June 1st, 2021
Aspirin and death in Covid-19
A systematic review and meta-analysis
Nicholas Moore MD, PhD, Nicolas Thurin PharmD, PhD, Pauline Bosco-Levy,
MD, PhD, Patrick Blin MD, Cecile Droz-Perroteau PhD
Bordeaux PharmacoEpi, Univ. Bordeaux, Inserm CIC1401, Bordeaux, France
Correspondance to Nicholas Moore,
Université de Bordeaux
Rue Leo Saignat, 33076 Bordeaux Cedex, France
Nicholas.moore@u-bordeaux.fr
Word Count:
Key points:
Question: Because COVID-19 is associated with thrombotic events, might
aspirin be beneficial?
Findings: from a meta-analysis of eight observational studies and one
open randomized trial providing adjusted or matched results of exposure
to aspirin in COVID-19, the random-effects odds ratio for death in
aspirin users compared to non-users was 0.63, 95% confidence interval
[0.40-0.99], I2 94%. Sensitivity analyses were
consistent.
Conclusion: Exposure to aspirin was associated with a lower risk of
dying in COVID-19 patients. This would warrant further randomized
clinical trials.
Abstract:
: Thrombotic events are common during COVID-19 infection. Aspirin might
be beneficial.
Objective: Systematic review and meta-analysis of deaths in users and
non-users of aspirin.
Data sources: Pubmed Medline, Google scholar, Clinicaltrials.gov,
Cochrane, to June 8, 2021,
Study selection: Studies providing adjusted or matched evaluation of
association of exposure to aspirin and death in COVID-19 patients were
included.
Data extraction and synthesis: Data were used as published, as Odds
ratio, hazard ratio or relative risks and 95% CI from which log(OR) and
SE were recalculated. These were entered in an inverse variance odds
ratios random-effects model, using RevMan 5.4 (the Cochrane
Collaboration).
Main outcomes and measure: The prespecified outcome studied was death.
Results: Nine studies (8 observational, one interventional) included
14989 patients exposed to aspirin and 15857 unexposed. Overall Odds
Ratio of death in aspirin exposed patients in a random effects model was
0.63, 95% confidence interval [0.40-0.99], I294%. Using a fixed-effect model did not change much the result (0.76
[0.71-0.81], removing the Recovery trial (OR 0.43 [0.38-0.49],
I271%, or the two largest studies (0.66
[0.47-0.93], I2 38%) reduced heterogeneity
without materially altering the results. The funnel plot showed no
evident publication bias
Conclusion: this meta-analysis suggests that the use of aspirin may be
associated with a lower risk of death in COVID-19. Considering the
results of the Recovery Study, it would appear preferable to continue
aspirin in patients who have a non-covid indication, but possibly
useless to add it if they don’t.
Introduction
COVID-19, the clinical manifestation of SARS-COV2 infection, is often
accompanied by thrombotic events that contribute to the severity of the
symptoms, suggesting a possible benefit from aspirin. (1-3) A previous
meta-analysis had found no effect of aspirin on mortality in COVID-19,
but included only 3 observational studies, and the statistical analysis
did not take into account adjusted results,(4) whereas aspirin users
would be expected to be older and with more cardiovascular diseases
associated with severe COVID-19 outcomes. Another more recent
meta-analysis included 6 of the studies we had already selected, and
found a decreased risk of death in patients prescribed aspirin, with an
OR of 0.46, 95%CI[0.35-0.61]. (5) We therefore updated the
meta-analyses, including more studies, and taking into account only
adjusted results.
Problem definition: What is the effect of exposure to aspirin on
mortality in COVID-19?
Hypothesis statement: Aspirin might have a favorable effect in COVID-19
Methods
Study outcome of interest was death, as reported.
Exposure of interest was aspirin, at any time or dose.
Data sources: Pubmed Medline, Clinicaltrials.gov, and Google scholar
were queried by the authors for studies mentioning COVID-19 in the
title, and aspirin in any field. We also scanned the references of the
retrieved papers, of two meta-analyses (4, 5) and of other papers
discussing the role of aspirin in the treatment of COVID-19 (1-3, 6-8).
Medline was queried with EndNote 20 (Clarivate analytics).
