Pooled prevalence of deep vein thrombosis among coronavirus disease 2019 patients

To the editor, The article by Ren et al. reported that there was an extremely high incidence (85.4%) of lower extremity deep venous thrombosis (DVT) among 48 patients with severe coronavirus disease 2019 (COVID-19) in Wuhan, China [1]. As the global pandemic of COVID-19, there have been several studies on the incidence, risk factors, and preventive strategies of DVT [1–4]. However, the incidence of DVT has been reported diversely among different clinical centers. Thus, we performed a meta-analysis to estimate the pooled prevalence of DVT in confirmed COVID-19 patients. We searched PubMed, EMBASE, Web of Science, and medRxiv databases until June 22, 2020, for relevant studies, using the keywords (“coronavirus” or “COVID-19” or “SARS-CoV-2” or “2019-nCoV”) and (“thrombosis” or “thrombi” or “thrombus”). In addition, we screened out the relevant potential articles in the references of selected studies. Articles reporting the prevalence of DVT in confirmed COVID-19 patients were included. The pooled prevalence and its 95% confidence interval (CI) were used to estimate the combined effects. We calculated the prevalence estimates with the variance stabilizing double arcsine transformation [5, 6]. The heterogeneity among studies was assessed with the I statistic and Cochran’s Q test. The meta-regression and subgroup analysis were used to investigate the potential heterogeneity sources (such as sample size, prevalence of prophylaxis in COVID-19 patients, location, design of studies, screening methods of DVT, and COVID-19 patients in intensive care unit (ICU)). We chose Egger’s test and Begg’s test to assess publication bias. All analyses were performed using the Stata 11.2 (StataCorp, College Station, TX), and a twotailed P value < 0.05 was considered to be statistically significant. A total of 1202 records were initially identified by our searches. We finally included 28 articles in our meta-analysis. The basic characteristics of included studies are shown in Table 1. There were 397 DVT cases in a total of 4138 COVID-19 patients. The pooled estimate of the prevalence for DVT was 16% by using a random-effects model (95% CI 10–23%, P < 0.01, I = 96.81, Q = 846.41, P < 0.01) (Fig. 1a). According to patients’ geographic location, the much higher pooled prevalence of DVT was found in COVID-19 patients from China (30%, 95% CI 2– 72%, P = 0.02, I = 98.73%, Q = 313.90, P < 0.01) compared with those from western countries (13%, 95% CI 8–19%, P < 0.01, I = 95.62%, Q = 502.07, P < 0.01) (Fig. 1b). Twenty articles clearly reported the prevalence of DVT in COVID-19 patients treated in ICU or non-ICU. The pooled prevalence of DVT in

To the editor, The article by Ren et al. reported that there was an extremely high incidence (85.4%) of lower extremity deep venous thrombosis (DVT) among 48 patients with severe coronavirus disease 2019  in Wuhan, China [1]. As the global pandemic of COVID-19, there have been several studies on the incidence, risk factors, and preventive strategies of DVT [1][2][3][4]. However, the incidence of DVT has been reported diversely among different clinical centers. Thus, we performed a meta-analysis to estimate the pooled prevalence of DVT in confirmed COVID-19 patients.
We searched PubMed, EMBASE, Web of Science, and medRxiv databases until June 22, 2020, for relevant studies, using the keywords ("coronavirus" or "COVID-19" or "SARS-CoV-2" or "2019-nCoV") and ("thrombosis" or "thrombi" or "thrombus"). In addition, we screened out the relevant potential articles in the references of selected studies. Articles reporting the prevalence of DVT in confirmed COVID-19 patients were included.
The pooled prevalence and its 95% confidence interval (CI) were used to estimate the combined effects. We calculated the prevalence estimates with the variance stabilizing double arcsine transformation [5,6]. The heterogeneity among studies was assessed with the I 2 statistic and Cochran's Q test.
The meta-regression and subgroup analysis were used to investigate the potential heterogeneity sources (such as sample size, prevalence of prophylaxis in COVID-19 patients, location, design of studies, screening methods of DVT, and COVID-19 patients in intensive care unit (ICU)). We chose Egger's test and Begg's test to assess publication bias. All analyses were performed using the Stata 11.2 (StataCorp, College Station, TX), and a twotailed P value < 0.05 was considered to be statistically significant.
In conclusion, more attention should be paid to the prevention and clinical management of DVT, especially for COVID-19 patients in ICU, and timely assessment of DVT is essential. However, there was considerable heterogeneity in our meta-analysis. In addition, there was significant publication bias in our meta-analysis, although we searched four databases as many and as carefully as possible. Finally, we included non-survival patients who were seriously ill and may exaggerate the prevalence of DVT in COVID-19 patients.