Increased arterial and venous thrombotic events have been reported in patients with COVID-19compared to those with other respiratory viruses. These thrombotic events are thought to be related to a thromboinflammatory state brought about by the virus. Observational data has suggested that starting patients on therapeutic or prophylactic anticoagulation on admission to the hospital may lower in-hospital mortality for patients with COVID-19. However, there is currently not enough data to know the optimal strategy in terms of type, dose, and duration of anticoagulation treatment.
This study aimed to determine whether therapeutic anticoagulation is effective in preventing complications in patients hospitalized with COVID-19 and elevated d-dimer concentrations. It was an open-label multicenter, randomized controlled trial in patients hospitalized in Brazil with COVID-19 diagnosis, symptoms for up to 14 days prior to randomization and elevated d-dimer concentrations. Patients were randomized via a 1:1 ratio in permuted blocks of variable size, stratified based on clinical condition (stable vs unstable) to either receive therapeutic anticoagulation for 30 days (with rivaroxaban if clinically stable, or with enoxaparin if clinically unstable) or prophylactic anticoagulation (with either enoxaparin or unfractionated heparin). Neither patients nor investigators were blinded to group allocation. The clinically unstable group received subcutaneous enoxaparin 1mg/kg twice per day or unfractionated heparin dosed to achieve a target anti-Xa concentration (0.3-0.7IU/mL). Once stabilized, these patients were transitioned to oral rivaroxaban dosed at 20mg daily. All patients in the therapeutic group continued treatment with rivaroxaban to day 30. The prophylactic group received standard venous thromboembolism prophylaxis dosing. If patients developed an indication for therapeutic anticoagulation, they were allowed to receive it. The prophylactic group of patients was only kept on anticoagulation while inpatient. Follow up was performed at 30 and 60 days.
The primary outcome studied was a hierarchical composite of time to death, duration of hospitalization or duration of supplemental oxygen use through 30 days. The primary safety outcome was major or clinically relevant non-major bleeding. Intention-to-treat analysis was used, and the primary outcome was reported using a win ratio method. Each patient in the treatment group was incrementally compared to each control patient to determine the “winners” or “ties” in relation to time to death, length of stay, and days of oxygen-free support in escalating challenges if “ties” occurred. The win ratio was then reported.
There were 310 patients in the therapeutic group's and 304 patients in the prophylactic group's primary analysis. Baseline characteristics were similar between the therapeutic and prophylactic groups. The total number of wins between the groups was not statistically different (win ratio 0.86, p=0.40, 95% CI 0.59-1.22) with 28,899 (34.8%) wins in the therapeutic group versus 34,288 (41.3%) wins in the prophylactic group. There were 19,837 (23.9%) ties. There was no statistical difference between the 8-point ordinal scale at day 30, disease progression measured on day 7, 15, and 30, or duration of invasive mechanical ventilation at the end of 30 days. There was also no statistical difference between individual thrombotic events or composite VTE, myocardial infarction, stroke, systemic embolism, or major adverse limb events. Differences remained insignificant when comparing clinically stable and unstable patients in each group. As to safety outcomes, there were 26 (8%) major or clinically relevant bleeding events in the therapeutic group compared to 7 (2%) in the prophylactic group (RR 3.64, 95% CI 1.61-8.27, p=0.001).
The authors concluded that therapeutic anticoagulation did not result in clinically better outcomes than prophylactic anticoagulation in the specific population of patients studied, and that therapeutic was in fact associated with a higher risk of major or clinically relevant bleeding than was prophylactic anticoagulation. They therefore recommended against using therapeutic anticoagulation in this population unless specifically indicated for other clinical reasons. The authors do note limitations such as performing follow-up interviews and pill counts over a phone call may have influenced their results or that the open label type study may have introduced bias although possibly mitigated by blind adjudication process.
Elizabeth M. Hanson, MD
Zachary B. Lewis, MD
University of Arkansas for Medical Sciences
Little Rock, AR
Comment: Thus far, there is a limited collection of data regarding the best management of hospitalized COVID-19 patients especially when considering thrombotic events. This study provides moderate quality evidence that therapeutic anticoagulation likely provides more harm than benefit and should not be used for patients with COVID-19 unless they are being treated for another indication. There is still much work to be done before we can definitively create high quality evidence practice guidelines on the role of anticoagulation in COVID-19 patients.