The Chimeric Antigen Receptor-Intensive Care Unit (CAR-ICU) Initiative: Surveying Intensive Care Unit Practices in the Management of CAR T-Cell Associated Toxicities

Abstract

Purpose:

A task force of experts from 11 United States (US) centers, sought to describe practices for managing chimeric antigen receptor (CAR) T-cell toxicity in the intensive care unit (ICU).

Materials and Methods:

Between June-July 2019, a survey was electronically distributed to 11 centers. The survey addressed: CAR products, toxicities, targeted treatments, management practices and interventions in the ICU.

Results:

Most centers (82%) had experience with commercial and non-FDA approved CAR products. Criteria for ICU admission varied between centers for patients with Cytokine Release Syndrome (CRS) but were similar for Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS). Practices for vasopressor support, neurotoxicity and electroencephalogram monitoring, use of prophylactic anti-epileptic drugs and tocilizumab were comparable. In contrast, fluid resuscitation, respiratory support, methods of surveillance and management of cerebral edema, use of corticosteroid and other anti-cytokine therapies varied between centers.

Conclusions:

This survey identified areas of investigation that could improve outcomes in CAR T-cell recipients such as fluid and vasopressor selection in CRS, management of respiratory failure, and less common complications such as hemophagocytic lymphohistiocytosis, infections and stroke. The variability in specific treatments for CAR T-cell toxicities, needs to be considered when designing future outcome studies of critically ill CAR T-cell patients.

INTRODUCTION:

Chimeric antigen receptor (CAR) T-cell therapy has proven to be a promising area in cancer treatment. Durable remissions in up to 50% of patients have been demonstrated in relapsed refractory acute lymphoblastic leukemia (ALL), large B-cell lymphoma, mantle cell lymphoma and most recently in multiple myeloma.^1–5 Toxicities associated with CAR T-cell therapy, particularly cytokine release syndrome (CRS) and Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS), are reversible but can be associated with significant morbidity, with up to 47% of patients requiring Intensive Care Unit (ICU) admission.^1,2,4,6,7 CRS presents as a range of clinical symptoms including fever to hypoxia and/or hypotension which can progress to circulatory shock, acute respiratory failure and multi-organ failure ^8–10. ICANS is a complex syndrome characterized by varying degrees of encephalopathy, dysgraphia, aphasia, seizures, motor deficits and, in severe cases status epilepticus and cerebral edema.^8,9,11 Numerous guidelines for grading and management of these toxicities exist, ^{1,6,8,9,12–15} and a recent consensus has achieved some homogeneity in grading of CRS and ICANS across centers.⁸

Despite the high acuity of illness observed with CRS and ICANS, there is limited data and consensus in management of critically ill patients with CAR T-cell therapy toxicities. Current treatment recommendations of grade 3 and 4 toxicities are based on clinical experience and extrapolation of practices and guidelines from other causes of critical illness (e.g. sepsis).¹² It is unclear how these practices and interventions in the ICU affect the outcomes of this patient population, where the cause of CRS and ICANS is different from common causes of shock, respiratory failure and encephalopathy. In order to improve ICU care, data driven interventions are necessary.

In an effort to address these knowledge gaps, we created a working group of critical care, hematology-oncology and infectious disease specialists from 11 centers in the United States (US) called the CAR-ICU initiative. Our focus is to conduct research on monitoring and management strategies to improve the care of critically ill patients following CAR T-cell therapy. In order to optimize the care of these patients, we must first understand current practices in the ICU management of CAR mediated toxicities. Utilizing a survey, we assessed the structure and practices of care for CAR patients in different ICUs. Our results, reported here, identified areas for further research into novel interventional strategies that could improve patient outcomes.

