Risk of Clinically Relevant Venous Thromboembolism in Critically Ill Patients With COVID-19: A Systematic Review and Meta-Analysis

Gratz, Johannes; Wiegele, Marion; Maleczek, Mathias; Herkner, Harald; Schöchl, Herbert; Chwala, Eva; Knöbl, Paul; Schaden, Eva

Background: Early during the course of the ongoing COVID-19 pandemic, reports suggested alarmingly high incidences for thromboembolic events in critically ill patients with COVID-19. However, the clinical relevance of these events was not reported in several studies. Additionally, more recent research showed contradictory results and suggested substantially lower rates of venous thromboembolism. Thus, the aim of the present study was to summarize evidence on the incidence of clinically relevant venous thromboembolism (VTE)—defined as VTE excluding isolated subsegmental pulmonary embolism (PE) and distal deep vein thrombosis (DVT)—in adult critically ill patients with COVID-19. Methods: We performed a systematic review of studies reporting the incidence of clinically relevant PE and/or DVT in critically ill patients with COVID-19. Scientific reports published in the English language between January and October 2020 were included. We conducted a random-effects model meta-analysis to calculate incidence estimates of clinically relevant VTE and bleeding events. We also performed exploratory meta-regression and subgroup analyses of different diagnostic approaches and additional factors that possibly influenced the incidence of these outcomes. Results: Fifty-four articles (5,400 patients) fulfilled the predefined inclusion criteria, of which 41 had a high risk of bias. The majority of included patients were male, > 60 years, and overweight. Twenty-one studies reported the use of prophylactic doses of heparin. Pooled incidences for clinically relevant PE were estimated at 8% (95% CI, 4–11%), for proximal DVT at 14% (95% CI, 9–20%), and—after exclusion of studies with a high risk of bias—for the composite outcome of VTE at 18% (95% CI, 13–24%). Clinically relevant bleeding occurred at a rate of 6% (95% CI, 2–9%). Conclusions: We summarized currently available data on the rate of clinically relevant VTE in critically ill patients with COVID-19. Pooled incidence estimates were lower than those reported by previous review articles. In the absence of evidence-based anticoagulation guidelines for critically ill patients with COVID-19, the results of our study provide clinically important information for an individual risk-benefit assessment in this context. Registration: The study protocol was prospectively registered in PROSPERO on June 22, 2020 (CRD42020193353; https://www.crd.york.ac.uk/prospero).