Inhibition of Platelet Aggregation by Inhaled Nitric Oxide in Patients with Acute Respiratory Distress Syndrome

This study was approved by the Comite Consultatif de Protection des Personnes dans la Recherche Biomedicale of La Pitie-Salpetriere Hospital and supported by l'Assistance Publique-Hopitaux de Paris. Written informed consent was obtained from each patient's next of kin. Consecutive patients diagnosed with ARDS on or after admission to the surgical intensive care unit of La Pitie Hospital in Paris (Department of Anesthesiology) were included in this prospective study during an 8-month period. Only patients with severe ARDS responding to inhaled nitric oxide in terms of decreased pulmonary artery pressure and increased arterial oxygenation were enrolled in the study. Inclusion criteria were (1) lung injury severity score greater or equal to 2.2 and (2) positive response to inhaled nitric oxide defined as a decrease in mean pulmonary artery pressure of at least 2 mmHg and an increase in Pa^O2(FI^O21) of at least 40 mmHg after nitric oxide inhalation at a concentration of 10 ppm. Exclusion criteria were (1) circulatory shock, defined as a systolic arterial pressure < 90 mmHg or dependence on exogenous catecholamines; (2) documented cardiac dysrhythmias; (3) treatment with antiplatelet or anticoagulant agents (aspirin, nonsteroidal antiinflammatory agents, heparin, oral anticoagulants) or drugs interfering with hemostasis; (4) underlying hemostasis disorder (constitutive or acquired); (5) a platelet count < 100 Giga/l; and (6) a decreased platelet aggregation defined as a maximal intensity of platelet aggregation less than the lowest value observed in healthy volunteers (< 20% for adenosine diphosphate (ADP). < 75% for collagen, and < 70% for ristocetin).

The trachea of each patient had been intubated with an endotracheal tube that incorporates two side ports, one that runs to the distal tip of the endotracheal tube and one more proximal that ends 6 cm from the tip (Hi-Lo Jet Tracheal Tube no. 8, Mallinckrodt, Argyle, NY). These additional channels were used for continuous monitoring of tracheal pressure and tracheal concentrations of inhaled nitric oxide.

Once identified, all patients were sedated and paralyzed with a continuous intravenous infusion of 250 micro gram/h fentanyl, 1 mg/h flunitrazepam, and 4 mg/h vecuronium, and their lungs were ventilated using continuous positive pressure in a conventional volume-controlled mode (Cesar ventilator, Taema, Antony, France). Minute ventilation was adjusted to maintain Pa^CO2between 35 and 45 mmHg. The fraction of inspired oxygen (FI^O2) was continuously monitored using an oxygen analyzer (Seres 2000, precision = 0.5%) and maintained at 0.85 for the duration of the study. Positive end-expiratory pressure was maintained at 10 cmH²O and the inspiratory time at 30% to provide optimal alveolar recruitment. In all patients, hemodynamic monitoring included the use of a fiberoptic thermodilution pulmonary artery catheter (Oximetrix Opticath catheter, Abbott Critical Care System-France, Rungis, France) and a radial or femoral arterial catheter.

