Risk factors for and long-term outcomes in people with covid-19 admitted to hospital in England

Date of ISAC Approval: 
Lay Summary: 
Covid-19 has rapidly spread across the world with many more cases and fatalities than in previous coronavirus outbreaks. Although most people with covid-19 experience a mild/moderate self-limiting disease, it has been estimated that some 20% of covid-19 infections detected are admitted to hospital and that 1 in 4 of those hospitalised will have a severe illness that needs intensive care and/or results in death. Current evidence indicates that the risk of severe illness is higher in men than women, at older ages, and in people with at least one pre-existing health condition (co-morbidity). We would like to understand better the extent to which different factors increase the likelihood of covid-19 infection that leads to hospital admission, admission to an intensive care unit, or death. Among those who are admitted to hospital and survive, we would like to better understand the long term risk of developing new lung or heart and other disease, and of deterioration in pre-existing conditions. This information will help to guide preventive strategies (e.g. modification of factors that increase vulnerability or shielding those who are most vulnerable), and also long-term clinical care for patients who have recovered from covid-19 infection.
Technical Summary: 
Evidence from around the globe indicates that most people infected by covid-19 will have only mild illness, but in approximately 20% of cases, the disease is more severe, requiring hospital admission. Among those admitted to hospital, ~1 in 4 require intensive care/and or die. Others may die from covid-19 without being admitted to hospital. To inform preventive strategies, there is a need to understand better what determines vulnerability to more severe illness, and to what extent. It is currently not clear if there are avoidable risk factors for admission, severe disease and/or death. For example, early papers suggest a low representation of smokers among those admitted, which is unexpected, but greater risk of severe disease. This information might allow modification of some factors that increase vulnerability, and help to work out those who could benefit most from shielding. In addition, it is important to find out whether people who recover from such illness incur an increased long-term risk of cardio-respiratory morbidity or of other types of illness. To explore risk factors for hospital admission for covid-19, we will undertake a case control study using CPRD Gold and Aurum primary care data linked with HES inpatient data and ONS mortality data. Cases will be those individuals admitted to hospital with covid-19 and/or dying from the disease, with 5 controls per case matched on GP practice, sex and age. Using conditional logistic regression, we will investigate potential risk factors such as smoking, pre-existing conditions and medication. To explore long-term health risks following hospital admission for covid-19, we will then follow up the cases and controls as a cohort, and use conditional Poisson regression to determine whether or not the cases have higher subsequent risk of deterioration in pre-existing disease or development of new cardiovascular, respiratory or other morbidity (e.g. heart failure, obstructive lung disease, mental health problems). The first analysis of outcomes will be at least 12 months after entry to follow-up.
Health Outcomes to be Measured: 
In the case-control study, the main outcome will be admission to hospital for, or death from, covid-19 infection. Within this, a sub-set of more severe disease will be defined by admission to an intensive care unit and/or death from the disease. For the cohort study, the exact specification of outcomes for analysis will depend on their frequency in the cohort as a whole (covid-19 patients and controls) at the time, and on the scientific evidence that has by then emerged regarding possible long-term effects of infection (which will be monitored continually over the follow-up period). To the extent that numbers are sufficient for meaningful statistical analysis, they will include measures of: - Cumulative incidence of recurrent/new covid-19 infection (confirmed by serology or probable from clinical presentation) classified according to severity (e.g. leading to hospital admission, admission to intensive care or death) - New or exacerbated cardio-respiratory disease - New or exacerbated renal disease - New or exacerbated auto-immune disease - Cumulative frequency of respiratory infections other than serologically confirmed/probable Covid-19 infection - Cumulative frequency of non-respiratory infections - New or exacerbated mental illness

Jennifer Quint - Chief Investigator - Imperial College London
David Coggon - Collaborator - University of Southampton
Ian Douglas - Collaborator - London School of Hygiene & Tropical Medicine (LSHTM)
Jennifer Quint - Corresponding Applicant - Imperial College London
John Britton - Collaborator - University of Nottingham
Liam Smeeth - Collaborator - London School of Hygiene & Tropical Medicine (LSHTM)
Nicholas Hopkinson - Collaborator - Imperial College London
Paul Cullinan - Collaborator - Imperial College London

HES Admitted;ONS;Patient IMD