Hydrogen’s Feasibility and Safety as a Therapy in Extracorporeal Cardiopulmonary Resuscitation (Hydrogen-FAST): study protocol for a trial of inhaled hydrogen gas as an adjunctive therapy in refractory cardiac arrest

Abstract

Background

Extracorporeal cardiopulmonary resuscitation (ECPR) improves survival for patients experiencing refractory cardiac arrest but is often associated with severe ischemia-reperfusion injury (IRI). Hydrogen gas (H₂) has demonstrated promising preclinical efficacy in reducing IRI. The Hydrogen-FAST trial investigates the feasibility and safety of inhaled hydrogen gas as an adjunctive therapy during ECPR in patients with congenital heart disease (CHD).

Methods

This phase 1 multicenter randomized controlled trial enrolls patients of any age undergoing ECPR due to refractory cardiac arrest associated with underlying cardiac morbidity. Due to the emergent nature of ECPR events, the trial leverages an Exception from Informed Consent enrollment process. Fifty-three participants will be randomized to standard care or standard care plus inhaled 2% H₂ gas delivered continuously for 72 h via mechanical ventilator and extracorporeal membrane oxygenation (ECMO) circuit. Primary endpoints include feasibility, measured as the percentage of the first 72 h post-arrest in which hydrogen is successfully administered, with feasibility defined as a mean percentage of ≥90%, and safety, evaluated through incidence rates of treatment-related serious adverse events (SAEs). Secondary endpoints examine clinical outcomes including survival to discharge, ICU length of stay, functional status at 6 months, neurologic function, and markers of ischemic injury.

Discussion

The Hydrogen-FAST trial will provide essential data on the feasibility and safety of H₂ administration during ECPR, informing larger efficacy studies and potential broader clinical use in critical care settings.

Trial registration

ClinicalTrials.gov NCT05574296. Registered on October 6, 2022. https://clinicaltrials.gov/study/NCT05574296.

Administrative information

Primary registry and trial identifying number {4}	ClinicalTrials.gov, NCT05574296. Registered on 6 October 2022
Secondary identifying numbers	Not applicable
Source(s) of monetary or material support	National Heart, Lung, and Blood Institute (NHLBI), Grant numbers 1R33HL164373-01 and 1R61HL164373-01, and by the Mooney Family Foundation
Primary sponsor and contact information {3b}	National Heart, Lung, and Blood Institute jewel.joshua@nih.gov 301–827−1809
Role of sponsor and funder {3c}	The study sponsors had no role in or ultimate authority over the study design; collection, management, analysis, and interpretation of data; writing of the report; and the decision to submit the report for publication
Contact for public queries	https://research.childrenshospital.org/research-units/hydrogenfast-study-research hydrogenfast@childrens.harvard.edu
Contact for scientific queries	John N. Kheir, MD (john.kheir@cardio.chboston.org) and Lynn A. Sleeper, ScD (lynn.sleeper@cardio.chboston.org)
Public title	Hydrogen’s Feasibility and Safety as a Therapy in Extracorporeal Cardiopulmonary Resuscitation: Design of a Randomized Controlled Trial (Hydrogen-FAST Trial)
Scientific title	Hydrogen’s Feasibility and Safety as a Therapy in Extracorporeal Cardiopulmonary Resuscitation: Design of a Randomized Controlled Trial (Hydrogen-FAST Trial)
Countries of recruitment	USA
Health condition(s) or problem(s) studied	Cardiac arrest with resuscitation using extracorporeal cardiopulmonary resuscitation (ECPR)
Intervention(s)	Inhaled 2% hydrogen gas
Key inclusion and exclusion criteria	Inclusion criteria (all must be fulfilled) will be: (1) Patients admitted to the cardiac intensive care unit at a participating site with cardiac comorbidity, including congenital heart disease, myocarditis, cardiac arrhythmia, or rejection of a transplanted heart (2) Patient experiencing a cardiac arrest >5 min and receiving ongoing CPR in the ICU, cardiac catheterization lab, or cardiac operating room (3) The decision made by the clinical team to resuscitate using ECPR Exclusion criteria will be (any one disqualifies patient): (1) Enrollment in the opt-out program (2) Patients known to be pregnant (3) Patients who are prisoners (4) Prior ECPR episode during admission (5) Enrollment does not occur within 6 h of the decision to resuscitate using ECPR (6) Patients enrolled in the Trial of Indication-Based Transfusion of Red Blood Cells in ECMO (TITRE) within the last 18 months
Study type	Open-label randomized controlled trial
Date of first enrollment	29 March 2024
Sample size	56
Primary outcome(s)	Primary feasibility endpoint. We will compute the percentage of the first 72 consecutive post-arrest hours (starting at the time of first CPR initiation) in which H2 gas was administered via all of the applicable pathways (e.g., mechanical ventilator and ECMO membrane) Primary safety endpoint. The primary safety endpoint will be the incidence rate of SAEs of interest per day during the first 30 days post-randomization that have been classified as treatment-related or possibly treatment-related as determined by the Adverse Event Adjudication Committee
Key secondary outcome(s)	We will compute the percentage of the first 72 consecutive hours post-H2 initiation in which H2 gas was administered via all of the applicable pathways (e.g., mechanical ventilator and ECMO membrane) Survival to hospital discharge, ICU and hospital lengths of stay Functional status score (FSS) will be retrospectively computed at baseline (admission to the hospital and at 24 h before cardiac arrest), discharge (±3 days), and 6-month (±2 months) post-randomization Markers of ischemic injury, including lowest pH; highest lactic acid; lowest platelet count; and highest hepatic transaminases, blood urea nitrogen and creatinine, and international normalized ratio on post-randomization day 0, 1, 2, 3, 4, 5, 6, 7, 14, 21, and 28 Exploratory efficacy endpoints, including brain biomarkers (GFAP, Tau, and Neurofilament light) Findings on clinically obtained imaging (brain MRI, head CT, cranial ultrasound)
Ethics review	Approved by the Institutional Review Board at Boston Children’s Hospital (IRB-P00043374)
Individual trial participant data sharing statement	The sharing of deidentified individual clinical trial participant-level data is not applicable at this time

