Key words: Aerosol. Congenital heart disease. Newborn. Prostacyclin. Pulmonary hypertension. Surgery.
INCREASED pulmonary vascular resistance is common in patients with congenital heart disease [1]and may be further exacerbated by open heart surgery and cardiopulmonary bypass (CPB). [2,3]Successful management of pulmonary hypertension subsequent to CPB has been limited, because the intravenous administration of substances that decrease pulmonary artery pressure also decrease systemic arterial pressure [4,5]and may adversely affect pulmonary gas exchange. [6].
Inhaled nitric oxide has been shown to reduce pulmonary hypertension after CPB in newborns and children. [7-9]Unfortunately, nitric oxide is not generally available, its safe administration is technically demanding and expensive, and its use may be limited by its side effects.
Alternatively, it has been shown that aerosolized prostacyclin (PGI2), similar to nitric oxide, has the potential to elicit selective pulmonary vasodilation both in animals [10]and humans [11]and to improve right ventricular function by reduction of increased pulmonary artery pressure. [12]In contrast to inhaled nitric oxide, no data are available on the effects of inhaled PGI2on cardiopulmonary function in newborns presenting with severe pulmonary hypertension aggravated by cardiac surgery and CPB. In the current report, we describe the successful intra- and postoperative use of aerosolized PGI2in a newborn with congenital heart disease and severe pulmonary hypertension.
Case Report
A 3.9-kg male newborn presented with hypoxemia and cardiovascular collapse immediately after birth and was transferred to our institution on day 4 of life. Cardiac catheterization revealed unobstructed total anomalous pulmonary venous connection of the supracardiac type. The child was hemodynamically unstable, and emergent surgical repair was undertaken.
Anesthesia was induced with 0.15 mg fentanyl (Janssen, Neuss, Germany) and 1 mg pancuronium (Organon-Teknika, Eppelheim, Germany) and maintained by bolus injections of fentanyl and pancuronium. Other drugs administered at time of induction included dobutamine (5-8 micro gram *symbol* kg sup -1 *symbol* min sup -1; Dobutrex, Lilly Deutschland, Bad Homburg, Germany) and alprostadil (50 ng *symbol* kg sup -1 *symbol* min sup -1; Minprog Pad., Upjohn, Heppenheim, Germany).
Mechanical ventilation was performed (ID 3.5 mm, Vygon, Ecouen, France) using a Siemens Servo 900 D ventilator (Elema, Sweden). Ventilation parameters were FIO21.0, minute volume 2.0 l/min, respiratory rate 41 breaths/min, PEEP 2 cmH2O, maximal airway pressure 20 cmH2O, and l:E ratio 0.43:0.57. Monitoring included electrocardiogram, pulse oximetry, temperature (rectal, esophageal), invasive blood pressure via cannulation of the left femoral artery (22-G cannula, Arrow Deutschland, Erding, Germany), central venous pressure via the right anonymous vein (4-French, 2-lumen, Arrow Deutschland, Erding, Germany), and arterial blood gases (Blood Gas System 288, Ciba Corning Diagnostic, Medfield, MA).
Surgery consisted of median sternotomy, installation of CPB, total cardiocirculatory arrest, anastomosis of the pulmonary venous confluence to the left atrium, closure of the atrioseptal defect, and ligation of the connection vein. During that time, a distended truncus pulmonalis and a minimally working right ventricle were noticed. CPB was terminated during infusion of enoximone (10 micro gram *symbol* kg sup -1 *symbol* min sup -1; Perfan, Marion Merrell Dow, Russelsheim, Germany) and alprostadil (150 ng *symbol* kg sup -1 *symbol* min sup -1) into the pulmonary artery. Five minutes later (Figure 1), systolic arterial pressure (SAP) was 70 mmHg and systolic pulmonary artery pressure (SPAP) was 60 mmHg.
