Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Chronic use of non-steroidal anti-inflammatory drugs (NSAIDs) or acetaminophen and relationship with mortality among United States Veterans after testing positive for COVID-19

  • Heather M. Campbell ,

    Contributed equally to this work with: Heather M. Campbell, Allison E. Murata

    Roles Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Cooperative Studies Program, Clinical Research Pharmacy Coordinating Center, US Department of Veterans Affairs, Albuquerque, New Mexico, United States of America

  • Allison E. Murata ,

    Contributed equally to this work with: Heather M. Campbell, Allison E. Murata

    Roles Data curation, Investigation, Resources, Software, Validation, Writing – review & editing

    Affiliation Cooperative Studies Program, Clinical Research Pharmacy Coordinating Center, US Department of Veterans Affairs, Albuquerque, New Mexico, United States of America

  • Todd A. Conner,

    Roles Funding acquisition, Resources, Writing – review & editing

    Affiliation Cooperative Studies Program, Clinical Research Pharmacy Coordinating Center, US Department of Veterans Affairs, Albuquerque, New Mexico, United States of America

  • Greg Fotieo

    Roles Conceptualization, Methodology, Project administration, Resources, Supervision, Writing – review & editing

    Affiliation New Mexico VA Healthcare System, US Department of Veterans Affairs, Albuquerque, New Mexico, United States of America


Non-steroidal anti-inflammatory drugs (NSAIDs) and acetaminophen are among the most-frequently used medications. Although these medications have different mechanisms of action, they have similar indications and treatment duration has been positively correlated with cardiovascular risk although the degree of risk varies by medication. Our objective was to study treatment effects of chronic use of individual NSAID medications and acetaminophen on all-cause mortality among patients who tested positive for COVID-19 while accounting for adherence. We used the VA national datasets in this retrospective cohort study to differentiate between sporadic and chronic medication use: sporadic users filled an NSAID within the last year, but not recently or regularly. Using established and possible risk factors for severe COVID-19, we used propensity scores analysis to adjust for differences in baseline characteristics between treatment groups. Then, we used multivariate logistic regression incorporating inverse propensity score weighting to assess mortality. The cohort consisted of 28,856 patients. Chronic use of aspirin, ibuprofen, naproxen, meloxicam, celecoxib, diclofenac or acetaminophen was not associated with significant differences in mortality at 30 days (OR = 0.98, 95% CI: 0.95–1.00; OR = 0.99, 95% CI: 0.98–1.00; OR = 1.00, 95% CI: 0.98–1.01; OR = 0.99, 95% CI: 0.98–1.00; OR = 1.00, 95% CI: 0.98–1.01; OR = 0.99, 95% CI: 0.97–1.01; and OR = 1.00, 95% CI: 0.99–1.02, respectively) nor at 60 days (OR = 0.97, 95% CI: 0.95–1.00; OR = 1.00, 95% CI: 0.99–1.01; OR = 0.99, 95% CI: 0.98–1.01; OR = 0.99, 95% CI: 0.97–1.00; OR = 0.99, 95% CI: 0.97–1.01; OR = 0.99, 95% CI: 0.97–1.01; and OR = 1.01, 95% CI: 0.99–1.02, respectively). Although the study design cannot determine causality, the study should assure patients as it finds no association between mortality and chronic use of these medications compared with sporadic NSAID use among those infected with COVID-19.