Study selection: Studies were first selected on their title and
abstracts indicating the presence of study results. Full text was
retrieved for all selected studies and analyzed for studies with clearly
described methodology, providing numbers of patients receiving aspirin
and controls, the data analysis method, and numerical results including
adjusted and/or propensity score matched results, with Odds ratios,
hazard ratios or relative risks, and confidence intervals, from which
log(OR) and SE were recomputed. Because aspirin use is indicated in
cardiovascular disease that increase mortality, only studies reporting
adjusted or matched analyses including age and concomitant diseases were
included. Authors were contacted if there were discrepant results.
Studies concerning children were not included, because of negligible
mortality. (9)
Exposure was aspirin at any time or dose, prior to inclusion in the
study and present at inclusion, considering aspirin long-lasting action
on platelets. Outcome was death in aspirin users and non-users. Outcomes
were used as is from the publications without attempting
reascertainment.
Statistical analysis: These data were input into RevMan 5.4 ((Review
Manager (RevMan), Version 5.4), The Cochrane Collaboration, 2020), and
analyzed using inverse variance odds ratios in a random effects model.
Heterogeneity was assessed with I2 as provided by the
analysis program. Sensitivity analyses including testing a fixed effects
model and repeating the analysis after excluding the largest studies so
that heterogeneity was 0%. A funnel plot was used to check for possible
publication bias.
Results.
On June 6thth, 2021, of 98155 papers in Medline
mentioning COVID-19 in the title, 87 mentioned aspirin in the title or
other fields. Death or mortality were mentioned in respectively 14 ad 23
of these papers.
In Clinicaltrials.gov 20 studies were identified as involving aspirin
and COVID-19, none of which provided any usable data: One had been
withdrawn, 2 were not yet recruiting, 11 were still recruiting, one was
active not recruiting, Five studies are indicated as completed, but
provided no usable information. (NCT04425863, NCT04368370 (10),
NCT04757792, NCT04518735, NCT04492501)
Google Scholar provided one further study (11)
At the time this paper was written, the results of Recovery, a large
randomized open-label study, were provided and added to the analysis.
(12)
Nine studies provided usable data. (Table 1) One paper studying the
effect of a 5-Drug regimen including aspirin was not retained because
the effect of the individual drugs could not be identified. (7)
In total, nine studies were included in this analysis (Table 2),
representing 14989 patients exposed to aspirin and 15857 unexposed.
In 4 studies, users and non-users were matched using propensity scores,
and further adjusted (13-16). In one study the actual number of deaths
in each arm was not given, only the final adjusted odd ratio and
confidence intervals. (14) In Recovery (12), treatment allocation was
random, but the study was open-label and no placebo was used.
Meta-analysis of these studies using a random effect model showed that
the use of aspirin was associated with lower mortality (OR 0.63, 95%CI
[0.40-0.99]) (figure 1), with much heterogeneity
(I2 94%, p=0.00001). With a fixed effect model, the
odds ratio was 0.76[0.71-0.81]. (Figure 2)
Removing Recovery study (12) results in a lower Odds Ratio (0.43
[0,38-0.49]) with still meaningful heterogeneity
(I2=71%) (figure 3) Removing the two largest studies,
Recovery and Osborne et al (15) did not meaningfully affect the result
(OR 0.66 [0.47-0.93]) but reduced heterogeneity
(I2=38%, p=0.14).(Figure 4)
The funnel plot found no obvious publication bias (Figure 5).
Discussion:
We found in this meta-analysis of observational studies and one open
randomized trial, (12), that exposure to aspirin was associated with a
lower risk of dying than non-use. The sensitivity analyses did not alter
the results nor did the inclusion or not of the Recovery trial. This was
not unexpected, in a disease that seems to have a strong thrombotic
component. The result is consistent with the expected effects of
aspirin. (2, 17) It appears relatively robust with about 15000 patients
in either treatment arm.
A previous meta-analysis (4) included only 3 studies, (11, 18, 19) and
used the raw estimates of odds ratios without taking adjustments or
matching into account. The odds ratio reported for the meta-analysis was
1.14 [0.84-1.50]
The other meta-analysis (20) included adjusted values from 6 studies
(11, 13-15, 19, 21) It found an Odds ratio of 0.46 [0.35-0.61],
lower than the one we found, probably because of the additional studies
we included (12, 16, 18)
This meta-analysis included mostly observational studies, so that
despite the intrinsic quality of the individual studies, the result is
at best tentative and will need to be confirmed by proper clinical
trials, a number of which are under way, as indicated in the
clinicatrials.gov registry, but not completed yet.
Because aspirin is usually given to prevent cardiovascular disorders,
patients routinely treated with aspirin will be at higher risk of dying.
