key: cord-0860628-moawnect
authors: Mauri, Tommaso; Foti, Giuseppe; Fornari, Carla; Grasselli, Giacomo; Pinciroli, Riccardo; Lovisari, Federica; Tubiolo, Daniela; Volta, Carlo Alberto; Spadaro, Savino; Rona, Roberto; Rondelli, Egle; Navalesi, Paolo; Garofalo, Eugenio; Knafelj, Rihard; Gorjup, Vojka; Colombo, Riccardo; Cortegiani, Andrea; Zhou, Jian-Xin; D’Andrea, Rocco; Calamai, Italo; González, Ánxela Vidal; Roca, Oriol; Grieco, Domenico Luca; Jovaisa, Tomas; Bampalis, Dimitrios; Becher, Tobias; Battaglini, Denise; Ge, Huiqing; Luz, Mariana; Constantin, Jean-Michel; Ranieri, Marco; Guerin, Claude; Mancebo, Jordi; Pelosi, Paolo; Fumagalli, Roberto; Brochard, Laurent; Pesenti, Antonio
title: Sigh in patients with acute hypoxemic respiratory failure and acute respiratory distress syndrome: the PROTECTION pilot randomized clinical trial.
date: 2020-11-13
journal: Chest
DOI: 10.1016/j.chest.2020.10.079
sha: 05999a5162355c32294da311c817cc17462f21e7
doc_id: 860628
cord_uid: moawnect

Background Sigh is a cyclic brief recruitment manoeuvre: previous physiological studies showed that its use could be an interesting addition to pressure support ventilation to improve lung elastance, decrease regional heterogeneity and increase release of surfactant. Research Question Is the clinical application of sigh during pressure support ventilation (PSV) feasible? Study Design and Methods We conducted a multi-center non-inferiority randomized clinical trial on adult intubated patients with acute hypoxemic respiratory failure or acute respiratory distress syndrome undergoing PSV. Patients were randomized to the No Sigh group and treated by PSV alone, or to the Sigh group, treated by PSV plus sigh (increase of airway pressure to 30 cmH2Ofor 3 seconds once per minute) until day 28 or death or successful spontaneous breathing trial. The primary endpoint of the study was feasibility, assessed as non-inferiority (5% tolerance) in the proportion of patients failing assisted ventilation. Secondary outcomes included safety, physiological parameters in the first week from randomization, 28-day mortality and ventilator-free days. Results Two-hundred fifty-eight patients (31% women; median age 65 [54-75] years) were enrolled. In the Sigh group, 23% of patients failed to remain on assisted ventilation vs. 30% in the No Sigh group (absolute difference -7%, 95%CI -18% to 4%; p=0.015 for non-inferiority). Adverse events occurred in 12% vs. 13% in Sigh vs. No Sigh (p=0.852). Oxygenation was improved while tidal volume, respiratory rate and corrected minute ventilation were lower over the first 7 days from randomization in Sigh vs. No Sigh. There was no significant difference in terms of mortality (16% vs. 21%, p=0.342) and ventilator-free days (22 [7-26] vs. 22 [3-25] days, p=0.300) for Sigh vs. No Sigh. Interpretation Among hypoxemic intubated ICU patients, application of sigh was feasible and without increased risk.

Sigh is a cyclic brief recruitment manoeuvre: previous physiological studies showed that its use could be an interesting addition to pressure support ventilation to improve lung elastance, decrease regional heterogeneity and increase release of surfactant.

Research Question. Is the clinical application of sigh during pressure support ventilation (PSV) feasible?

Study Design and Methods. We conducted a multi-center non-inferiority randomized clinical trial on adult intubated patients with acute hypoxemic respiratory failure or acute respiratory distress syndrome undergoing PSV. Patients were randomized to the No Sigh group and treated by PSV alone, or to the Sigh group, treated by PSV plus sigh (increase of airway pressure to 30 cmH 2 Ofor 3 seconds once per minute) until day 28 or death or successful spontaneous breathing trial. The primary endpoint of the study was feasibility, assessed as non-inferiority (5% tolerance) in the proportion of patients failing assisted ventilation. Secondary outcomes included safety, physiological parameters in the first week from randomization, 28-day mortality and ventilatorfree days.

