This page has received over 300,000 page views (March-July 2020)
Preamble
There has been an unprecedented interest in this topic. Hence we have added more details and moved this section from the main page to a dedicated area with focus on hypertension and ACE2. While many questions remain unknown we want to arm people with the current state of knowledge. Since this is an evolving topic we will continue to update this when new evidence is brought to our attention. Below we review the pre-clinical and clinical studies linking coronavirus to the renin angiotensin system.
Before we start reviewing the literature it is important to recognize the current landscape of peer review (or lack thereof). Especially in the fast moving world of COVID-19 research. Brian Byrd provides a nice overview for our readers here. The differences between preprints, letters to editor, rapid response and other publication types are discussed in detail.
Links to workgroup statements/publications: ASN Kidney News Online, CJASN Perspective, Clinical Infectious Diseases, Hypertension, AJP Heart and Circ. Phys. Nature Reviews Nephrology and Lancet Respiratory Medicine Correspondence.
Curated by
Matthew Sparks, MD, Duke University
Swapnil Hiremath, MD, MPH, University of Ottawa
with additional expert input from
Andrew South, MD, MS, Wake Forest School of Medicine, Brenner Children’s Hospital
Paul Welling, MD, Johns Hopkins
Matt Luther, MD, Vanderbilt University
Jordy Cohen, MD, MSCE, University of Pennsylvania
Brian Byrd, MD, MS, University of Michigan
Louise M. Burrell, MD, University of Melbourne, Austin Health Australia
Daniel Batlle, MD, Northwestern University
Laurie Tomlinson, MD, London School of Hygiene & Tropical Medicine, UK
Vivek Bhalla, MD, Stanford University
María José Soler, MD, PhD, Hospital del Vall d Hebron. Barcelona, Spain
Sundar Swaminathan, MD, University of Virginia
April Pettit, MD, MPH, Vanderbilt University
Javid Moslehi, MD, Vanderbilt University Adam Bress, PharmD, MS, University of Utah Ricky Turgeon, PharmD, University of British Columbia
Updated on Oct 4, 12:53 PM EST
Last updates: Updated outcomes on two RCTs added
The above recommendation is not only the recommendation of the NephJC team but is consistent with all of the professional societies who have examined and posted their recommendations on the question. The situation is changing by the hour and this position may change in the future but currently with the available information this is where we are.
Professional societies statements on the topic
European Society of Cardiology, European Society of Hypertension, Hypertension Canada, The Canadian Cardiovascular Society and the Canadian Heart Failure Society. The UK Renal Association, International Society of Hypertension, American College of Physicians, Spanish Society of Hypertension, American Heart Association/American College of Cardiology/Heart Failure Society of America, European Renal Association/European Dialysis and Transplant Association, HBPRC of Australia, Australian Diabetes Society, British and Irish Society of Hypertension
The competing hypothetical mechanisms by which inhibition of the renin–angiotensin system (RAS) with an angiotensin-converting enzyme inhibitor (ACEi) or angiotensin-receptor blocker (ARB) might be harmful (upper panels) or protective (lower panels) in COVID-19. Hypothesis 1: severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) gains entry into the cell by binding to angiotensin-converting enzyme 2 (ACE2; upper left panel). The addition of an ACEi or ARB could increase ACE2 abundance and thus enhance viral entry (upper right panel). Hypothesis 2: angiotensin II (Ang II) drives lung injury by activating the type 1 angiotensin receptor (AT1R), causing inflammation and fibrosis (lower left panel). Diminishing production of Ang II with an ACEi or blocking Ang II–AT1R actions with an ARB enhances the generation of Ang-(1–7) by ACE2 and activation of the Mas receptor (MasR), which attenuates inflammation and fibrosis and therefore attenuates lung injury. From South et al Nature Reviews Nephrology 2020
Hypertension and COVID-19
Are patients with hypertension more likely to get COVID-19?
For this, we need a well-designed cohort study with incidence rates of COVID-19 in patients with hypertension (HT) and those without HT, in which exposure history is able to be carefully accounted for.
Instead, what we have (below, in a table) is history of hypertension vs not, in those with COVID-19. These data are all unadjusted (for age, as an example). So one cannot draw any conclusion from these numbers. The best data we have so far is from the largest preprint from NY (see below for more).
Are patients with hypertension more likely to get serious COVID-19 or die?
Among the patients with COVID-19, it seems the prevalence of prior h/o HT is higher in those who develop severe disease than those who do not. Same applies for development of ARDS or death - but mostly in unadjusted analysis. See table below for more. The last two rows are the best quality data so far, and suggest the association between COVID-19 severity and hypertension is attenuated after adjustment.
Sources: Guan et al, NEJM; Huang et al, Lancet; Wang et al, JAMA; Zhang et al, Allergy; Zhou et al, Lancet; Wu et al, JAMA IM; Italian report (PDF); Chen et al, BMJ; Shi et al, JAMA Cardiol; McMichael et al, NEJM; Guo et al, JAMA Cardiol; Bean et al, MedRxiv 2020; Petrilli et al, MedRXiv 2020
As can be seen, most of the studies except the last two, did not adjust for age. Even age-stratified association of hypertension would have been a more useful way to see these data to understand this issue a bit better. We hope more data in coming days will clarify this relationship. Unlike what we stated earlier, the Zhou et al study did not adjust for hypertension.
Source: AHA website, using NHANES data
As can be seen above, hypertension in the general populationclosely associates with increasing age, hence any association with hypertension may merely represent confounding due to age, and should be interpreted after careful analysis.
A recent Italian study (via Bloomberg) reports a high prevalence of hypertension amongst those who died. It is hard to know what to make of those raw rates, as can be seen above, the patients who died were also quite older. The prevalence of HT in those who died (76%) was higher than diabetes (36%), heart disease (33%) and CKD (18%). However, the median age of those who dies was 80.5 years. The prevalence of hypertension goes up with age, as an example was 76% in those 75 and older in this report. Hence this prevalence might just represent the expected prevalence for this age group. Additionally, from the Seattle outbreak study (McMichael et al, NEJM), it is clear that the prevalence of hypertension was indeed high in the older care facility residents who developed COVID-19, but not so much in the younger visitors, and staff who worked there.