Introduction

Background and rationale {9a}

Extracorporeal cardiopulmonary resuscitation (ECPR) is known to improve outcomes in refractory cardiac arrest [1, 2] but is associated with significant ischemia-reperfusion injury (IRI), contributing to morbidity and mortality. Preclinical studies suggest hydrogen gas reduces IRI via selective neutralization of hydroxyl radicals and modulation of inflammatory pathways [3–5].

Explanation for the choice of comparator {9b}

Prior human studies indicate safety [6, 7] and efficacy [8] of hydrogen gas in the treatment of cardiac arrest. However, rigorous evaluations in critically ill populations are lacking. Participants in this trial receive either standard care plus 2% inhaled hydrogen gas or standard care alone.

Objectives {10}

We designed a multicenter, randomized controlled trial to assess the safety and feasibility of H₂ administration to infants and children receiving ECPR, titled Hydrogen’s Feasibility and Safety as a Therapy in ECPR (Hydrogen-FAST). This article describes the design of this trial.

Methods: patient and public involvement, and trial design

Patient and public involvement {11}

Community consultation

Prior to trial initiation at each center, and following single IRB (sIRB) approval, community consultation and public disclosure activities are carried out to ensure awareness and understanding of the trial among affected communities. Community consultation involves meetings with community representatives familiar with the cardiac ICU environment to discuss the study objectives, procedures, potential risks, and benefits of the research. Feedback from these consultations measures community attitudes towards EFIC, understanding of study risks and benefits, concerns about research participation, and recommendations for study procedures. A separate set of consultations takes place with staff members to obtain feedback and improve procedures for public disclosure and opt-out. Results from these consultations are summarized, analyzed qualitatively, and reported to the sIRB along with a public disclosure plan for approval for the commencement of the trial.

Public disclosure

Subsequent to sIRB approval, public disclosure takes place. Public disclosure plans may vary by site, based upon feedback from local community consultation meetings. For example, at Boston Children’s Hospital and Children’s National Hospital, public disclosure takes place through hospital websites, social media platforms, and press releases, providing detailed information about the trial and the EFIC process, along with mechanisms for opting out of the trial. Further, informational flyers are posted in waiting rooms of the intensive care unit and procedural areas and in each ICU bedspace.

Trial design {12}

This is a phase 1, open-label, randomized controlled trial conducted. Randomization (hydrogen to standard of care 2:1 ratio) allocates patients to receive either standard care plus 2% inhaled hydrogen gas or standard care alone. Randomization occurs at the time of the decision to cannulate the patient to ECMO using sealed envelopes stratified by age (<28 days old versus ≥28 days old) and center.

Trial setting {13}

This study is being conducted across multiple pediatric cardiac centers in the USA. The list of study sites is available on ClinicalTrials.gov (NCT05574296).

Eligibility criteria for participants {14a}

Inclusion criteria

• Patients admitted to a cardiac intensive care unit at a participating site with cardiac comorbidity, including congenital heart disease, myocarditis, cardiac arrhythmia, and rejection of a transplanted heart.