Within the following minutes, right ventricular contraction further deteriorated, as judged by direct visual inspection. Because therapy had resulted in no improvement of cardiocirculatory function, the decision was made to administer epoprostenol, an analog of PGI2, via the tracheal route. For this purpose, a jet nebulizer (Servo Nebulizer 945, Siemens, Erlangen, Germany) was connected to a Siemens 900 D ventilator. The nebulizer chamber (Cirrus, Intersurgical, Twickenham, United Kingdom) was adapted directly to the endotracheal tube. Thereafter, 500 micro gram epoprostenol (Flolan, Wellcome, London, United Kingdom) were dissolved in 50 ml glycine buffer (pH 10.5) resulting in a PGI2concentration of 10 micro gram/ml. Six milliliters of this solution was filled into the nebulizer chamber, and the nebulizer was started operating at its low-flow mode. In this mode, the nebulizer delivers a peak inspiratory flow of 7 l/min, thereby adding notably to the minute volume applied to the patient. Therefore, the minute volume delivered by the ventilator was reduced from 2.2 to 0.91/min. Within the following 15 min, SPAP decreased from 60 to 50 mmHg, and SAP increased from 70 to 75 mmHg (Figure 1). Blood gas analysis revealed a PaO2of 343 mmHg, and PaCO2of 19 mmHg indicated severe hyperventilation. Therefore, the nebulizer was turned off after 15 min.
Within the following 10 min, SPAP increased again to 60 mmHg, and PaO2decreased to 123 mmHg. Therefore, the nebulizer was turned on again; ventilator volume, however, was set to 0.31/min, resulting in PaCO2of 28 mmHg and peak airway pressure of 22 cmH2O. Similar to the first trial, an increase of the SAP-to-SPAP gradient and of PaO2were observed (Figure 1). Despite this improvement, the cardiocirculatory situation remained critical, and it was decided to close the chest using a large Goretex-patch to avoid further compromising right ventricular function.
Surgery lasted for 320 min, with a CPB time of 150 min and cardiocirculatory arrest of 59 min. Transfusions included 500 ml packed erythrocytes, 150 ml platelets, and 320 ml fresh frozen plasma.
The patient was transferred to the pediatric intensive care unit using a transportable ventilator to allow continuous PGI2administration. PGI2was continuously aerosolized for 13 h, and SAP-to-SPAP gradient increased to 44 mmHg (Figure 2). Thereafter, PGI sub 2 was discontinued. Within the following 30 min, however, an increase of SPAP from 33 to 44 mmHg and a decrease of PaO2from 259 to 65 mmHg were observed. At this time, nitric oxide was available and was administered at a concentration of 40 ppm. This caused SPAP to decrease again to a level similar to that observed during aerosolization of PGI2. Within the next few days, the patient's condition continuously improved. Administration of nitric oxide was stopped on the 4th day, and the chest was definitely closed. Separation from mechanical ventilation was started on the 5th postoperative day, and extubation was performed 11 days later.
Discussion
To our knowledge, this is the first case to demonstrate that inhalation of PGI2is effective in reducing an increased pulmonary artery pressure in a newborn with congenital heart disease and CPB without impairing systemic blood pressure or gas exchange. Although selective pulmonary vasodilation has been described for inhaled nitric oxide (and was confirmed in our patient), the advantages of inhaled PGI sub 2 are its general availability and that administration is comparably easy.
Several factors independent of PGI2may have influenced pulmonary artery pressure in our patient. Because of the severity of the condition of the newborn, enoximone and PGE1have been administered in the perioperative phase to reduce pulmonary artery pressure. Both drugs were continuously infused for more than 45 min before the onset of PGI2aerosolization, and no changes of dosage have been performed. Therefore, it seems unlikely that the reduction of SPAP in the post-CPB phase was a pharmacologic effect of enoximone or PGE1.
This view is supported by the fact that a decrease of SPAP (from 60 to 50 mmHg) was observed within a few minutes after the onset of aerosolization of PGI2. The termination of PGI2inhalation reproducibly resulted in a reduction of the SAP-to-SPAP gradient (Figure 1and Figure 2), which could be reversed by PGI2aerosol (operating room) or inhalation of nitric oxide (intensive care unit).