The emergence of the COVID-19 virus presented an immediate challenge to the scientific community. Because of the novel nature of COVID-19, there was no established understanding of risk or protective factors for severity of disease. The scientific community began to search for potential risk and protective factors. NSAIDs were identified as a potential candidate therapeutic class due to their mechanism of action. Authors of an early correspondence suggested ibuprofen’s ability to stimulate production of ACE2, an enzyme by which SARS-CoV-2 binds to its target cell, places patients at increased risk for worse COVID-19 outcomes [1]. Contradictorily, another correspondence suggested ibuprofen might prevent COVID-19 fatalities by reducing cytokine release syndrome found in severely ill COVID-19 patients by reducing interleukin-6 levels [24]. Subsequently, the European Medications Agency stated there was no evidence to suggest ibuprofen or non-steroidal anti-inflammatory drugs (NSAIDs) put patients at risk: rather they recommended that clinicians perform a risk assessment reminding them that paracetamol (i.e., acetaminophen) is the first-line treatment for fever [5]. The U.S. Federal Drug Administration also stated there was no evidence of NSAIDs putting patients at risk, though their communication did not specify acetaminophen as a first-line treatment [6].

According to the Centers for Disease Control and Prevention, history of cardiovascular disease, history of systemic inflammatory disorder, older age, type 2 diabetes mellitus, and smoking are a subset of established risk factors of severe COVID-19; hypertension and type 1 diabetes mellitus are possible risk factors, among others [7, 8]. Unrelated to COVID-19, these factors also increase risk of cardiovascular events with NSAID use [912]. There is further evidence that cardiovascular risk is increased as NSAID exposure is increased and that cardiovascular risk varies by NSAID, potentially due to cyclooxygenase (COX) inhibition selectivity [13]. Similar evidence exists for acetaminophen use and duration of therapy increasing cardiovascular risk [13]. The investigators of a meta-analysis found patients with a higher risk of vascular disease who took high-dose diclofenac or COX-2 inhibitors had higher risk of major cardiovascular events compared to placebo while those taking high-dose naproxen did not [9]. Additionally, patients taking diclofenac, COX-2 inhibitors, or naproxen were at higher risk for major coronary events. Our objective was to study effects of chronic use of individual NSAID medications and acetaminophen on all-cause mortality among patients who tested positive for COVID-19 while accounting for adherence.

Materials and methods

This article follows the Strengthening the Reporting of Observational Studies in Epidemiology guideline. The New Mexico Veterans Affairs Health Care System’s Institutional Review Board approved this study (20-H319/H3234). Consent was not obtained; they approved a waiver as, among other things, the research could not be carried out practically with the number of patients in this retrospective database analysis. We used the national databases at the VA Corporate Data Warehouse (CDW) and the VA COVID-19 Shared Data Resource to extract data for this study. If a Veteran who received VA healthcare services died or reported testing positive outside the VA, this was also captured. The VA CDW has about 94% accuracy capturing date of death; when looking at month and year of death only, this increases to 99% [14].

We incorporated a quasi-experimental study design. We followed patients who tested positive for COVID-19 from March 2, 2020 to December 14, 2020 for 30- and 60- day all-cause mortality from time of diagnosis. December 14, 2020 was the end date to eliminate confounding associated with distribution of the first U.S. COVID-19 vaccine for clinical use. We selected 30- and 60- day mortality to assess robustness of results.

The VA pharmacy database was queried for the year before testing positive for COVID-19 to classify patients into two patterns of use: sporadic NSAID and chronic NSAID or acetaminophen use. Studied NSAIDs included aspirin > 150mg/day, ibuprofen, naproxen, meloxicam, celecoxib, and diclofenac. To help differentiate between sporadic and chronic use, we wanted to ensure sporadic users did not have an NSAID prescription within the 90 days before the 14-day symptomatic window; doing this also aids in NSAID use not being confounded by COVID-19 severity. We used medication possession ratio (MPR) to measure adherence by evaluating fill dates and days supply of individual prescriptions longitudinally to categorize users. A sporadic user was defined as filling a prescription one or more times over the year resulting in a MPR < 0.75. A chronic user was defined as filling one medication amounting to a MPR ≥ 0.75 with a supply on-hand within the 14 days before testing positive. We used documented sporadic NSAID use as the control rather than no documented use to limit confounding by undocumented NSAID consumption; if patients fill prescriptions for NSAIDs in the VA they may be less likely to buy them over-the-counter.

