Protocol for the Prone and Oscillation Pediatric Clinical Trial (PROSpect)

Abstract

Objectives:

Respiratory management for pediatric acute respiratory distress syndrome (PARDS) remains largely supportive without data to support one approach over another, including supine versus prone positioning (PP) and conventional mechanical ventilation (CMV) versus high-frequency oscillatory ventilation (HFOV).

Design:

We present the research methodology of a global, multicenter, two-by-two factorial, response-adaptive, randomized controlled trial of supine versus PP and CMV versus HFOV in high moderate-severe PARDS, the PRone and OScillation PEdiatric Clinical Trial (PROSpect, www.ClinicalTrials.gov, NCT03896763).

Setting:

Approximately 60 pediatric intensive care units (PICUs) with on-site extracorporeal membrane oxygenation support in North and South America, Europe, Asia, and Oceania with experience using PP and HFOV in the care of patients with PARDS.

Patients:

Eligible pediatric patients (aged ≥2 weeks and <21 years) are randomized within 48 hours of meeting eligibility criteria occurring within 96 hours of endotracheal intubation.

Interventions:

One of four arms, including supine/CMV, prone/CMV, supine/HFOV, or prone/HFOV. We hypothesize that children with high moderate-severe PARDS treated with PP or HFOV will demonstrate ≥2 additional ventilator-free days (VFD).

Measurements and Main Results:

The primary outcome is VFD through day 28; non-survivors receive zero VFD. Secondary and exploratory outcomes include nonpulmonary organ failure-free days, interaction effects of PP with HFOV on VFD, 90-day in-hospital mortality and, among survivors, duration of mechanical ventilation, PICU and hospital length of stay, and post-PICU functional status and health-related quality of life. Up to 600 patients will be randomized, stratified by age group and direct/indirect lung injury. Adaptive randomization will first occur 28 days after 300 patients are randomized and every 100 patients thereafter. At these randomization updates, new allocation probabilities will be computed based on intention-to-treat trial results, increasing allocation to well-performing arms and decreasing allocation to poorly performing arms. Data will be analyzed per intention-to-treat for the primary analyses and per-protocol for primary, secondary, and exploratory analyses.

Conclusions:

PROSpect will provide clinicians with data to inform the practice of PP and HFOV in PARDS.

Pediatric acute respiratory distress syndrome (PARDS) is associated with high morbidity and mortality(1, 2). Despite its clinical significance in critically ill mechanically ventilated children, respiratory management remains largely supportive (3) with little data to support one approach over another, as recently underscored in the 2023 update to the Pediatric Acute Lung Injury Consensus Conference (PALICC-2) statements(4). Notably, in PALICC-2, no recommendations for or against prone positioning (PP) and high-frequency oscillatory ventilation (HFOV) were made(4–6). Prone positioning improves oxygenation and has few serious adverse events(5). In adults, the 2008–2011 Proning Severe ARDS Patients (PROSEVA) randomized clinical trial (RCT) found that prone positioning significantly improved outcomes in severe ARDS (i.e., PaO₂/FiO₂ <150 mmHg)(7). In adults, based on aggregated and individual patient data meta-analysis, use of prone positioning is now recommended for patients with moderate-severe ARDS, defined as PaO₂/FiO₂ <150 mmHg and PEEP ≥5 cm H₂O, despite optimization of ventilation settings(8). Regarding HFOV, the 2007–2012 OSCillation for ARDS (OSCAR) trial found no difference in 30-day mortality(9). In contrast, the 2007–2012 OSCILLation for ARDS Treated Early (OSCILLATE) trial was prematurely stopped because of higher in-hospital (47% versus 35%) and 60-day mortality (47% versus 38%) in the HFOV group(10).

To date, pediatric data are inconclusive on the use of PP and HFOV in PARDS(3), and there is wide practice variation in clinical management(11). Only one pediatric RCT compares prone to supine positioning(12). This 2001–2004 study of 102 patients with acute lung injury was stopped for futility, but trial drawbacks were that it was not limited to high moderate-severe PARDS and co-interventions, such as HFOV, were mandated. Additionally, a meta-analysis of pediatric RCTs and observational studies through December 2020 showed that HFOV was not associated with improved outcomes including mortality (risk ratio [95% confidence interval] 0.93 [0.72–1.20]) relative to conventional mechanical ventilation (CMV)(13). However, the question is whether the results of the pediatric (and adult) RCTs confirm that HFOV is not beneficial, and its use should therefore be discouraged or if it is a matter of how the oscillator was used that determined patient outcomes(14, 15). In this report, we describe the methodology and protocol for the PROSpect (Prone and Oscillation Pediatric Clinical Trial) study testing supine versus PP and CMV versus HFOV in high moderate-severe PARDS.

