Additive Effect of Nitric Oxide Inhalation on the Oxygenation Benefit of the Prone Position in the Adult Respiratory Distress Syndrome 

The study protocol was approved by our institutional ethical committee, and written informed consent was obtained from each patient's next of kin. We prospectively studied 47 patients (30 male, 17 female) who were 37 +/− 16 yr old (mean +/− SD). Patients were enrolled in the study if they had a diagnosis of new onset of severe ARDS (according to the European American Consensus Conference criteria) confirmed by computed tomography of the lung that revealed patchy and lobar infiltrates, a pulmonary artery occlusion pressure less than 18 mmHg, a Murray lung injury score [15]more than 2.5, and in whom ARDS was the cause of primary organ failure. During a 2-yr period, 47 patients fulfilled the entrance criteria. The mean lung injury score score was 3.1 +/− 1, the APACHE II score was 17 +/− 6, and the mean total intensive care unit duration of stay was 28 +/− 19 days. The cause of ARDS was intrapulmonary in 33 cases (16 lung contusion, 11 pneumonia, 3 aspiration, 2 bronchiolitis obliterans obstructive pneumonia, and 1 near drowning) and extrapulmonary in 14 cases (7 polytrauma, 3 pancreatitis, 2 massive transfusion, and 2 sepsis).

All patients required mechanical ventilation (Evita 2; Drager, Lubeck, Germany) in a pressure-controlled mode. The ventilatory settings were respiratory rate, 18 +/− 4 breaths/min; tidal volume, 8.1 +/− 1.5 ml/kg; inspiration: expiration ratio, 1.1 +/− 0.7; peak inspiratory pressure, <or= to 35 cm water; positive end-expiratory pressure, 10 to 15 cm water; and FI (^O)(2), 0.75 to 0.95 to attain an arterial oxygen saturation rate >or= to 90%. These settings were not altered for the duration of the study. All patients were deeply sedated with midazolam and sufentanil, and muscle relaxation was provided with vecuronium. Patients were monitored continuously by electrocardiography, peripheral oxygen saturation (Sp^O(2); CareVue 9000; Hewlett-Packard, Boblingen, Germany) and indwelling peripheral arterial catheter. A 7.5-French fiberoptic pulmonary artery catheter (REF-Sat; Baxter, Irvine, CA) was placed via the internal jugular vein.

Cardiotonic (dobutamine) or vasoactive (dopamine, norepinephrine) drugs (or both) were used during the initial phase of stabilization to maintain mean arterial pressure more than 70 mmHg and cardiac index more than 3.51 [middle dot] min^-1[middle dot] m^-2, and dosages were not changed during the study itself. Directly measured hemodynamic indices included heart rate, mean arterial pressure, mean pulmonary artery pressure, and pulmonary artery occlusion pressure. Cardiac output was measured by the thermodilution technique and expressed as the mean of four injections of 10 ml saline at 5 [degree sign]C. Right ventricular ejection fraction was derived from the REF-Sat catheter. Systemic and pulmonary vascular resistance indices were calculated according to standard formulas. Arterial (Pa^O(2)) and mixed venous (Pv^O(2)) oxygen tensions were measured using standard blood gas electrodes (AVL 995-Hb; AVL, Graz, Austria), and the total hemoglobin concentration, hemoglobin oxygen saturation, and methemoglobin levels were measured using spectrophotometry (AVL 912; AVL). The ratio of the partial pressure of arterial oxygen (Pa^O(2)) and the FI^O(2) was used as a parameter for arterial oxygenation because the admixture of nitrix oxide/N²gas decreases the actual FI^O(2). FI^O(2) was kept constant with the aid of an oxygen analyzer (Oxydig^R, Draeger, Lubeck, Germany) placed immediately proximal to the patient's endotracheal tube. Hemodynamic variables (pulmonary and systemic vascular resistance indices, right ventricular ejection fraction, cardiac index) and the venous admixture were calculated using a cardiac output computer (Explorer; Baxter) and a pressure monitoring kit (Baxter) according to standard formulas.

Delivery and analysis of nitric oxide was provided by a newly developed system (Pulmonox; Messer-Griesheim, Gumpoldskirchen, Austria). [16]The device electronically controls the flow of nitric oxide to decrease the concentration from 900 ppm in the tank to the concentration desired at the airway. It adapts to the gas flow of the mechanical ventilator to ensure nitric oxide delivery during the inspiratory phase of the respiratory cycle. Continuous analysis of the concentration in parts per million of nitric oxide and nitric oxide degradation products was performed using the chemoluminescence detection method (ECO Physics, Durnten, Switzerland). Gas samples were withdrawn a short distance proximal to the patient's endotracheal tube, ensuring that a length of 50 cm was maintained between nitric oxide instillation and withdrawal of gas for analysis.

