Article Text

Use of chronic medications and risk of death due to COVID-19 in hospitalised patients
  1. María Larrosa-García1,
  2. Sonia Garcia-Garcia1,
  3. Javier Louro2,
  4. Adrián Sánchez-Montalvá3,
  5. Julia Sampol Sirvent4,
  6. Salvador Augustín Recio5,
  7. Alfredo Guillén del Castillo5,
  8. Judit Riera-Arnau6,
  9. Maria Queralt Gorgas1,
  10. Marta Miarons1
  1. 1 Pharmacy Department, Vall d'Hebron University Hospital, Barcelona, Spain
  2. 2 Department of Epidemiology and Evaluation, IMIM, Hospital del Mar INAD, Barcelona, Spain
  3. 3 Infectious Diseases Department, Vall d'Hebron University Hospital, Barcelona, Spain
  4. 4 Pneumology Department, Vall d'Hebron University Hospital, Barcelona, Spain
  5. 5 Internal Medicine Department, Vall d'Hebron University Hospital, Barcelona, Spain
  6. 6 Pharmacology Department, Vall d'Hebron University Hospital, Barcelona, Spain
  1. Correspondence to Sonia Garcia-Garcia, Vall d'Hebron University Hospital, Barcelona 08035, Spain;{at}


Objectives To evaluate the potential association between chronic exposure to medication and death related to COVID-19.

Methods This is a retrospective cross-sectional study that included all patients hospitalised due to COVID-19 from 11 March to 4 June 2020 in our centre. Chronic patient medication was classified by the Anatomical Therapeutic Chemical (ATC) classification; demographic and clinical data were analysed. Multivariate logistic regression models were used to estimate the adjusted odds ratios (aOR) of death for each drug exposure; each aOR represents an independent model adjusted by clinical factors related to COVID-19 mortality.

Results The study included 978 patients with a mean (SD) age of 64.5 (17.7) years who were predominantly male (531, 54.3%). Of all 978 patients, 182 (18.61%) died during the follow-up of the study. The most common Charlson Comorbidity Index (CCI) was 0, 4.2% were smokers, 16.7% were obese, 47.4% had hypertension, and 19.4% were diabetic. Most patients (70.8%) were prescribed at least one treatment, 32.5% used >5 treatments, and 8.6% >10. Our data suggest that COVID-19 hospitalised patients taking trimethoprim and analogues, leukotriene receptor antagonists, calcineurin inhibitors, aldosterone antagonists, selective immunosuppressants, propulsives, insulins and analogues, and benzodiazepine derivatives have a higher risk of death.

Conclusions This study investigated the association between chronic exposure to drugs and the risk of death in COVID-19 patients. Our results have shed some light on the impact of chronic drug exposure on the risk of severe COVID-19; however, further research is needed to increase the understanding about its relevance.

  • Drug Monitoring
  • COVID-19
  • Chemical Safety

Data availability statement

Data are available upon reasonable request.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • The global pandemic of COVID-19 requires the identification of possible clinical predisposing or protective factors.

  • It has been reported that male sex and older age, among other factors, are associated with a worse prognosis. However, data regarding drugs indicated for other diseases as being potentially beneficial or harmful are scarce.


  • The present study widely investigates the association between drugs and the risk of severe COVID-19, increasing the understanding of their association.

  • COVID-19 hospitalised patients previously using insulins and analogues, leukotriene receptor antagonists and calcineurin inhibitors, among other treatments, had significantly higher odds of death than patients not taking those drugs.


  • Our results have shed some light regarding the impact of chronic treatments on COVID-19, even if the evidence is not strong enough to advise against the use of any groups of drugs.

  • Further research is needed in order to increase understanding about the impact of chronic drug exposure on the risk of severe COVID-19.