Nitric oxide was administered to all patients as follows: Nitric oxide was released from three tanks of nitrogen with nitric oxide concentrations of 25, 900, and 2,235 ppm measured using chemiluminescence (Air Liquide, Paris-LaDefense, France). Nitric oxide was continuously delivered within the inspiratory limb of the ventilator before the Y piece by using a nitrogen flowmeter calibrated in the range 0.250 - 1 l/min. Three intratracheal concentrations of inhaled nitric oxide were administered in a randomized order: 1, 3, 10, 30, and 100 ppm. Nitric oxide concentrations of 1 and 3 ppm were obtained using the 25-ppm tank, nitric oxide concentrations of 10 and 30 ppm were obtained using the 900-ppm tank, and the concentration of 100 ppm was obtained using the 2,235-ppm tank. In each condition, the nitric oxide flow was adjusted to obtain the desired intratracheal nitric oxide concentration, and tidal volume and FI^O2were adjusted to compensate for the added volume of nitric oxide gas. Thus, by keeping respiratory frequency constant, minute ventilation and FI^O2delivered to the patient remained unchanged regardless whether nitric oxide was administered. Control measurements were systematically performed before nitric oxide administration (C¹) and after the last nitric oxide concentration (C sub 2). Endotracheal concentrations of nitric oxide and nitrogen dioxide were continuously measured using a chemiluminescence apparatus (NOX 2000 Seres, Aix-en-Provence, France, precision = 0.005 ppm). An operating range of 0-5 ppm was selected for measuring intratracheal nitric oxide concentrations of 1 and 3 ppm and the NOX 2000 was calibrated using a tank of nitrogen containing 0.9 ppm of nitric oxide (Air Liquide). An operating range of 0-100 ppm was selected for measuring intratracheal nitric oxide concentrations of 10, 30, and 100 ppm, and the NOX 2000 was calibrated using a tank of nitrogen containing 25 ppm of nitric oxide (Air Liquide). When a different operating range of measurement was used, systematic recalibration of the NOX 2000 was performed to increase the precision of measurement. Intratracheal gas was continuously sampled using a continuous aspiration of 150 ml/min through the proximal side port of the Mallinckrodt endotracheal tube. At this aspiration flow rate, the time response of the NOX 2000 was about 40 s. Therefore, only mean intratracheal concentrations of nitric oxide were measured, and fluctuations between inspiratory and expiratory concentrations were not evaluated. The NOX 2000 also was used for continuous monitoring of oxygen concentration to ensure that FI^O2was maintained close to 0.85 during nitric oxide administration. This technique of nitric oxide administration (high nitric oxide concentration tank and low flow of nitric oxide) was intended to avoid any significant decrease in FI^Osub 2 during nitric oxide inhalation.

In each patient, systolic and diastolic arterial pressures and systolic and diastolic pulmonary arterial pressures were measured simultaneously using the arterial cannula and the fiberoptic pulmonary artery catheter connected to two calibrated pressure transducers (91 DPT-308 Mallinckrodt) positioned at the midaxillary line. Systemic and pulmonary arterial pressures, electrocardiogram, and tracheal pressure (Paw), measured through the distal port of the endotracheal tube, were recorded simultaneously on a Gould ES 1000 recorder (Cleveland, OH) for two different conditions: control (without nitric oxide) and during administration of nitric oxide at a concentration of 3 ppm.

In each phase, when a steady-state was obtained--defined as a relatively constant pulmonary arterial pressure--systolic and diastolic arterial pressures, systolic and diastolic pulmonary arterial pressures, pulmonary capillary wedge pressure, right atrial pressure, and Paw were recorded at a speed of 50 mm/s. Mean arterial pressure was calculated as one-third systolic arterial pressure plus two-thirds diastolic arterial pressure. Mean pulmonary artery pressure was measured by planimetry as the mean of four measurements performed at end-expiration. Systolic and diastolic arterial pressures, systolic and diastolic pulmonary arterial pressures, pulmonary capillary wedge pressure, and right atrial pressure also were measured at end-expiration. Cardiac output was measured using the thermodilution technique and a bedside computer allowing the recording of each thermodilution curve (Oximetrix 3 SO2/CO computer). Four serial injections of 10 ml of saline at room temperature performed at random during the respiratory cycle were used to average the variations in cardiac output related to the inspiratory and expiratory phases of continuous positive-pressure ventilation. Heart rate was measured from the recorded electrocardiogram. Systemic and pulmonary arterial blood samples were simultaneously withdrawn within 1 min after the measurements of cardiac output (after discarding an initial 10-ml heparin-contaminated aliquot). Arterial pH, Pa^O2, Pv^O2, and Pa^CO2were measured using a BGE blood gas analyzer (Instrumentation Laboratory-France, Paris, France). Hemoglobin concentration, methemoglobin concentration, and hemoglobin oxygen saturations (Sa^O2and Sv sub O²) were measured using a calibrated OSM3 hemoximeter (Radiometer Copenhagen-France, Neuilly-Plaisance, France). We used standard formula to calculate cardiac index, pulmonary vascular resistance index, systemic vascular resistance index, intrapulmonary shunt, and oxygen delivery.

