Integration of Industrial Hygiene Statistical Tools into TSCA Risk Evaluations

1477 

Professional Poster 

N Jacobs¹

¹Stantec, Washington, DC

Neva Jacobs  
Stantec

In recent years, EPA has released several risk evaluations under the Toxic Substances Control Act (TSCA) covering both consumer and worker exposures and associated risks. EPA has used various sources of exposure data and modeling approaches to characterize these exposures. In some instances, EPA has selected point values from the exposure data to describe exposures rather than using statistical tools to characterize the exposure profile. As a demonstration of the potential application of such statistical tools in enhancing EPA Risk Evaluations for prospective exposure assessments, we present a case study using Expostats with the available consumer and occupational data on exposures associated with vermiculite attic insulation. Vermiculite attic insulation that was historically sourced from the Libby, Montana, vermiculite mine may contain asbestiform amphibole minerals. Homeowners and contractors may continue to be exposed to these amphiboles during disturbance of existing attic insulation. EPA has previously evaluated exposures associated with vermiculite attic insulation under TSCA but relied upon selected exposure values from a single published study. We supplement this exposure dataset with additional published exposure data and perform statistical analysis on the full dataset to update cancer risk estimates and calculate the probability that exposures would exceed the OSHA PEL of 0.1 f/cc and that cumulative exposure would exceed a working lifetime at the OSHA PEL.

EPA evaluated occupational and consumer exposures associated with disturbance of asbestiform amphibole-containing vermiculite attic insulation under TSCA. Based on their selection of exposure point estimates from one study for the consumer scenario, EPA concluded that cancer risk estimates exceeded their benchmark of 1E-06. For occupational exposure, EPA grouped vermiculite attic insulation with other building materials and did not rely upon any vermiculite attic insulation-specific exposure data for their risk evaluation of this occupational exposure scenario (OES). Nevertheless, EPA concluded that the risk estimate for high-end occupational exposures exceeded their benchmark of 1E-04 but central tendency estimates did not. We sought to: 1) supplement the exposure data for vermiculite attic insulation relied upon by EPA with additional identified studies, 2) update risk estimates based on the additional exposure data following the EPA approach, and 3) characterize this larger dataset by estimating the probability of exceeding the OSHA 8-hour TWA PEL and the cumulative exposure associated with a working lifetime at the OSHA PEL.

We reviewed the EPA Risk Evaluation and conducted a literature search to identify additional exposure data relevant to the exposure scenarios of interest, namely consumer (i.e., homeowner) and worker (i.e., contractor) exposures associated with asbestiform amphibole-containing vermiculite attic insulation. Using EPA's risk model, we calculated lifetime excess cancer risks using central tendency (50th percentile) and high-end (95th percentile) concentrations from the assembled exposure database based on PCME values.

Using the traditional OEHS approach, we extracted the relevant exposure data from these studies and calculated 8-hour TWAs for the airborne asbestos concentrations as determined by PCM and PCME. Because the majority of amphibole minerals within Libby vermiculite are not regulated as asbestos by OSHA (i.e., winchite, richterite), we also calculated exposure concentrations for only the OSHA-regulated amphiboles (i.e., tremolite asbestos), which has been estimated to represent six percent of total Libby amphiboles. Measurements noted to be below the level of detection (LOD) were conservatively assumed to be at the LOD. Expostats was used to estimate the probability that the 8-hour TWA measurements would reach the OSHA PEL.

Cumulative exposure was estimated based on exposure factors from the vermiculite attic insulation disturbance scenario within the EPA Risk Evaluation, again assuming that six percent of the fibers were regulated asbestos. For homeowners, typical cumulative exposure was calculated assuming exposure to the 50th percentile of the available exposure measurements for one hour per day, once per year. High-end exposure was calculated assuming exposure to the 95th percentile of the available data for three hours per day, once per year. For contractors, typical cumulative exposure was estimated assuming exposure to the 50th percentile of the available data for one hour a day, 50 days per year. High-end cumulative exposure for this group was estimated assuming exposure to the 95th percentile of the available data for three hours a day, 50 days per year. The number of exposure years needed to exceed the cumulative exposure associated with a working lifetime at the OSHA PEL of 0.1 f/cc (i.e., 4.5 f/cc-year) was calculated in R. Expostats was then used to estimate the probability that these estimates would exceed 4.5 f/cc-year.

For homeowners, we calculated that 50th and 95th percentiles of exposure concentrations (PCME) were 0.25 and 5.63 f/cc, respectively. For comparison, EPA relied upon a point estimate of 1.16 f/cc for both scenarios. Using EPA's approach for lifetime cancer risk estimates and these updated exposure values, the central tendency risk estimate would decrease while the high-end risk estimate would increase compared to EPA's original calculations; both would still be above EPA's risk benchmark of 1E-06. For contractors, 50th and 95th percentile exposure concentrations (PCME) were 0.20 and 2.65 f/cc, respectively. The excess lifetime cancer risks for the contractor scenario using the EPA's risk model were calculated to exceed EPA's benchmark of 1E-04 for high-end (9.32E-04) but not for central tendency exposures (7.12E-05).

Of the 25 calculated 8-hour TWA values (PCM), six (24%) exceeded the current OSHA PEL of 0.1 f/cc. However, when considering that only approximately 6% of fibers are regulated amphiboles, no samples exceeded the current OSHA PEL as PCME values. It was estimated that, based on PCME, the probability was 26.4% that the true value of both the arithmetic mean and 95th percentile were ≤10% of the current OSHA PEL. Although OELs would not apply to a homeowner scenario, using the exposure factors from the EPA Risk Evaluation for disturbance of vermiculite attic insulation and the 6% PCME values, cumulative exposure for a homeowner (62 years) were estimated to be 0.00046 and 0.023 f/cc-year, respectively, for the central tendency and high-end estimate. It was estimated there was a >99.9% probability that the true value of the 95th percentile of cumulative exposure to asbestos was ≤10% of the cumulative exposure for a working lifetime at the OSHA PEL (i.e., ≤4.5 f/cc-years). For contractors, based on 6% PCME values, central tendency and high-end cumulative exposures after 45 years were calculated to be 0.017 and 0.83 f/cc-year, respectively. The cumulative exposure values for contractors exhibited more variability than for homeowners and it was estimated that the probability that the true value of the 95th percentile was ≤10% of a working lifetime at the OSHA PEL was 16.8%.

This research enhances approaches used in the EPA TSCA Risk Evaluation by applying an OEHS statistical approach to characterize the exposure profiles and probability of exceeding exposure benchmarks. We present the distribution of possible exposures and demonstrate the value of using statistical tools in prospective exposure scenarios to characterize uncertainty. Notably, exposure values relied upon herein assume no PPE use. This analysis highlights challenges in different regulatory definitions as most asbestiform amphiboles in Libby vermiculite are not regulated by OSHA. Notably, while we relied upon EPA's risk model for purposes of comparison, researchers have also published cumulative exposure benchmarks for Libby amphiboles based on epidemiological data.

Exposure Assessment

Health Regulations

Risk Assessment