• Refractory cardiac arrest lasting >5 min and receiving ongoing CPR in the intensive care unit (ICU), cardiac catheterization lab, or cardiac operating room.

• A decision made to resuscitate using ECMO. Note that ECMO cannulation must occur within 20 min of the last chest compression to qualify as ECPR.

Exclusion criteria

• Enrollment in the opt-out program

• Known pregnancy

• Prisoners

• Prior ECPR during the same hospitalization

• Enrollment would occur beyond 6-h window post-decision for ECPR

• Patients enrolled in the Trial of Indication-Based Transfusion of Red Blood Cells in ECMO (TITRE) within the last 18 months

Eligibility criteria for sites and those delivering interventions {14b}

Sites with pediatric cardiac ICUs are eligible to participate. Sites must complete applicable protocol certification, data management training, clinical staff training, and IRB requirements before they may enroll patients. Sites must have an ECMO care team (e.g., respiratory therapists, nurses) who can assist with randomization procedures and administration of hydrogen gas.

Who will take informed consent? {32a}

Exception from informed consent (EFIC) and opt-out procedures

Given the emergent and life-threatening nature of ECPR, this study employs an Exception from Informed Consent (EFIC) method of enrollment (Fig. 1). The community consultation and public disclosure procedures, required as part of EFIC, are described earlier in this report.

Fig. 1.

The process of setting up and enrolling patients in the Hydrogen-FAST trial using Exception from Informed Consent includes six sequential steps. (1) A community consultation plan must be approved by both the FDA and single IRB (sIRB). (2) Following initial approval, a series of community consultation meetings take place with both patients and family members as well as staff. (3) Thereafter, comments and feedback are compiled and reviewed by the sIRB. (4) Following approval, a period of public disclosure takes place, after which enrollment can begin. Steps 1–4 are repeated at each participating center. (5) When an eligible patient is identified, a member of the care team checks the opt-out list to ensure the patient’s absence. (6) When deemed appropriate by the clinical staff (a minority of cases), a pre-enrollment opportunity to object or traditional informed consent may be provided to the LAR. (7) The patient is then randomized and enrolled. (8) When deemed appropriate by the clinical team, the LAR is provided an opportunity to object to ongoing participation by a member of the research team

Opt-out procedures

Individuals learning about the trial may opt out by contacting the study team via phone, email, or an online form (linked using a QR code on each informational flyer) specific to each participating center. If opt-out is requested, the patient’s information is added to a “Do Not Enroll” list that is located by the hydrogen tanks and randomization envelopes (Fig. 2). Inpatients who have opted out are also identified via a physical opt-out bracelet and identifying signage on the door. All opt-out methods must be checked and reconciled at least once daily. In the event of discrepant opt-out identifiers, staff will be trained to prioritize the presence or lack thereof of “No hydrogen” bracelets. Prior to enrollment, the enrolling care provider verifies that the patient is not on the opt-out list and is not wearing an opt-out bracelet; such patients are excluded from randomization.

Fig. 2.

A secure and central location is used for storage of the investigational product. A “Do Not Enroll” list is affixed to an adjacent wall that contains a patient sticker for any patient who has opted out or who is known to fulfill an exclusion criterion. Two sets of hydrogen tanks (one primary, including a manifold, and one secondary) are each affixed to a cart for ease of transport. The primary cart contains randomization envelopes for on-site randomization by a member of the care team

Opportunity to object

If clinical conditions permit, legally authorized representatives (LARs) are given an opportunity to object prior to enrollment; in rare circumstances, proper informed consent may be offered. After enrollment, once the situation is stabilized, families are notified and given an opportunity to object to ongoing participation. The opportunity to object discussion will take place in-person whenever possible, but it may take place by phone if necessary. If requested, study intervention is immediately discontinued (for patients in the hydrogen arm), and the patient is withdrawn from the study.

Additional consent provisions for collection and use of participant data and biological specimens {32b}

When available, discarded samples are collected and frozen from hospital admission to 30 days post-enrollment. For the current study, these samples will be tested for brain biomarkers (GFAP, Tau, Neurofilament light).

During the opportunity to object discussion as described above, LARs are notified that the trial includes the collection of data and of blood specimens. LARs may decline the collection of their child’s blood specimens without declining continued participation in the rest of the trial.

Intervention and comparator

Intervention and comparator description {15a}

Participants randomized to hydrogen treatment receive 2% hydrogen gas via an approved mechanical ventilator and ECMO circuit for 72 continuous hours. Hydrogen is initiated as soon as possible after the decision is made to cannulate to ECMO. The methodology of hydrogen administration through the ventilator has been previously described [9].

