Key words: Acute respiratory distress syndrome, Arterial oxygenation, Dose-reponse relationship, Inhaled vasodilators, Nitric oxide, Prostacyclin, Pulmonary hypertension, Ventilation-perfusion distribution.
INHALATION of low concentrations of nitric oxide causes selective pulmonary vasodilation. [1,2]In patients with severe acute respiratory distress syndrome (ARDS), inhalation of nitric oxide resulted in a decrease of the intrapulmonary right-to-left shunt and increased arterial oxygenation. The lack of systemic hemodynamic effects is explained by the immediate inactivation of inhaled nitric oxide by hemoglobin. Recent reports suggest that aerosolized prostacyclin (PGI2) also may induce selective pulmonary vasodilation and a redistribution of pulmonary blood flow away from non-ventilated lung regions despite its half-life of 2-3 min. [5,6]We describe the short-term effects of aerosolized PGI2in comparison to inhaled nitric oxide in three children with severe ARDS.
This investigation was performed with the approval of the institutional ethics committee. Informed consent was obtained from each patient's family.
The first set of measurements evaluated the effects of three concentrations of aerosolized PGI2(0.4, 2, and 4 micro gram *symbol* ml sup -1 equivalent to about 2, 10, and 20 ng *symbol* kg sup -1 *symbol* min sup -1, respectively) and of three concentrations of inhaled nitric oxide (0.1, 1, and 10 ppm). Each concentration of a vasodilator was administered for 30 min, preceded and followed by a baseline period.
In a second set of measurements, we directly compared the effects of aerosolized PGI2with those of inhaled nitric oxide. For this set of measurements, the concentration causing the greatest increase in the ratio of arterial oxygen partial pressure to the fraction of inspired oxygen (PaO2/FIO2) during the first set of measurements was used. Systemic and pulmonary hemodynamics, gas exchange, and the ventilation-perfusion ratio (V with dotA/Q with dot) were determined before, during, and after each vasodilator administration. To assess V with dotA/Q with dot distributions, the multiple inert gas elimination technique (MIGET) was used. [7,8]Intrapulmonary right-to-left shunt (Q with dotS/Q with dotT) was defined as the fraction of blood perfusing unventilated lung regions (V with dotA/Q with dot ratio < 0.005); low V with dotA/Q with dot was defined as the fraction of blood flow in lung regions with V with dotA/Q with dot between 0.005 and 0.1; V with dotA/Q with dot ratios from 0.1 to 10 were taken as the normal range of V with dotA/Q with dot; and V with dotA/Q with dot ratios of infinity were defined as dead space.
Each step in each sequence lasted approximately 30 min, and measurements were performed toward the end of each period, when hemodynamic function was stable.
Technique of PGI sub 2 and Nitric Oxide Delivery
Three concentrations of PGI2(Flolan, Wellcome Laboratories, Beckenham, Kent, United Kingdom) were prepared in a glycine buffer (0.188% glycine, 0.147% sodium chloride, pH 10.5) and diluted with saline immediately before use. Each preparation of PGI2was administered with an infusion rate of 10 ml/h into a nebulizer chamber. The particles produced by the nebulizer had a mean mass diameter of 2.5 micro meter (range 1.5-5.5 micro meter), determined by phase-Doppler-anemometry; the maximum temperature achieved in the chamber was < 40 degrees Celsius. The ultrasound nebulizer (prototype, Siemens, Lund, Sweden) was inserted into the inspiratory limb of the ventilator circuit and delivered aerosol particles about 15 cm in front of the endotracheal tube. Nitric oxide was delivered using a prototype of a modified Siemens 300 ventilator (Siemens, Lund, Sweden) equipped with a built-in computerized nitric oxide delivery system, consisting of an additional digital-controlled nitric oxide valve. Inspiratory nitric oxide concentrations were measured by chemiluminescence (AL 700, ECO Physics, Duernten, Switzerland).
Child 1, an 8-yr-old boy (25 kg), was admitted to our intensive care unit on day 15 of mechanical ventilation due to severe ARDS induced by Epstein-Barr virus pneumonia and, in the course of the disease, aspergillosis. Because PaO2/FIO2remained < 50 mmHg despite maximum conventional therapy and nitric oxide inhalation, venovenous extracorporeal membrane oxygenation with a heparin-coated system was started. The short-term effects of aerosolized PGI2in comparison to inhaled nitric oxide were studied on day 22 of extracorporeal membrane oxygenation with a constant extracorporeal blood flow at 2.1 l/min. In the following days, the child was treated intermittently with inhaled nitric oxide. On day 43 of mechanical ventilation, extracorporeal membrane oxygenation was discontinued, but ventilatory support was required for an additional 17 days. The boy was discharged from our hospital 3 months after onset of ARDS.
Child 2, an 8-yr-old girl (26 kg), was admitted to our intensive care unit for treatment of severe ARDS 11 days after near-drowning with cardiac arrest. The effects of aerosolized PGI2were studied 4 days after referral to our intensive care unit. Using permissive hypercapnia, dehydration, prone position, and nitric oxide inhalation, gas exchange improved, and the FIO2could be reduced gradually. After 39 days of mechanical ventilation, the girl could be extubated. Ten days later she was discharged.
Child 3, an 8-yr-old girl (24 kg), was diagnosed with ARDS after reoperation for a large ependymoma of the fourth ventricle. She was referred to our hospital on day 41 of mechanical ventilation, and the effects of the vasodilators were studied on day 55. Although ARDS resolved within the next days, weaning from mechanical ventilation was difficult. During the following weeks, the child experienced recurrent episodes of ARDS and died on day 100 of mechanical ventilation.
