Physiologic impact of closed-system endotracheal suctioning in spontaneously breathing patients receiving mechanical ventilation.

BACKGROUND Endotracheal suctioning is required but can have adverse effects, and could affect cardiorespiratory variables that are used to predict whether the patient is ready for extubation. METHODS In a prospective cohort study in a university hospital's medical intensive care unit, we measured the impact of closed-system suctioning on cardiopulmonary variables in spontaneously breathing patients weaning from mechanical ventilation. All spontaneously breathing, mechanically ventilated patients were screened for enrollment at the initiation of weaning from mechanical ventilation. Before, during, and after standardized closed-system endotracheal suctioning we measured minute volume, heart rate, arterial oxygen saturation, mean arterial pressure, respiratory frequency, oxygen saturation, and tidal volume. RESULTS Twenty-nine patients were enrolled after a median of 5 (interquartile range [IQR] 3-9) ventilator days. Twenty-five patients (86%) were spontaneously breathing on pressure-support ventilation when suctioned. The median post-suctioning recovery time was > 5 min for minute volume, tidal volume, respiratory rate, and ratio of respiratory rate to tidal volume. The post-suctioning median values of the maximum deviations in the ventilatory variables were clinically important: minute volume -2.4 (IQR 1.6-3.7) L/min, respiratory rate 8 (IQR 2-14) breaths/min, tidal volume -175 (108-220) mL. Heart rate, mean arterial pressure, and oxygen saturation increased after suctioning (P < .05), but the increases were not clinically important. CONCLUSIONS Post-suctioning changes in the measured variables persisted longer in these spontaneously breathing patients weaning from mechanical ventilation than in patients who are sedated and paralyzed. The effects of suctioning on cardiopulmonary function should be considered in practice and during the design of future studies on weaning and extubation prediction variables.