INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inhaled nitric oxide (NO) and prone position (PP) are two of the new therapeutic strategies proposed for acute respiratory distress syndrome (ARDS) patients. However, the beneficial effects of PP on arterial oxygenation were described by Bryan (1) more than 20 years ago. These findings were subsequently confirmed by other investigators (2). The mechanisms that induce an increase in PO₂ remain controversial. In early ARDS, lung injury is characterized by pulmonary edema at all lung levels, and regional FRC decays from nondependent to dependent lung regions (8). Inhaled NO has been shown to be beneficial in patients with ARDS by increasing PO₂ and reducing pulmonary arterial pressure (9). Due to its selective pulmonary vasodilatory effects, NO improves the ventilation-perfusion relationship by directing pulmonary blood flow toward better ventilated lung regions (9, 10). The prone position may help to keep dependent lung areas better aerated because of its more negative associated pleural pressure. The improvement in oxygenation linked to PP could be enhanced when PP is combined with inhaled NO. Indeed, NO exerts its vasodilating properties by acting in noncollapsed alveoli. The aim of this prospective study was therefore to evaluate the hemodynamic and respiratory effects of inhaled NO and PP in patients with ARDS.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

During a 6-mo period, 20 patients presented ARDS diagnosed on or after admission to the polyvalent ICU of Sainte-Marguerite University Hospital in Marseille, France. Four of these 20 patients presenting ARDS were not included in the study. Two of the four patients presented an arteriopathy contraindicating femoral artery catheterization. As for the two other patients, their families refused the protocol. Two more patients were rapidly excluded because of a worsening in their hemodynamic status (hypokinetic septic shock), which meant that we could not complete the protocol. Therefore, 14 patients (11 male, 3 female; age 45 ± 14 yr; body weight 73 ± 11 kg; body height 172 ± 6 cm) were prospectively investigated early in the course of their respiratory disease after written informed consent was obtained from each patient's next of kin. The study was approved by our Ethics Committee (Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale de Marseille) and supported by l'Assistance Publique Hôpitaux de Marseille. Acute respiratory distress syndrome was defined according to the recommendations of the American-European Consensus Conference (11). Among the 14 patients enrolled in the study, five were admitted to the ICU after multiple trauma, two were admitted with postoperative complications following major surgery, and seven were admitted for an acute medical illness (Table 1). ARDS was related to lung contusion (four patients), nosocomial bronchopneumonia (four patients), community-acquired pneumonia (three patients), aspiration pneumonia (two patients), or acute pancreatitis (one patient). One patient presenting a status asthmaticus on admission (Patient 4) developed a severe life-threatening nosocomial pneumonia (while receiving corticosteroids) 15 d after admission to the ICU and 11 d after the beginning of mechanical ventilation. Moreover, at that time his bronchospasm had been resolved as demonstrated by peak pressure that had been lower than 30 cm H₂O (29 cm H₂O). On admission, mean SAPS II score was 38 ± 13 and mean APACHE III score was 73 ± 27. The duration of mechanical ventilation preceding the study was 6 ± 5 d. All patients were tracheostomized, sedated, and paralyzed with a continuous infusion of sufentanil midazolam, and vecuronium bromide, and the lungs were ventilated using conventional volume-controlled mechanical ventilation (Puritan Bennett 7200 series, Carlsbad, CA).

Instrumentation and Measurements

Hemodynamic parameters. All patients had a radial artery catheter (Seldicath, Plastimed, Saint Leu la Forêt, France) and a pulmonary artery catheter equipped with a fast response thermistor (model 93 A-434H-7.5F; Baxter Health Care Corporation, Irvine, CA), which was inserted percutaneously through the right jugular or the left axillar vein and positioned so that the distal port was in the pulmonary artery and the proximal port in the right atrium, just above the tricuspid valve.

Systolic arterial pressure, diastolic arterial pressure, systolic pulmonary arterial pressure, diastolic pulmonary arterial pressure, pulmonary artery occluded pressure (PAOP), and right atrial pressure (RAP) were measured at end-expiration. The supine zero reference level was the mid-axilla. Cardiac output (CO) was measured by thermodilution using three to five boluses of 10-ml glucose solution between 6° and 10° C, injected via a closed system (Co-set; Baxter Health Care Corporation) at end-inspiration to improve the reproducibility of the measurement and also to minimize the influence of changes in intrathoracic pressure. Injection temperature was measured by a thermistor located at the proximal port of the right atrial lumen. A mean of three to five measurements is reported. Cardiac index (CI), oxygen delivery index (DO₂I), oxygen consumption index (O₂I), oxygen extraction ratio (OER), right (RVSWI) and left ventricular stroke work indices (LVSWI), and true pulmonary shunt (S/T) were calculated using standard formulas. Systemic vascular resistance (SVRI) and pulmonary vascular resistance (PVRI) were calculated using the following standard formulas: SVRI = (mean arterial pressure  RAP) × 79.9/ CI; PVRI = (mean pulmonary arterial pressure  PAOP) × 79.9/CI.

