Mapping the MicroRNA Landscape of Arsenic Exposure: A Bibliometric and Evidence-Synthesis Analysis of Directional Consistency and Regulatory Themes

Arsenic exposure has been associated with altered microRNA (miRNA) expression in human and experimental settings, yet the structure of this literature and the consistency of reported expression directions remain insufficiently characterized. This study combined a three-database bibliometric analysis with an audit of published directional evidence. Searches of PubMed/MEDLINE, Scopus, and Web of Science on 27 August 2026 identified 1396 records, yielding 752 unique records and a 248-publication bibliometric corpus. Author-keyword analyses highlighted arsenic, microRNA regulation, carcinogenesis, and epigenetic themes, with miR-21 the most frequent individual miRNA keyword. A historical biological inventory contained 319 coded study–miRNA observations from 109 publications; its significance-conditioned selection prevents interpretation as a completed, unbiased evidence synthesis. Although legacy miR-21 directional concordance was 26/27, these codes do not establish independent biological reproducibility. Source-level adjudication of 11 of 137 previously unresolved records identified 10 exclusions and one eligible publication awaiting extraction; 126 records remain unresolved. Re-extraction of a human urinary-miRNA study recovered 18 signed arsenic-association estimates across three miRNAs, one of which (miR-423) was absent from the legacy inventory altogether. All were nonsignificant, but positive and negative estimates occurred across adjustment models, demonstrating that nonsignificance does not imply absence of effect. Aggregating the legacy miR-21 codes into senior-author units reduced the concordance to 8/9 (p = 0.039), and under a modest assumption of upward-reporting asymmetry (p0 = 0.65) the signal was no longer significant (p = 0.121). Mechanism-frequency ranking, aggregated false-discovery-rate control and mature-arm assignment are not supported by the available source data and are therefore not reported. The bibliometric mapping and source audit clarify research themes and reporting limitations, but definitive directional-consistency and biomarker-prioritization conclusions require completed source-level screening, significance-independent extraction, and verification of study independence.