor between the groups.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Effects of 1 h of inhaled NO 37.6 ppm on lung gas exchange in awake endotoxemic sheep. Time from baseline is in hours; PaO2 = arterial oxygen tension; AaPO2 = alveolar-arterial oxygen tension difference; S/ T = venous admixture. Data are presented as means ± SEM. @ denotes p < 0.05 from intragroup baseline in both groups; + denotes p < 0.05 from intragroup baseline in one group; § denotes p < 0.05 from intragroup value before NO; * denotes p < 0.05 between groups. For further information, see Figure 1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates that inhalation of NO reduces lung fluid filtration and protein efflux in awake endotoxemic sheep. Concomitantly, inhaled NO normalizes PVR, improves arterial oxygenation, and partially reverses the endotoxin- induced fall in blood leukocyte counts. NO attenuates pulmonary hypertension, particularly during the late phase of endotoxemia, less so during the early phase. Based on the latter finding and the observation that inhaled NO may influence microvascular permeability following oxidant injury in isolated rabbit lungs (16), we sought evidence that NO might modify the increase in fluid filtration during the late phase of increased lung vascular permeability and pressure after endotoxin. We were also stimulated, in part, by reports showing that endogenous NO attenuates increased microvascular permeability in ischemia-reperfusion injury to gut (23) and bradykinin injury of the hamster cheek pouch (24).

In lymphing sheep, the most remarkable effect of NO was the complete reversal of the endotoxin-induced enhanced pressure gradient between the pulmonary artery and the left atrium (Ppa-Pla). Compared with control subjects, NO rapidly reduced Ppa-Pla from 31 to 16 cm H₂O. This reduction is consistent with our own finding in non-lymphing sheep and with that of investigators employing NO in anesthetized piglets exposed to beta hemolytic streptococci (1). Using NO 150 ppm, the latter workers found reduction in PVR both in the early and late phase reaction to streptococcus (1).

During the early phase of the present experiments, an approximately 60 cm H₂O pressure gradient occurred between the pulmonary artery and left atrium in concert with a nearly 300% increase in L. The severity of this hemodynamic response exceeds that observed in man. Despite normalization of the Ppa-Pla gradient by NO, it was striking to find that the 270% increase in late-phase L was reduced by only 59% after 1 h of NO inhalation. We are not aware of studies specifically designed to evaluate endotoxin-induced changes in L and protein content with inhaled NO. On the other hand, investigators studying oleic acid lung injury in pigs could find no significant effect on lung extravascular water content (EVLW) of NO 10 to 80 ppm inhaled for 10- to 30-min periods. In the latter investigation, EVLW was determined by means of a thermal double-indicator technique (25).

We did not measure the longitudinal pressure profile between the pulmonary artery and left atrium in lymphing sheep. Estimation of lung microvascular pressures indicated more than 50% reduction during NO. Because the calculation of Pmv, according to Gaar and colleagues (18), is based on a fixed ratio between post-capillary and total lung vascular resistance, determination of microvascular pressure was addressed specifically in a group of non-lymphing sheep. These experiments revealed that NO acts both as pre- and post-capillary vasodilator in the constricted pulmonary circulation after endotoxin, reducing Pmw to baseline levels. The effect was slightly larger on upstream than on downstream resistance. The results suggest a reduced pressure gradient between the downstream microvessels and left atrium during NO, albeit not sufficiently low to prevent the enhanced filtration associated with a higher vascular permeability during the late phase of endotoxemia (26). However, it was noteworthy (Figure 3) that the shift of the L/P-L relationship toward more rightward displaced iso-PS lines during endotoxemia was hampered by inhaled NO. We have no specific information about vascular recruitment during inhalation of NO but assume that surface area is grossly unchanged (27). Thus, we suggest that NO precludes but does not reverse the endotoxin-induced rise in microvascular permeability. The latter observation is also consistent with the finding that L/P remained nearly unchanged despite a reduction in L by nearly 60% in response to inhalation of NO.

Figure 3 shows that the mean L/P-L relationships are located apart from the corresponding iso-PS lines. This is because the mean iso-PS lines are calculated on the basis of constant PS values being derived from mean L/P and L values, as calculated at each time point, and not based on individual values. In situ the L/P-L relationships do not result purely from dissipative transport but also involves hydrostatic pressure gradients across the microvasculature. Consequently, determination of permeability in lungs in situ by studying protein flux should be interpreted with caution, unless microvascular pressure is reasonably controlled.