Investigational product handling

Hydrogen gas mixtures (2% hydrogen in air, oxygen, carbon dioxide, or “90/10” mixture of carbon dioxide and nitrogen) are provided as certified, GMP-manufactured (Good Manufacturing Practice) mixtures in aluminum cylinders manufactured by a commercial gas vendor (Scott Medical Products, Plumsteadville, PA) and handled under a controlled, study-specific accountability system. Each tank is accompanied by a Certificate of Analysis (CoA) verifying the hydrogen concentration (acceptable range: 1.95–2.05%) by gas chromatography. Study personnel perform a two-person verification to match the tank serial number to the CoA and confirm compliance with protocol specifications. Accepted tanks are labeled with unique study identifiers and stored in locked, access-restricted rooms that are physically separate from clinical gas storage. To limit use to only study patients, tanks are stored in dedicated locations and require a proprietary regulator incompatible with standard hospital fittings, which is maintained exclusively by the study team. All tank movement and use is documented in a detailed accountability log that includes location, patient identifier, identifying tank numbers, and remaining pressure. To ensure uninterrupted gas delivery, tanks are daisy-chained using a dual regulator manifold system that automatically switches to a backup tank when one tank is depleted. Tanks are returned to the vendor when empty.

Criteria for discontinuing or modifying allocated intervention/comparator {15b}

Participants will be removed from hydrogen therapy prior to the 72 consecutive hours of administration in the following scenarios:

• Patient, LAR, or family member objects to participation.

• Identification of an SAE that is thought to be probably or definitely related to hydrogen therapy.

• Patient not cannulated for ECMO within 6 h of randomization.

• Transport of patient.

An additional patient will be enrolled into the study if either of the following conditions is met with respect to an enrolled patient (participant):

• The participant does not receive ECMO (i.e., ECMO does not begin within 6 h of the onset of cardiac arrest).

• The participant is notified of enrollment and opts out of further participation.

Strategies to improve adherence to intervention/comparator {15c}

Prior to activation of a study site for enrollment, in-person training simulations are held with the clinical staff. Refresher trainings are held on a quarterly basis to ensure continued adherence to the randomization procedures and assigned randomized treatment strategy. Adherence to the assigned strategy is monitored through submitted reports of hydrogen interruptions and discontinuation.

Concomitant care permitted or prohibited during the trial {15d}

Not applicable.

Ancillary and post-trial care {34}

Participants are monitored daily for 30 days following the ECPR event for the onset of any hydrogen-related AEs.

Outcomes {12}

Primary endpoints

• Feasibility: Percentage of first 72 consecutive post-arrest hours (starting at the time of first CPR initiation) with successful hydrogen administration. Hydrogen delivery will be determined feasible if the mean percentage in the hydrogen-treated group is ≥90%.

• Safety: Safety assessment in the Hydrogen-FAST trial centers on rigorous, ongoing monitoring of adverse events (AEs) and serious adverse events (SAEs) in the 30 days following randomization. The primary safety endpoint is the incidence rate of treatment-related or possibly treatment-related SAEs, adjudicated by an independently by an expert blinded to treatment allocation. Adverse events are systematically identified through daily review of the electronic health record and classified into three prespecified categories: (1) AEs of interest—specifically defined and graded per a study-specific taxonomy that includes hydrogen-specific complications (Table 1); (2) disproportionate AEs (i.e., events that occur with unexpected severity or timing for the ECPR population); and (3) unanticipated problems. This approach enables attribution of events in a high-risk, critically ill population where disease-related events are common. Hydrogen-related SAEs are calculated per patient-day and compared between study arms to determine non-inferiority, with a prespecified threshold of Et/Ec ≤ 1.125. This classification strategy is designed to detect safety signals that could inform subsequent efficacy trials, while minimizing false attribution in this complex clinical context.

Table 1.

