Hemostatic Changes in Patients with COVID-19: A Meta-Analysis with Meta-Regressions

Matteo Nicola Dario Di Minno, Ilenia Calcaterra, Roberta Lupoli, Antonio Storino, Giorgio Alfredo Spedicato, Mauro Maniscalco, Alessandro Di Minno, Pasquale Ambrosino

Research output: Contribution to journalArticlepeer-review


BACKGROUND: Complications of coronavirus disease 2019 (COVID-19) include coagulopathy. We performed a meta-analysis on the association of COVID-19 severity with changes in hemostatic parameters.

METHODS: Data on prothrombin time (PT), activated partial thromboplastin time (aPTT), D-Dimer, platelets (PLT), or fibrinogen in severe versus mild COVID-19 patients, and/or in non-survivors to COVID-19 versus survivors were systematically searched. The standardized mean difference (SMD) was calculated.

RESULTS: Sixty studies comparing 5487 subjects with severe and 9670 subjects with mild COVID-19 documented higher PT (SMD: 0.41; 95%CI: 0.21, 0.60), D-Dimer (SMD: 0.67; 95%CI: 0.52, 0.82), and fibrinogen values (SMD: 1.84; 95%CI: 1.21, 2.47), with lower PLT count (SMD: -0.74; 95%CI: -1.01, -0.47) among severe patients. Twenty-five studies on 1511 COVID-19 non-survivors and 6287 survivors showed higher PT (SMD: 0.67; 95%CI: 0.39, 0.96) and D-Dimer values (SMD: 3.88; 95%CI: 2.70, 5.07), with lower PLT count (SMD: -0.60, 95%CI: -0.82, -0.38) among non-survivors. Regression models showed that C-reactive protein values were directly correlated with the difference in PT and fibrinogen.

CONCLUSIONS: Significant hemostatic changes are associated with COVID-19 severity. Considering the risk of fatal complications with residual chronic disability and poor long-term outcomes, further studies should investigate the prognostic role of hemostatic parameters in COVID-19 patients.

Original languageEnglish
Article number2244
JournalJournal of Clinical Medicine
Issue number7
Publication statusPublished - Jul 15 2020

Fingerprint Dive into the research topics of 'Hemostatic Changes in Patients with COVID-19: A Meta-Analysis with Meta-Regressions'. Together they form a unique fingerprint.

Cite this