Objective: To determine whether diagnosis and bedside observations predict outcomes of patients who have declined intubation but accept noninvasive positive pressure ventilation (NPPV) to treat their respiratory failure.
Design: Prospective multiple-center cohort trial.
Setting: Two teaching hospitals and two community-based hospitals in southeastern New England from January through October 1999.
Patients: All patients receiving NPPV for acute respiratory failure were screened and enrolled if they had a written do-not-intubate (DNI) order.
Interventions: Patients were begun on NPPV with mean inspiratory and expiratory pressures of 13.4 +/- 0.3 and 5.0 +/- 1 cm H2O, respectively. Respiratory therapists recorded demographic information, blood gases, and ventilator type and settings, and they made bedside assessments of cough strength, presence of airway secretions, awake state, and agitation. Patients were followed until discharge for duration of NPPV, survival status, and disposition.
Measurements and main results: Of 1,211 screened patients, 114 had a DNI status and were enrolled into the study. Of these, 49 (43%) survived to discharge. Age, gender, location in a community vs. teaching hospital, and initial pH and PaO2 did not affect survival, but a higher baseline PaCO2 was associated with a favorable odds ratios for survival to discharge. Diagnosis was an important determinant of survival, with congestive heart failure patients having significantly better survival rates than those with chronic obstructive pulmonary disease, cancer, pneumonia, or other diagnoses. A stronger cough and being awake were also associated with increased probability of survival.
Conclusion: Patients with respiratory failure and a DNI status have a high overall mortality rate when treated with NPPV, but those with diagnoses such as congestive heart failure or chronic obstructive pulmonary disease, who have a strong cough, or who are awake have better prognoses. These data should be useful when counseling DNI patients and their families on use of NPPV.