Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID- 19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Articles Use of renin–angiotensin–aldosterone system inhibitors and risk of COVID-19 requiring admission to hospital: a case-population study Francisco J de Abajo, Sara Rodríguez-Martín, Victoria Lerma, Gina Mejía-Abril, Mónica Aguilar, Amelia García-Luque, Leonor Laredo, Olga Laosa, Gustavo A Centeno-Soto, Maria Ángeles Gálvez, Miguel Puerro, Esperanza González-Rojano, Laura Pedraza, Itziar de Pablo, Francisco Abad-Santos, Leocadio Rodríguez-Mañas, Miguel Gil, Aurelio Tobías, Antonio Rodríguez-Miguel, Diego Rodríguez-Puyol, on behalf of the MED-ACE2-COVID19 study group* Summary Background Concerns have been raised about the possibility that inhibitors of the renin–angiotensin–aldosterone Lancet 2020; 395: 1705–14 system (RAAS) could predispose individuals to severe COVID-19; however, epidemiological evidence is lacking. We Published Online report the results of a case-population study done in Madrid, Spain, since the outbreak of COVID-19. May 14, 2020 https://doi.org/10.1016/ S0140-6736(20)31030-8 Methods In this case-population study, we consecutively selected patients aged 18 years or older with a PCR- See Comment page 1671 confirmed diagnosis of COVID-19 requiring admission to hospital from seven hospitals in Madrid, who had been *Listed at the end of the Article admitted between March 1 and March 24, 2020. As a reference group, we randomly sampled ten patients per case, Clinical Pharmacology Unit individually matched for age, sex, region (ie, Madrid), and date of admission to hospital (month and day; index (Prof F J de Abajo PhD, date), from Base de datos para la Investigación Farmacoepidemiológica en Atención Primaria (BIFAP), a Spanish S Rodríguez-Martín PhD, primary health-care database, in its last available year (2018). We extracted information on comorbidities and V Lerma RN, prescriptions up to the month before index date (ie, current use) from electronic clinical records of both cases and A Rodríguez-Miguel PhD) and Department of Nephrology controls. The outcome of interest was admission to hospital of patients with COVID-19. To minimise confounding (Prof D Rodríguez-Puyol PhD), by indication, the main analysis focused on assessing the association between COVID-19 requiring admission to University Hospital Príncipe de hospital and use of RAAS inhibitors compared with use of other antihypertensive drugs. We calculated odds ratios Asturias, Alcalá de Henares, (ORs) and 95% CIs, adjusted for age, sex, and cardiovascular comorbidities and risk factors, using conditional Madrid, Spain; Department of Biomedical Sciences logistic regression. The protocol of the study was registered in the EU electronic Register of Post-Authorisation (Pharmacology Section) Studies, EUPAS34437. (Prof F J de Abajo, S Rodríguez-Martín, Findings We collected data for 1139 cases and 11 390 population controls. Among cases, 444 (39·0%) were female and A Rodríguez-Miguel, A García-Luque PhD, the mean age was 69·1 years (SD 15·4), and despite being matched on sex and age, a significantly higher proportion Prof M Puerro PhD) and of cases had pre-existing cardiovascular disease (OR 1·98, 95% CI 1·62–2·41) and risk factors (1·46, 1·23–1·73) than Department of Medicine did controls. Compared with users of other antihypertensive drugs, users of RAAS inhibitors had an adjusted OR for (Prof D Rodríguez-Puyol), COVID-19 requiring admission to hospital of 0·94 (95% CI 0·77–1·15). No increased risk was observed with either University of Alcalá (IRYCIS), Alcalá de Henares, Madrid, angiotensin-converting enzyme inhibitors (adjusted OR 0·80, 0·64–1·00) or angiotensin-receptor blockers (1·10, Spain; Clinical Pharmacology 0·88–1·37). Sex, age, and background cardiovascular risk did not modify the adjusted OR between use of RAAS Department, University inhibitors and COVID-19 requiring admission to hospital, whereas a decreased risk of COVID-19 requiring admission Hospital La Princesa, Teófilo Hernando Institute, to hospital was found among patients with diabetes who were users of RAAS inhibitors (adjusted OR 0·53, 95% CI Autonomous University of 0·34–0·80). The adjusted ORs were similar across severity degrees of COVID-19. Madrid, Institute of Health La Princesa, Madrid, Spain Interpretation RAAS inhibitors do not increase the risk of COVID-19 requiring admission to hospital, including fatal (G Mejía-Abril MD, Prof F Abad-Santos PhD); cases and those admitted to intensive care units, and should not be discontinued to prevent a severe case of COVID-19. Clinical Pharmacology Unit, University Hospital Ramón y Funding Instituto de Salud Carlos III. Cajal (IRYCIS), Madrid, Spain (M Aguilar MD, M Ángeles Gálvez PhD, Copyright © 2020 Elsevier Ltd. All rights reserved. I de Pablo MD); Department of Clinical Pharmacology, Defense Introduction increase expression of ACE2.3–5 These findings have led Central Hospital, Madrid, Spain Severe acute respiratory syndrome coronavirus 2 some researchers to postulate that the use of these drugs (A García-Luque, Prof M Puerro); Department of Clinical (SARS-CoV-2) uses the angiotensin-converting enzyme 2 might enhance the access of SARS-CoV-2 into cells, Pharmacology, Clinic Hospital (ACE2) as the receptor for its spike protein to invade cells predisposing patients to infection or increasing severity of San Carlos, Department of and replicate.1 ACE2 presents a high homology with ACE, COVID-19.6–8 This hypothesis was fuelled by results from Pharmacology and Toxicology, a key enzyme in the regulation of blood pressure.2 In some the first case series that was published in which age, Complutense University of Madrid (IdISSC), Madrid, Spain animal studies, renin–angiotensin–aldosterone system hypertension, diabetes, and coronary heart disease— (L Laredo PhD, (RAAS) inhibitors (a category that includes ACE inhibitors conditions associated with the use of RAAS inhibitors— E González-Rojano PhD); and angiotensin-receptor blockers) have been reported to were identified as potential risk factors for severe cases and Geriatric Department www.thelancet.com Vol 395 May 30, 2020 1705 Articles (Prof L Rodríguez-Mañas PhD) and Institute of Biomedical Research in context Research (L Pedraza MD), University Hospital of Getafe, Evidence before this study Added value of this study Getafe, Madrid, Spain; Centre Inhibitors of the renin–angiotensin–aldosterone system (RAAS) Using a case-population study design, we found no difference of Network Biomedical have been hypothesised to predispose patients to more severe between the risk of COVID-19 requiring admission to hospital Research on Frailty and Healthy Ageing (CIBERFES), Institute COVID-19. This hypothesis is based on two facts: these drugs associated with the use of RAAS inhibitors and the risk of Health Carlos III, Madrid, have been reported to upregulate the expression of angiotensin- associated with the use of other antihypertensive drugs, once Spain (O Laosa PhD, converting enzyme (ACE) 2, the gateway used by severe acute fully adjusted for age, sex, and cardiovascular comorbidities Prof L Rodríguez-Mañas); respiratory syndrome coronavirus 2 to enter cells, and patients and risk factors. Additionally, we found no increased risk with Clinical Pharmacology Department, University with severe COVID-19 have a high prevalence of cardiovascular either ACE inhibitors or angiotensin-receptor