METHODS

PROSpect tests the hypothesis that children with high moderate-severe PARDS treated with PP or HFOV will demonstrate ≥2 more ventilator-free days (VFD) through day 28. The trial is registered with ClinicalTrials.gov (October 2018; NCT03896763), and recruitment started May 2019 with anticipated completion in 2026.

Primary objectives are to compare the effects of PP versus supine positioning as well as HFOV versus CMV on VFD. The secondary objective is to compare the impact of these interventions on nonpulmonary organ failure-free days. Exploratory objectives include evaluating the interaction effects of PP with HFOV on VFD and investigating the impact of these interventions on 90-day in-hospital mortality and, among survivors, duration of mechanical ventilation (MV), pediatric ICU (PICU) and hospital length of stay (LOS), and post-PICU functional status and health-related quality of life (HRQL). PROSpect also collects and stores blood and endotracheal aspirate samples for future studies of genomics, proteomics, and metabolomics.

Design and population

PROSpect is a global, multicenter, two-by-two factorial, response-adaptive RCT in pediatric patients ≥2 weeks of age (≥42 weeks post gestational age) and <21 years of age with high moderate-severe PARDS per the first Pediatric Acute Lung Injury Consensus Conference (PALICC-1)(16) consensus recommendations (i.e., chest imaging consistent with acute pulmonary parenchymal disease and oxygenation index [OI] ≥12 or oxygen saturation index [OSI] ≥10) (Figure 1). The research design is efficient testing the efficacy of two interventions that are commonly used together – prone positioning and HFOV – for high-moderate to severe PARDS. Data generated during the trial will be used to modify randomization allocation assigning more patients to a more efficacious intervention(s). Patients must be intubated and undergoing MV for <96 hours and meet oxygenation criteria for <48 hours. Inclusion criteria include two OI ≥12 (or OSI ≥10) separated by 2–6 hours or one OI ≥16. We operationalize rapid improvement as meeting HFOV to CMV conversion criteria within 6 hours of qualifying for PROSpect; specifically, mPaw ≤20 cm H₂O and FiO₂ < 0.50 with PEEP per PEEP/FiO₂ grid. Excluded are patients with underlying disorders that preclude PROSpect’s oxygenation or ventilation targets, already receiving PP or HFOV for >6 hours, or supported on extracorporeal membrane oxygenation (ECMO). Complete eligibility criteria are shown in eTable 1. Prior to consent, the treating physician is required to attest that all study arms are reasonable for the care of the patient.

Figure 1:

PROSpect study scheme

Infinity sign reflects randomization updates

Definition of abbreviations: CMV = conventional mechanical ventilation; HFOV = high-frequency oscillatory ventilation; HRQL = health-related quality of life; OFFD = organ failure-free days; PARDS = pediatric acute respiratory syndrome; VFD = ventilator-free days.

Human subjects research sites and approvals

Approximately 60 PICUs with on-site ECMO support in North and South America, Europe, Asia, and Oceania with experience using PP and HFOV in the care of patients with PARDS are participating. PROSpect study procedures are in accordance with the ethical standards of the responsible committee on human experimentation and the Helsinki Declaration of 1975. The University of Pennsylvania (PENN) institutional review board (IRB) serves as the IRB of record or reviewing Single IRB (sIRB) (Prone and Oscillation Pediatric Clinical Trial; #831295; 9/20/18). PENN conducts the ethical review for all U.S. sites under a reliance agreement. International sites execute regulatory review and approval to conduct human subject research according to national and institutional guidelines.

The parent or legal guardian provides written informed consent. Prior to hospital discharge, children ≥8 years of age who are cognitively capable are asked to provide assent for follow-up using age-appropriate assent/consent forms. Adhering to international norms, Canada allows the use of deferred consent until the clinical team determines that the parent or legal guardian can participate in the consent process.

