DOI: 10.1097/CCM.0b013e3181c585fb

,  

,  

PMID: 20083937

Issn Print: 0090-3493

Publication Date: 2010/02/01


Print

Central venous saturation as a predictor of extubation failure*

Angela T. Wratney

    loading  Checking for direct PDF access through Ovid

Excerpt

Tolerance of a spontaneous breathing test (SBT) indicates weaning success, but variably predicts extubation success. After successful SBT, the need for reintubation within the subsequent 24 hrs to 72 hrs occurs in 5% to 30% of patients, depending on the population (1). The difficult-to-wean patient, by definition, has failed to tolerate the initial SBT with successful weaning requiring up to three SBTs or up to no more than 7 days from the initial SBT evaluation. After a successful SBT in this population, 15% to 33% will fail the extubation trial and require reintubation (2). Thus, identifying a noninvasive measure during the SBT, which could accurately predict extubation outcome for these difficult-to-wean patients, would inform the clinical decision to extubate.
In this issue of Critical Care Medicine, Teixeira et al (3) evaluate the change in central venous saturation (Scvo2) during the SBT as a predictor of extubation failure in difficult-to-wean patients. The authors hypothesized that Scvo2 could be a “reliable and convenient tool to rapidly warn about the acute changes in oxygen supply and demand of the patient during weaning.”
Across three medical-surgical intensive care units, 73 patients mechanically ventilated for >48 hrs who had previously failed a 2-hr SBT were enrolled in the study. All patients were weaned, using a standardized two-step process which included daily readiness to wean assessments, a 2-hr T-piece SBT with measure of frequency/tidal volume (f/Vt) ratio during the first 30-min interval, and measures of various physiologic and ventilatory parameters at 1 min and the 30 mins of the SBT. All patients had a subclavian or internal jugular central venous catheter present for Scvo2 measurement. Patients with an f/Vt >105 or intolerance of the SBT were returned to ventilatory support and reassessed the next day (4, 5). Patients who had an f/Vt ≤105 and who successfully passed an SBT were extubated and monitored for 48 hrs. The protocol among all three institutions included immediate use of noninvasive ventilation on extubation for all patients with chronic obstructive pulmonary disease. Extubation failure (EF) was defined as the need for reintubation within 48 hrs of extubation.
EF occurred in 31 patients (42.5%). Significant differences between the EF and extubation success (ES) group were detected at the 30th minute of the SBT including an increased heart rate, lower Pao2, Sao2, and Scvo2 in the EF group. Multivariate regression identified Scvo2 as the only independent variable able to discriminate extubation outcome. At 30 mins of the SBT, the Scvo2 in the EF group was 60 ± 8 as compared with 70 ± 7 in ES group (p = .009). A reduction in Scvo2 by >4.5% during the SBT was an independent predictor of reintubation, with an odds ratio of 49.5 (95% confidence interval, 12.1–201.5), sensitivity of 88% and specificity of 95%. These results indicate that Scvo2 measured during a T-piece trial may be an important part of the assessment of extubation readiness in the difficult-to-wean patient.
The utility of Scvo2 as a surrogate for SJOURNAL/ccme/04.02/00003246-201002000-00058/ENTITY_OV0456/v/2017-07-20T221634Z/r/image-pngo2 in clinical-decision making has been evaluated with conflicting results, particularly in the critically ill patient (6–9). The Scvo2 has performed well as a surrogate target for hemodynamic resuscitation, yet it seems to have a variable correlation with the SJOURNAL/ccme/04.02/00003246-201002000-00058/ENTITY_OV0456/v/2017-07-20T221634Z/r/image-pngo2, depending on the patient's disease and severity of cardiopulmonary status at the time of measurement. In 60 patients with low cardiac index, the limits of agreement between the SJOURNAL/ccme/04.02/00003246-201002000-00058/ENTITY_OV0456/v/2017-07-20T221634Z/r/image-pngo2 and Scvo2 were −19% to +18% with a correlation coefficient of 0.46 to 0.72 (7). The correlation between these two measures has been evaluated among various populations of intensive care patients, resulting in estimates between 0.78 and 0.95 (6–9).
    loading  Loading Related Articles