Deep vein thrombosis in COVID-19 patients in general wards: prevalence and association with clinical and laboratory variables

Abstract

Background
Preliminary reports suggest a hypercoagulable state in COVID-19. Deep vein thrombosis (DVT) is perceived as a frequent finding in hospitalized COVID-19 patients, but data describing the prevalence of DVT are lacking.Objectives
We aimed to report the prevalence of DVT in COVID-19 patients in general wards, blinded to symptoms/signs of disease, using lower extremities duplex ultrasound (LEDUS) in random patients. We tested the association of DVT with clinical, laboratory and inflammatory markers and also reported on the secondary endpoint of in-hospital mortality.Patients/Methodsn = 263 COVID-19 patients were screened with LEDUS between March 01, 2020 and April 05, 2020 out of the overall n = 1012 admitted with COVID-19.ResultsDVT was detected in n = 67 screened patients (25.5%), n = 41 patients (15.6%) died during the index hospitalization. Multiple logistic regression demonstrated that only C-reactive protein (odds ratio 1.009, 95% CI 1.004–1.013, p < 0.001) was independently associated with the presence of DVT at LEDUS. Both age (odds ratio 1.101, 95% CI 1.054–1.150, p < 0.001) and C-reactive protein (odds ratio 1.012, 95% CI 1.006–1.018, p < 0.001) were instead significantly independently associated with in-hospital mortality.Conclusions
The main study finding is that DVT prevalence in COVID-19 patients admitted to general wards is 25.5%, suggesting it may be reasonable to screen COVID-19 patients for this potentially severe but treatable complication, and that inflammation, measured with serum C-reactive protein, is the main variable associated with the presence of DVT, where all other clinical or laboratory variables, age or D-dimer included, are instead not independently associated with DVT.

Full text

Vol.:(0123456789)
1 3
La radiologia medica
https://doi.org/10.1007/s11547-020-01312-w
ULTRASONOGRAPHY
Deep vein thrombosis inCOVID‑19 patients ingeneral wards:
prevalence andassociation withclinical andlaboratory variables
AnnaMariaIerardi1 · NicolaGaibazzi2· DomenicoTuttolomondo2· StefanoFusco3· VincenzoLaMura4·
FloraPeyvandi4· StefanoAliberti5,6· FrancescoBlasi5,6· DilettaCozzi7· GianpaoloCarraello2,8·
MassimoDeFilippo9
Received: 24 August 2020 / Accepted: 15 November 2020
© Italian Society of Medical Radiology 2021
Abstract
Background Preliminary reports suggest a hypercoagulable state in COVID-19. Deep vein thrombosis (DVT) is perceived
as a frequent finding in hospitalized COVID-19 patients, but data describing the prevalence of DVT are lacking.
Objectives We aimed to report the prevalence of DVT in COVID-19 patients in general wards, blinded to symptoms/signs
of disease, using lower extremities duplex ultrasound (LEDUS) in random patients. We tested the association of DVT with
clinical, laboratory and inflammatory markers and also reported on the secondary endpoint of in-hospital mortality.
Patients/Methods n =263 COVID-19 patients were screened with LEDUS between March 01, 2020 and April 05, 2020
out of the overall n = 1012 admitted with COVID-19.
Results DVT was detected in n=67 screened patients (25.5%), n=41 patients (15.6%) died during the index hospitalization.
Multiple logistic regression demonstrated that only C-reactive protein (odds ratio 1.009, 95% CI 1.004–1.013, p < 0.001)
was independently associated with the presence of DVT at LEDUS. Both age (odds ratio 1.101, 95% CI 1.054–1.150, p <
0.001) and C-reactive protein (odds ratio 1.012, 95% CI 1.006–1.018, p < 0.001) were instead significantly independently
associated with in-hospital mortality.
Conclusions The main study finding is that DVT prevalence in COVID-19 patients admitted to general wards is 25.5%,
suggesting it may be reasonable to screen COVID-19 patients for this potentially severe but treatable complication, and that
inflammation, measured with serum C-reactive protein, is the main variable associated with the presence of DVT, where all
other clinical or laboratory variables, age or D-dimer included, are instead not independently associated with DVT.
