Evaluation of historic in vivo data to characterise the emetic properties of liquid cleaning products and provide a framework for the development of an in silico predictive algorithm.

Accidental ingestion of household cleaning products frequently results in emesis but the physicochemical properties responsible are not known. To investigate whether data collected during in vivo animal studies performed >30 years ago could provide novel insights into the components responsible, we re-analysed original studies from a total of 74 liquid cleaning formulations. The incidence of emesis was dose-related with ED values between 0.012 and 8.4 ml/kg and 57% of formulations having an ED ≤ 1 ml/kg. The median latency for emesis was 10.0 min (95% CI, 8-12 min) and number of vomits in 60 min ranged from 1 to 10 (median 2). From the ED, latency and number of vomits we derived a "vomiting index" (VI) for a subset of 15 formulations which revealed an association between a high VI, a high percentage of non-ionic surfactants/high ionic strength, and a pH of ~10 which we propose are causally linked with the possible mechanism(s) discussed. The limitations of using historic data are discussed but analysis of such data has provided novel insights into the emetic characteristics of this class of products and has informed the development of an in silico model to predict the emetic liability of novel formulations without additional in vivo studies.