Since low-dose aspirin does not systemically inhibit COX enzymes, the mechanism of action of NSAIDs, patients taking low-dose aspirin were not considered NSAID users unless they otherwise met the criteria for another NSAID. We also compared chronic acetaminophen use to sporadic NSAID use since acetaminophen can be taken for similar indications as NSAIDs. Comorbidities were captured if ICD-10 code documentation existed for the two years before testing positive.

Statistical analysis

All analyses were conducted using StataMP 15 (StataCorp LP, College Station, TX). We included age, sex, race and ethnicity, urban/suburban versus rural living, body mass index, smoking status, cardiovascular disease, systemic inflammatory disorder (defined as rheumatoid arthritis, systemic lupus erythematosus, or spondyloarthritis), hypertension, hyperlipidemia, type 2 diabetes mellitus, cancer, chronic kidney disease, chronic obstructive pulmonary disease, immunocompromised status (categorized as due to solid organ transplant or not), sickle cell disease, asthma, cerebrovascular disease, liver disease, dementia/Alzheimer’s disease, type 1 diabetes mellitus, and number of Elixhauser comorbidities as covariates in the study. Sex, race, and ethnicity were preferentially defined based on patient declaration; if data were missing the algorithm based these categories on majority declared by VA staff. We also incorporated test month and test season to account for the changing clinical management, weather, and population health behaviors during the study timeframe.

We used propensity scores analysis to adjust for differences in baseline characteristics between treatment groups. If confounders across groups become similar after such adjustment researchers can obtain an unbiased estimate of the treatment effect [15]. Through an iterative process of modeling the propensity score, one can attempt to achieve balance. If any covariates remain unbalanced, a second step is used for further adjustment, though bias will still exist [15]. A standardized mean difference > 0.1 is considered important. We used two-stage propensity scores analysis: in the first stage, the regression model predicts the treatment selection; in the second stage, the regression model predicts the outcome by using the inverse of the weight of the propensity score. Covariates related to the outcome should be used in both stages [16]. This method is considered doubly-robust, meaning even if one of the two models is mis-specified the estimator is still consistent. We started by including all covariates mentioned above, then selected the model with the best balance based on established approaches.

Additional analyses.

We stratified analyses of treatment effects for covariates with odds ratios (ORs) ≥ 2.0 to evaluate distinct effects among those having, and not having, the characteristic. Due to differences in their mechanisms of action, we also compared chronic NSAID medication use to chronic acetaminophen use.


Our cohort consisted of 28,856 patients: 20,311 sporadic NSAID users, 6,480 chronic NSAID monotherapy users, and 2,074 chronic acetaminophen users (Fig 1). The mean ± standard deviation age of patients who tested positive for COVID-19 and used NSAIDs was 57.88 ± 15.39 years. These patients were 87.40% male. Whites, Blacks, and American Indians/Alaska Natives comprised 60.22%, 29.83%, and 1.03% of this group, respectively (Table 1). For acetaminophen, age was 67.10 ± 14.36 years, 93.15% were male, and 56.36%, 34.19%, and 1.11% were White, Black, and American Indian/Alaska Native, respectively. At 30 days from testing positive for COVID-19, all-cause mortality was 4.67% for sporadic users, 3.01% for chronic NSAID users, and 8.68% for chronic acetaminophen users. By 60 days, 5.54% of sporadic users, 3.57% of chronic NSAID users, and 10.56% of chronic acetaminophen users had died from any cause. Among chronic NSAID users, 30-day mortality was 7.64% for aspirin, 2.54% for ibuprofen, 2.99% for naproxen, 2.53% for meloxicam, 3.10% for celecoxib, and 2.89% for diclofenac. At 60 days, mortality was 8.83%, 3.25%, 3.26%, 3.06%, 3.36%, and 3.37%, respectively. Across treatment groups, there were substantial differences in baseline characteristics before propensity score weighting. Table 2 reveals unadjusted ORs of the baseline characteristics.

