Comparison of renin–angiotensin–aldosterone system inhibitors with other antihypertensives in association with coronavirus disease-19 clinical outcomes: systematic review and meta-analysis

Introduction: The effects of renin–angiotensin–aldosterone system (RAAS) inhibitors on the clinical outcomes of coronavirus disease-19 (COVID-19) have been conflicting in different studies. This meta-analysis was undertaken to provide more conclusive evidence. Methods: A systematic search for published articles was performed in PubMed and EMBASE from January 5 2020 till May 5 2020 . Studies that reported the clinical outcomes of patients with COVID-19, stratified by the class of concomitant antihypertensive drug therapy, were included. The Mantel-Haenszel random effects model was used to estimate pooled odds ratio (OR). Results: A total of 6,997 patients with COVID-19 were included, and all of them had hypertension. The overall risk of poor patient outcomes (severe COVID-19 or death) was lower in patients taking RAAS inhibitors (OR=0.84, 95% CI: [0.73, 0.96]; P=0.017) compared with those receiving non-RAAS inhibitor antihypertensives. Patients taking angiotensin-I-converting enzyme inhibitors (ACEIs) were less likely to experience poor clinical outcomes (OR=0.73, 95% CI: [0.58-0.92]; P=0.01) compared with those receiving angiotensin-II receptor blockers (ARBs). In addition, comparison of ACEIs to the rest of non-ACEI antihypertensives gave a consistently decreased risk of poor COVID-19 outcome (OR=0.77, 95% CI: [0.63-0.93]; P=0.002). However, ARBs did not decrease the risk of poor COVID-19 outcomes compared to all other non-ARB antihypertensives (OR=1.13, 95% CI: [0.95-1.35]). Conclusion: The risk of developing severe illness or death from COVID-19 was lower in patients who received RAAS inhibitors compared with those who took non-RAAS inhibitors. ACEIs might be better in decreasing the severity and mortality of COVID-19 than ARBs.


Introduction
The effect of renin-angiotensin-aldosterone system (RAAS) inhibitors on the clinical outcomes of coronavirus disease-19 (COVID-19) is of great interest 1 . This is because RAAS blockers, one of the most commonly prescribed antihypertensive drug groups, were previously reported to have some interactions with the pathophysiology of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) 1,2 .
Experimental studies have shown that, blockage of RAAS by either angiotensin-I-converting enzyme inhibitors (ACEIs) or angiotensin-II receptor blockers (ARBs) substantially upregulates the expression of host angiotensin-converting enzyme 2 (ACE2) 3 , a transmembrane enzyme used by SARS-CoV-2 as a receptor to enter and infect cells 4 . On the other hand, ACE2 catalyzes the degradation of potentially harmful angiotensin-II to a vasodilator angiotensin (1)(2)(3)(4)(5)(6)(7), which has antiarrhythmic and cardioprotective effects 3,2 . In addition, RAAS inhibitors may also prevent some complications of COVID-19, such as hypokalaemia. Hence, despite concerns that overexpression of ACE2 with RAAS inhibitors could facilitate infection of tissues by SARS-CoV-2, these drugs could also have a therapeutic role.

Methods
This study was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2009 checklist 13 .

Data sources and search terms
We searched PubMed and EMBASE to identify potentially relevant articles published between January 5 2020 to May 5 2020. A grey literature search was also performed to find additional articles that may have not been indexed. We used three main search keywords: (1) clinical outcome OR death OR mortality, (2) angiotensin and (3) COVID. These key words were combined with Boolean operators to make the following search term: (( (((clinical outcome) OR death) OR mortality)) AND angiotensin) AND COVID. We found 53 and 64 articles indexed in PubMed and EMBASE, respectively. One additional article was found from a manual search. Two authors (Y. B., W. B.) selected studies by screening titles and abstracts. A third author (E. A) served as a mediator to reach a consensus for discrepancies.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 25, 2020. ; https://doi.org/10.1101/2020.05.21.20108993 doi: medRxiv preprint

