INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In patients with chronic obstructive pulmonary disease (COPD), exercise is often accompanied by an abnormal increase in pulmonary artery pressure (Ppa). Different studies have identified a number of mechanisms, including hypoxic vasoconstriction, reduction of the capillary bed by emphysema, vascular remodeling, and extramural compression by increased alveolar pressure (1) that may combine and contribute to the development of pulmonary hypertension during exercise. More recently, it has been recognized that endothelium plays a pivotal role in modulating both systemic and pulmonary vascular tone (5), and that endothelial cells act as major mediators in the dilator response to increases in blood flow (6, 7). Nitric oxide (NO) has been identified as the most powerful endothelium-derived relaxing agent (8). Inhibition of NO synthesis with analogues of L-arginine has been shown to reduce exercise-induced vasodilation in the human forearm (9) and to increase the slope of the pressure-flow relationship in an isolated lung model (10), thus suggesting that NO may play an important role in the adaptation of pulmonary circulation to the increase in blood flow induced by exercise.

Impaired NO-mediated endothelium-dependent pulmonary artery relaxation has been demonstrated in lungs from patients with end-stage COPD (11), probably as a consequence of a reduced expression of endothelial NO synthase (NOS) (12). To what extent a defective release of NO by pulmonary arteries could contribute to exercise-induced pulmonary hypertension in COPD is unknown. In this regard, the exogenous supply of NO might act as a substitute for the endogenous one, thus reversing or mitigating the increase in Ppa that takes place during exercise in these patients.

We have previously shown that in COPD patients studied at rest (13), inhaled NO exerts a selective pulmonary vasodilator effect, but that at the same time it may worsen hypoxemia because it also inhibits the hypoxic regulation of the matching between ventilation and perfusion (A/), in a manner similar to that of systemic vasodilators (1, 14). The administration of systemic vasodilators during exercise attenuates the increase in Ppa; however, this effect is usually accompanied by an impairment of A/ matching (1). Since on exertion the ventilation of patients with COPD may be distributed more homogeneously with respect to the A/ ratio (15), we hypothesized that a vasodilator driven by ventilation, such as NO, could exert a different effect on A/ relationships during exercise than at rest. Accordingly, we undertook the present study to examine the effects of selective pulmonary vasodilation with inhaled NO during submaximal exercise on pulmonary hemodynamics and gas exchange in a group of patients with advanced COPD.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Nine male patients who fulfilled the clinical and functional criteria for the diagnosis of COPD (16) were studied. The patients' general characteristics and pulmonary function data are shown in Table 1. Reference values used are those of our own laboratory (17, 18). Overall, the patients had severe airflow obstruction, increased FRC, and moderate to severe reduction of D L_CO. Inhaled short-acting bronchodilators and oral theophylline were withdrawn 12 and 24 h before the study, respectively, and no patient was receiving vasodilator treatment. The study was approved by the Ethical Committee of the Hospital Clinic of the University of Barcelona, and written informed consent was obtained from each participant.

Procedures

A transvenous, balloon-tipped Swan-Ganz catheter (Edwards Laboratories, Santa Ana, CA) was placed into the pulmonary artery under pressure-wave monitoring (M1166A; Hewlett-Packard, Boeblingen, Germany), and a polyethylene catheter (Seldicath 3F; Plastimed, Saint Leu-la-Forêt, France) was inserted into the radial artery. Intravascular pressures were continuously monitored and registered (7754B; Hewlett Packard, Palo Alto, CA). The external zero reference level was positioned at the midchest. Two measurements were made under each study condition and the mean value was reported as the final result. Measurements of pulmonary pressure were taken at the end of expiration. Cardiac output was determined by the thermodilution technique (M1012A; Hewlett-Packard, Boeblingen, Germany), and was expressed as the mean of three measurements.