Fig 1. Patients testing positive who had documented NSAID or acetaminophen use.

NSAID, non-steroidal anti-inflammatory drug; MPR, medication possession ratio.

Table 2. Unadjusted odds ratios and p-values associated with 30-day and 60-day mortality.

Our final model comparing chronic to sporadic use included covariates with ORs ≥ 2.0, denoting a moderate association with mortality [17]; the only exception being immunocompromised due to solid organ transplant as the number of patients was too small: three groups each had zero patients. This approach resulted in very similar baseline covariates: all standardized mean differences were ≤ 0.01 (Table 3); we also had good overlap in propensity scores among the treatment groups. When comparing chronic use of aspirin, ibuprofen, naproxen, meloxicam, celecoxib, diclofenac, and acetaminophen to sporadic NSAID use, we found no significant difference in all-cause mortality at 30 or 60 days apart from meloxicam (OR = 0.99 [95% CI: 0.98–1.00] and OR = 0.99 [95% CI: 0.98–1.00], respectively, p = 0.005 for both). See Table 4.

Table 3. Standard mean differences before and after propensity score weighting.

Table 4. Odds ratios of all-cause mortality by treatment group.

Similarly, analyses stratified by age, gender, history of cardiovascular disease, hypertension, type 2 diabetes mellitus, chronic kidney disease, chronic obstructive pulmonary disease, cerebrovascular disease, dementia, or number of comorbidities showed no clinically significant differences though meloxicam was marginally statistically significant at 30 and 60 days. When the comparison group was changed from sporadic NSAID users to chronic acetaminophen users, aspirin and meloxicam were statistically, though not clinically, significant for all-cause mortality at 30 and 60 days.


We found no clinically significant difference comparing chronic use with each of seven medications with sporadic use of an NSAID. This bore out in main and stratified analyses. We also found no clinically significant difference comparing chronic use of individual NSAID medications with chronic acetaminophen use. Our use of MPR rather than simple prescription refill history reflects actual adherence and use patterns more accurately. With our sample size, the analyses had narrow confidence intervals. Due to the standardized mean differences being negligible, we are confident in an unbiased estimate of treatment effect among observed variables with at least a moderate association with all-cause mortality. Further, there was sufficient overlap in propensity scores across treatment groups, also necessary for reduced bias and variance of estimates [18]. Collectively, this study showed that the risk of death in patients who take these medications chronically is no different than patients who sporadically used NSAIDs. This information can also be used by clinicians caring for patients as they make clinical treatment decisions based on their risk assessment. Our attempt to address this was by controlling the month or season in which patients tested positive: with an unadjusted OR = 0.90 and OR = 0.91 for 30-day and 60-day mortality, respectively, month did not meet the criteria for a moderate association to be included in the final propensity score-adjusted model. With season having OR = 0.66 and OR = 0.68 for 30- and 60- day mortality, respectively, it also did not meet the criteria of OR = 0.50 to be included.

At the time of this writing, seventeen studies have focused on potential effects of NSAIDs among COVID-19 patients: three of ibuprofen, one of naproxen, one of celecoxib, and seven of aspirin; the rest kept analysis at the therapeutic class [1935]. Our study adds compelling evidence to the literature with the second-largest sample size to evaluate chronic NSAID use and the third-largest to evaluate any NSAID use among COVID-19 patients: one study assessed NSAID effects on mortality in the rheumatoid arthritis and osteoarthritis population in the United Kingdom, though adherence was assumed [30]; the other focused on aspirin use at time of testing positive [28]. Three of the seventeen have assessed treatment effects of chronic NSAID use, with only one evaluating a medication effect rather than the therapeutic class and none accounting for adherence: baseline characteristics were not comparable or controlled for in two of the four studies [20, 21, 23, 30]. Among these studies, none found a significant difference in mortality. However, believing this about individual NSAIDs without a more granular assessment can lead to ecological fallacy: what was found for the NSAID class would not necessarily be found for each NSAID medication. In the context of COVID-19 patients, acute use of NSAIDs or acetaminophen was probably used for fever and muscle aches and pains related to the virus in the majority of cases. NSAIDs are taken chronically, however, for osteoarthritis, rheumatoid arthritis, and other chronic musculoskeletal conditions, more prevalent among older adults. Acetaminophen is also used for chronic pain conditions including osteoarthritis and back pain. Therefore, it is clinically important to ascertain possible effects in these patients, especially with the knowledge of associations of increased cardiovascular risk with chronic NSAID or chronic acetaminophen use.

