Abstract
Purpose Randomised trials of oral ivermectin have not shown benefit against COVID-19. The explanation most often advanced is pharmacokinetic: concentrations achievable at the respiratory epithelium after oral dosing fall well below those at which antiviral activity is seen in cell culture. If so, local application to the site of initial viral replication is a hypothesis oral trials cannot test. We characterised the registered evidence base to ask whether it has been evaluated. Methods We analysed all 253 ClinicalTrials.gov records for the intervention term “ivermectin”, then repeated the analysis against the WHO ICTRP to test dependence on a single registry. We derived indication, purpose and prophylactic intent from registry fields. Because automated inference misattributes a nasal route to records whose nasally applied agent is a different substance, we adjudicated the route of ivermectin administration at record level. Results Viral indications accounted for 97 of 253 ClinicalTrials.gov records (38.3%), of which 93 concerned COVID-19. Fourteen (5.5%) were adjudicated as primary prophylaxis of infection acquisition, none delivering ivermectin to the respiratory mucosa. Five gave ivermectin nasally or by inhalation (planned n = 467); all carried the status unknown, none posted results, none had a preventive objective, and no record described a semisolid nasal dosage form. Results were posted for 16 of 97 viral-indication studies (16.5%), against 30 of 129 (23.3%) for parasitic indications. ICTRP added 184 records from fourteen further registries, a combined corpus of 437 with 173 viral-indication and 28 prophylaxis studies. All 14 additional prophylaxis studies gave ivermectin orally; none administered it nasally. Conclusion These data do not establish that ivermectin prevents viral infection; they establish, across two registries, that the local-delivery hypothesis is untested. A nasal semisolid is a legitimate object of early-phase investigation, provided epithelial-model activity and adequate target-site exposure are shown before any efficacy claim.

This publication has 37 references indexed in Scilit: