Hindawi

International Journal of Clinical Practice

Volume 2022, Article ID 7325060, 9 pages
https://doi.org/10.1155/2022/7325060

Research Article
Chronic Oral Anticoagulation Therapy and Prognosis of Patients
Admitted to Hospital for COVID-19: Insights from the HOPE
COVID-19 Registry

José Miguel Rivera-Caravaca ,1,2 Iván J. Núñez-Gil,3 Gregory Y. H. Lip ,2,4

Aitor Uribarri,5 Marı́a C. Viana-Llamas,6 Adelina Gonzalez,7 Alex F. Castro-Mejı́a,8
Berta Alonso González,9 Emilio Alfonso,10 Juan Fortunato Garcı́a Prieto,11
Chiara Cavallino,12 Bernardo Cortese,13 Gisela Feltes,14 Inmaculada Fernández-Rozas,15
Jaime Signes-Costa,16 Jia Huang,17 Marcos Garcı́a Aguado,18 Martino Pepe,19
Rodolfo Romero,20 Enrico Cerrato,21 Vı́ctor Manuel Becerra-Muñoz,22
Sergio Raposeiras Roubin,23 Francesco Santoro,24,25 Rodrigo Bagur,26
Luciano Sposato,26,27,28,29 Ibrahim El-Battrawy,30 Alvaro López Masjuan,31
Antonio Fernandez-Ortiz,3 Vicente Estrada,3 Carlos Macaya,3 and Francisco Marı́n 1
1
Department of Cardiology, Hospital Clı́nico UniversitarioVirgen de la Arrixaca, University of Murcia,
Instituto Murciano de Investigación Biosanitaria (IMIB-Arrixaca), CIBERCV, Murcia, Spain
2
Liverpool Centre for Cardiovascular Science, University of Liverpool, Liverpool Heart & Chest Hospital, Liverpool, UK
3
Hospital Clı́nico San Carlos Universidad Complutense de Madrid,
Instituto de Investigación Sanitaria del Hospital Clı́nico San Carlos (IdISSC), Madrid, Spain
4
Department of Clinical Medicine, Aalborg University, Aalborg, Denmark
5
Hospital Clı́nico Universitario de Valladolid, Valladolid, Spain
6
Hospital Universitario Guadalajara, Guadalajara, Spain
7
Hospital Universitario Infanta Sofı́a, San Sebastian de los Reyes, Madrid, Spain
8
Hospital General del Norte de Guayaquil IESS Los Ceibos, Guayaquil, Ecuador
9
Hospital Universitario la Paz, Madrid, Spain
10
Instituto de Cardiologı́a y Cirugı́a Cardiovascular, La Habana, Cuba
11
Hospital de Manises, Valencia, Spain
12
Sant’ Andrea Hospital, Vercelli, Italy
13
San Carlo Clinic, Milano, Italy
14
Hospital Nuestra Señora de América, Madrid, Spain
15
Hospital Severo Ochoa, Leganés, Spain
16
Hospital Clı́nico Universitario, INCLIVA, Universidad de Valencia, Valencia, Spain
17
The Second People’s Hospital of Shenzhen, Shenzhen, China
18
Hospital Universitario Puerta de Hierro, Majadahonda, Spain
19
Azienda Ospedaliero Universitaria Consorziale Policlinico di Bari, Bari, Italy
20
Hospital Universitario de Getafe, Madrid, Spain
21
San Luigi Gonzaga University Hospital, Rivoli, Turin, Italy
22
Unidad de Gestión Clı́nica Área del Corazón, Instituto de Investigación Biomédica de Málaga (IBIMA),
Hospital Universitario Virgen de la Victoria, Universidad de Málaga (UMA),
Centro de Investigación Biomédica en Red Enfermedades Cardiovasculares (CIBERCV), Málaga, Spain
23
University Hospital Álvaro Cunqueiro, Vigo, Spain
24
Azienda Sanitaria Locale della Provincia di Barletta-Andria-Trani, Andria, Italy
25
Department of Medical and Surgery Sciences, University of Foggia, Foggia, Italy
26
London Health Sciences Centre, London, Ontario, Canada
27
Schulich School of Medicine and Dentistry, Western University, London, Ontario, Canada
28
Lawson Health Research Institute, London, Ontario, Canada
29
Robarts Research Institute, London, Ontario, Canada
2 International Journal of Clinical Practice

30
University of Mannheim, Mannheim, Germany
31
Hospital Universitario Juan Ramón Jiménez, Huelva, Spain

Correspondence should be addressed to Francisco Marı́n; [email protected]

Received 11 February 2022; Revised 21 March 2022; Accepted 21 April 2022; Published 26 May 2022

Academic Editor: Melanie Deutsch

Copyright © 2022 José Miguel Rivera-Caravaca et al. This is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
Background. Most evidence regarding anticoagulation and COVID-19 refers to the hospitalization setting, but the role of oral
anticoagulation (OAC) before hospital admission has not been well explored. We compared clinical outcomes and short-term
prognosis between patients with and without prior OAC therapy who were hospitalized for COVID-19. Methods. Analysis of the
whole cohort of the HOPE COVID-19 Registry which included patients discharged (deceased or alive) after hospital admission for
COVID-19 in 9 countries. All-cause mortality was the primary endpoint. Study outcomes were compared after adjusting variables
using propensity score matching (PSM) analyses. Results. 7698 patients were suitable for the present analysis (675 (8.8%) on OAC
at admission: 427 (5.6%) on VKAs and 248 (3.2%) on DOACs). After PSM, 1276 patients were analyzed (638 with OAC; 638
without OAC), without significant differences regarding the risk of thromboembolic events (OR 1.11, 95% CI 0.59–2.08). The risk
of clinically relevant bleeding (OR 3.04, 95% CI 1.92–4.83), as well as the risk of mortality (HR 1.22, 95% CI 1.01–1.47; log-rank p
value � 0.041), was significantly increased in previous OAC users. Amongst patients on prior OAC only, there were no differences
in the risk of clinically relevant bleeding, thromboembolic events, or mortality when comparing previous VKA or DOAC users,
after PSM. Conclusion. Hospitalized COVID-19 patients on prior OAC therapy had a higher risk of mortality and worse clinical
outcomes compared to patients without prior OAC therapy, even after adjusting for comorbidities using a PSM. There were no
differences in clinical outcomes in patients previously taking VKAs or DOACs. This trial is registered with NCT04334291/
EUPAS34399.

