Study objective: Evaluate the interest of the response to a therapeutic optimization as a predictor of prognosis in ARDS.
Design: Prospective study.
Setting: ICU of a University Hospital.
Patients: Thirty-six consecutive patients with severe ARDS addressed for extracorporeal carbon dioxide removal (ECCO2R).
Interventions: We studied the response during the first 2 days after arrival to the therapeutic optimization strategy consisting in a combination of the following: (1) decrease in extravascular lung water (diuretics or hemofiltration); (2) selection of the best ventilatory mode; (3) permissive hypercarbia; and (4) correction of hypoxemia by alveolar recruitment, additional continuous oxygen insufflation, body position changes (prone position), inhaled nitric oxide, enhancement of hypoxic pulmonary vasoconstriction with almitrine, and drainage of pleural or mediastinal effusions. In patients remaining severely hypoxemic despite these modalities, ECCO2R was then proposed.
Measurements and results: Thirty-six patients were addressed after 8.3+/-5.5 days of mechanical ventilation. On arrival, mean simplified acute physiologic score was 46.8+/-14.2, multiple system organ failure score was 1.8+/-1.6, Murray score was 3.4+/-0.4, PaO2 was 75.3+/-31.3 (fraction of inspired oxygen [FIO2]=1) for a positive end-expiratory pressure level of 12.3+/-3.4 cm H2O. Nineteen of 36 patients improved their gas exchange within 2 days and their mortality was 21%. The seventeen remaining patients did not improve PaO2/FIO2; PaCO2 and airway pressures remained high and their mortality was 88%. This different response to therapeutic optimization appeared using stepwise logistic regression as the most predictive factor for mortality (p<0.05).
Conclusions: In patients with severe ARDS, the response to an early performed therapeutic optimization used to improve hypoxemia appeared to be a highly discriminant factor distinguishing deceased from surviving patients.