Study interventions

Once randomized, patients are transitioned to their allocated positioning then ventilation intervention within 4 hours (Table 1). Patients are managed per assigned arm until weaned from MV or protocol failure is declared. Essential protocol elements are outlined in eTable 2. Ventilator strategy failure is defined as a 2-hour pattern of: 1) persistent hypoxemia (pulse oximetry saturation [SpO₂] <85%) with fractional inspired oxygen (FiO₂) of 1.0 and either maximum positive end-expiratory pressure (PEEP) as per the PEEP-FiO₂ grid (CMV) or mean airway pressure (mPaw) >35 cm H₂O (HFOV); or 2) persistent hypoventilation (with pH <7.15) with peak inspiratory pressure (PIP) >32 cm H₂O and a respiratory rate that does not cause intrinsic PEEP (CMV) or maximum power/amplitude at a frequency <8 Hz (HFOV). Positioning strategy failure is defined as persistent hypoxemia per ventilation mode or a three consecutive day pattern of no effect (prone). When protocol failure is declared, clinicians may consider an alternative therapy in a sequence based on their clinical judgment while considering ECMO cannulation. These treatments, when used, are managed according to the PROSpect protocols. Patients cannulated for ECMO are discontinued from study treatments but followed to describe ventilator management and study outcomes. Co-interventions, including endotracheal tube suctioning, hemodynamic management, sedation guidelines, and extubation readiness testing for all groups are managed per PALICC-1 recommendations(17–19).

Table 1.

PROSpect interventions

Intervention	Description
Supine	Supine position 24 hours/day
Prone	Prone position ≥16 hours/day
CMV	Includes 1) low tidal volume to obtain exhaled tidal volume of 5–7 mL/kg (ideal body weight); 2) PIP goal limited to ≤28 cm H2O; (3) lung recruitment maneuver to identify best PEEP then maintained per PEEP-FiO2 grid; and (4) use of SIMV or AC, PCV or PRVC, or equivalent. Protocols delineate ongoing CMV support, escalation of support, and weaning of support.
HFOV	Includes 1) frequency at 8–15 Hz; 2) amplitude (ΔP) 60–90 cm H2O; 3) mPaw titration maneuver Q12H; and use of the SensorMedics 3100A (patient <35 kg) or 3100B (patient ≥35 kg). Protocols delineate ongoing HFOV support, escalation of support, weaning of support, and conversion to CMV.

Definition of abbreviations: AC = assist control; CMV = conventional mechanical ventilation; FiO₂ = fraction of inspired oxygen; HFOV = high-frequency oscillatory ventilation; mPaw = mean airway pressure; PCV = pressure control ventilation; PEEP = positive end-expiratory pressure; PIP = peak inspiratory pressure; PRVC = pressure regulated volume control; SIMV = synchronized intermittent mandatory ventilation.

Randomization

Patients are randomized to one of four arms. The randomization process is managed centrally. For the first 300 randomized patients, the allocation is 1:1:1:1 stratifying by age (<1; 1–7; 8–17; 18–20 years) and direct/indirect lung injury (8 strata in total). Stratification will allow for the balance of potentially important subgroups among the intervention arms. Randomization will occur in permuted blocks with random block sizes of 4 and 8. Randomization update analyses will first occur 28 days after 300 patients are randomized and every 100 patients thereafter. For these analyses, the Bayesian posterior probabilities of each arm (i.e., highest median VFD based on ongoing intention-to-treat trial results) will be used to calculate new allocation probabilities, with increased allocation to well-performing arms and decreased allocation to poorly performing arms. Furthermore, there is potential permanent dropping of arms or early stopping of the trial due to efficacy or futility at these time points based on Bayesian posterior probabilities at the current or maximum sample size. Balance among the age and lung injury strata will be maintained in the response-adaptive randomization phase by the method described by Saville and Berry(20), using simple randomization within each stratum.

Clinical outcomes

PROSpect outcomes are summarized in Table 2. The primary outcome is VFD through day 28 with VFD defined as the number of days within 28 days that a patient is alive and free of MV(21). In computing VFD, day 0 is the time of endotracheal intubation. Duration of MV continues until the first time the endotracheal tube is continuously absent for at least 24 hours. Patients will be assigned zero VFD if they do not achieve the qualification for end of MV or died prior to day 28. To accommodate the use of non-invasive ventilation (NIV; i.e., continuous or bi-level ≥5 cm H₂O or humidified high-flow nasal cannula ≥5 L/min) as a separate outcome, the total duration of assisted breathing, which includes the use of NIV pre-intubation and post-extubation, will be computed.

