The effect of arterial blood gas values on extubation decisions.

BACKGROUND No studies have examined the clinical utility of arterial blood gas (ABG) values during spontaneous breathing trials (SBTs) for making extubation decisions. Nonetheless many intensive care units measure ABGs during an SBT to determine, in conjunction with other data, whether the SBT was successful. OBJECTIVE Determine how often ABG values change extubation decisions. METHODS Each patient was assessed at the bedside by one of 3 critical care attending physicians. Each assessment was made at the end of the SBT in which ABGs were obtained with the expectation of extubation. When a nurse, respiratory therapist, or resident was available, he or she also simultaneously and independently evaluated the patient and then reassessed with the ABG values. Physiologic data were recorded before and during the SBT, as the attending physician performed the bedside assessment. RESULTS Eighty-three patients were studied during 100 SBTs. Heart rate, respiratory rate, and mean blood pressure increased. Blood oxygen saturation (measured via pulse oximetry [F(IO)(2)]) and the P(aO)(2)/F(IO)(2) ratio decreased significantly during SBTs. Concordant clinical and ABG assessment (to attempt a trial of extubation) led to extubation in 86 patients, 76 of whom remained extubated after 72 hours. Following 14 of the SBTs, the patients were not extubated. Seven of these were judged by the attending physicians to fail both clinically and following review of ABG values. Seven were judged successful clinically, but the patients were not extubated because the attending physician thought the ABG values were unfavorable. Bedside nurses, who assessed 72 SBTs, had 1 case in which the ABG values changed their extubation opinion. Resident-trainees assessed 76 cases and had 5 changes of opinion based on ABG values. Respiratory therapists assessed 77 patients and had 4 such cases. The average changes in heart rate and blood pressure of patients not extubated were double those of extubated patients, but there were no significant differences in respiratory rate or P(aO)(2)/F(IO)(2) ratio between those groups. Ten patients (11.6%) were extubated based on the attending physician's decision but required re-intubation within 72 hours. CONCLUSION These data suggest that ABG values did not change extubation decisions in 93% of cases. However, in 7 cases the ABG values changed the extubation decision. If even a few of those cases would have failed extubation without knowledge of the ABG values, the increased patient risk and cost associated with failed extubation would more than offset the relatively small cost of collecting ABG values from all patients who undergo SBT.