COVID-19 (SARS-CoV-2) infection and thrombotic conditions: A systematic review and meta-analysis

Abstract

Background: COVID-19 is an infectious disease caused by SARS-CoV-2 associated with haematological manifestations (thrombolytic events).

Aims: Considering the high prevalence of the thrombotic scenarios associated with COVID-19, the aim of this study was to perform a systematic review of the available literature, concerning the relation of COVID-19 and the thrombotic events, and identify prognostic factors for these events.

Materials & methods: PubMed, Web of Science and Scopus databases were searched. Independent reviewers conducted all flow diagram steps. For qualitative analysis, Oxford level of evidence and Newcastle-Ottawa scale were used in the eligible articles. For the prognostic factors, a meta-analysis was conducted to age, number of neutrophils and platelets, and levels of ferritin, C-reactive protein, lactate dehydrogenase and D-dimer. Publication bias was accessed by funnel plot and by trim-and-fill test. Trim-and-fill test was also applied to evaluate meta-analysis bias.

Results: Twenty articles were included in the qualitative analysis, and 6 articles were included in the meta-analysis. Case-control studies showed bias related to exposure, and the main bias in cohort studies were related to selection and outcome. All articles received score 4 for the level of evidence. Hypertension and diabetes were the comorbidities more frequently associated with thrombolytic events. Significant results were found regarding D-dimer (P < .0001) and age (P = .0202) for thrombotic events in patients diagnosed with COVID-19.

Conclusion: Patients older than 60 years, with hypertension, diabetes and D-Dimer values above 3.17 µg/mL, can be considered prognostic factors for developing thrombotic events due to COVID-19.

Keywords: COVID-19; D-dimer; SARS-CoV-2; meta-analysis; prognostic factors; systematic review; thrombosis.

FIGURE 1

Flow diagram for the systematic review according to PRISMA Guidelines

FIGURE 2

Meta‐analysis illustrated in a forest plot showing significant results related to age in fixed effect model (A). Trim‐and‐fill test illustrated in a forest plot showing that Demelo‐Rodrígues et al 2020 and Artifoni et al 2020 are responsible for publication and meta‐analysis biases. There was no significant result for random effect model (B). Meta‐analysis illustrated in a forest plot performed after the exclusion of the articles responsible for biases. Significant results can be observed for fixed effect model (C). Experimental group was formed by patients tested positive for COVID‐19 with thrombolytic event. Control group was formed by patients tested positive for COVID‐19 without thrombolytic event. CI, confidence interval; MD, mean difference; SD, standard deviation; SeTE, standard error of treatment estimate; TE, estimate of treatment effect. Arrows indicate the direction of the effect

FIGURE 3

Meta‐analysis illustrated in a forest plot showing no significant results related to platelet count in fixed effect model (A). Trim‐and‐fill test illustrated in a forest plot showing that Demelo‐Rodrígues et al 2020 are responsible for publication and meta‐analysis biases related to platelet count in a random effect model (B). Meta‐analysis illustrated in a forest plot showing no significant results related to neutrophil count in fixed effect model (C). Funnel plot showing no publication bias for neutrophil count (D). Meta‐analysis illustrated in a forest plot showing no significant results related to C‐reactive protein in fixed effect model (E). Funnel plot showing no publication bias for C‐reactive protein (F). Experimental group was formed by patients tested positive for COVID‐19 with thrombolytic event. Control group was formed by patients tested positive for COVID‐19 without thrombolytic event. CI, confidence interval; MD, mean difference; SD, standard deviation; SeTE, standard error of treatment estimate; TE, estimate of treatment effect. Arrows indicate the direction of the effect

FIGURE 4

Meta‐analysis illustrated in a forest plot showing significant results for D‐dimer in fixed effect model (A). Trim‐and‐fill test illustrated in a forest plot showing significant results in a random effect model and no publication and meta‐analysis biases for D‐dimer (B). Meta‐analysis illustrated in a forest plot showing no significant results for lactate dehydrogenase in fixed effect model (C). Trim‐and‐fill test illustrated in a forest plot showing that Demelo‐Rodrígues et al 2020 and Artifoni et al 2020 are responsible for publication and meta‐analysis biases related to lactate dehydrogenase (D). Meta‐analysis illustrated in a forest plot showing no significant results for ferritin in fixed effect model (E). Funnel plot showing no publication bias for ferritin (F). Experimental group was formed by patients tested positive for COVID‐19 with thrombolytic event. Control group was formed by patients tested positive for COVID‐19 without thrombolytic event. CI, confidence interval; MD, mean difference; SD, standard deviation; SeTE, standard error of treatment estimate; TE, estimate of treatment effect. Arrows indicate the direction of the effect