(Figure 1) summarizes the study design. Measurements and calculations were performed with the patient in the supine position 6 h after initial stabilization (T1), after 1 h inhalation of 10 ppm nitric oxide in the supine position (T2), after 1 h with nitric oxide off in the supine position (T3), after 2 h in the prone position (T4), and after 1 h inhalation of 10 ppm nitric oxide in the prone position (T5). A patient was defined as a responder if the Pa^O(2):FI^O(2) increased by 20% from baseline.

Values are expressed as the mean +/− SD. In the case of the Pa^O(2):FI^O(2) ratio, the first median quartile (Q1) and the third median quartile (Q3) are used for description because of the skewed distribution of the ratio. Treatment effects are reported as the mean of the values in the respective groups. Analysis of variance (randomized block design)[17]was used to assess the independent effect of nitric oxide inhalation and positioning (supine or prone) on oxygenation and pulmonary hemodynamics. Because of the skewed distribution, log-transformed values of the Pa^O(2):FI^O(2) ratio are used for analysis. A P value < 0.05 was considered significant.

Of 47 patients, 4 died of multiple-organ failure while in the intensive care unit.

(Table 1) shows the measurements made during the study. The inhalation of 10 ppm nitric oxide in the supine position induced a significant change in several parameters compared with control (T2 vs. T1): the venous admixture decreased by 16%, the Pa^O(2):FI^O(2) ratio increased by 37%, the mean pulmonary artery pressure decreased by 15%, the pulmonary vascular resistance index decreased by 16%, and the right ventricular ejection fraction increased by 20%. After termination of nitric oxide inhalation, all these parameters returned toward baseline values (T3 vs. T1). Prone positioning (T4 vs. T3) caused a significant 35% increase in the Pa^O(2):FI^O(2) ratio. However, mean pulmonary artery pressure, pulmonary vascular resistance, right ventricular ejection fraction, and venous admixture did not change significantly. The inhalation of 10 ppm nitric oxide in the prone position yielded significant changes in several parameters compared with the prone position without nitric oxide inhalation (T5 vs. T4). Venous admixture decreased by 21%, mean pulmonary artery pressure decreased by 16%, pulmonary vascular resistance index decreased by 15%, and right ventricular ejection fraction increased by 19%. In addition, the Pa^O(2):FI^O(2) ratio increased by 40% compared with that achieved by the prone position alone. Systemic hemodynamic indices (heart rate, mean arterial pressure, cardiac index, and systemic vascular resistance index) did not change at any of the study time points. Methemoglobin levels were measured before the end of the procedure and never exceeded 0.8%.

As a first step, we performed within-group analyses of variance, which are tests of the overall therapeutic effect of nitric oxide inhalation. We found a significant effect of nitric oxide on the Pa^O(2):FI^O(2) ratio, independent of body positioning. As a second step, we tested for an interaction between nitric oxide and positioning. No statistical significance was found related to this interaction. We interpreted these results to indicate that nitric oxide effect is independent of supine or prone positioning, and that the significant difference in the Pa^O(2):FI^O(2) ratio at T5 vs. T4 represents an additive effect of nitric oxide and positioning (Table 2).

Gattinoni et al. [18]were the first to observe that lung density on computed tomographic images in supine patients with acute respiratory failure is frequently maximal in the dorsal (i.e., dependent) zone. When patients were turned from the supine to the prone position, the lung density ultimately redistributed from the dorsal to the ventral component. This suggested that the improvement in oxygenation in ARDS observed with assumption of the prone position results from decompression of dependent lung regions previously compressed by overlying lung structures, which then improves ventilation-perfusion matching. However, a beneficial effect is not observed in all cases, [11]and it has been speculated that the true benefit of the prone position is that it creates a transpulmonary pressure sufficient to exceed the airway opening pressure in dorsal lung regions. [12]In ARDS, the prone position induces redistribution of lung density, regardless of whether patients respond with improvement arterial oxygenation. [11,18]Thus, it remains speculative whether the beneficial effect on arterial oxygenation results from redistribution of blood flow to nonatelectatic lung regions, recruitment of previously atelectatic lung areas, or more efficient bronchial drainage. [18] 

Regardless of its mechanism, it appears to be appropriate to turn patients who have ARDS to the prone position and observe their responses. If improvement in oxygenation does occur-and in most cases it will within 30 min of the time when the prone position is assumed [19]-it allows a decrease in high levels of FI^O(2), airway pressure, or both. As also described by Friedrich et al., [20]most of the improvement in oxygenation occurs within the first hour after the patient is turned to the prone position.

