INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inhaled nitric oxide (NO) is a potent and selective pulmonary vasodilator (1) that has attracted considerable clinical interest in recent years. In normal lungs, NO reverses hypoxic pulmonary vasoconstriction, but it has a minimal effect on gas exchange (3). However, in patients with lung disease the reported effects have been variable. For example, in studies of patients with adult respiratory distress syndrome (ARDS) inhaled NO uniformly reduced intrapulmonary shunting by about 5% and improved oxygenation (5). In contrast to patients with ARDS, patients with chronic obstructive pulmonary disease (COPD) exhibited marked intersubject variability in response to NO inhalation, and there was no consensus as to the effect of inhaled NO. Various studies of patients with COPD have reported small increases in the partial pressure of oxygen in arterial blood (Pa_O2) (9), no effect on Pa_O2 (10), and decreased Pa_O2 with increased A/ inequality (11).

In the patient with ARDS, the reduction of intrapulmonary shunting and subsequent improvement in arterial P O₂ is expected since the underlying pathophysiology of the disease would preclude the distribution of inhaled NO to most if not all abnormal areas of the lung. In ARDS, the predominant gas exchange abnormality is intrapulmonary shunting (i.e., perfusion of areas devoid of ventilation) with a much smaller (and sometimes no) contribution from A/ inequality (12). Thus, the distribution of NO is limited to well-ventilated areas of the lung, and consequently these areas are vasodilated, blood flow from unventilated regions is redirected to them, and gas exchange improves. The underlying physiologic mechanism of the effect of NO on gas exchange in COPD is uncertain. The vasodilatory effect of NO is dependent on delivery of NO to the alveolus and subsequently to the pulmonary vascular smooth muscle. In conditions such as COPD where areas of low ventilation-perfusion are often present (13) without a substantial intrapulmonary shunt, the balance between (1) the preexisting vascular tone in the well-ventilated and poorly ventilated regions and (2) the amount and thus effect of inhaled NO reaching these different regions via ventilation may determine the net effect on gas exchange. Depending on the overall net effect of these influences, perfusion of the low A/ regions could fall or rise and overall gas exchange would accordingly improve or worsen.

We therefore hypothesized that inhaled NO has an effect on pulmonary gas exchange consistent with the vasodilatory effect on these areas that receive the gas. Specifically, as pre viously concluded (5), in conditions where intrapulmonary shunt is the predominant gas exchange abnormality, inhaled NO will improve pulmonary gas exchange and Pa_O2 by reducing blood flow to the areas of shunt and increasing blood flow to normal areas of the lung. In lungs with A/ inequality and areas of low A/ ratio, inhaled NO will ( 1) worsen Pa_O2 and pulmonary gas exchange if blood flow to areas of low A/ ratio is increased and (2) improve Pa_O2 and pulmonary gas exchange if blood flow to areas of low ventilation-perfusion ratio is reduced. This will vary according to the balance between the relative potential for vasodilation (of normal and low A/ areas) and the relative concentrations of NO reaching the different types of unit.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This experiment was approved by the Animal Subjects Committee of the University of California, San Diego. Fourteen adult mongrel dogs of either sex weighing 23.2 ± 0.9 kg (mean ± SE) were studied. Anesthesia was introduced intravenously with pentobarbital sodium 30 ml/ kg, and the animals were monitored for depth of anesthesia every 15 min by observing movement, heart rate, and blood pressure response to painful stimulus. Maintenance doses of 50 to 100 mg of the drug diluted in saline were given as required to maintain the animal in an unresponsive state. This level of anesthesia was sufficient to abolish spontaneous ventilation. The animals were intubated with a cuffed endotracheal tube (ID, 6 to 8 mm) and ventilated with a volume-cycled ventilator (Harvard 613; Harvard Apparatus Co., South Natick, MA) at a rate of 14 breaths/min with 5 cm H₂O of positive end-expiratory pressure (PEEP). To further prevent nonspecific atelectasis, animals were sighed with three tidal breaths at 1-h intervals. Airway pressure was monitored using a pressure transducer connected to the respiratory circuit. Tidal volume (mean ± SE, 17 ± 1 ml/kg) was adjusted to maintain arterial PCO₂ ~ 30 mm Hg (mean ± SE, 31 ± 0.4 mm Hg) and pH between 7.3 and 7.4 (mean ± SE, 7.35 ± 0.002). A temperature probe was placed in the esophagus to monitor body temperature. Body temperature was maintained above 37° C by the use of heating pads.