blockers. The Hospital Puerta de Hierro, diseases. Many scientific societies adopted the position of lack of association between RAAS inhibitors and risk of Majadahonda, Madrid, Spain not recommending the discontinuation of treatment. COVID-19 requiring admission to hospital was observed in (G A Centeno-Soto PhD); Yet, epidemiological evidence is lacking and the hypothesis has both the most severe (fatal cases and those needing Division of Pharmacoepidemiology and not been confirmed or refuted. We searched PubMed on admission to an intensive care unit) and less severe inpatients. Pharmacovigilance, Spanish April 15, 2020, for publications in English since Jan 1, 2020, Implications of all the available evidence Agency for Medicines and using the search terms “COVID-19” and “ACE inhibitors OR Medical Devices, Madrid, Spain The available evidence supports that RAAS inhibitors are safe angiotensin-converting enzyme inhibitors OR angiotensin (M Gil PhD); and Institute of and should not be discontinued for fear of an increased risk of Environmental Assessment receptor blockers OR RAAS inhibitors OR RAS inhibitors OR RAAS COVID-19. and Water Research (IDAEA), blockers OR RAS blockers”, filtering by “human” and Spanish Council for Scientific “observational study”, and no result was returned. Research (CSIC), Barcelona, Spain (A Tobías PhD) Correspondence to: in-hospital deaths.9–12 By contrast, other authors have Here, we collected data on patients (cases) aged 18 years Prof Francisco J de Abajo, Department of Biomedical proposed use of angiotensin-receptor blockers as a or older with a PCR-confirmed diagnosis of COVID-19 Sciences, University of Alcalá, preventive measure, or even a therapy, for COVID-19 who were admitted to seven hospitals in Madrid, Spain, 28805 Alcalá de Henares, Madrid, because of their potential to reduce lung injury caused by between March 1 and March 24, 2020. During the study Spain angiotensin II.13 RAAS inhibitors are among the most period, the Madrid region was in the upward part of the francisco.abajo@uah.es widely used drugs globally for indications such as outbreak, with an average of 465 inpatient admissions, hypertension, heart failure, kidney complications of 48 intensive care unit (ICU) admissions, and 76 deaths diabetes, and myocardial infarction; hence their per day associated with COVID-19 (official data are discontinuation because of COVID-19 could cause patients available online). The seven participating hospitals For more on COVID-19, ACE2, harm.14 Scientific societies and drug regulatory agencies belong to the National Health System (NHS) and are and hypertension see alike have advised against their discontinuation until distributed all over the Autonomous Community of http://www.nephjc.com/news/ covidace2 sound evidence is available.15 Madrid (appendix p 2). Cases were selected consecutively To provide a solution to this urgent issue, we designed according to the date of admission to hospital until the For official data on numbers of COVID-19 cases in Spain see a pharmacoepidemiological study with the aim of planned sample size was met in a competitive manner. https://covid19.isciii.es/ assessing whether the odds of exposure to RAAS For our population controls, we used a random sample of See Online for appendix inhibitors relative to other antihypertensive drugs was ten individuals per case, individually matched to our higher among patients with COVID-19 admitted to cases by sex, age (exact), region (ie, Madrid), and date hospital than in the general population, adjusted for age, (day and month), extracted from the last available year sex, and cardiovascular comorbidities and risk factors. (2018) from the primary health-care database Base de datos para la Investigación Farmacoepidemiológica en For BIFAP database see Methods Atención Primaria (BIFAP; figure 1). www.bifap.org Study design and participants The Ethics Research Committee of the University In this case-population study, we used an epidemiological Hospital Príncipe de Asturias (the coordinating centre) approach, conceived as a surveillance method to assess assessed the study protocol and granted a favourable adverse drug effects,16 to provide a rapid answer to our opinion on March 18, 2020, including a waiver for the study question. Briefly, this method uses data from a series informed consent of patients taking part in the study. of patients with an illness from a specific region (cases) The data extracted were fully anonymised and no attempt and data from patients randomly sampled from a primary was made to interview patients or their relatives. The health-care database in the same region (population study complied with the provisions of the Spanish controls). Assuming that the primary health-care database legislation and the Declaration of Helsinki 2013. The represents the source population of the cases, a random Scientific Committee of BIFAP granted access to sample of controls from that database would provide a pseudonymised data for controls on March 26, 2020. valid estimate of the prevalence of the exposure and The study protocol is registered in the EU electronic For study protocol see http://www.encepp.eu/encepp/ covariates in the source population, approaching the Register of Post-Authorisation Studies, EUPAS34437, viewResource.htm?id=34501 primary base paradigm of case-control studies.17 and is available online. 1706 www.thelancet.com Vol 395 May 30, 2020 Articles Data sources and collection breathing rate >30 breaths per min, while breathing Information on disease and the clinical disease course of ambient air); abnormal chest x-ray compatible with cases was retrieved from hospital medical records, while COVID-19-associated pneumonia (bilateral pneumonia information on drug exposure and comorbidities before or unilateral pneu­monia with damage in different lung admission were mainly obtained from electronic primary lobes); and relevant clinical alterations including haemo­ health-care records that can be accessed through the dynamic, hepatic, renal, or haematological derangements, NHS’s primary health-care data platform, HORUS, by together with clinically significant laboratory abnor­ authorised health-care workers in Madrid. All electronic malities (such as abnormal increase in D-dimer, ferritin, case report forms were sent on a weekly basis from each lactate dehydrogenase, or C-reactive protein), or severe hospital to the coordinating centre, where data quality lymphocytopenia. The date of admission was considered control was done. This quality control involved the the index date. Then, we followed up patients and selection of a random sample of clinical records from each identified those who had been admitted to the ICU or hospital and double-checking the information collected on who died in hospital. When considering a severity drug exposure and comorbidities through HORUS. analysis, these patients were considered to be the most Information on drug exposure and comorbidities of severe cases. population controls was obtained from the database The antihypertensive drugs examined were ACE inhibi­ BIFAP, owned by the Spanish Agency for Medicines and tors, angiotensin-receptor blockers, renin inhibitors, Medical Devices, which extracts infor­ mation from the aldosterone antagonists, calcium-chan­ nel blockers, electronic primary health-care records of the NHS from diuretics, β-blockers, and α-blockers for cardiovascular