Keywords Deep vein thrombosis· Screening· COVID-19· Duplex ultrasound· C-reactive protein
* Anna Maria Ierardi
amierardi@yahoo.it
1 Radiology Department, Fondazione IRCCS Cà Granda
Ospedale Maggiore Policlinico, Milan, Italy
2 Cardiology Department, Azienda Ospedaliero-Universitaria
di Parma, Parma, Italy
3 School ofRadiology, University ofMilan, Milan, Italy
4 Fondazione IRCCS Ca’ Granda, Ospedale Maggiore
Policlinico, U.O.C. Medicina Generale Emostasi e Trombosi,
University ofMilan, Milan, Italy
5 UOC Pneumologia, Fondazione IRCCS Ca’ Granda
Ospedale Maggiore Policlinico, Milan, Italy
6 Department ofPathophysiology andTransplantation,
Università degli Studi di Milano, Milan, Italy
7 Department ofRadiology, Azienda Ospedaliero-Universitaria
Careggi, Florence, Italy
8 Department ofHealth Sciences, Università degli Studi di
Milano, Milan, Italy
9 Department ofMedicine andSurgery (DiMeC), Unit
ofRadiology, University ofParma, Parma, Italy

La radiologia medica
1 3
Introduction
Acute respiratory disease from SARS-CoV-2 (COVID-19)
or coronavirus disease 2019, is an infectious, mainly respira-
tory disease caused by a virus named SARS-CoV-2 and it
is the cause of the ongoing worldwide pandemic in 2020.
One of the reasons that explains the rapid and uncon-
trolled spread of the virus is due to the relatively long
incubation period, in which the infected subjects typically
show no, mild or non-respiratory symptoms and whose
duration can vary between 2 and 14 days during which the
host can be contagious. Inter-human transmission occurs
by contact with infected secretions or by air (coughing or
sneezing) droplets [1, 2].
A key step in the diagnostic process of patients sus-
pected for COVID-19 is the use of chest computed tomog-
raphy. The HRCT (high-resolution computed tomography)
without the use of contrast rapidly allows the radiographic
diagnosis of interstitial pneumonia and provide helpful
prognostic information [3–5]. Chest X-ray has also proved
useful in the emergency setting as a quantitative method of
the extent of SARS-CoV-2 pneumonia, correlating with an
increased risk of intensive care unit admission [6].
There are several reasons why a patient affected by
COVID-19 may be predisposed to concomitant thrombotic
and thromboembolic disease.
Preliminary reports suggest the presence of a hyper-
coagulable state in COVID-19, with several (and partly
conflicting) haemostatic abnormalities being documented,
such as high D-dimer and other fibrin degradation prod-
ucts, high fibrinogen, prolonged activated partial throm-
boplastin time, positivity for lupus anticoagulant and other
abnormalities, suggesting some forms of undetermined
coagulopathy [7–9]. It is unknown whether these hae-
mostatic changes are directly caused by SARS-CoV-2 or
rather a consequence of the cytokine storm following the
systemic inflammatory response syndrome, as observed in
other viral disease [10–13].
In this context, deep vein thrombosis (DVT) is per-
ceived by clinicians on the field as a frequent finding in
hospitalized COVID-19 patients, although data describing
the true prevalence of DVT in COVID-19 are completely
lacking. The few existing published reports on DVT in
hospitalized patients with COVID-19 are either specifi-
cally collected only in intensive care units or they describe
overall thromboembolic events. Existing reports either
enrolled patients based on clinical signs or symptoms of
DVT or of thromboembolic events in general, perform-
ing duplex ultrasound because of this very high index of
clinical suspicion, or addressed only patients admitted
to intensive care units [14–18]; by so doing such stud-
ies select highest-risk, symptomatic patients, in particular
excluding the ones with mild or no symptoms or signs of
DVT, finally reporting biased data on DVT prevalence.
