COVID-19 Hyperinflammation Syndrome (COV-HI): Protocol for a Rapidly Executed Cohort Study

Early information from the UK ICNARC (Intensive Care National Audit and Research Centre)- COVID-19 Study Case Mix Programme Database 27 March 2020 report that critical care unit admissions with COVID-19 totaled 775 patients of whom 79 patients have died, 86 patients were discharged alive from critical care and 609 patients were last reported as still being in critical care. It is imperative to reduce this high death rate. Emerging evidence from China, Italy and the US suggests that in a small subset of patients with severe COVID-19 respiratory disease there is a hyperinflammatory syndrome (COV-HI), which may be contributing to the high mortality and morbidity.

Terminology regarding hyperinflammation is heterogenous. When caused by genetic abnormalities in children it is called primary or familial haemophagocytic lymphohistiocytosis (fHLH). Secondary HLH (sHLH) is a hyperinflammatory process driven by infection, rheumatic disorders and malignancy (usually lymphoproliferative disorders). sHLH is termed macrophage activation syndrome (MAS) when associated with rheumatic disease, macrophage activation-like syndrome (MALS) in sepsis, and cytokine release syndrome (CRS) following immune effector cell therapies (haplo-identical allogeneic stem cell transplantation or CAR-T cell therapy). It is likely that such hyperinflammatory states lie on a spectrum and collectively the clinical phenomenon has been called cytokine storm syndrome (CSS) or hyperinflammation (HI).

Current diagnostic criteria for HI are based on fHLH (HLH-2004 guidelines) and the H score (a weighted composite generating a probability). Recently published management algorithms in adults are based on consensus expert opinion and clinical experience, extrapolated from paediatric literature in fHLH.

The HLH Across-speciality Collaboration (HASC) is a 162-strong group of clinicians/scientists who specialize in the management of patients with hyperinflammation. HASC member include Rheumatologists, Haematologists, Intensivists, Infectious Disease specialists, Immunologists, Virologists, Nephrologists and Gatroenterologists from both adult and paediatric backgrounds. Members of HASC have led the set-up of multidisciplinary team meetings (MDTs) to raise awareness of and manage hyperinflammatory illness, completed national audits to demonstrate hyperinflammation is often missed and have set up HLH registries and bioresources (https://research.ncl.ac.uk/ukhr/).

COVID-19 hyperinflammation (COV-HI): rationale for treatment and summary of available evidence In COVID-19 respiratory failure from acute respiratory distress syndrome (ARDS) is the leading cause of mortality. ARDS is a late pre-terminal event in these patients but there is a reported clinical phenotype of severe COVID-19 where hyperinflammation complicates the respiratory failure, termed here COVID-19 hyperinflammation or COV-HI.

A cytokine profile resembling sHLH (a variant of HI) is associated with COVID-19 disease severity, characterised by increased interleukin (IL), -2, -6 -7, -8 granulocyte-colony stimulating factor, interferon-ginducible protein 10, monocyte chemoattractant protein 1, macrophage inflammatory protein 1-α and tumour necrosis factor-α[i]. Predictors of fatality from two recent retrospective, multi-centre studies of confirmed COVID-19 cases included elevated ferritin, c reactive protein (CRP) and serum IL-6. In the first study of 150 confirmed cases, the mean ferritin was 1297.6 ng/ml in non-survivors (n=68) versus 614.0 ng/ml in survivors (n=82); p<0.001) 9. Subsequently in 193 confirmed cases, the mean ferritin was 1453 ng/ml in non-survivors (n=54) versus 503 ng/ml in survivors (n=137) (p<0.0001). Ferritin >300ng/ml (p=0.0038) and raised IL-6 (p=0.0080) were both associated with in-hospital death.