participating regions. This database contains information indications (appendix p 8). Also, we grouped ACE inhibi­ from 1·15 million patients from Madrid, and the distri­ tors, angiotensin-receptor blockers, renin inhibitors, and bution of age and sex among patients is comparable to the aldosterone antagonists (alone or combined with any population census of the region (appendix pp 3–4). The drug) in a variable called RAAS inhibitors; and calcium- HORUS and BIFAP databases access the same primary channel blockers, β-blockers, diuretics, and α-blockers health-care data of patients in the Madrid NHS, and the (alone or combined with other drugs different from RAAS catchment population of the seven hospitals taking part inhibitors) in a variable called other antihypertensive in the study and the population attending primary care drugs. were the same. We defined exposure to the drug or drugs of interest as current use when an individual had a prescription lasting Outcomes until the month before the index date; otherwise exposure The main outcome variable was admission to hospital of was defined as non-use. For the main analysis, we patients with COVID-19 confirmed by a positive PCR test. generated a variable with the following mutually exclusive Hospitals posted in-house protocols for clinical manage­ categories: non-use of any antihypertensive drug, current ment and hospital admission based on criteria issued by use of RAAS inhibitors, and current use of other anti­ the Ministry of Health, including, but not limited to, hypertensive drugs and we used current use of other respiratory failure (oxygen saturation <90%, severe antihypertensive drugs as the reference category (unless hypoxaemia [partial pressure of oxygen <60 mm Hg], or otherwise specified). When a patient used a RAAS Madrid region has Seven hospitals Inpatient admission of patients Hospital records (diagnosis, test, approximately 6 million with COVID-19 (n=1139) and clinical evolution) inhabitants HORUS Primary health-care records BIFAP database for Madrid (2018) Population controls (comorbidities and drug exposure) 1·15 million patients (n=11 390) Quasi-random sample BIFAP Random sample of ten patients per case, matched for age, sex, and index date (month and day) of admission of cases to hospital Figure 1: Case-population study design COVID-19 cases requiring admission to hospital were selected consecutively from seven hospitals in Madrid, Spain. Data were collected for ten individuals per case who were matched for age, sex, and index date (day and month) of hospital admission of cases (matched controls) from the 2018 Madrid region database of BIFAP, a national primary health-care database. Drug exposure and comorbidities before the index date (2020 for cases and 2018 for controls) were collected from primary health-care records of the NHS in Madrid: for cases through HORUS (an online platform to access primary-care clinical records from any health-care centre of the NHS in Madrid) and for controls through BIFAP. BIFAP=Base de datos para la Investigación Farmacoepidemiológica en Atención Primaria. NHS=National Health System. www.thelancet.com Vol 395 May 30, 2020 1707 Articles inhibitor concomitantly with any other anti­hypertensive pharmacological subgroups: ACE inhibitors, angiotensin- drug, they were always assigned to the RAAS inhibitor receptor blockers (excluding current users of ACE category. Subsequently, we disaggregated the category inhibitors), aldosterone antagonists (excluding current current use of RAAS inhibitors into their different users of ACE inhibitors or angiotensin-receptor blockers), and renin inhibitors (excluding current users of any other Cases (n=1139) Controls (n=11 390) Crude odds ratio* RAAS inhibitor). For some analyses, we also disaggregated the category current use of other antihypertensive drugs Sex into its respective components (calcium-channel blockers, Male 695 (61·0%) 6950 (61·0%) ·· β-blockers, diuretics, and α-blockers), and used current Female 444 (39·0%) 4440 (39·0%) ·· use of calcium-channel blockers as the reference category. Age, years 69·1 (15·4) 69·1 (15·4) ·· Among current users of RAAS inhibitors, we distin­ Comorbidities guished when they used the drugs in monotherapy or Hypertension 617 (54·2%) 5644 (49·6%) 1·27 (1·10–1·46) combined with other antihypertensive drug (either in Diabetes 310 (27·2%) 2311 (20·3%) 1·50 (1·30–1·73) fixed-dose combinations or concomitant use as separate Dyslipidaemia 444 (39·0%) 3530 (31·0%) 1·49 (1·30–1·70) medicinal products). The date of first prescription of the Ischaemic heart disease 119 (10·5%) 862 (7·6%) 1·46 (1·19–1·80) current treatment episode was also extracted to estimate Heart failure 80 (7·0%) 400 (3·5%) 2·18 (1·68–2·82) the duration of treatment (categorised as up to 1 year and Atrial fibrillation 138 (12·1%) 970 (8·5%) 1·54 (1·26–1·88) longer than 1 year). Thromboembolic disease 44 (3·9%) 290 (2·6%) 1·55 (1·12–2·14) Cerebrovascular accident 73 (6·4%) 569 (5·0%) 1·32 (1·02–1·70) Statistical analysis COPD 119 (10·5%) 923 (8·1%) 1·35 (1·10–1·66) We estimated the sample size needed for different Asthma 78 (6·9%) 630 (5·5%) 1·26 (0·99–1·61) scenarios of effect size (ie, odds ratio [OR]) and prevalence Cancer 200 (17·6%) 1573 (13·8%) 1·35 (1·14–1·60) of use of antihypertensive drugs in controls, assuming Chronic renal failure 89 (7·8%) 573 (5·0%) 1·65 (1·29–2·09) an α error of 0·05 and a power of 80%. According to this Background cardiovascular risk calculation, a sample size of 1000 cases and 10 000 con­ Cardiovascular diseases† 312 (27·4) 2403 (21·1%) 1·98 (1·62–2·41) trols would allow us to detect an OR of 1·5 or greater if Cardiovascular risk factors‡ 504 (44·3) 4983 (43·8%) 1·46 (1·23–1·73) the prevalence of use of RAAS inhibitors among controls No cardiovascular disease or risk factors 323 (28·4) 4004 (35·2%) 1 (ref) was at least 5%. Current use§ We express quantitative variables as mean (SD) and RAAS inhibitors 497 (43·6%) 3822 (33·6%) 1·63 (1·43–1·87) qualitative variables as frequencies and percentages. ACE inhibitors 240 (21·1%) 2192 (19·2%) 1·13 (0·97–1·31) Differences in means were assessed using the Student’s Angiotensin-receptor blockers 244 (21·4%) 1616 (14·2%) 1·70 (1·45–1·98) t test and differences in percentages were assessed using Aldosterone antagonists 38 (3·3%) 218 (1·9%) 1·78 (1·25–2·53) the χ² test. We describe the distribution of comorbidities Renin inhibitors 1 (0·1%) 8 (0·1%) 1·25 (0·16–9·99) among cases and controls and their association with Other antihypertensive drugs 529 (46·4%) 3844 (33·8%) 1·90 (1·66–2·18) COVID-19 requiring admission to hospital was assessed Calcium-channel blockers 212 (18·6%) 1459 (12·8%) 1·59 (1·35–1·87) through univariable conditional logistic regression to Diuretics 347 (30·5%) 2579 (22·6%) 1·58 (1·37–1·83) calculate crude ORs) and 95% CIs (adjusted for age and β-blockers 200 (17·6%) 1303 (11·4%) 1·69 (1·43–1·99) sex due to matching). α-blockers 40 (3·5%) 183 (1·6%) 2·24 (1·58–3·18) When assessing potential confounding factors (covariates), we considered the presence of the fol­lowing Participating hospitals comorbidities at the index date: history of hypertension, Hospital Universitario Príncipe de 315 (27·7%) 3150 (27·7%) ·· Asturias diabetes, dyslipidaemia (defined as use of lipid-lowering Hospital Universitario de La Princesa 200 (17·6%) 2000 (17·6%) ·· drugs), ischaemic heart disease, atrial fibrillation, heart