A 3-F fiberoptic thermistor catheter (3-F FT Pulsiocath PV 2023; Pulsion Medizintechnik, München, Germany) was advanced via the right or the left femoral artery into the descending aorta and connected to the integrated fiberoptic monitoring system. A 15-ml bolus of 5% glucose containing 1 mg/ml of indocyanine green dye (ICG-Pulsion; Pulsion Medizintechnik) at 0° C to 2° C was injected into the right atrium with a temperature-controlled syringe. Injection was started at the end of inspiration. The thermodilution curves for dye and temperature were recorded simultaneously in the aorta with the thermistor-tipped fiberoptic catheter. A computer (system Cold Z-021; Pulsion Medizintechnik) determined the mean transit time for the thermal indicator and for the dye indicator and calculated intrathoracic blood volume index (ITBVI) and pulmonary blood volume (PBV) (12). ITBVI is obtained from the product of the thermodilution cardiac index and the mean transit time of the dye. Pulmonary blood volume was also determined from the exponential decay time for the indicators. Mean thermal dye dilution was calculated from two to three successive measurements.

On inclusion, seven patients were receiving cardiovasoactive drugs: norepinephrine (four patients); dobutamine (one patient); norepinephrine and dobutamine (one patient); and epinephrine (one patient). The infusion rate was kept constant throughout the study period. Fluid support was limited to 30 to 35 ml/kg/d of glucose 10%.

Blood gas analysis. Systemic and pulmonary arterial blood samples were simultaneously withdrawn within 3 min before the measurement of CO. Arterial pH, Po₂, mixed venous O₂ pressure (PO₂), and PCO₂ were measured using a blood gas analyzer (278-blood gas system; Ciba Corning, Medfield, MA). Hemoglobin concentration, arterial and mixed venous oxygen saturations (SO₂ and SO₂ ), and methemoglobin levels were measured using a calibrated hemoximeter (270-CO-oxymeter; Ciba Corning).

Respiratory parameters. The following respiratory parameters were recorded: exhaled tidal volume VT, peak inspiratory pressure, mean inspiratory pressure, and respiratory rate. Respiratory dynamic compliance was calculated as: (peak inspiratory pressure  positive end-expiratory pressure)/VT.

Prone Position

All patients were treated on special beds with low pressure support (Flexicair MC3; SSI Le Couviour, Montpellier, France). Change of position was manually performed by three nurses and two staff members. In the PP, the arms were positioned parallel to the body. Attention was paid to avoid eye damage or any nonphysiologic movement of the limbs during posture changes.

Nitric Oxide Administration

Nitric oxide was released from a tank containing nitric oxide in nitrogen at a concentration of 450 ppm (Air Liquide, Meudon, France) and was delivered within the inspiratory limb of the ventilator just after the Cascade II humidifier via a flowmeter delivering flows within a range of 1-999 ml/min (Taema, Antony, France). Intratracheal gas was sampled using continuous aspiration through the endotracheal tube, permitting continuous determination of inspiratory, expiratory, and mean concentrations of NO and NO₂ using a fast response time chemiluminescence apparatus (NOX 4000; Sérès, Aix-en-Provence, France) as previously described (15). For each patient, VT, respiratory rate, and fraction of inspired oxygen (FI_O2) were adjusted to maintain minute ventilation constant throughout the study period. To detect changes in FI_O2 induced by inhalation of NO, FI_O2 was monitored continuously using an O₂ analyzer (NOX 4000, Sérès).

