Intravenous Imatinib in Mechanically Ventilated COVID-19 Patients

1. INTRODUCTION 1.1 Background of ARDS due to COVID-19 The COVID-19 pandemic has led to an unanticipated increase of the number of patients with ARDS admitted to the ICU, contributing to high morbidity and mortality, as well as an unprecedented consumption of medical resources. COVID-19 is caused by a coronavirus (scientific name: SARS-CoV-2), a non-segmented, positive sense RNA virus. Although most SARS-CoV-2 infections have an asymptomatic or mild course of disease (80%), COVID-19 has a detrimental course in a minority of patients (20%), particularly in older patients or patients with pulmonary or cardiovascular comorbidities. In these cases, COVID-19 infection is characterized by damage to the alveolocapillary wall and extensive pulmonary capillary leak. The alveolar flooding results in impairment of oxygen diffusion and severe hypoxemic respiratory failure. In Chinese registries these cases of COVID-19 disease were classified as either 'severe' in case of low oxygen saturation or 'critical' when invasive mechanical ventilation was needed or multi-organ failure occurred. Radiological imaging demonstrates extensive ground glass opacities, consistent with alveolar oedema. Radiological as well as pathological examination demonstrated that critical COVID-19 infections closely mimic Acute Respiratory Distress Syndrome (ARDS), a condition characterized by damage to the alveolo-capillary membrane by various insults. Mortality in ICU treated 'critical' COVID-19 patients is comparable to mortality in ARDS patients. Of the patients admitted with COVID-19 infection to hospital, 17-35% develop ARDS, requiring ICU admission or even invasive mechanical ventilation (29-91%). According to recent reviews, mortality may mount up to 15-20% (hospitalized patients) to even 40% in ICU patients. For these reasons, the current COVID-19 pandemic has resulted in a huge increase in the incidence of ARDS with a homogenous aetiology i.e., SARS-CoV2 infection. According to the Berlin definition, ARDS is '…an acute diffuse, inflammatory lung injury, leading to increased pulmonary vascular permeability, increased lung weight, and loss of aerated lung tissue…[with] hypoxemia and bilateral radiographic opacities, associated with increased venous admixture, increased physiological dead space and decreased lung compliance'. It is characterised by an acute onset, with bilateral infiltrates on chest imaging due to pulmonary oedema, and with severe hypoxemia despite mechanical ventilation. Pathophysiologically, ARDS results from an overwhelming inflammatory process involving alveolar epithelial and vascular endothelial injury in the lung which can be infective and non-infective in origin. The early phase of ARDS is characterized by alveolar flooding with protein-rich fluid due to increased vascular permeability. Pulmonary oedema then leads to the clinical manifestation of poor lung compliance, severe hypoxaemia, and bilateral infiltrates on chest radiograph. It also leads to alveolar epithelial injury of type I cells, which contributes further to the pulmonary oedema. Despite decades of efforts, there is currently no registered drug to target pulmonary vascular permeability in ARDS. Current treatment of patients with 'critical' COVID-19 consists of supportive measures, including oxygen supplementation, diuretics and (non)-invasive mechanical ventilation. Current guidelines prescribe dexamethasone and remdesivir for patients with hypoxemia. Although remdesivir and dexamethasone target key pathophysiological processes of COVID-19 like viral replication and inflammatory damage, respectively, there is currently no proven benefit of pharmacological intervention to reverse pulmonary vascular leak and oedema. 1.2 Rationale for intervention Earlier studies demonstrate that the anti-leukemic drug imatinib effectively and consistently protects against pulmonary vascular leak and alveolar oedema during inflammatory stimuli. In 2008, a patient with acute respiratory failure due to widespread pulmonary oedema was treated with imatinib mesylate for another indication than the acute respiratory failure itself. However, initiation of the imatinib treatment (200mg/day) was followed by a surprisingly fast reversal of respiratory failure. The reversal of respiratory failure was paralleled by decrease in pulmonary oedema on radiological imaging, resulting in the hypothesis that imatinib directly protects the pulmonary