Results. Two-hundred fifty-eight patients (31% women; median age 65 [54-75] years) were enrolled. In the Sigh group, 23% of patients failed to remain on assisted ventilation vs. 30% in the No Sigh group (absolute difference -7%, 95%CI -18% to 4%; p=0.015 for non-inferiority). Adverse events occurred in 12% vs. 13% in Sigh vs. No Sigh (p=0.852). Oxygenation was improved while tidal volume, respiratory rate and corrected minute ventilation were lower over the first 7 days from randomization in Sigh vs. No Sigh. There was no significant difference in terms of mortality (16% vs. 21%, p=0.342) and ventilator-free days (22 [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] vs. 22 Multiple physiological studies showed that use of sighs could be an interesting addition to pressure support ventilation. Sigh may improve lung function through improved lung elastance 4 , decreased regional heterogeneity 5 ,increased release of active surfactant 6 and decreased effort 5 , the latter being protective also for the diaphragm. Moreover, sigh has been shown to allow a reduction in tidal volume and respiratory rate, reducing the ventilation load applied to the lungs 4, 5, 7 . These studies generated the hypothesis that addition of sigh to PSV might improve clinical outcomes of patients with AHRF and ARDS. However, no randomized clinical trial (RCT) on sigh addition to PSV has ever been performed, and, before conducting a larger trial aimed at verifying improved survival, we first conceived a pilot RCT to verify the clinical feasibility of sigh in comparison to standard PSV 8 and to have preliminary estimates of adverse events, lost to followup, outcomes and its variabilities. A non-inferiority approach was chosen to demonstrate that application of sigh in the clinical setting is as feasible as standard PSV, which is the most widely adopted assisted ventilation mode.

In the present trial, sigh was applied early after switching to PSV in intubated AHRF or ARDS patients and maintained until successful weaning, death or day 28. The study aimed attesting the non-inferiority of sigh, as compared to standard PSV without sigh, in terms of failure of assisted ventilation. Failure was defined as the occurrence of any of these conditions: switch back to controlled ventilation; use of rescue therapies for refractory hypoxemia; re-intubation.

Secondary outcomes included comparison between the two study arms in the incidence of adverse events, physiological parameters, survival and ventilator-free days. Richmond Agitation-Sedation Scale (RASS) 9 value at enrolment had to be between -2 and 0.

Exclusion criteria can be found in the Online Supplement 1 (Page 2).

After enrollment, all patients underwent a 30-minute test of addition of Sigh to clinical PSV to assess the prevalence of sigh responders vs. nonresponders as defined by improved oxygenation. Briefly, the ventilator inspired oxygen fraction (FiO 2 ) was titrated to obtain a peripheral oxygen saturation (SpO 2 ) of 90-96%, while keeping the same clinical PEEP and PSV levels. Then, sigh was added as a pressure control phase set at total end-inspiratory pressure of 30 cmH 2 O for 3 seconds insufflation time, once per minute. At the beginning and after 30 minutes, the SpO 2 /FiO 2 ratio was collected. Based on previous physiological study, the expected prevalence of sigh responders (i.e., patients improving SpO 2 /FiO 2 by >1%) was estimated to be 50% 5 .

After completion of the sigh test, patients were randomized by a 1:1 ratio to a strategy of PSV titrated following a predefined protocol with addition of sigh (Sigh group) or to a strategy of PSV titrated following the same protocol but without Sigh (no Sigh group). The local investigators randomized patients using a central, dedicated, password-protected, web-based, automated randomization system. The randomization sequence was generated using a permuted blocks randomization scheme (block size of 6).

After randomization, in the Sigh group, PSV was targeted to a tidal volume (Vt) of 6-8 mL/Kg of predicted body weight (PBW), with respiratory rate (RR) 20-35 breaths/min (bpm) and clinical PEEP. FiO 2 was left as selected during the pre-randomization sigh test. Sigh was promptly added as J o u r n a l P r e -p r o o f a pressure control breath at total end-inspiratory pressure of 30 cmH 2 O for 3 seconds delivered once per minute. Ventilators were switched to biphasic synchronized positive airway pressure mode (also known as synchronized intermittent mandatory ventilation combining pressure control and PSV) with the lower pressure level set at clinical PEEP and the higher-pressure level set at 30 cmH 2 O with a 3-second inspiratory time. Sigh settings were left unchanged until switch to controlled ventilation, day 28, death or performance of a successful spontaneous breathing trial (SBT, see below). In the No Sigh group, after randomization, PSV was set to obtain the same targets as above with clinical PEEP and the FiO 2 selected during the pre-randomization sigh test.

Then, in both groups at least every 8 hours, the PSV level was adjusted to maintain Vt of 6-8 ml/kg PBW and RR of 20-35 bpm, while PEEP and FiO 2 were managed to keep SpO 2 of 90-96%.