MATERIALS AND METHODS:

This study was conducted in accordance with the amended Declaration of Helsinki. Local institutional review boards or independent ethics committees approved the protocol; the study was approved by the Institutional Review Board (PA19-0289) at MD Anderson Cancer Center (MDACC). This survey was developed and reviewed by a multidisciplinary group of physician and pharmacy experts in CAR T-cell toxicity assessment and treatment. The survey was distributed electronically through REDCap hosted by MDACC.^16,17 The survey tool addressed: a) institutions and their ICUs; b) CAR products, the toxicities observed and targeted treatments; c) specifics of practices and interventions within the ICU when caring for critically ill CAR patients (Supplement 1). All data collected examined the latest ICU practices, from initiation of each center’s CAR program until distribution of the survey (June 2019) and did not include any patient specific data or protected health information (PHI). Practices in the pediatric population were excluded. No interventions were performed after analysis of these data. Each survey evaluated the center’s protocols and practices but not individual clinician practices.

Within the US, 15 of the largest centers participating in research and treatment of adult patients with CAR therapy were invited to participate in the CAR-ICU initiative; 11 centers (73%) replied to the invitation and joined the working group. These 11 centers are well recognized, large oncological centers that play an important role in research and treatment with CAR cell therapy in the US; 10 of the centers are accredited by the Foundation for the Accreditation of Cellular Therapy (FACT). During the period of June 21 to July 30, 2019, one set of survey questions was distributed to each of the participating centers via electronic mail; key ICU and hematology-oncology representatives of each institution completed the survey. Survey responses were analyzed using standard descriptive statistics.

RESULTS:

All 11 hospitals (100%) completed the survey. Participating centers included: MD Anderson Cancer Center, Memorial Sloan Kettering Cancer Center, Seattle Cancer Care Alliance/University of Washington Medical Center, Massachusetts General Hospital, Barnes Jewish Hospital, National Institutes of Health Clinical Center, Mayo Clinic, Roswell Park Comprehensive Cancer Center, University of Miami Cancer Center, The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins Hospital and Cleveland Clinic. Specifics of each participating center are summarized in Table 1. Two centers had specialized intensivists taking care of CAR patients, while the rest of the institutions had trained all ICU practitioners to treat this population. Nine centers (82%) had a dedicated committee to monitor, review and create guidelines for the care of CAR T-cell patients. The number of total treated patients varied within centers and on average 2 patients a month required ICU admission post-CAR therapy (range 1-4) (Table 1).

Table 1.

Characteristics of Participating Centers and CAR Products

Characteristic	Centers, n(%) (n=11)
Total Hospital Beds
1-50	1 (9)
50-100	1 (9)
100-250	2 (18)
250-500	1 (9)
>500	6 (55)
Total Medical ICU Beds
1-10	1 (9)
11-20	4 (36)
21-30	1 (9)
> 30	5 (46)
Type of ICU
Exclusively oncological	8 (73)
Mixed	3 (27)
CAR Products Used
Commercial	9 (82)
Clinical trial protocols	9 (82)
Institutional products	4 (36)
Toxicity Grading Guidelines Utilized
ASTCT	10 (91)
Other	1 (9)
Treatment Guidelines Utilized
Institutional	5 (46)
Neelapu 20179	3 (27)
Lee 201413	2 (18)
Specific to the clinical trial protocol	1 (9)
CAR Infusions to Date*
Commercial products
0-50	6 (55)
51-100	3 (27)
>100	2 (18)
Non-commercial products
0-50	7 (64)
51-100	0 (0)
>100	4 (36)
Infection Prophylaxis Guidelines for CAR patients	7 (64)

^*Patients treated to date as of distribution of survey (June 2019)

ASTCT - American Society for Transplantation and Cellular Therapy; CAR - chimeric antigen receptor; ICU - intensive care unit;

Specifics of CAR T-cell products:

The type of CAR products used at each institution varied (Table 1). Nine centers (82%) were using commercial products: axicabtagene ciloleucel and/or tisagenlecleucel. Ongoing protocols from pharmaceutical companies were being studied at 9 centers (82%); 4 centers (36%) had products created at their institution. Three of the 11 centers (27%) had experience using all 4 different types of products. In addition to treating patients with CAR T-cells, 7 centers (64%) were also treating patients with T-cell receptor (TCR) transduced T-cell therapy. Ten centers (91%) currently use the American Society for Transplantation and Cellular Therapy (ASTCT) grading system for CRS and ICANS.⁸ However, significant variability was observed in treatment guidelines used across centers (Table 1).