1. Introduction antagonist (VKA) therapy and patients on DOACs before

admission.
Vascular inflammation, hypercoagulable state, and endo-
thelial dysfunction have been described in patients with 2. Methods
severe acute respiratory syndrome coronavirus-2 (SARS-
CoV-2) infection [1, 2]. As a result, thromboembolic A detailed description of the HOPE COVID-19 Registry has
complications are common in patients with coronavirus been published elsewhere [10, 11]. Briefly, the HOPE
disease 2019 (COVID-19) [3–5]. Therefore, antithrombotic COVID-19 is an ambispective international registry, real-life
therapy, particularly anticoagulation, gained attention in the cohort “all comers” type, including more than 8100 patients
context of COVID-19. Indeed, some studies demonstrated from 9 countries (Canada, China, Chile, Colombia, Cuba,
that anticoagulation may be associated with improved Ecuador, Germany, Italy, and Spain). The study was an
outcomes among patients with COVID-19 [6, 7]. However, initiative without conflicts of interest, no financial remu-
most of the evidence in relation to anticoagulation and neration, and methodological support from the Institute for
COVID-19 refer to the acute hospitalization setting, whereas the Improvement of Health Care (IMAS) foundation
the role of stable oral anticoagulation (OAC) therapy before (Madrid, Spain).
the admission for COVID-19 has not been well explored. All patients discharged (deceased or alive) after hospital
One study suggested a protective role of chronic direct- admissions for COVID-19 were suitable for the study. There
acting OAC (DOAC) therapy in elderly patients with were no exclusion criteria, except for patients’ explicit refusal
COVID-19 [8]. In a preliminary analysis of the International to participate. The first patient was included in February
COVID-19 Clinical Evaluation (HOPE COVID-19) Registry 2020. Clinical and demographic data were collected at in-
published previously, we observed that COVID-19 patients clusion and during the hospitalization in an anonymized
on OAC therapy at hospital admission had lower survival database presented in the electronic format, to be filled in at
and higher mortality risk compared to patients without prior each participating center (NCT04334291/EUPAS34399).
OAC [9]. Reporting of the study conforms to broad EQUATOR
In the present study, we aimed to compare clinical guidelines. The study was performed according to the ethical
outcomes and in-hospital prognosis between patients on principles of Declaration of Helsinki and Good Clinical
prior OAC therapy and patients not on OAC therapy who Practice Guidelines and has been approved by Ethics Re-
were admitted for COVID-19 and enrolled in the HOPE search Committee from the Hospital Clı́nico San Carlos
COVID-19 Registry, using a propensity score matching (Madrid, Spain) (20/241-E) and the Spanish Agency for
(PSM) approach. Second, we aimed to compare clinical Medicines and Health Products (EPA-0D). Given the
outcomes and prognosis between patients on vitamin K anonymous characteristics of the registry and the health
International Journal of Clinical Practice 3

alarm situation generated by the virus, in principle, written survival of the primary outcome depending on the use (or
informed consent was waived. However, at least verbal not) of prior OAC therapy and depending on the use of prior
authorization from the patient (or familiar or caregiver, VKA or DOAC therapy. The risk of suffering from the
when unavailable) was required. primary outcome was assessed by Cox proportional hazard
regression, and results were reported as hazard ratio (HR)
with 95% confidence interval (CI). The risk of suffering from
2.1. Laboratory Analyses. Laboratory parameters were other study outcomes was investigated by logistic regression
considered elevated as defined by local laboratory cutoff analyses, since the exact date for these events was not
levels. However, the HOPE COVID-19 Registry protocol recorded. In these analyses, results were reported as odds
suggested the following as “elevated:” for D-dimer (≥0.5 mg/ ratio (OR) with 95% confidence interval (CI).
L), for procalcitonin (≥0.5 ng/mL), for C-reactive protein Two-sided p values <0.05 were accepted as statistically
(≥10 mg/L), for troponins (>99th percentile), for transam- significant. Statistical analyses were performed using SPSS v.
inases (≥40 U/L), for ferritin (≥336 ng/mL), and for lactate 24.0 (SPSS, Inc., Chicago, IL, USA) and MedCalc v. 16.4.3
dehydrogenase (≥280 U/L). (MedCalc Software bvba, Ostend, Belgium) for Windows.