Hospital Universitario Ramón y Cajal 176 (15·5%) 1760 (15·5%) ·· failure, thromboembolic disease, cerebrovascular acci­dent, Hospital Clínico San Carlos 127 (11·2%) 1270 (11·2%) ·· asthma, chronic obstructive pulmonary disease, chronic Hospital Central de la Defensa Gómez Ulla 123 (10·8%) 1230 (10·8%) ·· renal failure, and cancer. With the cardiovascular risk factors Hospital Universitario Puerta de Hierro- 99 (8·7%) 990 (8·7%) ·· and comorbidities, we constructed a composite variable of Majadahonda background cardiovascular risk with the following three Hospital Universitario de Getafe 99 (8·7%) 990 (8·7%) ·· categories: patients with history of any of ischaemic heart Data are n (%), mean (SD), or odds ratio with 95% CI in parentheses. ACE=Angiotensin-converting enzyme. disease, cerebrovascular accident, heart failure, atrial fibril­ COPD=chronic obstructive pulmonary disease. RAAS=renin–angiotensin–aldosterone system. *The exposure to the lation, or thromboembolic disease; patients with any of the specific comorbidity or drug was compared with non-exposure of that specific comorbidity or drug; crude odds ratios following cardiovascular risk factors (and no cardiovascular are adjusted for sex and age. †Includes ischaemic heart disease, cerebrovascular accident, heart failure, atrial disease): hypertension, dyslipidaemia, diabetes, or chronic fibrillation, and thromboembolic disease. ‡Includes hypertension, dyslipidaemia, diabetes, and chronic renal failure. §Patients can be counted several times if they were current users of two or more antihypertensive drugs belonging to renal failure; and the remainder of patients without any different pharmacological classes, hence total exceeds 100%. cardiovascular diseases or risk factors. We assessed the association between current use of Table 1: Demographic and clinical characteristics of cases with COVID-19 and population controls RAAS inhibitors and risk of COVID-19 requiring 1708 www.thelancet.com Vol 395 May 30, 2020 Articles admission to hospital, compared with either current use Cases Matched controls Crude odds Adjusted odds of other antihypertensive drugs (main analysis) or non- (n=1139) (n=11 390) ratio* ratio† use of any antihypertensive drug, through univariable Current use of other 155 (13·6%) 1129 (9·9%) 1 (ref) 1 (ref) and multivariable conditional logis­tic regressions. In antihypertensive drugs the multivariable model, we included all covariates Current use of RAAS inhibitors 497 (43·6%) 3822 (33·6%) 0·94 (0·77–1·14) 0·94 (0·77–1·15) described as potential confounding fac­ tors (with the ACE inhibitors 240 (21·1%) 2192 (19·2%) 0·78 (0·63–0·97) 0·80 (0·64–1·00) exception of hypertension) to calculate adjusted ORs Monotherapy 82 (7·2%) 757 (6·7%) 0·75 (0·57–1·00) 0·83 (0·62–1·12) and 95% CIs. Current users of RAAS inhibitors had a Combinations 158 (13·9%) 1435 (12·6%) 0·80 (0·63–1·01) 0·78 (0·62–0·99) prevalence of hypertension over 90% (appendix p 13); Angiotensin-receptor blockers 237 (20·8%) 1552 (13·6%) 1·11 (0·89–1·38) 1·10 (0·88–1·37) for this reason, we did not include hypertension in Monotherapy 38 (3·3%) 328 (2·9%) 0·82 (0·56–1·20) 0·87 (0·60–1·28) the multivariable model because the potential for Combinations 199 (17·5%) 1224 (10·8%) 1·18 (0·94–1·48) 1·15 (0·92–1·45) confounding was minimal. Aldosterone antagonists 19 (1·7%) 71 (0·6%) 2·05 (1·20–3·49) 1·68 (0·97–2·91) We assessed potential effect modification by age (<70 vs Renin inhibitors 1 (0·1%) 7 (0·1%) 1·08 (0·13–8·86) 1·04 (0·13–8·62) ≥70 years), sex, hypertension, diabetes, and background Non-use 487 (42·8%) 6439 (56·5%) 0·47 (0·38–0·58) 0·55 (0·44–0·68) cardiovascular risk through a stratified analysis. We assessed the statistical interaction by comparing the Data are n (%) or odds ratio with 95% CI in parentheses. The different pharmacological classes examined are mutually exclusive categories, so that patients who used combinations are counted only once, applying the following criteria: adjusted ORs across different strata using the Altman ACE inhibitors include current users of any ACE inhibitors alone or combined with any other antihypertensive drug and Bland test of interaction.18 Statistical significance (in fixed-dose combinations or in different medicinal products); angiotensin-receptor blockers include current users of was set at a p value of less than 0·05. However, for any angiotensin-receptor blocker, alone or combined with any other antihypertensive drug that is not an ACE inhibitor (in fixed-dose combinations or in different medicinal products); aldosterone antagonists include current users of any stratified analyses we applied Bonferroni correction to antagonist of aldosterone, alone or combined with any other antihypertensive drug that is not an ACE inhibitor or allow for multiple testing, for which we set the statistical angiotensin-receptor blocker (in fixed-dose combinations or in different medicinal products); other antihypertensive significance level at a p value of less than 0·008 (0·05 drugs include calcium-channel blockers, diuretics, β-blockers, and α-blockers, alone or combined (excluding RAAS divided by 6—ie, the number of stratified analyses done). inhibitors; either in fixed-dose combinations or in different medicinal products). ACE=angiotensin-converting enzyme. RAAS=renin–angiotensin–aldosterone system. *Adjusted for sex and age. †Adjusted for the matching variables plus According to COVID-19 severity we differentiated two history of diabetes, dyslipidaemia, ischaemic heart disease, heart failure, atrial fibrillation, thromboembolic disease, groups of cases: fatal cases and those requiring admission cerebrovascular accident, chronic obstructive pulmonary disease, asthma, cancer, and chronic renal failure. to an ICU (ie, most severe), and less severe cases. In each Table 2: Risk of COVID-19 requiring admission to hospital and current use of RAAS inhibitors compared group, we analysed the association with RAAS inhibitors with current use of other antihypertensive drugs (main analysis) as described earlier and compared the adjusted OR across the strata using the Altman and Bland test of interaction.18 We planned four sensitivity analyses. First, an analysis Cases (%) Controls (%) Adjusted odds pinteraction for the correction for secular trends. The control series ratio (95% CI) was sourced from a 2018 population, the last available year in the BIFAP database, while cases were from 2020. Thus, Sex in case the prevalence of use of RAAS inhibitors increased Female 188 (42·3%) 1395 (31·4%) 1·01 (0·74−1·38) 0·54 Male 309 (44·5%) 2427 (34·9%) 0·89 (0·69−1·15) over time, the exposure among controls would be an Age (years) underestimate of 2020 data and, as a result, the adjusted <70 149 (29·9%) 906 (18·2%) 1·20 (0·78−1·84) 0·21 ORs could be overestimated. To assess the magnitude of ≥70 348 (54·3%) 2916 (45·5%) 0·88 (0·70−1·10) this potential problem, we examined trends of use of dif­ Hypertension ferent antihypertensive drug sub­ groups in the Madrid No 33 (6·3%) 336 (5·9%) 0·77 (0·43−1·37) 0·51 population in BIFAP over the period 2012–18 and forecast Yes 464 (75·2%) 3486 (61·8%) 0·95 (0·75−1·21) the result corresponding to 2020 via a linear regression Diabetes model. Then, we divided the prevalence of use estimate No 315 (38·0%) 2564 (28·2%) 1·09 (0·85−1·41) 0·004 for 2020 by the prevalence observed in 2018 for every Yes 182 (58·7%) 1258 (54·4%) 0·53 (0·34−0·80) antihypertensive