MATERIALS AND METHODS
In an academic 24-bed medico-surgical intensive care unit, patients were screened over a 2-year period. Adult patients were enrolled if they had hypoxemic respiratory failure meeting standard moderate to severe ARDS criteria: Ratio of the partial pressure of arterial oxygen (PaO2) to the fraction of inspired oxygen (FiO2) of 200 or less, pulmonary capillary wedge pressure (Pcwp) ≤ 18 mmHg and bilateral infiltrates on frontal chest radiograph. Recruited patients also had a pulmonary artery catheter and an arterial line. Patients with severe hemodynamic instability (defined as the need for more than one vasopressor or the use of more than 0.5 μg/kg per minute of norepinephrine), on intravenous milrinone or nitrate derivatives that could not be weaned for study purposes and patients on high frequency oscillatory ventilation were excluded. Patients with a history of hypersensitivity to study medications, pregnant patients and those who participated in another study involving oxymetric values or pulmonary hemodynamics were also excluded.
Patients were randomly administered sequential nebulization of iNO (20 ppm) or epoprostenol (10 μg/mL for a total volume of 5 mL). Thereafter, milrinone (1 mg/mL for a total volume of 5 mL at each nebulisation) alone and in association with iNO was administered. Each drug was nebulized for 20 min and a 30 min washout was allowed between each drug. We used a jet nebulization device ventilator synchronized to nebulize epoprostenol and milrinone (MicroMist® Nebulizer model 1880; Hudson RCI, Temecula, CA, United States). The nebulizer was attached to the inspiratory limb of the ventilator near the endotracheal tube. The mass median diameter obtain with this nebulizer is 2.1 μm and the nebulization flow is between 0.25 and 0.3 mL/min. The ventilatory circuit humidification was stopped during the nebulization. Adjustment of the tidal volume was done during nebulization to obtain the same minute ventilation. The iNO was administered using a standard iNO Delivery system (INOvent®, Ohmeda, Madison, WI, United States) attached to the inspiratory limb of the ventilator. A constant dose of 20 ppm of iNO was used and monitored by the injection device. Blood pressure and oxygenation status were continuously monitored. If hemodynamic instability occurred, defined as a lowering of systolic arterial pressure ≥ 10 mmHg or lowering of mean arterial pressure (MAP) ≥ 5 mmHg, the study medication was stopped. If oxygenation status worsened, defined as a lowering of arterial oxygen saturation ≥ 10% (measured with continuous pulse oximetry and confirmed by arterial gas), the study medication was stopped. Adverse events potentially related to study medications such as increase hemodynamic instability and renal failure was prospectively evaluated. Using a previously published milrinone assay, we determined the plasma level of milrinone at the end of the administration of inhaled milrinone or the combination of iNo and inhaled milrinone in eight samples from our last four patients.
Demographic data, APACHE II and SOFA scores were collected. The following parameters were measured before and during each specific drug nebulization: MAP, mean pulmonary arterial pressure, thermodilution cardiac output, PaO2, heart rate, central venous pressure and Pcwp. The following parameters were calculated: Systemic and pulmonary vascular resistances (SVR and PVR), PVR/SVR Ratio, indexed pulmonary vascular resistance, transpulmonary gradient, PaO2/FiO2, cardiac index, shunt and oxygenation index. A patient was considered a responder to study medications if is PaO2 increased of more than 20% from the pre-inhalation value.
As the primary goal of this pilot study was safety and feasibility and no preliminary data existed in this population, a convenience sample of 15 patients was chosen. Given a within patient standard deviation of 11%, the sample size enabled the detection of a 12% difference in PaO2 between baseline and post-milrinone PaO2. Given the small sample-size, demographic and baseline data were described as medians and interquartile range. Continuous data were analysed using Wilcoxon signed rank tests.
The study was performed at Hôpital du Sacré-Coeur de Montréal, Canada. The local ethics committee approved the protocol and consent was obtained from patients or their next of kin. The protocol was submitted to and approved by Health Canada.