This project aims to undertake a high-resolution analysis of previous cases of COVID19 infection to determine from all available clinical data recorded over the timeline of admission, who might benefit best from specific interventions designed to target a hyper inflammatory response in this condition. This approach offers a timely contribution to the field where the first phase of unselected clinical trials is already underway, and the second phase will require a more targeted approach. Southampton offers a unique opportunity to undertake this work, as it is widely accepted that the investigators lead the UK in clinical data informatics. This project links the investigators research efforts at the clinical level to current understanding of disease pathways, for which the investigators have effective interventions and the problem the investigators aim to solve is, who will benefit from the available novel anti-inflammatory approaches and when should this treatment be given?

Purpose:

The primary purpose of this research is to identify an early warning for hyperinflammation in COVID19 so that interventional trials of anti-inflammatory agents can target this sub-group.

Background Mortality from SARS-CoV-2 infection causing COVID-19 is estimated to be 3.7% globally. The principal cause of death due to COVID-19 is respiratory failure due to acute respiratory distress syndrome. Early reports have suggested that a subgroup of individuals suffer a hyperinflammatory state with a high mortality which is associated with high levels of IL-6 and CRP. Recent randomised controlled trial data has shown that the anti-inflammatory agent dexamethasone can reduce mortality in severe COVID-19 in an unselected COVID-19 cohort. Whilst impressive, these results suggest that if the anti-inflammatory interventions could be targeted early to individuals with hyperinflammation, even greater benefit in mortality may be seen, and this approach may additionally reduce the morbidity of COVID-19 by preventing escalation to high dependency and intensive care. The purpose of our study is to identify hyperinflammation early because there are specific therapies which are in clinical use which treat hyperinflammation.

Hyperinflammation has been previously described secondary to acute infection and termed cytokine release syndrome / cytokine storm (CRS / CS), macrophage activation syndrome (MAS), macrophage-cytokine self-amplifying loop (MCSAL) and secondary haemophagocytic lymphohistiocytosis (sHLH). Viral infections are the commonest cause of sHLH, and symptoms of hyperinflammation resemble those of general sepsis, therefore hyperinflammation has generally been under-recognised at an early stage leading to a high mortality. Blockade of the key pathways in hyperinflammation such as IL-1 have been shown to be effective in sepsis triggered cases without increased adverse events . Early data in a single arm open label study of 21 severely ill COVID-19 patients with increased IL-6 expression, showed that inhibition of IL-6 signaling with tocilizumab caused rapid clinical improvement in 75% of cases, with some improvement in all cases .

During COVID19 infection, the initial viral invasion of epithelial tissues causes direct cytotoxic damage but subsequently progresses to an inflammatory response at 7-10 days. While an inflammatory response is likely to be needed to eliminate virus, there is a risk of ongoing activation with inflammatory mediators causing end organ damage. Therefore, it seems that targeting hyperinflammation (e.g. IL-6) during acute SARS-CoV-2 infection requires the timing of blockade to be finely balanced to avoid early on impairment of host antiviral responses. Similarly, intervening too late is likely to show minimal benefit. Due to the urgency, despite this uncertainty, clinical trials to address the effectiveness of anti-IL-6 are already underway with Tocilizumab (ClinicalTrials.gov Identifier: NCT04317092, REMAP-CAP, RECOVERY) and Sarilumab (ClinicalTrials.gov Identifier: NCT04315298). The entry criteria for the UK trials of such interventions have largely been unselected, ie not specifically targeting individuals who have hyperinflammation.