drug class and took this ratio as an Background cardiovascular risk indicator of the overestimation of adjusted ORs. For RAAS Cardiovascular risk factors 300 (59·5%) 2509 (50·4%) 1·04 (0·76−1·41) 0·53 inhibitors, the ratio was 1·04 overall (1·05 for ACE Cardiovascular diseases 186 (59·6%) 1260 (52·4%) 0·90 (0·64−1·25) inhibitors and 1·01 for angiotensin-receptor blockers), Study period March 1–16 335 (46·7%) 2453 (34·2%) 0·98 (0·77−1·25) 0·62 whereas for other antihypertensive drugs, including March 17–24 162 (38·4%) 1369 (32·4%) 0·88 (0·62−1·24) calcium-channel blockers, the ratio was close to 1·00 (appendix pp 5–7). In the sensitivity analysis, we corrected 0·5 1·0 1·5 2·0 for these patterns of use, dividing the adjusted OR of Adjusted odds ratio (95% CI) RAAS inhibitors, ACE inhibitors, and angiotensin- receptor blockers by their respective estimated ratios Figure 2: Association between current use of RAAS inhibitors and risk of COVID-19 requiring admission to hospital compared with current use of other antihypertensive drugs, stratified by different variables 2020:2018. Second, we assessed the potential effect of Number (%) of cases and controls given is the number exposed to RAAS inhibitors in each stratum and the resulting media alerts. Concerns about the association between odds ratio, adjusted for the matching variables and other comorbidities different from the one examined. RAAS inhibitors and severe COVID-19 were highlighted RAAS=renin–angiotensin–aldosterone system. www.thelancet.com Vol 395 May 30, 2020 1709 Articles Finally, we included hypertension in our multivariable Cases Matched controls Adjusted odds (n=1139) (n=11 390) ratio* model. We did all statistical analyses using STATA/SE Most severe cases (version 15). n 393 3930 ·· Sex Role of the funding source Male 264/393 (67·2%) 2640/3930 (67·2%) ·· The funder of the study had no role in study design, data Female 129/393 (32·8%) 1290/3930 (32·8%) ·· collection, data analysis, data interpretation, or writing of Age, years 75·3 (12·3) 75·3 (12·3) ·· the report. The corresponding author had full access to Current use all the data in the study and had final responsibility for Other antihypertensive drugs 64/393 (16·3%) 484/3930 (12·3%) 1 (ref) the decision to submit for publication. RAAS inhibitors† 215/393 (54·7%) 1592/3930 (40·5%) 1·08 (0·80–1·47) ACE inhibitors 101/393 (25·7%) 905/3930 (23·0%) 0·92 (0·65–1·29) Results Angiotensin-receptor blockers 105/393 (26·7%) 655/3930 (16·7%) 1·25 (0·89–1·77) We collected data on 1139 cases and 11 390 matched Non-use 114/393 (29·0%) 1854/3930 (47·2%) 0·48 (0·34–0·69) controls. Despite being matched by age and sex, cases Less severe cases had a greater prevalence of diverse comor­bidities than n 746 7460 ·· did controls (table 1). Sex The prevalence of use of different subgroups of Male 431/746 (57·8%) 4310/7460 (57·8%) ·· antihypertensive drugs was higher in cases than in Female 315/746 (42·2%) 3150/7460 (42·2%) ·· controls, which yielded positive associations with risk of Age, years 65·8 (15·9) 65·8 (15·9) ·· COVID-19 requiring admission to hospital in both Current use crude and fully adjusted analyses compared with non- Other antihypertensive drugs 91/746 (12·2%) 645/7460 (8·7%) 1 (ref) use of any antihypertensive drug (adjusted OR for RAAS RAAS inhibitors† 282/746 (37·8%) 2230/7460 (29·9%) 0·86 (0·66–1·11) inhibitors 1·71, 95% CI 1·46–2·01; and for other ACE inhibitors 139/746 (18·6%) 1287/7460 (17·3%) 0·74 (0·56–0·99) antihypertensive drugs 1·82, 1·47–2·26; appendix p 9). Angiotensin-receptor blockers 132/746 (17·7%) 897/7460 (12·0%) 0·99 (0·74–1·33) When use of other antihypertensive drugs was set as the Non-use 373/746 (50·0%) 4585/7460 (61·5%) 0·57 (0·43–0·75) reference category, the adjusted OR associated with the Data are n (%), mean (SD), and odds ratio with 95% CI in parentheses. Most severe cases are those who died and those current use of RAAS inhibitors was 0·94 (0·77–1·15; admitted to an intensive care unit. Less severe cases are all other inpatients. The different pharmacological classes table 2). No significant increase in the risk of COVID-19 examined are mutually exclusive categories, so that patients who used combinations are counted only once, applying requiring admission to hospital was observed with the following criteria: ACE inhibitors include current users of any ACE inhibitors alone or combined with any other either ACE inhibitors or angiotensin-receptor blockers, antihypertensive drug (in fixed-dose combinations or in different medicinal products); angiotensin-receptor blockers include current users of any angiotensin-receptor blocker, alone or combined with any other antihypertensive drug that nor did we find any difference when these drugs were is not an ACE inhibitor (in fixed-dose combinations or in different medicinal products); and other antihypertensive used in monotherapy or in combination with other drugs include calcium-channel blockers, diuretics, β-blockers, and α-blockers, alone or combined (excluding RAAS drugs (table 2). inhibitors; either in fixed-dose combinations or in different medicinal products). ACE=angiotensin-converting enzyme. RAAS=renin–angiotensin–aldosterone system. *Adjusted for the matching variables plus history of diabetes, The adjusted ORs associated with other antihyper­ dyslipidaemia, ischaemic heart disease, heart failure, atrial fibrillation, thromboembolic disease, cerebrovascular tensive drugs compared with non-use by different accident, chronic obstructive pulmonary disease, asthma, cancer, and chronic renal failure. †Including ACE inhibitors, pharma­cological subgroups were 1·96 (95% CI 1·43–2·69) angiotensin-receptor blockers, aldosterone antagonists, and renin inhibitors. for calcium-channel blockers, 1·79 (1·32–2·43) for Table 3: Risk of COVID-19 requiring admission to hospital and current use of RAAS inhibitors compared diuretics, 1·68 (1·16–2·43) for β-blockers, and 1·96 with current use of other antihypertensive drugs, by severity of disease (0·57–6·71) for α-blockers (appendix p 9). When the current use of cal­cium-channel blockers was set as the reference category, the adjusted OR was 0·87 (0·65–1·18) in several letters to medical journals8–10 and news began to for RAAS inhibitors, 0·74 (0·54–1·02) for ACE inhibitors, appear in the Spanish press and social media networks and 1·02 (0·74–1·40) for angiotensin-receptor blockers after March 16, 2020, when the Spanish Agency for (appendix p 10). Medicines and Medical Devices published a note Compared with long-term (longer than 1 year) users recommending not to discontinue these drugs.15 Because of other antihypertensive drugs, long-term users of this news might have had an influence on adherence to RAAS inhibitors had an adjusted OR of 0·96 (95% CI recorded treatment, we stratified the results into two study 0·76–1·21) for risk of COVID-19 requiring admission to periods: March 1–16, and March 17–24. Third, we excluded hospital. Likewise, compared with short-term users (up aldosterone antagonists and renin inhibitors from the to 1 year) of other antihypertensive drugs, short-term RAAS inhibitor group. For the main analysis we included users of RAAS inhibitors had an adjusted OR of 1·39 aldosterone antagonists in the RAAS inhibitor group (95% CI 0·92–2·10). because these drugs have also been reported to upregulate The results of the assessment of the potential effect ACE2 expression;19 