Surgical Preparation

The right femoral artery was cannulated for arterial blood gas sampling, and the left femoral vein was cannulated for infusion of an inert gas solution (see below). A No. 5F Swan Ganz catheter (Baxter Healthcare Corp., Irvine, CA) was inserted into the left external jugular vein and advanced into the pulmonary artery using direct-pressure monitoring. It was used for sampling of mixed venous blood and measurement of pulmonary arterial and pulmonary arterial wedge pressures. A tracheotomy was performed above the level of the endotracheal tube cuff to allow creation of one of the two types of endobronchial obstruction (see below), producing either a shunt or an area of low A/ ratio.

Lung Lesions

Partial and complete endobronchial obstructions were created sequentially in each animal in a balanced order. We attempted to create both lung lesions in all animals, but because of the difficulty in obtaining stable patterns of gas exchange with the lung lesions, not all animals were studied under all three conditions ( see RESULTS). Partial endobronchial obstruction (creating a primary gas exchange abnormality of A/ inequality [low A/ lesion]) was created using No. 00 to 3 rubber laboratory stopper through which a blunt 14 to 18-gauge needle allowed a small amount of ventilation. The cuffed endotracheal tube was transiently withdrawn past the site of the tracheotomy and, using a flexible guide wire, the stopper was manipulated past the carina and into a lobar bronchus. The lumen of the needle was protected from obstruction by secretions by attaching to it a piece of plastic tubing and clearing the lumen with puffs of air after the stopper was in place. Then the tubing and flexible wire were removed and the endotracheal tube was replaced beyond the tracheotomy site. Partial obstruction was confirmed by noting appearance of a low A/ mode receiving more than 5% of total blood flow determined by the multiple inert gas analysis ( see below), without appreciable amounts of intrapulmonary shunt (< 2% of pulmonary blood flow). After hemodynamic, blood gas, and inert gas measurements, made before, during, and after the administration of inhaled NO, the stopper was retrieved by traction on an attached cord.

After 10 min of preventilation with 100% O₂, complete endobronchial obstruction (intrapulmonary shunt condition) was next obtained in a similar fashion, using an intact stopper that did not allow distal ventilation. Complete obstruction was confirmed by the creation of a greater than 15% intrapulmonary shunt with less than 2% blood flow to areas of low A/ ratio. After all measurements were obtained before, during, and after inhaled NO ( see below), the stopper was retrieved, and the collapsed area of the lung was reinflated by the administration of 15 cm PEEP for 6 to 10 min. After each such experimental series, blood gas and inert gas measurements were made to confirm the restoration of normal pulmonary gas exchange.

Study Protocol

A graphic representation of the study design is shown in Figure 1. It can be seen that each measurement made during NO administration in the presence of the desired induced lung lesion was bracketed by similar measurements without inhaled NO to ensure stability of the lesion and reversibility of NO effects. These measurements were, in turn, bracketed by additional measurements before and after the induced lesion to ensure the reversibility of the induced lesions and the stability of the animal's gas exchange over the course of the experiments. Each set of measurements included mean arterial, pulmonary arterial, and pulmonary arterial wedge pressure measurements and sampling of pulmonary mixed venous blood, arterial blood, and mixed expired gases for the multiple inert gas analyses, blood gas determinations, cardiac output calculations, and metabolic measurements.

View larger version (46K): [in this window] [in a new window]   Figure 1.   Graphic representation of the study design. Condition 1 was either the lesion of low ventilation-perfusion ratio or the shunt, in random order. Condition 2 was the lesion not induced in Condition 1. At each time point pulmonary arterial pressure, pulmonary arterial wedge pressure, mean arterial pressure, ventilation, oxygen consumption, and carbon dioxide production were measured. The concentration of inhaled NO was measured when applicable. Blood gas and inert gas measurements were also made, and cardiac output was calculated.