Minute ventilation (E) and respiratory rate were recorded on a minute-by-minute basis with a calibrated Wright spirometer (MK8; BOC-Medical, Essex, UK). Low dead space, low resistance, nonrebreathing valves were used to collect the expired gas through a heated mixing box during rest (Hans Rudolph, Kansas City, MO) and exercise (E. Jaeger, Würzburg, Germany). Oxygen uptake and CO₂ production were calculated from mixed expired O₂ and CO₂ concentrations (CPX System; Medical Graphics, St. Paul, MN). Arterial and mixed venous PO₂, PCO₂, and pH were analyzed in duplicate with standard electrodes (IL 1302; Instrumentation Laboratories, Milano, Italy). Hemoglobin and methemoglobin concentrations were measured with a cooximeter (IL 482; Instrumentation Laboratories). A/ distributions were estimated with the multiple inert gas elimination technique (19). Data are reported for only eight patients because a technical error in the chromatographic process precluded the proper analysis of A/ distributions for one patient. The position of both perfusion and ventilation distributions was described by the A/ ratio at their mean, and the dispersions of the two distributions on a logarithmic scale (logSD Q and logSD V, respectively) were used as indices of A/ mismatch (upper normal limit = 0.6) (19). As an overall descriptor of the combined dispersion of both blood flow and ventilation distributions, we used the difference among measured retentions and excretions of the inert gases corrected for the elimination of acetone (DISP R  E*; normal value < 3.0), which includes intrapulmonary shunt whenever this is present (20). Intrapulmonary shunt was defined as the fraction of cardiac output perfusing lung units with A/ ratios < 0.005, normal A/ units were defined as those with A/ ratios between 0.1 and 10, and dead space was defined as the ventilation to units with A/ ratios > 100.

Study Design

The greatest workload that each patient could tolerate (Wmax) was determined on a preceding day with a progressive exercise test. The test was done with the patient on a cycle ergometer (E. Jaeger), and the load was increased by 20 W every 2 min until spontaneous interruption of exercise. Mean results of the incremental exercise test are shown in Table 1.

Patients were studied in a semirecumbent position. Thirty minutes after starting the inert gas infusion, they were connected to the breathing circuit and inhaled either room air or 40 ppm of NO in air for 20 min. We chose a dosage of 40 ppm of NO in accordance with our previous experience (13). Measurements during rest were made when steady-state conditions were assessed. Following this, each subject was asked to cycle at 50% of his Wmax. After 3 min of cycling, a second set of hemodynamic and gas-exchange measurements was made. Patients were allowed to rest for a minimum of half an hour, during which time systemic and pulmonary hemodynamics, as well as arterial blood gas values, were checked in order to guarantee that they had returned to baseline levels. The patients were then reconnected to the breathing circuit, and inhaled the second gas mixture (air or NO). Twenty minutes later, resting and exercise measurements were repeated as before. The order of inhalation of room air and NO was established randomly.

System of NO Administration

NO was delivered through a nonrebreathing circuit. Particular features of this setup in our laboratory have been reported elsewhere (22). Inspired NO was a mixture of O₂, N₂, and NO, obtained with a set of calibrated rotameters (Tecfluid, Barcelona, Spain). NO was obtained from a stock tank containing 800 ppm of NO in N₂ (Abelló, Barcelona, Spain). We continuously monitored the inspired concentrations of NO and NO₂ with a chemiluminescence analyzer (CLD 700 AL; Eco Physics, Dürnten, Switzerland) at the inspiratory port. Additionally, the inspired O₂ concentration was continuously controlled with a zirconium analyzer (CPX System; Medical Graphics) and was kept at 21% during NO administration.

Statistical Analysis

Data are expressed as mean ± SE. All analyses were performed with version 6.01 of the SPSS statistics package (SPSS Inc., Chicago, IL). The effects of exercise and NO inhalation were assessed with a two-way repeated measures analysis of variance (ANOVA). Whenever an interaction between NO inhalation and exercise was found, the difference between rest and exercise with and without NO and air was compared, using a paired t test. All tests of hypotheses were two tailed at the p = 0.05 level of significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Exercise Test

Patients exercised for the same period and at the same workload while breathing room air or 40 ppm NO (39 ± 6 W under both conditions), according to the design of the study. Oxygen uptake at the end of the exercise test was similar under the two conditions (Table 2), averaging 62 ± 6% (room air) and 61 ± 7% (NO) of the predicted maxima (21). Furthermore, the values of heart rate, E, breathing frequency, CO₂ output, and respiratory quotient recorded during exercise were also equivalent with the two inspired gas mixtures (Tables 2 and 3).

