Review characteristics
Forty-six reviews (97.9%) compared COVID-19 related outcomes between ACEI/ARB users vs. non-users among patients with COVID-19 (4-6, 16-51, 53-59), one study (2.12%%) compared outcomes between ACEIs/ARBs users in patients with and without COVID-19 infection (52)), and 16 studies (34.0%) explored both (6, 18, 24-26, 39, 40, 42, 43, 47, 49, 50, 53, 55, 57, 59). Definition criteria for COVID-19 diagnosis was reported by only six (12.8%) reviews as laboratory confirmed diagnosis based on a reverse transcriptase–polymerase chain reaction, whereas the remaining 41 (87.2%) reviews did not report any criteria for COVID-19 diagnosis definition. Most of the included reviews were peer-reviewed publications (68.1%; n=32), whereas the remining 15 (31.9%) reviews were non-peer reviewed publications (i.e. were published in a pre-print database) (16-18, 20-22, 29, 31-33, 35, 45, 49, 53, 59). The time the searches were conducted ranged from April 2020 to October 2020, with 21 (44.7%) review searches conducted in the month of May 2020 (4-6, 16, 20, 22, 23, 27, 29-31, 34, 35, 39-41, 43, 45, 47, 49, 53) Pre-print articles were included in 28 (59.6%) reviews (4, 16, 18-21, 24, 25, 29, 32, 36, 40-44, 46-52, 54, 55, 58, 59), and 10 (21.3%) reviews adjusted for retracted studies (4, 17, 30, 39, 44, 46-49, 55). Full details of the 47 reviews are presented in Supplementary file 3 .
A total of 213 meta-analyses were conducted by the 47 reviews (Supplementary file 4 ). In terms of number of COVID-19 related outcomes reported in each review, one outcome was reported by 13 reviews (27.7%) (17, 19, 20, 22, 23, 27, 28, 37, 38, 46, 51, 52, 60), two outcomes by 15 reviews (31.9%) (4, 16, 25, 30, 31, 33-36, 39, 41, 48, 53, 54, 57), three outcomes by 11 reviews (23.4%) (6, 21, 24, 26, 32, 43-45, 49, 55, 59) and 4-9 outcomes by eight reviews (17%) (18, 29, 40, 42, 47, 50, 56, 58). Overall, the 47 eligible reviews reported data on 18 unique pooled outcome estimates including death in 36 reviews, reviews (4, 6, 16-18, 21, 23, 24, 26, 29-38, 40-48, 53-55, 57-59), ICU admission in nine reviews (26, 27, 29, 40, 42, 47, 50, 55, 58), death/ICU admission as a composite outcome in 16 reviews (4, 19, 20, 22, 25, 28, 30, 31, 39, 40, 42, 44, 50, 54, 58), risk of acquiring COVID-19 infection in 15 reviews (18, 24, 26, 39, 40, 42, 43), severe COVID-19 infection in 22 reviews (6, 16, 18, 21, 24, 29, 32-36, 40-45, 47, 58, 59), hospitalisation in nine reviews (18, 29, 40, 42, 47, 58), length of hospital stay in five reviews (18, 21, 29, 45, 58), use of mechanical ventilator in three reviews (29, 40), risk of severe acute respiratory syndrome (SARS) in two reviews (25, 58), and each of hospital discharge (29), ICU admission/mechanical ventilator use (40), risk of COVID-19 infection/hospitalisation (52), severe pneumonia (40), level of serum creatinine (56), d-dimer (56), cough (56), fever (56) and renal dialysis (58) in one review; accordingly, nine out of these 18 outcomes were included in the meta-analysis as they were reported by at least two reviews. In terms of the exposure, ACEIs and ARBs were evaluated as one class (ACEIs/ARBs) in all the eligible 47 reviews apart from three (25, 52, 56), and as separate classes in 17 (4, 6, 22, 24-26, 29, 30, 37, 39, 40, 42, 46, 49, 52, 53, 57) and 16 (4, 6, 22, 24-26, 29, 30, 37, 39, 40, 42, 49, 52, 53, 57) reviews, respectively. Majority of the reviews (66%; n=31) only evaluated one exposure, mainly ACEIs/ARBs combined as one class (n=30); whereas one third of them (29.8%; n=14) reported data for the three level of exposure (ACEIs/ARBs, ACEIs, ARBs).