Division 3.8.4 Consideration of systemic effects and calculation of the margin of safety

The safety assessment of a product for systemic toxicity is highly dependent on data on each substance, since there will be no data on systemic toxicity for the finished cosmetic product.

Risk characterisation usually involves an expert evaluation of the potential non-quantifiable adverse effects, followed by calculation of an uncertainty factor or margin of safety^f1. This calculation depends on the systemic exposure to the substance and its toxicological parameters.

In accordance with Point 8 part A of Annex I to Regulation (EC) No 1223/2009, systemic effects and margin of safety are to be considered in Part A of the safety report. As they are mandatory, the omission of these steps is to be duly justified. An example where this could apply would be the presence of a substance in the cosmetic product at a low level, with the expected (worst case) exposure levels being below the appropriate threshold of toxicological concern (TTC) values^f2. Another example could be the inclusion of food materials for which a much higher innocuous ingestion level is known.

When the requirement to calculate the margin of safety cannot be met, a different way of expressing the safe dose for each substance may be appropriate, where justified. When a NOAEL is not available, other reference toxicology values such as No Observed Effect Level (NOEL), LOAEL, Lowest Observed Effect Level (LOEL), can be used to calculate the margin of safety; Benchmark Dose (BMD) or Virtually Safe Dose (VSD), used to qualify and quantify a risk in other fields, may be used in the context of cosmetic products safety assessment, provided a relationship with exposure is established, by comparing the exposure from cosmetics and these reference doses.

Otherwise, the safety of a particular substance in a particular product cannot be demonstrated.

According to the procedures described in the SCCS Notes of Guidance^f3, the margin of safety (MoS) for a specific route of exposure can be calculated using the following formula:

where the Systemic Exposure Dose (SED) is obtained by combining the external exposure (mg/kg bw/day) with the absorption rate (typically expressed in % or μg/cm²), frequency and retention factors.

It is generally accepted that the margin of safety should be at least 100 to declare a substance safe for use in a finished product.

In the case of route-to-route extrapolation, the respective bioavailability via each route should ideally be taken into consideration. The assumption of 100 % oral bioavailability might overestimate the systemic exposure in a toxicity study via the oral route. Therefore, in the absence of data, it should be assumed that not more than 50 % of an orally administered dose is systemically available. If there is evidence to suggest poor oral bioavailability, for example if the substance is a poorly soluble particulate, it may be more appropriate to assume that only 10 % of the administered dose is systemically available^f4. Whenever oral absorption data are available, these should be included in the calculations.

The NOAEL chosen for calculating the margin of safety is taken from long-term repeated dose toxicity studies (sub-acute, sub-chronic, and/or chronic toxicity tests, carcinogenesis tests, teratogenesis tests, reproduction toxicity, etc.).

The value used will be the lowest NOAEL obtained by the most pertinent study with respect to the conditions of use of the substance, to species sensitivity, etc.

From the complete toxicological profile, a NOAEL should be determined for the systemic effects. In general, the lowest relevant NOAEL of the most relevant endpoint is selected for calculating the margin of safety.

The calculation of the margin of safety based only on Median Lethal Dose (LD50) data derived from single dose tests (instead of a NOAEL from at least sub-acute tests) cannot be used to justify safe use.

When the absence of bioavailability can be clearly demonstrated, the calculation of the margin of safety is not necessary. In these cases the possible local effects on skin or mucous membranes should still be considered.