key: cord-1051483-jy9k90wl authors: Soler, Maria José; Ribera, Aida; Marsal, Josep R; Mendez, Ana Belen; Andres, Mireia; Azancot, Maria Antonia; Oristrell, Gerard; Méndez-Boo, Leonardo; Cohen, Jordana; Barrabés, Jose A; Ferreira-González, Ignacio title: Association of renin–angiotensin system blockers with COVID-19 diagnosis and prognosis in patients with hypertension: a population-based study date: 2021-09-03 journal: Clin Kidney J DOI: 10.1093/ckj/sfab161 sha: a29c75935a83090a12362acd04ade50c8e41608e doc_id: 1051483 cord_uid: jy9k90wl BACKGROUND: The effect of renin-angiotensin(RAS) blockade either by angiotensin-converting enzyme inhibitors (ACEi) or angiotensin-receptor blockers (ARBs) on coronavirus disease 2019(COVID-19) susceptibility, mortality and severity is inadequately described. We examined the association between renin-angiotensin system (RAS) blockade and COVID-19 diagnosis and prognosis in a large population-based cohort of patients with hypertension. METHODS: This is a cohort study using regional health records. We identified all individuals aged 18-95 years from 87 health care reference areas of the main health provider in Catalonia(Spain), with a history of hypertension from primary care records. Data were linked to COVID-19 test results, hospital, pharmacy and mortality records from 1 March 2020 to 14 August 2020. We defined exposure to RAS blockers as the dispensation of ACEi/ARBs during the three months before COVID-19 diagnosis or 1 March 2020. Primary outcomes were: COVID-19 infection, and severe progression in hospitalized patients with COVID-19(the composite of need for invasive respiratory support or death). For both outcomes and for each exposure of interest (RAAS blockade, ACEi or ARB) we estimated associations in age-sex-area-propensity matched samples. RESULTS: From a cohort of 1,365,215 inhabitants we identified 305,972 patients with hypertension history. Recent use of ACEi/ARBs in patients with hypertension was associated with a lower 6 month-cumulative incidence of COVID-19 diagnosis (3.78% [95% CI: 3.69% - 3.86%] vs 4.53% [95% CI: 4.40% - 4.65%]; p < 0.001). In the 12,344 patients with COVID-19 infection, the use of ACEi/ARBs was not associated with a higher risk of hospitalization with need for invasive respiratory support or death (OR = 0.91 (0.71 – 1.15); p = 0.426). CONCLUSION: RAS blockade in patients with hypertension is not associated with higher risk of COVID-19 infection or with a worse progression of the disease. There has been a growing interest and speculation regarding the effect of the widely used classes of drugs that inhibit the renin-angiotensin-aldosterone system (RAS) on the novel coronavirus (severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)) infection or coronavirus disease 2019 (COVID-19) severity from early 2020 [1] . Angiotensin-converting enzyme (ACE) 2 is a homolog of ACE, with 40% identity, which was discovered in 2000 [2] [3] [4] . Due to its important role as the receptor of SARS-CoV-2, recent studies have been focused on the potential implication of ACE2 in COVID-19 disease [5] [6] [7] [8] . Prior data suggested that RAAS blockade with ACE inhibitors (ACEi) or angiotensin II receptor blockers (ARBs) may increase ACE2 expression or activity, though these data are mixed. Accordingly, during the last several months, various observational studies and clinical trials have been designed and conducted to address the effect of RAS blockade on COVID-19 severity [9] [10] [11] [12] [13] . Most studies that have evaluated the association between RAS blockade with the risk of development and severity of COVID-19, suggest a neutral effect of this therapy [14] [15] [16] . Yet, to assess the association between RAS blockade and the risk of diagnosis of COVID-19, the majority of published studies have a used case-control design [14] [15] [16] . This design could bias to some extent risk estimates, especially if COVID-19 cases requiring hospitalization are overrepresented [11, 17] . Furthermore, to assess whether ACEi or ARBs are associated with worse outcomes in COVID-19 infected patients, most of the published studies included exclusively hospitalized patients, which constitutes a highly selected population, especially during the massive stroke of pandemic [18] [19] [20] [21] . This is especially true for the subpopulations with high cardiovascular risk profiles. To overcome potential selection bias in this context, the ideal approach would be to use a closed community population-based cohort. In this sense, although there are several cohort studies carried out in different populations, all of them have selected the final study population in ways 4 that could limit the