Inert Gas Measurements

Ventilation-perfusion distributions were obtained using the multiple inert gas elimination technique in the usual fashion (14, 15). A mixture of six inert gases (SF₆, ethane, cyclopropane, enflurane, diethyl ether, and acetone) were dissolved in normal saline and infused via the left femoral vein (rate in ml/min = 0.25 × E in L/min). Duplicate 5-ml blood samples for the inert gas analysis were collected in heparinized glass syringes from the pulmonary artery and right femoral artery, and 30-ml mixed expired gas samples were collected from a heated mixing chamber into gas-tight glass syringes.

Solubilities, retentions (R, equal to the ratio of arterial to mixed venous partial pressure), and excretions (E, equal to the ratio of mixed expired to mixed venous partial pressure) for the inert gases were determined using gas chromatography (HP-5890; Hewlett-Packard, Wilmington, DE) (14), and ventilation-perfusion distributions were calculated from the inert gas data (15). Using the multiple inert gas elimination technique, compartmental ventilation or blood flow to areas of different ventilation-perfusion ratios can be calculated and graphically represented. Because the ventilation-perfusion ratios of interest span several decades, a log scale is used for these data, and the second moment of the perfusion distribution, exclusive of intrapulmonary shunt (logSD), and the second moment of the ventilation distribution, exclusive of dead space (logSD), are used as indicators of the degree of ventilation-perfusion inequality (i.e., the greater the logSD or the logSD the greater the ventilation-perfusion inequality). The residual sum of squares (RSS) was used as an indicator of the adequacy of fit of the data to the 50-compartment model of the lung (15). Once the ventilation-perfusion distribution had been obtained, the respiratory gas exchange compatible with the recovered distribution was calculated assuming end-capillary diffusion equilibrium. The mixed arterial PO₂ expected from the ventilation-perfusion distribution, Pa_O2(p), was calculated from the compartmental end-capillary values (16). When the measured Pa_O2 is lower than that predicted from the inert gas exchange, this is suggestive of pulmonary diffusion limitation.

Hemodynamic Measurements

The pressure transducers (Statham P23 ID; Statham Instruments, Oxnard, CA) were zeroed to the level of the right atrium, and calibration was checked prior to each measurement. Mean arterial, pulmonary arterial, and pulmonary artery wedge pressures were recorded on a strip chart recorder (Model 200; Gould, Valleyview, OH) immediately before each set of inert gas measurements. Cardiac output was calculated from the mixed venous arterial blood and mixed expired inert gas concentrations using the Fick principle, and the total pulmonary vascular resistance was calculated. Pulmonary vascular resistance of the abnormal areas of the lung created by the induced lung lesion was calculated assuming that the pulmonary arterial and pulmonary artery wedge pressures were not different from those of the lung as a whole and blood flow perfusing the poorly ventilated or unventilated regions obtained from the inert gas analyses.

Blood Gas and Metabolic Measurements

Arterial and mixed venous samples (2 ml each) were collected immediately after the collection of each inert gas arterial and mixed venous blood sample and maintained on ice until analyzed for PO₂, PCO₂, and pH using an IL1306 blood gas analyzer (Instrumentation Laboratories, Lexington, MA). Each sample had hemoglobin and oxygen saturation measured using an IL 282 Co-oximeter (Instrumentation Laboratories), and the hematocrit was determined. Ventilation (E, measured using a calibrated Wright respirometer) and the mixed expired concentrations of O₂ and CO₂, (mass spectrometer 1100; Perkin-Elmer, Pomona, CA) were measured during each inert gas sample collection period, and oxygen consumption and carbon dioxide production were calculated.

NO Administration

NO (1,000 ppm) in nitrogen was obtained and diluted with room air and 100% O₂ to achieve 80 ppm in 21% O₂. The gas mixture was passed through a soda lime absorber prior to administration to remove NO₂ and added at the gas inlet of the ventilator. NO concentration was monitored just proximal to the endotracheal tube, using the Perkin-Elmer mass spectrometer on the mass 30 setting. Inhaled NO concentration and FI_O2 was determined prior to each administration.