Coronavirus disease 2019 (COVID-19) is a potentially fatal disease caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The estimated mortality of diagnosed cases with COVID-19 ranges from 2.3–15.2%, depending on patient population, diagnostic strategies and, probably, additional poorly characterised factors.1 Several studies have reported that clinical conditions such as hypertension, cardiovascular disease, cerebrovascular disease, diabetes and chronic kidney disease are associated with a higher risk for severe complications and death in the case of COVID-19.2

The use of certain chronic medications has also been suggested as a risk factor influencing COVID-19 prognosis. Initial concerns related antihypertensive medication with poor COVID-19 outcome; however, those fears were rapidly disproved.3 4 There is accumulating data suggesting that antidiabetic oral drugs, such as sodium-glucose cotransporter-2 inhibitors (SGLT2i) and metformin, act as protective factors in the case of SARS-CoV-2 infection.5 6 The impact of immunomodulation therapies on COVID-19 disease remains controversial. On the one hand, they may impair the natural immune response against the virus,7 8 and on the other hand, immunosuppressants could minimise immune system over-activation; in fact, dexamethasone, tocilizumab and anakinra have been used to control exacerbated immune responses to COVID-19.9–13 Finally, anticoagulant therapy appears to be associated with a better prognosis and lower mortality in patients meeting sepsis-induced coagulopathy criteria or with notably elevated D-dimer.14–18

Given the uncertainties, the current research aims to analyse the potential effect of chronic treatments on clinical outcomes of hospitalised patients with COVID-19.


Study setting and COVID-19 pharmacological treatment

We conducted a retrospective cross-sectional study analysing patients with SARS-CoV-2 infection. This cohort included 978 patients who were admitted to the emergency department because of COVID-19 from 11 March to 4 June 2020 in the Vall d′Hebron University Hospital, which is located in northern Barcelona; it provides care for >430 000 people and has an emergency department with 71 beds. Patients with mild SARS-CoV-2 infection were discharged following the home care institutional protocol. Hospitalised patients with moderate-severe COVID-19 were treated according to the hospital’s clinical protocol available at the time of hospitalisation. During hospitalisation, maintaining usual medication for chronic diseases was at the discretion of the treating physicians.

Ethical consideration

The Institutional Review Board provided ethical clearance for this study (reference number: PR (AG) 240/2020). Patients were asked for an oral consent. If patients were unable to give oral consent, the Institutional Review Board granted a waiver of informed consent. The study was carried out in accordance with the Declaration of Helsinki of the World Medical Association as established in the 18th General Assembly and its modifications.19 The Organic Law 3/2018, of 5 December, on the Protection of Personal Data and Guarantee of Digital Rights was applied.

Microbiological diagnosis

Laboratory-confirmed COVID-19 was defined as a positive SARS-CoV-2 polymerase-chain-reaction (PCR) result from any respiratory sample (nasopharyngeal swab, sputum, bronchoalveolar lavage or aspirate or tracheal aspirate). The commercial Allplex 2019-nCoV multiplex real-time PCR assay (Seegene, South Korea) was used for the detection of three target genes of SARS-CoV-2 (E and N genes and RNA-dependent RNA polymerase).

Data collection and outcomes

Data were collected from the time of COVID-19 diagnosis until an outcome (death, discharge or hospitalised after 28 days of hospitalisation) was reached. We obtained patients’ sociodemographic characteristics, epidemiological history, comorbidities and chronic medications from electronic medical records. Information about therapy during hospitalisation and tests on respiratory samples and supportive measures needed was also collected. Blood test results, vital signs, symptoms and physical examination were evaluated on admission and weekly during hospitalisation. We used Research Electronic Data Capture software (REDCap, Vanderbilt University) for data recording. Three investigators independently reviewed all the collected data.

Definition of study variables and outcomes

Age and gender were collected at the time of hospitalisation. Smoking status was self-reported and stratified into active smokers, ex-smokers and no smokers. Previous medical conditions were recorded from the electronic medical record system. The age-adjusted Charlson Comorbidity Index (CCI)20 was calculated. The main outcome was in-hospital death due to COVID-19.

Drug exposure

Chronic medication was defined as drugs taken for at least a 3-month period before admission. Information about chronic treatments was obtained from the Integrated Electronic prescription System (SIRE), a drug management programme that coordinates drug prescription and dispensing within the Catalan Public Healthcare System. Chronic treatments were analysed according to the Anatomical Therapeutic Chemical (ATC) classification system.21

In case <10 people used any drug from an ATC group it was excluded from analysis.

Statistical analysis

Descriptive statistics are expressed as median and range or mean and SD for continuous variables and absolute number and percentage for categorical variables. We evaluated the differences in the mean age of those patients who died using the Mann-Whitney U test. For the remaining variables we assessed the differences in the proportion of patients who died using the χ2 test. Multivariate logistic regression models were used to estimate the adjusted odds ratio (aOR) and 95% confidence interval (95% CI) of death for each drug exposure. Each aOR represents an independent model adjusted by age, gender, CCI, arterial hypertension, smoking status, diabetes and obesity. We represented the different aOR for every drug exposure using forest plots. Since there were a high number of drug categories, we stratified the results by statistical significance. All analyses were performed using the statistical software R version 4.0.3 (Development Core Team, 2014).


