Factors affecting the measurement of stability and safety of cosmetic products.

Abstract

There are a large number of products that come under the heading ‘cosmetic product’. Each is required, according to the EU Cosmetic Regulation, to demonstrate formulation stability to a EUROTOX Safety Assessor before being placed on the market. The regulation places a requirement on the Safety Assessor to take into account the long-term stability of the product in question but does not specify any protocol by which to obtain this data. Various guideline documents have been written, most notably by ISO 18811:2018 and Cosmetics Europe (Colipa), which use elevated temperature stress testing to accelerate reactions, and using the Arrhenius model to extrapolate duration of shelf lives from the results. More specifically the assumption is made that any reaction observed displayed 1st order rate kinetics with respect to temperature and that that behaviour can be quantified as each 10⁰C increase in temperature doubles the rate of reaction (or Q10=2). This research challenged the accuracy of the recommended accelerated stability tests with regard to emulsions. To do this, 65 emulsions were made on the laboratory scale which altered by emulsifier type and concentration; oil phase ratio and work done during emulsification. These emulsions were tested according to the recommended protocols of accelerated testing given in the guidance documents and put on long-term ambient temperature test for direct comparison with accelerated results. Three new parameters were introduced to measure the accuracy, precision and predictive threshold of the accelerated tests. It was found that for the emulsions studied, four measurement parameters out of the five tested showed that the assumptions made for elevated testing were both inaccurate and imprecise for the prediction of long-term stability. Indeed, in three of these parameters: viscosity; appearance and colour; the predictive threshold did not extend beyond the extent of the accelerated testing time, 16 weeks, let alone up to the 96-week+ shelf-life of a cosmetic product. It was also demonstrated, however, that one parameter, pH, which is more aligned to the original Arrhenius studies had a good adherence to the accelerated testing extrapolation, showing a predictive threshold beyond the 96-week target for the formulations tested. This showed that the parameters of measure need to be more critically considered before being subjected to accelerated stability extrapolations.