This is the first published study assessing aspirin restricted to a dose > 150mg in which it exhibits the effects of the non-aspirin NSAIDs [3638]. In the literature, about half studied exclusively low-dose while the other half studied all doses of aspirin. Only one study assessed differential dose effects, separating patients in low- and high- dose aspirin groups [33]. In it, among those taking aspirin, only 51 patients received > 81mg; with mortality rate point estimates of 33.3% and 21.6% for those taking and not taking aspirin, respectively, doubt remained if the authors’ finding of no significance was an artifact of small sample size.

This is also the first study assessing effects of meloxicam or diclofenac among COVID-19 patients for acute or chronic use, before or after testing positive. In our study, 41.90% of chronic NSAID users took these medications. Additionally, we did not identify a study of chronic acetaminophen use among COVID-19 patients. In our study, 24.25% of chronic users were taking acetaminophen. With this study providing evaluation of meloxicam, diclofenac, or acetaminophen in the literature in COVID-19 patients, this is crucial information for providers, as this pertains to 55.99% of chronic users in our study. This also is the first study assessing effects among a documented highly adherent group of patients who tested positive for COVID-19 and who have been taking NSAIDs or acetaminophen chronically.


This study has several limitations. First, since this study is retrospective we can only derive associations from the data; a clinical trial would be needed to establish causal relationships. Also due to its retrospective nature, we were dependent on documentation related to patient characteristics, comorbidities, and mortality; we note the high capture rate of deaths in the CDW. Second, endemic to any propensity scores analysis, we do not know if there are baseline differences between groups for unobserved variables associated with mortality. Having said this, propensity scores analysis reduces the bias of potential unobserved covariates [39]. A specific example of this is our inability to capture potential medications altering the disease course, such as dexamethasone, remdesivir, monoclonal antibody therapy, and experimental administration of COVID-19 vaccines. Third, undoubtedly there was undocumented medication use due to some non-selective NSAIDs’ and acetaminophen’s over-the-counter status in the U.S. The VA may be the ideal healthcare organization to assess NSAID use and effects in the U.S. as many over-the-counter prescriptions are frequently prescribed within the VA: in 2017, about 50% of Veterans had $0 medication copayments while others had a $700/year cap [40]. This means, compared to no usage, having sporadic use as a control reflects a conservative estimate. However, NSAIDs are one of the classes taken most frequently, so this indeed may have more external validity to the general population, serving as an apt control. Fourth, our population has a high proportion of males, though a study has shown similarities between the VA population and the Medicare population [41]. Fifth, although we studied seven different medications, due to small numbers we were not able to look at treatment effects for COVID-19 mortality for every FDA-approved NSAID medication. Lastly, data from other continents is not compared, further limiting generalizability.


The results of this study show no association between chronic use of any of the six NSAIDs studied or with acetaminophen and all-cause mortality in Veterans diagnosed with COVID-19 infection. Stratified analyses revealed no clinical importance as well, negating the potential of differential effects of chronic use of each medication using the observed covariates in our study. Lastly, a clinically important association was also not revealed when chronic acetaminophen use was substituted for sporadic NSAID use.