Statistical Analyses

Repeated-measures analysis of variance with preplanned contrasts (Super ANOVA 1.11; Abacus Concepts Inc., Berkeley, CA) was used to statistically test changes in the dependent variables with NO administration. Regression analysis was used to determine the relationship between the change in pulmonary vascular resistance, logSD, and Pa_O2 in the lungs with ventilation-perfusion inequality. Significance was accepted at p < 0.05, two-tailed.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We studied animals under the following conditions: (1) normal (n = 8; Dogs 2, 3, 4, 6, 7, 8, 9, and 10), (2) with intrapulmonary shunt but no A/ inequality (n = 9; Dogs 1, 2, 3, 5, 6, 7, 8, 10, and 11), and (3) with A/ inequality (n = 8; Dogs 1, 5, 6, 9, 11, 12, 13, and 14) but no intrapulmonary shunt. Representative examples of each condition are given in Figure 2. All results are given as mean ± SE. Animal preparations were stable over the course of the experiment, and all of the lung lesions were fully reversible; there were no significant changes in ventilation-perfusion inequality, intrapulmonary shunting, airway pressure, mean arterial pressure, pulmonary arterial and pulmonary arterial wedge pressures, or measured Pa_O2 comparing data from immediately before creation of and after the removal of the partial or complete endobronchial obstruction. There was a small but significant (p < 0.05) fall in Pa_CO2 of 2 mm Hg comparing data before the lung lesion and after resolution of the lung lesion.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Representative ventilation-perfusion distributions obtained sequentially from one animal. (Top panel ) Normal distribution with 0.4% shunt, no areas of low ventilation-perfusion ratio, and a logSD of 0.33. ( Middle panel ) Shunt distribution with 19.9% shunt, no areas of low ventilation-perfusion ratio, and a normal logSD of 0.46. ( Bottom panel ) Low ventilation-perfusion lesion with 0.3% shunt, 15.7% of blood flow to areas of low ventilation-perfusion ratio, and a markedly increased logSD of 1.28.

Metabolic and Hemodynamic Data

O₂, CO₂, cardiac output, and mean arterial pressure data for normal lungs and each of the lung lesions are given in Table 1. The creation of the lung lesions did not cause a significant change in O₂, CO₂, cardiac output, or mean arterial pressure. Averaged over all conditions, there was no significant change in cardiac output with inhaled NO. Averaged over all conditions, there was a small but highly significant (p  < 0.0001) reduction in mean arterial pressure, resulting in a fall from 140 ± 3 mm Hg before NO inhalation to 136 ± 3 mm Hg during NO inhalation. Averaged over all conditions, there was a significant (p < 0.0001) reduction in pulmonary arterial pressure, from 19.5 ± 0.6 mm Hg before NO inhalation to 17.4  ± 0.5 mm Hg during NO inhalation, accompanied by a fall in pulmonary vascular resistance, from 370 ± 20 dyn-s/cm⁵ to 334 ± 14 dyn-s/cm⁵ (p < 0.01) without a significant change in pulmonary arterial wedge pressure. There were no changes in airway pressure with the inhalation of NO.

Blood Gas and Inert Gas Data

Inert gas data for each of the experimental conditions are given in Tables 1 and 2. Averaged over all the data sets, the mean RSS was 2.5, and 91% were less than 5.48, the 50th percentile expectations of the sum of squares for six gases, indicating excellent technical quality of the data and good fit of the data to the inert gas model. Prior to NO inhalation, normal lungs (Figure 2) had a logSD of 0.44, indicating minimal A/ inequality, normal for anesthetized dogs, with no appreciable areas of low A/ ratio or areas of intrapulmonary shunting. Pa _O2 was 99 ± 3 mm Hg, and Pa_CO2 was 30 ± 1 mm Hg. There was no significant difference between the Pa_O2 predicted from the recovered A/ distribution, Pa _O2(p), and the measured Pa_O2, indicating absence of pulmonary diffusion limitation. There were no significant effects of inhaled NO on the arterial blood gases, logSD, or the minimal areas of low A/ ratio and shunt.

