Natural history of coagulopathy and use of anti-thrombotic agents in COVID-19 patients: a cohort study

Date of Approval: 
2021-05-07 00:00:00
Lay Summary: 
People with coronavirus disease-2019 (COVID-19) may be at a high risk of thrombotic disease, where blood clots block veins or arteries. There is a need to better understand how common these thromboembolic events among patients with COVID-19 are, their implications for health outcomes, and whether individuals with a particularly high risk for them can be identified. We plan to perform such research using primary care data linked to hospital records from the UK. This will be part of a network study where the same analyses will be performed using similar data from other countries (facilitated by using a common data model, with no patient-level data needing to be shared between sites). First we will assess how common thromboembolic events are among patients with COVID-19, second we will consider the impact of these events on the likelihood of patients having worse outcomes, third we will describe the impact of patient characteristics on their risks of having a thromboembolic event, and fourth we will develop algorithms to identify those people most at risk of having one of these events. The findings from our study will inform the management of COVID-19 patients.
Technical Summary: 
We will investigate the risks of thromboembolic and rare coagulopathy events among patients with COVID-19, the impact of thromboembolic events on prognosis in COVID-19, the association between various risk factors and rates of thromboembolic events, and to develop and externally validate prediction models for thromboembolic events for patients with COVID-19. The study will be a distributed network study using datasets mapped to the Observational Medical Outcomes Partnership (OMOP) common data model. Primary care data, linked to hospital records, from the UK will be one of the contributing databases. Linked data will be used to maximize completeness for hospital study outcomes. By using a common data model, no data will need to be shared between sites. Four study cohorts will be defined: 1) Persons tested positive for SARS-CoV-2,2) Persons tested positive for SARS-CoV-2 or with a clinical diagnosis of COVID-19, 3) Persons hospitalised with COVID-19, and 4) Persons requiring intensive services during a hospitalisation with COVID-19. The occurrence of venous thromboembolic and arterial thromboembolic events will be identified in the 30-, 60- and 90-days post-index date. COVID-19 worsening will be defined as increasing care intensity (e.g. from outpatient to inpatient, from inpatient to receiving intensive care services) and/or mortality. We will summarise the incidence of venous and arterial thromboembolic and coagulopathy events at 30, 60, and 90 days following the relevant index date for study cohorts as a whole and for various strata of interest. We will use a multi-state model to summarise risks of worsening stratified by those with and without thromboembolic events of interest. The impact of risk factors on risks of venous and arterial thromboembolic events will be assessed using Cox models. Patient-level prediction models for thromboembolic events will be developed using a data-driven approach, with external validation performed across the network of contributing databases.
Health Outcomes to be Measured: 
The occurrence of venous thromboembolic and arterial thromboembolic events will be identified in the 30-, 60- and 90-days post-index date. COVID-19 worsening will be defined as increasing care intensity (e.g. from outpatient to inpatient, from inpatient to receiving intensive care services) and/or mortality.
Application Number: 
21_000391
Collaborators: 

Daniel Prieto-Alhambra - Chief Investigator - University of Oxford

Edward Burn - Corresponding Applicant - University of Oxford

Antonella Delmestri - Collaborator - University of Oxford

Nathan Jones - Collaborator - University of Oxford

Linkages: 
HES APC, ONS, Patient IMD