Design
The PROTECTION (Pressure support ventilation + sigh in acute hypoxemic respiratory failure patients) trial is an investigator-initiated, international, multicenter, parallel randomized controlled two-arm trial that will be performed in intubated and ventilated patients with AHRF, who are admitted to the intensive care unit (ICU). The study will be conducted in adherence to the principles of the World Medical Association’s Declaration of Helsinki and in accordance with the Medical Research Involving Human Subjects Act (WMO). Ethics approval will be sought from each participating institution before starting enrollment and consent will be obtained for each patient following local regulations. The Institutional Review Board of the coordinating center approved the protcol on 12 June 2017 under reference number 318_2017bis. The trial was registered at www.clinicaltrials.gov with code NCT03201263 in June 2017. The Standard protocol items: recommendation for interventional trials (SPIRIT) checklist can be found in Additional file 1 and the SPIRIT figure is included in the main body of the manuscript (Fig. 1).
Setting
The PROTECTION trial will be performed in the ICUs of 23 centers around the world (12 in Italy, 10 in other European countries, 1 in China). The coordinating center will be Policlinico Maggiore Hospital in Milan, Italy.
Study population
Consecutive patients admitted with AHRF to each participating ICU, who have been intubated and switched to PSV between 1 and 7 days after intubation will be screened between 4 and 24 h from the start of PSV. Further inclusion criteria will be partial arterial pressure of oxygen (PaO2)/fraction of inspired oxygen (FiO2) ratio ≤ 300 mmHg (measured at clinical positive end-expiratory pressure (PEEP) and FiO2 values); clinical PEEP ≥ 5 cmH2O and stable Richmond Agitation-Sedation Scale (RASS) value of − 2 to 0. Exclusion criteria will be clinical PEEP ≥ 15 cmH2O; partial pressure of arterial carbon dioxide (PaCO2) > 60 mmHg; arterial pH < 7.30; PaO2/FiO2 ratio ≤ 100 mmHg (measured at clinical PEEP and FiO2 values); age < 18 years; presence of central nervous system or neuromuscular disorders; history of severe chronic obstructive pulmonary disease or fibrosis; AHRF fully explained by cardiac failure or fluid overload; evidence of active air leak from the lung (e.g., pneumothorax); cardiovascular instability (e.g., systolic blood pressure (SBP) < 90 mmHg despite vasopressors); clinical suspicion of elevated intracranial pressure; extracorporeal support; moribund status and refusal by the attending physician.
If an eligible patient is excluded from participation, the reason(s) for exclusion will be registered.
The Consolidated Standards of Reporting Trials (CONSORT) diagram of PROTECTION trial is presented in Fig. 2.
Sigh test
To explore the possibility of implementing predictive enrichment in a larger future RCT [17], we planned a pre-randomization test to assess the prevalence of patients with AHRF with improved oxygenation after the introduction of sigh. After enrollment, FiO2 will be titrated to obtain pulse arterial oxygen saturation (SpO2) of 90–96% maintaining clinical PSV settings including PEEP; then, all patients will undergo a clinical test of PSV + sigh to assess the prevalence of improved oxygenation in sigh responders vs. non-responders. To this end, we will first record the SpO2/FiO2 ratio at the start of the test; then, 30 min after the introduction of sigh as the cyclic pressure control phase set at 30 cmH2O for 3 s once per minute [14], SpO2/FiO2 will be collected again to quantify the number of patients in whom this ratio increased (i.e., “sigh responders”) [18]. To deliver sigh, ventilators will be switched to biphasic positive airway pressure mode with the lower pressure-level set at clinical PEEP and the higher pressure-level of the sigh set at 30 cmH2O with a 3-s inspiratory time and then a 57-s expiratory time [16].
Randomization
At the end of the sigh test, patients will be randomized through an online automatic centralized and computerized system to one of the two study groups (1:1 ratio): PSV or PSV + sigh. Due to the nature of the intervention, blinding will not be feasible.
Interventions: ventilation strategies for each study arm
PSV group
After randomization, clinicians will set PSV to meet the following targets: tidal volume (Vt) of 6–8 mL/kg of predicted body weight (PBW), with respiratory rate (RR) 20–35 bpm. FiO2 will be left as selected before the pre-randomization sigh test, while PEEP will be left as clinically set.
PSV + sigh group
PSV will be set with the same protective targets of the PSV group (see above) but cyclic pressure control phase at 30 cmH2O for 3 s delivered once per minute (i.e., sigh) [14] will be added. FiO2 will be left as selected before the pre-randomization sigh test with clinical PEEP.