The inhalation of nitric oxide in ARDS appears to exert its beneficial effects on oxygenation because it is preferentially delivered to areas of better ventilation. There it induces pulmonary vasodilation, enhances ventilation-perfusion matching, and decreases the magnitude of the intrapulmonary shunt. [13]It is rapidly inactivated by hemoglobin, so that its actions are confined to the pulmonary circulation, where it decreases the elevated mean pulmonary artery pressure [21]and right ventricular afterload. [14]These actions may also decrease elevated microvascular filtration pressure [22]and thereby curb accumulation of extravascular lung water. [23]We hypothesized that when used together these two therapies would have an additive effect, by combining the benefit of redistribution of previously atelectatic areas (prone position) with enhancement of ventilation-perfusion matching (nitric oxide). Responders to inhaled nitric oxide are recognized within minutes of its application. Therefore, we designed the study to allow more than enough time (1 h after nitric oxide exposure-cessation, 2 h after prone positioning) for stabilization at each setting.

Administration of 10 ppm nitric oxide to our patients in the supine position induced a significant improvement in oxygenation and pulmonary hemodynamics (Table 1). The prone position by itself provided a similar improvement in oxygenation, although pulmonary hemodynamics were relatively unaffected. When nitric oxide was inhaled in the prone position, a further significant improvement in oxygenation followed, with a 21% decrease in venous admixture and a significant increase in the Pa^O(2):FI^O(2) ratio (Table 2). In addition the beneficial effects of nitric oxide on pulmonary vascular resistance and mean pulmonary artery pressure were realized. We suggested that the observed improvement in oxygenation and pulmonary hemodynamics from T3 (supine, nitric oxide off) to T5 (prone, with nitric oxide inhalation) represents a clinical scenario in which the benefit of prone positioning is enhanced by the addition of inhaled nitric oxide.

The notion of combining two or more interventions in ARDS is not new and certainly reflects current clinical practice. One such approach is to combine inhaled nitric oxide with a vasoconstrictor. Almitrine bismesylate, a selective pulmonary arterial vasoconstrictor, has been shown to have an additive effect with inhaled nitric oxide on arterial oxygenation in ARDS. [24] 

Jolliet et al. [19]found that combining nitric oxide, prone positioning, and almitrine bismesylate improved oxygenation in a small group of patients. Approximately 60% of our patients responded to prone positioning, compared with 85% of patients in our study. Their lower response rate may be explained by the finding that, in their group of 12 patients, 8 died and most patients were in a state of severe immunosuppression (e.g., they had received bone marrow transplants or had Pneumocystis carinii pneumonia). Chatte et al., [25]in a series of 32 consecutive patients, observed a positive response to prone positioning (Pa^O(2):FI^O(2) ratio increased by > 20) in 78% of their patients. This is comparable to the response rate in our population, although we used a threshold of a 20% increase in the Pa^O(2):FI^O(2) ratio as an index of response.

Our study shows that in patients with severe isolated ARDS the combination of nitric oxide inhalation with the prone position offers a greater improvement in oxygenation than either treatment used alone. Nitric oxide also provides a favorable effect on pulmonary hemodynamics, which is not accomplished by using the prone position alone. Improved oxygenation can allow a decrease in high FI^O(2) and airway pressures, thus potentially reducing the risk of oxygen toxicity and barotrauma. Pulmonary vasodilation induced by nitric oxide facilitates permissive hypercapnia by minimizing its effect on pulmonary vascular resistance. Furthermore, the combination of the prone position with inhaled nitric oxide could help to decrease the necessary dose of nitric oxide, with its potential for the accumulation of toxic proinflammatory degradation products, such as nitrogen dioxide. [26-28] 

The major limitation of this study lies in its relatively short duration, therefore long-term conclusions about outcomes cannot be drawn. Indeed, there is no conclusive evidence that either inhaled nitric oxide or the prone position decrease the morbidity and mortality rates associated with ARDS. Although the prone position appears to be a relatively simple intervention, it necessitates skilled nursing and vigilance to prevent “disconnects” or pressure injury. Nonetheless, Albert [29]suggests that “complications severe enough to raise concerns about employing [it] occur infrequently.” A second limitation is that the effect of the time course of the disease could not be tested. Although disease progression or improvement is unlikely to have been a factor during the relatively short time course of this study and is supported by the finding that there were no significant differences in parameters between T1 (supine control) and T3 (supine, 2 h later), it cannot be completely discounted.

Statistical analysis of our data revealed a significant effect of nitric oxide on the Pa^O(2):FI^O(2) ratio, independent of body placement in the supine or prone position. This suggests that the significant increase in Pa^O(2):FI^O(2) from T4 (prone position without nitric oxide inhalation) to T5 (prone position with nitric oxide inhalation) represents an additive effect of nitric oxide on the enhancement of oxygenation by the prone position in patients with ARDS.

In patients with severe ARDS, clinicians should consider combining nitric oxide inhalation with the prone position, even if an improvement in oxygenation has already been achieved by using one of these measures used alone. Our data further emphasize the notion that it is unlikely that a single treatment modality will cost-effectively alter outcome in ARDS. Rather, it appears to be prudent for us to develop coherent treatment algorithms for ARDS in which different interventions are added in a methodical and rational sequence. Future investigations of the management of severe ARDS should evaluate such combined treatment algorithms and monotherapy.