Objective: To compare termination criteria (TC) of pressure support ventilation (PSV).

Design: Randomized cross-over clinical trial.

Setting: Intensive Care Units.

Patients: Sixteen patients on PSV with ability to answer a visual analog scale (VAS).

Protocol and Measurements: Each patient was ventilated with two different TC ventilators, fixed and automatic, with measurements of ventilatory mechanic variables, breathing comfort and asynchrony patterns.

Two ventilators with different flow termination criteria (TC) were compared: Servo 300 (Siemens-Elema, Solna, Sweden) with fixed TC (5% of peak inspiratory flow) and Newport E500 (Newport Medical Instruments, Costa Mesa, CA) with automatic TC (varies between 5% to 55%). Each patient remained three hours in the protocol, one hour in each ventilator, after been randomized to one of two sequences of 3 steps: Fixed 5% / Automatic / Fixed 5% or Automatic / Fixed 5% / Automatic (Figure 1). The transition among the ventilators was realized through a three-way flow directional valve (Three-Way T-Shape TM, Hans Rudolph Incorporation, Kansas City, EUA).

The PS level was determined by clinical staff to partially unload the respiratory muscle without respiratory distress based on observation of the patient's breathing pattern. The PS, the positive end expiratory pressure (PEEP), the inspiratory oxygen fraction (FiO2) and the pressure trigger sensitivity levels were unchanged during the protocol.

Study variables: Patients' demographic, anthropometric and clinical data were obtained (Table 1).

The study variables were detected noninvasively using only airway pressure (Paw) and flow signals, obtained with a pressure differential pneumotachometer (CO2SMO Plus; Novametrix Medical Systems, Wallingford, CT) located at the distal end of the ventilator circuit . Three records of five minutes were realized in each study step: time = 0' (0-5 min), t = 30' (30-35 min) and t = 55' (55-60 min). Data were digitalized at 100 Hz and recorded on a personal computer for subsequent analysis (LabVIEW 7.1, National Instruments Corporation, Austin, TX). A total of nine record phases of five minutes from each patient were analyzed .The variables values were obtained by one hundred ventilatory cycles from each recorded phase

Measured variables:

The following variables were analyzed (detailed descriptions of each one were in the Electronic Supplementary Material - ESM):

• Ventilatory parameters: PS (PSadjusted); measured PS (PSeffective) and PEEP (PEEPeffective); peak airway pressure; effective flow termination criterion (TC - V'ti/V'peak, %) .
• Breathing pattern: tidal volume obtained by mathematical integration of the flow signal; respiratory rate; minute ventilation; inspiratory time; expiratory time; duty cycle (Ti/Ttot); time constant obtained by measuring the slope off the expiratory limb of the flow-volume plot .
• Respiratory effort: pressure-time product required to ventilator trigger (PTPtrigger) ; inspiratory delivered assistance pressure-time products at 300 ms (PTPdelivered-assistance300) and 500 ms (PTPdelivered-assistance500) after the onset of inspiratory effort; over assistance pressure-time product (PTPover-assistance) (see Figure 2 for details); airway occlusion pressure at 0.1 second (P0.1) .
• Breathing comfort: evaluated by a modified VAS, labeled 0 to 10 (0 = best comfort and 10 = worst comfort) .
• Asynchrony patterns: ineffective efforts, both during inspiration and expiration, (IEinspiratory and IEexpiratory), double-triggering (D-T) and autotriggering (A-T) were detected by visual inspection of the flow and airway pressure signals recordings .

Statistical Analysis:

The statistical analysis was performed using the Statistical Package for Social Science (SPSS 15.0, Chicago, EUA) and the significance level was established as p < 0.05. Tests are detailed in the ESM.

Inclusion Criteria:

• Patients under PSV for more than 24 hours,
• Age > 18 years
• Ability to answer to a symptom visual analog scale (VAS)

Exclusion Criteria:.

• Hemodynamic instability (use of vasoactive drugs like dopamine, dobutamine or norepinephrine)
• Bronchopleural fistula