We aim to (a) describe the prevalence of DVT by screen-
ing COVID-19 patients independently from their symp-
tomatic status for DVT, using lower extremities duplex
ultrasound (LEDUS) scan in random patients admitted to
general wards (low or mid intensity care units), where the
vast majority of COVID-19 patients are initially admitted,
(b) test the potential association of DVT with clinical, labo-
ratory and inflammatory markers, (c) report on the second-
ary end point of in-hospital mortality.
Methods
This cross-sectional, single-centre study used a complete
radiologist-performed LEDUS scan to assess the prevalence
of DVT of the lower extremities in laboratory-confirmed
COVID-19 patients between March 01, 2020 and April 05,
2020. In-hospital mortality was also collected as a second-
ary end point.
Patients selection
We screened with LEDUS a random sample of patients with
laboratory-confirmed COVID-19, admitted and treated in
non-intensive care units (low-intensity and mid-intensity
care units) dedicated to COVID-19; such patients were ini-
tially admitted because of clinically-suspected COVID-19,
but only the ones in whom the diagnosis was then confirmed
by SARS-CoV-2 viral nucleic acid assay in nasopharyn-
geal swabs were selected for screening. We did not include
patients who were diagnosed with COVID-19 during hos-
pital stay, but were admitted for other medical conditions.
Index screening test
Patients were selected for LEDUS screening in a quasi-ran-
dom fashion. In fact LEDUS random screening was clini-
cally felt appropriate ad spontaneously implemented by the
radiology department, after the notion spread that COVID-
19 could be a systemic pro-thrombotic disease. Based on the
voluntary availability of one of the radiologists in charge on
each day of the week, he/she was randomly assigned one of
the designated COVID-19 units of the Fondazione IRCCS
Cà Granda Ospedale Maggiore Policlinico, rotating on a
daily basis, to perform bedside LEDUS in a minimum of 10
random patients for each working day. Patients to be scanned
were chosen starting alternatively from the first or last room
in the ward depending on the day.

La radiologia medica
1 3
Clinical variables
The demographics, traditional cardiovascular risk factors
and the available laboratory blood tests (C-reactive protein,
D-dimer and Fibrinogen) were the independent variables to
be tested for an association with the primary endpoint (DVT)
or with the secondary endpoint (in-hospital death).
Endpoints
The presence of DVT at the full LEDUS scan was the pri-
mary endpoint; in-hospital death, collected at least 30 days
after admission was the secondary endpoint.
This study complied with the edicts of the 1975 Dec-
laration of Helsinki and was approved by our Institutional
Review Board.
Statistical methods
No statistical sample size calculation was performed a priori,
and sample size was equal to the number of patients enrolled
during the study period. Categorical variables are expressed
as number of patients (percentage) with 95% CIs, and con-
tinuous variables as mean (SD) or median (interquartile
range [IQR]) as appropriate. The means for continuous vari-
ables were compared using independent group t tests when
the data were normally distributed, otherwise, the Mann-
Whitney test was used. Proportions for categorical variables
were compared using the χ2 test, although the Fisher exact
test was used when data were limited. Stepwise multi-
ple logistic regression was used to assess the relationship
between, first the demographic and clinical variables and
then adding laboratory variables, and the end point of the
detection of DVT at LEDUS; secondarily we tested the end
point of in-hospital mortality. All variables with p<0.1 on
univariable analysis were considered for the inclusion into
multivariable logistic regression models. A 2-sided p<0.05
was considered statistically significant. All statistical analy-
ses were performed with Stata statistical software, version
15.0 (StataCorp LLC, USA).
Results
Two hundred and sixty three patients laboratory-confirmed
COVID-19 patients were randomly screened with LEDUS,
out of the overall 1012 admitted to the hospital with con-
firmed COVID-19 in the same period of time. Mean age
was 63± 15, n = 175 patients were male (66.5%) and
n=41 patients (15.6%) died during their index hospitaliza-
tion, n=222 (84.4%) were discharged home alive. DVT
was detected in 67 of the 263 screened patients (25.5%),
among which 22 DVT were bilateral (32.8%). In 21 patients
DVT was found in the femoral veins (31.3%), 18 in the
popliteal veins (26.9%) and 28 in the calf veins (41.8%)
(Table1 reports baseline characteristics and frequencies of
Table 1. Demographics,
clinical, laboratory tests,
LEDUS results and frequency
of end points.
