Prophylactic anticoagulants for non-hospitalised people with COVID-19

The coronavirus disease 2019 (COVID-19) pandemic has impacted healthcare systems worldwide. Multiple reports on thromboembolic complications related to COVID-19 have been published, and researchers have described that people with COVID-19 are at high risk for developing venous thromboembolism (VTE). Anticoagulants have been used as pharmacological interventions to prevent arterial and venous thrombosis, and their use in the outpatient setting could potentially reduce the prevalence of vascular thrombosis and associated mortality in people with COVID-19. However, even lower doses used for a prophylactic purpose may result in adverse events such as bleeding. It is important to consider the evidence for anticoagulant use in non-hospitalised people with COVID-19.

To evaluate the benefits and harms of prophylactic anticoagulants versus active comparators, placebo or no intervention, or non-pharmacological interventions in non-hospitalised people with COVID-19.

We used standard, extensive Cochrane search methods. The latest search date was 18 April 2022.

We included randomised controlled trials (RCTs) comparing prophylactic anticoagulants with placebo or no treatment, another active comparator, or non-pharmacological interventions in non-hospitalised people with COVID-19. We included studies that compared anticoagulants with a different dose of the same anticoagulant. We excluded studies with a duration of under two weeks.

We used standard Cochrane methodological procedures. Our primary outcomes were all-cause mortality, VTE (deep vein thrombosis (DVT) or pulmonary embolism (PE)), and major bleeding. Our secondary outcomes were DVT, PE, need for hospitalisation, minor bleeding, adverse events, and quality of life. We used GRADE to assess the certainty of the evidence.

We included five RCTs with up to 90 days of follow-up (short term). Data were available for meta-analysis from 1777 participants. Anticoagulant compared to placebo or no treatment Five studies compared anticoagulants with placebo or no treatment and provided data for three of our outcomes of interest (all-cause mortality, major bleeding, and adverse events). The evidence suggests that prophylactic anticoagulants may lead to little or no difference in all-cause mortality (risk ratio (RR) 0.36, 95% confidence interval (CI) 0.04 to 3.61; 5 studies; 1777 participants; low-certainty evidence) and probably reduce VTE from 3% in the placebo group to 1% in the anticoagulant group (RR 0.36, 95% CI 0.16 to 0.85; 4 studies; 1259 participants; number needed to treat for an additional beneficial outcome (NNTB) = 50; moderate-certainty evidence). There may be little to no difference in major bleeding (RR 0.36, 95% CI 0.01 to 8.78; 5 studies; 1777 participants; low-certainty evidence). Anticoagulants probably result in little or no difference in DVT (RR 1.02, 95% CI 0.30 to 3.46; 3 studies; 1009 participants; moderate-certainty evidence), but probably reduce the risk of PE from 2.7% in the placebo group to 0.7% in the anticoagulant group (RR 0.25, 95% CI 0.08 to 0.79; 3 studies; 1009 participants; NNTB 50; moderate-certainty evidence). Anticoagulants probably lead to little or no difference in reducing hospitalisation (RR 1.01, 95% CI 0.59 to 1.75; 4 studies; 1459 participants; moderate-certainty evidence) and may lead to little or no difference in adverse events (minor bleeding, RR 2.46, 95% CI 0.90 to 6.72; 5 studies, 1777 participants; low-certainty evidence). Anticoagulant compared to a different dose of the same anticoagulant One study compared anticoagulant (higher-dose apixaban) with a different (standard) dose of the same anticoagulant and reported five relevant outcomes. No cases of all-cause mortality, VTE, or major bleeding occurred in either group during the 45-day follow-up (moderate-certainty evidence). Higher-dose apixaban compared to standard-dose apixaban may lead to little or no difference in reducing the need for hospitalisation (RR 1.89, 95% CI 0.17 to 20.58; 1 study; 278 participants; low-certainty evidence) or in the number of adverse events (minor bleeding, RR 0.47, 95% CI 0.09 to 2.54; 1 study; 278 participants; low-certainty evidence). Anticoagulant compared to antiplatelet agent One study compared anticoagulant (apixaban) with antiplatelet agent (aspirin) and reported five relevant outcomes. No cases of all-cause mortality or major bleeding occurred during the 45-day follow-up (moderate-certainty evidence). Apixaban may lead to little or no difference in VTE (RR 0.36, 95% CI 0.01 to 8.65; 1 study; 279 participants; low-certainty evidence), need for hospitalisation (RR 3.20, 95% CI 0.13 to 77.85; 1 study; 279 participants; low-certainty evidence), or adverse events (minor bleeding, RR 2.13, 95% CI 0.40 to 11.46; 1 study; 279 participants; low-certainty evidence). No included studies reported on quality of life or investigated anticoagulants compared to a different anticoagulant, or anticoagulants compared to non-pharmacological interventions.

We found low- to moderate-certainty evidence from five RCTs that prophylactic anticoagulants result in little or no difference in major bleeding, DVT, need for hospitalisation, or adverse events when compared with placebo or no treatment in non-hospitalised people with COVID-19. Low-certainty evidence indicates that prophylactic anticoagulants may result in little or no difference in all-cause mortality when compared with placebo or no treatment, but moderate-certainty evidence indicates that prophylactic anticoagulants probably reduce the incidence of VTE and PE. Low-certainty evidence suggests that comparing different doses of the same prophylactic anticoagulant may result in little or no difference in need for hospitalisation or adverse events. Prophylactic anticoagulants may result in little or no difference in risk of VTE, hospitalisation, or adverse events when compared with antiplatelet agents (low-certainty evidence). Given that there were only short-term data from one study, these results should be interpreted with caution. Additional trials of sufficient duration are needed to clearly determine any effect on clinical outcomes.