The lungs containing the intrapulmonary shunt (Figure 2) had a logSD of 0.48 (not significantly different from control values), indicating normal A/ inequality. Recall that logSD is a parameter unaffected by shunt. Again, there was no evidence of pulmonary diffusion limitation as measured Pa_O2 and Pa_O2(p) were not significantly different. Less than 1% of cardiac output perfused areas of low A/ ratio, but almost 25% of cardiac output perfused areas of the lung devoid of ventilation. This shunt was sufficient to reduce the measured Pa _O2 from 106 ± 2 to 68 ± 3 mm Hg. Inhaled NO resulted in a small but highly significant (p < 0.005) reduction in shunt, from 24.8 ± 1.9 to 20.7 ± 2.3% without change in A/ inequality or blood flow to areas of low A/ ratio. The effect of the reduction in intrapulmonary shunt was a significant (p < 0.005), if small, increase in the measured Pa_O2, from 68 to 74 mm Hg.

By contrast, before NO inhalation, the lungs with the low A/ lesions ( see Figure 2) had a logSD of 1.53 ± 0.12, which is markedly abnormal. Approximately 13% of the cardiac output in these lungs perfused areas of the lungs with a A/ ratio of less than 0.1, a A/ ratio that conservatively estimates the lower limit of normal. Less than 2% of blood flow was to areas of intrapulmonary shunt. This was reflected in the arterial P O₂, which was reduced from 107 ± 2 mm Hg, prior to the creation of the lung lesion to 76 ± 3 mm Hg with partial endobronchial obstruction. There was no evidence of pulmonary diffusion limitation as measured Pa_O2 was not significantly different from Pa_O2(p).

For all eight animals with the low A/ lesion taken together, inhaled NO did not result in a significant change in A/ inequality, blood flow to areas of low A/ ratio, intrapulmonary shunt, or Pa _O2. There was considerable interanimal variability in the response to inhaled NO (see Figure 3). With inhaled NO, five animals had a deterioration in gas exchange (Group 1), with mean Pa_O2 falling from 79 to 67 mm Hg, and three of the animals had a beneficial response (Group 2), with an increase in mean Pa_O2 from 72 to 79 mm Hg. There were no differences between these two groups of responders for any of the metabolic or hemodynamic data; however, blood flow was reduced to areas of low ventilation-perfusion ratio in all animals, with an increase in Pa_O2 accompanied by a decrease in logSD from 1.78 to 1.57. Conversely, those animals that had a fall in Pa_O2 had an increase in blood flow to the areas of low ventilation-perfusion ratio and a corresponding increase in logSD from 1.38 to 1.63 (interaction statistic Pa_O2 response by NO administration, p < 0.01). There was also a significant (R = 0.85, p < 0.01) negative relationship between the change in pulmonary vascular resistance of the abnormal area of the lung with NO inhalation and the change in logSD (Figure 4). The change in Pa_O2 was also significantly positively related to the change in pulmonary vascular resistance of the abnormal area of the lung (R = 0.78; p < 0.05) and negatively related to the change in logSD (R = 0.97; p < 0.0001) (Figure 4).

View larger version (25K): [in this window] [in a new window]   Figure 3.   Effect of 80 ppm inhaled NO in animals with areas of low ventilation-perfusion ratio. Open circles = Group 1 animals (n  = 5), which had a fall in PaO2 in response to inhaled nitric oxide. Closed circles = Group 2 animals (n = 3), which had an increase in PaO2 in response to inhaled NO. For each dependent variable there was a significant group by NO administration interaction: Blood flow to areas of low ventilation-perfusion ratio p < 0.01, LogSD, p < 0.05; PaO2, p < 0.005. For comparison, data for all animals together (closed squares, n = 8) are shown.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Relationship between the change in LogSD, the change in pulmonary vascular resistance, and the change in PaO2 in response to inhaled NO in the animals with areas of low ventilation-perfusion ratio.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has demonstrated in a "two-compartment" dog lung that the effect of 80 ppm inhaled NO on pulmonary gas exchange can be predicted by understanding the underlying gas exchange abnormality. There are several advantages of this model of abnormal pulmonary gas exchange. First, we were able to create areas of either shunting or low ventilation-perfusion ratio so that the effects of inhaled NO on pulmonary gas exchange could be examined without influence from more than one pattern of lung lesion or underlying pulmonary disease at any time. Second, all lung lesions were noninjurious to the lung and completely reversible, allowing the effect of different patterns of gas exchange to be examined in the same animals. Third, the lung lesions were stable over time, and therefore any changes that occurred during inhalation of NO were entirely due to the effect of the gas on the lung lesion and were not confounded by the continued deterioration of gas exchange often seen when agents such as oleic acid are used to create abnormal lungs. Finally, the lung abnormalities were mechanically induced rather than pharmacologically induced, eliminating any confounding influence of such drugs on the response to inhaled NO.

