Systematic review of preclinical evidence reveals the anti-inflammatory potential of Icariin based on meta-analysis and machine learning

Abstract

Purpose: The aim of this study was to construct a reference system for preclinical evidence and drug superiority characterization of the anti-inflammatory effects of icariin glycosides and their derivatives through meta-analysis combined with machine learning, and to excavate the biological mechanisms behind them. Methods: The data were obtained from databases of PubMed, Embase, Cochrane Library, and Web of Science. STATA software was used for meta-analysis of indicators, and subgroup analysis was conducted based on animal species, gender, type of disease, drug dosage, and course to obtain more particulars. Furthermore, model construction were performed using R software to explore influential features on drug efficacy. In addition, the pharmacological mechanisms of icariin and its derivatives in anti-inflammation were summarized based on a comprehensive understanding of relevant literature. Results: After searching and screeningThe results showed that icariin and its derivatives significantly inhibit inflammation indicators such as TNF-α and IL-1β. Besides, machine learning with TNF-α as the output variable showed that icariin and its derivatives had stronger anti-inflammatory effects when the type of disease was respiratory, urological, neurological, and digestive, and when the dose and duration of Icariin were greater than 27.52 mg/kg/day and 31.22 days, respectively. Conclusion: Icariin and its derivatives demonstrate strong anti-inflammatory effects, particularly for respiratory, urinary, neurological, and digestive disorders. When given at doses of 27.52 mg/kg/day or more, with treatment lasting 31.22 days or beyond, these compounds hold significant potential as drugs for inflammation inhibition across multiple dis