Thus hypertension does seem to be a common comorbidity, even more so than diabetes - but these are mostly data from China, and one brief report from Italy (latter with much higher hypertension rate) and one from US, so one should be careful before generalization. It is also not clear how hypertension was coded - we can speculate that it might be based on use of hypertension medications rather than actual BP measurement. We do have data on actual BP from two studies:
In one study by (Huang et al, Lancet), the median systolic BP, on admission, in the 13 patients with COVID-19 who required ICU care later was 145 mm Hg compared to 122 in the 22 patients who did not. The patients needing ICU care were similar in age to those who did not need ICU care (median age 49 years)
In another study (Chen et al, BMJ) the median arterial BP, on admission, was 137 mm Hg among those who died compared to 125 mm Hg amongst those who survived. In this study, the patients who died were older (median age 68 years) than those who survived (median age 51 years)
Without better adjusted or stratified data, it is hard to say what this represents.
Does being on ACEi or ARB increase risk of contracting COVID-19?
We don’t yet know that having hypertension is a risk factor (independent of age) for developing COVID-19, or serious disease. Roughly 15 - 30% of patients with COVID-19 had hypertension, in the reported literature so far (see table above). The proportion of these who might have been on an ACEi or ARB will be a smaller subset of this 15-30%. There are no clinical data to support that being on an ACEi or ARB increases a risk, nor does the basic science discussed below support a clear association.
But even if there is a doubtful or dubious association, what’s the harm in switching to a different BP medication?
From the science, it’s not just that we don’t if know ACEis or ARBs (looked at separately) increase the risk of COVID19, or the severity, we don’t even know whether these drugs might be beneficial. To reiterate:
We don’t know if there is an association between ACEi/ARB and COVID19
If there is one, we don’t know the direction of association (beneficial or harmful)
We don’t know the magnitude of any possible association
ACEis and ARBs are useful drugs. They don’t just reduce blood pressure, they decrease risk of progressive kidney disease, of dying if one is at high cardiovascular risk, and if one has heart failure.
Stopping these drugs for an unknown potential benefit can make these conditions worse. The patient will need a replacement drug (more contact with an already stretched medical system), visit pharmacies (when one should be practicing social distancing), and may even need follow up lab investigations or medical visits to ensure the replacement drugs are working as they should.
Moreover, how long does it take for any biological effect on ACE2, to wear off, and make a tangible difference?
Should these drugs be started or stopped in patients with suspected or confirmed COVID-19?
The same discussion largely applies here: in addition that if the virus is in, and has bound to ACE2, will stopping the ACEis or ARBs now have enough of an effect of ACE2 to make a difference?
There is several clinical trials ongoing, or planned to answer these questions in patients with COVID-19, along with several large scale observational studies (see below). Until then, the actual decision for an individual patient should be an individualized one, based on the specific needs, benefits, and hemodynamics. We do hope more data become available in the coming days to allow us to make more informed choices.
RAS activation and long term outcomes
Patients who survived SARS-CoV have altered lipid metabolism even 12 years later and thus may have worse long-term outcomes including cardiovascular disease. Thus data on long-term outcomes of patients who survive SARS-CoV-2 will be important, especially because the RAS can be programmed to chronically upregulate ACE/Ang II at the expense of ACE2/Ang-(1-7) in otherwise healthy adolescents and young adults.
ACEi/ARB use and Clinical Outcome Data
This section started off with a few case series, but has some large observational studies now, and the initial trial data. The caveats we mention at the start on being careful in drawing conclusions from these studies are even more necessary here. Nevertheless, as these data start trickling in, expect this section to grow, and check back here for updates.
A previous version of this blog page had a longer discussion of the results of all the studies conducted in this area. In the pursuit of brevity, we have archived most of that discussion, which is now available here. Below, we summarize the some recent epidemiological studies published, which are of higher quality and large numbers.
RCT Data
Data from randomized controlled trials (RCTs) are beginning to emerge, and we now have results from 2 trials previously included in our tables of registered trials: The completed BRACE-CORONA trial (NCT04364893) and an interim analysis of a trial from Argentina (NCT04355936; Duarte et al, MedRXiv, 2020).
The first of these trials, BRACE-CORONA (unpublished yet, see a tweet thread here from #ESCCongress), was an open-label RCT of 659 patients hospitalized with a confirmed diagnosis of COVID-19 who used chronic ACEI/ARB that randomized patients to temporarily suspend or continue ACEI/ARB for 30 days. Key exclusion criteria were hemodynamic instability in the first 24 hours, hospitalization for decompensated heart failure in the last year, use of sacubitril-valsartan, or use of more than 3 antihypertensives. Of the 1352 patients with suspected COVID-19 screened for this study, 620 were excluded for not being on an ACEI/ARB, not having confirmed COVID-19, or errors in randomization/consent, and an additional 73 were excluded post-randomization for being at a site with good clinical practice violations. At baseline, the average age was ~56 years, 40% were women, and patients had been on ACEI/ARB for a median of 5 years. In terms of comorbidities, all patients had hypertension, 42% were obese, 32% had diabetes, ~5% had CAD and 1-2% had HF. On admission, ~70% had fever, ~70% had cough, 54% had dyspnea, and 27% had SpO2 <94% on room air. Within the first 24 hours, severity of COVID-19 was rated as mild in 57% and moderate (SpO2 <94%, PF<300 or CT lung infiltrate >50%) in 43%. The primary outcome, number of days alive and out of hospital (DAOH) through 30 days, was 21.9 days in those suspending ACEI/ARB vs 22.9 days in those continuing ACEI/ARB, with a mean difference that was non-significantly in favor of continuing ACEI/ARB (-1.1 days, 95% confidence interval [CI] -2.33 to +0.17). Death from any cause at 30 days was 2.7% and 2.8%, respectively, in those suspending vs continuing ACEI/ARB (hazard ratio 0.97, 95% CI 0.38-2.52). Reporting of secondary outcomes from the registered protocol are pending. At present, this trial represents the best-available evidence on this question and does not suggest any short-term harm with continuing ACEI/ARB in patients hospitalized for mild-to-moderate COVID-19, provided that they are hemodynamically stable. In terms of risk of bias elements, sequence generation and allocation concealment are unclear (not commonly reported in presentations), patients and treating clinicians were not blinded, there was no loss-to-follow-up, and patients were analyzed according to the intention-to-treat principles. Using the GRADE approach, we grade the certainty of evidence as moderate for DAOH (rated down 1 category for serious risk of bias owing to the subjective nature of this outcome), and moderate for death from any cause (rated down 1 category for serious imprecision.