However, it is not yet clear how to identify those COVID-19 cases in whom anti-hyperinflammation strategies are required. Retrospective examination of IL-6 expression in 150 cases of COVID-19, estimated that in survivors median IL-6 level was approximately 6.9ng/ml (5-8.8 ng/ml), whereas in those who died was 11.25 ng/ml (7.75-16.25) (p<0.001) . Overall, 46 cases with IL-6 >7.75 ng/ml 63% died whereas 47 cases IL-6 <7.75ng/ml showed only 23% mortality. Similar data were identified in a second study to examine IL-6 levels, but only showed these to be elevated in 6 out of 48 cases (mortality 50% in IL-6 high) . This suggests that IL-6 expression levels may provide a useful biomarker of outcome, but IL-6 measurement alone is not sufficient and additionally this is not widely available. Although dexamethasone has been reported to reduce mortality in unselected cases with severe COVID-19, dexamethasone has a well-known side effect profile and risk of complications. Indeed, it is not surprising that for those patients with less severe COVID the non-significant trend was that dexamethasone caused harm (death compared to controls, OR 1.22, [0.86 to 1.75]). Therefore, it is likely that strategies to identify hyperinflammation and targeted intervention will offer the most effective approach to management of COVID-HI. Indeed, as well as anti-IL-6 other cytokines released in hyperinflammation for which existing biologic therapies are available are also potential targets for intervention including TNF-α (Infliximab), IL-1 (Anakinra), and JAK-inhibitors (e.g. Ruxolitinib). In addition, secondary events may prove potential check-points in hyperinflammation for blockade and IL-6 is also known to regulate Th17 differentiation and induction of CD8 cytotoxic T cells which are important in anti-viral responses.

To facilitate diagnosis of hyperinflammation, the 'HScore' which is a validated score and was developed in the context of secondary sHLH because of evidence that early recognition and intervention is beneficial. Some authors have recommended using the HScore in COVID-19 to identify hyperinflammation. However, very few COVID patients fulfil the current HScore criteria. So, it is unclear whether COVID hyperinflammation (COVID-HI) is an attenuated form of sHLH but with a similar mechanism or whether it is a distinct entity. Indeed, there are unique aspects to SARS-CoV-2 infection such as the unusual coagulopathy that arises in up to 67% of cases which suggests that perhaps COVID-HI is a unique entity.

This proposal aims to examine the HScore parameters in NHS clinical and laboratory records of patients admitted to Southampton with confirmed SARS-CoV-2 infection, with an aim to determine the relevance to COVID19, and will explore whether other routine NHS parameters could be incorporated to enhance the sensitivity of hyperinflammation diagnosis. To facilitate a validation of this scoring approach, the study will confirm how such a scoring system differentiates between the COVID-19 cohort and control cohort of hospitalised patients stratified for key parameters. This analysis primary objective is to develop a clear algorithm (COVID-HI score) for identification of the COVID19 subset with COVID-HI, and expect these results will inform the next phase of COVID19 interventional clinical trial design in anti-hyperinflammation therapy, such as the ACCORD study (CI, Prof T Wilkinson, Southampton) . To identify hyperinflammation, the investigators will specifically examine the data for trajectory changes in individual patients that reflect a hyperinflammatory response.

Hypothesis Our primary hypothesis is that it is possible to develop a scoring system from existing NHS clinical and laboratory parameters to facilitate early identification of COVID-HI. Our secondary hypothesis, for examination in a subsequent study, is that this score will offer an important pre-screening tool for COVID19 cases for entry to clinical trials of anti-hyperinflammation therapies.

To address the primary hypothesis, the aims are:

1. Correlate the existing HScore with outcomes in our dataset.
2. Examine in granular detail those cases who meet or partially meet the HScore threshold for hyperinflammation.
3. Utilise non-hypothesis driven approaches to determine other clinical or laboratory data which would improve the sensitivity of the identification of hyperinflammation in COVID19 to identify a Soton HI score.

Inclusion Criteria:

• Sars-CoV-2 confirmed
• Hospitalized case

Exclusion Criteria:

• Sars-CoV-2 not confirmed

All individuals with possible COVID19 at University Hospitals NHS Foundation Trust will be included. As a further control group, a control cohort will be recruited, reflecting various key stratification aspects of the COVID19 patient cohort (e.g. by age and sex, or by hyperinflammation status) from hospital inpatients without a COVID19 diagnosis.

Data will be disclosed only upon request and approval of the proposed use of the data by the chief investigator of the INFLAME study, and in alignment with the relevant governance and ethical approval. Data will be de-identified for participant, key identifying factors and will be available with a signed data access agreement.

starting 6 months after publication