Restrictive vs. Liberal Oxygen Therapy for Trauma Patients (TRAUMOX)

The objective of this trial is to investigate whether treatment with oxygen in the early phases after a trauma plays a role in recovery.

Oxygen is probably the most frequently used drug both in the prehospital and emergency setting. It is cheap, easily administered, and, at least for shorter time frames, widely believed to be without any risk of harm. Treatment with oxygen is indicated in the state of hypoxemia in order to prevent hypoxic organ damage, however, oxygen is administered in many other situations too, sometimes in a non-consistent manner and very often without even being prescribed. Notably, administration of un-titrated oxygen, introduces a risk of hyperoxia which has been associated with serious adverse effects. Retrospective studies on patients with cardiac arrest, myocardial infarction (MI) and traumatic brain injury have found hyperoxia to be associated with increased mortality and ICU length of stays. A randomized trial also found an increased rate of recurrent myocardial infarction in a high-concentration oxygen group. Furthermore, in patients undergoing surgery, high intraoperative FiO2 has been associated with major respiratory complications and with 30-day mortality. In ICU patients, an observational study as well as a randomized trial found high levels of oxygen to be associated with higher mortality compared to conservative oxygen treatment. In particular, the randomized trial found that patients with PaO2 between 70 and 100 mmHg had a mortality of 11.6% compared to 20.2% in patients managed in accordance with standard ICU practice (which allows PaO2 up to 150mmHg) (p=0.01). For people between the ages of 5 and 44 years, trauma is one of the top three causes of death. Almost twice as many men as women die as a result of trauma. Victims of trauma are often healthy individuals prior to the incident, but acquire numerous complications including sepsis and pulmonary complications as well as long-term complications and diminished quality of life after the trauma. Understanding the reasons behind these complications and poor outcomes and optimizing care is therefore essential to increase survival rates and quality of life. In the trauma population, oxygen administration is often standard of care. However, the evidence supporting oxygen administration in this population appears to be extremely limited. Nevertheless, absorption atelectases have been shown to develop within minutes under an oxygen fraction of 1.0, and are strongly related to the development of pneumonia. Indeed, a recent retrospective study found hyperoxemia to be an independent risk factor for ventilator associated pneumonia (VAP). As the incidences of hyperoxemia in mechanically ventilated patients range from 16-50% and the incidence of pneumonia in general in the Intensive Care Unit (ICU) is close to 26%, and reported to be up to twice as high in the trauma population, studies are warranted to establish the impact of a restricted, but sufficient, oxygen strategy vs a liberal oxygen strategy in the trauma population. The primary objective of this trial is to evaluate whether the maintenance of pragmatic normoxia, avoiding both hyperoxic and hypoxic phases, is feasible within the first 24 hours after trauma, as it may result in a reduction of 30-day mortality and major respiratory complications within 30 days (respiratory failure, pulmonary edema, and pneumonia). The investigators will therefore conduct a pilot study, where 40 evaluable patients are randomized to 24 hours of: A. Restrictive, but sufficient oxygen treatment: o Lowest oxygen delivery that obtains a saturation of ≥ 94% (within the low limits of standard of care and does thus not put patients at risk of receiving worse treatment. Hypoxic phases will be avoided.) B. Liberal oxygen treatment: 15 L/min oxygen flow initially/ FiO2 ≥0.8. (Considered as close as possible to the standard of care. As no precise guidelines on oxygen delivery for trauma patients within the first 24 hours exist, this will however inevitably vary in practice. We have therefore set the delivery to be in the high end of the varying standard of care.) The investigators hypothesize that hospitalized trauma patients treated with liberal oxygen therapy during the first 24 hours after trauma will have a greater risk of death and major respiratory complications compared to trauma patients treated with restrictive oxygen. The investigators realize that this cannot be assessed in this pilot study, but it is crucial to know if the study is feasible before doing a large and adequately powered trial. Composite Primary Outcomes: • 30-day mortality and major respiratory complications within 30 days (pneumonia, acute respiratory distress syndrome (ARDS), acute lung injury (ALI)). Secondary Outcomes Intensive Care Unit length of stay (ICU LOS) Hospital length of Stay Days on ventilator 30 days post-trauma (through telephone follow-up if discharged) Glasgow Outcome Scale Extended (GOSE) Pneumonia