Criteria for ICU admission:

The ASTCT grading for CRS and ICANS was used to assess criteria for ICU admission across centers (Table 2).⁸ Eight centers (73%) admitted patients with grade 1 and 2 CRS for close monitoring. One institution admitted all patients being infused with CARs to the ICU because of no available intermediate care unit to monitor these patients. Patients requiring vasopressors, continuous renal replacement therapy (CRRT), high flow nasal cannula (HFNC), bilevel positive airway pressure ventilation (BiPAP) or mechanical ventilation during CRS were admitted to the ICU. Other reasons for admission included: 1) patients who received fluid resuscitation and whose blood pressure had not returned to baseline or concerns for further deterioration and need for further interventions were present; 2) patients with increasing oxygen requirements, raising concerns for evolving non-cardiogenic pulmonary edema; 3) patients at risk for rapid deterioration due to large tumor burden; and 4) concurrent ICANS and CRS that required closer monitoring of respiratory status.

Table 2.

ICU Admission Criteria and ICU Management Practices of CAR T-Cell Patients

Characteristic	Centers, n(%) (n=11)
ICU Admission Criteria for CRS
Grade 1-2	8 (73)
Grade ≥ 3	11 (100)
ICU Admission Criteria for ICANS
Grade 1	1 (9)
Grade 2	9 (81)
Grade ≥ 3	10 (91)
ICE score ≤3	10 (91)
Seizures
Easy to control (Grade 3)	9 (81)
Status epilepticus (Grade 4)	11 (100)
Motor deficits	8 (73)
Cerebral edema	11 (100)
ICU Management Practices:
Fluid Resuscitation Guidance
mL/kg fluid bolus	8 (73)
Ultrasound guided	6 (55)
Pulse pressure variation on arterial line	4 (36)
Non-invasive modes of SVV, CO, SVR	5 (46)
Other	2 (18)
Choice of Fluid Resuscitation
Normal saline	10 (91)
Lactated ringers	7 (64)
Albumin 5%	5 (46)
Other	2 (18)
Vasopressor of Choice
1st line
Norepinephrine	11 (100)
2nd line
Vasopressin	9 (82)
Phenylephrine	2 (18)
3 rd line
Phenylephrine	3 (27)
Vasopressin	2 (18)
Epinephrine	6 (55)
Inotrope of Choice
Dobutamine	8 (73)
Epinephrine	2 (18)
Clinical situation dependent	1 (9)
Respiratoiy Failure Treatment
Trial of BiPAP prior to intubation	8 (73)
<12 hours	5 (46)
12-24 hours	2 (18)
24-48 hours	1 (9)
Transition from HFNC to MV	3 (27)
Brain Edema/Intracranial Pressure Monitoring
Fundoscopy	3 (27)
EVD	2 (18)
Ocular ultrasound	2 (18)
Brain imaging	7 (64)
Centers that have treated cerebral edema	9 (82)
Cerebral Edema Management	(n=9)
Acetazolamide	3 (33)
Hypertonic saline	7 (78)
Mannitol	5 (56)
Pharmacologic coma	1 (11)
Short course of hyperventilation	3 (33)
EVDs to guide CPP	1 (11)
Hypothermia	0 (0)

BiPAP- bilevel positive airway pressure ventilation; CAR - chimeric antigen receptor; CPP- cerebral perfusion pressure; CO- cardiac output; EVD- external ventricular devices; HFNC- high flow nasal cannula; ICU-Intensive Care Unit; MV- mechanical ventilation; SVR-systemic vascular resistance; SVV-stroke volume variation