Table 2.

PROSpect outcomes

Outcome	Definition
Primary
Ventilator-free days through Day 28	• VFD is defined as the number of days within 28 days that a patient is alive and free of mechanical ventilation. Patients will be assigned zero VFD if they do not achieve the qualification for end of mechanical ventilation or died prior to day 28.
Secondary
Nonpulmonary organ failure-free days through Day 28	• Nonpulmonary OFFD is defined as the number of days within 28 days that a patient is alive and free of clinically significant nonpulmonary organ failure. Nonpulmonary OFFD will be calculated for the clinically important nonpulmonary organ systems (neurologic, cardiovascular, renal, and hematologic) using nonpulmonary PEdiatric Logistic Organ Dysfunction-2 (PELOD-2) scores to Day 28.
Exploratory
Interaction effects between the positioning and ventilation strategies	• An evaluation of the interaction effects of prone positioning with high-frequency oscillatory ventilation on VFD.
90-day in-hospital mortality	• Deaths from all causes will be monitored through hospital discharge or day 90 (whichever occurs first).
Duration of mechanical ventilation (among survivors)	• Duration of mechanical ventilation is defined as the time from endotracheal intubation (day 0) to the first time the endotracheal tube is continuously absent for at least 24 hours. • For patients with tracheostomies, duration of mechanical ventilation is defined as the time of initiation of assisted breathing to the first time positive pressure is <5 cm H2O (continuous or bi-level) for at least 24 hours. • For ventilator-dependent patients, duration of mechanical ventilation is defined as the time the patient exhibits a second PROSpect-qualifying OSI of 10 to the first time that the patient maintains the prespecified OSI criteria to initiate extubation readiness testing (i.e., OSI ≤6) for 24 hours. • Duration of mechanical ventilation will be considered to be 28 days for patients still intubated on day 28, or in patients with tracheostomies those still supported on positive pressure ≥5 cm H2O, or in ventilator dependent patients those with OSI >6.
PICU and hospital length of stay (among survivors)	• PICU LOS is defined as the time from day 0 to the time of PICU discharge. • Hospital LOS is defined as the time from day 0 to the time of hospital discharge. • PICU and hospital LOS will be considered to be 90 days for patients still in the PICU/hospital on day 90.
Post-PICU discharge functional status (among survivors in the U.S.)	• Pediatric Cerebral Performance Category (PCPC) • Pediatric Overall Performance Category (POPC) • Functional Status Scale (FSS) score
Post-PICU health-related quality of life (among survivors in the U.S.)	• Pediatric Quality of Life Inventory (PedsQL™; Version 4.0 Generic Core Scales for patients ≥2 years of age; Infant Scales for patients <2 years of age).

Definition of abbreviations: LOS = length of stay; OFFD = organ failure-free days; OSI = oxygen saturation index; PICU = pediatric intensive care unit; U.S. = United States; VFD = ventilator-free days.

The secondary outcome is nonpulmonary organ failure-free days through day 28, defined as the number of days within 28 days that a patient is alive and free of clinically significant nonpulmonary organ failure calculated using PEdiatric Logistic Organ Dysfunction-2 (PELOD-2) scores(22). For the clinically important nonpulmonary organ systems (i.e., neurologic, cardiovascular, renal, and hematologic), any score >0 will be considered failure.

Exploratory outcomes include the interaction effects between the positioning and ventilation strategies on VFD, 90-day in-hospital mortality, and, among survivors, duration of MV, PICU and hospital LOS, and post-PICU functional status and HRQL. English or Spanish-speaking U.S. families participating in the post-PICU follow-up study will complete functional status and HRQL measures(23–26) at 1-, 3-, 6-, and 12-months post-PICU discharge.

Trial oversight

Standard operating procedures governing PROSpect’s study organization and responsibilities along with adverse event and unanticipated problem monitoring and reporting are included as supplemental digital content eAppendices 1 and 2. Of note, a NHLBI-appointed independent data and safety monitoring board (DSMB) is responsible for safeguarding the interests of study participants, assessing the safety and efficacy of study procedures, ensuring data quality, reviewing interim and final data, and monitoring overall study conduct. This includes review of the initial protocol and proposed protocol modifications and assessment of the primary manuscript for adherence to the protocol and truthfulness of data. The DSMB meets at least every 6 months and is required to provide recommendations about starting, continuing, and stopping the study.

