The Effect of Chronic and Inhospital Exposure to Renin-Angiotensin System Inhibitors on the Outcome and Inflammatory State of Coronavirus Disease 2019 Adult Inpatients

Background Controversies exist about the effect of renin-angiotensin system inhibitors (RASi) on coronavirus disease 2019 (COVID-19) outcome. The inhospital use of RASi and its effect on inflammatory sate are still poorly studied during the COVID-19 pandemic. Objectives We aimed to compare the impact of previous and inhospital RASi exposure on the outcome and inflammatory response of COVID-19 patients. Methods Single-centre, ambispective analysis of hospitalized adult COVID-19 patients at Hospital de Santa Maria, Lisbon, between March and August 2020 was performed. We excluded asymptomatic patients and those admitted due to another disease. The primary outcome was inhospital all-cause mortality. Illness severity was assessed based on the development of acute respiratory distress syndrome/acute lung injury (ARDS/ALI), intensive care unit (ICU) admission, and need for invasive mechanical ventilation (IMV). We used C-reactive protein (CRP), ferritin, and interleukin 6 (IL-6) as surrogate markers of the inflammatory response. Results From a total of 432 patients, 279 were selected, among whom 133 (47.7%) were receiving a RASi. Chronic treatment with RASi was not associated with the risk of death (OR 1.24, 95% CI 0.66–2.31, p=0.500), ARDS/ALI development (OR 1.12, 95% CI 0.67–1.86, p=0.676), ICU admission (OR 1.11, 95% CI 0.67–1.84, p = 0.686), and IMV need (OR 1.03, 95% CI 0.58–1.84, p=0.917) in a univariable and multivariable analysis. Inhospital RASi withdrawing was associated with the risk of death (OR 4.38, 95% CI 1.11–17.21, p=0.035) and ARDS/ALI development (OR 4.33, 95% CI 1.49–12.6, p=0.007), the latter remaining significant after adjustment. Previous exposure to RASi was associated with lower CRP levels at admission (p=0.018). IL-6 levels were significantly higher in those patients whose RASi were stopped (p=0.024). Conclusion Previous and inhospital exposure to RASi was not associated with mortality nor severity of COVID-19. This study supports current guidance on RASi management during the COVID-19 pandemic.


Introduction
e coronavirus disease 2019 (COVID- 19) has caused more than two and a half million deaths around the globe. Despite the experimental treatments that have been tried, the pandemic is still evolving, so every effort must be made to reduce its morbidity and mortality. e renin-angiotensin system inhibitors (RASi), namely, angiotensin-converting enzyme inhibitor (ACEi) and angiotensin II receptor type 1 blocker (ARB), are widely used drugs and represent first-line therapies for the treatment for high blood pressure (HBP) [1], a common comorbid condition in COVID-19 patients and a known risk factor for inhospital death [2]. e discovery that the COVID-19 pathogen, the severe acute respiratory syndrome new coronavirus 2 (SARS-CoV-2), uses angiotensin-converting enzyme 2 (ACE2) for viral entry [3] sparked a lively discussion about whether RASi exposure could modulate the clinical course of COVID-19. e renin-angiotensin system (RAS) relies on the balance between two opposite arms. On one side, angiotensin (Ang) II exerts vasoconstrictive, pro-proliferative, and proinflammatory effects via Ang II receptor type 1. On the other side, Ang 1-7, produced from the cleavage of Ang II by ACE2, has opposite actions [4]. Preclinical models showed that chronic exposure to RASi increased tissue ACE2 expression [5,6]. Conversely, ACE2 is significantly downregulated after SARS-CoV-2 binding [3], creating an imbalance between the two RAS axes towards Ang II effects. e acute respiratory distress syndrome (ARDS) is the main cause of death in COVID-19 patients [7]. Its pathophysiology seems to depend at least in part on the RAS instability. Severe ARDS secondary to impaired ACE2 activity has been identified in other viral pneumonias, like H7N9 influenza and SARS-CoV, where elevated circulating Ang II was associated with disease progression and higher mortality rates [8,9]. In a group of 35 ARDS patients, concentrations of Ang 1-7 were higher in patients who survived, suggesting a higher ACE2 activity in such patients [10]. In line with this, serum levels of Ang II were also significantly elevated in COVID-19 patients and exhibited a linear positive correlation with viral load and lung injury [11].
Despite that the safety of RASi use during the pandemic has been assured, data shows conflicting evidence regarding its effect on disease severity, possibly due to a lack of standardized definitions [12]. e growing body of evidence on this topic comes from observational retrospective studies and, as far as we know, no prospective analysis has been published. In addition, most of the data come from Asian countries, mainly from China. ere is a lack of data from European countries and, in particular, from Portugal, where no similar study has been published so far.
In the present study, we aim to investigate the effect of chronic RASi exposure on inhospital mortality and clinical severity of COVID-19 as measured by the development of ARDS/acute lung injury (ALI), intensive care unit (ICU) admission, and the requirement for invasive mechanical ventilation (IMV). We also sought to explore how RASi inhospital management influences those clinical outcomes and the inflammatory response.