The second trial from Argentina is an interim analysis of 68 patients (of a target 400 patients) hospitalized with confirmed COVID-19 with symptoms <=4 days randomized to telmisartan 80 mg BID for 14 days or open-label usual care. In addition to being smaller than BRACE-CORONA, this report has several important methodological issues that preclude clear interpretation. First, the results were provided based on an interim analysis without clear parameters or rules to stop or continue the trial. Second, the trial did not employ methods to conceal allocation (as confirmed by the authors in the MedRxiv comments), used simple randomization, and did not blind participants or their clinicians to allocated treatment. These factors result in high risk of allocation, performance and detection bias, which can exaggerate differences between groups and lead to spurious results, particularly for subjective endpoints. Third, data for the stated primary outcome of C-reactive protein at days 5 and 8 were missing for many patients in both groups, without proper methods to deal with these missing data in this interim analysis. Given these limitations, these interim results constitute evidence with (at best) low certainty of evidence for all outcomes (rated down 2 categories for very serious risk of bias).
Epidemiological Studies
We previously discussed a large observational study (Mehra et al, NEJM 2020) here, which purported to discuss results from a large observational database. After an initial expression of concern (Rubin E., NEJM 2020), that study has now been formally retracted (Mehra et al, NEJM 2020).
Reynolds conducted a propensity score-matched analysis (Reynolds et al, NEJM 2020) to determine if prior use of anti-hypertension medications was associated with the likelihood of COVID-19 or severe illness. This study included 12,594 patients tested for COVID-19 at the NYU Langone health system, of whom 5894 were COVID positive and 1002 had severe disease (ICU admission, ventilation or death). Of the COVID-19-positive patients, 954 were receiving an ACEi and 1238 were receiving an ARB. After propensity score matching, the median difference in the likelihood of testing positive or having severe disease by ACEi or ARB use was not significant. Prior ACEi/ARB use was ascertained using electronic medical records, though we don’t have data on whether these drugs were continued or stopped subsequently, nor do we have data on actual BPs.
A case-control study from Milan (Mancia et al, NEJM, 2020) investigated whether prior anti-hypertensive medication use was associated with the risk of developing COVID-19 infection and subsequent outcomes. This study compared 6272 patients with COVID-19 to 30,579 age/sex/municipality of residence-matched controls (1:5 ratio). ACEi/ARB use was ascertained using a regional drug database (current until Dec 2019). Logistic regression was performed including a variety of covariates. There was no association for ACEi or ARB use with developing COVID-19 (adjusted OR 0.97, 95% CI 0.87 to 1.07 for ACEi; and 0.95, 95% CI 0.86 to 1.05 for ARB) or for severe COVID-19 (adjusted OR 0.83, 95% 0.63 to 1.10 for ACEi; and 0.91, 95% CI 0.69 to 1.21 for ARBs). Similar results were reported when additionally stratified for age (≥/< 60 years) or sex and subgroup analysis based upon the number of prescriptions over set time periods. Data on ACEi/ARB use during the pandemic, especially after hospitalization, or data on actual BP was not available.
A case-control study from Madrid (de Abajo et al, Lancet 2020 ) investigated whether use of renin-angiotensin-aldosterone system (RAAS) inhibitors (including ACE inhibitors, ARBs, mineralocorticoid receptor antagonists and renin inhibitors) were associated with hospitalization for COVID-19 among the general population, and whether they were associated with risk of severe COVID-19 (admission to intensive care or in-hospital death) among those hospitalized with COVID-19. This study compared 1139 patients hospitalized with confirmed COVID-19 in March 2020 (cases) to 11,390 patients from the general population in March 2018 (controls). Cases were matched in a 1:10 ratio to controls based on age, sex, region and date (day of the month), and compared based on their exposure to RAAS inhibitors versus non-RAAS antihypertensive agents. Exposure to medications was based on outpatient prescription fills lasting until 1 month before the index date using administrative databases. The logistic regression was adjusted for several potential confounders. There was no association between RAAS inhibitor use compared to non-RAAS with COVID-19 hospitalization (adjusted OR 0.94, 9% CI 0.77-1.15) or severe COVID-19 (adjusted OR 1.08, 95% CI 0.80-1.47), whether RAAS inhibitors were considered as a group or as individual classes. Although the control group was selected from a different year before the COVID-19 pandemic, the results were robust to accounting for possible “secular trends” in sensitivity analysis. Key strengths of this study include more accurate ascertainment of medication exposure via prescription fills (rather than history collected on hospital admission), and minimization of confounding by indication by using other antihypertensives (rather than non-use of RAAS inhibitors) as a comparator. Notably, compared to no use of any antihypertensives, both RAAS inhibitors (adjusted OR 1.71, 95% CI 1.46-2.01) and other antihypertensives with no known effect on ACE2 (adjusted OR 1.82, 95% CI 1.47-2.26) were associated with COVID-19 hospitalization, pointing to residual confounding in analyses that do not account for confounding by indication.
