Supplementary Data — Ivermectin reduces COVID-19 risk: real-time meta-analysis of 106 studies (c19ivm ivmmeta)

@CovidAnalysis, January 22, 2026, Version 246V246

Fig. S1. Random-effects meta-analysis for primary outcomes (as defined before the trial started).

Fig. S2. Random-effects meta-analysis for peer-reviewed studies after exclusions. Effect extraction is pre-specified, using the most serious outcome reported, see the appendix for details. Analysis validating pooled outcomes for COVID-19 can be found below.

Fig. S3. Random-effects meta-analysis for RCT studies after exclusions. Effect extraction is pre-specified, using the most serious outcome reported, see the appendix for details. Analysis validating pooled outcomes for COVID-19 can be found below.

Fig. S4. Random-effects meta-analysis for mortality after exclusions.

Fig. S5. Random-effects meta-analysis for mechanical ventilation after exclusions.

Fig. S6. Random-effects meta-analysis for ICU admission after exclusions.

Fig. S7. Random-effects meta-analysis for hospitalization after exclusions.

Fig. S8. Random-effects meta-analysis for recovery results only after exclusions.

Fig. S9. Random-effects meta-analysis for COVID-19 case results after exclusions.

Fig. S10. Random-effects meta-analysis for viral clearance after exclusions.

Fig. S11. Random-effects meta-analysis for studies grouped by strongyloides prevalence. Data is by country and from Buonfrate. Effect extraction follows the same pre-specified protocol as detailed in the appendix.

Fig. S12. Random-effects meta-analysis with SSC exclusions. SSC has not reviewed late treatment and prophylaxis trials. SSC exclusions are from Nov 27, 2021.

Fig. S13. Random-effects meta-analysis with GMK exclusions. Our main exclusion analyses already exclude all studies where the GMK team believes there are major data issues. This analysis corresponds with GMK's recommendation for meta-analysis as of Oct 26, 2021. GMK excludes most non-RCT studies, with the notable exception of several studies with major issues that report negative or relatively poor results — Szente Fonseca which is likely affected by multicollinearity among treatments, Elavarasi which reports unadjusted results with no group details and is subject to confounding by indication, and Soto-Becerra which has several major issues described in the details.

Fig. S14. Comparison of results for RCTs versus observational studies. For COVID-19 treatments, there is no significant difference between the results of RCTs and observational studies. Observational studies do not systematically over or underestimate efficacy. For high-cost treatments, there is a non-significant trend towards RCTs showing greater efficacy.