Procedure

The protocol consisted of seven consecutive phases. Baseline measurements (T0) were made after 1 h of steady-state conventional mechanical ventilation in supine position (SP). After this period, NO (5 ppm for patients not receiving norepinephrine, 10 ppm for patients receiving norepinephrine or epinephrine) was administered during a 1-h period. The difference in concentration of inhaled NO was related to the results of a previous study that showed that the maximum dose effect was observed at 10 ppm for patients receiving norepinephrine while patients not receiving norepinephrine presented a maximum effect for a concentration of 2 to 5 ppm of NO (16). Measurements were made just before stopping NO (T1). A second baseline (T2) was recorded just prior to turning the patients prone, 1 h after NO discontinuation. The patients were then turned to PP, and measurements repeated at 5 h (T3). After these measurements, NO was initiated and measurements were performed after a 1-h period (T4). The patients were then turned to SP (while maintaining NO). After 6 h of SP and NO, measurements were taken (T5). The patients were then turned to PP for 6 h and the last measurements were made (T6) while patients were still receiving NO. The total duration of the protocol was approximately 21 h.

Volume-controlled mechanical ventilation settings (except adjustments to keep constant minute ventilation) as well as vasoactive agents and fluid administration rates remained constant during the study.

A response to NO inhalation and/or PP was defined by an increase of the PO₂/FI_O2 ratio by at least 20% when compared with T0.

Statistical Methods

All the statistics were performed by an experienced statistician (XT). Data are expressed as the mean ± SD. Statistical calculations were performed using the standard analysis software package (SAS Institute) and the SPSS 6.1 package (SPSS Inc., Chicago, IL). Data collected at T1, T3, and T4 were compared with the first baseline (T0). Statistically significant differences were analyzed by parametric or nonparametric tests when required. General factorial analysis of variance was performed to analyze the different times. When normal distribution was present, comparison between two times was performed by Student's t test for paired samples. When nonparametric distribution was observed, we performed Wilcoxon signed-rank test. When a correlation was calculated, Pearson's coefficient of variation was used. When distribution was not normal, Spearman's rank correlation was used. A p value of less than 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

The severity of ARDS was assessed by a Lung Injury Score (LIS) greater than 2.5 in all cases (mean 3.04 ± 0.43). Computed tomography (CT) scan was performed in 10 of the 14 patients included in the study in the 12-h period preceding the beginning of the protocol. They showed diffuse bilateral infiltrates, predominantly in dependent lung regions. The four patients who did not have CT scan were those who presented the most severe respiratory failure. On inclusion, mean positive end-expiratory pressure (PEEP) was 10.5 ± 1.8 cm H₂O with a mean FI_O2 of 0.74 ± 0.19. Mean tidal volume was 530 ± 86 ml. Inhaled NO was given to the patients during a period of 12.9 ± 8.8 d (range 5 to 34 d) while prone positioning was used for 14.7 ± 11 d (range 4 to 42 d). The total duration of mechanical ventilation was 35.2 ± 27.6 d (range 6 to 93 d). Of the 14 patients who completed the study, two subsequently died. Two of the four patients not included in the study died. When the two patients secondarily excluded from the analysis because of early death and the two latter patients were considered, mortality was 30%. The other patients survived and were discharged from the hospital.

Evolution of PO₂/FI_O2

In the supine position, NO inhalation (T1) induced a significant increase of PO₂/FI_O2 when compared with baseline (Table 2). Eight patients (57%) were considered as responders to NO administration (increase of PO₂/FI_O2 of at least 20% when compared with T0). Two other patients presented an increase of PO₂/FI_O2) of 19%. When PO₂/FI_O2 was evaluated just before turning the patients prone (T2, second baseline), it was not different from the first baseline (T0) (128 ± 44 mm Hg versus 128 ± 46 mm Hg). Prone position (T3) resulted in a significant increase of PO₂/FI_O2 when compared with T0. Nine patients (64%) were considered responders to PP when compared with baseline measurements (T0). There was no difference between the improvement in arterial oxygenation caused by NO (180 ± 75 mm Hg) and that due to PP (193 ± 83 mm Hg). The association of NO with PP (T4) resulted in a dramatic and significant improvement in PO₂/FI_O2 when compared with either T0, T1, and T3 (Figure 1). Moreover, we performed a general factorial analysis of variance, taking into account three different factors: NO, PP, and the association of NO and PP. General factorial analysis of variance showed a significant effect of both NO and PP on PO₂/FI_O2 (p < 0.000) but was unable to demonstrate a synergistic effect. We looked for time-related variability in various models. We were unable to demonstrate a significant time-related change, which suggests that the improvement of oxygenation was linked to NO and PP rather than being spontaneous. Thirteen patients (93%) were considered as responders with the association of NO and PP.