Method of Sampling. Arterial blood samples were collected in 3.8% trisodium citrated tubes (9:1 vol/vol, Becton Dickinson-France, Le Pont de Claix, France) for platelet aggregation measurements and in EDTA tubes (Becton Dickinson) for platelet count and hematocrit. To prevent platelet activation inside the arterial line, 20 ml of arterial blood was first slowly and smoothly sampled and discarded; then, a 20-ml arterial blood sample was collected for platelet aggregation analysis. Blood samples were collected before administration of nitric oxide (control 1), after each concentration of nitric oxide when a 30-min steady-state was obtained, and after nitric oxide cessation (control 2).

Platelet Aggregation. In a previous study, we were unable to detect any nitric oxide-induced inhibition of platelet aggregation and agglutination in patients with acute respiratory failure receiving 2 ppm of inhaled nitric oxide. [10]In fact, this "negative" result was due to a methodologic artifact: any delay between arterial sampling and platelet aggregation analysis may result in a false "normalization" of nitric oxide-induced inhibition of platelet aggregation. This effect is likely related to an in vitro loss of nitric oxide. In the current study, the following technique was used for measuring platelet aggregation and agglutination. After arterial sampling, the tubes were centrifuged and platelet aggregation analysis was performed immediately. Platelet-rich plasma (PRP) was obtained by centrifugation of whole blood at 1,500 g for 1 min. Platelet-poor plasma was prepared from the same blood sample by centrifuging blood at 1,500 g for 15 min. The platelet count in PRP was adjusted between 250 and 350 Giga/l by addition of autologous platelet-poor plasma. Aggregation and agglutination tests were performed according to the turbidimetric method of Born. [11]The PRP was incubated under continuous stirring (900 rounds per minute) at 37 degrees Celsius. Platelet aggregation was induced by 2.5 and 5 micro Meter ADP and 10 micro gram/ml collagen (type I; Helena-France, St. Leu, France). Platelet agglutination was induced by 1.5 mg/ml ristocetin (Helena-France). These automated tests were performed on a Helena Packs 4 aggregometer (Helena-France). The increase in light transmission was recorded for 4 min after addition of the different aggregating agents (agonists). Aggregation and agglutination induced by the different agonists in PRP were evaluated by measuring light transmission in stimulated PRP, assuming that light transmission was 100% in platelet-poor plasma and 0% in nonstimulated PRP. As shown in Figure 1, maximal intensity of platelet aggregation was defined as the maximal increase in light transmission and velocity of platelet aggregation as the speed of the increase in light transmission increase, after each aggregating agent, as computed by the software. The software and the aggregometer/computer interface for this operation and the generation of printed platelet aggregations reports were designed and developed by Helena. This technique was first tested in ten healthy volunteers who had not taken any antiplatelet agent for the preceding 10 days. Normal values and ranges for maximal intensity and velocity of platelet aggregation for each agonist are shown in Table 1.

Platelet Count and Hematocrit. Whole blood platelet count and hematocrit were measured with an Argos monitor (Roche, Hoffman La Roche Diagnostic Systems, Basel, Switzerland).