Data publicly available on MedRciv demonstrates that in a retrospective study of 100 patients with COVID-19 pneumonia, disease activity correlated with parameters known to be elevated in HLH, including IL-6 (p<0.001) and ferroprotein (p<0.001). Publicly available data on Chinaciv of 21 patients with severe COVID-19 treated with tocilizumab (anti-IL6 blockade), demonstrated that all patients became afebrile within 24 hours of treatment. In this study following tocilizumab treatment there was a reduction of oxygen requirement (15/20 patients), resolution of CT lesions (19/21 patients), normalization of lymphocyte count (10/19 patients), reduction of CRP levels (16/19 patients) and hospital discharge (19/21 patients) with an average hospitalization duration of 13.5 days. There were no adverse events attributable to tocilizumab. Personal communication from the investigators in Italy and treating physicians within the UK, has confirmed clinical benefit using tocilizumab in a group phenotypically in the COV-HI group, without overt adverse event. Trials are opening internationally using tocilizumab in the COV-HI group.

IL-1 is central to the hypercytokinaemia in HLH. Off-licence anakinra, a recombinant humanised IL-1 receptor antagonist, is recommended in treatment algorithms for HLH. There are active trials looking at the role of anakinra (see for example https://www.sobi.com/sites/default/files/pr/202003183346-1.pdf) and tocilizumab (see for example http://www.chictr.org.cn/showprojen.aspx?proj=494090) in the management of COV-HI patients and insertion of cytokine inhibitor arms into the adaptive platform REMAP CAP (https://www.remapcap.org). There is increasingly reported anecdotal experience of successful use of tocilizumab and anakinra in COV-HI and scientific rationale for plausibility.

Taken together, these findings suggest that high mortality may be due to virally-driven hyperinflammation in COVID-19 (COV-HI), that could be identified with routinely available clinical and laboratory data. Of particular note are C-reactive protein - CRP a surrogate for IL-6 - and ferritin, both tested in general medical assessment of unwell patients. Outcome could be improved if the COV-HI phenotype can be characterised; such phenotyping will inform trial design (in which HASC members are heavily involved) as well as identifying patients in hospitals overwhelmed by clinical cases who can be recruited to those trials. In particular, understanding the clinical course of those admitted to hospital (who by definition have severe COVID-19 and are at risk of COV-HI) and the inflexion point at which COV-HI emerges would enable targeted treatment. Such characterisation of phenotype is currently not available but HASC have mobilised the UK network to collect the data to inform this gap in the evidence.

Identifying people who, at the point of needing respiratory support, could potentially benefit from immune modulation may be possible by interrogating continuous clinical & laboratory data collected from hospital admission to the time of recovery/death.

This will be a retrospective cohort study. Due to the urgent nature of this research and the need to rapidly gather information about each patient's clinical experience and outcome during the COVID-19 pandemic it will not be possible to gain informed written consent.

It is proposed to:

1. Collect demographic information on consecutive people admitted to selected sites across the UK (initially UCLH, Leeds, Newcastle, Sheffield) with COVID-19 starting with the institution's first case. Hospital admission is proxy for severe COVID 19. Map their clinical journey by collecting serial clinical and laboratory data from routine clinical investigations during the course of their illness at 24-hour intervals

2. Retrospectively analyse data to try to identify patients who subsequently fit the COV-HI phenotype. Look in detail at these cases to understand whether there is an inflexion point early in their disease course such that a point for intervention with immune modulation could be have been made.

   It will be important to differentiate between people deteriorating due to organ failure because of co-morbidity and poor reserve, compared to those with overwhelming inflammatory response fitting with the proposed hypothesis. For the former group, the risk associated with immune modulation is likely to outweigh the benefit.

3. Use data collection to define criteria that could be used as threshold for treatment

De-identified data will be accessed by all members of the team who have contributed their patients to the study and to collaborators within the UK for statistical analysis. Patient age will be retained in the de-identified data as this is an essential clinical variable but NHS number, DOB, ICNARC number and date of admission will not be visible to members of the team who are not the normal care team of a given patient. The risk of identification of a patient from their age and diagnosis alone is very low.

Everyone with access to person-identifiable information will hold an NHS contract or honorary contract and have received necessary training regarding GDPR and Good Clinical Practice, and will be fully aware of their responsibilities to patient confidentiallity.

Understand and comply with the law. This has been set up as a research study using project specific data stored on the REDCap platform at Newcastle University. The data will be held on secure servers with dual password protection (login and SMS OTP) at Newcastle University.