This was a prospective multicenter randomized controlled trial in 6 closed ICUs with 24/7 in-house intensivist coverage at 6 academic-affiliated medical centers in a resource limited setting from October 18, 2007 to May 03, 2014. All parts of the study were reviewed according to the Consolidated Standards for Reporting Trials (CONSORT) statement. Crossover was not allowed. Participants were blinded to randomization group, as were the healthcare providers and statistician. Randomization was accomplished using Random Allocation Software© (RAS; Informer Technologies, Inc.). Block randomization was performed by a computer generated random number list prepared by a statistician. Allocation consignment occurred through confidential communication between the patient's nurse and a third party not involved in recruitment. The data analyzer was blinded to group randomization and was not present during ventilator weaning. There were no important changes to methods after trial commencement. The study ended because it achieved the necessary sample size.
Pparticipants were randomly assigned to ventilator weaning either by protocolized methodology or by usual care (UC). All patients were ventilated using Dräger Evita® XL or Evita® 4 ventilators (Dräeger Medical, Inc., Lubeck, Germany). In the PW group, weaning and extubation readiness was determined by the RT using pre-defined criteria and the result of a SBT. Pre-defined weaning criteria included: (1) patent upper airway; (2) ability to protect airway (defined by mental status and presence of adequate gag and cough reflexes); (3) ability to clear secretions; (4) decreasing secretion burden requiring suction not more frequently than every 2 hours; (5) level of support (FiO2 < 50%, PEEP = 5); and (6) hemodynamic stability not requiring chemical (vasopressors, inotropes) or mechanical (e.g. intra-aortic balloon pump, extracorporeal life support) circulatory support. Until the SBT, management of each group was the same. For patients in the UC group, SBT type and determination to extubate was determined by attending physician preference. For participants in the PW group, the SBT type was protocol determined, and extubation decisions were driven by RT. The SBT consisted of continuous positive airway pressure (CPAP) of 5 mmHg with an FiO2 ≤ 0.4. After 3-minutes patients were assessed for appropriateness to continue the SBT by arterial oxygen saturation (SaO2) ≥ 92% without arrhythmias, and a rapid shallow breathing index (RSBI = RR/VT) < 105 breaths/min/L. Specified signs of respiratory distress included respiratory rate (RR)> 30 breaths/minute, SaO2< 90%, heart rate (HR)> 140 breaths/minute, or a sustained increase or decrease of HR of >20%, systolic blood pressure >200 mmHg or <80 mmHg, or agitation, anxiety, or diaphoresis without other identified cause. The SBT was performed for 120 min in accordance with prior studies. At the end of the SBT, the RSBI was re-measured and an arterial blood gas (ABG) was obtained.
In the UC group, the SBT type and extubation decision was determined by the attending intensivist on service based upon neurologic status, airway competence (gag, cough, suction requirements), and negative inspiratory force (NIF) or RSBI measurements. Extubation success was defined as remaining extubated for 48 hours without need for re-intubation or other forms of non-invasive MV.