Adjusting ventilation settings
In both groups, PSV will be adjusted, at least every 8 h, as follows:
-
PSV support will be decreased by 2 cmH2O if Vt > 8 mL/kg PBW and/or RR < 20
-
PSV support will be increased by 2 cmH2O if Vt < 6 mL/kg PBW and/or RR > 35 and/or in the presence of respiratory distress (e.g., marked use of the accessory muscles)
-
PEEP and then FiO2 will be increased by 2 cmH2O and 0.1 if SpO2 is < 90%
-
FiO2 and then PEEP will be decreased by 0.1 and 2 cmH2O if SpO2 is > 96%.
The PSV group will be treated by protective PSV settings until day 28 or death or performance of a spontaneous breathing trial (SBT); the PSV + sigh group will be treated by protective PSV settings with the addition of sigh until day 28 or death or performance of a spontaneous breathing trial (SBT). Sigh settings will be left unchanged until day 28, death or the SBT.
Switch to controlled mechanical ventilation
In both groups, switch to protective controlled ventilation will be indicated if patient develops at least one of the following conditions: PSV support > 20 cmH2O; PEEP ≥ 15 cmH2O; unstable hemodynamic status (SBP < 90 mmHg with vasoactive drug); active cardiac ischemia (dynamic ST changes on cardiac monitor or electrocardiogram); unstable arrhythmias; uncontrolled hypertension (SBP > 180 mmHg); abrupt decrease in the level of consciousness (RASS <− 3); dangerous agitation (RASS > + 2); pH < 7.30; PaO2/FiO2 ratio ≤ 100 mmHg; necessity to perform diagnostic test (e.g., computed tomography (CT scan) or bronchoscopy).
Controlled ventilation will be set on volume mode with Vt 6–8 mL/kg PBW, RR to control pH, unchanged PEEP and FiO2. Controlled ventilation will be thereafter adjusted according to clinical evolution. Patients switched to controlled ventilation will be reassessed at least every 8 h and they will be switched back to PSV or PSV + sigh (to maintain study group assignment) targeting the aforementioned settings and adjustments as soon as all the following conditions are met: patient is able to trigger ventilator breaths; PaO2/FiO2 > 100 mmHg; PEEP < 15 cmH2O; pH ≥ 7.3; stable hemodynamic status with stable or decreasing doses of vasopressors for ≥ 6 h.
Rescue therapies
In the case of desaturation (SpO2 ≤ 90%) it will be crucial to exclude hemodynamic impairment as a possible cause. Also, airway obstruction and ventilator malfunction must be ruled out. Provided those factors are excluded, a rescue step-up strategy will be allowed: institution of protective controlled mechanical ventilation (see above for settings) and performance of recruitment maneuvers at 40–50 cmH2O, PEEP ≥ 15 cmH2O, prone positioning and inhaled nitric oxide, extracorporeal membrane oxygenation. Patients undergoing controlled ventilation and rescue treatments will be reassessed at least every 8 h and switched back to PSV or PSV + sigh (to maintain study group assignment) with the aforementioned settings and adjustments as soon as all the following conditions are met: patient can trigger ventilator breaths; PaO2/FiO2 > 100 mmHg; PEEP < 15 cmH2O; pH ≥ 7.3 and stable hemodynamic status with stable or decreasing doses of vasopressors for ≥ 6 h.
Weaning protocol and criteria for reintubation
SBT
Patients with SpO2 ≥ 90% on FiO2 ≤ 0.4 and PEEP ≤ 5 cmH2O, no agitation, hemodynamically stable with norepinephrine ≤ 0.1 μg/kg/min or equivalent and at a stable or decreasing dose for 6 h [19] and without any of the aforementioned criteria for switch to controlled ventilation will undergo the SBT. For patients in the PSV group, the attending physician will perform the SBT directly. For patients in the PSV + sigh group, the attending physician will first withdraw sigh, wait 60 min and confirm the criteria: if confirmed, the SBT will be performed; if not, sigh will be reintroduced and the clinical criteria will be checked again to repeat the procedure after at least 8 h.
The SBT will last at least 60 min with a combination of PEEP 0–5 cm H2O and PSV 0–5 cm H2O. At the end of the 60 min, the patient will fail the SBT if any of the following are present: criteria to start the SBT not confirmed; sustained (> 5 min) respiratory rate > 35 bpm; heart rate > 140 bpm; SBP > 180 or < 80 mmHg; marked complaint of dyspnea; increased somnolence with elevated PaCO2 and/or pH < 7.3; active cardiac ischemia (dynamic ST changes on cardiac monitor or electrocardiogram); abrupt decrease in the level of consciousness with RASS <− 3 or if cough is not strong enough to clear secretions.