Adverse events of interest

Hydrogen-specific complications
Tank-related complications (e.g., trip hazard, gas running out)
Grade 1	Intervention not indicated
Grade 2	Change in therapy required, tank change required
Grade 3	Severe or medically significant but not immediately life-threatening
Grade 4	Life-threatening consequences; urgent intervention indicated
Grade 5	Death
Ventilator-related complications (e.g., ventilator malfunction)
Grade 1	Ventilator audibly alarming, intervention not indicated
Grade 2	Change in therapy required, e.g., ventilator change required
Grade 3	Severe or medically significant but not immediately life-threatening
Grade 4	Life-threatening consequences; urgent intervention indicated
Grade 5	Death
Evidence for ability to support combustion
Grade 2	Increased spark with electrocautery
Grade 3	Visible flame with electrocautery
Grade 4	Life-threatening consequences; urgent intervention indicated
Grade 5	Death
ECMO-specific complications
Bleeding
Grade 3	Grade III/IV intraventricular hemorrhage, intracranial hemorrhage, epidural hemorrhage
Grade 4	Life-threatening bleeding found not to be at surgical site
Grade 5	Death due to hemorrhage
Thombosis
Grade 3	Thrombus requiring change of component of ECMO circuit within 48 h
Grade 4	Circuit thrombosis, urgent circuit change required
Grade 5	Death
Other ECMO circuit complications (e.g., gas diffusion abnormalities)
Grade 3	Severe or medically significant but not immediately life-threatening; hospitalization or prolongation of existing hospitalization indicated
Grade 4	Life-threatening consequences; urgent intervention indicated
Grade 5	Death
Infectious complications
Grade 3	Positive blood culture with pathogen (not contaminant), surgical site infection, no sepsis physiology
Grade 4	Positive blood culture with pathogen and signs of sepsis
Grade 5	Death

Secondary endpoints

• Feasibility: The percentage of the first 72 consecutive hours post-hydrogen initiation in which hydrogen gas was administered via all of the applicable pathways (e.g., mechanical ventilator and ECMO membrane)

• Survival to hospital discharge

• ICU and hospital lengths of stay (for survivors only)

• Functional status score (FSS) [10] at discharge and 6 months

• Laboratory markers of ischemia (e.g., pH, lactate, hepatic enzymes) as drawn at the primary team’s clinical discretion; data is collected from days 0–30

• Brain biomarkers (GFAP, Tau, Neurofilament light)

Harms {17}

As described earlier in this report, AEs are systematically identified through daily review of the electronic health record and classified into three prespecified categories: (1) AEs of interest—specifically defined and graded per a study-specific taxonomy that includes hydrogen-specific complications (Table 1); (2) disproportionate AEs (i.e., events that occur with unexpected severity or timing for the ECPR population); and (3) unanticipated problems. This approach enables attribution of events in a high-risk, critically ill population where disease-related events are common.

Participant timeline {18}

The full schedule of measurements is available in Table 2. When available, discarded samples are collected and frozen from hospital admission to 30 days post-enrollment. Patient records are screened for endpoints through 30 days post-enrollment.

Table 2.

Trial schedule of measurements

Procedures	Admission to time of ECPR	Randomization (time 0)	Post-ECPR day 0–3 (hour 0–72)	Post-ECPR day 4–30	Hospital discharge (±3 days)	Post-ECPR 6 months (±2 months)
Identify opt-out patients	X
Randomization		X
Enrollment		X
Administer study intervention			X
Review of concurrent therapies and studies		X	X	X	X
Arterial or venous blood gasa			X	X
Serum lactic acida			X	X
Complete blood counta			X	X
Serum chemistrya			X	X
Hepatic transaminasesa			X	X
Coagulation profilea			X	X
Brain biomarkers	X	X	X	X
Adverse event review and evaluation		X	X	X
Functional status score	X				X	X

ECPR extracorporeal cardiopulmonary resuscitation

^aMarkers of ischemic injury, including lowest pH; highest lactic acid; lowest platelet count; and highest hepatic transaminases, blood urea nitrogen and creatinine, and international normalized ratio on post-randomization day 0, 1, 2, 3, 4, 5, 6, 7, 14, 21, and 28

Sample size {19}

There will be a total of 56 participants. The first three are non-randomized vanguard participants and receive hydrogen treatment. The remaining 53 participants will be randomized in a 2:1 ratio (32 hydrogen, 21 control), in order to maximize information regarding the intervention. The trial is powered to demonstrate the safety of the intervention. The required sample size was based on a test of non-inferiority, i.e., that the SAE rate in the hydrogen arm is no worse than that in the usual care arm, with non-inferiority defined as a 12.5% margin. With group sizes of 32 (hydrogen) and 21 (usual care), a one-sided 0.05 level test and assumed 10% relative difference in SAE rates, there is 85% power to demonstrate the non-inferiority of the hydrogen intervention. This sample size will detect safety signals that will inform progression to later-phase efficacy trials.

Feasibility will be demonstrated using a one-sided 90% confidence interval for the mean number of hours of administered hydrogen, commencing with CPR. With the target of 32 participants in the hydrogen arm, the lower limit of the confidence interval will be 80%, demonstrating feasibility (mean administration time of at least 90%).

Recruitment {20}

Study and clinical staff have been trained on screening and randomization procedures so that eligible participants may be promptly randomized. Recruitment is largely done using EFIC procedures, and we expect that recruiting a participant using traditional informed consent will be very rare.

