Catheter suction (CS), the standard method for airway secretion management during mechanical ventilation, is invasive and has significant hemodynamic and traumatic side effects. In-line mechanical insufflation-exsufflation (IL-MIE) is a new, noninvasive technology that clears secretions by cough-simulation, without interrupting ongoing ventilation. It is not known whether IL-MIE can be safely and effectively used as an alternative to CS in ventilated patients. Methods: A randomized, controlled, non-inferiority study comparing a standard protocol of CS, with automatic IL-MIE (CoughSync, Ruxin Medical Systems, Beijing) performed every 30 minutes, with CS added only if needed, in post-operative ventilated patients.

The study was designed as a randomized, open-label, parallel, non-inferiority, controlled trial. The study was carried out in two Intensive Care Units - the Cardiac Surgery ICU at Anzhen Hospital, and the Critical Care Medicine ICU at Tian Tan Hospital - both in Beijing, China, over a 36 month period between July 2015 and July 2018. The study group included patients aged 18 to 75 years undergoing mechanical ventilation during the immediate recovery period following a cardiac or neuro-surgical procedure.

Subjects were randomized to either a control group, managed for 8 hours with CS whenever the subject showed signs of airway secretion accumulation, as per standard clinical practice for those ICU's, or a study group, managed for 8 hours with automatic IL-MIE treatments (CoughSync, Ruxin Medical Systems Company Ltd, Beijing, China) performed automatically every 30 minutes, and with CS performed only if signs of airway secretion accumulation manifested despite ongoing IL-MIE. IL-MIE was performed using standard IL-MIE parameters (exsufflation pressure = -60 cm H2O, with flutter, 10 coughs/treatment).

Before each CS treatment, 100% oxygen was administered for one minute, regardless of the patient's baseline oxygen requirement, as per standard operating protocol in that ICU. Before IL-MIE treatments, no additional oxygen was administered beyond the patient's baseline oxygen requirement.

Pharmacological management in both groups included analgesia and sedation as routinely used postoperatively in those ICU's.

Demographic information and vital signs were recorded for all subjects. Arterial partial pressure of oxygen (PaO2), arterial oxygen saturation (SaO2 ) and oxygenation index (PaO2/FiO2) were defined as the primary end points of the study, and arterial partial pressure of carbon dioxide (PaCO2), pulseoximetry (SpO2), heart rate (HR), and ventilator parameters (inspired oxygen, tidal volume [Vt], peak inspiratory pressure [PIP], airway plateau pressure [Pplat], and PEEP) as secondary end points.

All primary and secondary end-point data were recorded at baseline (2 minutes prior to starting the trial), and at 5 minutes, 4 hours and 8 hours after commencement of the trial. The number of CS treatments performed on each subject was recorded throughout the trial. Follow up for adverse events was performed during, and 48 hours after completion of, the trial.

Statistical Analysis:

The minimum sample size required to demonstrate non-inferiority, was calculated to be 49 in each cohort, or a total of 98 subjects in total.

Collected data were analyzed with a mixed model with repeated measures (MMRM) considering all observations (2 minutes before commencement of the trial protocol, and at 5 minutes, 4 hours and 8 hours thereafter) and accounting for the baseline value of oxygenation. For the indices of oxygenation derived from blood gas measurements (PaO2, SaO2 , and oxygenation index), non-inferiority was evaluated by comparison to the two-sided 95% confidence interval of the treatment effect in the MMRM model, with non-inferiority between the IL-MIE and control cohorts established if the lower limit of the 95% confidence interval for the intergroup difference in least squares mean for a measured index was found to be higher than the pre-determined non-inferiority margin for that index. For comparison of other quantitative data between groups, a two-sample t-test or Wilcoxon rank sum test was used, based on the data distribution. For between-group comparison of subjects with significant adverse events, a chi-square test was used. The number of CS treatments performed in each cohort was analyzed post-hoc as an exploratory analysis.

Inclusion Criteria:

• patients undergoing mechanical ventilation during the immediate recovery period following a cardiac or neuro-surgical procedure
• Ventilation expected to last more than 8 hours

Exclusion Criteria:

• acute spinal cord shock
• recent airway trauma or surgery
• cardiogenic pulmonary edema or ARDS necessitating ventilation with a peak end expiratory pressure (PEEP) equal to or greater than 6 cmH2O
• pneumothorax
• hemoptysis
• severe ischemic heart disease
• lung tumors
• pulmonary Tuberculosis
• history of lung transplantation
• pregnant or breastfeeding women