A cohort study from the United States (Mehta et al. JAMA, 2020) investigated the association between ACE inhibitors/ARBs and vulnerability to COVID-19 (likelihood of testing positive for COVID-19 among those tested) and severity of COVID-19 (hospital admission, ICU admission and death) among those testing positive. The study included 18,472 patients tested for COVID-19, including 1322 ACEi users and 982 ARB users. Exposure to medications was based on the medications listed in the electronic medical record. Notably during the timeframe, COVID-19 testing was limited, and therefore was prioritized toward higher-risk patients, including several conditions that are indications for ACEi/ARB use. Propensity scores were generated using a list of known potential confounders, and these were used for statistical adjustment using using overlap propensity score weighting. In this adjusted analysis, use of ACEi/ARB was not associated with testing positive for COVID-19 (OR 0.97, 95% CI 0.81-1.15), but was associated with higher risk of hospital and ICU admission among those testing positive. Important limitations include the use of a comparator (non-use of RAAS inhibitors) that does not account for confounding by indication, and a limited list of potential confounders considered in the propensity score, which may explain the association identified between ACEI/ARB use and hospitalization/ICU admission.
Finally, Khera et al. performed 2 cohort studies assessing the association between ACEi/ARB and severity of COVID-19 (Khera et al, MedRXiv, 2020). In the first cohort, they matched 441 patients taking an ACEi to 441 patients taking a non-ACEi/ARB antihypertensive agent, and 412 patients taking an ARB to 412 patients taking a non-ACEi/ARB antihypertensive agent. There was no association between either ACEi or ARB with COVID-19 hospitalization. However, the authors highlighted a subgroup analysis with a weak interaction (p-value for interaction=0.09) suggesting that patients insured under Medicare Advantage, which services a population that is more likely to be elderly and have comorbidities, had an associated lower risk of COVID-19 hospitalization, whereas there was no difference in patients with a commercial insurance plan. No such interaction was observed for ARBs. Furthermore, the second cohort that focused on patients hospitalized with COVID-19, did not find an association between ACEi or ARB use and mortality. Overall, this study shows no association between ACEi or ARB use and COVID-19 outcomes, and the subgroup effect for ACEi use in more vulnerable patients has not yet been replicated and is likely attributable to chance.
This larger and more recent literature is clearly of higher quality, uses some causal inference methods, and provides reassurance for the stance taken that there is no need for preemptive change in ACEi or ARBs, unless clinically indicated. The ideal study method to establish the role of prior ACEi/ARB use and developing COVID-19 would be of a prospective design, with careful and accurate ascertainment of drug use over time, complete follow up, and using causal inference methodology. See the list of ongoing trials at the end of this blog post that we expect will provide cleaner and clearer data of this issue.
Current known clinical trials examining RAS blockade in COVID-19
Ongoing Trials
Registered trials, not yet recruiting
Why might there be a link between high blood pressure, ACEi/ARB use and COVID19?
As you can read so far, we are not convinced the data show a strong, robust link. However, the virus uses the renin-angiotensin system - hence all the speculation. Read about the science behind the speculation below.
Letters in Lancet Respiratory Medicine and The BMJ and the Journal of Hypertension and Journal of Travel Medicine are speculative and below we provide a detailed description of the science.
Terminology
These terms can be bewildering to the non-RAS nerd. Here is a cheat sheet. Detailed explanations follow beneath.
The Science
What is ACE2 anyway?
Angiotensin (Ang) converting enzyme 2 (ACE2) is a homolog of ACE. ACE2 negatively regulates the renin angiotensin system (RAS) by converting Ang II to Ang-(1-7). ACE2 has only 1 active enzymatic site. It functions as a carboxypeptidase. The first description, interestingly enough, of ACE2 was its ability to generate Ang-(1-9) (no known biologic activity) from Ang I (2000 Circ Research). However, the conversion of Ang II to Ang-(1-7) has two effects- 1. Diminishing the primary effector of the RAS, Ang II, thus decreasing vasoconstriction and 2. Production of the vasodilatory Ang- (1-7) (binding to the Mas receptor). The net effect is less vasoconstriction. ACE2 also cleaves other peptides. These include des-Arg9-bradykinin, neurotensin 1–13, kinetensin, apelin-13 and dynorphin A 1–13. ACE2 is also membrane bound. There is considerable interest in how ACE2 functions in cardiovascular disease and in particular the role of membrane bound versus soluble ACE2.
from Ghazi et al 2017
Now… What in the world is the link between ACE2 and coronavirus?
First, the coronavirus gains entry to a cell by utilizing ACE2 and type II transmembrane serine proteases (TMPRSS2).
How does it do this? The coronavirus S (spike) protein utilizes ACE2 as a receptor for host cell entry. The major viral coronavirus target cells are type II pneumocytes and enterocytes. The S protein binds the catalytic domain of ACE2 with high affinity. Binding of the coronavirus S protein to ACE2 triggers a conformational change in the S protein of the coronavirus, allowing for proteolytic digestion by host cell proteases (TMPRSS2) which is a potential target for intervention. Zhou et al Nature (2020) described the full-length genome sequences from five patients early in the outbreak.
shares 79.6% sequence identity to SARS-CoV.
is 96% identical at the whole-genome level to a bat coronavirus
Wrapp et al. Science (2020) provide the Cryo-EM structure of the 2019-nCoV spike. This paper also documented a 10-20-fold higher affinity of ACE2 for SARS-CoV-2 compared to SARS-CoV.
ACE2 has been shown to be expressed in human lung tissue (Hamming et al The Journal of Pathology, 2004). Utilizing immunostaining they showed ACE2 expression on type II alveolar epithelial cells and capillary endothelium (below).
Figure 1 of Hamming et al Normal lung tissue: overview (A) and larger magnification (B). Positive staining for ACE2 is present on alveolar epithelial cells (arrow) and capillary endothelium (arrow‐head). Fibrotic lung tissue (C) and larger magnification (D). Positive staining for ACE2 is present on type II cells (arrow). Cultured lung type II alveolar epithelial cells (A549) are positive for ACE2 (E). Control section stained with anti‐ACE2 in the presence of the synthetic ACE2 peptide shows no staining of lung tissue (F)
A preprint from ChinaXiv from Wu et al (2020) shows how the enzyme furin plays an important role in this viral life cycle of SARS-CoV-2 and this is distinctly different than SARS-CoV. Below is a cartoon of this.