endothelial barrier. This unexpected effect of imatinib was published in 2008. The hypothesis was tested in an extensive preclinical study, evaluating the effect of imatinib on endothelial barrier function in several conditions. These studies demonstrated that imatinib protects the endothelial barrier under inflammatory conditions, both in in vitro models using several types of endothelial cells and various inflammatory mediators, and in vivo, using various animal models for vascular leak. The protective effects of imatinib were confirmed by several studies by independent research groups. After the publication of the first case report and the preclinical evidence for a protective effect of imatinib on the endothelial barrier, imatinib (300mg/day) was associated with clinical improvement of acute respiratory failure in 1 patient in an independent hospital. In addition, imatinib has been used in a compassionate use setting in 2 additional patients in the investigators' hospital - in both cases initiation of imatinib therapy (200-400mg/day) was followed by reversal of vascular leak and/or respiratory failure. Although these cases should be interpreted with caution, they consistently suggest a beneficial effect of imatinib on pulmonary vascular leak and/or respiratory failure. To date, one case report using oral imatinib to successfully treat COVID-19 pneumonia has been published. The protective effects of imatinib at the endothelial level 2-10μM were found at concentrations that are comparable to plasma levels found in patients treated with imatinib for CML (2-5μM), indicating that regular dosing schemes of imatinib are sufficient to induce its protective effect on the endothelial barrier. Indeed, preliminary data from the COUNTER-COVID study, which evaluates the effect of oral imatinib mesylate in hospitalized patients with COVID-19, demonstrate that dosing of 400mg/day orally is sufficient to reach target levels described above. Based on the observations that the optimal protective effect of imatinib on the endothelial barrier ranges between 2-10μM, together with previous pharmacokinetic studies after intravenous imatinib, the investigators propose a dosing scheme of 200mg b.i.d. This dosing scheme was shown to provide plasma levels that correspond to the 2-10μM range found in prior in vitro studies. The current study will evaluate the effect of intravenous imatinib on pulmonary oedema as measured by EVLWi. Several arguments drive the use of intravenous imatinib in the current study: 1) since most ICU patients are intubated, oral administration is suboptimal. 2) Intestinal uptake in mechanically ventilated ICU patients is often impaired due to intestinal oedema. 3) by intravenous administration, imatinib directly reaches the target organ, i.e., the endothelium, requiring lower dosing and fewer side-effects. A detailed analysis on dosing and route of administration is provided in the Investigators Brochure. Concerning safety, imatinib was shown to have mild or no effects on the immune response. Although developed to target leukemic cells, imatinib hardly affects healthy leucocytes and a normal lymphocytic response was observed in lymphocytes from patients treated with imatinib. Of particular relevance for the current protocol, it was shown that treatment with imatinib does not affect the control of primary viral infections. The investigators demonstrated in a healthy volunteer model of lung injury that imatinib did not affect the immune response to LPS inhalation in healthy human subjects. Finally, the first interim analysis of the COUNTER-COVID study, which tests oral imatinib in COVID-19 patients admitted to hospital with hypoxemia, did not reveal safety concerns after inclusion of the first 140 patients. Altogether, substantial evidence indicates that imatinib is an ideal candidate for treatment of the pulmonary complications of SARS-CoV2 like ARDS, since it protects against vascular leak and alveolar oedema thereby reducing hypoxemic respiratory failure. The current study seeks to directly measure the effect of intravenous imatinib on pulmonary vascular leak in COVID-19 ARDS. COVID-19 offers a homogenous cause of ARDS, however, the available evidence indicates that ARDS due to other causes will respond similarly to imatinib. Therefore, the identification of a compound that reverses pulmonary vascular leak may benefit well beyond COVID-19 ARDS and provide a first treatment for other forms of ARDS as well.
2. OBJECTIVES Primary:

• Efficacy: To evaluate the effect of intravenous imatinib compared to standard of care on limiting development of extravascular lung water in invasively mechanically-ventilated subjects with COVID-19-related ARDS. Secondary:
• Efficacy: To evaluate the effect of intravenous imatinib compared to standard of care on patient outcomes in mechanically-ventilated subjects with COVID-19-related ARDS.
• Safety: To evaluate the safety and tolerability of intravenous imatinib compared to standard of care in mechanically-ventilated subjects with COVID-19-related ARDS
• Pharmacokinetics: To determine imatinib pharmacokinetics in subjects with COVID-19-related ARDS.

1. STUDY DESIGN This is a randomised, double-blind, parallel-group, placebo-controlled, multi-centre clinical study comparing intravenous imatinib mesylate with placebo in invasively mechanically-ventilated subjects with COVID-19-related ARDS. The study will enrol 90 subjects (45 subjects/treatment arm; see sample size calculation). In this two-arm study, eligible subjects will be randomly allocated to receive imatinib mesylate or placebo in a 1:1 ratio. Importantly, the inclusion will be extended to all-cause ARDS if recruitment drops to below a predefined threshold (see inclusion criteria). Besides the randomisation to treatment of placebo, all subjects will be cared for using local treatment protocols as standard of care. At each of the participating centres, standard treatment protocols may include: COVID-19-specific medication (e.g. remdesivir and dexamethasone), low tidal volume ventilation, conservative fluid management and prone positioning in case of persistent low PaO2/FiO2. For eligibility criteria and outcome measures, see separate section.
2. SAMPLE SIZE CALCULATION The number of subjects scheduled for inclusion in the study is 90, including 45 subjects in the placebo arm and 45 subjects in the imatinib mesylate arm. Sample size calculations were done with the formula: zα/2 - zπ = (n/2)1/2 * (μ1 - μ2)/σ, in which zα/2-zπ = 2,8 (α = 5% en π = 80%). The change (Δ) in EVLWi between Day 0 and Day 4 is the primary outcome. The investigators expect the baseline EVLWi to be around 17ml/kg, as previously described for patients with moderate-severe ARDS. In the placebo group the ΔEVLW Day 0 and Day 4 is expected to be 0.5 (μ1), based on previously literature from collaborating groups; in the imatinib group the ΔEVLWi is expected to be -4 (μ2). This is considered a clinically relevant difference as this difference was found to independently predict ARDS mortality in another clinical study. The expected treatment effect is based on the imatinib effect observed in preclinical data: 25% reduction in vascular leak. The sigma (σ) for ΔEVLWi is set at 7.0, based on previous EVLWi studies. Using the equation above, this yields 76 subjects with 38 subjects/arm. Taking into account a drop-out rate of 15%, a total of 90 subjects will be recruited.
3. TREATMENT OF SUBJECTS 5.1 Investigational product/treatment Subjects will undergo 1:1 randomisation to receive intravenous imatinib (200mg bid) or placebo for 7 days. 5.2 Interactions Imatinib is a substrate for CYP3A4 and P-gp. It inhibits CYP3A4, CYP2D6 and CYP2C9. Most relevant interactions for the study population include:

• Strong induction of CYP3A4. The following drugs are contra-indicated for the study: Carbamazepine, efavirenz, enzalutamide, phenobarbital, phenytoin, hypericum, mitotane, nevirapine, primidone, rifabutin, rifampicin.
• CYP3A4 inhibition. In general, imatinib is relatively insensitive to CYP3A4 inhibition, as imatinib may rely on enzymes other than CYP3A4 for its metabolism. Therefore, there is no need for adjustment, but recording of the following drugs in the eCRF is important: azoles.
• Imatinib may change the concentration of other CYP3A4, CYP2D6 and CYP2C9-dependent drugs, including anticoagulants, cyclosporine, simvastatin etc. All relevant co-medication will be registered.

1. INVESTIGATIONAL PRODUCT 6.1 Dosage and Administration A 25ml volume of IMP will be administered over 2-hours as an intravenous infusion. This corresponds to a dose of 200mg imatinib (100mg/h), or 25ml placebo (12.5ml/h). Treatment will be administered twice daily (400 mg total daily imatinib dose) for up to 7 days, or until patient is discharged from critical care, if earlier. The first dose of IMP will occur on Day 1, as soon as possible after randomisation. IMP should be administered 12h (± 2h) apart between 06:00 - 10:00 in the morning and 18:00 - 22:00 in the evening. On Day 1, the first dose of IMP may be administered between 04:00 and 12:00 (08:00± 4h) and the second dose between 16:00 and 24:00 (20:00± 4h). Patients commencing treatment after 12:00 on Day 1 should receive only one dose of IMP (between 16:00 and 24:00).
2. Randomisation, blinding and treatment allocation Randomisation: Before randomisation, the study physician will check all inclusion and exclusion criteria and will enters these in the web-based application (Castor). Screening and placement of the PiCCO catheter is performed as part of a deferred consent procedure. For subjects who meet the eligibility criteria, the patient's legal representative will be asked for informed consent(§11.2.2). When informed consent is provided, subjects will be randomized 1:1 to receive placebo or imatinib. Randomisation will take place via Castor using variable block sizes with stratification per participating centre. The pharmacist will check relevant drug interactions before dispensing medication. Blinding: Subjects, clinical staff and investigators will be blinded for study medication. Blinding will be guaranteed be covering the syringes containing the IMP so that the content will not be visible from the outside. Code breaking is allowed in any of the following circumstances:

• Treatment of an individual in a medical emergency where knowledge of the treatment allocation is required.
• In the event of a Suspected Unexpected Serious Adverse Reaction (SUSAR) the subject will be unblinded, if this is required for treatment of the SUSAR.
• In the event that the ICU staff / researcher is accidentally exposed to study medication.

