Table 2.

Published evidence (as to May 12, 2020) addressing the suggestions that the use of ACE-inhibitors and ARBs might be harmful in the context of SARS-CoV-2 susceptibility to infection and in the prognosis of patients with Covid-19.

Author (Country)	Period	Population	Study design	Main results	Additional results	Authors' conclusions	Strengths and Limitations
Bean [33] (United Kingdom)	01-Mar to 22-Mar, 2020	Inpatients with Covid-19 at King's College Hospital and Princess Royal University Hospital in London. Mean age 63 yrs; 52% men; 51.2% with hypertension.	Case-series (n = 205) follow-up 7-days from symptoms onset to primary end point (death or transfer to ICU). 53 patients reached end-point.	Lower death rate or transfer to ICU in patients on an ACE-i (OR 0.29, 95% CI 0.10 to 0.75, p < 0.01), adjusting for age, gender, comorbidities (hypertension, diabetes mellitus, ischaemic heart disease and heart failure).	–	No evidence for ACE-i increasing the short-term severity of Covid-19 disease.	Small sample size. Short follow-up.
Conversano [11] (Italy)	27-Feb to 17-Mar, 2020	Patients admitted to San Raffaele Hospital in Milan with confirmed SARS-CoV-2 pneumonia. Mean age: 63.4 yrs; 68.6% men.	Retrospective case-series (n = 191), 96 (50.2%) with hypertension	Predictors of death in Cox models (HR, 95% CI). All (n = 191): Age (1.1, 1.0–1.2); HF (3.1, 1.3–7.5); CKD (2.1, 1.1–4.0). Hypertensive (n = 96): Age 1.1 (1.0–1.1); HF 2.8 (1.1–6.9).	No association between use of ACE-i/ARB with mortality in both groups in multivariate models.	In multivariate model, age, heart failure and CKD (but not hypertension, diabetes and current use of ACE-i/ARB) were independent predictors of mortality.	Multivariate models deal with important confounders. Small sample size. Short follow-up.
Guo [8] (China)	23-Jan to 23-Feb, 2020	Hospitalised patients with Covid-19 at the Seventh Hospital of Wuhan City. Mean age 58.5 yrs; 48.7% men.	Retrospective case series (n = 187), 32.6% with hypertension.	–	Death rates with and without use of ACE-i/ARB 36.8% (7/19) and 25.6% (43/168), respectively.	Not primary objective of the study.	Incidental finding. Small sample. Risk of bias and confounding.
Li [9] (China)	15-Jan to 15-Mar, 2020	Hospitalised patients with Covid-19 at the Central Hospital of Wuhan. Median age 55.5 yrs; 46.3% men.	Retrospective case series (n = 1178), 362 (30.7%) with hypertension.	In-hospital mortality 21.3%. Percentage taking ACE-i/ARB did not differ between those with severe and non-severe infections (32.9% vs 30.7%; P = 0.645) nor between non-survivors and survivors (27.3% vs 33.0%; P = 0.34).	Similar findings when data analysed for patients taking ACE-i and ARB.	ACE-i/ARB not associated with severity or mortality of Covid-19 in patients with hypertension.	Small number with hypertension taking ACE-i/ARB who were hospitalized with Covid-19. Not generalisable to all patients with hypertension. Risk of bias and confounding.
Mancia [10] (Italy)	21-Feb to 11 Mar, 2020	Residents in Lombardy, 40+ years, beneficiaries of the Regional Health Service as target population (>6 M, ~17% of the Italian population of that age). Mean age 68 yrs; 37% women.	Population-based case–control study (6272 cases matched to 30,759 controls). Case: confirmed infection with SARS-CoV-2. Control: non-infected matched by age, sex and municipality.	[1] No association between use of RAAS blockers and Covid-19 (adjusted OR 0.95, 95% CI: 0.86 to 1.05) for ARB and 0.96, 0.87 to 1.07) for ACE-i or among patients who had severe or fatal disease (adjusted OR 0.83, 0.63 to 1.10) for ARB and 0.91, 0.69 to 1.21 for ACE-i. [2] No sex difference.	[1] Use of ACE-i and ARBs more frequent among patients with Covid-19 than controls. [2] Results apply to both sexes, younger and older. [3] Loop diuretics associated with an increased risk of Covid-19.	No evidence that ACE-i or ARB affect the risk of Covid-19 or its severity.	Large sample size, well-matched controls. Drugs prescribed to outpatients accurately recorded. Information on drug use limited to prescriptions, not actual consumption. Drugs purchased by the patients and those dispensed after Dec 31, 2019, not captured. Unmeasured confounders.
Mehra [12] (Asia, Europe, N America)	20-Dec 2019 to 15-Mar 2020, deaths at 28-Mar 2020	Patients with Covid-19 hospitalised in 169 hospitals in 3 continents.	Observational database (n = 8910), 1536 from N America, 5755 from Europe and 1619 from Asia).	515 died and 8395 survived to discharge. No increased risk of in-hospital death associated with the use of ACE-i (2.1% vs. 6.1%; OR 0.33; 95% CI, 0.20 to 0.54) or ARBs (6.8% vs. 5.7%; OR 1.23; 0.87 to 1.74).	Age>65 yrs, CHD, cardiac arrhythmia, COPD, smoking independently associated with increased risk of in-hospital death.	Results did not confirm concerns of harmful association of ACE-I or ARB with in-hospital death.	Large sample size. International representation of patients. Cannot exclude confounding. No pre-specified hypothesis, increased probability of chance finding. Cause–effect relationship between therapy and survival should not be inferred.