TMRPCC2 = TMPRSS2
Emerging data utilizing single cell RNA-seq
A preprint on bioRxiv describes single nuclei RNA sequence (snRNAseq) and single cell RNAseq analysis of healthy lung tissue from lung adenocarcinoma resections and healthy primary human bronchial epithelial cells grown in vitro (HBECs), respectively. A large number of cells were analyzed and sequencing depth was adequate enough to detect weakly expressed genes, like the SARS-CoV-2 receptor, ACE2, in the broad array of cell types of lung.
ACE2 was found to be preferentially expressed in alveolar type II (called AT2 but we will stick to full name so as to not confuse with angiotensin receptor) cells.
TMPRSS2 was also found in alveolar type II cells, as reported before.
In the HBECs, a significant enrichment of the number of ACE2+ /TMPRSS2+ double-positive cells was observed. Hierarchy mapping suggested that these cells are a transitional secretory cell type. These cells not only have high co-expression levels of ACE2 and TMPRSS2, they are also enriched in RHO GTPases and related pathways, suggesting they may be especially vulnerable to SARS-CoV-2 infection.
The FURIN cleavage site in the SARS-CoV-2S protein may provide an priming mechanism, and Alveolar type II cells were strongly positive for FURIN while transient secretory cells had an intermediate level of expression.
A marked enrichment of the number of double and triple positive cells for any combination of ACE2, TMPRSS2, and/or FURIN expression was observed.
The authors acknowledge obviously limitations of the study, including lack of infection history, comparison of different sequencing modalities, and small sample size. Nonetheless, the dataset is likely to provide a useful tool to understand more about the molecular determinants of SARS-CoV-2 infection and identify new drug targets.
What is known about ACE2 in the lung?
One of the hypotheses for enhanced infectivity and severity of illness of SARS-CoV-2 is the abundance of ACE2 expression in a given tissue (such as the lung). Currently, no studies have demonstrated this to be the case. This is an area of research that is badly needed. It has been observed that older people and men have worse disease and higher death rates from COVID-19. Thus, a simple question to ask is, what happens to ACE2 levels in the lung over time and between sexes (male and female)?
A study from Xudong et al in Life Sciences (2006) examined ACE2 levels (via Western blot and immunohistochemistry) in Sprague Dawley rats (polyclonal rabbit anti-ACE2, Millenium Pharmaceuticals). The studied 3 different age groups and male/female. They reported that ACE2 levels decreased as the rats aged and that older male rats had lower levels than older female rats (Figure Below). The also show that ACE2 is predominantly expressed in alveolar epithelium, bronchiolar epithelium, endothelium and smooth muscle cells of pulmonary vessels with similar content; whereas no obvious signal was detected in the bronchiolar smooth muscle cells. Some important caveats are that these are rat studies with no disease. Thus, it is still unknown if this pattern occurs in humans or if the expression pattern is altered in disease.
Figure 2 from Xundong et al Life Sciences (2006)- rat lung tissue of male (A) & female (B). means ± SEM (n = 10). 3 m, young-adult group (3 months); 12 m, middle-aged group (12 months); 24 m, old group (24 months). **, P < 0.01; ***, P < 0.001.
What about ACE2 expression in human lungs?
The Genotype-Tissue Expression (GTEx) project aims to provide to the scientific community a resource with which to study human gene expression and regulation and its relationship to genetic variation. Below you can see the relative gene expression of ACE2 relative to other genes. You can see from the red box the lung expression in post mortem is low and not different between male and female.
From GTEx Portal (Broad Institute)
Does the use of ARBs/ACEi lead to increased ACE2 expression and does this change viral entry?
Kreutz et al in Cardiovascular Research 2020 have a very comprehensive table showing all known studies examining use of ARBs and ACEi and tissue/serum levesl in mostly animal models.
It has been shown in some animal models that both ACEis and ARBs can upregulate ACE2
Ferrario et al in Circulation (2005) showed that ACE2 mRNA expression increased in the left ventricle of normotensive Lewis rats after 12 days of lisinopril and losartan.
This Rat study from Ocaranza et al in Hypertension (2006) showed increased levels and activity of ACE2 in rat heart tissue after enalapril (ACEi).
This rat study of coronary artery ligation by Ishiyama et al in Hypertension (2004) showed increased levels of ACE2 mRNA after ARBs.
Mouse study from Soler et al in AJP Renal Physiology (2009) showed that ACE2 protein and mRNA was increased after the ARB temisartan
However, animal models have also shown no increase in ACE2 after ACEi or ARB.
Burrell et al in European Heart Journal (2005) showed no increase in ACE2 mRNA in hearts of Sprague–Dawley rats after coronary artery ligation and treatment with ramipril compared to control.
Burchill et al in Clin Sci (Lond) (2012) showed no increase in ACE2 mRNA and protein in rats after coronary artery ligation and treatment with valsartan, ramipril or both compared to control
Let’s take a look at the evidence in humans.
This study by Walters et al in EP Europace (2017) (in humans, not mice) did not find any association (increase of decrease) of circulating ACE2 levels with the use of ACEi or ARBs. from Dr. Louise Burrell.
Another study in PLOS one (Ramchand et al 2018) evaluated 79 patients with with coronary heart disease and measured plasma ACE2 levels. They did not find any correlation between RAS blockers or age with plasma ACE2 levels.
Cohort study performed in Japan by Furuhashi et al in American Journal of Hypertension (2015). 101 people with no medication (men/women: 40/61) who served as controls and 100 with hypertension (men/women: 42/58) who had been treated with a CCB, an ACE inhibitor, or an ARB for >1 year were enrolled in this study. amlodipine, long-acting nifedipine, enalapril, losartan, candesartan, valsartan, telmisartan, and olmesartan. Only the group taking olmesartan and not any other agent had increased urinary ACE2. Important caveats- This is a cohort study, so groups may not be comparable. This measures urinary concentration of ACE2 (soluble), not membrane bound ACE2. A cross-over study would be a better design.