We sincerely thank Glen H. Murata, M.D., who provided his insight and expertise, offered several ideas, helped curate and synthesize the data, guided in data collection, and aided in code validation. We would also like to acknowledge the resources and support of the VA COVID-19 Shared Data Resource and the VA Informatics and Computing Infrastructure (VINCI), VA HSR RES 13–457.


  1. 1. Fang L, Karakiulakis G, Roth M. Are patients with hypertension and diabetes mellitus at increased risk for COVID-19 infection? Lancet Respir Med. 2020;8: e21. pmid:32171062
  2. 2. Giollo A, Adami G, Gatti D, Idolazzi L, Rossini M. Coronavirus disease 19 (Covid-19) and non-steroidal anti-inflammatory drugs (NSAID). Ann Rheum Dis. 2021;80: e12. pmid:32321720
  3. 3. Gallelli L, Galasso O, Falcone D, Southworth S, Greco M, Ventura V, et al. The effects of nonsteroidal anti-inflammatory drugs on clinical outcomes, synovial fluid cytokine concentration and signal transduction pathways in knee osteoarthritis. A randomized open label trial. Osteoarthritis Cartilage. 2013;21: 1400–1408. pmid:23973155
  4. 4. Chmiel JF, Konstan MW, Accurso FJ, Lymp J, Mayer-Hamblett N, VanDevanter DR, et al. Use of ibuprofen to assess inflammatory biomarkers in induced sputum: Implications for clinical trials in cystic fibrosis. J Cyst Fibros. 2015;14: 720–726. pmid:25869324
  5. 5. EMA advice on the use of NSAIDs for Covid-19. Drug Ther Bull. 2020;58: 69. pmid:32234728
  6. 6. United States Food and Drug Administration. FDA advises patients on use of non-steroidal anti-inflammatory drugs (NSAIDs) for COVID-19. 2020 Mar 19 [cited 28 April 2020]. In: Drug Safety and Availability [Internet]. Silver Spring: United States Food and Drug Administration 2020 - [about 2 screens]. Available from: https://www.fda/gov/drugs/drug-safety-and-availability/fda-advises-patients-use-non-steroidal-anti-inflammatory-drugs-nsaids-covid-19
  7. 7. United States Department of Health and Human Services. Coronavirus disease 2019 (COVID-19): who is at increased risk for severe illness?—people of any age with underlying medical conditions. 2020 Jun 26 [cited 2 November 2020]. In: COVID-19 [Internet]. Atlanta: Centers for Disease Control and Prevention 2020 - [about 1 screen]. Available from:
  8. 8. United States Department of Health and Human Services. Coronavirus disease 2019 (COVID-19): evidence used to update the list of underlying medical conditions that increase a person’s risk of severe illness from COVID-19. 2020 Jun 26 [cited 2 November 2020]. In: COVID-19 [Internet]. Atlanta: Centers for Disease Control and Prevention 2020 - [about 1 screen]. Available from: Access date: November 2, 2020.
  9. 9. Coxib and traditional NSAID Trialists’ (CNT) Collaboration, Bhala N, Emberson J, Merhi A, Abramson S, et al. Vascular and upper gastrointestinal effects of non-steroidal anti-inflammatory drugs: meta-analyses of individual participant data from randomised trials. Lancet. 2013;382: 769–779. pmid:23726390
  10. 10. Helin-Salmivaara A, Virtanen A, Vesalainen R, Gronroos JM, Klaukka T, Idanpaan-Heikkila JE, et al. NSAID use and the risk of hospitalization for first myocardial infarction in the general population: a nationwide case-control study from Finland. Eur Heart J. 2006;27: 1657–1663. pmid:16731535
  11. 11. Schmidt M, Christiansen CF, Mehnert F, Rothman KJ, Sorensen HT. Non-steroidal anti-inflammatory drug use and risk of atrial fibrillation or flutter: population based case-control study. BMJ. 2011;343: d3450. pmid:21727167