but we did a sensitivity analysis to modification of RAAS inhibitors by age (<70 and estimate the adjusted ORs among RAAS inhibitors ≥70 years), sex, diabetes, hypertension, and background excluding aldosterone antagonists and renin inhibitors. cardiovascular risk are shown in figure 2. No significant 1710 www.thelancet.com Vol 395 May 30, 2020 Articles interaction was observed with any variable, except for users. Sex, age, and background cardiovascular risk did diabetes, for which a significantly reduced risk of not significantly affect the results, although use of RAAS COVID-19 requiring admission to hospital associated inhibitors was associated with a reduced risk of with RAAS inhibitors was observed (adjusted OR 0·53; COVID-19 requiring admission to hospital in patients 0·34–0·80; test of interaction, p=0·004; still significant with diabetes. after applying the Bonferroni correction for multiple Our results do not support the hypothesis that previous testing [corrected significance level fixed at 0·008]). intake of RAAS inhibitors facilitates or increases the 393 (34·5%) of 1139 cases had the most severe form of severity of COVID-19. This hypothesis emerged when two COVID-19: 283 (24·8%) died in hospital and 110 (9·7%) facts were linked: the finding that RAAS inhibitors survived but required admission to an ICU. The mean upregulate the expression of ACE2,3–5,19 and the observation duration of hospital stay was 13 days (SD 8) among that hypertension, diabetes, and ischaemic heart disease survivors and 11 days (SD 7) among non-survivors. were highly prevalent among patients with severe Broadly, compared with patients with less severe COVID-19,9–12 conditions for which RAAS inhibitors are COVID-19 in our study, patients with the most severe widely used. However, these facts are only two pieces of a COVID-19 were older (mean age 75·3 years [SD 12·3] vs more complex puzzle. Evidence to support the idea that 65·8 years [SD 15·9]; p<0·0001) and a higher proportion ACE2 might have a dual role in COVID-19 is increasing.20 were male (264 [67·2%] of 393 with severe disease vs On the one hand, ACE2 overexpression might increase 431 [57·8%] of 746 with less severe disease; p=0·001; the susceptibility of cells to SARS-CoV-2, but, on the other table 3). Compared with current use of other anti­ hand, its downregulation associated with older age, male hypertensive drugs, the adjusted OR associated with the sex, and cardiovascular comorbidities, and further current use of RAAS inhibitors was 1·08 (95% CI heightened by SARS-CoV-2 binding and internalisation21,22 0·80–1·47) among the most severe cases and 0·86 could increase the unopposed action of angiotensin II and (0·66–1·11) among the less severe cases (table 3). In both have a key role in the subsequent organ injury.23 ACE2 severity groups, the 95% CIs for the adjusted ORs of ACE counteracts the deleterious effect of RAAS axis. In patients inhibi­tors and angiotensin-receptor blockers compared with COVID-19, Liu and colleagues24 reported that serum with other antihypertensive drugs overlapped with each concentrations of angiotensin II were signifi­cantly higher other, indicating no significant differences (table 3). in infected individuals than non-infected individuals In our first sensitivity analysis, the adjusted OR of and was linearly associated with viral load and lung current use of RAAS inhibitors compared with current damage. ACE2 inactivates angiotensin II and increases use of other antihypertensive drugs corrected for the generation of angiotensin 1–7, a peptide that, acting secular trends was 0·90 (95% CI 0·74–1·11), and 0·76 on the Mas receptor, exerts a vasodilatory effect and (0·61–0·95) for ACE inhibitors and 1·09 (0·87–1·36) for anti-inflammatory and antioxidative actions.20 In patients angiotensin-receptor blockers. In our second sensitivity with acute respiratory distress syndrome (ARDS), analysis, up to March 16, the adjusted OR associated Reddy and colleagues25 observed a higher ratio of with current use of RAAS inhibitors compared with angiotensin 1–7:angiotensin I among survivors than non- current use of other antihypertensive drugs was 0·98 survivors, which adds evidence to the idea that the counter- (95% CI 0·77–1·25) and for March 17 onwards was 0·88 regulation exerted by the axis of ACE2–angiotensin (0·62–1·24; figure 2). In our third sensitivity analysis, 1–7–Mas receptor benefits patients with ARDS. In line the exclusion of aldosterone antago­ nists and renin with this theory, recom­ binant soluble ACE2 has been inhibitors from the RAAS inhibitor group hardly had shown in animal models of ARDS to protect subjects any effect (overall adjusted OR 0·92, 95% CI 0·76–1·12; from lung injury,26–28 and clinical trials are underway in among the most severe cases, 1·06, 0·78–1·44; and patients with COVID-19 (NCT04287686). Finally, to our among the less severe cases, 0·84, 0·65–1·09; appendix knowledge, the case series of patients with COVID-19 pp 11–12). And in our final sensitivity analysis, the published to date has not been adjusted for important inclusion of hypertension in the multi­variable model potential confounding factors, such as sex, age, and cardio­ had little effect on the adjusted ORs for RAAS inhibitors vascular comorbidities.29 In our study, we used controls (1·00, 0·82–1·23), ACE inhibitors (0·85, 0·68–1·06), or who were matched to cases for sex and age, and we also angiotensin-receptor blockers (1·17, 0·93–1·47). adjusted for a number of comorbidities. To our knowledge, two epidemiological studies have Discussion been published to date that aimed to explore the Here we show that the current use of RAAS inhibitors is association between RAAS inhibitors and COVID-19 not associated with an increased risk of COVID-19 comparing severe cases with less severe cases among requiring admission to hospital (including fatal cases inpatients.30,31 Neither of these studies found an increased and those admitted to an ICU) compared with other risk of severe outcomes associated with these drugs, and antihypertensive drugs. No substantial difference was in one study30 a substantially reduced risk of death and observed between ACE inhibitors and angiotensin- transfer to a critical care unit within 7 days of admission receptor blockers, nor among short-term and long-term to hospital was observed among users of ACE inhibitors www.thelancet.com Vol 395 May 30, 2020 1711 Articles (OR 0·29, 95% CI 0·10–0·75). Notably, in both studies, in both patients with most severe and less severe disease, the authors considered not only exposure to ACE suggesting that the outpatient use of these drugs neither inhibitors before (7 days) or at admission to hospital, but facilitates nor aggravates the infection. also during hospital stay. In our study we used population Our study has several limitations. First, we used controls, not just a case series, and only considered drugs different data sources to extract information from cases used before admission. (hospital records and primary health-care records through Several cardiovascular comorbidities are clearly linked to HORUS, consulted case by case by local researchers) and severe COVID-19, and RAAS inhibitors are often used to controls (BIFAP database, which automatically extracted treat such cardiovascular conditions; thus substantial data from primary health-care records); however, both confounding by indication can be