Document resource
Background and Importance Direct oral anticoagulants (DOACs) are used for the prevention of thromboembolic events. They may interact with concomitant substances, including herbal products, with variable severity. Drug interaction databases often provide inconsistent information, leading to uncertainty in clinical decision making.Aim and Objectives To assess the level of agreement in the classification of interactions between herbal or food products and DOACs across four drug interaction databases.Material and Methods A descriptive cross-sectional study was performed to identify potential interactions between herbal or food products and DOACs (apixaban, rivaroxaban, edoxaban, dabigatran) listed in four drug interaction databases (Lexicomp, Medinteract, Micromedex, Drugs.com). Interaction severity was standardised into four categories: Severe, moderate, minor and undetermined.Abstract 4CPS-026 Table 1Statistical analyses were performed with STATA BE/17.0. Overall agreement among the four databases was assessed using Gwet’s AC with quadratic weights, while Fleiss’ Kappa with quadratic weights was applied for sensitivity analysis. Agreement levels were interpreted according to Altman’s thresholds: ≤0.20 poor, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 good, ≥0.81 excellent.Results A total of 1992 potential interactions between DOACs and molecules were identified, of which 149 (7.5%) involved herbal or food products: apixaban 42 (28.2%), dabigatran 42 (28.2%), edoxaban 23 (15.4%), rivaroxaban 42 (28.2%).Overall agreement between the four databases was moderate with Gwet’s AC (range 0.55–0.57) but poor with Fleiss’ Kappa (range 0.05–0.10).Apixaban: Gwet’s AC: 0.57. Fleiss’ Kappa: 0.10.Dabigatran: Gwet’s AC: 0.55. Fleiss’ Kappa: 0.05.Edoxaban: Gwet’s AC: 0.56. Fleiss’ Kappa: 0.08.Rivaroxaban: Gwet’s AC: 0.56. Fleiss’ Kappa: 0.10.Conclusion and Relevance Drug interaction databases showed inconsistent classification of DOAC–herbal product interactions, with overall agreement ranging from poor to moderate depending on the statistical method applied. These discrepancies highlight the absence of harmonised criteria and limit the reliability of database alerts. Pharmacists should critically appraise interaction information before making clinical recommendations. Standardisation of classification systems and further research on DOAC–herbal interactions are needed to strengthen clinical decision support.Conflict of Interest No conflict of interest