Coagulopathy is an impairment in the blood’s ability to form clots. Coagulopathy associated with COVID-19 infection has emerged as a significant contributor to morbidity among those infected with the illness. Although the incidence of coagulopathy is unknown, COVID-19–related coagulopathy has been described as having a unique hemostatic signature [1-3]. Typically reported abnormalities include a mildly prolonged prothrombin time/international normalized ratio (PT/INR) as well as an activated partial thromboplastin time (aPTT), mild thrombocytopenia, and elevated D-dimer and fibrinogen [3,4]. Generally, COVID-19–associated coagulopathy results in a prothrombotic state and is nonconsumptive or progresses to hemorrhage [5]. As such, observational studies and autopsy reports have focused on characterizing thrombotic complications, including venous thromboembolism (VTE).

Potential approaches to treatment are continuing to develop with increasing experience with COVID-19 infection. Institutions have described updated VTE prophylaxis protocols aimed to more aggressively prevent clots [1,2,6]. For example, some describe utilization of standard dose VTE prophylaxis with unfractionated heparin (eg, unfractionated heparin 5000 units by subcutaneous injection twice daily or three times daily) or low molecular weight heparin (eg, enoxaparin 40 mg by subcutaneous injection every 24 hours) for all patients hospitalized with COVID-19 infection rather than based on risk assessment. Others deploy anticoagulation regimens that are higher intensity compared to standard dose VTE prophylaxis. For example, high dose, low molecular weight heparin (eg, enoxaparin 40 mg by subcutaneous injection every 12 hours) or empiric treatment anticoagulation (eg, unfractionated heparin by intravenous infusion; enoxaparin 1 mg/kg by subcutaneous injection every 12 hours) for high-risk patients, which is variably defined (eg, intensive care unit [ICU] level of care, clinical deterioration, rising d-dimer) and based on historical data in other high-risk populations (eg, bariatric surgery, third trimester pregnancy) [1,2,6]. Also, utilization of specific anticoagulation products will vary by country depending on regulatory approval and availability. In fact, early experience indicates that high-risk patients (eg, sepsis-induced coagulopathy score ≥4 or D-dimer >3 times the upper limit of normal [ULN]) are most likely to benefit from VTE prophylaxis [7]. However, evidence-based treatment protocols are needed to further improve in COVID-19 patients with VTE [1,2].

The mechanism of COVID-19–associated coagulopathy is not fully understood. However, the intense and sustained cytokine-mediated inflammatory response to SARS-CoV-2 infection is likely etiologic. As an example, elevated fibrinogen levels have been found to be associated with elevated interleukin-6 (IL-6) levels, while D-dimer also rises in parallel with C-reactive protein (CRP) [8]. Furthermore, inflammatory markers directly activate the clotting system as does tissue hypoxia, both on top of direct endothelial cell injury and subsequent dysfunction by SARS-CoV-2 cellular entry [7,9].

Serologic markers may be associated with severity of infection and may also be predictors of increased morbidity and coagulopathy. For example, D-dimer is a biomarker that increases as a result of thrombosis (eg, microvascular thrombosis, deep vein thrombosis [DVT], or pulmonary embolism [PE]) or systemic activation of hemostasis (eg, disseminated intravascular coagulation) and is associated with severe COVID-19 and mortality [10]. Tang et al [4] demonstrated that D-dimer elevation on admission and rising to at least 3-4 times ULN over the course of the hospital stay was associated with increased mortality. Other hemostatic markers such as PT and aPTT prolongation, elevated fibrinogen degradation product, and low platelets have also been associated with severe COVID-19 and mortality [4,11]. Governing bodies have recommended that D-dimer, PT, platelets, and fibrinogen be measured in patients with COVID-19 infection for risk stratification and prognosis [1,2]. 

We performed a systematic review of VTE in the setting of patients hospitalized with COVID-19 infection and summarized the potential treatment effects in VTE management in these patients. Our aim was to estimate the observed incidence of hospitalized VTE patients and analyze patient characteristics in the VTE cohort. 

We performed a systematic literature search in PubMed with the title search terms “COVID-19” or “SARS-CoV-2” or “Novel Coronavirus 2019” and “venous thromboembolism” or “deep vein thrombosis” or “pulmonary embolism” or “thrombosis” or “thromboembolic” or “anticoagulation” or “heparin” or “thromboprophylaxis” to identify primary research studies that report the rate of VTE in patients hospitalized with COVID-19 infection who are treated with standard dose pharmacologic VTE prophylaxis, high dose pharmacologic VTE prophylaxis, treatment dose anticoagulation, no anticoagulation, or no documentation. A supplementary search was performed on Google Scholar using the same search terms and journal article references were reviewed to identify additional studies. Studies of adult populations that were published in a PubMed peer-reviewed journal from June 22, 2019, to June 22, 2020 were included for review. Data were collected for each included study design, population studied, VTE event rate, VTE diagnostic strategy, VTE prophylaxis or treatment strategy, hemostatic lab abnormalities, and clinical outcomes including ICU level of care and survival. 