During the study, the lungs of all children were ventilated in a pressure-controlled mode with a peak airway pressure of 27-35 cmH sub 2 O, a positive end-expiratory pressure of 9-15 cmH2O, a respiratory rate of 13-30/min, and tidal volumes from 160 to 250 ml.
In the first set of measurements, at least one of the three tested concentrations of inhaled nitric oxide and aerosolized PGI2caused selective pulmonary vasodilation in each child (Table 1and Table 2). None of the concentrations of any vasodilator induced a decrease in systemic arterial pressure or an increase in cardiac output. In each child, all three concentrations of nitric oxide and at least one concentration of PGI2resulted in an increased arterial oxygenation (Table 1and Table 2). Whereas the effect of nitric oxide inhalation on PaO2started within 1-2 min after onset and ended within the same time period after discontinuation of nitric oxide, this required 5-10 min of aerosolized PGI2.
In the second set of measurements, selective pulmonary vasodilatory effect of aerosolized PGI2and inhaled nitric oxide was confirmed (Figure 1). In the first and third children, 10 ng *symbol* kg sup -1 *symbol* min sup -1 aerosolized PGI2and inhaled nitric oxide increased PaO2/FIO2(Figure 2) In the second child, 20 ng *symbol* kg sup -1 *symbol* min sup -1 aerosolized PGI2decreased the PaO2(Figure 2), which was in contrast to the results of the first set of measurements. In children 2 and 3, both vasodilators caused a slight decrease in PaCO2. The MIGET demonstrated that the increased arterial oxygenation in children 1 and 3 was caused by a decrease in Q with dotS/Q with dotTduring PGI2administration and during inhalation of nitric oxide. At the same time, perfusion to normal lung areas increased; the perfusion of low V with dotA/Q with dot areas remained unchanged.
To our knowledge, these case reports describe the first comparison of the effects of aerosolized PGI2and inhaled nitric oxide in children suffering from severe ARDS. At least one of the three tested concentrations of both vasodilators caused a decrease in pulmonary artery pressure without a clear decrease in mean systemic arterial pressure, suggesting a selective pulmonary vasodilation. The observed improvement in arterial oxygenation occurring at different concentrations of aerosolized PGI2may be explained by probable vasodilation, predominantly in ventilated lung areas. This hypothesis is supported by the results of the MIGET, which demonstrated that the increased arterial oxygenation in children 1 and 3 during aerosolized PGI2and during inhaled nitric oxide was due to an intra-pulmonary redistribution of blood flow away from shunt areas toward well ventilated lung regions, whereas the selective vasodilation of aerosolized PGI2in child 2 was accompanied by an increase of blood flow toward shunt areas and a decrease of blood flow to well ventilated lung areas. The increase in PaO2/FIO2and the decrease in pulmonary artery pressure achieved with aerosolized PGI2and inhaled nitric oxide was comparable. As shown for nitric oxide, aerosolized PGI2also demonstrated an individual dose-dependent effect with PGI2: the maximum improvement in PaO2/FIO2did not occur at the same dose that caused the maximum decrease in PAP. In child 2, the dose of PGI2that decreased PAP the most resulted in a decreased PaO2/FIOsub 2. Therefore, we believe that future studies dealing with effects of aerosolized PGI2and of inhaled nitric oxide on hemodynamics and gas exchange should be preceded by dose-response studies in each patient.
These case reports demonstrate that the inhalation strategy allows an agent such as PGI2, with a half-life of 2-3 min, to act as a selective pulmonary vasodilator and, at precise concentrations, as a selective vasodilator of well ventilated lung regions. With the exception of 10 and 20 ng *symbol* kg sup -1 *symbol* min sup -1 in the first set of measurements in child 3, none of the tested concentrations of aerosolized PGI2induced a decrease in mean systemic arterial pressure. Theoretically, an increase in the dose of aerosolized PGI2might cause a "spillover" of the prostanoid into systemic circulation, thereby reducing the mean systemic arterial pressure. .
In contrast to nitric oxide, the concentration of which can be measured by chemiluminescence or electrochemical fuel cells, the concentration of PGI2and amount absorbed cannot be precisely measured. The nebulized dose of PGI2was calculated from the concentration of the PGI2solution, the infusion rate into the nebulizer chamber, and the body weight. It is obvious that this does not reflect the actual amount of PGI2inhaled, because there will be losses in the nebulizer chamber, ventilator tubing, and the endotracheal tube. Others have estimated that the aerosol fraction deposited in the alveolar space during mechanical ventilation is less than 10%. [11,12]Therefore, the alveolar dose of inhaled PGI2probably was much less than that delivered and calculated. Because the effect of 20 ng *symbol* kg sup -1 *symbol* min sup - aerosolized PGI2differed between the first and second sets of measurement, the effective dose of inhaled PGI2may differ even if the same solution and the same infusion rate into the nebulizer chamber are used in the same patient at different times.
Because both vasodilators may cause selective pulmonary vasodilation and may improve arterial oxygenation, both inhaled drugs are clinically advantageous compared to intravenous vasodilators. However, before these drugs are inhaled as a routine long-term treatment of pulmonary hypertension, the side effects and toxicity of both drugs when inhaled should be studied more extensively.
In conclusion, we believe that aerosolized PGI2may become an alternative to inhaled nitric oxide in the treatment of severe ARDS.
The authors thank Professor P. Radermacher and the Siemens Comp., for the ultrasound nebulizer, and Gisela Kaufmann, for technical assistance.