In both groups, switch to protective controlled ventilation was indicated when patients developed specific pre-defined criteria 8 Criteria for success vs. failure of the SBT were pre-defined by study protocol 8 . Subjects successfully completing the SBT were promptly extubated or, in the presence of tracheostomy, mechanical ventilation was discontinued. Patients who failed the SBT were switched back to Sigh or No Sigh and criteria for SBT were checked again after at least 6 hours. After extubation, re-intubation was performed if at least one of the criteria pre-defined by the study protocol was present 8 .

Outcomes. The primary endpoint of this trial 8 was to assess non-inferiority of Sigh feasibility vs.

No Sigh by comparing the number of patients in each group experiencing at least one of the following criteria for failure of assisted ventilation: switch to controlled ventilation for ≥24 consecutive hours; use of rescue therapy; re-intubation within 48 hours.

Secondary outcomes included: comparison of selected physiological variables during the first 7 days from randomization in the two study groups; evaluation of the clinical safety of Sigh vs. No

Sigh by comparing incidence of pre-defined adverse events; quantification of responders and non-J o u r n a l P r e -p r o o f responders to the pre-randomization Sigh test; 28-day mortality and ventilator-free days in the two study groups and in responders and non-responders.

Statistical analysis. Based on previous data 10 , we computed that a sample size of 258 patients (with 129 patients per study arm) was sufficient to assess feasibility of the Sigh strategy (primary outcome) using a non-inferiority test with a tolerance of 5%, power of 0.8, alpha 0.05, 22% and 15% as the expected rate of failure of assisted ventilation in patients undergoing No Sigh and Sigh treatment, respectively. Failure of assisted ventilation in patients treated with Sigh was compared to patients with No Sigh using a one-tailed non-inferiority test for proportions with a 5% tolerance.

In details, non-inferiority of Sigh was established when failure in the Sigh group was lower than failure of No Sigh plus 5%. This is the standard alternative hypothesis for non-inferiority tests 11 .

Thus, in this study, p-value lower than 0.05 (type I error) for the non-inferiority test would reject inferiority of the new treatment (Sigh) compared to No Sigh. Survival at 28 th day was analysed using Kaplan Meier curves and Log-Rank test was used to test differences between curves. To test differences in time-trends of physiological and clinical parameters between the two study groups we used Generalized Estimating Equation (GEE) models to account for repeated measures.

The funding sources didn't have any role in the study design, data collection, analysis and interpretation, writing of the manuscript and decision to submit it.

One-thousand-sixty-four intubated ICU patients undergoing PSV were screened. A total of 806 were not enrolled, of whom 726 (90%) met at least one of the exclusion criteria and 80 (10%) were eligible but could not be enrolled for various reasons (Figure 1 ). Two-hundred-fiftyeight patients completed the Sigh test and were subsequently randomized, 129 to the Sigh group and 129 to the No Sigh group. None of the patients with drew consent after randomization. Sigh was applied for 4 [2] [3] [4] [5] [6] [7] [8] [9] days in the Sigh group. Follow-up until day 28 was complete for all patients. Baseline characteristics were well balanced between the two study groups (Table 1) . Males were 67% (87 patients) and 71% (92 patients) in Sigh group and in No Sigh group, respectively. Mean age of patients was 63±15 years, with no significant difference between groups. Prevalence of comorbidities and general severity at admission were comparable ( Table 1) . Prevalence of the diagnosis of ARDS was 46% in the Sigh and 53% in the No Sigh group, with non-significant difference ( Table 1) .

Outcomes. Twenty-eight days after randomization, 30 patients (23%) in the Sigh group vs. 39 (30%) in the No Sigh group (Table 2 ) experienced at least one criterion for failure of assisted ventilation were. Sigh treatment group was therefore non-inferior to No Sigh treatment group in terms of failure of assisted ventilation (absolute difference -7%; 95%CI -18% to 4%, p=0.015 for non-inferiority test) (Figure 2 ). Specific reasons for failure of assisted ventilation and type of rescue treatment are shown in Table 2 . Per-protocol analysis showed similar results with 29 (23%) patients failing to remain on assisted ventilation in the Sigh group vs. 37 (29%) in the No Sigh group (absolute difference -6%; 95%CI -17% to 5%, p=0.022 for non-inferiority test).

Adverse events (i.e. hemodynamic instability, arrhythmias and barotrauma) did not differ between the 2 study groups (16 (12%) patients in the Sigh group vs. 17 (13%) patients in the No Sigh group, p=0.852). Type of adverse events are described in Table 2 . Ventilator-free days were significantly higher in non-responders treated with Sigh vs. No Sigh (23 (9-26) vs. 10 (0-24) days, p=0.006).