generalizability of their findings [12, 16, 20 ]. In the present study, we used a large, unselected community cohort of 305,972 patients with a history of hypertension to assess whether chronic treatment with RAS inhibitors is associated with an increased risk of COVID-19 diagnosis or with worse outcomes in infected patients. We performed a cohort study using prospectively collected data stored in the Information Systems of the Institut Català de la Salut (ICS) in Catalonia, Spain. ICS is the main health provider of the Catalan health system, providing primary care to a representative >6.7 million people (>88% of the population in Catalonia), and hospital care, to over 1.5 million. The cohort data were obtained from primary-care electronic health records (ECAP). ECAP Registries of people with a diagnostic code related to hypertension were linked to the official repository of reverse transcriptase polymerase chain reaction (RT-PCR) tests for SARS-CoV-2), to hospital admissions, intensive care units (ICU) and mortality registries. The Vall d'Hebron Ethics Committee approved the study protocol, including a waiver for the informed consent of patients taking part in the study, and the data extracted were fully anonymized. 5 We included all individuals in all health care areas with primary and hospital care provided by ICS, aged 18 or older and younger than 95 years with a diagnosis of hypertension up to March 1 st 2020. For the first main study objective, the association of RAS inhibitors use with risk of COVID-19 diagnosis, participants were followed from 1 March 2020 to the earliest of a first positive RT-PCR test or clinical diagnosis or death or the end of the study period (14 August 2020). COVID-19 was identified by a positive RT-PCR test for SARS-CoV-2 (confirmed cases) and/or a clinical diagnosis (probable cases) recorded in primary care or hospital diagnoses, or in death certificates. ICD-10 CM codes used are listed in Supplementary Table 1 . For the second main study objective, the association of RAS inhibitors use with risk of worse outcomes among infected patients, patients with a diagnosis of COVID-19 were followed from the earliest COVID-19 diagnosis (confirmed or probable) until death or the end of the study period (14 August, 2020). The merged data allowed us to assess different key events throughout the progression of COVID-19: hospitalization, admission to the intensive care unit, need for respiratory support, need for invasive hemodynamic support and COVID-19 mortality. Our primary endpoint for this second objective was need for invasive organ/respiratory support or COVID-19 mortality (see the combinations of ICD-10 CM procedure codes for the need for invasive respiratory support). We assessed baseline characteristics and comorbidities on March 1st 2020, including: sex, age (in years), rurality (rural, urban), socioeconomic status (MEDEA [Mortalidad en áreas pequeñas Españolas y Desigualdades Socioeconómicas y Ambientales] deprivation Index) and pre-existing comorbidities. Rural areas were defined as areas with less than 10,000 inhabitants and a population density below 150 inhabitants/km 2 . The validated MEDEA deprivation index is calculated by census tract level in urban areas, categorized in quartiles, where the first and 6 fourth quartiles are the least and most deprived areas, respectively [24] . Comorbidities were defined as the presence of a diagnosis code recorded any time before the index date and still active on March 1st for a pre-specified list of conditions. Lists of ICD-10-CM codes for each of these conditions are provided in Supplementary Table 2 . We identified major classes of antihypertensive agents (ACEis, ARBs, calcium-channel blockers, diuretics, and beta-blockers) that were dispensed to patients in the three months prior to the date of COVID infection, or 1 March 2020 in the non-infected (see the list of ATC codes of included pharmacy products in Supplementary Table 3) . We also identified other drugs including lipid-lowering agents, oral antidiabetic agents, insulin, antiplatelet agents, antiarrhythmic agents, anticoagulant agents, digitalis, nitrates, inhaled glucocorticoids, nonsteroidal antiinflammatory drugs (NSAIDs), immunosuppressive agents, short-acting β-agonists, long-acting β-agonists, and other agents used for chronic respiratory diseases. The exposure of interest was defined as the dispensation of ACEis or ARBs at least once during the three months before COVID-19 diagnosis or before the start of the first pandemic wave (1 March 2020). Patients with dispensation of both agents represented only 0.4% (n=1,532) of the cohort and were excluded. We report descriptive statistics including means