Assessment of new and old probit functions for inhalation of acute toxic substances

Probit functions are used as part of a quantitative risk analysis (QRA) of hazardous events, such as acute emissions of hazardous substances. A probit function describes the relationship between the concentration of a substance, the duration of the exposure to the substance, and the part of the exposed population that suffers a certain effect, such as illness or death, from the exposure. The National Institute for Public Health and the Environment in the Netherlands (RIVM) has developed a new method for deriving probit functions. Changes of the probit functions due to the new method may have consequences when establishing risk contours, which can affect area plans and have economic consequences. In this assignment, the Norwegian Defence Research Establishment (FFI) has elucidated the basis for establishing the new probit functions. The new and old probit functions for the substances sulfur dioxide (SO2), ammonia (NH3), hydrogen fluoride (HF), chlorine (Cl2), carbon monoxide (CO), hydrogen chloride (HCl), hydrogen sulphide (H2S) and sulfuric acid (H2SO4) have been evaluated and compared. Probit curves for different exposure times have been calculated. FFI considers the basis for the establishment of the new probit functions to be expedient. The new method uses safety factors based on criteria related to adequacy of the database, nominal concentration and animal-to-human extrapolation. Comparison of the new and old probit functions showed that the new method had a relatively large impact for substances where the database was inadequate. FFI recognizes the need to deal with uncertainty in the database, but the use of safety factors on probit functions may not be the best way to deal with this uncertainty. The new method assumes a worst-case approach. This approach will provide a risk contour, which is very safe for the society, but very costly, and probably far from the realities. FFI suggests several approaches, each of which can contribute to more expectation-orientated risk contours. I) A Monte Carlo analysis of all uncertainties related to the establishment of risk contours to point out parameters that contribute the most for the total uncertainty. II) A cost-benefit analysis of improving the database for probit functions. III) Comparing the system with risk contours based on probit functions with the system in the US, which uses acute exposure guideline levels (AEGL).