1. SAFETY REPORTING 8.1 Temporary halt for reasons of subject safety In accordance to section 10, subsection 4, of the WMO, the sponsor may suspend the study if there are sufficient grounds that continuation of the study will jeopardise subject health or safety. The sponsor will notify the accredited METC without undue delay of a temporary halt including the reason for such an action. The study will be suspended pending a further positive decision by the accredited METC. The investigator will take care that all subjects are kept informed. 8.2 AEs, SAEs and SUSARs: Due to the nature of disease, the incidence of AEs and SAEs, as well as the risk of death due to the underlying condition is high (the hospital mortality in ventilated ICU patients is 21%). In clinical studies conducted in patients with ARDS, changes in vital parameters are frequent, including temporary changes in blood pressure and gas exchange parameters, as are deviations in laboratory values and ECG values. AE Reporting: Given the high incidence of adverse events inherent to the nature of the underlying condition (i.e., ARDS), given the safety of imatinib as observed in COVID-19 pneumonitis before (COUNTER-COVID study), and given the regular recording of vital signs and blood and ECG parameters as safety indicators in the eCRF, the investigators propose not to report AEs in a standard fashion, except for the following events:

• Pulmonary embolism, as detected on contrast enhance chest CT, not leading to circulatory of pulmonary instability.
• Occurrence of infections, requiring initiation of antibiotic therapy
• Non-life-threatening infusion reactions, including but not limited to skin rash. SAE Reporting: Given the high incidence of adverse events inherent to the nature of the underlying condition (i.e., ARDS), and given the safety of imatinib as observed in COVID-19 pneumonitis before (COUNTER-COVID study), the investigators propose not to report all serious events as SAEs. The following SAEs will be reported:
• Death due to any cause
• Cardiopulmonary: The need for extracorporeal membrane oxygenation; Cardiac events like arrythmias requiring CPR or medical resuscitation; Thrombo-embolic events with life-threatening circulatory or pulmonary instability
• Spontaneous bleeding, requiring blood transfusion or surgical intervention.
• Myocardial infarction
• Renal: the need for renal replacement therapy
• Hepatic: Liver failure (hepatic SOFA score >4)
• Haematology: Thrombocytopenia (<50^109/L), diffuse intravascular coagulation, leukocytopenia (<2^109/L), anaemia (haemoglobin <4mmol/L).
• Central Nervous System: Intracranial bleeding or ischemic stroke.
• Life-threatening infusion reactions, requiring intensification of existing intensive care treatment, including additional vasopressor, fluid support, corticosteroids and antihistamines.
• Any unexpected serious event judged as an "untoward medical occurrence". Reporting of Suspected unexpected serious adverse reactions (SUSARs):
• SUSARs that have arisen during this clinical trial that was assessed by the METC;
• SUSARs that have arisen in other clinical trials of the same sponsor and with the same medicinal product, and that could have consequences for the safety of the subjects involved in this clinical trial that was assessed by the METC.

1. STATISTICAL ANALYSIS The statistical analysis will be performed using the program SPSS version 22.0 or R. Continuous data will be tested for normal distribution using a Kolmogorov-Smirnov test. In case of normal distribution, continuous data will be presented as mean ± standard deviation, in case of non-normal distribution, continuous data will be presented as median ± interquartile (IQR) range. Categorical data will be presented as absolute number (%). The continuous variables and the categorical data will be presented quantitatively in tables and in figures. Missing data will be accepted after maximal efforts to retrieve data.

10.1 Primary study parameter(s) The primary endpoint is the change in extravascular lung water index (ΔEVLWi) between Day 1 (baseline) and Day 4. The primary endpoint will be represented mean ± standard deviation, in case of non-normal distribution, continuous data will be presented as median ± interquartile (IQR) range, and tested for statistical difference using a t-test or a Mann-Whitney U test in case of non-normal distribution. In addition, ANCOVA analysis will be performed.

10.2 Secondary study parameter(s) For all parameters, statistical comparison will be performed between placebo and imatinib group.

Plasma concentrations of imatinib at Ctrough will be described by descriptive statistics, including mean, SD, minimum, maximum, and median.

1. ETHICAL CONSIDERATIONS 11.1 Regulation statement The general principles of informed consent, ethics review and data management will be in line with good clinical practice (GCP). The study will be conducted according to the principles of the Declaration of Helsinki (2013) and in accordance with the Medical Research Involving Human Subjects Act (WMO).

11.2 Recruitment and consent 11.2.1 Recruitment Subjects will be recruited on the emergency units of the AUMC, locations AMC and VUMC, and the Onze Lieve Vrouwe Gasthuis, location Amsterdam Oost. Two additional centres are under consideration. With a minimum of 5 active centres and an anticipated duration of the trial of 18 months, and a target number of inclusions of 90 subjects, this means that 90/5=18 subjects per centre need to be enrolled, which amounts to <0.25 subjects/centre/week. When considering the AMC and VUMC as representative hospitals, the required rate of inclusion/centre is far below the current rate of presentations/centre.