Baseline characteristics of patients

At the end of the follow-up, 978 patients had been diagnosed with confirmed COVID-19 and admitted to the hospital. Table 1 shows baseline demographic and clinical characteristics of all patients during the time of the study.

Table 1

Baseline demographic and clinical characteristics of patients

The patients’ mean (SD) age was 64.5 (17.7) years and they were predominantly male (531 (54.3%)). The most common CCI was 0, and 58.2% of the patients had no or mild comorbidity (CCI 0, 1 and 2). The percentage of smokers was 4.2%, 16.7% of patients were obese, 47.4% had hypertension and 19.4% were diabetic. Regarding chronic treatments, 70.8% of patients included in the study used at least one chronic medication before hospital admission, 32.5% used >5 drugs and 8.6% >10. Of the 978 patients, 182 (18.6%) died during the follow-up of the study. Non-survivors were significantly older than survivors (mean (SD) age 78.2 (10.5) vs 61.4 (17.6) years, p<0.001). The proportion of non-survivors among male patients was significantly higher than among females (111 (20.9%) vs 71 (15.9%), p=0.044). The proportion of non-survivors increased significantly as the CCI increased (p<0.001). There were higher proportions of non-survivors among ex-smokers (65 (29.4%)) and active smokers (9 (22.0%)) than among no smokers (108 (15.3%)). No differences were found in patients with obesity (p=0.462). The proportion of non-survivors was also higher in patients with arterial hypertension (136 (29.3%) vs 46 (8.9%), p<0.001) and in diabetic patients (52 (27.4%) vs 130 (16.5%), p<0.001). There was a higher proportion of non-survivors as the number of drugs chronically used by patients increased, ranging from 17 (5.9%) non-survivors in the group of patients not exposed to any drug to 33 (39.3%) in those exposed to >10 drugs (p<0.001). Risk of death was not statistically different between patients admitted to the intense care unit or not (18.8% vs 17.4%, p=0.707).

Chronic drug exposure and COVID-19 related mortality

The median number of chronic drugs used by patients was 3 (range 0–18); 286 (29.2%) patients did not have any active prescription at the time of hospital admission and 318 (32.5%) patients were taking >10 drugs.

As shown in figure 1, we found statistically significant higher odds of death due to COVID-19 in those patients who had been exposed to trimethoprim and derivatives (aOR 5.15, 95% CI 1.89 to 14.02), leukotriene receptor antagonists (aOR 4.77, 95% CI 1.61 to 14.14), calcineurin inhibitors (aOR 3.65, 95% CI 1.58 to 8.41), aldosterone antagonists (aOR 3.13, 95% CI 1.05 to 9.35), selective immunosuppressants (aOR 2.99, 95% CI 1.31 to 6.78), propulsives (aOR 2.79, 95% CI 1.01 to 7.69), insulins and analogues (aOR 2.40, 95% CI 1.18 to 4.89) and benzodiazepine derivatives (antiepileptics) (aOR 2.05, 95% CI 1.27 to 3.31) in comparison with patients who did not use these treatments.

Figure 1

Adjusted odds ratios (aOR) of COVID-19 mortality by previous exposure to chronic treatments (significant results).

Other ATC groups showed a tendency towards either higher or lower risk of death due to COVID-19, but they were not statistically significant (online supplemental figure 1).

Supplemental material


The global pandemic caused by SARS-CoV-2 requires the identification of factors that influence the severity of COVID-19. The literature describes different pharmacology groups that could alter the prognosis of SARS-CoV-2 infection. The present study examines the association between the chronic use of drugs and death in COVID-19 hospitalised patients. We found that COVID-19 hospitalised patients previously using trimethoprim and derivatives, leukotriene receptor antagonists, aldosterone antagonists, propulsives, insulins and analogues, selective immunosuppressants, calcineurin inhibitors and benzodiazepine derivatives (antiepileptics) had potentially higher odds of death than other patients.