Criteria for ICU admission for ICANS was similar between centers (Table 2). Of the 11 centers, 9 (82%) admitted patients with grade 2 neurotoxicity for close monitoring. Ten centers (91%) admitted patients with depressed level of consciousness or with an Immune Effector Cell-Associated Encephalopathy (ICE) score ≤3. Patients with seizures who responded quickly to anti-epileptic drugs (AEDs) were admitted to the ICU for close monitoring in 9 centers (81%). All centers admitted patients with difficult to control seizures, status epilepticus or cerebral edema on imaging (brain CT or MRI) to the ICU.

CAR toxicities and management:

Assessment of practices specific to the treatment of CAR related toxicities was determined by evaluating the use of tocilizumab, siltuximab, anakinra and corticosteroids (Table 3). Use of tocilizumab, was similar amongst institutions and limited to CRS, or neurotoxicity with accompanying CRS symptoms (n=9; 82%). Siltuximab, utilized in 6 centers (55%), was reserved for patients with grade 3-4 CRS refractory to tocilizumab. Anakinra was used in 6 centers (55%) and commonly for refractory grade 3-4 CRS or neurotoxicity (Table 3). Dosing and choice of corticosteroids when treating CRS and ICANS varied amongst centers (Table 3). Although, nine centers did not follow a defined protocol for corticosteroid taper, their practice was to taper rapidly following symptom resolution while monitoring for recurring symptoms. For refractory cases, institutions had used rescue treatments such as anti-thymocyte globulin, suicide genes, ruxolitinib, infliximab, basiliximab and cyclophosphamide (Table 3).

Table 3:

Corticosteroid and Anti-Cytokine Therapy Practices for CAR Related Toxicities

Therapy	Centers, n(%) (n=11)
Tocilizumab Use
≤3 doses	11 (100)
Dosing frequency (every 8 hours)	10 (91)
Neurotoxicity use
Never	2 (18)
Only when associated with CRS	9 (82)
Siltuximab Use
Ever used	6 (55)
Refractory to tocilizumab	5 (46)
Grade 3-4 CRS	1 (9)
Anakinra Use
Ever Used	6 (55)
Indication
Refractory grade 3-4 CRS	5 (45)
Refractory grade 3-4 ICANS	5 (45)
Refractory to IL-6 therapy	1 (9)
All grade 3-4 patients	1 (9)
Corticosteroid Dosing
Dexamethasone 10mg every 6 hours	10 (91)
Dexamethasone 20mg every 6 hours	3 (27)
Methylprednisolone 250mg every 12 hours	4 (36)
Methylprednisolone 500mg every 12 hours	5 (45)
Higher doses if needed	5 (45)
Rescue Therapies
Ever used	4 (36)
ATG	1 (9)
Suicide genes	2 (18)
Ruxolitinib	1 (9)
Infliximab	1 (9)
Basiliximab	1 (9)
Cyclophosphamide	1 (9)

ATG – anti-thymocyte globulin; CRS – cytokine release syndrome; ICANS – immune effector cell-associated neurotoxicity syndrome.

C-reactive protein (CRP) and ferritin were measured daily at 10 centers (91%). Measurement of serum cytokine levels was performed routinely in 4 centers (36%) and as needed in 4 others (36%). All institutions indicated that measuring CRP, ferritin and cytokines could help guide management of CAR patients. Seven centers (64%) had the capability to perform flow cytometry to examine CAR expansion (Table 4)

Table 4.