Assignment of interventions: randomization

Sequence generation: who will generate the sequence {21a}

The vendor SealedEnvelope generated the randomization assignments.

Sequence generation: type of randomization {21b}

Restricted randomization will be completed with stratification according to (1) center and (2) age at the time of ECPR (neonate, defined as <28 postnatal days, versus non-neonate, defined as ≥28 days).

Allocation concealment mechanism {22}

Sealed, opaque envelopes conceal the randomization assignments. The strata and envelope number are printed on the outside of each envelope. Envelopes are opened sequentially within each set of strata to reveal assignments.

Implementation {23}

Patients who fulfill eligibility criteria are considered for randomization by a member of the care team present at the time of the resuscitation. This member proceeds to a centrally determined location (Fig. 1) which contains a clearly visible “Do Not Enroll” list; all patients who have opted out as described below and those who do not fulfill eligibility criteria are listed on the Do Not Enroll list, which is maintained by the research team. This improves the efficiency of screening and identification of non-eligible patients. If absent from the list, the enrolling care team member opens the physical randomization envelope appropriate to the center and age stratum.

Assignment of interventions: blinding

Who will be blinded {24a}

The AE adjudicator will remain blinded to treatment intervention.

How will blinding be achieved {24b}

The AE adjudicator will access only a redacted narrative and information describing the AE that does not disclose the participant’s treatment assignment.

Procedure for unblinding if needed {24c}

Not applicable.

Data collection and management

Plans for assessment and collection of outcomes {25a}

Clinical, physiologic, and lab data are collected from the electronic medical record. H₂ delivery logs are used to calculate hydrogen administration compliance. FSS is determined retrospectively using neurology and ICU notes, or by interview with the care team or LAR when notes are insufficient. All data are entered into a 21 CFR Part 11–compliant OpenClinica database.

Plans to promote participant retention and complete follow-up {25b}

We expect that incomplete follow-up will be exceptionally rare. Participants remain admitted to the hospital during most of the course of follow-up, and most data can be obtained from the electronic medical record. The final FSS assessment, at 6 months following the ECPR event, may be completed using one of multiple methods (electronic medical record review, interview with the care team, interview with the LAR).

Data management {26}

Validation checks and range limits are programmed into OpenClinica, allowing for automatic query generation. In addition, the trial data manager reviews entered data in real time and issues additional, manual queries. Question-by-question guides have been developed to advise sites on the entry of data, to ensure consistency. When data are exported to OpenClinica, they are saved on a password-protected, hospital-hosted server accessible only to select members of the study team.

Confidentiality {33}

The OpenClinica database automatically assigns participant IDs to each screened patient. Study sites maintain the link between participant IDs and patient identifiers; no patient identifiers other than dates are submitted in OpenClinica. Further, study sites have been trained on the redaction of patient identifiers from source documents submitted within OpenClinica.

Statistical methods

Statistical methods for primary and secondary outcomes {27a}

Primary endpoints

• For the feasibility endpoint, the percentage of time hydrogen was administered via ventilator and ECMO during the first 72 h will be calculated for each participant randomized to the hydrogen group. The mean proportion and its 95% confidence interval will be reported. Successful feasibility will be defined as a group mean ≥90%.

• For the safety endpoint, the incidence of treatment-related serious adverse events (SAEs) during the first 30 days will be calculated per patient. Poisson or negative binomial regression models will be used to estimate event rates and compare groups, adjusted for potential confounders. The AE adjudicator will adjudicate relatedness to treatment.

Secondary endpoints

The endpoints of ICU and hospital length of stay will be compared in primary analysis using a two-sample t-test of the log-transformed lengths of stay, as well as a logrank test for the treatment group comparison of the distributions of time to discharge. In-hospital mortality will be compared using a Fisher exact test for the two mortality rates and a logrank test for time to death. Functional status score (FSS) in the two treatment groups will be compared with a Wilcoxon rank sum test. Laboratory and biomarker data (laboratory markers of ischemic injury, as well as GFAP, S100B, and neurogranin) will be plotted over time. Longitudinal mixed-effects models will assess group differences over time, accounting for correlation of the repeated measures.

Who will be included in each analysis {27b}

This phase 1 study does not utilize an intention-to-treat approach for analysis. Analysis of the primary endpoints will be performed on the data from participants in the randomized component who were not removed from the study within 72 h of randomization. This per-protocol analysis ensures that patients who are analyzed were sufficiently exposed to hydrogen, so the data will reflect the effects of hydrogen exposure rather than their treatment allocation. Secondary analytic approaches for both primary and secondary endpoints are outlined in the trial protocol.

How missing data will be handled in the analysis {27c}

Minimal missing data are anticipated. Serial laboratory data may have some time points with missing values; however, imputation is not planned.