  12. 12. Andersohn F, Suissa S, Garbe E. Use of first- and second-generation cyclooxygenase-2-selective nonsteroidal antiinflammatory drugs and risk of acute myocardial infarction. Circulation. 2006;113: 1950–1957. pmid:16618816
  13. 13. Chan AT, Manson JE, Albert CM, Chae CU, Rexrode KM, Curhan GC, et al. Nonsteroidal antiinflammatory drugs, acetaminophen, and the risk of cardiovascular events. Circulation. 2006;113: 1578–1587. pmid:16534006
  14. 14. Arnold N. CDW date of death analysis and vital status file comparison update. Washington, D.C.: Department of Veterans Affairs Data Quality Program Office of Informatics and Analytics Health Insurance Governance; 2019. pp. 1–11.
  15. 15. Nguyen TL, Collins GS, Spence J, Daures JP, Devereaux PJ, Landais P, et al. Double-adjustment in propensity score matching analysis: choosing a threshold for considering residual imbalance. BMC Med Res Methodol. 2017;17: 78. pmid:28454568
  16. 16. Eeren HV, Spreeuwenberg MD, Bartak A, de Rooij M, Busschbach JJ. Estimating subgroup effects using the propensity score method: a practical application in outcomes research. Med Care. 2015;53: 366–373. pmid:25738381
  17. 17. Austin PC, Grootendorst P, Anderson GM. A comparison of the ability of different propensity score models to balance measured variables between treated and untreated subjects: a Monte Carlo study. Stat Med. 2007;26: 734–753. pmid:16708349
  18. 18. Petersen ML, Porter KE, Gruber S, Wang Y, van der Laan MJ. Diagnosing and responding to violations in the positivity assumption. Stat Methods Med Res. 2012;21: 31–54. pmid:21030422
  19. 19. Jeong HE, Lee H, Shin HJ, Choe YJ, Filion KB, Shin JY. Association between NSAIDs use and adverse clinical outcomes among adults hospitalized with COVID-19 in South Korea: A nationwide study. Clin Infect Dis. 2021;73: e4179–e4188. pmid:32717066
  20. 20. Bruce E, Barlow-Pay F, Short R, Vilches-Moraga A, Price A, McGovern A, et al. Prior routine use of non-steroidal anti-inflammatory drugs (NSAIDs) and important outcomes in hospitalised patients with COVID-19. J Clin Med. 2020;9: 2586.
  21. 21. Chandan JS, Zemedikun DT, Thayakaran R, Byne N, Dhalla S, Acosta-Mena D, et al. Non-steroidal anti-inflammatory drugs and susceptibility to COVID-19. Arthritis Rheumatol. 2021;73: 731–739. pmid:33185016
  22. 22. Lund LC, Kristensen KB, Reilev M, Christensen S, Thomsen RW, Christiansen CF, et al. Adverse outcomes and mortality in users of non-steroidal anti-inflammatory drugs who tested positive for SARS-CoV-2: A Danish nationwide cohort study. PLoS Med. 2020;17: e1003308. pmid:32898149
  23. 23. Abu Esba LC, Alqahtani RA, Thomas A, Shamas N, Alswaidan L, Mardawi G. Ibuprofen and NSAID use in COVID-19 infected patients is not associated with worse outcomes: a prospective cohort study. Infect Dis Ther. 2021;10: 253–268. pmid:33135113
  24. 24. Chow JH, Khanna AK, Kethireddy S, Yamane D, Levine A, Jackson AM, et al. Aspirin use is associated with decreased mechanical ventilation, intensive care unit admission, and in-hospital mortality in hospitalized patients with Coronavirus Disease 2019. Anesth Analg. 2021;132: 930–941. pmid:33093359
  25. 25. Hong W, Chen Y, You K, Tan S, Wu F, Tao J, et al. Celebrex adjuvant therapy on Coronavirus Disease 2019: an experimental study. Front Pharmacol. 2020;11: 561674. pmid:33312125