View larger version (28K): [in this window] [in a new window]   Figure 1.   Evolution of PO2/FIO2 during the study period: T0: supine position; T1: supine position + NO; T2: supine position; T3: prone position; T4: prone position + NO. Values are expressed as the percent of change compared with T0. Each line represents one patient.

Evolution of Hemodynamic Parameters

Nitric oxide produced a significant decrease of mean pulmonary arterial pressure in both supine and prone positions (Table 3). Pulmonary vascular resistances exhibited a significant decrease when NO was added to patients in PP (T4). The decrease of both mean pulmonary arterial pressure (MPAP) and PVRI induced by inhaled NO was confirmed by an analysis of variance (p < 0.01 and p < 0.05, respectively). We found that PO₂/FI_O2 was highly correlated (p < 0.001) with both MPAP, PVRI, and S/T. Prone position induced a slight but significant increase in heart rate, probably related to an insufficient level of sedation, whereas the increase in mean arterial pressure was not significant. General factorial analysis of variance showed an effect of both NO (p < 0.01) and PP (p < 0.01) on S/T. A sharp decrease in S/T was observed by Student's t test when PP and NO were combined (T4) when compared with T0, T1, and T3. The lack of change in oxygen delivery can probably be explained by the high level of Sa_O2 at baseline (96.7 ± 2.6), which could explain the very small changes in oxygen content. Cardiac index, ITBVI, and PBVI remained stable throughout the study period.

Evolution of Respiratory Parameters

Except for a slight but significant decrease in PCO₂ induced by NO (while maintaining minute ventilation constant) and a significant increase in peak inspiratory airway pressure induced by PP, there were only minor changes noted during the study period (Table 4).

Evolution of Respiratory and Hemodynamic Parameters after the First Episode of Prone Positioning

It is interesting to point out that no residual effect on PO₂/FI_O2 was observed 6 h after returning the patients to the supine position (T5), when compared with T4, while maintaining NO (Table 5). Indeed, we observed a significant drop in arterial oxygenation from 261 ± 98 mm Hg to 190 ± 70 mm Hg (p < 0.003). Moreover, the second episode of PP (T6) did not induce a greater increase in PO₂/FI_O2 than the first period of PP (T4) (262 ± 81 mm Hg and 261 ± 98 mm Hg, respectively).

Returning patients from PP to SP (while maintaining NO) resulted in a significant increase in S/T (p < 0.05). The second period of PP associated with NO (T6) induced a decrease in S/T (p < 0.05) and a slight but significant increase in PAOP (p < 0.05) when compared with T5. When the two periods of PP associated with NO were compared, no significant differences were found between the various parameters studied. We did not observe a significant variation of the various parameters when the 2 periods during which patients were receiving NO while positioned in SP were compared (T1 and T5).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study shows that PP and inhaled NO presented additive effects on oxygenation. No clinically relevant hemodynamic or respiratory deleterious effects related to these treatments were noted in the present work. We did not observe a persistent beneficial effect on oxygenation when patients were returned to SP. The second episode of PP was not characterized by a better improvement in oxygenation than the first positioning.

Our findings confirm previous reports of gas exchange improvements in patients with ARDS when turned from the supine to the prone position. Consistent with previous reports (5, 6, 17), approximately 65 to 75% of the patients responded to PP with an improvement in oxygenation. To our knowledge, no study has validated the "best" duration of prone positioning. This point was emphasized by Chatte and colleagues (6) in their study. In the Fridrich and coworkers study (7), patients were ventilated in PP for 20 h per day. We chose 6 h of PP because of our experience in clinical practice and because no study has convinced us to choose one duration rather than another. It was not possible to have a third baseline at the end of the protocol because of the study design. Indeed, the duration of the study was approximately 21 h. Respiratory status would have been changed compared with the first baseline (T0). The effects of NO (T1), PP (T3), and the combination of the two treatments (T4) were compared with T0 (duration approximately 9 h), while the potential persistent effects of PP and NO were evaluated over a 12-h period (this comparison does not necessitate a baseline). Moreover, we had no idea of the precise duration of SP after PP in order to be sure that the residual effect of PP would disappear.

