May 18th, 2020, by Labtoo's team
A study establishing correlations between the mortality rate of COVID-19 patients and their potential risk factors has been revealed on the 7th of May, though it has not been peer-reviewed yet. Conducted in England and based on electronic medical records from the NHS, Labtoo explores its preliminary results.
Background of the study
R&D research on COVID-19 intensifies, revealing more and more information on the virus and its consequences. An English research team, named the "OpenSafely Collaborative" and led by Elizabeth Williamson, has shed light on risk factors that are most likely to lead to the death of patients infected by COVID-19.
To do so, data from the medical records of more than 17.5 million patients has been analyzed thanks to the OpenSafely digital platform, on a period extending from the 1st of February 2020 to the 25th of April 2020. These patients were all hospitalized in the UK at some point during the past years, though most of them are not affected by coronavirus.
The primary outcome of this brand-new study is the death in hospital among people with confirmed COVID-19. The used method is a cohort study analyzed by Cox-regression, which generates hazard ratios adjusted on co-variates selected by clinical interest (such as age, sex, ethnicity, etc.).
What are the results?
Among the 17.5 million patients included in this study, 5683 of them died due to the coronavirus. The risk factors detected during the analysis are the following:
Age: patients above 80 years have more than 12-fold increased risk to die compared with those aged 50-59 years. Overall, risk grows exponentially as the age rises. Age constitutes the biggest risk factor regarding the mortality rate.
Sexe: men are at double risk than women.
Ethnicity: all non-white ethnic groups have higher risk than those with white ethnicity, and especially Black and Asian people.
Precarity: it is the only risk factor whose prevalence has risen during the study. People from deprived areas are at more risk than the others.
Smoking: many adjustments have been made to determine if smoking had an influence on the mortality rate due to coronavirus. Even though results slightly change depending on the added covariates, a positive correlation between former smokers and the risk of death remains. However, current smokers have not been detected to be at more risk than non-smokers after a multivariate adjustment; they even seem to be protected from COVID-19 by smoking. This is not the case anymore if the adjustment is made on sex and age only, or on ethnicity only, which might be one of the methodological limits of this study.
In addition to these potential risk factors, the study has also analyzed the comorbidities associated to patients. The main outcome is that you are more likely to die from COVID-19 if you suffer from the following conditions: diabetes, obesity, asthma, chronic heart disease, liver disease, stroke or dementia, neurology disease, kidney disease, organ transplant, immunosuppressive conditions.
People suffering from cancer are at more risk than the average, especially if diagnosis has been done in the past year. No association between hypertension and the main outcome has been found.
The conclusions we can draw from this study
The results of this unprecedented national study confirm most of the previous findings of other smaller studies: the risk of infection and death by COVID-19 is determined by many vulnerability factors, such as age, which might be the most important factor here after seeing the estimated risks, socio-economic factors such as ethnicity and deprivation, whose underlying causes have yet to be clarified, and also comorbidities as seen above.
While interesting results are featured in this non-reviewed study, it also presents some limits:
Censored clinical data: due to a high number of included patients, medical records do not all have the same information. For example, only 26% of the records mentioned the ethnicity of patients.
The explanation about the impact of ethnicity on the number of deaths is not clear: data analyses do not allow us to distinguish if ethnicity is a biological or socio-economic factor. There has not been any proven linkage between ethnicity and precarity.
The absence of a systematic screening among the cohort: this might lead to the exclusion of results belonging to untested patients whose death was linked to COVID-19.
The total number of cases (ongoing, cured, deceased) is not mentioned, which may induce a bias regarding the analysis of the death risk factors.
All sensitivity analyses seem to have proven the data robustness. However, some doubts still remain: for example, the small protective effects from smoking has still yet to be prove. While some studies suggest that hypertension is linked to a higher risk of death, this study have not found solid clues about this effect.
One of the main strengths of this study is its high number of analyzed criteria. However, this might also constitute a limit as it would underestimate some risk factors, due to the multivariate adjustment that can overwhelm their true impact. This is especially true for risk factors that affect many others, such as smoking or age.
In any case, it is obvious that many patients’ groups are particularly more vulnerable to COVID-19 than others, which we have to protect: scientific literature agrees on this point. The pandemic is still in progress all around the world, though at a slower pace than before; however, R&D research still has to continue its work to find a cure to this virus, and that is why Labtoo continues to mobilize its resources to do so.
Whether you need R&D service providers to work with you on coronavirus-related projects, or you want to bring your biotech expertise to researchers, Labtoo is here to support you. Our COVID-19 initiative aims to foster research by referencing biotech service providers that can work on COVID-19 projects. For more details about this, visit our COVID-19 special page right here.