Ivy Incision Bleeding Time. A sphygmomanometer was installed on the arm and inflated to maintain a constant pressure of 40 mmHg. One minute after, a standardized horizontal incision (5 mm long and 1 mm deep) was performed on the forearm with an automated device (Simplate, Organon Teknika, Durham, NC). The incision was blotted every 30 s by a filter paper until the bleeding stopped. Normal values of the bleeding time have been reported to be in the range 4-8 min [12]. Ivy bleeding time was measured immediately after each arterial sampling corresponding to C1, C2, and the different nitric oxide concentrations.

The effects of increasing concentrations of inhaled nitric oxide on platelet aggregation and agglutination were measured using a one-way analysis of variance for repeated measures followed by Fisher's exact test. Nitric oxide 3 ppm-induced inhibition of platelet aggregation was compared for the three agonists (ADP, collagen, and ristocetin) using a one-way analysis of variance for repeated measures followed by a Fisher's exact test. Hemodynamic and respiratory effects induced by nitric oxide 3 ppm were analyzed using a nonparametric Wilcoxon's paired test. All data are presented as mean plus/minus SD. A P value of less than 0.05 was considered statistically significant.

During the study period, 18 consecutive surgical critically ill patients who met the diagnostic criteria for severe ARDS (inclusion criteria 1) and who responded to inhaled nitric oxide (inclusion criteria 2) were screened for inclusion into the study. Among these patients, one was excluded because he was receiving aspirin, four because they had a baseline platelet count < 100 Giga/l, and seven because they had a baseline decreased platelet aggregating activity, when compared to the healthy volunteers group. Thus, six male patients (age 37 plus/minus 16 yr) were included in the study: Four were admitted after trauma and two after surgical procedures (orthopedic and gastrointestinal surgery). The initial clinical data for the patients, recorded on the day of inclusion into the study, just before initiation of the protocol and during intermittent positive-pressure ventilation (FI^O21.0, inspiratory time 30%, and 0 positive end-expiratory pressure) are presented in Table 2.

Hemodynamic and respiratory data measured with and without nitric oxide are summarized in Table 3. The administration of inhaled nitric oxide resulted in a significant reduction in mean pulmonary artery pressure and pulmonary vascular resistance index (P < 0.05), whereas heart rate, cardiac index, mean arterial pressure, right atrial pressure, pulmonary capillary wedge pressure, systemic vascular resistance index, oxygen consumption, oxygen delivery, and oxygen extraction ratio remained unchanged. There was a significant increase in Pa^O2and Sv^O2(P < 0.05), with a concomitant reduction in intrapulmonary shunt (P < 0.05) after nitric oxide administration. After the cessation of inhaled nitric oxide, pulmonary artery pressure and Pa^O2returned to control values in all patients (data not shown).

After nitric oxide administration at a concentration of 3 ppm, there was no significant increase in methemoglobin plasma concentrations. Mean endotracheal nitrogen dioxide concentration measured using chemiluminescence was found to be 0.05 plus/minus 0.003 ppm after inhalation of 3 ppm of nitric oxide at a FI^O2of 0.85.

Platelet aggregation and agglutination (maximal intensity and velocity) were significantly decreased after nitric oxide administration (Figure 1, Table 4and Table 5). The effect was maximal at 3 ppm of nitric oxide, reaching 50% of the control value but was not dose-dependent in the range 1-100 ppm (Figure 2and Figure 3). The effects of the three aggregating agents were inhibited by inhaled nitric oxide. At a nitric oxide concentration of 3 ppm, ADP-induced platelet aggregation and ristocetin-induced platelet agglutination were significantly more inhibited than collagen-induced platelet aggregation (P < 0.05, maximal intensity only). In four patients, platelet aggregation was measured in duplicate either immediately after blood sampling or after a 20-min delay; although a significant decrease in maximal intensity and velocity was evidenced when platelet aggregation was measured immediately after sampling, no inhibition could be found when platelet aggregation was measured after a 20-min delay.

No lengthening of Ivy bleeding time was observed after nitric oxide administration as compared to the control value. Furthermore, bleeding time remained in the normal range in all patients (Figure 4).