Our results differed from those of Chatte and colleagues (6), who found a persistent improvement of the PO₂/FI_O2 when the patients were returned from PP to SP. In the present study, PO₂/FI_O2 was not significantly different between T1 (SP, inhaled NO) and T5 (second supine position, inhaled NO). In the study by Chatte and colleagues (6), 13 of their 32 patients (40%) were considered as persistent responders. Using the same criteria as Chatte and colleagues (6), we found that in the present study, only four patients (29%) had a persistent improvement in PO₂/FI_O2 when returned to SP. The lack of persistent improvement compared with other studies could be explained in different ways. From a methodologic point of view, in the study by Chatte and colleagues (6), arterial blood gas analysis was performed 1 h after returning the patients to SP whereas, in our study, such an evaluation was made after 6 h in SP. If the physiopathologic basis of the action of PP is true, repositioning patients in SP restores the gravitational ventro-dorsal gradient. Therefore, after a few hours, it was not surprising to observe a decrease in oxygenation due to the decrease in ventilation of the posterior part of the lungs. Moreover, Chatte and colleagues (6) reported that some of the patients who respond to PP go back to initial PO₂/FI_O2 values when returned to SP. They have classified these patients as nonpersistent responders.

Although the initial reports of prone position-induced improvements in PO₂ did not explore the mechanisms by which the improvement could occur, a number of potential explanations has been suggested. Possible mechanisms include blood flow redistribution, ventilation redistribution, hydrostatic pressure changes, and increased FRC. Concerning FRC, Albert and coworkers (18) found substantial increases in PO₂ in dogs with acute lung injury when they were turned from SP to PP in absence of an increase in FRC. However, a recent study by Pelosi and associates (19) suggested that while the PP did not significantly alter either lung or chest wall mechanics, it markedly improved FRC (and oxygenation). It was also suggested (18, 20) that improvement in PO₂ is not related to alterations in regional diaphragmatic motion. In the present study, we did not observe any change in respiratory mechanics. Chatte and colleagues (6) noted that the prone position did not modify peak pressure in the controlled-volume mode. This fact could be explained by the type of bed used, limiting high abdominal pressure by favoring downward abdominal mass expansion. In a very recently published study, Fridrich and colleagues (7) did not report any change in compliance linked to PP. In other clinical studies devoted to ARDS and PP (4, 5, 17), no information was given about pressures or compliance.

Blood flow redistribution due to gravitational changes is one of the factors thought to be responsible for an improvement of oxygenation while patients are in PP by favoring dependent and intact lung regions. However, the heterogeneic distribution of lesions induced by ARDS should not allow for blood flow redistribution to healthy regions by gravitation alone. Blood flow also has been shown to be reduced in the atelectatic lung due to hypoxic pulmonary vasoconstriction, resulting in an increased pulmonary vascular resistance in this lung and thus diverting blood flow to the nonatelectatic lung (21). Glenny and coworkers (22) noted that because blood was preferentially directed toward the dorsal lung regions, regional perfusion had a distribution that followed the gravitational gradient when the animals were supine. In PP, this dorsally directed perfusion opposed the effects of gravity, allowing for a decreased gravitational perfusion gradient as described by certain authors (22, 23). However, using the multiple inert gas elimination technique, Pappert and colleagues (17) suggested that improvement in arterial oxygenation during PP resulted from a redistribution of blood flow away from unventilated areas to regions with normal ventilation-perfusion ratios. The effects of position on ventilation distribution have been studied by Mutoh and associates (24), who demonstrated in pigs that the gravitational pleural pressure gradient with the animals in PP was significantly lower when compared with that in SP. They also found that after creating pulmonary edema by volume infusion, pleural pressure in the dependent lung region became positive when the animals were supine, but was much less positive when they were turned prone. The less positive pleural pressure in dependent (i.e., ventral) regions that results from turning prone, along with the more uniform pleural pressure gradient that exists in this position, could cause much less of the lung to be below closing volume. The reasons for the larger vertical gradient in pleural pressure in SP than in PP are still very speculative. It may be due to the position of the heart relative to the lung, the mismatch between the natural shapes of the lung and the chest wall, the interaction between the lung and abdominal contents, or gravity acting on the lung, heart, chest wall, and/or abdominal contents.

The favorable effects of inhaled NO on oxygenation have been well established. The present study shows that the improvement of PO₂/FI_O2 related to the administration of inhaled NO was similar to the improvement in oxygenation related to PP.

The lack of synergistic effect between inhaled NO and PP could be explained by the absence of effect of PP on FRC. Indeed, an increase in FRC should theoretically result in an increase in lung areas reached by inhaled NO.

In the present study, PP and inhaled NO induced a comparable improvement in oxygenation. They presented additive effects on oxygenation. It is important to note that outcome was not an endpoint of our study and that only a well-designed comparative study would show a decrease in mortality due to the use of the association of PP to inhaled NO.