Study by Vuille-dit-Bille et al in Amino Acids (2015) showed that patients on ACEis and not ARBs had increased ACE2 gene expression in the duodenum compared to controls not on ACEi/ARB.
How does Ang II interact with ACE2?
Deshotels et al in Hypertension examined how Ang II alters ACE2 expression in a cellular and rodent model. They found that ACE2 interacts with the type I angiotensin receptor (AT1 receptor, the target of ARBs). Acutely, Ang II decreased ACE2 expression in a tissue-dependent manner, by internalization into lysosomes (Deshotels Figure 3B). This internalization was prevented by AT1 receptor antagonist losartan or by blocking proteolytic degradation and was association with ubiquitination of ACE2.
Key points of this study (Figure):
AT1 receptor and ACE2 physically interact to form complexes on the cell membrane in the absence of excess Ang II (left panel).
potentially reducing interaction of virus with ACE2 (speculation with no data)
Ang II administration decreases the physical interaction between ACE2 and the AT1 receptor, and also induces ubiquitination and ACE2 internalization, and lysosomal degradation.
AT1 receptor antagonism with losartan prevented the internalization/degradation and ubiquitination of ACE2.
AT2 receptor and Mas expression did not affect ACE2 degradation
The clinical impact of this study remains uncertain, but this it could provide another mechanism by which ACE inhibition or ARBs could prevent COVID-19 viral entry. For example, the stabilization of the ACE2-AT1 receptor complex on the cell membrane by ARBs could either increase or decrease cellular viral entry by affecting viral protein binding and internalization. The effects of Ang II and ARBs on viral entry are not known at this time. These studies did not assess Ang-(1-7) generation or Ang II degradation, which may also play a role in preventing severe lung injury. It remains unknown if prevention of ACE2 internalization through this mechanism could also prevent viral infection with the SARS or COVID-19 virus. Research is needed to clarify these issues.
Can ARB/ACEi use (Less Ang II and potentially increased ACE2) actually be beneficial in coronavirus and other viral pneumonias?
This is an interesting question and has been looked at in both animal model and human studies (retrospective). This study looked at patients (humans) with viral pneumonia and demonstrated an association with improved outcomes in patients with continued ACEi use during viral pneumonia. However, you could argue that patients with viral pneumonia who had continued use of ACEi while hospitalized were not “as sick” as patients in which it was discontinued.
An animal study in Scientific Reports (mouse) demonstrated the importance of ACE2 in viral pneumonia from H7N9 influenza virus pneumonia. This study showed worse pathology and survival in ACE2 KO mice. Thus, showing the potential benefit of ACE2 in viral mediated lung injury presumably from the removal of Ang II and generation of Ang I-7.
An important article from Kuba et al from 2005 in Nature Medicine titled “A crucial role of ACE2 in SARS coronavirus–induced lung injury (Mouse Model)”
We will highlight a few of the findings from this paper
ACE2 is essential to coronavirus infection (Beijing strain, PUMC01 isolate) and lung injury (from ACE2 KO studies)
Decreased ACE2 and normal ACE levels are seen in lungs of mice with acid aspiration–induced acute lung injury (with addition of Spike protein)
this is mediated by the SARS-CoV Spike (S) protein
This Spike S protein lung injury is mediated by enhancing the renin-angiotensin system
decreased ACE2, increased Ang II
when blocked with AT1R blocker (losartan) improved lungs
Adapted from Figure 4 of Kuba et al Nature Medicine
This study in Scientific Reports which included both mice and pediatric (human) patients, report improved outcomes in with recombinant ACE2 administration in patients with RSV (respiratory syncytial virus) infections.
Liu et al in Science China Life Sciences (2020) examined the clinical and biochemical parameters
12 patients with COVID19 (4 females and 8 males) in Shenzhen Third People’s hospital
Ang II levels
higher in patients with COVID-19 compared to healthy controls
viral load (by CT value RT-PCR)
correlated with lung injury (PaO2/FiO2)
though the latter did not reach statistical significance (p=0.06). While this finding supports the hypothesis that Ang II is driving acute lung injury in COVID-19, as it did in SARS-CoV, there are several methodological limitations that hinder interpretation of these findings.
This was a small study with bivariate analysis that could not address potential sources of bias. Importantly, the ELISA used (Cloud-clone) to measure Ang II is not necessarily well validated. Furthermore, the authors did not describe their blood sample collection and processing methods which is a critical aspect of measuring components of the RAS.
What about the use of Ang II as a vasopressor?
We are beginning to see anecdotes (social media) and reports (Chow et al Anesthesia & Analgesia (2020) of utilizing Ang II as a presser in patients with COVID-19 and shock. Exogenous Ang II (marketed as Giapreza, see manufacturer website) is a potent vasopressor and increases mean arterial pressure in patients with certain types of shock. So what’s special about it in COVID-19?
The hypothesis purported (untested in COVID-19) is that, in shock, there is dysfunction in ACE and thus, less Ang II production. This is reported in a paper by Bellomo et al Critical Care (2020) in a post-hoc analysis of the ATHOS-3 trial (Khanna et al NEJM 2017) in patients with catecholamine-resistant shock. The most common reason for shock in ATHOS-3 was sepsis (no mention of viral pneumonia). 28% of the patients had evidence of ARDS on chest X-ray. The post-hoc analysis showed that patients with more Ang II (via Ang I/II ratio) did better than if this ratio was lower (less Ang II). Additionally, the above review by Chow et al cites studies showing that Ang II can down-regulate ACE2 (this has not been demonstrated in humans with exogenous Ang II). Thus, then extrapolating that administering exogenous Ang II might be beneficial in this setting.
However, the role of Ang II in viral illness and in particular COVID-19 remains controversial. As discussed above and in preprint (Liu et al in Science China Life Sciences 2020) it is hypothesized that Ang II could be a driving mediator of worsening COVID-19 since levels of Ang II correlate with disease severity and and animal models suggest that blockade of AT1 receptors is beneficial.