Study Design and Participants.
is is an ambispective open cohort study of all adult COVID-19 patients admitted to Hospital de Santa Maria (Lisbon, Portugal), between March 3 rd and August 3 rd , 2020. e clinical outcomes were recorded up until September 3 rd , 2020. All patients had a positive test for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by reverse-transcriptase polymerase chain reaction (RT-PCR). We have excluded all patients that were hospitalized due to another disease (despite having tested positive for SARS-CoV-2 by RT-PCR), those admitted due to social/ sanitary reasons only (e.g., inability of social isolation), and pregnant women. e patients' selection process is schematized in Figure 1. Patients entered the study retrospectively and prospectively before and after the 27 th of April, respectively. Patients were followed until they died, were lost to follow-up, or were discharged. All patients reached the end of the study. Patients were stratified according to RASi exposure: RASi group vs. non-RASi group. Patients were classified in the RASi group as long as they were on any ACEi/ARB upon admission, despite the dose and duration of treatment and regardless of how RASi therapy was subsequently managed during the inhospital stay. For a subanalysis purpose, we then focused on how the RASi treatment was managed during the first three days of hospitalization. We have looked into whether patients had begun, stopped, or continued the RASi treatment regimen and categorized them into three groups, start-RASi group, stop-RASi group, and non-stop-RASi group, accordingly. e study was approved by the ethics committee of Hospital de Santa Maria (N°177/20) and complies with the Declaration of Helsinki statement.

Data Collection.
Patients' demographic characteristics, comorbid conditions, and clinical manifestations were obtained through clinical interviews and the analysis of individual patient electronic files. Laboratory results were collected using electronic software. Comorbidities were either self-reported by the patients and/or extracted from their medical records. Patients' treatment regimens were assessed through medical interviews and confirmed in the prescribed medication chart. e inhospital treatments were collected from electronic files and prescription software. All data were collected by four physicians and reviewed by other two. e anonymity of the collected data was warranted by the authors, and data were collected and stored according to the applicable legislation.

Definitions and Outcomes.
e primary outcome was inhospital all-cause mortality. e COVID-19 severity was assessed according to the secondary outcomes as follows: development of ARDS/ALI, ICU admission, and need for IMV. ARDS and ALI were defined according to Berlin criteria [13] and the American-European Consensus Conference on ARDS [14], respectively. e COVID-19 onset was defined as the time point when symptoms were first noticed. Symptoms were considered up until the third day of hospitalization. Laboratory results corresponded to the first available result up to the third day of inhospital stay. We considered the outcome date as the time the patient died or was clinically discharged (no fever and no supplemental oxygen requirements for more than two consecutive days). e total length of hospital stay was determined from the date of admission until the outcome date. Continuous variables were expressed as median ± interquartile range and categorical variables as number (%). We compared groups of patients using Pearson's chi-square test and Fisher exact test for categorical variables and the Mann-Whitney U test for continuous numerical variables. Odds ratios and multivariable analysis were performed using a logistic model. For multivariable analysis, we included variables from the univariable analysis if their between-group differences were significant. Patients were censored if they were lost to follow-up or reached the primary outcome. e significance level was defined at 0.05.

3.1.
Demographics, Comorbidities, and Disease Characteristics. From a total of 432 patients admitted during the study period time, 279 patients were selected for analysis. e retrospective and prospective parts of the study included 117 (41.9%) and 162 patients (58.1%), respectively. One hundred thirty-three patients (47.7%) belonged to the RASi group, whereas 146 patients (52.3%) were categorized in the non-RASi group. e baseline demographic, clinical, and laboratory characteristics as well as inpatient treatments and outcomes of our cohort are summarized in Table 1 Both groups showed similar signs and symptoms except for cough (74.5% vs. 58.8, p � 0.006), headache (21% vs. 7.7%, p � 0.002), ageusia (11.3% vs. 3.9%, p � 0.024), and anosmia (9.2% vs. 3.1%, p � 0.046) that were significantly more prevalent in the non-RASi group than in the RASi group. Patients under RASi therapy presented sooner to the hospital, with a significantly smaller median time from symptom onset to hospital admission (4 ± 4 vs. 6 ± 5 days, p � 0.041). e median duration of total hospital stay was similar between the two groups (11 ± 13 vs. 10 ± 10.5 days, p � 0.669) ( Table 1).