DVT deep vein thrombosis, LEDUS lower extremities duplex ultrasound
No.(%, if not otherwise specified) Total
Demographics
Number of patients 263
Age, median [lower–upper quartile], y 63 [54–76]
Female sex 88 (33)
Risk factors and patient history
Hypertension 128 (49)
Dyslipidaemia 27 (10)
Current Smoker 17 (6)
Diabetes mellitus 53 (20)
Obesity 45 (17)
History of prior DVT17 (3)
Laboratory blood tests
C-reactive protein, median [lower–upper quartile] (mg/l) 52 [13–115]
D-dimer, median [lower–upper quartile] ng/ml 1332 [809–3779]
Fibrinogen, [lower–upper quartile] ng/dl 536 [390–691]
End points
Patients with DVT at LEDUS 67 (25.5)
Bilateral DVT 22 (33)
Femoral veins DVT 21 (31)
Patients who died in-hospital 41 (16)

La radiologia medica
1 3
end points). LEDUS was performed at a mean 9±6 day
after hospital admission. All patients, from admission and at
least until the day LEDUS was performed, were per hospital
protocol treated with prophylactic doses of weight-adjusted
enoxaparin (100 international units per kilogram, once per
day, the dose being halved in severe chronic kidney disease).
Primary end point
Figure1 shows the box and whisker plots graphically dem-
onstrating the distribution of continuous independent vari-
ables—age, C-reactive protein, D-dimer and fibrinogen—in
patients with and without the finding of DVT at LEDUS.
When testing univariable association of the demographic
or clinical variables (Table2, left column) only the pres-
ence of hypertension (odds ratio 1.858, 95% CI 1.021–3.380,
p=0.042) was significantly associated with DVT; among
laboratory variables, C-reactive protein (odds ratio 1.009,
95% CI 1.005–1.012, p< 0.001), D-dimer (odds ratio
1.000, 95% CI 1.000–1.000, p=0.021) and fibrinogen
(odds ratio 1.003, 95% CI 1.002–1.004, p<0.001) were
also significantly associated with DVT. Since C-reactive
protein and fibrinogen were strongly and significantly cor-
related (r=0.610, p<0.001) only C-reactive protein and
not fibrinogen (mainly because C-reactive protein use is
more widespread) were inserted in the multiple logistic
regression analysis. Stepwise multiple logistic regression
demonstrated that only C-reactive protein (odds ratio 1.009,
95% CI 1.004–1.013, p<0.001) was finally independently
associated with the presence of DVT at LEDUS (Table2,
right columns).
Secondary end point
Regarding the secondary end point of in-hospital mortal-
ity, age (odds ratio 1.089, 95% CI 1.056–1.12, p<0.001),
dyslipidaemia (odds ratio 3.286, 95% CI 1.216–8.877,
p=0.019) and C-reactive protein (odds ratio 1.006, 95% CI
1.005–1.013, p<0.001) were significantly associated in uni-
variable assessment, while in multivariable stepwise logistic
Fig. 1 Distribution of main continuous independent variables in the groups with and without deep vein thrombosis (DVT) at lower extremities
duplex ultrasound. DVT: deep vein thrombosis

La radiologia medica
1 3
Table 2 Relationship of demographics, clinical factors, and serum biomarkers, with deep vein thrombosis at LEDUS
Bold values indicate the statistical significance of some of them; fibrinogen was not inserted in model 2 because of collinearity with C-reactive protein
DVT Deep vein thrombosis
* Fibrinogen has not been inserted in Model 2 because of collinearity with C-Reactive protein
Deep Vein thrombosis of the Lower Extremities
Univariable analysis Multivariable analysis
OR (95% CI) p value Model 1: clinical OR
(95% CI)
p- alue Model2:Clinical +
serum laboratory
markers OR (95% CI)
p value
Age 1.014 (0.996–1.034) 0.128 – –– –
Female sex 1.858 (0.770–2.439) 0.283 – – – –
Hypertension 1.25 (1.021–3.380) 0.042 0.781 (0.366–1.666) 0.523 – –
Dyslipidemia 2.439 (0.919–6.472) 0.073 2.801 (1.007–7.787) 0.048 2.184 (0.767–6.220) 0.144
Current Smoker 0.399(0.088–1.807) 0.233 – – – –
Diabetes mellitus 1.476 (0.676–3.220) 0.328 – – – –
Obesity 0.682 (0.247–1.882) 0.460 – – – –
History of prior DVT 4.085 (0.890–18.746) 0.070 3.692 (0.589–23.160) 0.163
D-dimer 1.00006 (1.00001–
1.00011)
0.021 –– – 1.00004 (0.999–
91.00009)
0.093
C-reactive protein 1.009 (1.005–1.012) <0.000 – – 1.009 (1.004–1.013) <0.000
Fibrinogen 1.003(1.002–1.004) <0.000 – – * –

La radiologia medica
1 3
regression, only age (odds ratio 1.101, 95% CI 1.054–1.150,
p<0.001) and C-reactive protein (odds ratio 1.012, 95% CI
1.006–1.018, p<0.001) were significantly and indepen-
dently associated with in-hospital mortality.