Study definitions
RAAS inhibitors in this study refer to only ACEIs and ARBs. Severe COVID-19 refers to the presence of any of the following: respiratory rate ≥30 breaths/minute, oxygen saturation at rest ≤93%, oxygenation index [PaO2/FiO2] ≤300 mm Hg, respiratory or other organ failure, mechanical ventilation, shock, or intensive care unit treatment 14 . We used the term 'poor clinical outcome' to indicate the presence of either severe COVID-19 or death. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

Outcome of Interest
The copyright holder for this preprint this version posted May 25, 2020. ; https://doi.org/10.1101/2020.05.21.20108993 doi: medRxiv preprint The main outcome of interest was the risk of having poor clinical outcomes in patients infected with COVID-19 while receiving RAAS inhibitors, compared with those taking other antihypertensive agents. The secondary outcome was the risk of severe COVID-19 or death in patients receiving ACEIs inhibitors compared with those receiving ARBs or other classes of antihypertensives.

Study selection: inclusion and exclusion criteria
Studies that reported the clinical outcomes of COVID-19 patients stratified by class of antihypertensive drug therapy for a similar comorbidity were included. Cohort (prospective or retrospective) studies, case series and editorials/letters that assessed COVID-19 clinical outcomes for patients taking RAAS inhibitors versus non-RAAS inhibitors were included. The included papers were either published or accepted original articles written in English. We excluded review papers and case reports. In addition, studies that examined COVID-19 clinical outcomes in heart failure patients on RAAS inhibitors were ineligible. This is because ACEs and ARBs are generally given to heart failure patients when the disease is worse (ejection fraction <40%) for the prevention of cardiac remodelling 15 .

Data extraction and quality control
In each study, the total number of patients taking RAAS inhibitors or other class(es) of antihypertensives was recorded. Then, for each antihypertensive class exposure, the total number of patients with a poor clinical outcome (severe COVID-19 or death) versus those with a good outcome (non-severe COVID-19 and survival) were recorded. In addition, year, design of study and nature of comorbidities were also documented ( Table 2).
The Newcastle-Ottawa quality assessment scale (NOS) 16 was used for quality assessment of the included studies (Supplementary Table S1). Two reviewers (W.B. and E.A.) independently performed the quality assessment and another author (Y.B.) brought consensus during . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 25, 2020. ; https://doi.org/10.1101/2020.05.21.20108993 doi: medRxiv preprint discrepancies. Articles which got a score of less than 7 stars in the NOS were considered poor quality and excluded (Supplementary Table 1).

Data Analysis
The Mantel-Haenszel random effects model was used to estimate pooled odds ratio (OR), and a two-side alpha value less than 0.05 was considered significant. Publication bias was assessed using the funnel plot asymmetry. All the analysis were performed using the OpenMeta (Analyst) 17 .