In our study, adjustments by age, gender, CCI, arterial hypertension, smoking status, diabetes and obesity were made because these factors have been identified as strong predictors of mortality in people with SARS-CoV-2 infection.22–26 We have found that the mortality rate was 18.6%, which is consistent with other studies performed in COVID-19 hospitalised patients. The study of Roso-Llorach et al found that the highest 30-day mortality rate was reported during the first wave (17%) and decreased afterwards, remaining stable at 13% in the second and third waves (overall 30% reduction).

Regarding the association of drugs and death, the use of the combination of trimethoprim and derivatives is related to a higher incidence of death due to COVID-19. Since this drug is most commonly used as prophylactic treatment for Pneumocystis jirovecii in patients with an impaired immune system due to oncohaematologic diseases,27 HIV28 or organ transplantation,29 we suggest that the increased mortality might be related to the immunosuppression status.

We have also found that the use of leukotriene receptor antagonists, which includes montelukast, was related to a higher mortality rate. Montelukast reduces inflammation and leads to smooth muscle relaxation in the small airways, and has been suggested as a treatment for COVID-19.30 31 The use of montelukast by patients with severe asthma could explain its relationship within higher COVID-19 mortality; however, the relationship between asthma and severe COVID-19 remains uncertain.32 33

Aldosterone antagonists have been identified as risk factors of death due to COVID-19. However, some authors have hypothesised that spironolactone could have a protective effect against COVID-19.34 35 A greater use of these treatments among polymedicated patients with cardiovascular diseases,36 which have been related to a worse prognosis in COVID-19 infection,37 could have impacted our results.

Selective immunosuppressants and calcineurin inhibitors were related to increased risk of death. Some studies have suggested that serum concentrations of pro-inflammatory cytokines are associated with COVID-19 severity, so it could be expected that immunosuppressive drugs could act as a protective factor for severe disease.38 Nevertheless, it is well known that these drugs increase the risk of severe infections39 and involve faster infection disease progression. During the early viral replication phase, immunosuppression therapy may lead to higher viral load peak, higher direct tissue damage and cytokine storm release. Consequently, immunosuppressant therapy (associated with comorbidities) could impair the immune system’s ability to fight the virus and involve a worse prognosis of COVID-19 in transplant recipients29 39 40 and patients with autoimmune diseases,41–43 as has been reported by others.44 45

The positive correlation found between propulsives and poor prognosis may be attributable to various unknown facts. Propulsive drugs are widely used in oncology patients46 47 and palliative care.48 In this group of fragile patients, a higher mortality after COVID-19 is to be expected.49

Insulins and analogues constitute a therapeutic regimen for diabetes mellitus, a chronic and low-grade inflammatory disease. Yu et al 50 conducted a retrospective study analysing hospitalised diabetic patients with COVID-19 and found that insulin was related to higher mortality, as we have observed. Authors have suggested that insulin may have a deleterious effect by increasing the pro-inflammatory cytokine levels and lung inflammation, as has been proven by several in vitro, in vivo and clinical studies.51 52

Benzodiazepine derivatives (antiepileptics) were related to a higher risk of death. Asaadi-Pooya et al 53 and other authors have stated that it is highly likely that some patients with severe COVID-19 have central nervous system involvement and neurological manifestations. 54–56

We did not identify any ATC group related to a lower risk of death, even if other antihistamines for systemic use and β-blocking agents, non-selective among others, showed a tendency towards having a protective effect, as shown in online supplemental figure 1. These ATC groups might have a protective effect that could not be identified as statistically significant in our study due to the limited sample and/or confounders effect.

The pathogenesis of distress syndrome in the clinical context of COVID-19 may be related to the direct effect of SARS-CoV-2 on alveolar epithelial cells and to the indirect effects of infection-related hypoxia, severe inflammatory response and microvascular pulmonary thrombosis.57 Given that the thrombogenic features of COVID-19 potentially cause a progressive disease course, anticoagulation may protect against severe COVID-19; however, evidence regarding the potential beneficial effect of anticoagulants is controversial. In a retrospective cohort study including severely ill COVID-19 patients, a decrease in 28 day-mortality was observed in heparin users versus non-users in specific subgroups of patients.14 Our results did not show any beneficial effect of anticoagulants in COVID-19 patients, as has been found in other retrospective studies.57–59

It has been hypothesised that drugs that act on the renin-angiotensin-aldosterone system (RAAS), such as angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin II receptor blockers (ARBs), may affect the susceptibility to and severity of COVID-19.3 4 60 61 However, currently available results show that the use of ACEIs or ARBs are not associated with an increased risk of severity of COVID-19,3 4 as the results of our study show. Other studies analysing our population, conducted in Spain62 and specifically in Catalonia,63 as well as other European countries,64 agree with our findings.