Thus, our suggestion is to avoid exogenous Ang II (if possible) in patients with shock/COVID-19 and only use in a research protocol especially given the interaction between SARS-CoV-2 and the renin-angiotensin system.
Does SARS-CoV-2 infect the kidney or other organs?
This is a concern because SARS-CoV-2 receptor, ACE2, is expressed at high levels in the proximal tubule. However, ACE2 is expressed on the apical membrane, and it is unclear if the virus is capable of gaining entry to the luminal compartment. Coronavirus entry into host target cells also requires fusion of the viral envelope with cellular membranes. Fusion-activate SARS-CoV peptides are created by specific proteolytic cleavage of the S proteins in a step called “priming.” As a consequence, cell infectivity not only depends on ACE2 expression but is also governed by types of proteases found in a given cell type. In the kidney, detectable levels of the TMPRSS2 (transcript mouse, Ransick et al. Developmental Cell (2019), which primes the SARS-CoV-2 S protein, are only detectable in the proximal tubule S3 segment. Even then, TMPRSS2 is expressed at very low levels. It remains to be determined if other TMPRSS or other proteases in the proximal tubule can mediate the priming of SARS-CoV-2.
A preprint by in medRxiv from Diao et al (2020) demonstrated in autopsies from patients who died with COVID-19 evidence of staining of SARS-CoV-2 nucleocapsid (NP) protein in the kidneys (tubules) via immunohistochemistry. However, the jury is still out on whether or not this leads to AKI or whether or not this is specific finding as no corroborating study (like RNA scope for instance) was performed. Go to the NephJC COVID-19 AKI Section for more information.
Now what about blocking ACE2 or TMPRSS2 or overwhelming with more soluble ACE2?
This is an interesting article looking at novel ways to block the interaction between coronavirus protein S and ACE2 from 2013.
A correspondence in The Lancet suggested the use of a AP2-associated protein kinase 1 inhibitor (regulator of endocytosis)/ janus kinase inhibitor called baricitinib could reduce viral entry.
Other strategies from this article
Spike protein-based vaccine
Inhibition of transmembrane protease serine 2 (TMPRSS2) activity
Blocking ACE2 receptor
Delivering excessive soluble form of ACE2
Dr. Dan Batlle has a paper in Clin Sci (Lond) (2020) discussing the potential therapeutic utility of overwhelming the body with more soluble ACE2 in an attempt to diminish viral infection into membrane bound ACE2.
Schematic of coronavirus (CoV) spike protein (S) binding to the surface receptor which is the full-length ACE2 from Batlle et al Clin. Sci (London) Soluble ACE2 administration may intercept the coronavirus.
Back to Humans, and what we can do
Do we have any evidence of renin-angiotensin system activation in COVID-19 so far?
A study, reported in pre-print form suggests that hypokalemia may be a prevalent feature of COVID-19.
The study was retrospective, small (175 patients) and focused on hospitalized patients in Wenzhou, China.
EMR data indicated ~22% of patients were severely hypokalemic (< 3 mM), ~40% had potassium of 3–3.5 mM, and ~38% were normokalemic (>3.5 mM).
Severity of hypokalemia seemed to associate with pre-existing disease, myocardial injury markers, and severity of COVID-19.
GI loss could not easily explain the hypokalemia.
Based on measurements of potassium/creatine ratios in spot urine of 20 patients with hypokalemia compared to 20 patients with normokalemia, it was concluded hypokalemia was caused by urinary potassium wasting.
It was suggested that hypokalemia does not respond well to oral potassium supplementation because kidney potassium losses persists until patients begin to recover, although this conclusion appears to be based on observations in two patients.
The authors speculated that urinary potassium wasting results from an increase in the RAS pathway secondary to the inhibitory effects of SAR-CoV-2 on ACE2. According to this scheme, elevated Ang II would increase aldosterone, which in turn, would activate the epithelial sodium channel, ENaC, in the distal nephron and drive potassium secretion. It should be pointed out that primary aldosteronism does not always lead to hypokalemia (see Welling aldosterone paradox thread). It will be important for future studies to careful measure all components of the RAS pathway. Pre-clinical studies on the role of ACE2 in potassium excretion should also be illuminating. Diuretics, which can be used treat hypertension or fluid overload in ARDS, might also contribute to potassium loss, and should be considered as important potential unmeasured confounders. Absence of disclosure about medications in the cohort is a major limitation of study.
This hypokalemia was also seen in patients with SARS-CoV infection in 2003. While this is strongly suggestive of increased Ang II and aldosterone activity via AT1 receptor and ENaC/MR, respectively, these initial studies did not measure components of the RAAS, including Ang II or aldosterone, which are necessary to confirm this likely pathophysiologic process. While there were no differences in baseline serum sodium in the patients described in the Chen preprint paper, other kidney sodium/water handling mechanisms may be intact.
Descriptions of baseline antihypertensive medications including ACEi and ARB as well as potassium-sparing diuretics, have not yet been reported in patients with COVID-19 to understand what is the exact mechanism of the hypokalemia reported in some studies. Interestingly, the largest case series (1099 patients) reported a fairly normal serum potassium (3.8 mmol/L) with no difference between those who had severe or non-severe disease.
In summary, this study provides reason to carefully monitor potassium levels in patients with COVID-19, although this observation must be confirmed in other centers. More data is needed to reach definitive recommendations. Nevertheless, efforts to maintain normokalemia may be warranted to limit muscle and cardiac sequela of COVID-19.
