BACKGROUND: In patients hospitalised with COVID-19, an increased incidence of thromboembolic events, such as pulmonary embolism, deep vein thrombosis and stroke, has been reported . It is unknown whether anticoagulation can prevent these complications and improve outcome .
METHODS: A systematic literature search was performed to answer the question: What is the effect of (prophylactic and therapeutic dose) anticoagulation therapy in COVID-19 patients on cardiovascular and thromboembolic complications and clinical outcome? Relevant outcome measures were mortality (crucial), hospital admission, length of stay, thromboembolic complications (pulmonary embolism, stroke, transient ischaemic attack), need for mechanical ventilation, acute kidney injury and use of renal replacement therapy . Medline and Embase databases were searched with relevant search terms until 17 July 2020 . After systematic analysis, eight studies were included . Analysis was stratified for the start of anticoagulation before or during hospital admission .
RESULTS: There was insufficient evidence that therapeutic anticoagulation could improve the outcome in patients hospitalised with COVID-19 . None of the studies demonstrated improved mortality, except for one very small Italian study . Furthermore, none of the studies showed a positive effect of anticoagulation on other outcome measures in COVID-19, such as ICU admission, length of hospital stay, thromboembolic complications, need for mechanical ventilation, acute kidney failure or need for renal replacement therapy, except for two studies demonstrating an association between anticoagulation and a lower incidence of pulmonary embolism . However, the level of evidence of all studies varied from 'low' to 'very low', according to the GRADE methodology .
CONCLUSION: Analysis of the literature showed that there was insufficient evidence to answer our objective on the effect of anticoagulation on outcome in COVID-19 patients, especially due to the low scientific quality of the described studies . Randomised controlled studies are needed to answer this question.