High rates of pulmonary artery occlusions in COVID-19. A meta-analysis

Background: COVID-19 patients are considered at high risk of venous thromboembolism (VTE). The real nature of pulmonary artery occlusions (PAO) in COVID-19 has been questioned, suggesting that it is caused also by in situ thrombi, rather than only by emboli (PE) from peripheral thrombi. Methods: We searched MEDLINE for studies published until 6 June 2020 that included COVID-19 patients or non-COVID-19 medical patients at VTE risk, treated with heparins, in whom VTE (PE and deep vein thrombosis, DVT) had been reported. Systematic review and results reporting were conducted in accordance with PRISMA guidelines. Data were independently extracted by two observers, and estimates were pooled using random-effects meta-analysis. Results: We identified 17 studies including 3224 COVID-19 patients and 7 including 11 985 non-COVID-19 patients. Two analyses were performed: in all COVID-19 patients and only in those (n = 515) who, like non-COVID-19 patients, were screened systematically for DVT. The latter analysis revealed that the prevalence of DVT was 15.43% (95%CI, 4.08-31.77) in COVID-19 and 4.21% (2.27-6.68) in non-COVID-19 patients (P =.0482). The prevalence of PE was 4.85% (40.33-13.01) in COVID-19 patients and 0.22% (0.03-0.55) in non-COVID-19 patients (P =.0128). The percentage of PE among VTE events was 22.15% (5.31-44.60) in COVID-19 and 6.39% (3.17-10.41) in non-COVID-19 patients (P =.0482). Differences were even more marked when all COVID-19 patients were analysed. Conclusions: The results of our meta-analysis highlight a disproportion in the prevalence of PE among all VTE events in COVID 19 patients, likely reflecting PAO by pulmonary thrombi, rather than emboli from peripheral vein thrombi.