Statistical analysis

There are two primary outcome analyses, one each for the positioning strategy and ventilation strategy. For positioning, analysis of the primary outcome, VFD, will be performed by intention-to-treat using a stratified Wilcoxon rank-sum test, adjusting for ventilation strategy. Similarly, for ventilation, analysis of the primary outcome will be performed by intention-to-treat using a stratified Wilcoxon rank-sum test, adjusting for positioning strategy. Differences between positioning or ventilation strategies will be considered statistically significant if a one-sided p-value is <0.022. This threshold was obtained by simulation to control Type I error at the 0.025 level, given the response-adaptive randomization design. Analysis of the primary outcome will also be performed on a per-protocol basis. The per-protocol data set will consist of all randomized subjects except subjects who were never transitioned to their allocated positioning and ventilation strategies, subjects withdrawn from the protocol during the first 24 hours post-randomization by a clinician or parent/legal guardian and subjects whose parent/legal guardian withdrew full consent for the protocol and data collection. Additionally, adjusting for age group and lung injury type will be explored using proportional hazards regression. We will also analyze VFD using competing risks regression, with time to successful extubation and time to death within 28 days as competing events. Analysis of the secondary outcome will be performed per-protocol received using stratified Wilcoxon rank-sum tests. Analyses of the exploratory outcomes will be performed using logistic, proportional hazards, and linear regression for binary, time-to-event, and continuous outcomes, respectively. For the secondary and exploratory outcomes, differences between positioning or ventilation strategies will be considered statistically significant if the two-sided p-value is <0.025. Appropriate methods for longitudinal outcomes, including random effects models or generalized estimating equations, will be used to model repeated measures outcomes from the follow-up aspect of the study. The PROSpect Statistical Analysis and Adaptive Design Plan is included as supplemental digital content eAppendix 3.

Sample size estimation

Sample size calculations are based on simulations using the data from the Randomized Evaluation of Sedation Titration for Respiratory Failure (RESTORE) trial(27) for a similar group of patients expected to be enrolled in the PROSpect trial. Of 2,449 RESTORE patients, 712 patients had severe PARDS with bilateral disease by the fourth day of intubation and were not intubated for asthma/reactive airways disease or bronchopulmonary dysplasia nor supported on ECMO. The mean VFD for these patients was 16.0 days with 14.2% patients assigned zero VFD (i.e., died or still intubated by day 28). Based on the full design including response-adaptive randomization and early stopping for a maximum total sample size of 600 patients, simulation results (based on 20,000 simulations using the RESTORE dataset) showed 73.1% power for detecting a clinically meaningful 2 VFD improvement(28) when one of the strategies was effective and 77.6% power for detecting either strategy as beneficial if both of the strategies were effective. In several scenarios, the total expected sample size was closer to 500 patients; specifically, 476 for the null scenario (i.e., no difference in VFD for positioning or ventilation strategies), 498 for the scenario where there is a superior positioning strategy or a superior ventilation strategy, and 507 for the scenario where there is both a superior positioning strategy and a superior ventilation strategy. In addition, response-adaptive randomization allocates more study patients to the superior intervention(s). For example, simulation results showed that approximately 61% of patients (304 of 498 on average) were assigned to the more efficacious intervention when one of the strategies was effective.

DISCUSSION

Research evidence is needed to bring about any major change in clinical practice. In PARDS, most of the 2015 PALICC-1 recommendations(16) and updated statements from the 2023 PALICC-2 report(4) are supported by low quality of evidence. This article presents the protocol and design of PROSpect, the first large-scale, multicenter RCT of interventions to improve clinical outcomes for PARDS. PROSpect aims to provide a definitive answer to the role of PP and HFOV for children with PARDS. The global nature of this investigation will improve generalizability and international implementation of the interventions.