Diagnostic Tests for CAR T-Cell Related Toxicities

Procedure	Centers, n(%) (n=11)
Lumbar Puncture	6 (55)
EEG Timing
Onset of neurotoxicity	3 (27)
Clinical suspicion of seizure	5 (46)
ICANS > grade 2	3 (27)
Routine MRI Brain
Yes	2 (18)
No	1 (9)
Attempt, but not always feasible	8 (73)
Daily CRP/Ferritin	10 (91)
Serum Cytokines
Yes, routine	4 (36)
No (not available)	3 (27)
As needed	4 (36)
Flow Cytometry for CARs	7 (64)

CAR - chimeric antigen receptor; CRP – c-reactive protein; ICANS – immune effector cell-associated neurotoxicity syndrome; EEG – electroencephalogram; MRI - magnetic resonance imaging

Interventions in the Intensive Care Unit:

Interventions for hemodynamic support were similar between all institutions. All centers used norepinephrine as the first vasopressor agent for grade 3 CRS but some variability was observed when choosing a second and third vasopressor (Table 2). Choice of intravenous fluids for resuscitation and assessment of volume responsiveness varied widely between institutions (Table 2). Seven centers (64%) routinely perform echocardiograms once shock occurs during CRS. Non-cardiogenic pulmonary edema was reported as a frequent finding in 7 institutions (64%). Ventilatory support practices for patients with respiratory failure varied widely between institutions (Table 2). Eight centers (73%) routinely gave a trial of BiPAP, and there was significant variability when describing its duration.

Practices of supportive care for specific presentations of ICANS such as seizures, status epilepticus and cerebral edema were surveyed (Table 2). Criteria for performing a diagnostic electroencephalogram (EEG) varied between institutions (Table 4). Seizure prophylaxis with anti-epileptic drugs (AEDs) was used in the majority of centers (n=10; 91%). The decision to introduce AEDs for seizure prophylaxis depended on factors such as the type of CAR product, prior neurological comorbidities, protocol specific requirements and high-risk patients (i.e. high disease burden). Methods used to monitor for signs of elevated intracranial pressure and cerebral edema included fundoscopy, ocular ultrasound and external ventricular devices (Table 2). Most centers (n=8; 73%) recognized the challenges in performing brain MRIs in this patient population, mostly due to the degree of encephalopathy, agitation necessitating sedation and endotracheal intubation for airway protection (Table 4). Despite this, nearly two-thirds of centers (64%) utilized brain imaging (mainly CT) to monitor for cerebral edema. Lumbar punctures were performed at 6 centers (55%) when there was no significant improvement of neurotoxicity and infectious causes needed to be ruled out. Nine (82%) centers had treated patients with cerebral edema and findings of focal versus diffuse cerebral edema on imaging influenced their decision regarding level of aggressive medical treatment (excluding corticosteroids) at 6 of these centers (64%). Amongst the 9 centers who had treated patients with cerebral edema, treatment practices varied (Table 2). Supportive care guidelines specific for CAR patients for management of cerebral edema, seizures and status epilepticus were available at 6 centers (55%).

Other ICU interventions used included renal replacement therapy at 9 institutions (82%). Cases of hemophagocytic lymphohistiocytosis (HLH) were observed at 7 centers. Other adverse events noted during ICU admission included infectious complications, hypogammaglobinemia, disseminated intravascular coagulation, thrombotic events and tumor lysis syndrome. Ischemic and hemorrhagic strokes concomitantly with neurotoxicity were observed at 5 centers (46%). Limitation of care and Do Not Resuscitate (DNR) conversations occurred at 7 centers.

DISCUSSION:

Our survey showed that participating centers in the CAR-ICU initiative use similar ICU admission criteria for neurotoxicity but not for CRS. Practices for vasopressor administration, neurotoxicity monitoring, utilization of EEG, prophylactic AEDs and tocilizumab use were comparable. In contrast, fluid resuscitation, respiratory support, methods of surveillance and management of cerebral edema, corticosteroid and other anti-cytokine therapies varied within centers. Our survey identified areas of collaboration and future research that could impact outcomes of critically ill CAR T-cell patients: 1) choice of intravenous fluids and vasopressor agents for CRS; 2) optimal management strategies of respiratory failure; and 3) recognition and management of other complications such as infections, HLH and stroke.

