A Randomized Trial of GM-CSF in Patients With ALI/ARDS

BACKGROUND: Respiratory failure due to ALI/ARDS remains a major health problem, despite significant progress in intensive care unit care and ventilator management. ALI/ARDS is characterized by unacceptably high mortality despite enormous expenditure of health care resources. Survivors face long-term consequences that may affect their quality of life. New therapies are needed to improve early survival and to decrease long-term sequelae of this syndrome. GM-CSF is a naturally occurring cytokine that is present in the normal lung, with important roles in pulmonary homeostasis. GM-CSF is essential for normal maturation and function of alveolar macrophages (resident inflammatory cells that are responsible for initial defense against pneumonia). Alveolar epithelial cells line the gas exchange surface of the lung. Acute lung injury and subsequent abnormal healing is linked to delayed repair of damage to the epithelium following initial injury. This can then lead to pulmonary fibrosis. GM-CSF has potent effects on alveolar epithelial cells, promoting proliferation and limiting epithelial cell death. Thus, GM-CSF has a distinctive combination of activities that make it an excellent candidate for a therapeutic intervention in ALI/ARDS. Preliminary studies for this project demonstrate that GM-CSF can protect experimental animals against acute lung injury, can decrease susceptibility to pneumonia, and is protective against pulmonary fibrosis following acute lung injury. There is extensive experience with the administration of recombinant human GM-CSF to human patients (this biological is approved by the FDA and has been well-tolerated in trials involving critically ill patients). This project is based on the hypothesis that administration of GM-CSF will improve clinical outcomes for patients with ALI/ARDS. DESIGN NARRATIVE: With the assent of the attending physician, informed consent will be obtained from the patient or next of kin as soon as possible after case identification. Physiologic measurements and specimen collection will begin at the time of entry into the study. Three days after the patient has met criteria for ALI/ARDS or at entry into the study (whichever is later), he/she will be randomized to receive recombinant human GM-CSF (250 mcg/M2) or placebo, administered by slow intravenous infusion once daily for 14 days. This study will allow entry of patients who have fulfilled criteria for ALI/ARDS for up to 7 days. Treatment will be initiated after patients have met criteria for at least 3 days. Treatment with GM-CSF may prove both safe and effective within the first 1-2 days of lung injury. However, the present study will not address that question. It is unlikely that the opportunity for improved outcome will be lost by delaying therapy for up to 3 days (based on the proposed mechanisms by which GM-CSF might benefit this patient population). Similarly, the decision to treat for 14 days will allow for improved outcome in patients with non-resolving ARDS by reducing the incidence of ventilator-associated pneumonia and by decreasing pathologic fibroproliferation. The primary endpoint for this study will be the duration of mechanical ventilation. Additional important endpoints will include changes in the severity of physiologic derangements of respiratory gas exchange, non-respiratory organ failure, and incidence of ventilator-associated pneumonia. Additional assessments designed to determine the mechanism of benefit of GM-CSF treatment will include measures of lung epithelial cell integrity and measures of alveolar macrophage (lung inflammatory cell) function.