Abstract & Commentary
An Evidence-Based Extubation Bundle Improved Care Outcomes in Mechanically Ventilated Brain-Injured Patients
By Leslie A. Hoffman, RN, PhD
Professor Emeritus, Nursing and Clinical & Translational Science, University of Pittsburgh
SYNOPSIS: Implementation of an evidence-based extubation-readiness bundle was associated with a decrease in mechanical ventilation days and pneumonia in brain-injured patients.
SOURCE:Roquilly A, et al. Implementation of an evidence-based extubation readiness bundle in 499 brain-injured patients. A before-after evaluation of a quality improvement project. Am J Respir Crit Care Med 2013;188:958-966.
Brain injury is a frequent cause of prolonged mechanical ventilation. The authors hypothesized that use of a systematic management protocol, termed an extubation-readiness bundle, could reduce the duration of mechanical ventilation for brain-injured patients. Patients were required to have: 1) evidence of acute neurological injury, e.g., extradural hematoma, subarachnoid hemorrhage, brain contusion, brain edema, skull fracture, stroke, or abscess, 2) require mechanical ventilation for > 24 hours, and 3) Glasgow Coma Score of ≤ 12. Patients were excluded if there was a decision to limit care within 24 hours of ICU admission or if they died within 24 hours of admission. The study used a before-after design. The control phase included all patients (n = 299) admitted to the two study ICUs over a 3-year period. The intervention was then introduced over a 12-month period (no data collected). The intervention phase included all patients (n = 200) who met entry criteria during the following 22-month period. The bundle included four components: 1) lung protective ventilation (tidal volume 6-8 mL/kg of ideal body weight; PEEP > 3 cm H20; respiratory rate set to achieve normocapnia or moderate hypocapnia), 2) early enteral nutrition (Day 1), 3) optimized antibiotic therapy (predefined criteria), and 4) a systematic approach to extubation. Sedation management was not included in the bundle because it was already standardized.
Patients enrolled in the bundle experienced fewer days of mechanical ventilation (12.6 ± 10.3) compared to control (14.9 ± 11.7 days; P = 0.02). At day 28 and day 90, the intervention group also experienced more ventilator-free days (P = 0.01), with no difference in mortality in the ICU (P = 0.51) or at day 90 (P = 0.22). The rate of hospital-acquired pneumonia declined from 57.5% in the control phase to 47.5% in the intervention phase (P = 0.03). There was a nonsignificant increase in reintubation for intervention vs control patients (13.8% vs 9%; P = 0.11). Unplanned extubation decreased from 9.4% (control) to 4.5% (intervention) (P < 0.01).
Commonly, patients with an acute brain injury are explicitly excluded from clinical trials. Therefore, guidance regarding weaning from mechanical ventilation is limited. This is concerning, as brain injury is a frequent cause of prolonged mechanical ventilation. The protocol tested in this study consisted of four components: lung protective ventilation, early enteral nutrition, protocolized antibiotic therapy, and a standardized extubation protocol. Lung protective ventilation was included based on two rationales: lung compliance of brain-injured patients is frequently not altered and hypercapnia can be easily prevented by increasing respiratory rate instead of increasing tidal volume. Early enteral nutrition was included because a prior study, conducted by the authors, suggested early enteral nutrition was a protective factor in regard to the risk of hospital-acquired pneumonia. International guidelines, adapted to local epidemiology, were used to guide antibiotic therapy. Criteria used to indicate extubation readiness included tolerance of inspiratory support < 10 cm H2O or spontaneous breathing and FIO2 ≤ 40% for ≥ 30 minutes. Tube removal occurred if the Glasgow Coma Score was ≥ 10 and the patient had a cough that was spontaneous or caused. Full awakening was not required. These criteria were selected based on findings from prior studies that suggested a low risk of extubation failure when the Glasgow Coma Score was between 8 and 10, provided patients can cough.
Using these criteria, 13.8% of patients required reintubation in the intervention phase, a percentage slightly higher than goal (≤ 10% in neurological patients). However, there were significantly fewer unplanned extubations. As with all before-after studies, the study design prevents attributing causality to positive outcomes. Notably, the authors included several statistical procedures designed to exclude the possibility that extraneous factors caused their results in their analytic strategy. These included a time series analysis to rule out changes due to improvements in patient care and two sensitivity analyses designed to balance covariates in the two phases and reduce bias. Future studies are needed to confirm findings. In the interim, study findings suggest ways to potentially speed weaning of this complex and challenging patient population.