Study selection and characteristics
Figure 1 shows the study selection process. We initially retrieved 1248 articles using the search strategy, leaving 966 articles after removing duplication. Screening by titles and abstracts, 34 articles were selected full-text assessment. 22 articles were excluded because the population of patients infected with SARS-CoV-2 in these articles was less than 10. Meeting the inclusion criterias, 12 articles and 1083 patients were included for quantitative meta-analysis.(15-20, 24, 26-30) The main characteristics of the included studies were shown in Table 1. These studies enrolled patients from France, China, England, and other countries, with small sample sizes ranging from 34 to 198. Among these included studies, eight were retrospective study, three were prospective study, and left one was case series and case-control study.
There was a certain degree of heterogeneity among these studies. Six studies detailed patients who needed mechanical ventilation or were in ICU or seriously ill, one studies reported patients from the general ward, and the other five studies included patients both in the general ward and intensive care units. Almost all patients were treated with thromboembolism prophylaxis using low molecular weight heparin or nadroparin. These studies mainly focus on venous thromboembolism, especially deep venous thrombosis, while only one article reported both venous thrombosis and arterial thrombosis.(19)
Quality Assessment
The total score of quality assessment of included studies was shown in Table 1 according to the STROBE guideline.(25) The overall score of the included studies ranged from 19 to 22, which indicated the high quality of these studies.
Prevalence of thrombosis in patients with COVID-19
As mentioned above, the conditions of patients were not completely consistent in different studies, and admission to the ICU was related to an increased risk of thrombosis. Therefore, we calculated the prevalence of thrombosis twice (Figure 2). The prevalence is 22% (95% CI 0.08-0.40) in patients infected with SARS-CoV-2, and increased to 43% (95% CI 0.29-0.65) after admission to ICU, both by the random-effects model.
Risk factors associated with thrombosis for patients with COVID-19
Meta-analysis was carried out on several factors related to thrombosis, and the results are shown in Figure 3. Compared with non-thrombotic patients, thrombotic patients had higher level of D-dimer (MD=2.79, 95% CI 2.27–3.31), LDH (MD=112.71, 95% CI 62.40–163.02), and WBC (MD=1.14, 95% CI 0.47–1.81) while decreased lymphocytes (MD= -0.20, 95% CI -0.38 – -0.02) (Fig 3A-3D). Age (MD=1.91, 95% CI -1.58–5.40), platelets count (MD=13.06, 95% CI -1.62–27.73), and male (RR=1.10,95% CI 0.92–1.33) tended to be risk factors of thrombotic complications for COVID-19 patients though P>0.05 (Fig 3E-3G). Increased thrombotic complications were not observed in patients with obesity (body mass index (BMI) MD=0.24, 95% CI -0.14–1.51), hypertension (OR=0.98, 95% CI 0.66–1.46), coronary artery disease (OR=0.79, 95% CI 0.36-1.73), current-smoking (OR=0.90, 95% CI 0.31-2.64) and malignancy (OR=1.01, 95% CI 0.50–2.01) (Fig 3H-3L). C-reactive protein (CRP) (MD=16.58, 95% CI -22.67–55.82), hemoglobin (mean difference=-2.38, 95% CI -7.13–2.37), and fibrinogen (MD=0.23, 95% CI -0.23–0.69) did not show significant difference between thrombotic and non-thrombotic patients with COVID-19 (Fig 3M-3O). Diabetes (OR=0.73, 95% CI 0.47–1.15) tended to be protective factors in our study (Fig 3P). Finally, consisted with our knowledge, patients accompanied by thrombotic events were at higher risk of death (OR=2.39, 95% CI 1.36–4.20) (Fig 3Q).