PROSpect has several unique features related to its design and methodology (Tables 3 and 4)(7, 9, 10, 12, 29, 30). We designed the prone positioning protocol in PROSpect to mirror the PROSEVA trial. Additionally, an individual patient data meta-analysis of adult HFOV trials (publications through 2013) showed that HFOV was associated with improved outcomes for patients with severe hypoxemia (i.e., PaO₂/FiO₂ <100 mmHg)(14). Driven by these observations, we designed our trial to include high-moderate to severe PARDS, defined by OI ≥12 or OSI ≥10. Unlike the OSCAR and OSCILLATE trials, PROSpect uses a physiologic-based HFOV approach with individualized mPaw titration (instead of a standardized recruitment maneuver [RM] and mPaw/FiO₂ grid) and higher frequencies (i.e., ≥8 Hz) to maximize lung volume and deliver the smallest tidal volume for each patient. The response to a lung volume optimization maneuver in pediatric HFOV is highly variable because of the heterogeneity observed in PARDS lung mechanics(15). Consequently, the RM used in, for example, the OSCILLATE trial (i.e., sustained inflation of 40 cm H₂O for 40 seconds) ignores this heterogeneity in lung mechanics and may lead to an under- or over-recruited lung with subsequent desaturation of hemodynamic instability((31). Therefore, we decided to implement the staircase incremental-decremental mPaw titration to map the pressure-volume loop and identify the “optimal” mPaw after the RM. In the 1990–1994 pediatric trial by Arnold et al.(29), only an incremental mPaw titration was used. However, limiting the RM to only the incremental phase prevents oscillating on the deflation limb of the pressure-volume loop, making use of hysteresis of the lungs. We also chose not to use an mPaw/FiO₂ table as occurred in the OSCILLATE trial, since there are no data supporting a relationship between mPaw values and oxygenation in PARDS, especially considering the heterogeneity in lung volume behavior. For our HFOV protocol, we use higher frequencies starting at 10–12 Hz with 8 Hz as the lowest acceptable frequency. This choice was driven by the concept of corner frequency (Fc), which states the pressure-cost of ventilation is the lowest at a certain frequency defined by lung mechanics(32). For patients with disease conditions with reduced compliance, Fc is usually >8 Hz. We have previously reported that our physiology-based approach is safe in terms of hemodynamics and feasible in terms of oxygenation and ventilation(33). Failure criteria were chosen to align with those used in clinical decision-making to transition a patient to ECMO. Lastly, PROSpect is the first clinical trial in PARDS uses adaptive randomization in a population that most likely benefits from escalation in interventions, such as prone positioning and HFOV.

Table 3.

Unique features of PROSpect compared with previous prone positioning randomized controlled trials

Feature	PROSEVA	Pediatric Prone	PROSpect
Setting	• Only experienced centers	• Experience not required	• Only experienced centers
Study entry criteria	• Adults (≥18 years) • PaO2/FiO2 <150 mmHg • FiO2 ≥0.60 • PEEP ≥5 cm H2O • Vt close to 6 ml/kg predicted body weight	• Children (2 weeks to <18 years) • PaO2/FiO2 <300 mmHg	• Children (2 weeks to <21 years) • OI ≥12 or OSI ≥10
Timing of intervention	• Within 36 hours of ARDS diagnosis	• Within 48 hours of ALI diagnosis	• Within 48 hours of PARDS diagnosis
Duration of prone positioning	• 16 hours/day to day 28	• 20 hours/day to day 7	• 16 hours/day to day 28
Abdominal support	• No	• Yes	• If <8 years of age
Mechanical ventilation	• Protocolized, LPV	• Protocolized, LPV ± HFOV	• CMV/HFOV randomized
Enrollment dates	• 1/2008 to 7/2011	• 8/2001 to 4/2004	• 5/2019 to present

Definition of abbreviations: ALI = acute lung injury; ARDS = acute respiratory distress syndrome; CMV = conventional mechanical ventilation; FiO₂ = fraction of inspired oxygen; HFOV = high-frequency oscillatory ventilation; LPV = lung protective ventilation; OI = oxygenation index; OSI = oxygen saturation index; PaO₂ = partial pressure of oxygen in arterial blood; PARDS = pediatric acute respiratory distress syndrome; PEEP = positive end-expiratory pressure; PROSEVA = Proning Severe ARDS Patients; Vt = tidal volume.

Table 4.