There was no measurement of pulmonary blood flow, and we cannot be certain what the natural history of post-CPB pulmonary hypertension was without prostacyclin treatment for this patient. Although the small increase in SPAP from 50 to 60 mmHg when the aerosolized therapy was discontinued is supportive of the efficacy of this agent, it is well known that there is a transient overshoot in pulmonary artery pressure when almost any vasodilator is abruptly discontinued. Therefore, in this child, we may be overestimating the extent of the underlying pulmonary hypertension during the discontinuation of prostacyclin at 13 h.
Because no degradation products of epoprostenol (e.g., 6-keto PGF1alpha) have been measured, we cannot exclude the possibility that part of the observed reduction of SPAP was due to systemic resorption of inhaled PGI2. Yet, this possibility seems unlikely because systemic administration of PGI2elicits systemic hypotension and may impair pulmonary gas exchange. [4-6]Systemic arterial pressure, however, remained unchanged in our patient, whereas gas exchange was improved with PGI2aerosol. As a possible mechanism of pulmonary selectivity of inhaled PGI2, it may be proposed that inhaled PGI2, after inhalation and deposition within the bronchoalveolar tree, diffuses to its target vessels in the pulmonary vasculature resulting in local vasodilation. Part of the highly concentrated, local PGI2then is absorbed by the systemic circulation. At the same time, however, it is considerably diluted by the large intravasal compartment, which probably makes its concentration fall below that necessary to cause systemic side effects.
PGI2was put into the nebulizer chamber at a concentration of 10 micro gram/ml. Because 6 ml was sufficient for approximately 5 h, it can be calculated that 12 micro gram PGI2left the nebulizer chamber per hour. Taking into account a rate of disappearance of fluid from the nebulizer chamber of approximately 1.2 ml/h and the newborn's weight, the amount of PGI2aerosolized per minute approximated 51 ng *symbol* kg sup -1 *symbol* min sup -1. This dose is large compared to intravenous doses. [13]However, because only a small fraction (less than 5%) of an aerosol is deposited in the alveoli of the lung during mechanical ventilation, [14]the effective dose administered to our patient was much less than 51 ng *symbol* kg sup -1 *symbol* min sup -1 (probably in the range of 2.5 ng *symbol* kg sup -1 *symbol* min sup -1), explaining the absence of effects on central hemodynamics and gas exchange.
Possible side effects of PGI2include inhibition of platelet aggregation, thereby increasing the risk of intra-and postoperative bleeding. In our patient, significant bleeding was not observed.
Several technical problems were encountered during aerosolization. First, the Siemens 945 nebulizer delivers a peak inspiratory flow of as much as 7 l/min, even if operating in the low-flow mode. As a consequence, the nebulizer contributes considerably to the total minute volume. In the current case, hyperventilation and an increase of airway pressures during PGI2aerosol could be avoided only by reducing the minute volume delivered by the ventilator to minimum (0.3 l/min). With this setting, a critical reduction of body temperature might occur in small children and newborns during long-term aerosolization, because a device to heat the gas delivered to the nebulizer is not available.
Furthermore, the nebulizer output may be too high in babies weighing less than 4 kg, and a modification of the set-up (e.g., the introduction of an outlet valve at the tube) would be required when treatment with PGI2aerosol is considered. Finally, endotracheal suction had to be performed frequently in our patient. This may be explained by the amount of fluid nebulized throughout the duration of therapy. Because intratracheal PGI2may cause an irritation of the airways, [15-17]we cannot exclude, however, that part of the hypersecretion observed was due to a direct irritant effect of PGI2on the bronchial system.
In summary, the current case suggests that intratracheal administration of PGI2may decrease a critically high and otherwise untreatable pulmonary artery pressure in newborns without causing systemic hypotension or deterioration of gas exchange. Further studies to establish dose-response relationships and to define potential toxic effects of inhaled PGI2are being performed in our laboratory.