2.2. Study Outcomes. The primary endpoint for this analysis 3. Results
was in-hospital all-cause mortality. Any thrombotic/
thromboembolic event and any clinically relevant bleeding A cohort of 8168 patients was included. After excluding
were the secondary outcomes. Bleeding was defined as patients with insufficient or not reliable data on previous
“relevant” at the discretion of the attending medical team OAC, 7698 patients remained in the study (4500 (58.5%)
and classified using the BARC bleeding score as type 2, 3, or male; median age of 65 (IQR 51–77) years). Of these, 675
5. (8.8%) were on OAC therapy at hospital admission, 427
Although not classified as primary or secondary out- (5.6%) were on VKAs, and 248 (3.2%) were on DOACs.
comes, other adverse events during hospitalization were
recorded, including renal failure, respiratory insufficiency,
upper respiratory tract infection, heart failure, sepsis, and 3.1. Outcomes on Prior OAC Therapy. In the overall cohort of
systemic inflammatory response syndrome (SIRS). 7698 patients, we found that patients on prior OAC therapy
Local researchers identified, confirmed, and recorded all were less commonly admitted in the intensive care unit
adverse events. The clinical management was decided, in all (ICU) compared to patients not previously taking OACs
cases, by the attending team and researchers had no role in (6.7% vs. 10.1%, p � 0.004). During hospitalization, the
this point. prognosis of patients on prior OAC therapy was also poor,
and these patients had more incident heart failure, renal
failure, sepsis, and SIRS (all with p value <0.001). As ex-
2.3. Statistical Analysis. Quantitative variables were pected, the risk of any clinically relevant bleeding in patients
expressed as mean ± standard deviation (SD) or median and with previous OAC therapy was higher compared to patients
interquartile range (IQR), as appropriate according to the not taking OAC previously (11.6% vs. 3.4%, p < 0.001; OR
Kolmogorov–Smirnov test, whilst categorical variables were 3.71, 95% CI 2.83–4.85), without differences in terms of
expressed as absolute frequencies and percentages. Pearson’s thromboembolic events (3.1% vs. 2.7%, p � 0.493). The risk
chi-squared test was used to compare proportions. Differ- of mortality was found to be significantly increased in pa-
ences between two groups regarding a quantitative variable tients on prior OAC therapy (39.1% vs. 17.0%, p < 0.001; HR
were tested with Student’s t or the Mann–Whitney U tests, as 2.45, 95% CI 2.14–2.79); however, there were significant
appropriate if normally or not normally distributed. differences between patients on prior and not on prior OAC
To compare the risk of the study outcomes among pa- in terms of several comorbidities. We therefore performed
tients on prior OAC therapy and patients without prior PSM to adjust these analyses (Table 1).
OAC therapy, we conducted a propensity score matching After PSM, 1276 patients remained in the study (638 :
(PSM) adjusting for demographics and baseline comor- 638 paired comparisons), with no significant differences
bidities. The risk of the study outcomes among patients on regarding admission to the ICU in patients on prior OAC
prior VKA therapy or DOACs was also evaluated by another compared to patients not previously taking OACs (6.9% vs.
PSM. In both PSMs, those variables that were significantly 6.3%, p � 0.652). The prognosis of patients on prior OAC
different between both cohorts were included in the model therapy during hospitalization was still poor even after
to adjust for differences. Patients were matched 1 : 1 across adjustment, and these patients suffered more commonly
each cohort on a propensity score generated by logistic from heart failure, renal failure, and SIRS (all with p value
regressions using the nearest neighbour technique without <0.05). No significant differences were found in terms of
replacement with a maximum caliper of 0.2, thus avoiding at respiratory insufficiency (67.2% vs. 64.7%; p � 0.280), upper
least 98% of the bias due to the measured confounders. The respiratory tract infection (13.9% vs. 14.1%; p � 0.987), or
value of absolute standardized mean difference <10% in- sepsis (15.0% vs. 12.1%; p � 0.299) (Table 2).
dicated balance of matched cohorts [12, 13]. Similar to the finding observed before PSM, the risk of
Survival analyses by Kaplan–Meier estimates were any clinically relevant bleeding was higher in patients with
performed after PSM to assess differences in event-free previous OAC therapy compared to patients not taking
4 International Journal of Clinical Practice

Table 1: Comparison of clinical characteristics of the study cohort before and after propensity score matching.
Before propensity score matching After propensity score matching
Patients without Patients with prior Patients without Patients with prior
P P
prior OAC OAC prior OAC OAC
value value
N � 7023 N � 675 N � 638 N � 638
Demographic
Male sex, n (%) 4097 (58.3) 403 (59.7) 0.491 386 (60.5) 372 (58.3) 0.425
Age (years), median (IQR) 63 (50–75) 80 (72–86) <0.001 80 (72–86) 80 (72–86) 1.000
Race (non-Caucasian), n (%) 1603 (22.8) 59 (8.7) <0.001 49 (7.7) 59 (9.2) 0.315
Body mass index (kg/m2),
27.1 (24.2–30.7) 27.7 (25.0–31.2) 0.011 26.9 (24.5–30.5) 26.7 (25.0–31.3) 0.168
median (IQR)
Baseline comorbidities, n (%)
Hypertension 3176 (45.2) 542 (80.3) <0.001 433 (68.0) 516 (80.9) 0.053
Diabetes mellitus 1257 (17.9) 198 (29.3) <0.001 168 (27.0) 190 (30.3) 0.198
Heart failure 128 (1.8) 46 (6.8) <0.001 35 (5.5) 39 (6.1) 0.632
Stroke/TIA 439 (6.3) 131 (19.4) <0.001 92 (14.4) 122 (19.1) 0.437
Chronic kidney disease 369 (5.3) 115 (17.0) <0.001 69 (11.0) 109 (17.0) 0.487
Vascular disease∗ 543 (7.7) 102 (15.1) <0.001 93 (14.6) 88 (13.8) 0.688
Hypercholesterolemia 2096 (29.8) 344 (51.0) <0.001 288 (45.1) 326 (51.1) 0.085
Current smoking habit 407 (5.8) 35 (5.2) 0.243 21 (3.3) 31 (4.9) 0.071
COPD/SAHS 419 (6.0) 104 (15.4) <0.001 81 (12.7) 84 (13.2) 0.802
History of malignant disease 822 (11.7) 139 (20.6) <0.001 129 (20.2) 129 (20.2) 1.000
Liver disease 238 (3.4) 33 (4.9) 0.001 30 (4.7) 31 (4.9) 0.795
Dysthyroidism 334 (4.8) 40 (5.9) 0.177 37 (5.8) 40 (6.3) 0.724
Any dependency level 819 (11.7) 210 (31.1) <0.001 177 (28.2) 194 (30.6) 0.365
Concomitant treatment at admission, n (%)
Beta-blockers 865 (12.3) 328 (48.6) <0.001 132 (20.7) 311 (48.7) <0.001
ACEi/ARBs 2320 (33.0) 369 (54.7) <0.001 311 (48.7) 350 (54.9) 0.086
Antiplatelet therapy 1229 (17.5) 74 (11.0) <0.001 199 (31.2) 72 (11.3) <0.001
Laboratory parameters at admission
Creatinine (mg/dL), median
0.90 (0.72–1.17) 1.19 (0.90–1.64) <0.001 0.98 (0.78–1.42) 1.20 (0.88–1.66) <0.001
(IQR)
Hemoglobin (g/dL), median
14.0 (12.0–15.0) 13.0 (11.0–14.0) <0.001 13.0 (12.0–15.0) 13.0 (11.0–14.0) <0.001
(IQR)
9
Platelet count (×10 /L), 179.0
203.0 (155.0–265.8) <0.001 195.0 (145.0–260.8) 181.0 (138.0–241.0) 0.019
median (IQR) (136.0–240.0)
Elevated D-dimer, n (%) 3921 (55.8) 358 (53.0) 0.036 425 (66.6) 342 (53.6) <0.001
Elevated procalcitonin, n (%) 1048 (14.9) 126 (18.7) 0.001 103 (16.1) 123 (19.3) 0.299
Elevated C-reactive protein, n
5841 (83.2) 608 (90.1) <0.001 566 (88.7) 576 (90.3) 0.657
(%)
Elevated troponins, n (%) 527 (7.5) 107 (15.9) <0.001 54 (8.5) 100 (15.7) <0.001
Elevated transaminases, n (%) 2598 (37.0) 220 (32.6) 0.009 216 (33.9) 210 (32.9) 0.023
Elevated ferritin, n (%) 2306 (32.8) 207 (30.7) 0.424 198 (31.0) 198 (31.0) 1.000
Elevated lactate
4414 (62.9) 464 (68.7) 0.005 427 (66.9) 440 (69.0) 0.466
dehydrogenase, n (%)
ACEi, angiotensin-converting enzyme inhibitors; ARBs, angiotensin II receptor blockers; IQR, interquartile range; TIA, transient ischemic attack; COPD/
SAHS, chronic obstructive pulmonary disease/sleep apnea-hypopnea syndrome. ∗ Coronary artery disease and/or peripheral artery disease.

