Introduction: Our systematic review and meta-analysis address gaps in UV-C viral inactivation protocols in liquid solutions to understand both the influencing characteristics of viruses and related experimental setups. Key issues include inconsistent reporting of experimental setups, virus identification, terminology, and dose–effectiveness relationships, which have all bee examined in the literature (Chiappa et al., 2021; Freeman et al., 2022; Masjoudi et al., 2021; Raeiszadeh & Adeli, 2020). A clear distinction between "inactivation" and "disinfection" is crucial. We propose defining complete inactivation as an undetectable viral load while preserving viral and protein structures to improve research applications.

Methods: A PICO search (April 10, 2024) across five databases (N=2814) yielded 1,686 records after duplicates were removed adhering to the PRISMA, GRADE, and MECIR protocols. We reviewed 309 full texts, with 33 studies meeting our inclusion criteria for strain-specific viruses, defined UV-C setups, and quantifiable viral reductions. Random-effects meta-regression of 18 studies analyzed strain-specific inactivation, while Spearman correlation was used to examine terminology usage.

Results: UV-C effectively (254nm being the most effective) reduced viral loads but faced inconsistencies in experimental setups, viral conditions, and dosages. Virus-specific dose–response relationships varied, leading us to focus on only SARS-CoV-2 strains for further calculations. Our semantic analysis revealed the interchangeable use of "inactivation" and "disinfection". The issues identified in our meta-analysis—0-0 distortions, dose saturation, and PFU/TCID inconsistencies—necessitate non-linear models and data refinement. Despite the variability, the preliminary regression (R² = 0.49) suggested predictable outcomes.

Conclusions: We propose a reporting framework to improve reliability and cross-study comparability. Inactivation studies must start with higher viral loads to account for plaque assay and TCID₅₀ detection limits, and the “delivered dose” should only be reported when measured; otherwise, “calculated dose” is correct. Reports must include TCID/plaque assay details, log-transformed data (e.g., log₃ = base 3), and at least five data points for inactivation curves. "Inactivation" should be used for laboratory studies maintaining viral structures, while "disinfection" suits hospitals and water treatments.