and standard deviations (SD) for continuous variables, and frequencies and percentages for binary and categorical variables. Differences in means were assessed using the Student's t-test and differences in percentages were assessed using the Chi square test. The association between use of RAS inhibitors and risk of COVID-19 diagnosis was assessed using mixed effects Cox regression analysis, and the association between the use of RAS inhibitors and outcomes in infected patients was assessed using mixed effects logistic models. We used 7 random slope models and a non-structured covariance matrix for random effects. In both analyses we adjusted for the effect of clustered structure of data among patients cared for by different health providers located in different geographic locations (i.e. health care areas; n=89) that might differ in socio-economic status and in the level of exposure to the virus. We also adjusted all analyses by the estimated prevalence of COViD-19 at each health care area as an indicator of COVID-19 impact in the area or a surrogate marker of the probability of each individual living in each area to be exposed to the virus. To minimize confounding, all comparisons performed were matched by age, sex, health care area and the propensity score for the corresponding exposure. Propensity scores were estimated using logistic regression and they included the variables age, gender, MEDEA deprivation index, diabetes, cardiovascular disease history (i. e. myocardial infarction, angina, stroke, heart failure, peripheral vascular disease), asthma/COPD history, dementia, hypercholesterolemia, atrial fibrillation, cancer, renal impairment, obesity (diagnosis or BMI>30kg/m 2 ) and the dispensed medications. Each patient with the exposure of interest was 1:1 matched without replacement to the non-treated patient of the same age, sex, and health care area with the lowest propensity score difference. Balance in patient characteristics between exposure groups in matched data was assessed using standardized differences (Supplemental Tables 4, 5, 6). Outcomes between matched cohorts were compared with conditional logistic regression models. We performed further analyses within the matched cohorts to adjust for those variables with standardized differences higher than 10%. All analyses were performed using the statistical package RStudio. From a reference population of 1,365,215 inhabitants, ≥18 years old, living in areas with both primary and hospital care delivered by ICS centers (a total of 87 health areas) we identified 305,972 patients with hypertension (Fig 1) . Among these, 201,131 (66%) were dispensed RAAS were on an ACEi and 73,599 (24%) were on an ARB. The cumulative incidence of COVID-19 infection during the first six months of the pandemic in this population was 4.03% (95% CI: 3.96 -4.10). Table 1 shows the baseline differences between RAS inhibitors users vs non-users and, within the RAS inhibitors group, between ACEi and ARB users. Compared to non-users, RAS inhibitor users were older, mostly male, with higher deprivation indexes, higher prevalence of obesity, hypertension and hypercholesterolemia, and lower prevalence of diabetes. Other comorbidities were in general also more common in patients on RAS inhibitors, although stroke history and dementia were less frequent in these patients. Similarly, RAS inhibitor users were prescribed more other drugs on average (mean total number of other drugs 2.16 ± 1.87 vs 1.55 ± 1.8, p<0.001), except for NSAIDs. Patients lived in health care areas with a COVID-19 impact fairly similar in both groups, although with a trend of higher cumulative incidence in the group of patients on RAS inhibitors. Differences were smaller when comparing ACEi with ARB users, although the latter tended to have a higher prevalence of baseline risk factors, comorbidities and concomitant therapies. The cumulative incidence of COVID-19 diagnosis at 6 months (table 3) was lower in RAS inhibitor users compared to non-users (3.8% [95% CI: 3.7% -3.8%] vs 4.53% [95% CI: 4.4% -4.7%]; p<0.001). Within the RAS inhibitors group, cumulative incidence was slightly higher in patients on ARBs (3.9%, 95% CI: 3.8% -4.0%) than in those on ACEis (3.7%, 95% CI: 3.6% -3.8%); p=0.036). Baseline covariates between RAS inhibitor users vs non-users were well-balanced in 93,662 pairs of age-sex-area-propensity-matched patients (Supplemental Table 4 ). In these matched cohorts, cumulative incidence of COVID-19 diagnosis ( of ACEi users vs non-users (n=60,964 pairs) and ARB users vs non-users (n= 32,698 pairs) covariates were also well balanced (Supplemental Tables 5 and 6 ). In these matched cohorts, cumulative incidence of COVID-19 