OAC previously (11.4% vs. 4.1%, p < 0.001; OR 3.04, 95% CI SIRS-related: 4.9% vs. 3.6%; sepsis-related: 3.3% vs. 7.6%;
1.92–4.83), without differences in the risk of thromboem- other reasons or combined causes of death: 29.6% vs. 23.4%;
bolic events (3.3% vs. 3.0%, p � 0.748; OR 1.11, 95% CI p � 0.187).
0.59–2.08). There was increased mortality in patients who
were on previous OAC therapy in comparison to patients
who were not on previous OAC (38.1% vs. 30.9%, 3.2. Impact of OAC Type. In patients on prior OAC therapy,
p � 0.007), with a significantly higher risk of death (HR 1.22, we observed significant differences regarding age and
95% CI 1.01–1.47), also confirmed by the Kaplan–Meier comorbid conditions between patients who were previously
analysis (log-rank p value � 0.041) (Figure 1). There were no taking VKAs and those who were on prior DOAC therapy. We
differences between patients on prior or non-prior OAC performed another PSM to balance these characteristics. This
therapy regarding specific causes of death (cardiovascular analysis demonstrated no differences in the remaining 464
death: 2.6% vs. 2.5%; respiratory-related: 59.7% vs. 62.9%; subjects: 232 on VKAs and 232 on DOACs, as given in Table 3.
International Journal of Clinical Practice 5

Table 2: Clinical outcomes during hospitalization after propensity score matching.

Patients without prior OAC (N � 638) Patients with prior OAC (N � 638)
P
Incidence per 100 patients- Incidence per 100 patients- OR (95% CI) value
N (%) N (%)
days (95% CI) days (95% CI)
1.11
Intensive care unit admission 40 (6.3) 0.52 (0.37–0.71) 44 (6.9) 0.58 (0.42–0.77) 0.652
(0.71–1.73)
151 212 1.61
Renal failure 1.97 (1.67–2.31) 2.77 (2.41–3.17) 0.001
(23.7) (33.2) (1.26–2.06)
413 429 1.09
Respiratory insufficiency 5.39 (4.89–5.94) 5.60 (5.09–6.16) 0.280
(64.7) (67.2) (0.86–1.38)
Upper respiratory tract 90 89 0.99
1.18 (0.95–1.44) 1.16 (0.93–1.43) 0.987
infection (14.1) (13.9) (0.72–1.35)
65 115 1.93
Heart failure 0.85 (0.66–1.08) 1.50 (1.24–1.80) <0.001
(10.2) (18.0) (1.39–2.68)
77 96 1.29
Sepsis 1.01 (0.79–1.26) 1.25 (1.02–1.53) 0.299
(12.1) (15.0) (0.93–1.78)
Systemic inflammatory 129 181 1.55
1.69 (1.41–2.00) 2.36 (2.03–2.73) 0.003
response syndrome (20.2) (28.4) (1.20–2.02)
197 243 1.38
All-cause mortality 2.57 (2.23–2.96) 3.17 (2.79–3.60) 0.007
(30.9) (38.1) (1.09–1.74)
Any thrombotic/ 1.11
19 (3.0) 0.25 (0.15–0.39) 21 (3.3) 0.27 (0.17–0.42) 0.748
thromboembolic event (0.59–2.08)
73 3.04
Any clinically relevant bleeding 26 (4.1) 0.34 (0.22–0.50) 0.96 (0.75–1.20) <0.001
(11.4) (1.92–4.83)

100 p � 0.703), with a non-significant difference in mortality

Event-Free Survival (%)

90 risk amongst previous VKA users (HR 0.84, 95% CI

80 0.62–1.12; p � 0.233) (Figure 2).
70
60
50 3.3. Anticoagulation Management during Hospitalization.
40 Regarding anticoagulation during hospitalization in the
30 PSM cohort of previous vs. no previous OAC, most patients
Log-rank test p-value = 0.041
20 not taking OAC previously were prescribed heparin (79.9%,
0 10 20 30 40 50 60 70 382/478) and 19.2% (92/478) did not receive anti-
Time (days) coagulation. In patients who were previously on OAC, 65.5%
Number at risk (330/504) were switched to heparin, 25% (126/504) con-
Non-prior OAC 638 341 176 83 43 25 12 7 tinued on OAC, and 9.5% (48/504) did not receive any
Prior OAC 638 336 185 114 71 42 21 15 anticoagulation therapy. These proportions were signifi-
Non-prior OAC cantly different (p < 0.001).
Prior OAC In the PSM cohort of previous VKA vs. DOAC, most
Figure 1: Comparison of survival curves between patients on prior patients under either therapy received heparin during
OAC and nonprior OAC. Solid line, nonprior OAC; dashed line, hospitalization, without differences between drug families
prior OAC. (116 vs. 114; p � 0.567). Those patients who were maintained
on OAC during admission were predominately treated with
the same OAC that they were before (95.5% for previous
With this matched cohort, the rate of ICU admission VKAs users and 93.2% for previous DOACs users;
between patients on VKAs (14, 6.0%) and patients on p < 0.001).
DOACs (14, 6.0%) was similar (p � 1.000). There were no
differences in terms of respiratory insufficiency, heart fail- 4. Discussion
ure, development of renal failure, upper respiratory tract
infection, sepsis, or SIRS, in patients on prior VKAs or In this study of the HOPE COVID-19 Registry, including a
DOACs (all p > 0.005) (Supplementary Table 1). large cohort of patients hospitalized for COVID-19, we
No significant differences in the incidences of clinically demonstrate that the risk of in-hospital worse clinical
relevant bleeding or thromboembolic events were observed outcomes was higher in patients with prior OAC therapy,
in patients previously taking VKAs compared to DOACs even after adjustment by a PSM. Importantly, this study
(1.01 vs. 0.83 per 100 patient-days (p � 0.458) and 0.29 vs. population showed a 22% higher risk of mortality and
0.11 per 100 patient-days (p � 0.127), respectively) (Sup- bleeding, without significant differences in the prognosis
plementary Table 1). Mortality rate between previous VKA with regard to the particular anticoagulant drug, i.e., VKAs
and DOAC users was also similar (37.9% vs. 39.7%, versus DOACs.
6 International Journal of Clinical Practice