(4) For this reason we only considered studies that provided adjusted or
propensity score matched studies, and the one open-label randomized
study, Recovery.
The open-label platform trial Recovery (12) has prepublished that random
attribution of aspirin did not improve the outcomes of patients with
COVID-19, in patients with no previous use of aspirin and no formal
indication or contra-indication for aspirin. Without Recovery the result
of the meta-analysis show a lower event rate in patients using aspirin,
including Recovery in the analysis did change the results, the global OR
going from 0.43 [0.38-0.49-0.89] to 0.76 [0.71-0.81] for the
fixed effects model. Of the many reasons that might explain the apparent
lack of effect of aspirin in Recovery, the main may be that patients
randomized to aspirin in Recovery were not currently on aspirin and had
no indication or contra-indication for the use of aspirin, which is in
contrast with the observational studies, where there most probably was a
non-covid indication for aspirin. This may restrict the conclusions of
both the trial and the meta-analysis: patients diagnosed with COVID not
currently on aspirin may not benefit from adding aspirin to their
treatment, and may have a higher risk of bleeding. Patients on aspirin
at the time of diagnosis appear to have a lower risk of dying than
patients who were not on aspirin, taking into account various
confounders. It would appear preferable not to stop aspirin in patients
who have a non-covid indication, but possibly useless to add it if they
don’t.
Study limitations.
In addition to the limitations inherent in observational studies, or in
open-label trials noted above, only English language publications were
found and analyzed. Considering the topic and its timeliness, it is
unlikely that non-English language publications would alter the results.
There may be other English-language studies that were not identified,
but it is unlikely that they would modify the results in a relevant
manner. The funnel plot does not indicate publication bias. Studies in
children were not included, because of negligible mortality (9)
Conclusion
Aspirin is a widely used antiplatelet drug, and its risks are well
known, mainly an increased risk of upper GI bleeding consistent with its
irreversible non-selective inhibition of Cox-1, resulting in the
anti-platelet activity. From our results and the Recovery trial, it
would appear preferable not to stop aspirin in patients who have a
non-covid indication, but possibly useless to add it if they don’t.
Ethical considerations
This study did not require ethical approval.
This paper was unfunded.
The authors report performing studies regularly with pharmaceutical
industry, required by regulatory authorities, but none concerning
COVID-19 and aspirin. They are involved in the surveillance of the
COVID-19 vaccination program in Europe, with the ACCESS project. The
list of these studies can be found atwww.Encepp.eu, in the European PAS
registry, and atwww.pharmacoepi.eu.
References
1. Bianconi V, Violi F, Fallarino F, Pignatelli P, Sahebkar A, Pirro M.
Is Acetylsalicylic Acid a Safe and Potentially Useful Choice for Adult
Patients with COVID-19 ? Drugs. 2020;80(14):1383-96.
2. Coppola S, Chiumello D. Aspirin in COVID-19 related ARDS: an old,
low-cost therapy with a strong rationale. Anesth Analg. 2021.
3. Diaz T, Trachtenberg BH, Abraham SJK, KosagiSharaf R,
Durant-Archibold AA. Aspirin Bioactivity for Prevention of
Cardiovascular Injury in COVID-19. Front Cardiovasc Med. 2020;7:562708.
4. Salah HM, Mehta JL. Meta-Analysis of the Effect of Aspirin on
Mortality in COVID-19. Am J Cardiol. 2021;142:158-9.
5. Martha JW, Pranata R, Lim MA, Wibowo A, Akbar MR. Active prescription
of low-dose aspirin during or prior to hospitalization and mortality in
COVID-19: A systematic review and meta-analysis of adjusted effect
estimates. Int J Infect Dis. 2021;108:6-12.
6. Canzano P, Brambilla M, Porro B, Cosentino N, Tortorici E, Vicini S,
et al. Platelet and Endothelial Activation as Potential Mechanisms
Behind the Thrombotic Complications of COVID-19 Patients. JACC Basic
Transl Sci. 2021.
7. Kevorkian JP, Lopes A, Sene D, Riveline JP, Vandiedonck C, Feron F,
et al. Oral corticoid, aspirin, anticoagulant, colchicine, and
furosemide to improve the outcome of hospitalized COVID-19 patients -
the COCAA-COLA cohort study. J Infect. 2021.
8. Mohamed-Hussein AAR, Aly KME, Ibrahim MAA. Should aspirin be used for
prophylaxis of COVID-19-induced coagulopathy? Med Hypotheses.
2020;144:109975.