Study characteristics and quality assessment
A total of 117 potentially relevant articles through EMBASE (64), PubMed (53), and manual search (1) were found. Of these, 7 articles were included in our final analysis ( Figure 1). All the included articles were of good quality (NOS score ≥7), and study characteristics and quality assessment are shown in Table 1 and Supplementary table S1, respectively.
A total of 6,997 COVID-19 patients were included, and all of them had hypertension. The majority (58% or 4074/6,997) were taking non-RAAS inhibitors, whereas 41% (2,858/6,997) were receiving RAAS inhibitors. The remaining 1% were on non-drug treatment (included for descriptive purposes only and not included in the meta-analyses) for their hypertension. Most of the studies (5 of the 7 studies) 5,8-11 categorized patients based on severity of COVID-19, of which 28% patients (1,107/4,018) developed severe COVID-19. The remaining two studies 6,7 classified patients by survival status, and 6% (181/2,979) died. Only four studies 5,[7][8][9] documented the number of patients taking the specific drug class within the RAAS inhibitor and non-RAAS inhibitor groups. In these studies, the total number of patients taking ACEIs . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Comparison of the risk of poor COVID-19 clinical outcomes with different antihypertensives
We found that the overall risk of poor patient outcomes was lower in patients taking RAAS  (Tables 1 and 2 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Discussion:
Evidence on the safety of antihypertensive medications is of paramount importance as about one-third of the world's population is estimated to have hypertension 18 and this comorbidity is associated with increased mortality in patients with COVID-19 19 . Since RAAS inhibitors were reported to affect the clinical outcome of COVID-19 either for good or worse 6,12,20 , we pooled the recent studies to provide stronger evidence on the effects of these drugs. In addition, we also performed sub-meta-analyses to identify drug classes associated with better outcomes. We found that COVID-19 patients taking RAAS inhibitors had an overall decreased risk of poor outcomes compared to those receiving non-RAAS inhibitors. Interestingly, we also found that only the ACEIs (and not ARBs) were associated with decreased risk of poor patient outcomes (Table 2).
The decreased risk of COVID-19 severity or mortality with the use of RAAS inhibitors could be related to the blockage of a rapidly progressing systemic inflammation that is frequently seen in severe COVID-19 21 . For example, COVID-19 patients taking ACE/ARBs had lower levels of inflammatory markers, such as interleukin 6 (IL-6) 10 , C-reactive protein (CRP) and procalcitonin 11 , than those not taking these drugs. In addition, both classes of drugs could also help prevent hypokalaemia, a complication that was reported to occur in COVID-19 patients 22 .  Figure S10).
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 25, 2020. ; https://doi.org/10.1101/2020.05.21.20108993 doi: medRxiv preprint Further sub-analysis suggested that only ACEIs (and not ARBs) decreased the risk of poor clinical outcomes in COVID-19. This could be due to the differing mechanism of action of these drugs. While ACEIs block the synthesis of angiotensin-II, ARBs only block its action in certain receptors and do not affect its synthesis. Therefore, a high level of angiotensin-II in blood and tissues still occurs with the use of ARBs (versus low levels with ACEIs) 23 . It is known that there are different kinds of receptors for angiotensin-II, and the angiotensin-II receptor 1 (AT1R) is mainly blocked by ARBs. Hence, ARBs may not block the inflammatory effects of angiotensin-II mediated through receptors other than AT1R. Animal studies have shown that in order to fully stop the inflammatory effects of angiotensin-II, a combined blockade of both AT1 and AT2 receptors, as well as inhibition of its effects through the nuclear factor-κB (NF-κB) pathway, are required 24 . Therefore, our results do not support the claims that ARBs could have a therapeutic role in COVID-19 25 . However, ACEIs warrant further study as a potential repurposed drug therapy for COVID-19.
This meta-analysis has several limitations. First, we included a small number of studies that may affect the power of our conclusions. Second, even though all of the included papers were of good quality, propensity matching to address common confounders was performed in only two studies. In addition, there could also be a possibility of publication bias as all odds ratios were less than one (Figure 1), giving an asymmetry in the funnel plot. The other intrinsic problem of comparing one drug therapy to the other is that the indication of a certain antihypertensive usually depends on the presence of another coexisting comorbidity. For example, a comparison of RAAS inhibitors with non-RAAS inhibitors such as BBs might give a false sense of good outcome with RAAS blockers. This is because BBs might be commonly indicated as a rate controller for arrythmias that directly affect the prognosis. However, ACEIs and ARBs have fairly similar indications and our comparison of the risk between these two drugs should be more reliable.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 25, 2020. ; https://doi.org/10.1101/2020.05.21.20108993 doi: medRxiv preprint In conclusion, the risk of severe COVID-19 or death was less likely in patients receiving RAAS inhibitors compared to those taking non-RAAS inhibitor antihypertensive agents. ACEIs could potentially decrease the severity and mortality of COVID-19.