diagnosis (table 3) was also lower in those patients on ACEis and in those patients on ARBs comparing their respective non-user pairs: 1.7% (95% CI 1.6% -1.8%) vs 1.8% (95% CI: 1.7% -1.9%) (p<=0.011) in the ACEi cohort and 2.2% (95% CI: 2.0% -2.4%) vs 2.6% (95% CI: 2.4% -2.7%) (p=0.023) in the ARBs cohort. Figure 2 shows the subgroup analyses for each therapy stratified by age, sex, diabetes, hypertension, cardiovascular disease, obesity, number of concomitant treatments and the COVID-19 impact in the corresponding health care areas. Association of RAS inhibitors with a lower cumulative incidence of COVID-19 diagnosis was consistent across the different subgroups. Table 2 shows the baseline differences in the 12,344 COVID-19 infected patients between RAS inhibitor users (n=7,598) and non-users (n=4,746) and, within the RAS inhibitor group, between ACEi (n=4,731) and ARB (n=2,867) users. In general, differences observed in the whole study population were also reproducible in infected patients, although without statistically significant differences in angina, cancer, and concomitant treatment with digitalis, loop diuretics, vitamin K, nitrates and steroids. As in the whole population group, differences between ACEi and ARBs users were smaller. Table 3 shows the cumulative incidence of main outcomes according to RAS inhibitors therapy. Hospitalization rates, need for intensive care, and the need for invasive respiratory support were higher in RAS inhibitors users vs non-users. By contrast, the death rate was lower in RAS inhibitors users (4.8% [95% CI: 4.4% -5.3%] vs 7.2% [95% CI: 6.5% -8.0%]; p<0.001). The same pattern was observed in the composite endpoint of death or need for invasive respiratory support (7.3% [95% CI: 6.7% -7.90]) in RAS inhibitor users vs (8.7% [95% CI: 7.9% -9.5%]) in non-users, p=0.006. Within the RAS inhibitor group, higher rate of 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 10 hospitalization, death, and in the composite endpoint death or need for invasive respiratory support was observed in in patients on ARBs compared with patients on ACEis. Table 4 shows the crude and adjusted effects of RAS (ACEi and ARB) use on main outcomes. In the age-sex-area-propensity-matched analyses, baseline differences between patients on RAS inhibitors (ACEi users n=1295, ARB users n=728, any RAS inhibitor n=2023) were well-balanced (Supplemental Tables 7, 8 ). In the matched cohorts there were no differences between RAS inhibitor users and non-users regarding hospitalization, death, need for invasive respiratory support and the composite of death or need for invasive respiratory support (Table 4) . Interestingly, the death rate was lower in ACEi users as compared with non-ACEi users (4.40% [95% CI: 3.28% -5.52%] vs 6.25% [95% CI: 4.94% -7.57%]; p=0.044). Subgroup analyses for the risk of severe COVID-19 disease showed consistent results across the different subgroups ( Figure 3 ). In this large population-based, unselected cohort of 305,972 patients with hypertension, recent use of ACEis or ARBs was associated with a lower 6-month-cumulative incidence of COVID-19 diagnosis. In addition, in the 12,344 patients with COVID-19 infection, the use of ACEis or ARBs was not associated with a higher risk of hospitalization, need for invasive respiratory support, or death. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 11 speculation motivated researchers to investigate this hypothesis, while medical societies urged against withdrawal of RAS blockade in patients with or at risk of cardiac or renal disease in the COVID-19 pandemic era [30] . The primary function of ACE2 is to counterbalance the actions of ACE. Thus, whereas ACE degrades angiotensin I to Angiotensin II, ACE2 degrades angiotensin II into angiotensin (1-7), exerting counterregulation of the renin-angiotensin system [31] . Circulating ACE2 activity is increased in patients with increased cardiovascular risk characteristics such as male sex, heart failure, myocardial infarction and diabetes[28, [32] [33] [34] . Some authors postulate that this increase in circulate ACE2 may be in part due to a compensatory mechanism, where in conditions with high cardiovascular risk the excess of angiotensin II accumulation may lead to a circulating rise of ACE2 to degrade angiotensin II to angiotensin-(1-7) [28, 35] The potential harm related to RAS blockade in patients with COVID-19 infection has now been widely studied. In the last 6 months, more than 15 observational studies have been focused in the effect of prior RAS blockade on the severity of COVID-19 disease [36, 37] . Consistently, no association between ACEi or ARB use and illness severity has been found, in studies