In our model, when the underlying gas exchange problem was one of intrapulmonary shunting, inhaled nitric oxide uniformly reduced the pulmonary vascular resistance of the normal, ventilated areas of the lung, resulting in increased blood supply to well-ventilated areas of the lung and a consequent improvement in Pa_O2. The abnormal areas of the lung in this instance were not vasodilated because they were not ventilated, and therefore the NO was presumably not delivered to the pulmonary vasculature of these areas.

In lungs where areas of low ventilation-perfusion ratio are present, the situation is more complex. In our study, in the lungs with ventilation-perfusion inequality, the mean logSD was 1.53 before NO inhalation, and 13% of the pulmonary blood flow was to areas of the lung with a ventilation-perfusion ratio of less than 0.1. Inhaled NO improved pulmonary gas exchange in three animals and worsened it in five animals. In the animals where gas exchange was worsened, inhaled NO resulted in a reduction in the pulmonary vascular resistance of both the normal and the abnormal areas of the lung. The result was an increase in blood flow to the areas of low A/ ratio, an increase in ventilation-perfusion inequality, and a worsening of gas exchange. However, when the pulmonary vascular resistance fell only in the normal areas of the lung and not in the areas of the low ventilation ratio, the blood flow to the areas of low A/ ratio was reduced and was shifted towards normal areas of the lung. This resulted in a reduction in ventilation-perfusion inequality and an improvement in pulmonary gas exchange. This is consistent with a complex competition between vasoconstriction secondary to local hypoxia and vasodilation resulting from delivery of different amounts of NO to areas of low and normal ventilation perfusion ratio.

Clinical Implications

In this study it was our intent to create and examine the effects of inhaled NO on specific patterns of gas exchange. We believe that this analysis can be extended to patients with lung disease with known gas exchange abnormalities. For example, in patients with severe ARDS, the pattern of the gas exchange lesion is of lung units filled with fluid and/or debris, resulting in marked intrapulmonary shunting with some normal lung units interspersed (12). Although areas of low ventilation-perfusion ratio may be present in ARDS, they are not the predominant underlying problem. Our model of intrapulmonary shunt confirms the results of the clinical literature. It has been consistently shown in patients with ARDS that inhaled NO ranging from 10 to 80 ppm results in a reduction of pulmonary arterial pressure, redistribution of blood flow to areas of normal pulmonary gas exchange accompanied by a reduction in intrapulmonary shunting, and an improvement in Pa _O2 (5).

In patients with chronic obstructive pulmonary disease (COPD) ventilation-perfusion inequality is the most common gas exchange abnormality, and marked intrapulmonary shunting is absent (13). Therefore, from the present study, we would expect a variable alteration in pulmonary gas exchange in these patients, depending on the amount of local hypoxic pulmonary vasoconstriction present and the amount of inhaled NO delivered to both normal and abnormal areas of the lung. Again this is supported by the reported effects of inhaled NO in patients with COPD. Adnot and coworkers (9) have reported that 40 ppm of inhaled NO increased Pa_O2 in their patients with COPD. However, examination of the individual data reveals that seven of their subjects increased their Pa_O2, whereas four had a reduction in Pa_O2. Moinard and colleagues (10) reported no effect of 15 ppm inhaled NO in patients with COPD, but again (as we found in the present study) grouped data may not reveal the entire pattern of response as nine subjects worsened their Pa_O2, whereas five subjects improved. Recently, Barbera and coworkers (11) reported a worsening of gas exchange in patients with COPD, with 10 patients experiencing a fall in Pa_O2 and only three having an improvement in Pa_O2 in response to 40 ppm of inhaled NO. Thus, we believe that in lung disease when areas of low A/ ratio are present, the variable clinical effects of inhaled NO in this patient population are not random and can be explained by an understanding of the basic physiologic mechanisms as discussed in this report.