Methods for additional analyses (e.g., subgroup analyses) {27d}

Descriptive subgroup analyses by location of arrest (ICU vs. catheterization lab vs. operating room), age category (neonate vs. non-neonate), and duration of CPR will be conducted. However, tests of interaction with treatment group will be underpowered to detect differential treatment effects within the subgroup.

Interim analyses {28b}

For this early phase study, there are no planned formal stopping rules for premature termination of the trial for safety, efficacy, or futility. Trial stopping will be in the purview of the DSMB, to be recommended based on the totality of evidence provided in DSMB reports.

Protocol and statistical analysis plan {31c}

Access to the statistical analysis plan is not applicable at this time. Access to the full protocol may be provided upon reasonable request to Dr. Kheir and Dr. Sleeper.

Oversight and monitoring

Composition of the coordinating center and trial steering committee {3d}

Hydrogen-FAST is led by two coordinating centers. The Clinical Coordinating Center includes two pediatric cardiac ICU physician investigators and a research assistant. The Data and Statistical Coordinating Center includes a doctoral statistician investigator, project manager, and data manager.

Composition of the data monitoring committee, its role and reporting structure {28a}

An independent Data and Safety Monitoring Board (DSMB) oversees the study. The DSMB includes five members: two pediatric critical care specialists, a pediatric nephrologist and bioethicist, a biostatistician, and a cardiovascular surgeon. DSMB meetings are held biannually. The role of the DSMB is to assess the progress of the trial, the safety data, and critical efficacy endpoints, and to provide recommendations to the Hydrogen-FAST trial leadership.

Frequency and plans for auditing trial conduct {29}

Each study site is audited by a monitor external to the study teams. Monitoring is completed remotely, first after the first enrollment occurs, and then every 6 months afterward.

Protocol amendments {31}

The coordinating centers jointly implement protocol amendments. When a new amendment is drafted, it is promptly submitted to the sIRB. Once a new protocol amendment is approved by the sIRB, the new protocol version is distributed to study sites along with an Operations Memorandum detailing the changes since the last version. Each protocol version and Operations Memorandum is archived for study sites on the trial administrative website.

Substantive protocol amendments are submitted to the FDA for review promptly. Amendments that are not substantive are instead submitted at the time of the annual review.

Dissemination policy {8}

Results will be disseminated via peer-reviewed publications, scientific meetings, and public summary reports.

Discussion

The Hydrogen-FAST trial represents an important next step in translating promising preclinical findings regarding the protective effects of hydrogen gas into a clinical setting for patients undergoing ECPR. IRI remains a significant contributor to morbidity and mortality in the pediatric CICU population, particularly in those with underlying CHD who already face increased vulnerability due to complex cardiovascular physiology. Despite advances in ECMO technology and resuscitation science, neurologic injury remains a major cause of death or disability after ECPR, with limited therapeutic options available to mitigate reperfusion-associated damage [11].

Hydrogen gas has emerged as a biologically plausible therapy for IRI due to its selective scavenging of cytotoxic oxygen radicals—especially hydroxyl radicals—without impairing beneficial reactive oxygen species involved in cellular signaling [12]. Its favorable pharmacokinetics, diffusibility, and minimal toxicity have been demonstrated across a range of preclinical models including cardiac arrest, myocardial infarction, stroke, and organ transplantation [4, 7, 13–15]. Notably, in a swine model of circulatory arrest closely mimicking ECPR physiology, hydrogen administration was associated with improved neurobehavioral outcomes, decreased brain injury on magnetic resonance imaging, and reduced histologic evidence of ischemic damage [5].

Translationally, the use of hydrogen in critically ill humans remains in early stages. A series of phase I/II studies, including the HYBRID-II trial, demonstrated the safety and potential neuroprotective effect of hydrogen gas when administered to comatose survivors of cardiac arrest [8, 16], an important trial that described a technique for delivering hydrogen in mechanically ventilated adults. The approach to hydrogen administration being used in the Hydrogen-FAST trial is distinct: pre-certified hydrogen gas mixtures are administered through approved, but unmodified, ICU ventilators [9]. This approach permits the free titration of the fraction of inspired oxygen as well as spontaneous breathing, both limitations to previously described approaches to hydrogen administration.

The Hydrogen-FAST trial is also unique in its integration of this new intervention into the highly acute, ethically complex context of pediatric ECPR. The study utilizes an Exception from Informed Consent (EFIC) mechanism to enable enrollment in time-critical scenarios where obtaining proper informed consent is almost always precluded. EFIC has been increasingly recognized as a necessary and appropriate regulatory tool in resuscitation research, provided that robust community consultation and opt-out procedures are implemented [17, 18]. The Hydrogen-FAST trial adheres to FDA 21 CFR 50.24 regulations by incorporating layered stakeholder engagement through community consultations, public disclosure, and opportunities to object before and after enrollment—establishing a framework that could be adapted by other pediatric trials in emergent settings.