In isolated lamb lungs investigators recently found that NO reduces both small arterial and venous vascular resistance induced by hypoxia and U-46619, a thromboxane A₂ analog, whereas sodium nitroprusside also dilated larger veins (28). Likewise, effective arterial dilation has been demonstrated in isolated rat lungs following endothelin-induced vasoconstriction (29). In the latter study, relaxation of pulmonary veins was relatively greater in lungs perfused with Krebs solution than in blood-perfused lungs, probably because NO is rapidly inactivated by hemoglobin (15). However, in contrast to these animal experiments, a recent investigation in patients with acute lung injury showed that 3 ppm NO exerts its predominant vasodilating effect on the pulmonary veins (30). If NO lowers the microvascular pressure, it may reduce fluid filtration in humans with acute lung injury, as was recently suggested by Frostell (31). Our findings in sheep support this suggestion, but in addition to a reduction in microvascular pressure, NO may influence other factors affecting lung microvascular filtration, such as permeability and surface area. We cannot exclude the possibility, albeit it is unlikely, that decreased surface area, due to reduced inflow pressure to recruitable vessels, contributed to the observed reduction of L during NO inhalation (27). It is interesting that L did not rise to that of the control group precisely after cessation of NO, and that Ppa also returned slowly to pre-NO levels. The latter observations support the hypothesis that inhaled NO exerts some protection against lung injury lasting far beyond its effective half-life of only a few seconds (15). We did not follow values beyond 4 h after endotoxin, except for an additional lymph sample at 4.5 h.

Previous studies in unanesthetized sheep have shown that the endotoxin-induced increases in L, CL, and extravascular fluid volume are caused by a combination of both increased capillary pressure and permeability (1, 27). In the present work, late-phase L/P protein ratio increased slightly in both groups and became significantly elevated above baseline in the NO group. This indicates that relatively more protein or relatively less water enters the lymph during NO inhalation. We did not observe that plasma concentration decreased; in contrast, it increased in concert with a slight but not significant elevation of the hemoglobin concentration (Table 5). The high CL suggests severe damage to the integrity of lung microvasculature. In other vascular beds, beneficial NO effects have been interpreted as due to improvement in microvascular permeability, perhaps by radical scavenging or micropore tightening effects (16) or by inhibiting leukocyte (32, 33) and platelet adhesion to the endothelium (34, 35). Our findings support such an NO-related reduction in permeability, although the mechanisms are unknown as yet.

There is growing evidence that inhibition of NO production leads to increased leukocyte adhesion and that formation of leukocyte aggregates are attenuated by NO donors (32, 36). We observed indirect evidence for a similar inhibitory effect of NO on leukocyte adhesion, based on the higher leukocyte counts at 3.5 and 4 h in the NO group. This change in leukocyte count may be partly responsible for the reduction in lung microvascular permeability in this model. It is therefore conceivable that NO given early in endotoxemia might exert a protective effect (36).

The changes in gas exchange subsequent to injection of endotoxin have been described in several earlier studies using the same model (4, 37, 38). Improvement of gas exchange during administration of NO has been reported in animal studies of acute lung injury (2, 39), and also recently in studies of adult respiratory distress syndrome (ARDS) in man (11). In an investigation of NO effects in an ovine model of ARDS, produced by lung lavage (39), NO 80-120 ppm for 10 min improved Pa_O2 and significantly reduced the increase in S/T. Improvements in gas exchange during the late phase have also been noted with NO 10 ppm in pigs exposed to E. coli endotoxin (2). However, improvement in gas exchange during inhalation of NO has not been unanimously accepted. Thus, workers studying lung damage to streptococci in piglets, employing the multiple inert gas elimination technique, found no decrease in intrapulmonary shunt during NO inhalation (1). We determined venous admixture at FI_O2 0.25, which does not allow any conclusion to be drawn about the presence of "true" shunt. Nevertheless, by comparing the results immediately before and after inhalation of NO, Pa_O2, AaPO₂, and venous admixture (calculated as S/T), all showed significant improvement. Furthermore, Sa_O2 and Sv_O2 improved slightly, albeit not significantly (Table 4), whereas oxygen delivery and oxygen consumption remained unaffected by inhaled NO (Table 5). Thus, the present experiments confirm evidence from studies in animals and even in patients with ARDS that NO improves gas exchange in endotoxin-injured lungs.

In summary, inhalation of nitric oxide significantly reduces lung microvascular filtration in sheep with endotoxin-induced lung injury. This reduction is associated with reduced pulmonary arterial pressure at a high L/P ratio. It appears that inhaled NO protects the lung, both by reducing microvascular pressure via dilation of upstream, as well as downstream peri- alveolar vessels and by improving lung microvascular permeability. Whether NO protects against edema formation by a similar mechanism in human sepsis is not yet known.