Unique features of PROSpect compared with previous HFOV randomized controlled trials

Feature	OSCAR	OSCILLATE	Arnold et al.	El-Nawawy et al.	PROSpect
Setting	• Little experience	• Only experienced centers	• Only experienced centers	• Experienced center (single-center study)	• Only experienced centers
Device	• Novalung	• SensorMedics 3100B	• SensorMedics 3100A/B	• Fabian	• SensorMedics 3100A/B
Study entry criteria	• Adult (≥16 years) • PaO2/FiO2 <200 mmHg • PEEP >5 cm H2O	• Adult (≥16 years) • PaO2/FiO2 <200 mmHg • PEEP >10 cm H2O	• Pediatric • OI >13 and/or radiologic evidence of barotrauma	• Pediatric • PaO2/FiO2 ≤200	• Pediatric • OI ≥12 or OSI ≥10 • FiO2 ≥0.60
Recruitment strategy	• Not allowed	• Sustained inflation (40 cm H2O for 40 seconds)	• Staircase incremental mPaw recruitment maneuver	• Staircase incremental mPaw recruitment maneuver or sustained inflation	• Staircase incremental-decremental mPaw recruitment maneuver
Initial mPaw	• 5 cm H2O above plateau pressure on CMV	• 30 cm H2O	• Dependent on “optimal” mPaw during recruitment	• Dependent on lung inflation on chest radiograph	• Dependent on “optimal” mPaw during recruitment
mPaw adjustment	• Not specified	• mPaw/FiO2 table	• Gradual decrease in mPaw	• Gradual decrease in mPaw	• Individualized mPaw maneuver every 12 hours
Initial frequency	• 10 Hz	• 3–12 Hz	• 10–15 Hz	• 5–12 Hz	• 8–12 Hz
pH adjustment	• Cycling volume	• Frequency	• Frequency	• Amplitude	• Frequency
Control group	• Not protocolized, local practice	• Protocolized	• Protocolized	• Protocolized	• Protocolized
Enrollment dates	• 12/2007 to 7/2012	• 7/2009 to 8/2012	• 7/1990 to 1/1994	• 7/2011 to 4/2016	• 5/2019 to present

Definition of abbreviations: ARDS = acute respiratory distress syndrome; CMV = conventional mechanical ventilation; FiO₂ = fraction of inspired oxygen; HFOV = high-frequency oscillatory ventilation; mPaw = mean airway pressure; OI = oxygenation index; OSCAR = Oscillation in ARDS; OSCILLATE = Oscillation for ARDS Treated Early; OSI = oxygen saturation index; PaO₂ = partial pressure of oxygen in arterial blood; PEEP = positive end-expiratory pressure.

Limitations

Trial recruitment has not been immune to the effects of the COVID-19 pandemic on clinical research. The pandemic had a major impact on the capacity to screen, enroll, consent, and manage patients and launch new clinical sites. In addition, the COVID-19 pandemic also reduced the number of potentially eligible patients with the significant decrease in PICU census across the world from 2020 to 2022(34–36). eTable 3 outlines the strategies implemented to address COVID-19 related enrollment challenges at the patient, unit, and study level.

It may be argued that the level of protocolization in our trial comes at a cost of altering the utility of the answer the study can produce, thus potentially making it less generalizable to the highly variable real world without such a protocolized approach. However, the protocols utilized in this study are based on PALICC-2 practice recommendations for PARDS(4).

CONCLUSIONS

PROSpect is the first multicenter, two-by-two factorial, response-adaptive RCT of interventions designed to improve clinical outcomes for PARDS. Uniquely, the protocol is physiology-based in its use of PP and management of HFOV. Testing these interventions will help establish a standard of care that will influence the care of many pediatric patients supported on mechanical ventilation around the world.

“Research in Context”

• No pediatric-specific data support the use of prone positioning (PP) or high-frequency oscillatory ventilation (HFOV) in pediatric acute respiratory distress syndrome (PARDS).

• Pediatric Intensive Care Units around the globe have joined together to answer this clinically relevant question.

• PROSpect’s factorial design and response-adaptive randomization improve trial efficiency. Data generated during the trial will be used to modify randomization allocation assigning more patients to a more efficacious intervention(s).

“At the Bedside”

• Testing PP and HFOV will help to establish a standard of care that will influence the management of many critically ill pediatric patients supported on mechanical ventilation.

• The global nature of this investigation will contribute to the generalizability of the findings.