  26. 26. Kragholm K, Gerds TA, Fosbol E, Andersen MP, Phelps M, Butt JH, et al. Association between prescribed ibuprofen and severe COVID-19 infection: a nationwide register-based cohort study. Clin Transl Sci. 2020;13: 1103–1107. pmid:32970921
  27. 27. 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: 471–479. pmid:33476420
  28. 28. 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: e0246825. pmid:33571280
  29. 29. Sahai A, Bhandari R, Koupenova M, Freedman J, Godwin M, McIntyre T, et al. SARS-CoV-2 receptors are expressed on human platelets and the effect of aspirin on clinical outcomes in COVID-19 patients. Res Sq: rs-119031 [Preprint]. 2020 [cited 2022 February 18]. Available from: pmid:33398263
  30. 30. Wong AYS, MacKenna B, Morton CE, Schultze A, Walker AJ, Bhaskaran K, et al. Use of non-steroidal anti-inflammatory and risk of death from COVID-19: an OpenSAFELY cohort analysis based on two cohorts. BMJ. 2021;80: 943–951. pmid:33478953
  31. 31. Asadi M, Sayar S, Radmanesh E, Naghshi S, Mousaviasl S, Jelvay S, et al. Efficacy of naproxen in the management of patients hospitalized with COVID-19 infection: a randomized, double-blind, placebo-controlled clinical trial. Diabetes Metab Syndr. 2021;15: 102319. pmid:34700294
  32. 32. Sahai A, Bhandari R, Godwin M, McIntyre T, Chung MK, Iskandar JP, et al. Effects of aspirin on short-term outcomes in hospitalized patients with COVID-19. Vasc Med. 2021;26: 626–632. pmid:34010070
  33. 33. Zhao X, Gao C, Dai F, Treggiari MM, Deshpande R, Meng L. Treatments associated with lower mortality among critially ill COVID-19 patients: a retrospective cohort study. Anesthesiology. 2021;135: 1076–1090. pmid:34597362
  34. 34. Son M, Noh MG, Lee JH, Seo J, Park H, Yang S. Effect of aspirin on coronavirus disease 2019: a nationwide case-control study in South Korea. Medicine. 2021;100(30):e26670. pmid:34397693
  35. 35. Chandan JS, Zemedikun DT, Thayakaran R, Byne N, Dhalla S, Acosta-Mena D, et al. Arthritis Rheumatol. 2021;73: 731–739. pmid:33185016
  36. 36. Fuster V, Sweeney JM. Aspirin: a historical and contemporary therapeutic overview. Circulation. 2011;123: 768–778. pmid:21343593
  37. 37. Antithrombotic Trialists Collaboration. Collective meta-analysis of randomized trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002;324: 71–86. pmid:11786451
  38. 38. Schmidt M, Lamberts M, Olsen AMS, Fosboll E, Niessner A, Tamargo J, et al. Cardiovascular safety of non-aspirin non-steroidal anti-inflammatory drugs: review and position paper by the working group for Cardiovascular Pharmacotherapy of the European Society of Cardiology. Eur Heart J. 2016;37: 1015–1023. pmid:26984863
  39. 39. Rosenbaum PR. Propsensity score. In: Armitage P, Colton T, editors. Encyclopaedia of biostatistics. 2nd ed. Boston, Massachusetts: Wiley; 2005. pp. 4267–4272.
  40. 40. Department of Veterans Affairs Veterans Health Administration Office of Policy and Planning. Number of Veteran patients by healthcare priority groups. 2018 Jan 24 [cited 16 April 2021]. In: Healthcare Services [Internet]. Washington: Department of Veterans Affairs National Center for Veterans Analysis and Statistics 2000–2017. [about 1 screen]. Available from:
  41. 41. Wong ES, Wang V, Liu CF, Hebert PL, Maciejewski ML. Do Veterans Health Administration enrollees generalize to other populations? Med Care Res Rev. 2016;73: 493–507. pmid:26589675