Relevance of non-guideline studies for risk assessment: the coverage model based on most frequent targets in repeated dose toxicity studies.

A common challenge for human risk assessment is the quality of the available animal studies. Non-guideline studies are often limited for different aspects of study design and documentation. Within this publication the relevance of a limited scope of examination is discussed. Preliminary analyses of the RepDose database have shown that liver, body weight, kidney and clinical symptoms are frequently affected in oral repeated dose toxicity in rats and mice (Bitsch et al., 2006), while many other targets are seldom affected. As most of these targets are investigated frequently also in non-guideline studies, it is likely that they provide a reliable NOEL, although the full spectrum of endpoints has not been covered. Based on RepDose data we investigate the relevance of individual targets for determining the LOEL and the consequences for risk assessment. The resulting coverage model for subchronic oral rat studies includes up to six targets and an additional assessment factor for LOEL extrapolation. It can be applied to assess the reliability of non-guideline studies with respect to the scope of examination. Furthermore the application of the coverage model to other databases will increase and/or specify the chemical domain and reveal respective targets as well as effects.

The relation between extrapolated risk, expressed as potentially affected fraction, and community effects, expressed as pollution-induced community tolerance.

The results of toxicity tests can be used to calculate the potentially affected fraction (PAF) of species in an ecosystem at a given pollutant concentration using statistical extrapolation methods. The PAF curve indicates the fraction of species from the original community that may become inhibited at each elevated pollutant concentration and is a measure of the ecotoxicological risk. Pollution-induced community tolerance (PICT) is a true community response that is measured under controlled conditions in the laboratory, using organisms from contaminated field sites. Microorganisms from experimental field plots with added Zn were exposed to various concentrations of Zn in the laboratory and the mineralization of 14C acetate was monitored. Microorganisms from plots with Zn concentrations above 124 mg/kg showed a significant increase in the effect concentration 10% (EC10) and, therefore, had a significant PICT. The pore-water concentrations of Zn in these field soils were in the same magnitude as the EC10 of the microorganisms from these soils. The PAF curve was calculated from previously reported toxicity tests with five different microbial species using the average and the standard deviation of the logarithmically transformed EC10 values. The average sensitivity of this PAF curve was similar to the EC50 of the acetate mineralization curve from the field plot without added Zn2+, but the PAF curve was less steep. Our experiments indicated that 27 to 84% of the original microbial species were inhibited at Zn concentrations from 334 to 1,858 mg/kg soil, respectively. Our results suggest that the PICT method can now also be used to quantify the fraction of the original species composition that is inhibited at a specific pollutant concentration.

Congener-specific model for polychlorinated biphenyl effects on otter (Lutra lutra) and associated sediment quality criteria.

A model for risk assessment was built for simultaneous, congener-specific PCB bioaccumulation from sediment to fish to otters (Lutra lutra). Toxic equivalence factors (TEFs) were used to sum individual congeners in otters to a toxic equivalent concentration (TEQ) relative to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Microbial dechlorination and burial in sediments and bioaccumulation are modeled to predict time trends of PCB concentrations in otters. Effects of accumulated PCBs on otters were assessed for vitamin A metabolism and reproduction, taking model uncertainty into account. Three locations in The Netherlands were modeled with PCB levels in sediment of 1 to 171 pg TEQ/g organic carbon (OC). Almost 100% reduction in litter size was predicted for the most polluted area in 1996. Due to large associated uncertainty, a period of 25 to 80 years may be needed for recovery of otter vitamin A levels and litter size at this site. Calculated median sediment quality criteria (SQC) range between 1 and 12 pg TEQ/g OC, depending on the chosen effect criterion. Uncertainty in calculated effects and SQCs is substantial and is mainly caused by uncertainty in PCB congener 126 accumulation.

Uncertainty of the hazardous concentration and fraction affected for normal species sensitivity distributions.

Species in the environment vary according to their sensitivity to a toxicant. Because these differences in sensitivity are unique to the toxicant at consideration and laboratory data sets to assess this variability are very small due to cost, it is important to provide uncertainty estimates of (1) environmental quality objectives (hazardous concentrations) derived from these laboratory data and (2) fraction of species affected at given, or predicted, laboratory or environmental concentrations. This article focuses on the normal (Gaussian) distribution of species sensitivity. It examines and compares results of Problems (1) and (2) from two opposing statistical philosophies, Bayesian and Classical, leading to vastly different numerical approaches. For the normal model, both approaches lead to identical answers, numerically. Extrapolation factors for the lower, median, and upper estimates of the hazardous concentration at six levels of protection are derived. Furthermore, upper, median, and lower estimates of the fraction affected at given, standardized, logarithmic concentrations have been tabulated. This table can be used directly for risk assessment without reference to protection levels or hazardous concentrations. The confidence limits for hazardous concentration and fraction affected depend heavily on the number of species tested and are independent of the toxic substance involved (provided the model is right), due to correction for the mean and standard deviation of the toxicity data. The equivalence of confidence limits for hazardous concentration and fraction affected is captured in the law of extrapolation: the upper (median, lower) confidence limit for the fraction affected at the lower (median, upper) confidence limit of the hazardous concentration is equal to the fraction affected (e.g., 5%) used to define the hazardous concentration. The upper confidence limit for the fraction affected at the median estimate of the hazardous concentration for 5% of the species is a fixed number depending on the sample size of the toxicity data only. It amounts to 46% at n=3, down to 20% at n=10, and still 12% at n 30.

Confidence limits for hazardous concentrations based on logistically distributed NOEC toxicity data.

This paper deals with the calculation of Hazardous Concentrations of toxic substances from small sets of laboratory toxicity data, e.g., NOECs. A procedure due to Van Straalen and Denneman, as adapted from Kooijman (case n = 1), in which one seeks a concentration that protects 95% of the biological species is modified to account for the uncertainty in the estimates. New constants are obtained by simulation. These allow the calculation of the one-sided 95% left confidence limit of the Hazardous Concentration, from the mean and standard deviation of a sample of (laboratory) toxicity data. This 95% confidence limit is always lower than the 95% certainty value calculated with the Kooijman (n = 1)/Van Straalen method. The authors also derive constants to calculate a one-sided 50% confidence value, that overpredicts as often as it underpredicts. This value may be used as a median guess of the Hazardous Concentration. It will always be higher than the 95% certainty value of the Kooijman (n = 1)/Van Straalen method. However, by using the 50% value, one runs the risk of protecting substantially less than 95% of the biological species.

The application of QSARs, extrapolation and equilibrium partitioning in aquatic effects assessment for narcotic pollutants.

QSAR estimates of toxicity of relatively unreactive organic chemicals for 19 species of bacteria, algae, fungi, protozoans, coelenterates, rotifers, molluscs, crustaceans, insects, fish and amphibians were used to predict 'no-effect-levels' (NELs) at the ecosystem level by means of a recently developed extrapolation model. Equilibrium-partitioning theory for sediment and water was used to derive NELs for aquatic sediments. A simple table is given from which NELs for water and sediments can be predicted on the basis of only the octanol-water partition coefficient and molecular weight.