Estimation of Incremental Unit Risks for Inhaled Nickel Subsulfide and Nickel Oxide, Based on a Combination of Human and Animal Data

ABSTRACT

Both animal and human exposure–response data are used to estimate the incremental unit risks (IURs) of lung cancer for Ni₃S₂ and NiO, which are constituent compounds of nickel refinery dust. The animal experiments are used to determine relative lung cancer potencies for Ni₃S₂ and NiO, and the human epidemiological data are used as the best estimate of overall risk for refinery dust exposure. The animal data for Ni₃S₂ are fit with a linear model, while the nonlinear animal data for NiO are fit with a Weibull model. The lower 95% confidence limit at a 5% point of departure is used to calculate a tumorigenic potency ratio of Ni₃S₂ to NiO of 5.6. Analyses of actual nickel refinery dust indicate the weight% of Ni₃S₂ and NiO to be 82% and 9%, respectively. This information is used with the previously determined IUR for nickel refinery dust to calculate IUR_NiO = 5.1 × 10⁻⁵ (μg Ni/m³)⁻¹ and IUR_Ni3S2 = 2.9 × 10⁻⁴ (μg Ni/m³)⁻¹.

Key Words:

• nickel subsulfide
• nickel oxide
• inhalation
• carcinogenic unit risk

ACKNOWLEDGMENTS

The authors thank Vale Inco for providing funding for writing this article.

Notes

¹Number of tumor bearing animals/number of animals.

Figure 3 US National Toxicology Program's rat dose–response data for the two-year bioassay of Ni₃S₂. Vertical bars show 95% confidence limits for observed tumor response (pooled across genders). Dashed line shows fitted quantal-linear dose–response function across all doses. Dotted line shows fitted (upper bound) quantal-linear dose–response excluding highest dose.

Figure 4 US National Toxicology Program's rat dose–response data for the two-year bioassay of NiO. Vertical bars show 95% confidence limits for observed tumor response (pooled across genders). Curved dashed line shows fitted Weibull dose–response function. Straight dashed line shows linear extrapolation from USEPA default point of departure (lower 95% confidence limit on dose associated with 10% excess risk) to the model-estimated background risk (between observed risks at 0 and 500 μg Ni/m³). Dots show points of departure ranging from 0.01% to 10% (Table 1). Solid straight line shows linear extrapolation to the estimated background risk for the 5% point of departure.

²Model has many parameters as data points; goodness of fit statistic could not be calculated.

¹Shaded row reflects point of departure selected for potency determination (see text).

²Benchmark dose.

³Point of departure (lower 95% confidence limit on BMD).