It is currently recognized that an injudicious strategy in the last decades has been not only focusing of research typically on caries in children, but also the narrow focusing on fluoride, because despite sufficient availability of fluoride in water and oral healthcare products, caries levels escalate steadily as people get older and caries remain a main public health issue to be settled. In the last two decades the scientific community intensified efforts of exploring other products for caries prevention, herbal products being one of these approaches. Because preliminary evidence indicated that clinical trials for caries prevention with herbal products are heterogeneous in design, quality and products evaluated, we performed a scoping review intended to explore the main characteristics of such clinical trials. From an initial collection of 1986 unique papers from different literature databases, 56 articles satisfied the inclusion and exclusion criteria. The species investigated, dosage forms, study designs, duration of intervention, controls, endpoints, quality of reporting and risk of bias are discussed. 85.71% of the trials reviewed here reported positive results but given the methodological flaws and biases affecting them, it is difficult to conclude on the efficacy of those products based on the studies published thus far.

Drug induced liver injury (DILI) remains one of the challenges in the safety profile of both authorized drugs and candidate drugs and predicting hepatotoxicity from the chemical structure of a substance remains a challenge worth pursuing, being also coherent with the current tendency for replacing non-clinical tests with in vitro or in silico alternatives. In 2016 a group of researchers from FDA published an improved annotated list of drugs with respect to their DILI risk, constituting “the largest reference drug list ranked by the risk for developing drug-induced liver injury in humans”, DILIrank. This paper is one of the few attempting to predict liver toxicity using the DILIrank dataset. Molecular descriptors were computed with the Dragon 7.0 software, and a variety of feature selection and machine learning algorithms were implemented in the R computing environment. Nested (double) cross-validation was used to externally validate the models selected. A number of 78 models with reasonable performance have been selected and stacked through several approaches, including the building of multiple meta-models. The performance of the stacked models was slightly superior to other models published. The models were applied in a virtual screening exercise on over 100,000 compounds from the ZINC database and about 20% of them were predicted to be non-hepatotoxic.