Hemodynamic Measurements

At rest and during breathing of room air, mean Ppa was within normal limits (< 20 mm Hg) in all but two patients, and pulmonary artery occluded pressure (Ppao) was normal in all of them (Table 3). As expected, Ppa and Ppao increased significantly when patients exercised while breathing room air, to 38 ± 3 mm Hg and 19 ± 2 mm Hg, respectively (Table 3). The mean slope of the relationship between pulmonary vascular pressure gradient (Ppa  Ppao) and the increase in cardiac output from rest to exercise was 1.36 ± 0.21 mm Hg · L¹ · min (Figure 1), which is more than twice the normal value (0.53 mm Hg · L¹ · min) (23).

View larger version (19K): [in this window] [in a new window]   Figure 1.   Relationship between mean values (± SE) of pulmonary vascular pressure gradient (mean Ppa  Ppao) and cardiac output, at rest and during exercise, with breathing of room air (open symbols) and inhalation of 40 ppm NO (closed symbols). The pressure- flow relationship was shifted downward during NO inhalation (p = 0.03).

NO inhalation at rest decreased Ppa by 13 ± 6%, whereas Ppao, cardiac output, and systemic arterial pressure (Psa) remained essentially unchanged (Table 3). During exercise, NO inhalation resulted in a smaller increase in Ppa than that seen during breathing of air (Table 3). Further, the relationship between the pulmonary vascular pressure gradient and cardiac output during NO inhalation was shifted slightly downward compared with that for room air (Figure 1). The increase in Ppao seen during exercise was not modified by NO inhalation (Table 3).

Gas-exchange Measurements

At rest while breathing room air, patients showed moderate to severe hypoxemia (Pa_O2 range: 55 to 77 mm Hg), and all but one had normal Pa_CO2 values (Table 2). All patients exhibited a moderate to severe degree of A/ inequality, as shown by increased dispersions of both ventilation (logSD V) and blood flow (logSD Q) distributions (Table 2). The amount of intrapulmonary shunt was trivial in all patients (0.5 ± 0.1%).

During exercise while breathing room air, Pa_O2 decreased and both Pa_CO2 and (A-a)PO₂ increased (Table 2). Yet there was better A/ matching, as shown by a decrease in logSD V (Table 2), resembling previous observations made with the inert gas technique in COPD patients (1, 15).

At rest, NO inhalation decreased Pa_O2 and increased (A-a)PO₂ (Table 2). This effect resulted from the worsening of A/ distributions, as shown by the increase in the overall index of A/ mismatch (DISP R  E*), in the dispersion of blood flow distribution (logSD Q), and in the percentage of blood flow perfusing poorly or nonventilated lung units with low A/ ratios (Table 2). Pa _CO2 and mixed venous PO₂ were not altered by NO inhalation.

In contrast to the decrease in Pa_O2 shown on exertion during breathing of room air, no further change in Pa_O2 was observed when exercise was performed during NO administration (Table 2, Figure 2). Changes in (A-a)PO₂ showed a similar trend to those in Pa_O2 (Table 2), although the interaction between NO and exercise did not reach statistical significance (ANOVA; p = 0.09). Neither the increase in Pa_CO2 nor the decrease in mixed venous oxygen pressure (Pv_O2) induced by exercise was modified by NO inhalation (Table 2).

View larger version (14K): [in this window] [in a new window]   Figure 2.   Change in PaO2 from rest to exercise during breathing of room air and NO; *p < 0.05 compared with room air. Values are mean ± SE.

Exercise done during NO inhalation resulted in a more homogenous distribution of alveolar ventilation with respect to A/ ratio, similar to that shown during breathing of room air (Table 2). Moreover, contrasting with the differences shown at rest, the proportion of blood flow to alveolar units with low A/ ratios and that to units with normal A/ ratios was similar during exercise with the breathing of both room air and NO (Table 2). Therefore, NO administered during exercise contributed to the redistribution of blood flow from alveolar units with low A/ ratios (a decrease of 3.1 ± 1.1%) to units with normal ratios (an increase of a similar proportion) (Figure 3). NO inhalation did not modify the changes induced by exercise in the mean A/ ratio of both the ventilation and perfusion distributions (Table 2).

View larger version (20K): [in this window] [in a new window]   Figure 3.   Change from rest to exercise, during breathing of room air (open bars) and NO (hatched bars), in perfusion of alveolar units with low A/ ratios ( < 0.1, including shunt) (left panel ), and in perfusion of alveolar units with normal A/ ratios (between 0.1 and 10) ( right panel ). *p < 0.05 compared with room air. Values are mean ± SE.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of the present study show that in COPD patients, the inhalation of NO during exercise attenuates the increase in Ppa and, contrasting with the effects shown at rest, produces no further impairment of gas exchange.