Below is a nice explainer video from Med Cram
Here is link to a podcast featuring Andrew South, MD (member of the team) on Casual Inference Podcast discussing ARBs and ACEis and COVID-19
Other Reviews and Viewpoints
A discussion from Malha et al in KI Reports (2020)
Verma and Patel discuss these issues in a JAMA Viewpoint (2020)
Vaduganathan et al in NEJM Special Report (2020)
ACE2 as a double-edged sword (Wang et al, Circulation, 2020)
ACEi/ARBs in COVID-19 (Sanchis-Gomar et al, Mayo Clinic Proc, 2020)
ACE/ACE2 imbalance as a hypothesis in COVID-19 (Sriram et al. British Journal of Pharmacology)
Kallikrein-kinin blockade in patients with COVID-19 (van de Beerdonk et al. eLife)
Potential harm of stoping ACEi and ARBs in COVID019 (Rossi et al eLife 2020)
Use of Renin-Angiotensin System Blockers During the COVID-19 Pandemic: Early Guidance and Evolving Evidence: What does this all mean? (Turgeon et al, Can J Cardiol, 2020)
We have more questions than answers right now. It is unclear whether or not continuing or stopping ARBs or ACEi’s is warranted in COVID-19. However, it is clear that ACE2 and TMPRSS2 are important targets for therapy. The link between hypertension is very much confounded by age. If you have more to add please let us know.
Patients who are taking ACEi or ARBs for hypertension are not advised to change their therapy unless advised to do so by their physician
The following is our opinion of the issue, as of now, we will update this if new evidence comes out.
Summary
The data on hypertension, RAS blockade, and COVID-19 are currently inconclusive. In humans, our current evidence comes from unadjusted or inadequately adjusted studies, making them highly prone to selection bias, misclassification bias, and confounding (e.g., due to age, duration and type of antihypertensive exposure, and comorbidity burden). We agree with the societal recommendations above not to change RAS blockade in those patients who are on it, unless there is another clinical indication to do so, due to a lack of sufficient evidence to draw firm conclusions at this time. Stay tuned as new data become available.
How to cite this
Sparks MA, Hiremath S et al. "The Coronavirus Conundrum: ACE2 and Hypertension Edition” NephJC http://www.nephjc.com/news/covidace2 with access date
List of Updates
(we are keeping track of updates from March 16th onwards in this list)
March 16: updated original description of SARS-CoV-2, Links to outpatient and inpatient clinical trials using losartan in COVID-19, added explainer video from Med Rants
March 17: preprint section of single cell RNA seq data added, Link to Casual Inferences Podcast added, Updated rACE2 pilot trial withdrawn.
March 18: Table of terms added, and data from Italian study of 355 deaths from COVID19
March 19: Added link to ASN Kidney News Online NephJC statement.
March 21: Minor typos fixed (terms, single cell RNA sections); link to EHJ review added, link to Drug Development Research commentary added (ARBs for COVID-19), added Furuhashi et al in American Journal of Hypertension (2015). Add section on SARS-CoV-2 and potential kidney involvement.
March 22: Added Brian Byrds discussion on peer review, added Xudong et al in Life Sciences (2006) examining ACE2 levels spanning age and sex in Sprague Dawley rats
March 23: Added Journal of Travel Medicine article RE speculative articles about ACEi and ARBs.
March 24: Study by Vuille-dit-Bille et al in Amino Acids (2015) showed that patients on ACEis and not ARBs had increased ACE2 gene expression in the duodenum compared to controls not on ACEi/ARB. Added a preprint from ChinaXiv from Wu et al (2020) shows how the enzyme furin plays an important role in this viral life cycle of SARS-CoV-2
March 26: Added section on Ang II as a presser (Chow et al Anesthesia & Analgesia (2020) and discussion surrounding this.
March 28: Added the NephJC workgroup’s CJASN Perspective, Clinical Infectious Diseases, and Lancet Respiratory Medicine Correspondence.
March 29: Added clinical outcome data from Cheng et al and Liu et al, added GTEx Human ACE2 expression data. Moved clinical ACEi/ARB to the clinical section. added Malha et al in KI Reports (2020), added retrospective/prospective studies of ACEi and/or ARB in COVID-19; Italian retrospective NCT04318418, UK Prospective Nested Case Control NCT04322786, Wuhan Retrospective Study ACEi/ARB NCT04318301
March 30: Added data from 3 studies to table on HT and COVID-19 association. Added statement from Australian Diabetes Society. Added Vaduganathan et al in NEJM Special Report (2020)
April 2: Added Coronavirus ACEi/ARB Investigation (CORONACION) randomizing patients already on RASi to stop or continue.
April 4: preprint from Yang et al MedRXiv, (2020, April 4) showing outcomes in single center in Wuhon, China of ARB/ACEi vs non ARB/ACEi. Added AJP Heart and Circ. Phys. and Nature Reviews Nephrology
April 8: Added REPLACECOVID US Multicenter Stop or Continue ACEi/ARB Randomized Trial NCT04338009 added. Added new GTex Broad institute ACE2 expression data in humans.
April 11: Added Table on 4 studies so far on ACEi/ARB use and clinical outcomes with discussion of unreliability
April 14: Updated table on outcomes with Bean et al MedRXiv; Updated table on HT with Bean et al MedRXiv and Petrilli et al MedRXiv; added graph showing relation of age and HT
April 15: Updated Kreutz et al in Cardiovascular Research 2020 have a very comprehensive table showing all known studies examining use of ARBs, ACEi, spiro and tissue/serum levels in mostly animal models.
April 16: The BIHS statement on ACEi/ARB added
April 18: Data from the Zhang et al CircRes paper added
April 20: Data from the Rentsch et al preprint added
April 28: added review article- ACE/ACE2 imbalance as a hypothesis in COVID-19 (Sriram et al. British Journal of Pharmacology). Added summary figure from Nature Reviews Nephrology.
May 4th: added 3 NEJM epidemiological studies to the outcome data; updated table; shortened discussion with archiving of smaller studies and preprints
June 2nd: Added links to Rossi et al eLife 2020 and van de Veerdonk et al eLife 2020, removed discussion of Mehra et al, NEJM 2020 given expression of concern in NEJM, updated contributors
June 3rd: Added links to Mehta et al, de Abajo et al and Khera et al in discussion of ACEi/ARB use and outcomes
June 4th: Added links to Turgeon et al (review); Retraction for Mehra et al (NEJM); Tables for ongoing trials added
July 4th: Removed Review of Statins and ARBs in Ebola. The commentary in Drug Development Research (2020) from David Gurwitz discusses the possibility of using ARBs as a treatment in COVID-19
Oct 4: Updated outcomes on two RCTs added