Discussion
The main finding of this screening study is the descrip-
tion that DVT in random, laboratory-confirmed COVID-19
patients admitted to general wards is 25.5%. The prevalence
of DVT in patients admitted because of COVID-19 is here
reported for the first time, to the best of our knowledge,
using a LEDUS screening strategy in a random sample of
subjects with COVID-19. We did not assess patients pre-
selected based either on their extreme clinical severity (as
done when assessing only patients in intensive care units) or
based on the presence of symptoms and signs of DVT. We
think the current study supports the concept that approxi-
mately one quarter of COVID-19 patients admitted to gen-
eral wards in fact have DVT, by so doing suggesting it may
be reasonable to screen COVID-19 patients for this poten-
tially severe but treatable complication.
The second finding is that the grade of inflammation, in
this case measured with serum C-reactive protein, is the
main (and only independent) variable associated with the
presence of DVT, where all other clinical or laboratory
variables, age or D-dimer included, are instead not inde-
pendently associated with DVT.
Interestingly, the commonly used threshold of a
D-dimer >500 ng/ml to suspect DVT (or other thrombo-
embolic events) in general patients did not fit in the spe-
cific COVID-19 setting, in which D-dimer is in fact known
to be generally increased [19]. For example, in the current
study the median D-dimer value was 1332 ng/ml (lower-
upper quartile 809–3779 ng/ml) and the use of the <500
ng/ml cut-off would classify only 28 patients (11%) as low-
risk for DVT; unfortunately, this 500 ng/ml D-dimer cut-
off also demonstrated suboptimal sensitivity in our cohort,
“missing” the DVT diagnosis in 3 patients out of the 28
with a D-dimer value lower than 500 ng/ml. The 3 patients
with DVT and D-dimer <500 ng/ml, however, had very
high C-reactive protein values, which highlights the role
of inflammation. In fact, if we alternatively approached
COVID-19 patients starting from C-reactive protein val-
ues, it is of interest that in the 36 patients (14%) with nor-
mal C-reactive protein at admission (normal range is 0.5–5
mg/l), no one had DVT at LEDUS, in spite of a frequently
high D-dimer (>500 ng/ml in 26 out of 36 patients). On
the other side of the tail of C-reactive protein distribution,
if we consider the 18 patients with high inflammatory sta-
tus, according to a C-reactive protein >150 mg/l, 11 out of
18 patients (61%) had DVT at LEDUS. These observations
lead to speculate that the grade of systemic inflammation
may be the key determinant facilitating DVT in COVID-19
and this is confirmed by the results of the multiple logistic
regression analysis reported in Table2, showing that only
C-reactive protein (not D-dimer or other clinical variables)
is independently associated with DVT. According to our
study, the finding of an extremely high or extremely low
C-reactive protein value can certainly better inform the cli-
nician reinforcing the decision to indicate LEDUS or not
as a screening tool for DVT in a given COVID-19 patient.