This study demonstrates that platelet aggregation and agglutination induced ex vivo by three agonists--ADP, collagen, and ristocetin--is significantly inhibited in patients with ARDS receiving low doses of inhaled nitric oxide. This effect is obtained without any apparent prolongation of the bleeding time, in contrast to what has been reported in healthy volunteers. [13]Platelet aggregation inhibition by nitric oxide was first demonstrated in vitro by Mellion et al. [14]and then confirmed by Radomski et al. [15]The incubation of human PRP with nitric oxide resulted in a concentration-dependent inhibition of platelet aggregation induced by ADP, collagen, and thrombin. Nitric oxide was two- to threefold more potent in human washed platelets than in PRP. However, inhibition of aggregation of human washed platelets decayed (half-life approximately 2 min) and disappeared after 4 min. Furthermore, preincubation of platelets with hemoglobin or Iron²+ reduced the antiaggregating activity of nitric oxide. When comparing nitric oxide and prostacyclin-induced inhibition of platelet aggregation, the same authors found that nitric oxide was less potent than prostacyclin. Salvemini et al. also demonstrated that in vitro nitric oxide completely inhibited thrombin-induced platelet aggregation. [16]However, this inhibition was reversed by oxyhemoglobin administered in vitro 30 s to 10 min after thrombin-induced platelet aggregation, according to a time-dependent profile. [16]Nitric oxide is unstable, especially when stirred in solution, [17]as in the aggregometer cuvette. The current study confirms the in vitro findings that the inhibitory effect of inhaled nitric oxide on platelet aggregation disappears with time. After 20 min of prolonged centrifugation or conservation of the blood sample at room temperature, no platelet aggregation inhibition could be evidenced. Additional studies are required to elucidate the mechanisms by which inhaled nitric oxide-induced inhibition of platelet aggregation vanishes with time.

Despite the fact that, in the current study, inhaled nitric oxide was administered at increasing concentrations ranging between 1 and 100 ppm, nitric oxide-induced inhibition of platelet aggregation did not appear to be dose-dependent, the maximal inhibition being observed at 3 ppm. Further studies are needed to determine whether nitric oxide-induced inhibition of platelet aggregation is, like nitric oxide-induced decrease in pulmonary vascular resistance index, [6]dose-dependent in the range 0.1-3 ppm. The inhibition of platelet aggregation was less pronounced when collagen was used as the agonist as compared to ADP. It may be assumed that this difference was related to the dose of collagen used in the current study. It is likely that, if smaller concentrations of collagen had been chosen, nitric oxide-induced inhibition of platelet aggregation would have been the same with ADP and collagen. Because ristocetin-induced platelet agglutination was inhibited by inhaled nitric oxide to the same extent as was ADP-induced platelet aggregation, it is highly likely that inhaled nitric oxide inhibited both platelet adhesion and aggregation. However, these effects could be measured only ex vivo. The effect of inhaled nitric oxide on platelet function could be direct or indirect. Exogenous nitric oxide might increase platelet cGMP or induce the production of a substance that could modulate platelet function. Because nitric oxide-induced effects on platelet functions vanish with time, the true antiplatelet effect of inhaled nitric oxide in vivo cannot be assessed with certainty. Golino and Yao have shown that nitric oxide, either endogenous or exogenous, is able to inhibit intravascular platelet aggregation in the Folts model of carotid thrombosis in the rabbit. [18-20]After placing an external constrictor around endothelially injured arteries, cyclic flow reductions due to recurrent platelet aggregation were measured at the site of stenosis using a Doppler probe directly inserted on the artery. Soluble nitric oxide infused in the carotid completely abolished cyclic flow reductions in all animals. These effects were transient, and cyclic flow reductions were restored spontaneously within 10 min after cessation of nitric oxide infusion. Therefore, in the rabbit, nitric oxide was considered as an antithrombotic agent inhibiting platelet aggregation in vivo to an extent similar to aspirin and thromboxane synthase inhibitors. Because the antiaggregating potency of nitric oxide on washed platelets was found to be three- to fourfold lower in the rabbit than in humans, [15]it may be assumed that nitric oxide is a potent antithrombotic agent in humans.