expected.32 In our view, HORUS and BIFAP access the same primary health-care this bias explains the positive association we found data, which was the main source for determining between current use of any antihypertensive drug and risk comorbidity and drug exposure in our study, making of COVID-19 requiring admission to hospital compared major information bias unlikely. Second, cases were with non-use of such drugs, which did not disappear after recorded as of March, 2020, whereas controls were full adjustment for comorbidities. This problem can be sourced from records in March, 2018, which might have mitigated if current users of RAAS inhibitors are compared affected the results due to secular trends in use of with current users of other drugs that share, at least antihypertensive drugs; however, when we corrected for partially, their indications, but not their safety issues.32 For secular trends we saw little change in the estimated instance, some researchers have recommended calcium- adjusted OR. Third, we did not collect data on smoking channel blockers as a suitable alternative treatment to and other lifestyle habits (eg, exercise, diet, alcohol intake), RAAS inhibitors, because calcium-channel blockers have which might have affected the association between RAAS not been reported to increase the expression of ACE2.8 inhibitors and COVID-19 requiring admission to hos­ Following this reasoning, other antihypertensive drugs in pital; however, these habits are correlated with several general, and calcium-channel blockers in particular, can be comorbidities that were included in the regression model. considered a more valid comparator group than non-use of Fourth, we obtained information on drug prescriptions, such drugs. We followed this approach and reassuringly but adherence to treatment cannot be guaranteed. Hence, we found no increased risk when using either of these when we have used the term drug use, we should say drug drug groups as the reference group. Nevertheless, RAAS prescribed. The effect of non-adherence in our results is inhibitors do not share all indications with other difficult to measure but would most probably cause a non- antihypertensive drugs; hence, the comorbidity pattern differential misclassification of exposure that, if relevant, associated with use of these drugs might not overlap and would distort the estimates towards the null value. Fifth, some residual confounding by indication could remain. To we did not collect information on clinical and analytical explore the extent of this possibility, we examined the covariates during patients’ stay in hospital, nor did we comorbidity pattern associated with various types of collect information on in-hospital treatment (eg, if RAAS antihypertensive drugs among controls and found much inhibitors were continued or withheld in hospital), which fewer differences among the types of drug than when they might have been interesting for the assessment of the were compared with non-users, which reinforces the association of antihypertensive drugs with severity of and validity of our approach (appendix p 13). fatality due to COVID-19. Sixth, we did not analyse any The risk reduction associated with RAAS inhibitors dose effects. Finally, as in any other observational study, among individuals with diabetes who had been residual confounding due to unmea­sured or unknown admitted to hospital for COVID-19 (not explained by confounders cannot be ruled out. chance) deserves a comment. Diabetes has been In summary, the results of the present study suggest reported as a risk factor for severe COVID-19.33 The that the outpatient use of RAAS inhibitors does not biological underpinning for this clinical observation is increase the risk of COVID-19 requiring admission to not yet known, but experimental models of diabetes in hospital, including its most severe forms. This finding mice have shown that ACE activity is high in the lungs.34 should be confirmed using other data sources, study If this activity also occurs in humans, patients with designs, and populations. Meanwhile, the data available, diabetes who have COVID-19 might present a great along with the important role of ACE inhibitors and imbalance in the ACE:ACE2 ratio (ACE activity would angiotensin-receptor blockers in the management of be high due to diabetes and ACE2 would be low due to several cardiovascular diseases, do not support their downregula­tion induced by SARS-CoV-2) that could discontinuation as a preventive measure against explain the increased severity of COVID-19 among COVID-19. patients with diabetes and also the protective effect of Contributors RAAS inhibitors suggested by our data. FJdA, DR-P, and SR-M contributed to conceptualisation and study The lack of an increased risk of COVID-19 requiring design. FJdA, SR-M, MG, and AR-M contributed to the methods. VL, SR-M, GM-A, MA, AG-L, LL, OL, GAC-S, MAG, MP, EG-R, LP, IdP, admission to hospital associated with RAAS inhibitors FA-S, LR-M, MG, and DR-P contributed to data collection. SR-M, FJdA, (compared with other antihypertensive drugs) was found AR-M, MG, and AT contributed to data analysis. All authors contributed 1712 www.thelancet.com Vol 395 May 30, 2020 Articles to data interpretation. FJdA, SR-M, and AR-M wrote the first draft of the 7 Sommerstein R. Preventing a covid-19 pandemic. BMJ 2020; manuscript. All authors contributed to critical revision of the manuscript 368: m810. for important intellectual content and gave final approval. VL contributed 8 Fang L, Karakiulakis G, Roth M. Are patients with hypertension and to revision of English language. FJdA applied for funding. All authors are diabetes mellitus at increased risk for COVID-19 infection? accountable for all aspects of the work in ensuring that questions related Lancet Respir Med 2020; 8: e21. to the accuracy or integrity of any part of the work are appropriately 9 Yang X, Yu Y, Xu J, et al. Clinical course and outcomes of critically investigated and resolved. ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study. MED-ACE2-COVID19 study group Lancet Respir Med 2020; published online Feb 24. https://doi.org/ Spain: F J de Abajo, S Rodríguez-Martín, V Lerma, A Rodríguez-Miguel, 10.1016/S2213-2600(20)30079-5. D Rodríguez-Puyol, D Barreira-Hernández (Hospital Universitario 10 Guan W, Ni Z, Hu Y, et al. Clinical characteristics of coronavirus Príncipe de Asturias, Madrid); G Mejía-Abril, F Abad-Santos, P Zubiaur, disease 2019 in China. N Engl J Med 2020; 382: 1708–20. E Santos-Molina, E Pintos-Sánchez, M Navares-Gómez (Hospital 11 Zhang JJ, Dong X, Cao YY, et al. Clinical characteristics of Universitario de La Princesa, Madrid); A García-Luque, M Puerro, 140 patients infected by SARS-CoV-2 in Wuhan, China. Allergy 2020; R M Aparicio, V García-Rosado, C Gutiérrez-Ortega (Hospital Central de published online Feb 19. DOI:10.1111/all.14238. la Defensa Gómez Ulla, Madrid); L Laredo, E González-Rojano, 12 Zhou F, Yu T, Du R, et al. Clinical course and risk factors for C Pérez, A Ascaso, C Elvira (Hospital Clínico San Carlos, Madrid); mortality of adult inpatients with COVID-19 in Wuhan, China: M Aguilar, M A Gálvez, I de Pablo (Hospital Universitario Ramón y Cajal, a retrospective cohort study. Lancet 2020; 395: 1054–62. Madrid); O Laosa, L Pedraza, L Rodríguez-Mañas (Hospital Universitario 13 Gurwitz D. Angiotensin receptor blockers as tentative SARS-COV-2 de Getafe, Madrid); G A