Finally, glucocorticoids for systemic use have been associated with better outcomes when administered to patients with severe COVID-1913; specifically, the RECOVERY trial has shown that corticoid administration leads to lower mortality in patients requiring oxygen support or mechanical ventilation.65 However, chronic exposure to glucocorticoids was associated with higher odds of hospitalisation in patients with inflammatory bowel disease,66 and rheumatic and musculoskeletal diseases.67 We have found that there was a tendency towards higher risk of death in patients who were chronically exposed to corticoids, and we hypothesise that the time of administration plays a key role in corticosteroids effect: chronic use of systemic corticosteroids may worsen COVID-19 prognosis by suppressing the immune response and allowing virus replication, whereas corticoid use in the hyperinflammation phase of COVID-19 disease may prevent multiple organ failure and acute respiratory distress syndrome.68 69

We found that the mortality rate was 18.6%, which is consistent with other studies performed in COVID-19 hospitalised patients. The study of Roso-Llorach et al found that the highest 30-day mortality rate was reported during the first wave (17%) and decreased afterwards, remaining stable at 13% in the second and third waves (overall 30% reduction).70

Our study has some limitations that might reduce our ability to establish a causal relationship between chronic drug exposure and severe COVID-19, and should be considered when interpreting the results. First, there are biases on exposure; for example, we only analysed the impact of ATC groups that were used by >10 patients because we considered that the information available in these cases was insufficient to draw conclusions. Second, we considered that patients were exposed to treatment if they had an active electronic medical prescription from 3 months before hospital admission; however, we cannot affirm that patients were actually taking their medication. This could lead to an underestimation of these drugs’ impact on COVID-19. Third, we did not evaluate over-the-counter treatments, which could interfere with the results of the prescribed drugs and might cause interaction in their effect. Fourth, we included patients admitted to the hospital due to COVID-19; therefore, even if we detect that chronic exposure to drugs belonging to a certain ATC group is a risk factor for severe COVID-19, this might not apply to the whole population taking the drug in the community. We assumed that patients who were not hospitalised and had COVID-19 overcame the infection and survived, but we cannot affirm that there were not severe COVID-19 cases outside the hospital. Fifth, chronic treatment use is related to certain health conditions or chronic diseases; this implies that exposure to the drug could act as a confounder. We reduced the risk of bias and the confounder effect by adjusting for different parameters including age, gender, CCI, arterial hypertension, diabetes, smoking status and obesity; however, we cannot affirm that the risk of bias was completely erased. Finally, we are aware that multiple comparisons may increase the chances of finding significant results even if they do not exist. Therefore, our study should be interpreted with caution, and the results should help to generate hypotheses.

In this study, we analysed the relationship between exposure to chronic treatments before COVID-19 and the risk of death in hospitalised COVID-19 patients. COVID-19 prognosis depends on a myriad of factors, some of which remain to be identified, making it challenging to detect individual risk or protective factors; however, making efforts to detect drugs that might exert an effect on COVID-19 outcomes is crucial in order to optimise chronic treatments in the context of the SARS-CoV-2 pandemic and potentially identify treatments for COVID-19. We have observed that the use of certain drugs including trimethoprim and derivatives, leukotriene receptor antagonists, aldosterone antagonists, propulsives, insulins and analogues, selective immunosuppressants, calcineurin inhibitors and benzodiazepine derivatives (antiepileptics) are related to a potential higher risk of severe COVID-19. Evidence is not strong enough to advise against the use of any groups of drugs; however, our results have shed some light regarding the impact of chronic treatments on COVID-19 and they should be taken into consideration. Further research is needed in order to increase understanding about the impact of chronic drug exposure on the risk of severe COVID-19.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication


We want to thank Vall d′Hebron COVID-19 working group for their collaboration.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • EAHP Statement 4: Clinical Pharmacy Services.

  • X @laresiLarrosa, @J_Riera_Farcli

  • Correction notice This paper has been corrected since it was first published, there was an error in the manuscript title.

  • Contributors ML-G, SG-G, JL and MM participated in research design and writing the article. ML-G, SG-G, JL. AS-M, JS-S, SA-R, AG-C, JR-A, M-QG and MM critically reviewed the article. SG-G and M L-G are guarantors.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.