Objectives: There is increased awareness of imposed work of breathing contributing to apparent ventilatory dependency. This study evaluates the impact of tachypnea as an indicator of ventilatory failure during a room air-5 cm H2O continuous positive airway pressure, spontaneous breathing, preextubation trial when associated with increased imposed work of breathing.
Design: Prospective, descriptive, 1-yr data collection.
Setting: University hospital trauma intensive care unit (ICU).
Patients: Mechanically ventilated trauma ICU patients surviving to discharge.
Intervention: Patients were weaned to minimal mechanical ventilator support and underwent a 20-min room air-continuous positive airway pressure preextubation trial (FIO2 = 0.21, continuous positive airway pressure = 5 cm H2O [0.5 kPa]). When passed (PaO2 >/= 55 torr [>/= 7.3 kPa], PaCO2 </= 45 torr [</= 6.0 kPa] with prior eucapnea, arterial pH >/= 7.35, respiratory rate </= 30 breaths/min), extubation followed. If patients failed due to hypoxia, ventilatory support resumed. If tachypnea was the reason for failure, work of breathing was measured. If patient work of breathing was </= 1.1 joule/L, extubation proceeded despite tachypnea. If patient work of breathing was > 1.1 joule/L, imposed work of breathing was measured, and if residual "physiologic" work of breathing (patient work of breathing minus imposed work of breathing) was </= 0.8 joule/L, patients were extubated.
Measurements and main results: Of 589 extubations, 105 (18%) were classified as false negatives based on a preextubation rate of > 30 breaths/min. Of these, 97 were successfully extubated despite tachypnea ranging from 32 to 56 breaths/min, when combined with either a patient work of breathing </= 1.1 joule/L or physiologic work of breathing </= 0.8 joule/L. The rate of extubation failure within 72 hours was 7.8% (8/105) in the tachypneaic group, compared with 7.9% (38/484) for those patients with a respiratory rate of </= 30 breaths/min. Some of the stimulus for the tachypnea was possibly due to increased imposed work of breathing, as the increased respiratory rate usually abated within 18 hrs after extubation. The reliance on a respiratory rate of </= 30 breaths/min as an absolute preextubation criterion would have resulted in a sensitivity of 82%, a specificity of 17%, a positive predictive value of 92%, a negative predictive value of 8%, and an overall accuracy of 77%. The average duration of mechanical ventilation during the study period decreased by 2 days, from 8.6 to 6.3 days (p=.03).
Conclusions: Tachypnea as a marker of respiratory distress is sensitive, but is not sufficiently specific to be used as a criterion in preextubation trials. Reliance on tachypnea as a preextubation trial failure criterion is likely to prolong intubation and ventilatory support for a large number of patients. Patient risks, determined by the extubation failures and reintubation rate, are the same.