Key points are not available for this paper at this time.
Objective This study evaluates the safety profiles of Withania somnifera (Ashwagandha), an adaptogenic herb prevalent in Ayurvedic medicine, focusing on liver and reproductive toxicity. Utilizing advanced AI methodologies, we conducted a comprehensive meta-data analysis to assess the safety of the plant’s root and non-root parts, comparing Ashwagandha’s safety to other herbal supplements. Methods We employed natural language processing (NLP) to systematically review existing literature and utilized quantitative structure–activity relationship (QSAR) models to predict liver and reproductive toxicity at the molecular level. Special attention was given to withanolides, the bioactive compounds in Ashwagandha, due to conflicting safety information. Additionally, we reviewed case studies reporting liver toxicity, noting that many involved supplements containing both leaves and roots, complicating the identification of the toxicity source. Results Our analysis indicated that Ashwagandha root exhibits a superior safety profile compared to non-root parts, particularly concerning liver and reproductive toxicity. When compared to a broad set of other herbal supplements, Ashwagandha root was found to have a better safety profile than most, making it a first-choice ingredient for safe and effective use in supplements. While non-root parts of Ashwagandha showed higher toxicity potential than the root, their safety profile was still comparable to other edible plants and herbal supplements. Conclusion This study suggests that the root of Withania somnifera (Ashwagandha) demonstrates a favorable safety profile, particularly concerning liver and reproductive toxicity, when compared to other herbal supplements. Our findings support the traditional preference for root-based formulations and highlight the importance of distinguishing between plant parts in safety assessments. While these results strengthen the evidence supporting the safe use of Ashwagandha root, further experimental and clinical validation would be valuable to confirm these AI-driven predictions and literature-based findings. The study also illustrates how artificial intelligence approaches can complement traditional toxicological evaluations and enhance safety assessment frameworks in the herbal supplement industry.
Ebert et al. (Mon,) studied this question.