Hypotension and shock are the most common clinical findings of CRS.^5,18 The majority of CAR T-cell patients present with hypotension that responds to intravenous fluids, but approximately 20% of patients develop ≥grade 3 CRS, necessitating hemodynamic support with vasopressors.^2–4,7,18 Although tocilizumab and corticosteroids are the mainstays of treatment of CRS, fluid resuscitation and vasopressors are crucial to prevent further end-organ damage. Recent data suggest that catecholamine feedback loops may play an essential role in CAR T-cell mediated CRS.¹⁹ While these data are preliminary, it raises the question whether non-adrenergic vasopressors such as vasopressin and angiotensin-II, may be beneficial in this patient population. Our survey revealed that norepinephrine (adrenergic agonist) is the first-line vasopressor used for hemodynamic support in accordance with traditional clinical practice guidelines, followed by vasopressin.^20,21 Homogeneity on current practices could facilitate future research, and collection of baseline data from our institutions could support the design of such a study.

We observed significant variability in the choice of intravenous fluids for resuscitation. The choice of either normal saline, balanced solutions or albumin for critically ill patients is a topic of debate and can be an area of investigation in CAR T-cell patients. The use of normal saline over balanced solutions can increase the risk of hyperchloremic metabolic acidosis, acute kidney injury (AKI) and possibly increase mortality.^22–24 Patients with CRS are at risk of AKI by direct inflammatory injury, fluid shifts, hypoperfusion and shock. Therefore, evaluation of the choice of intravenous fluid for resuscitation and its effects on AKI could be valuable in this patient population. Additionally, low levels of serum albumin and increased capillary leak syndrome are described in patients with high grade CRS.¹⁸ Few studies have demonstrated that albumin administration, especially in patients with hypoalbuminemia, can reduce days on vasopressors and decrease respiratory, cardiovascular and neurological organ failure scores, all possibly beneficial in CAR patients.^25–27 Moreover, the use of albumin could reduce the severity of non-cardiogenic pulmonary edema that occurs during CRS, since data in the general ICU population suggest albumin can be beneficial in patients with acute lung injury.^28,29 Lastly, endothelial dysfunction is characteristic of patients with CAR related toxicities,^11,18,30 and use of albumin has shown to reduce endothelial dysfunction during inflammatory processes.^31,32 Thus, it is possible that albumin could have more favorable effects than crystalloids for fluid resuscitation in these patients.

The variability in management of respiratory failure that we observed in our survey was consistent with what is described in the literature when treating cancer patients with respiratory failure. Overall, the use of mechanical ventilation in cancer patients increases mortality when compared to use of non-invasive ventilation (NIV).^33,35 However, cancer patients who do not improve with BiPAP therapy and require intubation have increased mortality, reported to be as high as 80%.^33,36,37 Unfortunately, data defining the cohort of patients who benefit from atrial of BiPAP is limited. In our survey, 7 institutions believe that non-cardiogenic edema was frequent in CAR T-cell patients, however earlier studies showed that only 7-13% of patients require mechanical ventilation.^2,4,6,7,18 The inconsistency between this observation and available data has to be carefully examined by further outlining the etiology of respiratory failure in CAR patients. While the low incidence of respiratory failure requiring mechanical ventilation in these patients may limit future interventional studies, details describing the etiology, type of ventilatory support including HFNC and BiPAP and outcomes are necessary.