Patients with ARDS often experience pulmonary arterial microthrombosis, which may increase pulmonary artery pressure, compromise right ventricular function, promote lung ischemia, and result in the development of severe fibrosis. [3,21]Platelet activation is involved in such a process, [22,23]and antiplatelet agents have been proposed as a prevention of these deleterious effects. [3]However, most of antiplatelet agents, such as prostacyclin and nitroprusside, are associated with important hemodynamic side effects that limit their routine use in patients with ARDS. In addition, most of these agents are nonselective pulmonary vasodilators and worsen gas exchange when administered to patients with acute respiratory failure. [5]In contrast, inhaled nitric oxide, which inhibits platelet adhesion and aggregation, is a selective pulmonary vasodilator and offers the possibility of combining a potent antithrombotic effect with an increase in arterial oxygenation and a reduction of pulmonary hypertension. Platelets are involved in the early pulmonary hypertensive response observed in ARDS. [3,22]Infusion of ADP into sheep promotes platelet aggregation and generates pulmonary hypertension. The increase in pulmonary artery pressure is not observed if the animals are platelet-depleted. [24]Therefore, it can be hypothesized that an early use of inhaled nitric oxide, by preventing thrombi formation in the pulmonary circulation, could prevent, in part, the increase in pulmonary artery pressure characterizing late stages of ARDS. Inhaled nitric oxide could contribute indirectly to limiting the fixed part of pulmonary hypertension in addition to its well recognized ability to reverse pulmonary artery constriction.

Regarding the bleeding risk, nitric oxide plays a role in primary hemostasis. Remuzzi et al. showed that N-monomethyl-L-arginine, a specific inhibitor of nitric oxide formation, normalized bleeding when given to uremic rats. [25]This correction was reversed by giving the animals the nitric oxide precursor L-arginine. The authors have emphasized the role of nitric oxide as a mediator of the bleeding tendency of uremia. However, with respect to the potential effect of nitric oxide on primary hemostasis, in the current study performed in patients with ARDS and without preexisting coagulation disorders, no increase in the bleeding risk could be demonstrated because bleeding time was not prolonged. Several mechanisms might induce a discrepancy between inhaled nitric oxide-induced inhibition of platelet aggregation and its lack of effect on bleeding time. Low hematocrit and platelet count and edema, factors frequently encountered in critically ill patients, may interfere with primary hemostasis and induce a lengthening of Ivy bleeding time. It should be emphasized that, despite the fact that the patients of the current study had a decreased hematocrit (Table 2), no prolongation of the bleeding time was observed. Confirming a previous study, [26]we recently found that the level of von Willebrand factor (antigen and activity) frequently is increased in critically ill patients with ARDS. [10]It could interfere with primary hemostasis [27]and contribute to a decrease in the bleeding time. Our results contrast with Hogman's showing a 33% prolongation of the bleeding time 15 min after administration of 30 ppm of nitric oxide in the rabbit [28]and in six healthy volunteers. [13]Reasons for these opposite results are not clear, but it may be assumed that factors influencing primary hemostasis might be different in rabbits, in healthy volunteers, and in patients with ARDS. Furthermore, there is a general agreement for considering that bleeding time is neither sensitive nor specific for predicting the bleeding risk in many clinical settings. [29].

In conclusion, the beneficial effects of inhaled nitric oxide on arterial oxygenation and pulmonary circulation in patients with ARDS are associated with a significant inhibition of platelet aggregation and agglutination. This antithrombotic effect is not associated with a significant lengthening of the bleeding time in the absence of preexisting coagulation disorders.