Lower the residential soil lead preliminary remediation goal (PRG) from 80 mg/kg to 55 mg/kg

California’s Office of Environmental Health Hazard Assessment established a health-based benchmark of a 1.0 µg/dL incremental increase in children’s blood lead levels (BLLs), associated with an average loss of one IQ point, as the basis for establishing protective measures related to lead in soil [13]. This benchmark was used by California’s Department of Toxic Substances Control’s (DTSC) “LeadSpread9” model to create a corresponding Preliminary Remediation Goal (PRG) for lead in soil of 80 mg/kg [14].

DTSC’s LeadSpread 9 is the model used to estimate the residential lead soil screening level. The U.S. EPA’s Adult Lead Model (ALM) used the DTSC’s LeadSpread 9 model to estimate the blood lead concentration in a fetus of an adult worker exposed to lead-contaminated soil. This is the concentration that would correspond to an estimated increase in blood lead in a 90th percentile child of 1 µg/dL. The model reflects four exposure pathways: (1) incidental ingestion of outdoor soil and, (2) incidental dust ingestion from indoor dust from soil tracked into a home, (3) dermal uptake from contact with outdoor soil or indoor dust, and (4) inhalation of resuspended particles. In the current model, ingestion is the dominant pathway; there is little contribution from inhalation of resuspended soil and dust and dermal uptake of lead. Thus, the choice of exposure factors for dust ingestion has a large influence on the resulting BLL calculation. DTSC uses a soil+dust ingestion rate of 80 ug/day for children, based on the central tendency from the EPA Exposure Factors Handbook.

Issue 1: misalignment with California’s health-protective benchmark

DTSC acknowledges that they use the toxicity criterion from OEHHA of 1.0 ug/dL rise in children’s blood lead levels as the basis for their approach at setting soil limits for lead:

“The toxicity criterion on which LeadSpread 9 is based is CalEPA’s Office of Environmental Health Hazard Assessment’s (OEHHA) toxicity evaluation of lead with a source-specific “benchmark change” of 1 µg/dL which is the estimated incremental increase in children’s blood lead that would reduce IQ by up to 1 point.”

However, in their calculations using LeadSpread9, DTSC acknowledges that 70 mg/kg is the soil PRG that is estimated using their tool:

“Using the previous version of LeadSpread, LeadSpread 8, a Preliminary Remedial Goal of 77 mg/kg soil lead was estimated. A value of 70 mg/kg soil lead is estimated using LeadSpread 9.”

Despite this, DTSC attempts to explain that 80 mg/kg should be used, even though it is actually based on a rise in BLL of 1.14 μg/dL, not the 1.0 μg/dL benchmark from OEHHA:

“For most sites without special circumstances, such as markedly elevated soil lead bioavailability, the difference in predicted incremental blood lead and IQ change for exposures to soil lead between 70 mg/kg and 77 mg/kg is within the LeadSpread model uncertainty and does not exceed the de minimis level of 1 IQ point identified by OEHHA. The current DTSC residential lead (Pb) soil screening level is 80 mg/kg, based on an estimated increase in blood Pb in a 90thpercentile child of 1 µg/dL. At 80 mg/kg soil lead, LeadSpread 9, estimates the increase in blood Pb in a 90thpercentile child as 1.14 µg/dL which, in turn, is associated with an upper-bound estimate of a loss of 1 IQ point. The change is not discernable at one significant figure. Results of IQ tests are reported as an integer. Fractional IQ points are not measured. The blood lead level of 1.14 would have to rise to 1.5 (which would round up to 2.0) to be considered a significant increase. Therefore, HERO recommends that the remedial/mitigation level for residential soil exposure remain at the current residential default value of 80 mg/kg. Future development of better-defined childhood exposure parameters may change this recommendation.”

We disagree that the rounding is inconsequential and there is not a strong basis for DTSC to depart from the OEHHA benchmark of 1.0 μg/dL. Using OEHHA’s benchmark of 1.0 μg/dL, the soil lead PRG should be 70 mg/kg, with no other calculations changed.

Issue #2: uncertainty with EPA’s exposure factors used by DTSC

DTSC’s LeadSpread9 model uses several exposure factors to estimate the amount of lead in soil that would lead to the corresponding BLL. Here, we demonstrate the high level of uncertainty for one exposure factor, “ingestion constant”, as an example of how assumptions for these factors can lead to significantly different PRGs.

DTSC uses an ingestion contact of 0.16 (µg/dL)/(µg/day) in LeadSpread9. This exposure factor is ratio of blood level to lead that enters the body through the ingestion pathway, essentially capturing the fraction of lead that contributes to a rise in blood lead. There are several issues with this ingestion constant: it is from an old study conducted in the early 1980s [15], it is not technically scientifically accurate because it uses liquid ingestion and extrapolates to soil ingestion, and it has a small sample size, all of which indicate high uncertainty in this ingestion constant.

The estimate for the ingestion contact originates from a 1983 study of 29 breast-fed and formula-fed infants [15]. The authors measured the amount of lead the infants consumed in milk and the corresponding increase in BLLs. In addition to the small sample size and the age of the study, a critical issue is the assumption that this constant, derived from the ingestion of formula and breast milk, also applies to the ingestion of soil and dust. The original study included a high- and a low-exposure group, which were combined to calculate the 0.16 (µg/dL)/(µg/day) constant. However, if the data from each exposure group and age category are used separately, this value could be different. For example, among infants aged 112–195 days who remained in the study, the ingestion contact was approximately 0.2 or 0.4. If the LeadSpread9 model is used with this constant adjusted to either 0.2 or 0.4, the resulting PRG changes from 70 to 56 mg/kg and 28 mg/kg, respectively.