Patients who fail the SBT will be switched back to PSV or PSV + sigh (to maintain study group assignment) and clinical criteria will be checked again to repeat the procedure after at least 6 h.
Patients who pass the SBT will be extubated or, in the presence of tracheostomy, mechanical ventilation will be discontinued. If a patient is reintubated or mechanically ventilated through a tracheostomy again within 48 h, PSV or PSV + sigh (to maintain study group assignment) will be restored. If a patient remains extubated or separated from the ventilator for > 48 h data collection only will continue.
Reasons for reintubation
After elective extubation, reintubation should be promptly performed if at least one of the following criteria is present: cardiac arrest; respiratory arrest (respiratory pauses with loss of consciousness or gasping for air); respiratory failure with SpO2 < 90% and/or RR > 35 bpm despite non-invasive ventilation; decreased level of consciousness impairing ability to protect airway; hemoptysis or hematemesis impairing ability to protect airway; abundant secretions that cannot be effectively cleared or are associated with lobar collapse, acidosis, hypoxemia or change in mental status; surgical/invasive procedure requiring sedation/anesthesia +/− neuromuscular blockade such that patient is no longer be able to sustain unassisted breathing or hemodynamic instability with SBP < 80 mmHg despite vasoactive drugs.
Standard of care
In all patients, standard of care for intubated patients with hypoxemic acute respiratory failure (e.g., restrictive fluid strategy, early appropriate antibiotics, prophylaxis of gastric stress ulcer and deep vein thrombosis, semi-recumbent positioning, respiratory physiotherapy, adequate nutrition, monitoring of sedation, pain and delirium, tracheostomy and non-invasive ventilation for post-extubation respiratory failure) will be granted throughout the whole ICU stay in accordance to local protocols.
End of follow up
Enrolled patients will be observed until day 28, ICU discharge or death, whichever comes first.
Study endpoints
Primary endpoint
The primary endpoint will be to verify the clinical feasibility of PSV + sigh vs. PSV. Feasibility will be assessed by comparing the number of patients in each group experiencing at least one of the following failure criteria [20]: switch to controlled ventilation following presence of one of the aforementioned reasons for ≥ 24 h consecutively; use of PEEP ≥ 15 cmH2O, prone positioning, inhaled nitric oxide, extracorporeal membrane oxygenation or reintubation within 48 h from extubation following one of the aforementioned reasons.
Secondary endpoints
The study will also provide a preliminary evaluation of the safety of PSV + sigh by comparing the incidence of the following adverse events in the two study groups: hemodynamic instability with hypotension (i.e., SBP < 90 mmHg) despite vasoactive drugs; arrhythmias with heart rate < 40 or > 140 bpm; radiographic evidence of barotrauma (i.e., pneumothorax, pneumomediastinum, pneumatocoele or subcutaneous emphysema or new chest tube placement.
Furthermore, it will quantify the prevalence of short-term (i.e., within 30 min) and long-term (i.e., within 24 h in the PSV + sigh group) sigh responders defined by improved oxygenation. Finally, it will assess the following clinical outcomes: 28-day mortality; ventilator-free days; PSV level, PEEP value and oxygenation index on day 1–3; number of days on assisted ventilation until day 28; use of rescue treatments; number of quadrants involved on chest x-ray on day 1–5; ICU and hospital length of stay and tracheostomies. The listed variables will be analyzed by comparing the two study groups considering, first, the whole study population and, then, only patients who improved oxygenation during the pre-randomization sigh test (sigh responders).
Data collection
At enrollment, before the sigh test we will anonymously collect patients’ demographic information (e.g., age, sex, height, weight), past (e.g., hypertension, chronic medications) and recent (e.g., etiology of the acute respiratory failure, days since intubation) medical history, severity of lung injury (e.g., ventilation setting, arterial blood gases, respiratory system compliance, diagnosis of acute respiratory distress syndrome (ARDS)) and of systemic involvement (e.g., presence of shock, number of organ failures) and finally, ventilation data (e.g., RR, Vt, PEEP, FiO2, PSV level).
In both groups, for the first 24 h, we will assess the SpO2/FiO2 ratio, RR and PSV tidal volume every 4 h to further characterize physiologic response to sigh over time.
From day 1 (i.e., within 24 h from randomization) to day 28 or death or discharge from the ICU, the following data will be collected every day between 6:00 and 10:00 in the morning: arterial SpO2, arterial and central venous blood gas analyses, ventilation settings and pattern (i.e., sigh pressure level, sigh tidal volume, PSV level, PSV tidal volume, respiratory rate, PEEP, FiO2, minute ventilation, P0.1, mean airway pressure), heart rate, arterial blood pressure, central venous pressure, cumulative fluid balance, patient’s comfort assessed using a visual analog scale (VAS), RASS value, SOFA score and dosage of sedative agents and vaso-active drugs.