Measurements made at rest confirm previous observations that inhaled NO may worsen gas exchange in COPD (13, 24), and extend these observations to patients with less advanced COPD and a slightly lower degree of A/ mismatch, and without pulmonary hypertension, as compared with the patients in our former study (13). We ascribe such effect of NO to the inhibition of hypoxic pulmonary vasoconstriction, according to what has been shown in healthy subjects (25). The magnitude of worsening of A/ distributions induced by NO inhalation was similar to that found after the administration of a systemic vasodilator (1). Interestingly, the present study demonstrates that the effect of inhaled NO on gas exchange was different during exercise than at rest. Whereas Pa _O2 decreased when exercise was performed with breathing of room air, no further change in Pa_O2 was shown during exercise with breathing of 40 ppm NO (Figure 2). This finding is at variance with the effects of systemic vasodilators, which in COPD also worsen gas exchange during exercise (1), and is likely to be due to a different effect of NO on A/ distributions during exercise than during rest.

Despite the improvement of A/ relationships during exercise in our study, Pa_O2 decreased when patients exercised while breathing room air. This is consistent with a small increase in alveolar ventilation secondary to the patients' airflow obstruction, which was insufficient to increase the mean A/ distribution ratio to a level that might have adequately counterbalanced the decrease in Pv_O2 (26). The mean ratio of ventilation to cardiac output during exercise in the current subset of patients was 1.3, whereas in healthy subjects this ratio may rise to 4.8 (27). Moreover, during exercise with breathing of room air, 3.4% of pulmonary blood flow was diverted to alveolar units with low A/ ratios, thus explaining why, in face of the decrease in Pv_O2, blood passing through these units remained poorly oxygenated, hence further decreasing arterial PO₂ (28). When patients exercised while inhaling NO, the degree of A/ mismatch was slightly greater than that shown during breathing of room air (DISP R  E*, 9 ± 1 versus 7 ± 1), yet the percentage of perfusion to alveolar units with low A/ ratios was similar (4 ± 2% versus 3 ± 2%). Since the reduction in Pv_O2 was equivalent under the two conditions, the resulting Pa_O2 was very close (Table 2). These combined effects are illustrated in Figure 4, which shows the relationship between Pa_O2 and Pv_O2 under the different study conditions. It is well established that for a given Pv_O2, the value of Pa_O2 depends on the degree of A/ mismatch (28). In our patients the value of Pv_O2 during NO inhalation was similar to that observed during breathing of room air, both at rest and during exercise. At rest, Pa_O2 decreased when NO was administered because A/ inequality increased. However, during exercise, the dispersion of A/ distributions improved with both inspired gas mixtures and became closer, thus resulting in similar values of Pa _O2 (Figure 4, Table 2).

View larger version (20K): [in this window] [in a new window]   Figure 4.   Plot of PaO2 as a function of PvO2 during breathing of room air (open symbols) and 40 ppm NO (closed symbols). Dashed lines indicate relationship between PaO2 and PvO2 calculated at different degrees of A/ mismatch (logSD Q) from a normal A/ distribution (logSD Q = 0.6) to very severe A/ mismatch (logSD Q = 1.8). At rest (high PvO2 values), PaO2 was lower during NO inhalation, owing to greater A/ mismatching. During exercise (low PvO2 values), the degree of A/ mismatch during breathing of both room air and NO was similar, and so was the resulting Pa O2. Values are mean ± SE.