We excluded studies with arterial thrombosis, myocardial infarction or ischemic stroke, pediatric and fetal populations, and reviews, case reports, letters to the editor, or any study that had not yet undergone peer review. Clinical outcomes data for the included studies were pooled, and we conducted a systematic review and meta-analysis with a random effects model to measure a single group summary for VTE incidence as our primary outcome [12]. Confidence intervals were determined by the adjusted Wald method. Secondary outcomes included a single group summary for mortality with the same methodology as the primary outcome and patient demographics with clinical characteristics using descriptive statistics with weighted mean and weighted standard deviation. Assessment of variables associated with VTE incidence was conducted using univariate linear regression and multivariate linear regression. Assessment of binary variables associated with VTE occurrence was conducted using multiple logistic regression. Estimation of differences in continuous variables between patients with VTE and patients without VTE was conducted using the Z test (two samples for weighted means with weighted variance). Data were compiled using Google Sheets (Google LLC) and Microsoft Excel (Microsoft Corp).

The Cedars-Sinai Hospital Institutional Review Board requirement for approval was waived as this is a systematic literature review.

The initial PubMed literature review returned 212 journal articles, of which 12 studies were included in our review. The supplementary Google Scholar and journal article references search identified an additional 2 studies. In total, 14 studies were included in our review [6,9,13-24] (Figure 1).

Studies included were observational (Table 1) and predominantly based on experience at a single center (single center: n=10; multicenter: n=4). The total patient sample size was 1677 (range 26-388) and represented a multinational patient population (China: n=272; France: n=206; Italy: n=415; Netherlands: n=382; Spain: n=156; United States: n=44). The weighted median study duration was 37.2 (SD 17.4) days and 3 studies reported a median length of stay (LOS) (weighted mean LOS 9.5 [SD 1.8 ]days) with patients receiving both ICU and non-ICU levels of care (Table 2). Five studies (n=352) did not report the patient status (eg, discharged alive, expired, or admitted) at completion of the study period. For the other 9 studies, the designations and counts were as follows: nonsurvivors (n=244), discharged alive (n=717), admitted (n=369), and unknown (n=20).

In total, 13 studies reported utilization of VTE chemoprophylaxis or treatment anticoagulation. Most patients (n=1306, 82.4%) received anticoagulation; 17.6% (n=279) did not. Of the patients who received anticoagulation, standard dose VTE prophylaxis was most common (n=691, 52.9%). Patients were also prescribed high dose VTE prophylaxis (n=84, 6.4%) or treatment anticoagulation (n=197, 15.1%). In 25.6% (n=334) of patients prescribed anticoagulation, the dosage or intensity was not specified. VTE diagnosis was determined by systematic screening in 7 studies, 1 of which also implemented systematic screening for PE. For 7 other studies, VTE was diagnosed by usual practice. Three studies exclusively screened for DVT and did not report PE.

The combined estimate of VTE incidence was 26.9% (SE 3.1; 95% CI 20.8-33.1) (Figure 2). Occurrence of VTE (n=377) was more often attributed to DVT (n=262) and less often to PE (n=116). The combined estimate of mortality incidence was 24.4% (SE 7.1; 95% CI 10.5-38.2). Absolute values were 244 for nonsurvivors, 717 for discharged alive, 369 for admitted, and 20 for unknown.

Systematic screening for VTE (r²=.34, P=.03) and study duration (r²=–.33, P=.03) were both correlated with VTE incidence. There were no associations with VTE and mortality, percentage of patients prescribed anticoagulation, gender, age, or D-dimer level. Multivariate linear regression for the intensity of VTE prophylaxis and VTE incidence was not significant (r²=.64; F=.25) nor was a model that included the percentage of patients prescribed VTE prophylaxis or anticoagulation, the percentage of patients in the ICU, gender, age, D-dimer level, study duration, and implementation of systematic screening for VTE (r²=.67; F=.58).

Five studies compared clinical characteristics and outcomes for patients with VTE (n=157) to patients without VTE (n=296). D-dimer was significantly increased in patients with VTE compared to patients without VTE (5.62 [SD 0.9] vs 1.43 [SD 0.6]; P≤.001). VTE was decreased in patients receiving anticoagulation (either VTE prophylaxis or treatment anticoagulation) (OR .58, 95% CI .36-.92; P=.02) and was increased in patients receiving an ICU level of care during their admission (OR 6.38, 95% CI 3.67-1.11; P≤.001). There was no difference in VTE rates for nonsurvivors compared to survivors (OR 2.02, 95% CI .98-4.19; P=.06) (Figure 3).