Our survey demonstrated significant variability in the indications, timing and dosing of corticosteroids and use of anakinra, siltuximab and salvage therapy for refractory cases. Discussions between oncologists and ICU providers probably influence the threshold to initiate corticosteroids and other rescue therapies in complex cases. Balancing the desire to resolve CRS and neurotoxicity, while considering the effect corticosteroids and other therapies on long term remission, will be a continuous debate until additional research on their in-vivo effect on CAR T-cell proliferation and response rates is available. Ongoing efforts to create consensus guidelines on management for CAR T-cell toxicities will be valuable; however due to the variability of products, malignancies and severity of toxicities, creating a “one-size fits all” recommendation might not be feasible. As a result, variations in toxicity management need to be considered when designing multicenter studies and a careful choice of measurable outcomes is extremely important. Consider mortality, which may not be a feasible primary outcome in these patients as mortality associated with CAR related toxicities is low (1-3.8%) resulting in poor sensitivity to detect the impact of interventions being tested.^7,18 Secondary outcomes such as days on vasopressors, need for mechanical ventilation, and incidence of specific organ failure and ICU length of stay may be considered. Interventions influencing these variables could reduce cumulative doses of corticosteroids and possibly decrease infectious complications, critical illness myopathy/neuropathy and reduce inhibition of CAR proliferation and positively impact remission rates.

Our study has several limitations. First, the limited number of centers (n=11) may not be representative of the larger ICU community however it is an important start, especially considering that the participating centers have extensive experience managing CAR T-cell patients. As this treatment becomes widely available, assessment of differences not only within centers, but also within clinician groups, might become necessary. Second, the self-reporting nature of the survey could lead to desirability bias; therefore, objective data collection will be vital within this collaborative. Third, most centers use CD 19 products which could impact some of the results in practices observed in the survey. Lastly, additional details on more specific ICU interventions during neurotoxicity, the specifics of CAR products and treated malignancies at each center, were limited. It is important to note that specific clinical scenarios were not evaluated in the survey; therefore, some of the reported variability within centers, especially when it comes to interventions in the ICU such as initial choice for fluid resuscitation or respiratory support could also be due to variability of clinical scenarios commonly observed at each institution. With this in mind, one has to be careful when designing future studies to evaluate these differences. These variables were not measured due to constraints in the length of the survey but need to be addressed with objective data. Despite these limitations, this CAR-ICU initiative survey has identified priority areas of research and future interventions that may potentially standardize and create consensus in ICU care. Ongoing efforts by our group to collect patient data will help determine the focus of future research initiatives, and outcomes to be measured that could positively impact these patients.

CONCLUSION:

The CAR-ICU initiative survey revealed important similarities as well as differences in the management of CAR patients in select ICUs across the US. The variability in practices created opportunities for possible future collaborative investigations. Continuing this initiative will help us gather baseline data on critically ill CAR patients and better design interventions that could help improve outcomes of this complex patient population.

HIGHLIGHTS:

• Despite the high acuity of illness of CRS and ICANS, there is limited data in its ICU management

• The CAR-ICU (11 leading institutions within the US) focuses on studying critically ill CART-cell patients

• Our survey reports the structure and practices of care for CAR patients in different ICUs

• Practices in the management of CART-cell patients varies within institutions in the US

• We identify areas of future research that could impact outcomes of critically ill CAR T-cell patients.

Abbreviations list:

CAR

Chimeric antigen receptor

ICU

Intensive Care Unit

ALL

acute lymphoblastic leukemia

CRS

cytokine release syndrome

ICANS

Immune Effector Cell Associated Neurotoxicity Syndrome

ARDS

acute respiratory distress syndrome

ASTCT

American Society for Transplantation and Cell Therapy

CRRT

continuous renal replacement therapy

HFNC

high flow nasal cannula

BiPAP

Bilevel positive airway pressure ventilation

ICE

Effector Cell-associated Encephalopathy

CARTOX-10

CAR Toxicity-10

AEDs

anti-epileptic drugs

ATG

anti-thymocyte globulin

EEG

electroencephalogram

EVDs

external ventricular devices

HLH

Hemophagocytic lymphohistiocytosis

DNR

Do Not Resuscitate