Randomized Controlled Trial
Study population and setting
These were two international open-label, multiplatform, randomized controlled trials published concurrently comparing therapeutic dosing of heparin to usual standard thromboprophylaxis in patients hospitalized with Covid-19. The patients were determined to be critically ill based on the need for critical-care organ support (high-flow nasal cannula, mechanical ventilation, extracorporeal life support, vasopressors, or inotropes) at time of enrollment, or noncritically ill otherwise. The primary outcome was number of days free of organ support in the first 21 days, with patients who did not survive to discharge being automatically scored -1.
Summary of Main Findings
Enrollment began in April 2020 for both studies. For critically ill patients, the study enrolled 1207 patients through December 2020, when an interim analysis showed futility. Patients who received therapeutic dose heparin had a median of 1 organ support-free day (IQR -1 to 16) versus 4 (IQR -1 to 16) in the usual standard of care group (OR = 0.93; 95%CI 0.67 – 1.03; posterior probability of futility = 99.9%; posterior probability of inferiority = 95%). 62.7% of patients in the heparin group survived to hospital discharge, versus 64.5% in the usual standard of care group (OR = 0.85; 95%CI = 0.64 – 1.11; posterior probability of inferiority = 89.2%).
For noncritically ill patients, the study enrolled 2244 patients through January 2021, when an interim analysis showed superiority. The median number of organ support-free days was 22 in both groups; 80.2% of noncritically ill patients who received therapeutic dose heparin survived until hospital discharge without needing organ support, versus 76.4% in the standard of care group (absolute difference 4.0%, 95%CI = 0.5 – 7.2). The odds ratio of organ-support free days was 1.27 (95%CI 1.03 – 1.58; posterior probability of superiority 98.6%) and was similarly high after stratification by D-dimer level. Major bleeding occurred in 1.9% of patients in the treatment group, versus 0.9% in the standard of care group, with 3 cases of fatal bleeding in the treatment group, versus 1 in the standard of care group.
As the studies were paired, the results between the two can be readily controlled and compared. The sample size of both studies was large, allowing for stratified analyses. By using Bayesian analyses, the authors were able to set and use cutoffs to determine when the study was sufficiently powered. The multiplatform nature of the studies also suggests that these findings are generally reproducible in broad clinical contexts.
This was a pair of open-label studies, so participants were not blinded to treatment group. This introduced the risk of ascertainment bias, in which a patient on a higher dose of heparin may have been more aggressively monitored for bleeds. The study excluded patients with an increased risk of bleeding, which was important for study balance but excluded a nontrivial amount of people at risk for moderate to severe COVID-19. Standard-of-care thromboprophylaxis varied between sites; most critically ill participants were enrolled in the UK, where standard treatment guidelines included intermediate-level heparin treatment for all COVID-19 patients, which may have impacted between-group comparisons and study generalizability. Early trial conclusion prevented researchers from identifying long-term secondary benefits of treatment.
These paired studies provide strong evidence that treatment with therapeutic dose heparin (as if the patient had a known venous thromboembolism) is beneficial for reducing in-hospital mortality and the need for organ support among patients with noncritical COVID-19. However, this approach provides no benefit to patients with critical illness (patients needing ICU-level care at time of enrollment), and is likely inferior to standard thromboprophylaxis treatment.
This review was posted on: 1 September 2021