Publication bias
For the primary outcome referring to ACE inhibitors, funnel plot did not show apparent visual asymmetry (Supplementary figure 2). The Begg tests showed no evidence of publication bias (P=0.66). As for the primary outcome referring to ARBs, funnel plots did not show any asymmetrical tail (Supplementary figure 3). Likewise, Begg tests showed no evidence of publication bias (P=0.21). For the secondary outcome, funnel plot did not show visual asymmetry (Supplementary figure 4).; testing for publication bias showed similar results (Begg’s P-value for asymmetry 1.00). As for the COVID-16 studies referring to primary outcome, Begg’s test for asymmetry was also not significant (P =0.08).