Association between RASi Exposure with Mortality and
Severe Clinical Outcomes. Forty-eight patients (17.3%) died during the study period, and two were transferred to another hospital during ICU stay, hence lost to follow-up. Patients who died were older (84.5 ± 16.5 vs. 67 ± 25 years, p < 0.001) and presented sooner to the hospital (3 ± 4 vs. 6 ± 4 days, p � 0.008). CVD (50.0% vs. 21.8%, p < 0.001), COPD (29.2% vs. 10 Patients who reached these outcomes were more frequently of male gender (p � 0.001, p < 0.001, and p � 0.001, respectively), whereas there was no difference in age or in ethnic origin. In general, the proportion of comorbid conditions were similarly distributed between patients who reached any of these severe clinical outcomes compared to those who had a less severe clinical course, except for COPD which was significantly more prevalent in patients who were admitted to ICU than in those who were not (20.2% vs. 10.0%, p � 0.019). See Supplementary Tables 2-4 for further characterization.
Regarding the use of RASi, we observed no difference in inhospital mortality rate (18.9% vs. 15.9%, p � 0.499), development of ARDS/ALI (33.6% vs. 31.2%, p � 0.676), ICU admission (33.1% vs. 30.8%, p � 0.686), and need for IMV (21.1% vs. 20.6%, p � 0.917) between patients on the RASi group and those in the non-RASi group, despite a uniformly greater percentage in the former group. In a logistic model, the association between RASi use and any of the aforementioned outcomes remained without statistical significance after adjusting for confounder variables (Table 2). Additionally, we noticed that patients in the RASi group had shorter time from disease onset to ARDS/ALI (7 ± 5 vs. 11 ± 6 days, p � 0.002), to ICU admission (7 ± 5 vs. 8 ± 3.5 days, p � 0.044), and to IMV (6 ± 5 vs. 9 ± 4 days, p � 0.002) than in those in the non-RASi group. Time from