Intervention: restrictive oxygen treatment for the first 24 hours. This implies administration of the lowest dosage of oxygen possible in order to achieve a saturation of 94%, using mechanical ventilation (intubated patients), a nasal cannula, a non-rebreather or nothing. A saturation above 94% shall not be aimed for using supplemental oxygen, and thus only patients without oxygen requirement shall have saturations above 94%. Control: liberal oxygen treatment for the first 24 hours. In the trauma bay and during intrahospital transportation this implies an FiO2 of 1.0 for intubated patients and an oxygen flow on a non-rebreather with reservoir of 15 l/min for non-intubated patients. In the OR, patients will receive a FiO2 of ≥ 0.8 to obtain a saturation of ≥ 98%. Patients admitted to the ICU/PACU/floor will receive and FiO2 of ≥ 0.8 or more to obtain a saturation of ≥ 98% when intubated and for non-intubated patients a non-rebreather with reservoir will be set to 15 l/min.

The study will be an open-label randomized clinical trial with regards to treatment: treating staff will be aware of the patient's randomization group. Outcome assessors (attending anesthesiologist for pneumonia and research assistant for long term outcomes (GOSE after 30 days, ARDS* within 30 days, pneumonia diagnosed after discharge within 30 days)) will be blinded to the patients' randomization. To ensure blinding, prior to assessing the outcomes, all information indicative of the allocation will be censured by the PI. To validate the blinding, the attending anesthesiologist and research assistant assessing outcome variables will be asked to guess which group of randomization the patient is in. ARDS: PaO2/FiO2 will be calculated for the assessor in order to maintain blinding.

Liberal oxygen administration (to mimic current practice) for the first 24 hours without interruption. In the trauma bay and during intrahospital transportation this implies administration of a FiO2 of 1.0 for intubated patients and an oxygen flow on a non-rebreather with reservoir of 15 l/min for non-intubated patients. In the operating room, patients will receive a FiO2 of ≥ 0.8 to obtain a saturation of ≥ 98%. Patients admitted to the ICU/PACU/floor will receive and FiO2 of ≥ 0.8 or more to obtain a saturation of ≥ 98% when intubated and for non-intubated patients a non-rebreather with reservoir will be set to 15 l/min.

Titrated oxygen administration for the first 24 hours without interruption. Lowest dosage of oxygen possible in order to achieve a saturation of at least 94%, either using mechanical ventilation (intubated patients), a nasal cannula, a non-rebreather or nothing. A saturation above 94% shall not be aimed for using supplemental oxygen, and thus only patients without oxygen requirement shall have saturations above 94%. The intervention will only be interrupted in case the saturation becomes unmeasurable - if this happens, the treating physician shall treat the patient as he/she judges best fit. As soon as the saturation is measurable again, the intervention will resume. The treating physician must document and explain the situation.

30-day mortality and major respiratory complications within 30 days (pneumonia (CDC criteria), acute respiratory distress syndrome (ARDS), acute lung injury (ALI)).

From date of randomization until the date of first documented discharge from the intensive care unit, up to 100 months

From date of randomization until the date of first documented discharge from the hospital, up to 100 months

Eight point grading scale of recovery levels assessed through a questionnaire. 1= death, 2=vegetative state, 3=lower severe disability, 4=upper severe disability, 5=lower moderate disability, 6=upper moderate disability, 7=lower good recovery, 8=upper good recovery

Inclusion Criteria: Age ≥ 18 years, including fertile women* Blunt/Penetrating trauma patient Direct transfer from the scene of the accident to Rigshospitalets Traume Center Trauma team activation * There is no added risk for enrolment of fertile women as oxygen administration is approved for this group of patients. Exclusion Criteria: Patients in cardiac arrest before/at admission Patients with a suspicion of smoke inhalation No hospital admission after initial treatment in trauma center

Baekgaard JS, Isbye D, Ottosen CI, Larsen MH, Andersen JH, Rasmussen LS, Steinmetz J. Restrictive vs liberal oxygen for trauma patients-the TRAUMOX1 pilot randomised clinical trial. Acta Anaesthesiol Scand. 2019 Aug;63(7):947-955. doi: 10.1111/aas.13362. Epub 2019 Mar 25.