Table 3: Comparison of clinical characteristics of patients on VKA or DOAC prior admission after propensity score matching.
Patients on prior VKA Patients on prior DOAC
P value
N � 232 N � 232
Demographic
Male sex, n (%) 139 (59.9) 127 (54.7) 0.260
Age (years), median (IQR) 80 (72–87) 81 (73–86) 0.575
Body mass index (kg/m2), median (IQR) 28.0 (25.1–31.6) 27.3 (24.3–31.0) 0.445
Baseline comorbidities, n (%)
Hypertension 189 (81.5) 178 (76.7) 0.209
Diabetes mellitus 61 (26.3) 74 (31.9) 0.184
Heart failure 15 (6.5) 14 (6.0) 0.848
Stroke/TIA 42 (18.1) 47 (20.3) 0.555
Chronic kidney disease 31 (13.4) 31 (13.4) 1.000
Vascular disease∗ 33 (14.2) 34 (14.7) 0.895
Hypercholesterolemia 112 (48.3) 113 (48.7) 0.926
Current smoking habit 13 (5.6) 9 (3.9) 0.143
COPD/SAHS 30 (12.9) 29 (12.5) 0.889
History of malignant disease 38 (16.4) 38 (16.4) 1.000
Dysthyroidism 18 (7.8) 17 (7.3) 0.860
Any dependency level 68 (29.3) 72 (31.0) 0.686
Concomitant treatment at admission, n (%)
Beta-blockers 103 (44.4) 124 (53.4) 0.042
ACEi/ARBs 130 (56.0) 123 (53.0) 0.703
Antiplatelet therapy 24 (10.3) 27 (11.6) 0.656
Laboratory parameters at admission
Creatinine (mg/dL), median (IQR) 1.19 (0.87–1.56) 1.13 (0.87–1.56) 0.628
Hemoglobin (g/dL), median (IQR) 13.0 (12.0–14.0) 13.0 (11.0–14.0) 0.853
Platelet count (×109/L), median (IQR) 178.0 (138.0–244.8) 176.0 (134.0–233.0) 0.432
Elevated D-dimer, n (%) 121 (52.2) 118 (50.9) 0.954
Elevated procalcitonin, n (%) 45 (19.4) 40 (17.2) 0.795
Elevated C-reactive protein, n (%) 209 (90.1) 210 (90.5) 0.984
Elevated troponins, n (%) 35 (15.1) 31 (13.4) 0.711
Elevated transaminases, n (%) 86 (37.1) 66 (28.4) 0.138
Elevated ferritin, n (%) 75 (32.3) 61 (26.3) 0.344
Elevated lactate dehydrogenase, n (%) 167 (72.0) 150 (64.7) 0.182
ACEi, angiotensin-converting enzyme inhibitors; ARBs, angiotensin II receptor blockers; IQR, interquartile range; TIA, transient ischemic attack; COPD/
SAHS, chronic obstructive pulmonary disease/sleep apnea-hypopnea syndrome. ∗ Coronary artery disease and/or peripheral artery disease.

100 hospitalized COVID-19 patients [14]. Patients who received

high-intensity prophylactic anticoagulation have a down-
Event-Free Survival (%)

90
80 trend in D-dimer levels and improved 30-day mortality [15].
70 Indeed, a cross-sectional analysis showed that anti-
60 coagulation use was associated with delayed death, both at
50 prophylactic (HR 0.29, 95% CI 0.15–0.58; p < 0.001) and
40 therapeutic doses (HR 0.15, 95% CI 0.07–0.32; p < 0.001),
30
Log-rank test p-value = 0.225 compared with no anticoagulation [16]. In contrast, one
20
retrospective analysis of hospitalized COVID-19 patients
0 10 20 30 40 50 60 70 suggested that therapeutic anticoagulation provided no
Time (days) mortality benefit over thromboprophylaxis, independently
Number at risk
Prior DOAC use 232 107 50 28 14 6 3 1
of comorbidities or disease severity, and more adverse events
Prior VKA use 232 128 75 50 33 22 12 9
were observed with therapeutic anticoagulation [17]. On the
other hand, a large cohort study simulating an intention-to-
Prior VKA use treat clinical trial analyzed the effect on mortality of anti-
Prior DOAC use coagulation therapy chosen in the first 48 hours of hospi-
Figure 2: Comparison of survival curves between patients on prior talization showing that patients with moderate or severe
VKAs or DOACs. Solid line, prior VKA use; dashed line, prior illness benefited from anticoagulation and that apixaban had
DOAC use. a similar efficacy to enoxaparin in decreasing mortality
amongst these patients [18]. Another study showed that
Anticoagulation in the context of COVID-19 has been hospitalized COVID-19 patients suffered from more
widely debated, with some studies showing that prophylactic bleeding events in those on low-molecular-weight heparin
and therapeutic anticoagulation might reduce mortality in (LMWH) compared to DOACs, and DOAC use may be
International Journal of Clinical Practice 7

