I-SPY COVID-19 TRIAL: An Adaptive Platform Trial for Critically Ill Patients

This platform trial will provide access to repurposed and investigational agents for

critically ill patients infected with SARS-CoV-2 who have severe or life-threatening

COVID-19. The main focus of this trial is a platform study for identifying effective

agents for the treatment of COVID-19. Any critically ill patient with known or presumed

COVID-19 will be automatically entered into the screening phase of the trial until

SARS-CoV-2 infection is confirmed. Basic data will be assembled for each patient (such as

ventilatory status and survival). If interested in the therapeutic portion of the trial,

potential participants will be asked to sign a consent form describing the backbone

treatment and the two specific investigational agent arms to which they may be

randomized. The primary endpoints will be time to recover to a durable level 4 (or less)

on the WHO COVID-19 ordinal scale for clinical improvement and time to mortality (death).

For this trial, a durable level 4 is defined as at least 48 hours at COVID level 4 or

less (nasal prongs oxygen) without returning to high flow oxygen or intubation. Acute

care facility resource utilization will be automatically calculated (total length of stay

in a critical care setting, days intubated, and survival). Any change in status,

including intubation, extubation, death or discharge, will be recorded and verified by

the attending physician.

Patients will be evaluated based on their initial status (ventilation at entry vs. high

flow oxygen). Exploratory biomarkers will be evaluated over time (ARDS phenotypes and

other proposed markers) to facilitate clinical learning. A maximum of two investigational

arms may be open at a time. The anticipated accrual will be 50 patients per week. The

maximum number of participants assigned to an arm without graduation will be 125

patients. Agents can be dropped for futility after enrollment of 40 patients. As the

trial proceeds and a better understanding of the underlying mechanisms of the COVID-19

illness emerges, expanded biomarker and data collection can be added as needed to further

elucidate how agents are or are not working. The study design features comparison of

investigational agent efficacy using a Bayesian design, which will allow the detection of

strong efficacy signals with the fewest possible patients. Initially the control will be

patients given current standard of care (supportive care for ARDS, including lung

protective ventilation and remdesivir and dexamethasone as backbone therapy). As other

treatments (for example, anticoagulation) become part of standard supportive care across

sites, these will be added to the backbone therapy. If an agent meets the threshold for

graduation the company leadership will be informed as will the FDA. The arm with the

graduated agent will cease to enroll, allowing a new arm with a different investigational

agent to be added.

Every trial participant will have blood collected at trial enrollment, day 3, and day 7

for pre-specified biomarker and DNA and RNA analysis. Additional biomarkers can be added

as the trial proceeds. Patient outcomes will also be evaluated on the basis of whether

patients are ventilated initially or not.

Observational Component:

Initially, all COVID-19 confirmed patients who started high-flow oxygen (WHO COVID-19

level 5; ≥6L oxygen by nasal prongs or mask) were entered in an Observational Component

which collected data via extraction of medical records. Patients in this Observational

Component also had their daily COVID status and drug administration form CRFs completed.

An expanded Observational Study will replace the original Observational Component. The

expanded observational study (Supplement 1) will collect blood sample(s) and clinical

data from ARDS and AHRF patients (including COVID ARDS patients) to test the feasibility

of quantifying a set of biomarkers that will allow each patient to be classified into

either a hyper-inflammatory or hypo-inflammatory subtype in real time. If treatment of

these critically ill ICU patients is to be guided by subtype classification, it is

essential that the operational time for classification is as quick as possible.