performed across several continents [36] . However, most of these studies were performed exclusively in hospitalized patients, which are not necessarily representative of the whole spectrum of COVID-19 patients, or used case-control designs, which are prone to several sources of bias[12,14- There are several cohort studies that have been conducted in community populations cohorts. Our study performed in an unselected cohort of 305,972 patients with hypertension has demonstrated, using a properly balanced 93,662 age-sex-area-propensity matched pairs, that prior treatment with RAS inhibitors in this high cardiovascular risk cohort is associated with a lower frequency of COVID-19 diagnosis. Whether this association is related to the effect of RAS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 13 blockade, or the treatment with RAS inhibitors is merely a marker of other unknown factors should be elucidated in a randomized clinical trial. In any case, this finding is in concordance with a recently published meta-analysis that included three studies with a total of 8,766 patients with COVID-19, showing that ACEis or ARBs were not associated with a higher likelihood of positive SARS-CoV-2 test results in symptomatic patients [36] . In agreement with the majority of previous observational studies, in our large cohort in hypertensive patients from Catalonia (Spain) we have been able to replicate the same results and demonstrated that there was no association between ACEis or ARBs and negative outcomes such as hospitalization, total mortality, need for invasive organ/respiratory support and the composite of total mortality or need for invasive organ/respiratory support in COVID-19 infected 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 hypertension, diabetes or cardiovascular disease, RAS discontinuation may play a deleterious effect on cardio and renoprotection. Limitations of our work include the retrospective and observational study design that lack of the causal interpretation provided by interventional studies and randomized clinical trials. Another limitation of our study is that the number of people tested for COVID-19 varied over time during the study period. At the beginning, a larger amount of people with possible COVID-19 were not tested, especially those with no or mild symptoms and who were not hospitalized. In conclusion, the present study provides evidence that the use of RAS blockade in patients with hypertension does not increase the risk of COVID-19 diagnosis or of disease severity among patients with COVID-19. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Table 3 . Cumulative incidence of COVID-19 diagnosis and incidence of main outcomes in the COVID-19 infected, in the whole study sample and in the age-sex-area-propensity-matched samples 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Figure 2B 270x270mm (300 x 300 DPI) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Figure 2C 270x270mm (300 x 300 DPI) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Figure 3B 270x270mm (300 x 300 DPI) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Figure 3C 270x270mm (300 x 300 DPI) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Sound science before quick judgement regarding RAS blockade in COVID-19 A novel angiotensin-converting enzyme-related carboxypeptidase (ACE2) converts angiotensin I to angiotensin 1-9 Angiotensinconverting enzyme 2 is an essential regulator of heart function Angiotensin-converting enzyme 2 and the kidney ACE2 and diabetes: ACE of ACEs? Detection of soluble angiotensin-converting enzyme 2 in heart failure: insights into the endogenous counterregulatory pathway of the renin-angiotensin-aldosterone system Role of circulating angiotensin converting enzyme 2 in left ventricular remodeling following myocardial infarction: a prospective controlled study Circulating ACE2 activity is increased in patients with type 1 diabetes and vascular complications Circulating angiotensin converting enzyme 2 activity as a biomarker of silent atherosclerosis in patients with chronic kidney disease Risks and Impact of Angiotensin-Converting Enzyme Inhibitors or Angiotensin-Receptor Blockers on SARS-CoV-2 Infection in Adults: A Living Systematic Review Outcomes of COVID-19 Hospitalized Patients Previously Treated with Renin-Angiotensin System Loop diuretics, no. (%) Oral antidiabetic agents Immunosuppressive agents, no. (%) Non-steroid anti-inflammatory agents We are grateful to Manuel Medina-Peralta from SISAP in the Institut Català de la Salut for supervision of data extraction and Sònia Abilleira from Institut Català de la Salut for supervision of data extraction and for reviewing the manuscript.