The effect of NO administered during exercise on A/ relationships differed from that of systemic vasodilators, which usually increase A/ inequality (1). We hypothesize that this particular effect of NO was due to its delivery through inspired air concurrently with a more homogenous distribution of ventilation with respect to the A/ ratio during exercise. The latter facilitated the access of NO to those alveolar units with best A/ ratios, which received the greatest proportion of ventilation. It is well established that when breathing frequency increases in patients with heterogenous obstruction of the airways, ventilation is preferentially distributed to those units that impose less airflow resistance and have faster time constants (product of compliance and resistance) (29). Accordingly, during exercise, as the breathing frequency increases, more of the tidal volume (VT) is distributed to units with faster time constants, and less is distributed to those with slower time constants. This may explain why, when NO was added to the inspired air during exercise in our study, it was preferentially delivered to well ventilated alveoli with faster time constants, which are more efficient in terms of gas exchange, which was at variance with what happened during rest. Conceivably, the selective vasodilator action of NO in these well-ventilated alveoli facilitated the redistribution of blood flow from units with low A/ ratios to those with normal ratios. This hypothesis is consistent with previous observations made in mechanically ventilated patients with acute respiratory distress syndrome (ARDS), in whom the recruitment of alveolar units with either positive end-expiratory pressure (PEEP) or continuous positive airway pressure (CPAP) enhanced the effect of NO on gas exchange (30, 31). Accordingly, a potential clinical implication of our findings would be that targeting the delivery of NO specifically to those lung units that have faster time constants, by adjusting the ventilator settings or delivering NO at the beginning of inspiration (32), could result in an improvement in pulmonary gas exchange in patients with patchy airflow obstruction. Further studies are needed, however, to elucidate the extent to which this potential system for NO delivery might be beneficial in mechanically ventilated COPD patients, who usually do not respond to inhaled NO (33).

The results of the present study contribute to a better understanding of the role of hypoxic vasoconstriction in minimizing A/ mismatch in COPD. In a previous study (1), we showed that even after inhibition of hypoxic vasoconstriction with a systemically administered vasodilator (nifedipine), there was less A/ inequality during exercise than at rest. This suggested that the improvement in A/ relationships during exercise was essentially due to a more homogenous distribution of inspired air, rather than to a more efficient hypoxic vasoconstriction. The results of the present study, showing that inhaled NO did not exert any detrimental effect on gas exchange when it was administered by inhalation during exercise, are in agreement with this suggestion, and reinforce the concept that in COPD, the improvement in ventilation distribution is the major reason for the amelioration of A/ matching during exercise. Accordingly, we suggest that in COPD patients, hypoxic vasoconstriction plays an important role in modulating the matching between ventilation and perfusion at rest, whereas during exercise the ability to distribute inspired air more homogeneously is the major determinant of the improvement in A/ relationships.

The administration of NO shifted slightly downward the pressure-flow relationship of pulmonary circulation from that with room air (Figure 1). We ascribe this change to a decrease in pulmonary vascular tone induced by NO inhalation. The contribution to this change of additional factors that may modulate pulmonary vascular pressures (i.e., exercise workload, cardiac output, or ventilation) can be disregarded, since the patients in the present study were studied under similar conditions while breathing both room air and NO (Tables 2 and 3). However, the effect of NO on the exercise-induced increase in Ppa was modest, as shown by the fact that the slope of the pressure-flow relationship of pulmonary circulation remained about twice normal (23). Accordingly, the exogenous supply of NO during exercise to this group of patients did not significantly block the abnormal increase in Ppa, suggesting that the defective release of NO by pulmonary arteries in COPD (11) appears not to be a determinant factor of the pulmonary vascular hypertensive response induced by exercise. These findings would support the notion that in COPD, the increase in pulmonary vascular tone during exercise appears to be dictated primarily by mechanical rather than by chemical factors (34). To some extent, our findings concur with those reported by Koizumi and coworkers (35), who showed that NO inhalation in sheep did not enhance the reduction of pulmonary vascular resistance during exercise. Nevertheless, we cannot rule out the possibility that in patients with more advanced COPD and greater endothelial dysfunction, a reduced synthesis of NO by pulmonary endothelium might play a more important role in the abnormal increase in Ppa induced by exercise. Since in our population Ppa measured at rest was within the normal range in the majority of patients, it can be speculated that the stimulus for remodeling of pulmonary vessels was insufficient to alter the adaptation of the pulmonary circulation to exercise.

In summary, the results of the present study show that inhaled NO in COPD may exert a different effect on gas exchange at rest than during exercise. Whereas at rest NO inhalation reduced Pa_O2 due to the increase in A/ inequality that results from the inhibition of hypoxic vasoconstriction, no further decrease in Pa _O2 was seen with exercise during NO administration. This effect is at variance with that of systemic vasodilators, which decrease Pa_O2 during both rest and exercise, and is probably explained by a preferential distribution during exercise of NO to well-ventilated alveolar units with faster time constants. Furthermore, the administration of NO exerted a moderate effect on the exercise-induced increase of Ppa, hence suggesting that the defective release of NO by pulmonary endothelial cells appears to be of minor significance for the development of pulmonary hypertension during exercise in COPD.