9. Duarte-Salles T, Vizcaya D, Pistillo A, Casajust P, Sena AG, Lai LYH,
et al. Baseline characteristics, management, and outcomes of 55,270
children and adolescents diagnosed with COVID-19 and 1,952,693 with
influenza in France, Germany, Spain, South Korea and the United States:
an international network cohort study. medRxiv. 2020.
10. Viecca M, Radovanovic D, Forleo GB, Santus P. Enhanced platelet
inhibition treatment improves hypoxemia in patients with severe Covid-19
and hypercoagulability. A case control, proof of concept study.
Pharmacol Res. 2020;158:104950.
11. Chow JH, Khanna AK, Kethireddy S, Yamane D, Levine A, Jackson AM, et
al. Aspirin Use is Associated with Decreased Mechanical Ventilation, ICU
Admission, and In-Hospital Mortality in Hospitalized Patients with
COVID-19. Anesth Analg. 2020.
12. Recovery collaborative group. Aspirin in patients admitted to
hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label,
platform trial. MedRxiv. 2021.
13. Liu Q, Huang N, Li A, Zhou Y, Liang L, Song X, et al. Effect of
low-dose aspirin on mortality and viral duration of the hospitalized
adults with COVID-19. Medicine (Baltimore). 2021;100(6):e24544.
14. Meizlish ML, Goshua G, Liu Y, Fine R, Amin K, Chang E, et al.
Intermediate-dose anticoagulation, aspirin, and in-hospital mortality in
COVID-19: A propensity score-matched analysis. Am J Hematol.
2021;96(4):471-9.
15. Osborne TF, Veigulis ZP, Arreola DM, Mahajan SM, Roosli E, Curtin
CM. Association of mortality and aspirin prescription for COVID-19
patients at the Veterans Health Administration. PLoS One.
2021;16(2):e0246825.
16. Sahai A, Bhandari R, Godwin M, McIntyre T, Chung MK, Iskandar JP, et
al. Effect of aspirin on short-term outcomes in hospitalized patients
with COVID-19. Vasc Med. 2021:1358863X211012754.
17. Haque S, Jawed A, Akhter N, Dar SA, Khan F, Mandal RK, et al.
Acetylsalicylic acid (Aspirin): a potent medicine for preventing
COVID-19 deaths caused by thrombosis and pulmonary embolism. Eur Rev Med
Pharmacol Sci. 2020;24(18):9244-5.
18. Alamdari NM, Afaghi S, Rahimi FS, Tarki FE, Tavana S, Zali A, et al.
Mortality Risk Factors among Hospitalized COVID-19 Patients in a Major
Referral Center in Iran. Tohoku J Exp Med. 2020;252(1):73-84.
19. Yuan S, Chen P, Li H, Chen C, Wang F, Wang DW. Mortality and
pre-hospitalization use of low-dose aspirin in COVID-19 patients with
coronary artery disease. J Cell Mol Med. 2021;25(2):1263-73.
20. Martha JW, Pranata R, Lim MA, Wibowo A, Akbar MR. Active
Prescription of Low-dose Aspirin During or Prior to Hospitalization and
Mortality in COVID-19 - A Systematic Review and Meta-analysis of
Adjusted Effect Estimates. Int J Infect Dis. 2021.
21. Merzon E, Green I, Vinker S, Golan-Cohen A, Gorohovski A, Avramovich
E, et al. The use of aspirin for primary prevention of cardiovascular
disease is associated with a lower likelihood of COVID-19 infection.
FEBS J. 2021.
22. Meizlish ML, Goshua G, Liu Y, Fine R, Amin K, Chang E, et al.
Intermediate-dose anticoagulation, aspirin, and in-hospital mortality in
COVID-19: a propensity score-matched analysis. medRxiv. 2021.
Legends to the figures
Figure 1 meta-analysis of aspirin and COVID-19 associated deaths;
Inverse variance with random effects model.
Figure 2 meta-analysis of aspirin and COVID-19 associated deaths;
Inverse variance with fixed effects model.
Figure 3 meta-analysis of aspirin in COVID-19 associated deaths,
excluding the Recovery trial. Inverse variance with random effects
model.
Figure 4: funnel plot for meta-analysis of studies of COVID-19 deaths in
patients exposed to aspirin.
Table 1 Search results on June 7, 2021
COVID in TITLE 98155
AND
Aspirin (Title) 28
Aspirin (all fields) 87
AND (Death or mortality) 12
Studies with abstract and results 8
Table 2
Study first author