associated with better survival and lower invasive respiratory patients treated or not with antiplatelets or anticoagulants
support rate compared to LMWH [19]. Given such con- preadmission [39]. Likewise, anticoagulant use pre-COVID-
tradictory observations, there are a number of studies and 19 diagnosis was not associated with a decreased risk for all-
clinical trials with the aim to assess the role of antith- cause mortality, mechanical ventilation, or hospital ad-
rombotic therapy on mortality and thromboembolic events mission in a study from the New York City health system,
[20–26]. suggesting that previous anticoagulant use did not protect
OAC management in the setting of the COVID-19 against development of severe COVID-19 [40]. Also, our
pandemic is even more complex. VKAs have the limitation preliminary analysis of the HOPE COVID-19 Registry ob-
of routine monitoring and dose adjusting for maintaining served a significantly lower survival and higher mortality
good quality of anticoagulation. One study demonstrated a risk in COVID-19 patients on OAC therapy at hospital
significant increase in high INR results during the COVID- admission compared to patients without prior OAC at
19 pandemic, the majority of them after the introduction of a admission [9]. More recently, a nationwide register-based
lockdown [27]. In addition, patients on VKA hospitalized cohort study in Sweden demonstrated that ongoing DOAC
with SARS-CoV-2 showed greater instability of PT INR due use at the time of SARS-CoV-2 infection was not associated
to the inflammatory state and the interactions with nu- with reduced risk of COVID-19 hospitalization or the
merous drugs. On the other hand, DOACs avoid some of the composite of ICU admission or death due to COVID-19,
VKA limitations, but DOAC-treated patients have an in- indicating that the evidence for DOACs in this context is
crease in DOAC plasma levels when treated with antiviral controversial [41].
drugs for COVID-19 [28]. For these reasons, some groups Our results in the present study confirm our previous
have suggested replacing OAC with parenteral heparin observation about the higher risk of mortality in COVID-19
during hospitalization to avoid the risk of over/under patients with OAC therapy before hospital admission. Of
treatment [29, 30]. Nevertheless, other authors suggested note, our analysis is balanced by PSM, and there were no
that the indications for antiplatelet/anticoagulant use differences regarding admission to the ICU in patients on
(prevention, prophylaxis, and therapy) should be guided by prior and no prior OAC. However, not only mortality was
the clinical context and the COVID-19 severity and not increased in patients with prior OAC therapy but also other
based on a systematic change per protocol in all patients clinical outcomes. Despite an appropriate PSM adjusting for
[31, 32]. comorbidities, postadmission serum creatinine as a marker
Nevertheless, most of the evidence focused on hospi- of renal function (and injury) and troponins as markers of
talized patients, whereas the potential effect of chronic myocardial damage were higher in these patients, thereby
antithrombotic therapies in COVID-19 progression and showing increased rates of heart failure and renal failure
prognosis remains uncertain. The pathophysiology under- during hospitalization. This reinforces the hypothesis that
lying the prothrombotic state elicited by SARS-CoV-2 OAC-treated patients are particularly vulnerable and still
outlines possible protective mechanisms of antithrombotic have an inherent proinflammatory state.
therapy for this viral disease. In particular, aspirin and FXa
inhibitors have been postulated as potential prophylactic
and therapeutic treatment for high-risk patients with 4.1. Limitations. We should acknowledge some limitations
COVID-19 [31, 33]. Unsurprisingly, ongoing clinical trials in relation to this study. First, the constraints of an obser-
are comparing the effectiveness and safety of apixaban, vational registry study of this design need to be considered.
aspirin, and rivaroxaban versus heparin, placebo, and other Second, the HOPE Registry only included patients from the
therapies on progression, arterial, and venous thrombo- first wave of the pandemic, and therefore, our results
embolic events and mortality in patients with COVID-19 probably require further investigation during the subsequent
not yet admitted to hospital [34–36]. waves. A bias inherent in the first wave neither can be
To date, data in this particular context are scarce and excluded, given that hospitalization services throughout the
limited, with positive, negative, and neutral results. One world were overwhelmed. We also recognize that including
small study in an Italian cohort of elderly patients with several different indications for OAC may hinder and dis-
COVID-19 concluded that chronic DOAC intake was an sipate the specific effect that each indication has, since
independent parameter associated with a decreased mor- patients presented different risk profiles.
tality risk (HR 0.38, 95% CI 0.17–0.58; p � 0.010) [8]. In addition, the indication for OAC as a whole may have
Similarly, another study in Italy showed that elderly patients some influence on the risk of outcomes, but comparing
with COVID-19 on chronic OAC treatment for atrial fi- patients with prior OAC and no OAC was actually our aim,
brillation had lower all-cause mortality rate ratio compared so we cannot adjust for specific indications for OAC but only
to their PSM non-anticoagulated counterpart [37]. However, for demographics data and other comorbidities at baseline.
Sivaloganathan et al. demonstrated that patients taking The absence of INR determinations (and therefore the time
antithrombotic therapy (anticoagulant or antiplatelet in therapeutic range (TTR)) in VKA-treated patients is also a
agents) at the time of infection with COVID-19 did not have limitation since the efficacy and safety of VKA depend on the
a significantly different mortality risk to those patients not quality of anticoagulant control, as reflected by the average
taking these drugs [38]. Another study showed no difference TTR of INRs 2.0-3.0, and therefore may be related to the risk
in the risk of acute respiratory distress syndrome at ad- of worse outcomes. In addition, the type of DOAC was
mission or death during hospitalization between COVID-19 unknown in some cases, and this prevented us for analyzing
8 International Journal of Clinical Practice