This study was intentionally designed as a phase I feasibility and safety trial. By prioritizing rigorous data collection on protocol adherence, SAE rates, and biomarker trends, the Hydrogen-FAST trial will generate essential preliminary evidence to support or refute progression to a larger efficacy trial. While the sample size is not powered to detect small differences in neurologic outcomes or survival, it is adequate to identify feasibility barriers and major safety concerns. Moreover, the collection of banked serum and neuroimaging data will facilitate exploratory endpoints that may inform the design of future phase II/III trials.

If successful, the implications of this study extend beyond pediatric ECPR. Hydrogen therapy could be explored in other critical care conditions characterized by IRI, including out-of-hospital cardiac arrest, neonatal hypoxic-ischemic encephalopathy, cardiopulmonary bypass, and solid organ transplantation. Furthermore, the infrastructure developed for hydrogen administration—such as gas delivery systems, monitoring tools, and dosing protocols—can be leveraged across institutions for broader clinical implementation.

The Hydrogen-FAST trial addresses an urgent clinical need using a mechanistically targeted intervention within a novel regulatory framework. The trial administers 2% hydrogen through standard medical equipment using manufactured, certified mixtures that de-risk the administration of hydrogen in the clinical environment. The findings of this randomized trial will inform not only to our understanding of hydrogen therapy but also contribute to the evolving methodology of pediatric critical care research.

Trial status

Protocol version 2.5, July 21, 2025. Investigational New Drug (IND) 157341 activated June 2023. ClinicalTrials.gov ID NCT05574296. Recruitment began in March 2024. Estimated study completion date is August 31, 2027.

Abbreviations

ECPR

Extracorporeal cardiopulmonary resuscitation

ECMO

Extracorporeal membrane oxygenation

CHD

Congenital heart disease

IRI

Ischemia-reperfusion injury

ICU

Intensive care unit

EFIC

Exception from Informed Consent

SAE

Serious adverse event

GFAP

Glial fibrillary acidic protein

FSS

Functional status score

Authors’ contributions {3a}

JNK conceived the study, led protocol development, and secured funding. He provided oversight for all clinical and regulatory components of the study, including Exception from Informed Consent (EFIC) procedures, investigational product handling, and hydrogen delivery methodology. He co-wrote the manuscript and led its revisions. LAS contributed to study design and led the statistical planning, including sample size calculations, primary and secondary endpoint definitions, and statistical methodology. She also co-wrote the manuscript and contributed extensively to data management infrastructure and regulatory submissions. VH was a primary contributor to protocol development, manuscript drafting, and refinement of clinical procedures. She contributed to coordination across clinical sites and the development of training materials for EFIC implementation. TD contributed to protocol development, logistics, and management of the investigational product. He supported the preparation of randomization materials and study logistics across participating centers. KMB coordinated communication with study sites and contributed to public disclosure planning, recruitment procedures, and data entry protocols. She also helped draft and revise sections of the manuscript related to study conduct and feasibility metrics. AF provided scientific input into the rationale and background for hydrogen therapy, reviewed preclinical data supporting the intervention, and contributed to early feasibility assessments. She helped revise sections of the manuscript detailing biological mechanisms and biomarkers. BNP contributed to site-level implementation planning, patient screening protocols, and clinical data interpretation, particularly related to ECMO logistics and patient selection. He also reviewed and provided feedback on the final manuscript draft. All authors reviewed the study protocol, contributed to manuscript preparation, and approved the final version of the manuscript. Victoria Habet, Tysen DeWaard, and Krislyn M. Boggs contributed equally to this work.

Funding {7a}

National Heart, Lung, and Blood Institute (NHLBI), Grant numbers 1R33HL164373-01 and 1R61HL164373-01, and by the Mooney Family Foundation. The sponsors had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Data availability {6}

Not applicable, as this manuscript is a study protocol.

Declarations

Ethics approval and consent to participate {30}

This study complies with the International Council on Harmonisation’s Good Clinical Practice guidelines and has been approved by the IRB at Boston Children’s Hospital (IRB-P00043374), which serves as the sIRB for all Hydrogen-FAST centers under a reliance agreement. Participants are enrolled into Hydrogen-FAST under Exception from Informed Consent.

Consent for publication

Not applicable.

Competing interests {7b}

The authors declare that they have no competing interests.