11.2.2 Deferred consent For this trial deferred consent is asked and the investigators appeal to the emergency procedure for consent in medical research as stated in article 6, paragraph 4 of the WMO, similar to a presently running trial of ventilation in a similar patient cohort, the 'REstricted versus Liberal positive end-expiratory pressure in patients without Acute respiratory distress syndrome (RELAx)' - a multicenter randomized controlled trial (NL60402.018.17).

Pulmonary vascular leak is a phenomenon that is observed early in the course of ARDS. Previously, an early 'exudative' phase (the first 3-5 days in the course of ARDS) was distinguished from a 'reorganising/proliferative' phase (Day 5 and further in the course of disease). Although this distinction is formally abandoned, it is generally accepted that the early course of ARDS is characterised by pulmonary vascular leak and alveolar exudates, in which the pulmonary oedema is the main cause of hypoxemia. Based on preclinical and early clinical data there is a reasonable chance that patients will benefit from the vasculoprotective effect of imatinib, especially in the early, exudative phase. The short extent of this exudative phase in time (days) indicates a limited window-of-opportunity to apply therapies directed at preserving or restoring the alveolocapillary barrier. For this reason, the investigators consider it of utmost importance to initiate the vasculoprotective treatment with imatinib as soon as possible (i.e., within 48 hours after ICU admission), not doing so would largely reduce validity of this trial. Additionally, the potential benefit from the intervention (which would not be available outside of given trial) is likely highest in the early phases.

Subjects admitted to the ICU with ARDS are, without exception, incompetent to give informed consent. Persons who may take the role of legal representative in accordance with the WGBO are: a predefined representative, husband or wife, registered partner or other life partner, a parent or child, brother or sister, and incidentally a curator appointed the judge. However, obtaining informed consent from a legal representative in this situation usually takes much time, even by an experienced research team (see Textbox A). Reasons include the absence of a legal representative within the first 48h after admission, and early after admission to the ICU the legal representatives are far more concerned about the wellbeing of the patient then participation in a trial.

For these reasons, the investigators opt for using deferred consent, where informed consent from a legal representative must be obtained as soon as possible, but always within 48 hours after intubation. While awaiting deferred consent, subjects will undergo screening and first baseline measurements. A PiCCO catheter will be placed (if not already in place), and the baseline EVLW and PCP measurement, as well as thoracic ultrasound in a subpopulation will be performed. The placement of a central venous catheter is part of standard care in intubated subjects. When deferred consent is provided the patient will continue study participation by randomisation and administration of the first IMP dose. When deferred consent is not obtained due to absence of a legal representative in the first 48 hours, or if a legal representative denies participation within the time window of 48 hours, the patient will not be enrolled and screening data will no longer be used. In that case, the PiCCO catheter will be used as arterial line as part of standard care, resulting in no net additional study procedures on top of standard care with the exception of the procedures needed to assess eligibility for the study.

Further consent will be sought from the patient if they recover sufficiently and regain decision-making capacity while in hospital.

11.2.3 No consent in subjects who die before obtaining deferred consent In case a patient dies before informed consent could be obtained from the legal representative, the investigators propose that the data collected during screening and at baseline will no longer be used. This will be recorded as a screening failure, and the reason for failure will be recorded.

11.2.4 Conclusion deferred consent Critically ill subjects in need of ventilation are, without exception, incapable to give informed consent at the moment of ICU admission. Patient screening, eligibility checking and placing the PICCO catheter instead of a "normal arterial line" are part of the study procedures but cannot wait until informed consent is obtained from a legal representative. Therefore, a deferred consent procedure is relevant for the current study. IStudy medication will only be administrered after informed consent.

11.2.5 Consent procedure

The recruitment of study subjects is as follows:

1. Approach of subjects (if patient is a candidate for study participation)
2. In cases where a legal representative is present / accessible at ICU admission: The treating physician asks the legal representative if they would consider participation in a clinical trial for which they may be eligible and obtains their permission, for a member of the study team to contact them
3. If their legal representative is not present / accessible at ICU admission: The treating physician contacts the study team who initiate the deferred consent procedure. During the deferred consent period, a member of the study team will regularly attempt to contact the patient's legal representative to obtain consent.
4. Informed consent The investigator gives the patient information and consent statement to the patient's legal representative answering any questions about the study they may have and providing additional information when requested. The legal representative has at least 6 hours to read and consider whether to provide consent. The informed consent form is signed within 48 hours from ICU admission.