Effect of Inhaled Nitric Oxide on Lung Lesion Stability

It is important to recognize that our model of lungs with ventilation-perfusion inequality likely represents the best expected effect of this concentration of inhaled NO on pulmonary gas exchange, as lungs with ventilation-perfusion inequality were only accepted for study if they were stable over the course of data collection. Seven lungs (not used in the present analyses) with ventilation-perfusion inequality were unstable over the course of the experiment. There was a progressive decrease in perfusion of areas of low ventilation-perfusion ratio and an increase in intrapulmonary shunting, whereas the overall amount of blood flow to abnormal areas of the lung remained relatively constant. There are three possible explanations for this change in the pattern of gas exchange. First, the small needle allowing ventilation could have become obstructed by mucus or debris, causing the area to become unventilated. Second, the amount of ventilation delivered could have been insufficient to maintain adequate ventilation of these lung units over time, leading to progressive collapse (17). Finally, the inhaled NO may have, by increasing blood flow to the areas of low ventilation-perfusion ratio, allowed the A/ ratio to fall below the critical level for stability (17) and result in atelectasis. This critical A/ ratio is determined by the net flux of O₂, CO₂, and N₂ from alveolar gas to capillary blood and is about 0.001 breathing room air (17). Thus if a tenfold increase in blood flow were to affect a lung unit with an initial A/ ratio of 0.01, instability would be likely . Although this increase seems extreme since we have no way of knowing which of these possibilities actually occurred, we cannot rule out the possibility that the progressive increase in shunt was due to the effects of inhaled NO. If this were the case, this scenario would be undesirable in patients, as these lung units might be difficult to reexpand after the administration of inhaled NO.

Dose of Inhaled Nitric Oxide

On the basis of our model, in patients with ventilation-perfusion inequality as the major gas exchange abnormality, the minimum possible dose of inhaled NO required to vasodilate normal areas of the lung would be most likely to improve pulmonary gas exchange since the distribution of the gas to areas of low ventilation-perfusion ratio would be proportionally less. We chose 80 ppm of inhaled NO in order to maximize the effect of the gas on pulmonary arterial pressure and gas exchange, consistent with the dosage of inhaled NO used in other animal studies investigating the effect of inhaled NO on gas exchange in lung injury (18). Some investigators have reported a beneficial effect of inhaled NO in concentrations ranging from 2 ppm (22) to 60 parts per billion on gas exchange in patients with ARDS (6, 7). This improvement in gas exchange can occur even in the absence of the detectable change in pulmonary arterial pressure (6). To our knowledge, the effect of these low concentrations has not been investigated in patients with COPD. Therefore, it is possible that inhaled NO concentrations of less than 10 ppm may have improved gas exchange in more of the lungs with ventilation-perfusion inequality in our study.

Systemic Effects of Inhaled Nitric Oxide

We found a small but significant drop in mean arterial pressure with the administration of 80 ppm inhaled NO. This has been reported in the pig (18), and it is possibly related to reactions of the inhaled NO with protein thiols (23), forming relatively unstable S-nitrosothiols and releasing NO into the systemic circulation (24) under some circumstances. Although this effect of inhaled NO has not been reported in humans, it is another reason to use the minimum effective concentration of inhaled NO.

In conclusion, in a fully reversible animal model of pulmonary gas exchange abnormalities we found, consistent with the current literature, that inhaled NO improves pulmonary gas exchange in lungs with intrapulmonary shunt as the sole lesion by increasing blood flow to normal areas of the lung. In lungs with areas of low ventilation-perfusion ratio, animals had variable responses. Inhaled NO improved gas exchange when the normal areas of the lung increased blood flow but worsened it when both the normal areas and the areas of low ventilation-perfusion ratio had a fall in pulmonary vascular resistance. This variability underscores the need to examine more than average group responses to NO and it may explain the variable clinical response to inhaled NO of patients with COPD, a disease that is characterized by marked ventilation-perfusion inequality.