Issue: Appropriateness of Relocation Criteria
Public health and regulatory agencies typically recommend that continuous human exposures to toxic substances not exceed a dose that is tenfold to a hundredfold lower than a dose that causes no observable adverse effects (NOAEL) in study populations. EPA has formalized this process by including, when available, chemical-specific values, reference doses (RfD), in its toxicity database used by regulatory programs. A subchronic RfD for methyl parathion has been established by EPA. Although RfDs are generally considered to be screening levels and are not regulatory standards, they are often used to establish environmental cleanup levels.
The proposed urine levels based on other health criteria are about thirtyfold higher than EPA’s subchronic reference dose (RfD), threefold higher than some animal no-observed adverse effect levels (NOAELs), and about threefold lower than the lowest observed adverse effect levels (LOAELs) found in recent subchronic animal studies. In consideration of the chronic toxicity and degradation of methyl parathion and the likely exposure time and pattern for affected persons, are the proposed urine PNP criteria sufficiently protective for all persons? If not, what criteria would the panel recommend?
The proposed relocation criteria rely on
- environmental sampling to classify homes as contaminated or noncontaminated;
- biomonitoring of occupants of contaminated homes for urinary PNP as a biomarker of MP exposure; and
- relocation of an entire household if the urinary PNP of any single member exceeds the risk-group-specific urinary PNP action level.
The work group concurs with the steering committee that urinary PNP is an appropriate biomarker of low-level MP exposure. The work group also concurs with the steering committee that reference doses (RfDs) are screening values and that a weight-of-evidence approach is reasonable as applied to specific exposure situations. The work group concurs with the approach to base relocation criteria, as a minimum, on risk-group-specific urinary PNP biomonitoring.
The work group recognizes that where to set the urinary PNP action levels is a risk management decision faced by the involved governmental agencies. However, we have identified critical gaps that preclude our ability to endorse or refute specific action levels on a scientifically substantiated basis. We strongly feel that resources should be devoted to generating the data necessary to fill these gaps as quickly as possible. It is critical to conduct a dermal absorption study correlating MP dermal dose with urine PNP excretion pattern. Although such a dermal absorption study in human volunteers would be ideal, valuable data can also be obtained from the use of relevant animal models such as pigs, monkeys, or hairless guinea pigs. Such a study would provide the “link” between urine PNP and dose. In addition, a subchronic toxicity study (in rats) designed to determine a dermal Rfd would provide valuable data for risk assessment.
It must be realized that oral to dermal route extrapolation reflects two differences:
- slower dermal absorption and
- different first-pass metabolism.
Dermal absorption is nonlinear and saturates at high doses. Another difference is that oral subchronic studies are 90 days continuous exposure. The rodent no-observed-effect level (NOEL) presently used to establish RfD is based on a continuous oral exposure to MP. On the basis of data presented, dermal absorption is probably four times less than oral and thus, a dermal NOEL should also be lower. This assumption needs to be validated by a rodent dermal subchronic toxicity trial. In the human MP exposures, limited urine PNP monitoring results provided to the work group demonstrated fluctuating levels, which suggests intermittent exposures to household MP. Because the rat subchronic toxicity trials were continuous gavage feeding and the human exposures may be intermittent, the safety factor related to exposure may be conservative. This assumption also needs to be validated by better PNP monitoring data.