drug types as separate. Finally, although this cohort was Supplementary Materials
collected in a prospective manner, the results reported in this
study are based on a post hoc analysis and should be On behalf of HOPE COVID-19 investigators, full list of
regarded as hypothesis-generating. investigators is shown in the Supplementary Material
(HOPE participating hospitals, principal investigators,
5. Conclusion HOPE participating hospitals, coprincipal investigators,
scientific committee, and list of collaborators). Sup-
Hospitalized COVID-19 patients on prior OAC therapy had plementary Table 1. Clinical outcomes during hospi-
a higher risk of mortality and worse clinical outcomes talization after propensity score matching in patients on
compared to patients without prior OAC therapy, even after prior oral anticoagulation therapy. (Supplementary
adjusting for comorbidities using PSM. There were no Materials)
differences in clinical outcomes in patients previously taking
VKAs versus DOACs. References
Data Availability [1] F. Violi, D. Pastori, R. Cangemi, P. Pignatelli, and L. Loffredo,
“Hypercoagulation and antithrombotic treatment in coro-
The data used to support this study cannot be shared for navirus 2019: a new challenge,” Thrombosis & Haemostasis,
ethical/privacy reasons. vol. 120, no. 6, pp. 949–956, 2020.
[2] F. Del Sole, A. Farcomeni, L. Loffredo et al., “Features of
severe COVID-19: a systematic review and meta-analysis,”
Disclosure European Journal of Clinical Investigation, vol. 50, no. 10,
Article ID e13378, 2020.
This nonprofit institution had no role in the study design; in [3] G. Piazza, U. Campia, S. Hurwitz et al., “Registry of arterial
the collection, analysis, and interpretation of data; in the and venous thromboembolic complications in patients with
writing of the report; nor in the decision to submit the article COVID-19,” Journal of the American College of Cardiology,
for publication. vol. 76, no. 18, pp. 2060–2072, 2020.
[4] B. Bikdeli, M. V. Madhavan, D. Jimenez et al., “COVID-19
Conflicts of Interest and thrombotic or thromboembolic disease: implications for
prevention, antithrombotic therapy, and follow-up,” Journal
JMR-C has received an unrestricted educational grant from of the American College of Cardiology, vol. 75, no. 23,
Bristol-Myers Squibb-Pfizer Alliance (BMS protocol num- pp. 2950–2973, 2020.
ber: CV185-805) for this study. GYHL is a consultant and [5] R. D. McBane, V. D. Torres Roldan, A. S. Niven et al.,
speaker for BMS/Pfizer, Boehringer Ingelheim, and Daiichi- “Anticoagulation in COVID-19: a systematic review, meta-
Sankyo. Other authors declare that they have no conflicts of analysis, and rapid guidance from mayo clinic,” Mayo Clinic
interest. Proceedings, vol. 95, no. 11, pp. 2467–2486, 2020.
[6] I. Paranjpe, V. Fuster, A. Lala et al., “Association of treatment
dose anticoagulation with in-hospital survival among hos-
Authors’ Contributions pitalized patients with COVID-19,” Journal of the American
College of Cardiology, vol. 76, no. 1, pp. 122–124, 2020.
JMR-C performed statistical analyses and drafted the study. [7] N. Tang, H. Bai, X. Chen, J. Gong, D. Li, and Z. Sun, “An-
IJN-G, GYHL, and FM supervised the research and draft the ticoagulant treatment is associated with decreased mortality
article. AU, MCV-L, AG, AFC-M, BAG, EA, JFG-P, CC, BC, in severe coronavirus disease 2019 patients with coagulop-
GF, IF-R, JS-C, JH, MGA, MP, RR, EC, VMB-M, SR-R, FS, athy,” Journal of Thrombosis and Haemostasis, vol. 18, no. 5,
RB, LS, IE-B, ALM, AF-O, VE, and CM collaborated in data pp. 1094–1099, 2020.
collection and made a critical revision of the article. IJN-G, is [8] R. Rossi, F. Coppi, M. Talarico, and G. Boriani, “Protective
the guarantor of the article and takes responsibility for the role of chronic treatment with direct oral anticoagulants in
integrity of the work as a whole, from inception to published elderly patients affected by interstitial pneumonia in COVID-
article. 19 era,” European Journal of Internal Medicine, vol. 77,
pp. 158–160, 2020.
[9] J. M. Rivera-Caravaca, I. J. Núñez-Gil, D. Vivas et al., “Clinical
Acknowledgments profile and prognosis in patients on oral anticoagulation
The authors thank Cardiovascular Excellence SL for their before admission for COVID-19,” European Journal of
Clinical Investigation, vol. 51, no. 1, Article ID e13436, 2020.
essential support in the database and registry web page, all
[10] I. J. Núñez-Gil, V. Estrada, C. Fernández-Pérez et al., “Health
HOPE COVID-19 researchers (nonconditional grant outcome predictive evaluation for COVID-19 international
(Fundación Interhospitalaria para la investigación Cardio- registry (HOPE COVID-19), rationale and design,” Con-
vascular, FIC. Madrid, Spain)). This work was also supported temporary Clinical Trials Communications, vol. 20, Article ID
by the Spanish Ministry of Economy, Industry, and Com- 100654, 2020.
petitiveness, through the Instituto de Salud Carlos III after [11] International COVID-19 Clinical Evaluation Registry and
independent peer review (PI21/00607 cofinanced by the (HOPE COVID-19), “ClinicalTrials.gov Identifier:
European Regional Development Fund and group CB16/11/ NCT04334291,” 2020, https://clinicaltrials.gov/ct2/show/
00385 from CIBERCV). NCT04334291.
International Journal of Clinical Practice 9