Physiology. Over the first seven days from randomization, the PEEP level and set FiO 2 did not differ between groups. PaO 2 /FiO 2 ratio was significantly higher while respiratory rate, tidal volume and corrected minute ventilation (i.e., the minute ventilation multiplied by actual PaCO 2 divided by 40 mmHg, with lower values indicating higher efficiency to clear CO 2 by the respiratory system) were all significantly lower in the Sigh group (eTable 3 and eFigure 1 in the Online Supplement 1).

The tidal volume delivered by Sigh in the first seven days from randomization remained stable and around 15 ml/kg PBW (eFigure 2 in the Online Supplement 1). PaCO 2 and pH, RASS score and SOFA score were similar (eTable 3 and eFigure 1 in the Online Supplement 1).

This randomized clinical trial showed the feasibility of adding sigh to PSV: the rate of failure of assisted ventilation was non-inferior to conventional PSV. Secondary outcomes indicated safety of sigh with a similar rate of adverse events, a comparable mortality and number of ventilator-free days. Moreover, improved physiology was confirmed in the first week from randomization by addition of sigh.

Sigh is commonly performed during quiet breathing by healthy subjects, mainly acting as negative feedback on respiratory drive with positive functional and psychological consequences 12 . Many studies performed both in hypoxemic patients 13, 14 and in animal models of lung injury 15 showed that sigh is associated with improved physiology. Sigh induces recruitment of the collapsed lungs, restores surfactant production, decreases ventilation heterogeneity, improves regional mechanics, increases oxygenation and modulates the inspiratory effort 5, 16 . On the other hand, sigh cyclically delivers large inspiratory volumes in patients in whom current guidelines recommend mandatory reduction of tidal volume 1, 17 . Since no study existed on the feasibility and safety of long-term J o u r n a l P r e -p r o o f application of sigh to hypoxemic patients, it seemed important to conceive a large non-inferiority randomized controlled trial aimed at assessing the clinical feasibility and the safety of sigh.

The present trial indicates that addition of sigh to PSV leads to a rate of patients with acute hypoxemic respiratory failure or ARDS experiencing failure of assisted ventilation similar to traditional PSV. Moreover, number of adverse events were similar and low, with only 2 patients per group experiencing barotrauma; in only two patients sigh was stopped to continue with traditional PSV; mortality and ventilator-free days did not differ. Taken together, these results suggest that sigh could be added to PSV without causing any additional risk and yielding similar clinical outcomes in acute hypoxemic respiratory failure and ARDS patients. Possible explanations for these findings could be that sigh was not able to produce any clinical benefits in comparison to PSV alone; or that the non-significant difference in mortality showed in this trial might become significant in a study performed with the same protocol but with larger sample size.

Reduction of mortality with sigh in the subgroup of patients not responding in terms of oxygenation during a 30-minute sigh test performed before randomization is an additional intriguing finding that will require confirmation.

Assisted ventilation carries the intrinsic risk of additional patient self-inflicted lung injury (P-SILI) 18 and respiratory muscles myotrauma 19 , making lung and diaphragm protection a key clinical goal 20 .

Limiting the inspiratory volume and transpulmonary pressure is the recommended strategy for hypoxemic patients on PSV to minimize the risk of P-SILI 21, 22 . We confirmed that sigh improves oxygenation and decreases respiratory rate, tidal volume and minute ventilation during the first week, potentially decreasing the risk of additional P-SILI. As un-physiological high inspiratory pressure and volume leading to P-SILI increase the risk of prolonged ventilation and worse outcome 23 , the physiological analyses from this study might help generating a more solid hypothesis on the clinical effects of sigh.

Our results suggest that sigh is easy to implement and could be already seen as alternative ventilation mode for ICU physicians, even in resource-limited settings 24 .

Sigh can be delivered for longer time period (e.g., from intubation), at more physiological lower rate (e.g., once every other minute) and at different inspiratory pressure (e.g., personalized based on transpulmonary pressure) than in our study. Sigh isn't a general concept but rather a mechanical ventilation strategy with specific settings and variability in the delivery of Sigh may alter the results presented herein.

