Effect of therapeutic versus prophylactic anticoagulation therapy on clinical outcomes in COVID-19 patients: a systematic review with an updated meta-analysis

Background: Previous studies demonstrate a reduced risk of thrombosis and mortality with anticoagulant treatment in patients with COVID-19 than in those without anticoagulation treatment. However, an open question regarding the efficacy and safety of therapeutic anticoagulation (T-AC) versus a lower dose, prophylaxis anticoagulation (P-AC) in COVID-19 patients is still controversial. Methods: We systematically reviewed currently available randomized clinical trials (RCTs) and observational studies (OBs) from January 8, 2019, to January 8, 2022, and compared prophylactic and therapeutic anticoagulant treatment in COVID-19 patients. The primary outcomes were risk of mortality, major bleeding, and the secondary outcomes included venous and arterial thromboembolism. Subgroup analysis was also performed between critically ill and noncritically ill patients with COVID-19 and between patients with higher and lower levels of D-dimer. Sensitivity analysis was performed to decrease the bias and the impact of population heterogeneity. Results: We identified 11 RCTs and 17 OBs fulfilling our inclusion criteria. In the RCTs analyses, there was no statistically significant difference in the relative risk of mortality between COVID-19 patients with T-AC treatment and those treated with P-AC (RR 0.95, 95% CI, 0.78-1.15, P = 0.60). Similar results were also found in the OBs analyses (RR 1.21, 95% CI, 0.98-1.49, P = 0.08). The pooling meta-analysis using a random-effects model combined with effect sizes showed that in the RCTs and OBs analyses, patients with COVID-19 who received T-AC treatment had a significantly higher relative risk of the major bleeding event than those with P-AC treatment in COVID-19 patients (RCTs: RR 1.76, 95% CI, 1.19-2.62, P = 0.005; OBs: RR 2.39, 95% CI, 1.56-3.68, P < 0.0001). Compared with P-AC treatment in COVID-19 patients, patients with T-AC treatment significantly reduced the incidence of venous thromboembolism (RR 0.51, 95% CI, 0.39-0.67, P < 0.00001), but it is not associated with arterial thrombosis events (RR 0.97, 95% CI, 0.66-1.42, P = 0.87).