Mehta [13] (USA)	8-Mar to 12-Apr, 2020	All patients tested for SARS-CoV-2 at the Cleveland Clinic Health System in Ohio and Florida. Mean age 49 yrs; 40% men; 69% white.	Retrospective cohort analysis (n = 18,472), 2285 taking either ACE-i or ARB, with overlap propensity score (OPS) weighting.	OPS weighting showed no significant association of ACE-i and/or ARB use with SARS-CoV-2 test positivity (OPS–weighted OR 0.97; 95% CI, 0.81–1.15). For use of ACE-i only, OPS–weighted OR 0.89; 0.72–1.10.	Higher likelihood of hospital admission in patients with positive results who were taking either ACE-i/ARB (OR 1.84; 1.22–2.79), or ACEIs only (OR 1.77; 1.07–2.92).	No association between ACE-i or ARB use and SARS-CoV-2 test positivity.	Risk of confounding. No information on duration of drug use before and after testing. Data reflect medication list in the electronic medical record, not actual use, risk of non-ascertainment bias. Most patients white, limited generalisability to other ethnic groups. Small sample size and wide confidence intervals for data on severity.
Meng [14] (China)	11-Jan to 23-Feb, 2020	Covid-19 patients with treated hypertension. Median age 64.5 yrs; 57% men.	Retrospective review of medical records (n = 42 out of 417) admitted to Hospital in the period.	During hospitalisation, 12 patients in the non ACE-i/ARB group (48%) were severe and one died. In the ACE-i/ARB group only 4 patients (23.5%) were severe and none died.	ACE-i and ARB associated with increased CD3 and CD8 T cell counts and lower peak viral load.	Patients receiving ACE-i or ARB had lower rate of severe diseases.	Observational case-series. Small sample, open to bias and confounding.
Reynolds [15] (USA)	01-Mar to 15-Apr, 2020	All patients in the New York University (NYU) Langone Health electronic health record who had Covid-19 test results recorded.	Observational study using Electronic Health Records (n = 12,594), 5894 with positive test and 1002 with severe illness (defined as admission to ICU, mechanical ventilation or death).	[1] Among 12,594 patients, 4357 (34.6%) had hypertension; of these, 2573 (59.1%) were Covid-19–positive and 1784 (40.9%) were Covid-19–negative. The results in matched patients ruled out an absolute difference of >10% in the likelihood of a positive test with at least 97.5% certainty for ACE inhibitors, ARBs, beta-blockers, calcium-channel blockers, and thiazide diuretics. [2] Among the 2573 patients with hypertension and positive results on Covid-19 testing, 634 (24.6%) had severe Covid-19. The likelihood of severe Covid-19 among patients who tested positive, matched on propensity score for previous treatment, ruled out a ≥10% in the likelihood of severe illness with at least 97.5% certainty for all the medication classes.	–	No increase in likelihood of a positive test for Covid-19 or risk of severe Covid-19 in patients who tested positive in association with five common classes of anti-hypertensive medications, including ACE-i and ARB.	Propensity-score matching creates balance between comparison groups. Variations in diagnostic characteristics for Covid-19 testing methods may have led to misclassification of Covid-19 status. Some patients underwent multiple tests, increased likelihood of identifying disease. Possible Covid-19 are not tested, particularly those who are not hospitalised. Results may overestimate proportion of severe Covid-19. Ascertainment of medication use may not reflect actual drug exposure.
Richardson [5] (USA)	01-Mar to 04-Apr, 2020	Patients admitted to 12 hospitals in NYC, Long Island and Westchester County, within the Northwell Health system with Covid-19 confirmed positive to SARA-CoV-2. Median age 63 yrs; 39.7% women; 56.6% with hypertension.	Case-series (n = 5700). Outcomes assessed in 2634 who were either discharged or died (latter n = 553).	–	[1] Of 2,411, 189 were on ACE-i and 267 on ARBs. During admission, 91 and 136, respectively continued their RAAS blocker as in-patients. [2] Suppl. Table 2. Hypertensives on ACE-i: died 55, discharged 113; on ARB: died 75, discharged 170. Not on ACE-i/ARB: died 254; discharged 699.	This case series design cannot address the complexity of the question.	Results are unadjusted for common confounders. Study population only included patients within the NYC metropolitan area. Electronic health record database. Brief post-discharge follow-up. Clinical outcomes only available in 46.2% of patients.
Zhang [34] (China)	31-De, 2019 to 20-Feb, 2020. Final date of follow up 07-Mar, 2020	Patients with hypertension diagnosed with Covid-19 and admitted to 9 hospitals in Hubei Province.	Retrospective multi-center study (n = 1128), 18 taking ACE-I/ARB and 940 not on them.	Risk for all-cause mortality lower in the ACE-i/ARB group vs non-ACE-i/ARB group (adjusted HR 0.42; 95% CI 0.19–0.92; P = 0.03). In a propensity score-matched analysis followed by adjusting imbalanced variables, lower risk of Covid-19 mortality in patients who received ACE-i/ARB vs those who did not receive ACE-i/ARB (adjusted HR 0.37; 0.15–0.89; P = 0.03).	Compared to use of other antihypertensive drugs, ACEI/ARB was associated with decreased mortality (adjusted HR, 0.30; 95% CI, 0.12–0.70; P = 0.01).	Inpatient use of ACE-i/ARB was associated with lower risk of all-cause mortality compared with ACE-i/ARB non-users	May not reflect outpatient management. Modest sample size in those on ACE-i/ARB. Some parameters not available in all patients. Antihypertensive drugs not matched or adjusted when comparing ACE-i/ARB and non-ACE-i/ARB groups. [5] Not retrieving pre-hospital self-medications.