C-reactive protein, but also age in this case, were inde-
pendently significantly associated with the secondary end
point of in-hospital mortality. In this regard, our study
confirms the association between C-reactive protein and
mortality suggested at this time only in few pre-print
reports (not yet published in peer-reviewed journals), in
which age and inflammation severity were among the main
key variables associated with mortality [19–22].
Limitations
The current study is a cross-sectional screening study, with
the addition of the collection of short-term, in-hospital
death follow-up, in which the available clinical and labo-
ratory data were only the ones available from the chart or
other electronic records. Cross-sectional studies cannot
prove causation but only inform on associations detected.
The LEDUS was performed by voluntary trained radi-
ologists operating at the bedside in the complex clini-
cal contagious scenario of COVID-19 wards, so that the
LEDUS images are not available for review, but only the
written official report was stored, available from the elec-
tronic chart and radiology reporting system. Randomiza-
tion of patients to be screened with LEDUS was directly
performed by the radiologist in charge on the field, simply
starting to scan patients either from the first room of the
unit and on, or reverse from the last room, depending on
the day, so that while this is theoretically not an ideal ran-
domization process, it was the best that could be achieved
during an infective outbreak and it should have avoided
most types of selection bias. Several laboratory variables
were available only in a percentage of the study population
and consequently they could not be used in the current
analysis, and among them, granular body mass index was
not available in all patients, only the binary variable “obe-
sity” found in the charts, defined by a body mass index
>29 kg/m2 being available for all patients.
Funding This study was not supported by any funding.

La radiologia medica
1 3
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of
interest.
Ethical standards All procedures performed in studies involving
human participants were in accordance with the ethical standards of
the institutional and/or national research committee and with the 1964
Helsinki Declaration and its later amendments or comparable ethical
standards. The study was approved by our Institutional Review Board
(Radcovid03/2020).
Informed consent The informed consent has been obtained from all
patients.
References
1. Frizzelli A, Tuttolomondo D, Aiello M etal (2020) (2020) What
happens to people’s lungs when they get coronavirus disease
2019? Acta Biomed 91(2):146–149. https ://doi.org/10.23750 /
abm.v91i2 .9574
2. Tuttolomondo D, Frizzelli A, Aiello M etal (2020) Beyond the
lung involvement in COVID-19 patients. A review. Minerva Med.
https ://doi.org/10.23736 /S0026 -4806.20.06719 -1 (Published
online ahead of print, 2020 Jun 19)
3. Carotti M, Salaffi F, Sarzi-Puttini P etal (2020) Chest CT fea-
tures of coronavirus disease 2019 (COVID-19) pneumonia: key
points for radiologists. Radiol Med 125(7):636–646. https ://doi.
org/10.1007/s1154 7-020-01237 -4
4. Gaia C, Maria Chiara C, Silvia L etal (2020) Chest CT for early
detection and management of coronavirus disease (COVID-19):
a report of 314 patients admitted to Emergency Department with
suspected pneumonia. Radiol Med. https ://doi.org/10.1007/s1154
7-020-01256 -1 (Published online ahead of print, 2020 Jul 29)
5. Ierardi AM, Wood BJ, Arrichiello A etal (2020) Preparation of a
radiology department in an Italian hospital dedicated to COVID-
19 patients. Radiol Med 125(9):894–901. https ://doi.org/10.1007/
s1154 7-020-01248 -1
6. Cozzi D, Albanesi M, Cavigli E etal (2020) Chest X-ray in new
Coronavirus Disease 2019 (COVID-19) infection: findings and
correlation with clinical outcome. Radiol Med 125(8):730–737.
https ://doi.org/10.1007/s1154 7-020-01232 -9 (Epub 2020 Jun 9)
7. Helms J, Tacquard C, Severac F etal (2020) High risk of throm-
bosis in patients with severe SARS-CoV-2 infection: a multi-
center prospective cohort study. Intensive Care Med. https ://doi.
org/10.1007/s0013 4-020-06062 -x (Epub ahead of print)
8. Klok FA, Kruip MJHA, van der Meer NJM etal (2020a) Inci-
dence of thrombotic complications in critically ill ICU patients
with COVID-19. Thromb Res. https ://doi.org/10.1016/j.throm
res.2020.04.013 (Epub ahead of print)
9. Llitjos JF, Leclerc M, Chochois C etal (2020a) High incidence of
venous thromboembolic events in anticoagulated severe COVID-
19 patients. J Thromb Haemost. https ://doi.org/10.1111/jth.14869
(Epub ahead of print)
10. Mehta P, McAuley DF, Brown M etal (2020) COVID-19:
consider cytokine storm syndromes and immunosuppression.