The present study has limitations. First, at enrolment, the patients were on mechanical ventilation since 3 [2] [3] [4] [5] days and sigh was applied only for approximately half the total number of days spent on mechanical ventilation. We can't answer as to whether application of sigh earlier and for longer time period might lead to increased benefits (from improved physiology) or harm (from higher risks of cyclic over-distension and atelectrauma). However, application of sigh during controlled ventilation requires specific machines and we reasoned that sigh has specific advantages in patients on assisted ventilation (e.g., modulation of effort). Second, we delivered sigh at the same total inspiratory pressure in all patients, which, based on predictable differences in respiratory mechanics, could have determined variable levels of transpulmonary pressure. Response to the pre-randomization Sigh test might have been influenced by this, too, with non-responders receiving insufficient volume. Personalized sigh settings based on specific patients' characteristics could lead to higher number of responders and improved outcomes. Third, the rate of sigh in this study was one per minute, while physiological studies suggested that lower rate may be more effective 5 . Once again, to our knowledge, only few ventilators can deliver sigh during PSV once every two minutes. Fourth, because of the nature of the intervention, physicians and nurses attending patients enrolled in the study could not be blinded. However, we provided detailed protocols for changes in PSV settings, performance of rescue therapies, spontaneous breathing trials, extubation and re-intubation 8 ,which should have limited biases in primary outcomes. Fifth, we defined sigh responders based on improvement of the SpO 2 /FiO 2 by >1% during the prerandomization sigh test. This threshold could be seen as too low for being clinically meaningful, however, the analysis was exploratory and higher threshold would have yielded large imbalances in groups numerosity.

Addition of sigh to PSV in patients with acute hypoxemic respiratory failure or ARDS is as feasible as traditional PSV in terms of failure of assisted ventilation, and yields comparable adverse events, mortality and ventilator-free days. Results from the present trial could inform planning and design of larger clinical trials aimed at verifying reduced mortality by application of sigh. Results. The study showed that in mechanically ventilated patients with acute hypoxemic respiratory failure or acute respiratory distress syndrome addition of Sigh in comparison to No Sigh during PSV was feasible and safe: no increase in patients failing to remain on assisted ventilation (23% vs. 30%, respectively) and similar proportion of adverse events (12% vs. 13%, respectively).

Interpretation. Addition of sigh to PSV is feasible and safe in intubated ICU patients with acute hypoxemic respiratory failure or acute respiratory distress syndrome. 

Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome

Spontaneous breathing: a doubleedged sword to handle with care

Patterns of Care, and Mortality for Patients With Acute Respiratory Distress Syndrome in Intensive Care Units in 50 Countries

Sigh improves gas exchange and lung volume in patients with acute respiratory distress syndrome undergoing pressure support ventilation

Effects of Sigh on Regional Lung Strain and Ventilation Heterogeneity in Acute Respiratory Failure Patients Undergoing Assisted Mechanical Ventilation

Morphologic evidence that large inflations of the lung stimulate secretion of surfactant

Sigh improves gas exchange and respiratory mechanics in children undergoing pressure support after major surgery

Pressure support ventilation + sigh in acute hypoxemic respiratory failure patients: study protocol for a pilot randomized controlled trial, the PROTECTION trial

The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients

Proportional assist ventilation with loadadjustable gain factors in critically ill patients: comparison with pressure support

Understanding equivalence and noninferiority testing

Van den Bergh O. A sigh following sustained attention and mental stress: effects on respiratory variability

Comparison of optimal positive end-expiratory pressure and recruitment maneuvers during lung-protective mechanical ventilation in patients with acute lung injury/acute respiratory distress syndrome

Effects of periodic lung recruitment maneuvers on gas exchange and respiratory mechanics in mechanically ventilated acute respiratory distress syndrome (ARDS) patients

Acute Lung Injury Causes Asynchronous Alveolar Ventilation That Can Be Corrected by Individual Sighs

Effects of sigh during pressure control and pressure support ventilation in pulmonary and extrapulmonary mild acute lung injury

Acute respiratory distress syndrome

Mechanical Ventilation to Minimize Progression of Lung Injury in Acute Respiratory Failure

Mechanical Ventilation-induced Diaphragm Atrophy Strongly Impacts Clinical Outcomes

Respiratory Drive in Critically Ill Patients. Pathophysiology and Clinical Implications

A novel non-invasive method to detect excessively high respiratory effort and dynamic transpulmonary driving pressure during mechanical ventilation

Spontaneous breathing during lung-protective ventilation in an experimental acute lung injury model: high transpulmonary pressure associated with strong spontaneous breathing effort may worsen lung injury

Impact of spontaneous breathing during mechanical ventilation in acute respiratory distress syndrome

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study