Data Extraction and Quality Assessment
Data from the eligible studies were subsequently extracted by two authors (NA, AK) into a spreadsheet including information on the study characterises (study design, setting, sample size, population, exposure-ACEIs/ARBs, ACEIs, or ARBs) and outcome measures including death, intensive care unit (ICU) admission, risk of COVID-19 infection, severe COVID-19 infection, severe pneumonia, hospitalisation, hospital discharge, use of ventilators, duration of hospital stay, septic shock, acute kidney injury, cardiac injury, and hospital readmission. Since the need for using ventilators typically necessitates ICU admission, we combined studies that reported ICU admission and ventilator use as a further composite outcome measure. Two authors (NA and LA) independently conducted the assessment of risk of bias using the Newcastle-Ottawa Scale (NOS) for nonrandomised studies (32), whereby studies were classified into good, fair and poor quality (33). Some of the co-authors have used this approach before (34).