Lancet 395(10229):1033–1034. https ://doi.org/10.1016/S0140
-6736(20)30628 -0 (Epub 2020 Mar 16)
11. Borges AH, O’Connor JL, Phillips AN etal (2014) (2014) Factors
associated with D-dimer levels in HIV-infected individuals. PLoS
One. 9(3):e90978. https ://doi.org/10.1371/journ al.pone.00909 78
(eCollection)
12. Ramacciotti E, Agati LB, Aguiar VCR etal (2019) Zika and chi-
kungunya virus and risk for venous thromboembolism. Clin Appl
Thromb Hemost. https ://doi.org/10.1177/10760 29618 82118 4
13. Smither SJ, O’Brien LM, Eastaugh L etal (2019) Haemostatic
changes in five patients infected with Ebola virus. Viruses.
11(7):E647. https ://doi.org/10.3390/v1107 0647
14. Klok FA, Kruip MJHA, van der Meer NJM etal (2020b) Inci-
dence of thrombotic complications in critically ill ICU patients
with COVID-19. Thromb Res. https ://doi.org/10.1016/j.throm
res.2020.04.013 (Epub ahead of print)
15. Marone EM, Rinaldi LF (2020) Upsurge of deep vein thrombosis
in patients affected by COVID-19: preliminary data and possible
explanations. J Vasc Surg Venous Lymphat Disord. https ://doi.
org/10.1016/j.jvsv.2020.04.004 (Epub ahead of print)
16. Lodigiani C, Iapichino G, Carenzo L (2020) Venous and arterial
thromboembolic complications in COVID-19 patients admitted
to an academic hospital in Milan. Italy. Thromb Res. 191:9–14.
https ://doi.org/10.1016/j.throm res.2020.04.024 (Epub ahead of
print)
17. Middeldorp S, Coppens M, van Haaps TF (2020) Incidence of
venous thromboembolism in hospitalized patients with COVID-
19. J Thromb Haemost. https ://doi.org/10.1111/jth.14888 (Epub
ahead of print)
18. Llitjos JF, Leclerc M, Chochois C etal (2020b) High incidence of
venous thromboembolic events in anticoagulated severe COVID-
19 patients. J Thromb Haemost. https ://doi.org/10.1111/jth.14869
(Epub ahead of print)
19. Wynants L, Van Calster B, Bonten MMJ etal (2020) Prediction
models for diagnosis and prognosis of covid-19 infection: system-
atic review and critical appraisal. BMJ 7(369):m1328. https ://doi.
org/10.1136/bmj.m1328
20. Lu J, Hu S, Fan R etal (2020) ACP risk grade: a simple mortal-
ity indexfor patients with confirmed or suspected severe acute
respiratory syndrome coronavirus 2 disease (COVID-19) during
the early stage of outbreak in Wuhan, China. medRxiv. https ://doi.
org/10.1101/2020.02.20.20025 510 (Preprint)
21. Xie J, Hungerford D, Chen H etal (2020) Development and exter-
nal validation of a prognostic multivariable model on admission
for hospitalized patients with COVID-19. medRxiv. https ://doi.
org/10.1101/2020.03.28.20045 997 (Preprint)
22. Yan L, Zhang H-T, Xiao Y etal (2020) Prediction of criti-
cality in patients with severe Covid-19 infection using
three clinical features: a machine learning-based prognos-
tic model with clinical data in Wuhan. MedRxiv. https ://doi.
org/10.1101/2020.02.27.20028 027 (Preprint)
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.