Ultra-ECMO trial is a single-center, prospective clinical trial that will enroll 2 parallel groups of patients in a 1:1 ratio. It is designed to evaluate the 28-day number of ventilator-free days (VFDs) following an early, short-course application (72 h) of an ultra-protective VT vs. protective VT strategy in patients on VA-ECMO (Table 1) (1). Table 2 overviews the trial registration information. This trial protocol conforms to the Consolidated Standards of Reporting Trials (CONSORT) statements (10, 11) and has obtained the approval of the Research Ethics Committee at the First Affiliated Hospital of Nanjing Medical University. The legal guardians of eligible patients will sign the informed consent following the Declaration of Helsinki of 1964 (revised in 2013) (12).

This trial has the main objective of assessing the potential benefit of an ultra-protective VT strategy relative to a protective VT strategy on the 28-day VFD number (13) (Table 3).

Meanwhile, the secondary objectives of this study include the following: 1.

To examine whether ultra-protective VT can decrease the all-cause in-hospital mortality rate

We will conduct the Ultra-ECMO trial in the Department of Emergency Medicine at the First Affiliated Hospital of Nanjing Medical University, which is capable of initiating ECMO therapy for nearly 100 critically ill patients annually.

Patients ≥18 years of age receiving VA-ECMO therapy, such as extracorporeal cardiopulmonary resuscitation, will be eligible for inclusion in the Ultra-ECMO trial. Additional key inclusion criteria include (1) IMV onset of >72 h, (2) no pregnancy, (3) ECMO duration of >72 h, (4) RCS or CA with a cardiac origin, (5) no enrollment in another clinical trial, (6) a consent form signed by the guardian has been provided, and (7) a description of CPE on high-resolution chest computed tomography images has been given by two radiologists with ≥5 years of diagnostic experience who are unaware of the clinical diagnosis.

Meanwhile, the exclusion criteria for patient enrollment are as follows: (1) trauma; (2) written informed consent has not been provided by close relatives; (3) the patient has contraindications for fiberoptic bronchoscopy examination; (4) lung ultrasound images are blurry; (5) pneumothorax or massive pleural effusion; (6) ECMO is being used as a bridge for heart transplantation; and (7) the patient developed adverse events that complicate weaning from IMV (e.g., intracranial hemorrhage, thromboembolism, severe infection).

The Ultra-ECMO trial will include patients who meet the abovementioned eligibility criteria. Formal information about the trial and an official copy of the informed consent form will be sent to their guardians. We will assign participants to an intervention arm or a control arm in a random manner, with those in the former group receiving ultra-protective VT and those in the latter receiving 72 h of protective VT. Guardians will also be informed that they are allowed to drop out of the trial at any time.

The randomization list will be computer-generated according to a simple randomization principle using a ratio of 1:1 by the Clinical Research Board of the School of Public Health of Nanjing Medical University. Participants will be assigned to each group by an independent trial assistant after verification of patient eligibility. Once group allocation has been completed, local investigators will register the record, then reveal the group allocation to care providers, who will deliver the intervention.

Trial assistants responsible for group allocation will be blinded to both groups of patients, but blinding of care providers is impracticable as ventilatory settings and respiratory mechanics measurements are part of routine daily clinical practice. The outcome assessors, data analysts and statisticians will not know the group allocation.

At the onset of ECMO initiation, all participants will be screened by therapists. Once the patient meets the inclusion criteria and is finally enrolled, the therapists will provide the selected intervention. We will assign patients in the intervention group to receive ultra-protective VT after randomization (initial settings: VT, 4 ml/kg of PBW; PEEP, 10 cmH₂O; respiratory rate, 10 min⁻¹; Vt may be decreased stepwise by 0.5 ml/kg PBW if the plateau pressure is >30 cmH₂O) (17).

Patients in the control group will be managed using a protective VT strategy after randomization (initial settings: VT, 6 ml/kg of PBW; PEEP, 10 cmH₂O; respiratory rate, 10 min⁻¹; Vt may be decreased stepwise by 0.5 ml/kg of PBW if the plateau pressure is >30 cmH₂O).

We will set the sweeping gas flow of ECMO, targeting a PaCO₂ range of 35–45 mmHg, and the FiO₂ value of the ventilator may be increased if PaO₂ in the right upper limb decreases to <60 mmHg during the course of Vt and PEEP adjustments.

Furthermore, a hybrid veno-arterio-venous ECMO mode will be considered if differential oxygenation happens (18). Differential oxygenation refers to the lower body being perfused by ECMO-oxygenated blood while the upper body (coronaries, right arm, brain) receives hypoxemic blood pumped by the native heart. This pathophysiological phenomenon often results from the recovery of native left ventricular function and simultaneous severe lung failure. If optimized ventilation fails to solve this issue, it will be necessary to employ veno-arterio-venous ECMO in consideration of the potential for hypoxic damage to the brain and heart (19, 20).

At the moment of enrollment (T0), investigators will take charge of assessing and collecting pre-defined respiratory mechanics parameters; IL-6, IL-8, and MCP-1 levels; LUS; and other trial data as baseline information. The same procedure will be performed repeatedly at 24, 48, and 72 h (T1, T2, and T3, respectively) after ECMO initiation (Figure 2). After 72 h, the mechanical ventilation strategy in both groups will be changed at the intensivists' discretion.

Patients in both groups will receive deep sedation (i.e., targeting a Richmond Agitation Sedation Scale score of −3 or −4 points) during the intervention period. Although deep sedation will be adopted, it is still recommended to administer a neuromuscular blocking agent specific to the situation of patient–ventilator asynchrony.

In case adverse events occur due to the intervention, patients will be managed by the attending physicians at their discretion but still remain in the allocation group to support later data analysis.

The total duration of ECMO treatment, total time (in days) on IMV, ICU length of stay, analgesic/sedation dosage, all-cause hospital mortality, levels of inflammatory factors in BALF, LUS, amounts of fluid infusion, and total hospitalization cost will be calculated using patient medical records (Table 6).

We will not consider missing values for primary and secondary outcomes during hospitalization. However, missing values could appear specific to patients who survive to hospital discharge. If so, we will contact patients or their next of kin at day 28 to assess the pre-defined outcomes.

Clinical data will include but are not limited to general characteristics (age, sex, weight, height), comorbidities, original disease, the Acute Physiology and Chronic Health Evaluation II score, partial pressure of arterial oxygen/inspired oxygen fraction, PEEP, VT (in ml/kg of PBW), respiratory rate (in breaths/min), DP, P_plat, P_es, P_tp, P_mean (in cm of water), C_stat (in ml/cmH₂O), R (in cmH₂O/L/s), arterial blood PaO₂ and PaCO₂ values (in mmHg), and mechanical power (in J/min).

We will also monitor the clinical conditions as well as the ECMO running status and collect any records.

Delegated team members will take charge of inputting data in the electronic case report form, and experienced clinical research associates will take charge of the data monitoring. Team members will store research data under each participant's study identification code, which will include the first letter of the first and family names of the participant, together with their inclusion number. Only the research team will have the right to possess the key to the identification code list during the study. Detailed information regarding the identification of patients will not be available in any resulting publications.

PASS 11 was used for sample size calculation, where a sample size of 292 patients per group would achieve 80% power to detect a difference between the two groups, providing the means were 7.1 and 9.2 with group standard deviations of 8.8 and 9.3 and with a significance level α of 0.05 using a two-sided two-sample t-test. Accounting for an anticipated dropout rate of 5%, 614 patients were required for the study.