[12] P. C. Austin, “Balance diagnostics for comparing the distri- respiratory syndrome patients treated with antiviral agents:
bution of baseline covariates between treatment groups in the cremona experience,” Journal of Thrombosis and Hae-
propensity-score matched samples,” Statistics in Medicine, mostasis, vol. 18, no. 6, pp. 1320–1323, 2020.
vol. 28, no. 25, pp. 3083–3107, 2009. [29] S. Testa, O. Paoletti, M. Giorgi-Pierfranceschi, and A. Pan,
[13] P. C. Austin, “Optimal caliper widths for propensity-score “Switch from oral anticoagulants to parenteral heparin in
matching when estimating differences in means and differ- SARS-CoV-2 hospitalized patients,” Internal and Emergency
ences in proportions in observational studies,” Pharmaceu- Medicine, vol. 15, no. 5, pp. 751–753, 2020.
tical Statistics, vol. 10, no. 2, pp. 150–161, 2011. [30] G. T. Gerotziafas, P. Van Dreden, M. P. Colgan et al.,
[14] A. M. Kamel, M. Sobhy, N. Magdy, N. Sabry, and S. Farid, “Guidance for the management of patients with vascular
“Anticoagulation outcomes in hospitalized Covid-19 patients: disease or cardiovascular risk factors and COVID-19: position
a systematic review and meta-analysis of case-control and paper from VAS-European independent foundation in
cohort studies,” Reviews in Medical Virology, vol. 31, no. 3, angiology/vascular medicine,” Sang Thrombose Vaisseaux,
Article ID e2180, 2020. vol. 32, no. 6, pp. 241–259, 2020.
[15] A. Hsu, Y. Liu, A. S. Zayac, A. J. Olszewski, and J. L. Reagan, [31] C. Godino, A. Scotti, N. Maugeri et al., “Antithrombotic
“Intensity of anticoagulation and survival in patients hospi- therapy in patients with COVID-19?-rationale and evidence,”
talized with COVID-19 pneumonia,” Thrombosis Research, International Journal of Cardiology, vol. 324, pp. 261–266,
vol. 196, pp. 375–378, 2020. 2020.
[16] F. Ionescu, G. Grasso-Knight, E. Castillo et al., “Therapeutic [32] D. Vivas, V. Roldán, M. A. Esteve-Pastor et al., “Recom-
anticoagulation delays death in COVID-19 patients: cross- mendations on antithrombotic treatment during the COVID-
sectional analysis of a prospective cohort,” TH Open, vol. 04, 19 pandemic. position statement of the working group on
no. 03, pp. e263–e270, 2020. cardiovascular thrombosis of the Spanish society of cardiol-
[17] L. Lynn, J. A. Reyes, K. Hawkins et al., “The effect of anti- ogy,” Revista Espanola de Cardiologia, vol. 73, no. 9,
coagulation on clinical outcomes in novel Coronavirus pp. 749–757, 2020.
(COVID-19) pneumonia in a USA cohort,” Thrombosis Re- [33] G. H. Frydman, M. B. Streiff, J. M. Connors, and G. Piazza,
search, vol. 197, pp. 65–68, 2021. “The potential role of coagulation factor Xa in the patho-
[18] H. H. Billett, M. Reyes-Gil, J. Szymanski et al., “Anti- physiology of COVID-19: a role for anticoagulants as mul-
coagulation in COVID-19: effect of enoxaparin, heparin, and timodal therapeutic agents,” TH Open, vol. 4, no. 4,
apixaban on mortality,” Thrombosis and haemostasis, vol. 120, pp. e288–e299, 2020.
no. 12, pp. 1691–1699, 2020. [34] F. Sciurba, “COVID-19 positive outpatient thrombosis pre-
[19] G. N. Nadkarni, A. Lala, E. Bagiella et al., “Anticoagulation, vention in adults aged 40-80 2020,” 2020, https://clinicaltrials.
bleeding, mortality, and pathology in hospitalized patients gov/ct2/show/NCT04498273.
with COVID-19,” Journal of the American College of Cardi- [35] R. Whitlock, E. Belley-Cote, and J. Eikelboom, “Anti-coro-
ology, vol. 76, no. 16, pp. 1815–1826, 2020. navirus therapies to prevent progression of coronavirus
[20] Bill-&-Melinda-Gates-Medical-Research-Institute, “A trial to disease 2019 (COVID-19) trial (ACTCOVID19) 2020,” 2020,
evaluate safety and efficacy of rivaroxaban (COVID-19),” https://clinicaltrials.gov/ct2/show/NCT04324463.
2020, https://clinicaltrials.gov/ct2/show/NCT04504032. [36] Janssen-Research-&-Development-L, “A study of rivaroxaban
[21] U. Landmesser, “Effect of anticoagulation therapy on clinical to reduce the risk of major venous and arterial thrombotic
outcomes in COVID-19 (COVID-prevent),” 2020, https:// events, hospitalization and death in medically ill outpatients
clinicaltrials.gov/ct2/show/NCT04416048. with acute, symptomatic coronavirus disease 2019 (COVID-
[22] Fundació-Institut-de-Recerca-de-l’Hospital-de-la-Santa- 19) infection (PREVENT-HD),” 2020, https://clinicaltrials.
Creu-i-Sant-Pau, “Evolution of COVID-19 in anticoagulated gov/ct2/show/NCT04508023.
or antiaggregated patients (CORONA study) (CORONA),” [37] G. Denas, N. Gennaro, E. Ferroni et al., “Reduction in all-
2020, https://clinicaltrials.gov/ct2/show/NCT04518735. cause mortality in COVID-19 patients on chronic oral
[23] P. Horby, “Randomised evaluation of COVID-19 therapy anticoagulation: a population-based propensity score
(RECOVERY),” 2020, https://clinicaltrials.gov/ct2/show/ matched study,” International Journal of Cardiology, vol. 329,
NCT04381936. pp. 266–269, 2020.
[24] J. Hochman, “Anti-thrombotics for adults hospitalized with [38] H. Sivaloganathan, E. E. Ladikou, and T. Chevassut, “COVID-
19 mortality in patients on anticoagulants and antiplatelet
COVID-19 (ACTIV-4),” 2020, https://clinicaltrials.gov/ct2/
agents,” British Journal of Haematology, vol. 190, no. 4,
show/NCT04505774.
pp. e192–e5, 2020.
[25] S. Webb, C. McArthur, M. Bonten, L. Derde, M. Marshall, and
[39] V. Russo, M. Di Maio, E. Attena et al., “Clinical impact of pre-
A. Derek, “Randomized, embedded, multifactorial adaptive
admission antithrombotic therapy in hospitalized patients
platform trial for community-acquired pneumonia (REMAP-
with COVID-19: a multicenter observational study,” Phar-
CAP),” 2020, https://clinicaltrials.gov/ct2/show/
macological Research, vol. 159, Article ID 104965, 2020.
NCT02735707.
[40] D. Tremblay, M. van Gerwen, M. Alsen et al., “Impact of
[26] P. R. Lawler, E. C. Goligher, and R. Zarychanski, “Antith-
anticoagulation prior to COVID-19 infection: a propensity
rombotic therapy to ameliorate complications of COVID-19
score-matched cohort study,” Blood, vol. 136, no. 1,
(ATTACC),” 2020, https://clinicaltrials.gov/ct2/show/
pp. 144–147, 2020.
NCT04372589.
[41] B. Flam, V. Wintzell, J. F. Ludvigsson, J. Mårtensson, and
[27] V. Speed, R. K. Patel, R. Byrne, L. N. Roberts, and R. Arya, “A
B. Pasternak, “Direct oral anticoagulant use and risk of severe
perfect storm: root cause analysis of supra-therapeutic anti-
COVID-19,” Journal of Internal Medicine, vol. 289, no. 3,
coagulation with vitamin K antagonists during the COVID-19
pp. 411–419, 2020.
pandemic,” Thrombosis Research, vol. 192, pp. 73-74, 2020.
[28] S. Testa, P. Prandoni, O. Paoletti et al., “Direct oral antico-
agulant plasma levels’ striking increase in severe COVID-19