Moreover, each day we will collect information on use of rescue treatments (i.e., recruitment manoeuvers, use of PEEP ≥ 15 cmH2O, prone positioning, inhaled nitric oxide, extracorporeal membrane oxygenation), tracheostomy, switch from the allocated treatment to the other study arm for ≥ 24 h, switch to controlled ventilation for ≥ 24 h, reason for switch to controlled ventilation, adverse events (i.e., hemodynamic instability with hypotension with SBP < 90 mmHg despite vasoactive drugs; arrhythmias; radiographic evidence of barotrauma with pneumothorax, pneumomediastinum, pneumatocoele, or subcutaneous emphysema), SBT performance and reason for SBT failure, extubation or separation from mechanical ventilation, reintubation and reason for reintubation with time. We will collect data on day 28 for all enrolled patients on mortality, extubation that occurred outside the ICU and readmission to ICU.
Statistical considerations
Statistical support for the PROTECTION trial will be granted by collaboration with Dr Carla Fornari and Dr Sara Conti from the Research Center on Public Health, Department of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy. Drs Fornari and Conti are PhD statisticians with experience in public health and physiologic studies and they participated to PROTECTION study conception, draft of the study protocol, performed the power analysis to decide sample size and prepared the data analysis plan.
Sample size
The sample size was computed based on the hypothesis that PSV + sigh might decrease the rate of failure of assisted ventilation compared to standard PSV. Based on previous data [21], the expected proportion of failure in patients undergoing PSV will be 22% and we hypothesized a proportion of 15% for patients in the PSV + sigh group. Assuming non-inferiority of PSV + sigh treatment, with a tolerance of 5% we estimated that a sample size of 258 patients (with 129 patients per study arm) will be sufficient to assess feasibility of the PSV + sigh strategy in this pilot phase with power of 0.8 and alpha 0.05.
Statistical analysis
Continuous normally distributed variables will be expressed by their mean and standard deviation or as medians and their interquartile ranges when not normally distributed. Categorical variables will be expressed as number and proportion (%). To assess the feasibility of PSV + sigh as compared to PSV (i.e., the primary endpoint), we will compute the proportion of patients experiencing at least one of the following events in each arm: switch to controlled ventilation for ≥ 24 h consecutively; use of PEEP ≥ 15 cmH2O, prone positioning, inhaled nitric oxide, extracorporeal membrane oxygenation or reintubation within 48 h from extubation.
We will compare the two proportions using the one-tailed non-inferiority test for proportions with a 5% tolerance. We will use 0.05 as the significance level. The same method will be adopted to evaluate the safety of PSV + sigh, which is a secondary endpoint. Prevalence of sigh responders in the two arms will be compared using the chi-squared or Fisher’s exact test. Ventilator-free days will be calculated as the difference between the duration of follow up (up to 28 days) and the number of days of intubation before successful extubation or separation from mechanical ventilation for tracheostomized patients. For all the secondary endpoints, we will use the chi-squared or Fisher’s exact test in the case of binary variables or the t test or Wilcoxon signed rank test for continuous variables. Time to death or extubation will be analyzed using Kaplan-Meier survival curves. Adjustement for potential confounding factors will be evaluated using appropriated regression models.
Study organization
The steering committee is composed of two principal investigators (TM and LB) plus seven international experts on ventilation who contributed to the design and revision of the study protocol. Each participating center will indicate a local investigator in charge of the study. The principal investigators are responsible for administrative management and communication with the local investigators and for helping the participating clinical sites in trial management, record keeping and data management. The local investigators provide structural and scientific leadership. They guarantee the integrity of data collection and ensure timely completion of the case report forms.
Collected data will be entered in an electronic case report form (eCRF) available online at a dedicated website (https://sighprotection.digitalcrf.eu), with protected individual access for each participating center. Patient data will be anonymous and coded according to a number. The eCRF includes tools to promote data quality, such as range checks for data values. Data monitoring will be performed by means of queries on the database done by statisticians and analyzed to identify abnormalities and inconsistencies.
Based on clinical experience and results of physiologic studies, the additional risks for patients enrolled in the PSV + sigh arm is expected to be minimal in comparison to standard of care. Still, specific patient insurance will be granted to cover all unexpected adverse events caused by the study interventions and all adverse events will be monitored by and reported to the coordinating center.