Centeno-Soto, B Ruiz-Antorán (Hospital therapeutics. Drug Dev Res 2020; published online March 4. Universitario Puerta de Hierro Majadahonda, Madrid); M Gil (Agencia DOI:10.1002/ddr.21656. Española de Medicamentos y Productos Sanitarios, Madrid); and A Tobías 14 Vaduganathan M, Vardeny O, Michel T, McMurray JJV, Pfeffer MA, (Instituto de Diagnóstico Ambiental y Estudios del Agua, CSIC Barcelona). Solomon SD. Renin-angiotensin-aldosterone system inhibitors in patients with Covid-19. N Engl J Med 2020; 382: 1653–59. Declaration of interests 15 Medicamentos antihipertensivos que actúan sobre el sistema We declare no competing interests. renina-angiotensina e infeccción por COVID-19. Agencia Española Data sharing de Medicamentos y Productos Sanitarios, March 16, 2020. https://www.aemps.gob.es/informa/notasinformativas/ After publication, study data will be made available on reasonable medicamentosusohumano-3/seguridad-1/2020-seguridad-1/ request to the corresponding author. A proposal with a detailed medicamentos-antihipertensivos-que-actuan-sobre-el-sistema-renina- description of study objectives and statistical analysis plan will be needed angiotensina-e-infeccion-por-covid-19/ (accessed March 25, 2020). for assessment of requests. Additional materials might also be required 16 Rodríguez-Martín S, Martín-Merino E, Lerma V, et al. during the process of assessment. De-identified participant data will be Active surveillance of severe cutaneous adverse reactions: provided after approval from the principal researchers of the a case-population approach using a registry and a health care participating hospitals. database. Pharmacoepidemiol Drug Saf 2018; 27: 1042–50. Acknowledgments 17 Rothman KJ, Greenland S, Lash TL. Case-control studies. In: Rothman KJ, Greenland S, Lash TL, eds. Modern epidemiology, This study was funded by a research grant from the Institute of Health 3rd edn. Philadelphia, PA: Wolters Kluwer/Lippincott Williams & Carlos III (COV20/00027). We thank the health professionals, patients, Wilkins, 2008: 111–27. and administrative staff from the hospitals that participated in this study, 18 Altman DG, Bland JM. Interaction revisited: the difference between and the primary care practitioners and staff of BIFAP. We also thank two estimates. BMJ 2003; 326: 219. Arturo Álvarez, for his diligence and skill in extracting the information 19 Keidar S, Gamliel-Lazarovich A, Kaplan M, et al. from population controls, and Dolores Montero and Miguel A Maciá. Mineralocorticoid receptor blocker increases angiotensin- We also thank María José Bleda, Luis A García Rodríguez, and converting enzyme 2 activity in congestive heart failure patients. Alberto García-Lledó who provided helpful suggestions and comments Circ Res 2005; 97: 946–53. during preparation of the protocol, statistical analysis plan, and 20 Guo J, Huang Z, Lin L, Jv J. Coronavirus disease 2019 (COVID-19) manuscript. This Article is dedicated to all researchers’ relatives and and cardiovascular disease: a viewpoint on the potential influence of friends who have died due to COVID-19; they gave us moral strength to angiotensin-converting enzyme inhibitors/angiotensin receptor complete this research. The results, discussion, and conclusions in this blockers on onset and severity of severe acute respiratory syndrome Article are those of the authors and do not necessarily represent the coronavirus 2 infection. J Am Heart Assoc 2020; 9: e016219. position of their Institutions or the Spanish Agency for Medicines and 21 Kuba K, Imai Y, Rao S, et al. A crucial role of angiotensin converting Medical Devices. enzyme 2 (ACE2) in SARS coronavirus-induced lung injury. Nat Med 2005; 11: 875–79. References 22 Xie X, Chen J, Wang X, Zhang F, Liu Y. Age- and gender-related 1 Wan Y, Shang J, Graham R, Baric RS, Li F. Receptor recognition by difference of ACE2 expression in rat lung. Life Sci 2006; novel coronavirus from Wuhan: an analysis based on decade-long 78: 2166–71. structural studies of SARS coronavirus. J Virol 2020; 94: e00127–20. 23 Cheng H, Wang Y, Wang GQ. Organ-protective effect of 2 Li XC, Zhang J, Zhuo JL. The vasoprotective axes of the angiotensin-converting enzyme 2 and its effect on the prognosis of renin-angiotensin system: physiological relevance and therapeutic COVID-19. J Med Virol 2020; published online March 27. implications in cardiovascular, hypertensive and kidney diseases. DOI:10.1002/jmv.25785. Pharmacol Res 2017; 125 (PtA): 21–38. 24 Liu Y, Yang Y, Zhang C, et al. Clinical and biochemical indexes from 3 Ferrario CM, Jessup J, Chappell MC, et al. Effect of angiotensin- 2019-nCoV infected patients linked to viral loads and lung injury. converting enzyme inhibition and angiotensin II receptor blockers Sci China Life Sci 2020; 63: 364–74. on cardiac angiotensin-converting enzyme 2. Circulation 2005; 111: 2605–10. 25 Reddy R, Asante I, Liu S, et al. Circulating angiotensin peptides levels in acute respiratory distress syndrome correlate with clinical 4 Ocaranza MP, Godoy I, Jalil JE, et al. Enalapril attenuates outcomes: a pilot study. PLoS One 2019; 14: e0213096. downregulation of angiotensin-converting enzyme 2 in the late phase of ventricular dysfunction in myocardial infarcted rats. 26 Imai Y, Kuba K, Rao S, et al. Angiotensin-converting enzyme 2 Hypertension 2006; 48: 572–78. protects from severe acute lung failure. Nature 2005; 436: 112–16. 5 Soler MJ, Ye M, Wysocki J, William J, Lloveras J, Batlle D. 27 Zou Z, Yan Y, Shu Y, et al. Angiotensin-converting enzyme 2 Localization of ACE2 in the renal vasculature: amplification by protects from lethal avian influenza A H5N1 infections. angiotensin II type 1 receptor blockade using telmisartan. Nat Commun 2014; 5: 3594. Am J Physiol Renal Physiol 2009; 296: F398–405. 28 Yang P, Gu H, Zhao Z, et al. Angiotensin-converting enzyme 2 6 O’Mara G. Could ACE inhibitors and particularly ARBs increase (ACE2) mediates influenza H7N9 virus-induced acute lung injury. susceptibility to COVID-19 infection? BMJ 2020; 368: m406. Sci Rep 2014; 4: 7027. www.thelancet.com Vol 395 May 30, 2020 1713 Articles 29 Sommerstein R, Kochen MM, Messerli FH, Gräni C. Coronavirus 32 Psaty BM, Koepsell TD, Lin D, et al. Assessment and control for disease 2019 (COVID-19): do angiotensin-converting enzyme confounding by indication in observational studies. inhibitors/angiotensin receptor blockers have a biphasic effect? J Am Geriatr Soc 1999; 47: 749–54. J Am Heart Assoc 2020; 9: e016509. 33 Guo W, Li M, Dong Y, et al. Diabetes as a risk factor for the 30 Bean D, Kraljevic Z, Searle T, et al. Treatment with ACE-inhibitors progression and prognosis of COVID-19. Diabetes Metab Res Rev is associated with less severe disease with SARS-Covid-19 infection 2020; published online March 31. DOI:10.1002/dmrr.3319. in a multi-site UK acute hospital trust. medRxiv 2020; published 34 Roca-Ho H, Riera M, Palau V, Pascual J, Soler MJ. Characterization online April 11. DOI:10.1101/2020.04.07.20056788 (preprint). of ACE and ACE2 expression within different organs of the NOD 31 Li J, Wang Z, Chen J, Zhang H, Deng A. Association of renin- mouse. Int J Mol Sci 2017; 18: E563. angiotensin system inhibitors with severity or risk of death in patients with hypertension hospitalized for coronavirus disease 2019 (COVID-19) infection in Wuhan, China. JAMA Cardiology 2020; published online April 23. DOI:10.1001/jamacardio.2020.1624. 1714 www.thelancet.com Vol 395 May 30, 2020