Discussion
Our main findings are in accordance with the current state of the art, in the sense that chronic treatment with RASi did not affect the severity of the disease and the risk of death during hospitalization for COVID-19. To the best of our knowledge, this is the first hospital-based study to enroll patients prospectively. In addition, this is also one of the few observational studies that specifically correlates the potential effect of RASi maintenance vs. withdrawing on clinical outcomes during hospitalization. While the pandemic is rapidly evolving, the neutral effect of RASi exposure on mortality is increasingly recognized, although some disparities exist. Patoulias et al. [15] showed that, although the use of RASi does not increase the odds for SARS-CoV-2-related death in a global scenario (OR � 1.06, 95% CI 0.77-1.47, I 2 � 83%), it increases the odds for death in Europe by 68% (OR � 1.68, 95% CI 1.05-2.70, I 2 � 82%), while decreasing it in Asia by 38% (OR � 0.62, 95% CI 0.39-0.99, I 2 � 0%). Others have come to the same nonsignificant results in a global perspective, but no subgroup analysis by region was made [16]. Our study indicates that RASi exposure is not associated with an increase nor decrease in the risk of inhospital all-cause mortality (adjusted OR 0.66, 95% CI 0.31-1.40, p � 0.226), even though the odds showed a nonsignificant trend of decrease in the exposed group after adjusting for age and comorbidities (logistic regression coefficient of −0.42) ( Table 2). ese findings should contribute with new insights from a European country since most of the studies come from Asia where cardiovascular disease prevalence and its treatment strategies diverge greatly.
ARDS is the main cause of death in COVID-19 patients [7] and its prevalence is highly variable between studies, ranging from 15% to almost 33% [17]. In line with the literature, almost one-third of our patients developed ARDS/ALI. Decreased expression of ACE2 after viral infection can alter pulmonary vascular permeability leading to pulmonary oedema and acute respiratory failure [18]. is is believed to play a central role in SARS-CoV-2-related ARDS and justifies the theoretical protective role of RASi in this context. Unlike mortality, which is easily assessed and surely the most frequent outcome of interest in similar studies, illness severity criteria have been inconsistently defined across the studies addressing the impact of RASi in COVID-19 patients. In our study, we defined severe disease based on objective and easily accessible parameters, namely, the development of ARDS/ALI and the rates of ICU admission and IMV requirements. A recent meta-analysis [12] showed that only one study defined severe disease based on the development of ARDS. Unfortunately, the study was probably not powered enough to detect any difference between the two  [19]. In the present study, we observed that the use of a RASi was not associated with the odds of developing ARDS/ALI, even after multivariable adjustment. ese results should, however, be analysed with caution because it englobes not only ARDS patients diagnosed by Berlin criteria [13] but also patients that we classified as having ALI who would, otherwise, be classified as having ARDS should they be given a positive end-expiratory pressure. In our institution, this is only available in the ICU from where these subjects were not candidates due to their comorbid conditions to start with. We also observed no significant differences in the odds for ICU admission nor for IMV need between the two groups. We find these results expectable given that the development of ARDS/ALI is the main reason why these patients are admitted to an ICU and are ventilated. We did notice that the median time from symptom onset to any of the secondary outcomes, but not to death, was significantly shorter in the RASi group than those in the non-RASi group. Giving that patients under RASi therapy presented sooner to the hospital, and assuming that medical care was equally given to both groups of patients (with a comparable median time from admission to each secondary outcome and a comparable proportion of inhospital treatments between the two groups (Table 1)), this suggests that patients in the RASi group may have been admitted with more severe disease. e effect of inhospital withdrawing vs. maintaining the RASi prescription is poorly reported in the context of SARS-CoV-2 infection. Zhang et al. [20] showed that inpatient use of RASi was associated with a lower risk of 28-day all-cause inhospital mortality (adjusted hazard ratio 0.37; 95% CI, 0.15-0.89; p � 0.03) than in RASi nonusers. Due to the small number of patients (n � 5) who started an ACEi or ARB upon admission, we could not analyse it separately. Our adjusted results show, however, that the withdrawing of RASi in the first three days of hospitalization was associated with a higher risk of developing ARDS/ALI but not of dying (Table 3). e opposite is necessarily true since patients were dichotomized in a binary way. ese findings further support the guidance statements by several international societies on continuing current antihypertensive treatment during the COVID-19 pandemic [21]. Ongoing randomized clinical trials of RASi in COVID-19 patients (e.g., NCT04312009, NCT04366050, and NCT04355429) will certainly elucidate this in the near future. A hyperinflammatory state characterizes the severe cases of COVID-19 [22]. Consistent with this hypothesis, levels of CRP, ferritin, IL-6, IL-8, and tumour necrosis factor alpha are significantly increased in deceased patients with COVID-19 compared to those who survived [23] and are independent predictors of survival [24]. e anti-inflammatory effect of ACEi/ARB use has been increasingly demonstrated in several preclinical and clinical disease models [25], but studies on COVID-19 are scarce [26]. Even though we did not aim to specifically capture the full inflammatory profile of SARS-CoV-2 infection, our study is one of the few observational studies to explore with some detail the effect of RASi exposure and inpatient management on inflammatory markers. We found that CRP levels on admission were significantly lower in patients receiving chronic RASi treatment, whereas no significant change was observed in IL-6 and ferritin levels. ese findings have been inconsistently reported, with some authors showing similar [27,28] and opposite [29] results. We also found that CRP and ferritin values did not differ significantly whether patients had maintained or discontinued their RASi treatment upon admission. Curiously, IL-6 levels did change, being significantly lower in those patients who maintained their RASi therapy regimen. A multicentric study regarding the effect of inhospital use of ACEi/ARB reported similar findings regarding CRP levels, but neither ferritin nor IL-6 was measured [20]. To study the correlation between the inflammatory markers and the measured outcomes was beyond the scope of the current study.
is study has several limitations. First, its single-centre and observational nature limits the generalisability and the establishment of causality. Second, we have excluded patients who, despite being tested positive for SARS-CoV-2 by  International Journal of Hypertension 7 RT-PCR, were admitted due to social reasons. We assume this could have driven our results towards the null as asymptomatic and, thus, less severe patients were less likely to have an adverse outcome. ird, when examining the effect of RASi withdrawing vs. its maintenance, we did not take into consideration whether patients were actually given the prescribed drug nor the effect of its introduction or suspension beyond the first three days of hospitalization. Finally, although more than half of the enrolled patients were followed prospectively and despite our active role in their inpatient treatment, we were not able to have direct interaction with all patients, neither did we have control over when or which laboratory exams were performed; hence, some data are missing. We also have some strengths. As all our patients ended the study, our mortality and severe outcome rate are accurate. In addition, the ambispective nature of the study, which is a unique feature in this kind of work, warrants preciseness in data collection.

Conclusion
In conclusion, chronic treatment with RASi is not associated with either inhospital mortality or disease severity as measured by the risk of developing ARDS/ALI, ICU admission, and need for IMV in COVID-19 patients. Our findings support the maintenance of RASi during hospitalization. Future nationwide and multinational studies should corroborate our results and address these pitfalls, particularly patients' enrollment criteria as well as severity outcome definitions, as this can help uniformize and generalize knowledge.

Data Availability
e collected data used to support the findings of this study are restricted by the Ethics Committee of Hospital de Santa Maria in order to protect patient privacy. Data are available for researchers who meet the criteria for access to confidential data.

Conflicts of Interest
e authors declare no conflicts of interest.

Authors' Contributions
PG contributed to concept and design and drafting of the manuscript; PG, IP, FB, PAM, VDS, and AMA were responsible for acquisition, analysis, and interpretation of the data; PG and PAM performed statistical analysis; IP, FB, PAM, VDS, AMA, APL, and CM critically revised the manuscript; APL and CM supervised the study. All authors approved the final version of the manuscript.