Our main goal of mentioning this study is less about opening a debate about which constant to use from this study in 1983; rather, we are using it to show how sensitive the model is to exposure assumptions.

Issue #3: children are more susceptible to lead than the current model accounts for

The substantial advances in developmental science since the original model was developed suggest that LeadSpread 9 model likely underestimates the effects of lead under current exposure estimates, both in the adult models that account for sensitivity to the fetus and the model for childhood, which accounts for lead only as a single exposure (which in reality, never occurs in young children). They also fail to account for differential sensitivity to lead across childhood and assumes the effects on children ages 1–6 years of age, despite these ages having very different behaviors and sensitivities to exposure [16].

Both adult models (residential and industrial) account for fetal sensitivity to lead use the same standard that is used in infants, of protecting the fetus carried by an exposed adult to prevent an increase in blood lead of the fetus of >1 µg/dL. This logic is in direct opposition to the abundance of studies showing that fetal sensitivity of both the brain and other rapidly developing biological systems is greater than the sensitivity of infants on whom the original model was based [17]. The half-life of lead is also increased during pregnancy, leading to longer exposure times for the same dose of exposure [18]. Given this differential sensitivity, it is nearly a guarantee this same blood lead level increase in a fetus would have greater adverse effects than the same level rise in an infant of older child. Additionally, blood lead levels in the fetus have been identified simultaneously as higher than in the mother [19].

The risk of compounding environmental exposures, which are now felt to be critical in understanding and calculating risks to fetal and early childhood development, are not accounted for at all in the original model. For example, blood lead levels can increase more rapidly in children with iron deficiency anemia, a condition that is likely to worsen when children are displaced and have less consistent access to mitigating factors like iron-rich foods [20]. Science increasingly demonstrates that co-exposure to psychosocial stress, which is significantly increased following wildfire events, and lead increase the harmful effects of each of these exposures alone [21]. Because their brains and biological systems are so rapidly developing, children are especially sensitive to both direct psychosocial stress as well as the indirect stress experienced by their caregivers.

Furthermore, the model’s scope is confined to acute toxicity and does not adequately account for the systemic health consequences of chronic lead exposure. Substantial scientific evidence has demonstrated that lead negatively impacts multiple organ systems beyond the nervous system, including the cardiovascular system [22], renal function [23], endocrine signaling [24], and the immune system [25]. By neglecting these broader health effects, the current model fails to provide a comprehensive risk assessment, ultimately undermining efforts to fully protect vulnerable populations from the multifaceted harms of lead exposure.

Finally, both lead ingestion and lead absorption change throughout the first two years of life. Assuming that the model holds for children ages 1–6 is especially problematic given how different children are even within that range. For children consistently exposed to lead, blood levels increase rapidly between 6 and 12 months of age, are highest form 18 months to 36 months of age, and then decrease gradually [7]. After that time, dust ingestion from ‘hand to mouth’ behaviors become the dominant exposure route as children become more mobile.

Recommendation: adjust soil screening level from 80 mg/kg to 55 mg/kg

To account for DTSC’s misalignment with toxicity criterion from OEHHA, high levels of uncertainty in DTSC’s LeadSpread9 model, and to capture the current state of science with regard to the impacts of lead on children’s health, we recommend using a PRG of 55 mg/kg. The basis for this updated PRG is the following:

Re-aligning with OEHHA’s toxicity criterion and using OEHHA’s benchmark of 1.0 µg/dL BLL increase as the de minimis threshold.

Accounting for parameter uncertainty and updated science on kids health and using the high-end central tendency for children’s soil ingestion of 100 mg/day from the EPA Exposure Factors Handbook, rather than the 80 mg/day used by DTSC.

When using these parameters, without changing any other default values or calculations used by DTSC in LeadSpread9, the resulting PRG is 56 mg/kg, and we use 55 mg/kg for simplicity.

Other key considerations:

We note the following:

A PRG of 55 mg/kg is a soil remediation level below which no further action is needed on the site for full use; levels above this should be remediated.

This PRG does not account for other potential exposures in a residence. If there are other sources, a site-specific risk assessment may be warranted, as noted by DTSC:

“Because the lead benchmark dose is an incremental change in blood lead, background exposures to lead, and media other than soil, or dust from the site which may be impacted by lead are not considered in the worksheet. If lead is present in media other than soil (e.g., water, air) or if the home grown produce pathway is anticipated at the site, please contact the HERO toxicologist assigned to the site.”

The exposure assumptions are for exposed soil; if there is ground cover, or if there is fresh topsoil, exposure will be lower.

In addition to remediating soils, we recommend the following individual actions that can help reduce exposure:

Wash hands frequently, especially before eating

Remove shoes when entering a residence

Clean the paws of pets before entering a residence

Keep indoor surfaces clean

Damp wipe dirty surfaces, especially playroom floors, carpets, and foam or rubber mats, where children play on them more often

Use a vacuum with a HEPA filter